Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. 'l'he life h · lstory strategy of C arex pumila Thunb. (Cyperaceae), a rhizomatous perennial pioneer o species R. E. Burgess 19t34 \ ' I alns of the n the sand pl . dune system of coastal M anawatu A thesis presented . ln partial fulfilment Ulrements of the req . egree of Doctor of for the Philosoph in Botany at Ha . /sey � 1versity I . ,. . . .. .. ABSTRACT The life history strategy of Carex pumila Thunb . ( Cyperaceae ) � a major colonist of raw moist sand on the sand plains of coastal Manawatu , New Zealand , is outlined. By virtue of the continuous formation of sand plains � sites suitable for colonization are a permanent feature of this habitat and vegetation of increasing seral maturity is represented at any one time across a series of adjoining deflation hollows and low dunes . It is proposed that the species is an r-strategist well suited to exposure � nutrient stress and seasonal flooding . Amelioration of these conditions by deliberate perturbation treatments resulted in this seral species responding in a way that ultimately lead to its more rapid demise . The species has a rhizomatoas perennial growth habi t . The modular construction of its rhi zome system is described for the first time . Similarly , the occurrence of both long and short sympodial rhizome branches and of large-diameter sinker roots have not been previously des cribed in the literature on this species . Its floral development appears to be environmentally cued. Emergence of inflorescences occurrs in early October. Maximum size of dissemules is obtained by early January . Subsequently seeds are shed and the shoots bearing them die . The species is essentially allogamous� although in a laboratory experiment � it was found to be partially self-compatible . Self-pollination must be expected in the field since neighbouring shoots are likely to be part of the same genet . Field studies are reported in which the performance o f Carex pumila was monitored , firstly at sites of increasing seral maturity both in space and in time , and secondly in response to perturbation treatments . Populations showed a pattern of development that included a juvenile phase of rhizome expansion, an adolescent phase of increasing shoot density, a mature phase in which a proportion of the shoots were reproductive ; and a senile phase of diminished growth and seed production. Phasic development was more protracted on the more stressed and more exposed si tes . Other species more rapidly filled the space made available by the death and decay of Carex pumila shoots , than the colonis t itself . As a pioneer , the species is doomed to extinction on the sites it colonizes . In a perturbation experiment , the sward mass of the total vegetation per unit area was increased at all si tes by nitrogen ferti lizer, applied as ammonium ions at a rate of 50 kg N / ha . Where the Carex pumila population was in a senile phase in an old deflation hol low , the increase was made mainly by other species . In younger populations on a low dune , the density of shoots and expanding buds of Carex pumila were markedly increased by the fertilizer treatment . Associated with this , a significant increase occurred in the proportion of the total dry weight of vegetative branches in rhizomes and in green leaves . A nitrogen limitation to seed yeild was indicated at the older low dune site . Here nitrogen fertilizer addition increased seed output per uni t area by inc reasing bo th seed number per culm and seed size . By contrast on the younger low dune site , seed output per uni t area was unchanged by the perturbation. In this population , real location of resources within fertile shoots , which was seen as an increased number of seeds pe r culm , was offset however by a reduction in fertile shoot density . Seed reproductive effort varied between 0 and 1 6% of total biomass � whereas rhizome allocation was more variable ; up to 1 00% of biomass where the species was invading an embryonic deflation hollow. As a proportion of the biomass of fertile shoots alone ; seed reproductive effort estimates of up to 32% were obtained . The post-anthesis photosynthetic contribution of female spikes to final seed weight was estimated at 26%, in a growth room experiment . This estimate is considered conservative given that final seed weight was not significant ly reduced by defoliation and shading of the culm . Thus , the allocation of biomass to seeds cannot be considered a drain on the carbon resources of the plant that might otherwise be allocated to growth or some other plant function. Total nitrogen concentrations were dissimilar in different plant parts and , for comparable organs , between populations of different ages . Thus , allocation patterns to component parts based on dry weight and total nitrogen were different . Given that nitrogen was seen to be limiting growth in this seral habi tat � the allocation of this resource is likely to be of greater significance in the evolution of life history strategies than is that of dry weight . ACKNOWLEDGEMENTS I record with appreciation the assistance and encouragement given to me by the many people who made the completion of this thesis possible. I especially thank Dr J Skipworth for his supervision, Mr G Arnold for statistical guidance , Dr R R Brooks and Dr R Haselmore for guidance in analytical procedures , Mr B Campbell , Ms L Rhodes , Mr K Kelliher and Mr A Tucker for technical assistance , Mr I Gray and Dr P Gregg for plant and soil analyses , and Ms J Tipoki for typing. I am indebted to Dr W A Laing who made critical comments on drafts of this thesis . I am also grateful to the N . z. Forest Service for allowing me to undertake the field work on the Tangimoana State Forest , to Plant Physiology Division, D. S . I . R . for use of the C limate Laboratory , and to Mr and Mrs H Ellison for rainfall data and access to the field study area . I also record my appreciation to Massey University , Community Volunteers , Palmerston North ( Inc ) and Grasslands Division , D. S . I . R . who employed me during this period . Finally I thank my family for their fo rebearance . TABLE OF CONTENTS ACKNOWLEDGEMENTS TABLE OF CONTENT S LI ST OF FIGURES LI ST OF TABLES CHAPTER 1 . 1 1 . 2 1 . 3 1 . 4 1 . 5 1 . 6 I NTRODUCTION Life histories , allocation of resources and reproductive effort Successional trends of reproductive effort Quantification of life history strategies Sand plains and water-logged soils Carex pumila Thunb . Aims of the study CHAPTER 2 A SPECTS OF THE BIOLOGY OF CAREX PUMILA 2 . 1 2 . 2 2 . 3 2 . 4 2 . 5 2 . 6 Growth habit Germination Seedling morphology Shoot phenology Self incompatibility The contribution of reproductive assimilation to seed page V vi viii xiii 2 7 1 3 21 26 29 33 38 39 39 43 production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 CHAPTER 3 INITIAL FIELD STUDIE S 3 . 1 Des cription of the habitat The study area Climate Wind speeds 63 63 65 67 Vegetation cover 69 3 . 1 • 4 3 . 1 . 5 3 - 1 • 6 3 . 1 . 7 3 - 1 . 8 The water table . . . . . . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Soi l water profile 77 Soi l nutrient and organic matter status • • • • • • • • • • • 79 Nitrogen fixation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 Carex pumila on the sand plains Phasic development 89 89 Rates of clonal spread • • • • • • • • • • • • • • • • • • • • • • • • • • • • 95 Leaf litter decomposition • • • • • • • • • • • • • • • • • • • • • • • • 1 02 CHAPTER 4 FIELD PERTURBATIONS Introduction . . . . . . . . . . . . . . . . . . 4 . 2 Methods and materials . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 . 3 Results and dis cussion 1 1 0 1 1 6 4 . 3 . 1 Aerial shoot densities • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 1 26 4 . 3 . 2 Leaf area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 38 4 . 3 . 3 Age and size distributions • • • • • • • . • • • • • • • • • • • • • • • • • 1 39 4 . 3 . 4 Dry weight , energy , elemental contents • • • • • • • • • • • • • 1 43 4 . 3 . 5 Flowering and seed production 4 . 3 . 6 Allocation of dry mat ter and total nitrogen 1 85 1 99 4 . 3 . 7 The effect of nitrogen fertilizer addition • • • • • • • • • 225 CHAPTER 5 DI SCU SSION • • . . • . • • . • • • • . • • • . • • . • . . . • . . • . • . • . . • . . • . • . • 254 REFERENCE S • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 28 1 _, LI ST OF FIGURES 2 . 1 The rhizome axis of Carex pumila illustrating morphology. 2 .2 Vertical growth of a primary rhi zome axis . 2 . 3 Rhizome looping down into the substrata . 2 . 4 Short branch modules . 2 . 5 Basic linear pattern of rhizome architecture . 2 . 6 Short branch module (arrowed) abnormally elongated . its basic 2 . 7 Mean number and length of internodes on long and short rhizome modules . 2 . 8 Adventitious roots : ( a ) sinker roots and (b ) fine roots . 2 . 9 Seed and seedling morphology. 2 . 1 0 Male spike with anthers exserted . 2 . 1 1 Stages in flowering and seed production. 2 . 1 2 Defoliation and shading treatments . 2 . 1 3 Distribution of the mean aerial dry weight per fertile shoot to component parts under contrasting defoliation and shading treatments . 2 . 1 4 Female spikes at the end of the defoliation and shading experiment . 2 . 1 5 Mean dry weight per seed under contrasting defoliation and shading treatments . 3 . 1 Location of the study area . 3 . 2 Views of the study area. 3 . 3 Plan view of the study area showing the location of the field sites . 3 . 4 Panorama of the study area over time . 3 . 5 The field sites in December 1 977 . 3 . 6 Mean monthly (a ) rainfall � ( b ) temperature and ( c ) windspeed , at sites on the Manawatu plains . 3 . 7 Percentage frequency of surface wind directions . 3 . 8 Stratification of the vegetation. f\gu..-e 3 . 9 Topographic pro fi le of the study area . 3 . 1 0 Water table levels in the old deflation hollows over time . 3 . 1 1 Monthly rainfall at Tangimoana farm sett lement over time during 1 978, 1 979 and 1 980 . 3 . 1 2 Soil moisture profiles at five sites on the sand plain � in September 1 979. 3 . 1 3 Anabaena bloom in the flooded terminal deflation hollow� in August 1 981 . 3 . 1 4 Predicted nitrogen fixation rates during primary succession . 3 . 1 5 Carex pumila invading the terminal hollow . 3 . 1 6 Vegetation in the terminal hollow, 1 980 and 1 98 1 . 3 . 1 7 Profile diagram on the edge of the low dune . 3 . 1 8 Portions of single genets excavated from the sand plain . 3 . 1 9 Statistics of successive sympodial units along primary rhizome axes , at two sites on the sand plain. ( a ) Total length� ( b ) total number of internodes and ( c ) mean internode length per module. 3 . 20 Statistics of long sympodial units at five plain , in September 1 979. ( a ) mean length internodes and ( c ) mean internode length . 3 . 2 1 Rhizome diameter. sites on (b ) mean 3 . 22 Litter bag on an old deflation hollow on the sand_plain . 3 . 23 Decomposition of Carex pumila leaf litter over time . the sand number of 4 . 1 The sum of the density of vegetative � fertile and dead shoots of Carex pumila at fosites on the sand plain� in January 1 978. 4 . 2 The sum of the density of vegetative and ferti le shoo ts of Carex pumila over time at four sites on the sand plain , during 1 978 � 1 979 and 1 980 . 4 . 3 The sum of the density of dwarf and long shoot populations of Carex )umila over time at four sites on the sand plain between July (winter and December ( summer) 1 980 . 4 . 4 Views of the low dune showing newly deposited sand� ( a ) in spring 1 979 and ( b ) in autumn 1 980 . 4 . 5 Area of green leaves of vegetative shoots time at four sites during summer 1 979-80 . of Carex pumila over 4 . 6 Frequency distribution of the number of leaves per dead shoot in September 1 979 . F1g�re 4 . 7 Frequency distribution of shoot age ( ratio of the number of dead to total living plus dead leaves ) at harvests during 1 979 and 1 980 . 4 . 8 The sum of the sward mass of Carex pumila and other species , at each of four sites on the sand plain in January 1 978. 4 . 9 The sum of the herbage mass of vegetative ; ferti le and dead shoots of Carex pumila at four sites on the sand plain , in January 1 978. 4 . 1 0 ( a ) The mean dry weight per aerial shoot module of Carex pumila at four sites on the sand plain in January 1 978, for ( i) vegetative , ( ii ) fertile and ( iii ) dead shoots . 4 . 1 0 (b ) The mean dry weight per branch module ( aerial plus underground parts ) of Carex pumila at four sites on the sand plain in January 1 978, for (i ) vegetative and (ii ) fertile shoots . 4 . 1 1 The sum of the aerial biomass of fertile and vegetative shoots on the sand plain, during 1 978 and 1 979 . 4 . 1 2 The aerial biomass of fertile shoots of Carex pumila ( grams DM I shoot module ) over time at four sites on the sand plain, during two consecutive summers , 1 978-79 and 1 979- 1 980 . 4 . 1 3 The sum of the biomass of fertile and vegetative shoots of Carex pumila at four sites on the sand plain, during 1 979 and 1 980. 4 . 1 4 The ratio of the biomass of living to total living plus dead shoots on the sand plain over the duration of the study . 4 . 1 5 The sum of the biomass of dwarf and long shoot populations of Carex }umila over time at four sites on the sand plain between July (winter and December (summer ) 1 980 . 4 . 1 6 The sum of the biomass of component parts of (a ) dwarf and (b ) long branch modules of Carex pumila ( g DM I shoot ) on the sand plain in July , October and December 1 980 . 4 . 1 7 The mean aerial dry weight per shoot of dwarf ( D ) and long (1), fertile (F ) and vegetative (V ) shoot modules of Carex pumila at four sites on the sand plain in December 1 980 . 4 . 1 8 The effect of hares grazing. ( a ) View of the edge of the low dune and (b ) frequency distribution of leaf length in spring 1 980 . 4 . 1 9 The relationship between annual net biomass accumulation and standing biomass of Carex pumila ( a ) before and (b ) after a 1 2-th period of biomass accumulation. 4 . 20 Percent crude total nitrogen content of seeds over four consecutive summers at various sites on the sand plain . 4 . 21 The distribution of the total nitrogen content of the sward mass at four sites on the sand plain in January 1 978. 4 . 22 The distribution of the total nitrogen content of the standing shoot population of Carex pumila to component parts , at four sites on the sand plain in three successive years (a ) 1 977-78 , (b ) 1 978-79 , ( c ) 1 979-80 . Figure 4 . 23 Stage of development of fertile shoots of Carex pumila ( a ) over time at two sites on the sand plain� during summer 1 979-80 and ( b ) on the low dune . 4 . 24 Green leaf area ( cm / shoot ) of fertile shoots . 4 . 25 Mean seed weight of Carex pumila over time at three sites on the sand plain during summer (a) 1 978-79 and ( b ) 1 979-80. 4 . 26 Mean dry weight per inflores cence of Carex pumila over time at three sites on the sand plain � during summer (a) 1 978-79 and ( b ) 1 979-80 . 4 . 27 Mean number of spikelets ( seeds ) per inflorescence of fertile shoots of Carex pumila over time, at three sites on the sand plain, during summer (a) 1 978-79 and ( b ) 1 979-80 . 4. 28 Mean number of female spikes per inflorescence over time at three sites on the sand plain, during summer 1 979-80 . 4 . 29 Mean number of spikelets ( seeds ) per female spike over time at three sites on the sand plain, during summer 1 979-80 . 4 . 30 Seasonal maximum mean dry weight of seeds per inflores cence in four consecutive summers at various sites on the sand plain. 4 . 3 1 Seasonal maximum mean dry weight per seed of Carex pumila in four consecutive summers at various sites on the sand plain. 4 . 32 Seasonal maximum mean number of seeds per inflorescence on fertile shoots of Carex pumila in four consecutive summers at various sites on the sand plain. 4 . 33 The fre�uency distribution of the stage of development of fertile shoots in February 1 980, on sheltered and control plots . 4. 34 Proportional allocation of dry weight and total nitrogen of the sward mass of Carex pumila to aerial and underground components, at each of four sites (S1 , S2, S3 and S4) , in summer (a ) 1 977-78 and ( b ) 1 978-79 . 4. 35 The proportion of the total nitrogen content of the sward mass of Carex pumila found in vegetative, fertile and dead shoots, at four sites in summer 1 979-80 . 4 . 36 Proportional allocation of biomass of fertile shoots of Carex pumila to component parts over time, in ( a ) 1 978-79 and ( b ) 1 979-80 . 4 . 37 Proportional allocation of aerial biomass of ferti le shoots of Carex pumila to component parts in mid-summer in four consecutive seasons . 4 . 38 Proportional allocation of fertile shoot populations of December 1 979. dry weight and total nitrogen of Carex pumila to component parts, in 4. 39 Ratio ohe density of ferti le to vegetative shoots at four sites on the sand plain, in January 1 978. Figt..�re... 4 . 40 Proportional allocation of aerial biomass of Carex pumila to vegetative shoots (ABV ) and to vegetative ( AVR) and fertile fractions of reproductive shoots in four consecutive seasons . 4 . 41 Proportional allocation of aerial biomass of dwarf and of long shoots of Carex pumila to component organs� averaged over four sites in July 1 980 . 4 . 42 Proportional allocation of biomass of ( a ) dwarf and (b ) long branches of Carex pumila to component organs in July , October and December 1 980 . 4. 43 The distribution of biomass o f the sward to Carex pumila and to other species with and without nitrogen fertilizer in December 1 980 . 4. 44 Effect of nitrogen fertilizer on the distribution of ( a ) dry weight and ( b ) total nitrogen content of Carex pumila biomass to component organs at two sites in summer 1 980-81 . 4. 45 The effect of nitrogen fertilizer on the vegetation on the low dune . 4. 46 The effect of nitrogen fertilizer on the number of leaves per shoot , in December 1 980 . 4. 47 The effect of nitrogen ferti lizer on the proportional allocation of biomass of component shoot populations of Carex pumila to component organs , in December 1 980 . 4. 48 The effect of nitrogen fertilizer on the proportional allocation of aerial biomass of component shoot populations of Carex pumila to component organs , in December 1 980 . 4 . 49 The effect of nitrogen fertilizer on the stage of development reached by fertile shoots , in December 1 980 . LIST OF TABLES 2 . 0 Stages of development of fertile shoots of Carex pumila (after Waldren and Flowerday � 1 979) . 2 . 1 Dry weight of seeds (utricles ) per shoot . 2 . 2 Number of seeds per shoot . 2 . 3 Dry weight per seed (utricle ) . 2 . 4 Percentage of utricles containing mature nuts . 2 . 5 Analysis of variance . 2 . 6 Reproductive effort as a proportion of (a )aerial and ( b ) total biomass of fertile shoots . 2 . 7 Parameters of shoots under contras ting defoliation and shading treatments . 2 . 8 Contributions to propagule weight in Carex pumila . 3 . 1 Wind speed at four sites on the sand plain under two treatments (with ( S ) and without ( C ) shelter ) . 3 . 2 Percentage cover at five sites on the sand plain in November 1 980 . 3 . 3 Mean summer water table levels at six sites on the sand plain� November 1 978 to February 1 979 . 3 . 4 Percentage of total oven dry weight of various particle size classes of sand at 0- 1 00 mm depth at three deflation hollow sites on the sand plain . 3 . 5 Soi l pH at five sites on the sand plain in spring 1 978 and 1 980 . 3 . 6 Available nutrient status and percentage carbon � organic matter and nitrogen of the soil at five sites on the sand plain � in September 1 978. 3 . 7 Rates of acetylene reduction and corresponding values for N-fixation of soil and unialgal saes from a moist sand slack on the Manawatu coastal dune system , September 1 979 . 3 . 8 Characteristics of excavated portions of Carex pumila genets in winter 1 980 . 3 . 9 Spread of the clone into the terminal hollow over time . 3 . 1 0 Mean length and time of initiation of growth of successive long branch units on a genet excavated from the terminal hollow in July 1 980 . 3 . 1 1 Outline of the procedure to de termine the decomposition rate of Carex pumila leaves using litter bags in a deflation hollow on the sand plain. Ta.ble.- 3 . 1 2 Decomposition parameters for Carex pumila leaf litter. 3 . 1 3 Observed deviations of leaf litter. and Y , predicted re si duals values of Y ( ln Wt/Wo ) � standard and standard residuals for Carex pumila 4 . 1 Change in vegetative shoot density in autumn and in winter 1 978 at four sites on the sand plain� relative to the combined fertile plus vegetative shoot density at these sites in January 1 978 . 4 . 2 Mean density of fertile shoots Carex pumila at five sites on the sand plain, in four consecutive summers . 4 . 3 Mean age of Carex pumila shoot populations . 4 . 4 Mean vegetative 1 979-80 . dry and weight per rhizome segment attached to fertile , dead shoots at four sites on the sand plain in summer 4 . 5 Mean dry weight per rhi zome segment � December 1 980 . 4 . 6 Mean dry weight per unit length (g/m) of rhizome segments attached to living and dead shoot modules at five sites on the sand plain � in spring 1 979 . 4 .7 Net accumulation of biomass of Carex pumila at sites SO , S1 , S2 , S3 and S4 between successive harves ts . 4 . 8 Estimates of aerial shoot growth rates of long and dwarf shoots at sites on the sand plain during spring and early summer 1 980 . 4 . 9 Rates of leaf emergence and increase in leaf length at SO , between August and October 1 980 . 4 . 1 0 Estimated rates of annual net biomass accumulation of Carex pumila at five sites on the sand plain , during 1 978 , 1 979 and 1 980 . 4 . 1 1 Analysis of variance of the regression of standing biomass on annual net biomass accumulation. 4 . 1 2 Mean seasonal maximum aerial dry weight per shoot module of reproductive shoot populations of Carex uumila at five sites on the sand plain , in each of four summers . 4 . 1 3 Seasonal minimum and maximum mean dry weights per aerial shoot module of the vegetative shoot populations of Carex pumila at five sites on the sand plain. 4 . 1 4 Sward mass ( grams DM/ m2 ) of species other than Carex pumila in 1 979-80 and 1 980-81 at two sites on the sand plain. 4 . 1 5 Mean energy values ( joules / gram DM) of component parts of Carex pumila shoots . 4. 1 6 Mean elemental concentrations of component parts of Carex pumila in October 1 978. To.ble. 4 . 1 7 Percent crude total nitrogen values for component parts of live and dead branch modules of Carex pumila; averaged over four sites ; in summer 1 979-80. 4 . 1 8 Percent crude total nitrogen values for component parts of live and dead branch modules of Carex pumila . 4 . 1 9 Percent crude total nitrogen values for component parts of vegetative and fertile Carex pumila shoots in December 1 980 . 4 . 20 Analysis of variance of percent total nitrogen content of component organs of Carex pumila. 4. 21 Mean percent crude total nitrogen values for plant species other than Carex pumila found on the sand plain. 4 . 22 Total nitrogen content of aerial biomass of Carex pumi la during 1 978. 4 . 23 The dry weight and total nitrogen content of component organs of fertile shoots per unit area , at S 1 and S2 in December 1 979 . 4 . 24 Total nitrogen content ( grams N per unit area ) of the aerial biomass of Carex t�!la at five sites on the sand plain between 1 977-78 and 1980-81 summer means ) . 4 . 25 Total nitrogen content of species other than Carex pumila . 4 . 26 Mean number of spikelets per female spike on fertile shoots of Carex pumila at four sites on the sand plain , in December 1 979 , (a ) on the most distal and most basal female spikes on the culm and ( b ) on inflorescences wi th different numbers of spikes . 4 . 27 Estimated parameters of seed shed at various sites on the sand plain in late summer ( a ) 1 977-78 and ( b ) 1 978-79 . 4 . 28 Regression equations used to spike (y = seed number per spike ; estimate seed number per female x = spike length ) ; in 1 979-80. 4 . 29 Mean loss of seeds (number and dry weight ) from December 1 979 to ( a ) early January 1 980 and ( b ) early February 1 980 . 4 . 30 Effect of shelter on fertile shoot development . 4 . 3 1 The effect of shelter on ( a ) mean spike length number of seeds/spike in December 1 979 and ( c ) seeds/spike at S2 over time . and mean ( b ) mean number of 4 . 32 Proportional allocation of biomass of Carex pumila to component parts over time . 4 . 33 Proportional allocation of total nitrogen of Carex pumila biomass to seeds , rhizomes of vegetative shoots and the sum of the two ( a ) in December 1 979 and (b ) in December 1 980 . 4 . 34 Proportional allocation of biomass of fertile shoots of Carex pumila to component organs in December 1 980 . Table. 4 . 35 Proportional allocation of to tal nitrogen (TN ) and dry weight (DW ) of the aerial biomass o f fertile shoots of Carex pumila to component parts , summer 1 977-78 and 1 978-79 . 4 . 36 Ratio of the biomass of dwarf to total (dwarf plus long) branches of Carex pumila at four si tes in July, October and December 1 980 . 4 . 37 Proportion of the total nitrogen content of the biomass of Carex pumila allocated to dwarf and to long branches at four sites in December 1 980 . 4 . 38 Analysis of variance of the proportion of biomass of dwarf and of long branch modules in rhizomes , in July 1 980 . 4 . 39 Analysis of variance of seed reproductive effort as a proportion of aerial biomass of fertile shoots attached to dwarf and to long rhizome segments , in December 1 980 . 4 . 40 Total reproductive effort as a proportion of aerial biomass of fertile sheets attached to dwarf and to long rhizome segments , in Decemoe r l�ts • 4 . 41 Analysis of variance of the proportion of aerial biomass of fertile shoots attached to dwarf and of long rhizome segments in dead leaves , in December 1 980 . 4 . 42 Shoot and shoot bud densities at SO , S1 , S2 and S3 on the sand plain , in December 1 980 , with (N ) and without ( C ) nitrogen fe rtilizer addition . 4 . 43 The effect of nitrogen fertilizer on the %TN of component parts of Carex pumila shoots at two sites on the sand plain . 4 . 44 The effect of nitrogen fertilizer addition on the ratio of dwarf shoots to total ( long plus dwarf ) shoots ; ( a ) density, (b ) total biomass and ( c ) aerial biomass . 4 . 45 The effect of nitrogen fertilizer addition on (a ) total and (b ) aerial biomass of component shoo t populations of Carex pumila on the sand plain , in December 1 980 . 4 . 46 Mean dry weight per unit area of component organs of Carex pumila populations in table 4 . 45 . 4 . 47 The effect of nitrogen fertilizer on mean dry weight per branch module of component parts of Carex pumila shoots , in December 1 980 . 4 . 48 The weight , nitrogen and long effect of nitrogen fertil izer addition on ( a ) mean (b ) mean number of seeds per inflorescence and ( c ) content of seeds of Carex pumila on shoots at tached to rhizome branches , in December 1 980 . seed total dwarf 4 . 49 The effect of nitrogen fertilizer (N ) on the proportional allocation of biomass based on ( a ) dry weight and (b ) total nitrogen to component shoot populations , in December 1 980 . 4 . 50 Effect of nitrogen fertilizer (N ) on the proportional allocation of biomass based on (a ) dry weight and (b ) total nitrogen of Carex pumila to component organs � in December 1 980 . Table 4 . 5 1 The effect of nitrogen ferti lizer on the proportional allocation of biomass of ( a ) dwarf and (b ) long vegetative shoot populations of Carex pumila to rhi zomes ; in December 1 980 . 4 . 52 The effect of nitrogen fertilizer (N ) on the proportional allocation of aerial biomass of ( a ) dwarf and (b ) long vegetative shoots of Carex pumila to component organs , in December 1 980 . 4 . 53 Proportional allocation of dry weight and of total nitrogen of the aerial biomass of fertile shoots of Carex pumila to seeds ; with and without nitrogen fertilizer in December 1 980 . CHAPTER ONE Introduction 1 . 1 Life histories , allocation of resources and reproductive effort . 1 . 2 Successional trends of reproductive effort . 1 . 3 Quantification of life history strategies . 1 . 4 Sand plains and waterlogged soils . 1 . 5 Carex.pumila Thunb . 1 . 6 Aims of the study . Page 2 This thesis describes aspects of the life history strategy of Carex pumila , a rhizomatous perennial plant species , a colonizer of the seasonally flooded sand plains of the Manawatu coastal dune system. Carex pumila plays a pivotal role in the early development of these sand plains (Esler 1 978) . Thus � the behaviour of this species may not only reflect the early stages of development of the sere , but may also be a causal influence in the process . In this study , the life history strategy of Carex pumila on the sand plains is described in terms of the allocation of both biomass and mineral nutrients over time , along with the changes in reproductive effort of the species as the sere progresses . 1 . 1 Life histo ries , allocation of resources and reproductive effort Reproductive effort was originally defined by Harper and Ogden ( 1 970 ) as the efficiency of a plant as a "seed-producing machine" . The energy input of the machine was seen as the total assimilated carbon or gross production plus the energy stored in the o riginal seed from which the plant grew. The energy output was the energy contained in the seeds produced . The machine efficiency or reproductive effort was simply the ratio between input and output . The measurement of reproduc tive effort therefore involved the determination of gross production , starting capital and that fraction which was seeds or vegetative propagul�s. Detailed energy budgets � ideally expressed in energy units rather than grams dry weight� over the entire life cycle of the plant were thuS required . - I I Page 3 The measurement of reproduc tive effort in plants stemmed from the concept of energy allocation (Cody 1 966 ) ; that is the idea that organisms have a limited quantity of time or energy ( or other resource ) to partition between the various life activities of growth ; maintenance , predator avoidance and reproduction. The expenditure of resources on any one of these necessarily means a decreased amount available for expenditure on others . The way in which an organism allocates its resources between these functions was seen to be a characteristic of ecological and evolutionary significance . Reproductive effort as defined above is a measure of the energy allocation to reproduction, as opposed to the allocation to development of a competitive growth form or defense against predators � and thus has been seen as a quantifiable measure of the "relative value which natural selection has placed upon that function" (Harper and Ogden 1970 ) . Monocarpic grain crops and wild annual species have a high repr?ductive effort , usually greater than 20% and up to 40% of their annual net assimilation , whereas in po l cetrpic. po yoie herbaceous perennials and forest trees it may be less than 5% (Harper et al . 1 970 ) . The quantitative pattern of resource allocation integrated over the entire life of an individual is a means by which the life histories of different organisms can be compared . The resource allocation pattern will be a reflection of the individual's genotypic programme or strategy and is determined by the interaction of that strategy with the environment . A single strategy may include a range of developmental pathways as seen in plants by the plastic response of an individual to fluctuations in the environment (Hes lop-Harrison 1 964 ) . This developmental flexibi lity has itself been moulded by natural selection . Thus � the range of life history strategies Page 4 displayed by plants is considered to be adaptive , reflecting the sets of selective pressures under which these organisms have evolved . Examination of the literature reveals that the types of life his tories encountered in nature � as with the types of physical and biological environments � are limited. For example � Dobzhansky ( 1 950) � working with birds , distinguished between the ways in which natural selection acted in the tropics and at higher latitudes . He argued that in temperate zones physical environmental factors which are density-independent tend to limit population growth whereas in the tropics biological (density-dependent ) interactions predominate . He argued , further, that this led to selection for early onset of sexual maturity and larger clutches in temperate zones as opposed to selection for the ability to compete and to avoid predation in the tropics . MacArthur and Wi lson ( 1 967 ) developed these ideas , coining the term 'r-selection' for selection in uncrowded situations favouring rapi d population expansion and 'K-selection' for selec tion in saturated or crowded situations favouring ability to compete . During the last 1 5 years the meaning of r- and K-selection has developed and broadened t o include selection in relation to the degree of environmental uncertainty or habi tat disturbance ( Gadgil and Solbrig 1 972 ) and also in relation to the degree of density-dependent mortality (McNaughton 1 975 ) . I t has also been used erroneous ly to describe the combinations of life history characteristics that it has brought about . r-K-selection theory predicts the ass ociation of life-history traits into two groups ( 1 ) resulting from r-selection : early age at first reproduction ; large clutch size ; once only reproduction ; no parental care ; large reproductive effort ; many , small progeny ; low assimilation efficiency ; and short generation time (a specific - 7 ,. Page 5 combination of which is a single r-strategy) and ( 2 ) resulting from K-selection : delayed reproduction ; repeated reproduction ; clutches ; parental care ; smaller reproductive effort ; few � large progeny ; and high assimilation efficiency (a specific combination of which is a single K-strategy) ( Stearns 1 976) . The use of the terms r- and K-selection to refer to selection in relation to the amount of density-dependent mortality and degree of habitat stability in addition to the original meaning of selection in relation to the degree of crowding (competition) has led to considerable confusion. If crowding resul ts in density dependent mortality through depletion of resources or conversely if crowding is prevented by density-independent mortality fac tors , or if the stabi lity of the habitat affects the degree of crowding then the life history consequences may be predicted by r-k selection theory. The confusion arises , however , where the selective effects of these environmental factors are not mediated by their effect on crowding but where instead they directly affect life history parameters . For example , where the density-dependent source of mortality is predation , the life-history consequences are not likely to be the same as those that result from crowding (Wilbur et al . 1 974 ; Wilbur 1 976 ) . Reproductive effort has been the life history trait most commonly used to place a population on the r-K-strategy continuum for the purpose of comparison with populations living in other environments (e . g . Gadgil and Solbrig 1 972 ; Gaines et al . 1 974 ; Abrahamson ' 1979 ) . More recently , however , Parry ( 1 98 1 ) deemed that reproductive effort should not be used as an index of r- and K-selection. In support of this view he pointed to the lack of association between reproductive effort and egg size in both plants and other organisms (Wilbur 1 977 ; Bos tock and Benton 1 979) ; egg size being a life Page 6 history trait fundamental to r-K-selection theory. Further, Parry has questioned the explanations given by Gadgil and Solbrig ( 1 972 ) and Gaines et al . ( 1 974) for differences in reproduc tive effort between populations where the degree of density�dependent/density�independent mortality was inferred but not actually measured . He indicated that other causes of these population differences � for example � differences in adult as opposed to juvenile mortality rates � were not considered . The life history consequences of such differences in mortality schedules are discussed below. The possibility remains that r-K-selection which assumes constant mortality in both adults and juveniles , need not have been invoked . Alternative explanations for those groupings of life history attributes that have become known as r- and K-strategies and for other groupings of traits , have been hypothesi zed . "Bet -hedging" or stochastic models which deal with the consequences of fluctuations in juvenile and adult mortality and fecundity schedules ( Schaffer 1 974 ) predict that than adult when juvenile mortality or birth rate fluctuates more mortality, a syndrome of delayed maturity� reduced reproductive effort , fewer young and longer life should evolve ; whereas when adult mortality fluctuates more than juveni le mortality or birth rate then earlier maturity � larger reproductive effort and more young should evolve ( Stearns 1976 , 1 977 ) . Measurements of reproductive effort in situations where adult and juvenile mortality schedules have been estimated � confirm that higher adult mortality is correlated with high reproductive effort and lower adult mortality with lower reproductive effort ( Calow 1 979 ; Primack 1 979 ) . Page 7 Grime ( 1 977 ) recognized stress ( conditions that restrict production , e . g . mineral nutrient or water deficiency ) and disturbance ( conditions associated with the destruction of plant biomass e . g . wind or herbivores ) as factors which limit plant biomass in any habitat. He proposed the existence of three primary strategies in plants : the competitive ( C ) s trategy associated with low st ress and low disturbance conditions , the stress-tolerant ( S ) strategy with high stress and low disturbance , and the ruderal (R) strategy with low stress and high disturbance . Secondary strategies in response to combinations of these conditions were also recognised . It is evident , therefore , that sele ction regimes that determine plant strategies have many dimensions (Wilbur 1 976 ) , one of which is the continuum from r- to K-selec tion ( sensu MacArthur and Wilson 1 967 ) . Others include seasonality (Boyce 1 979) , degree of habitat disturbance ( Gadgil and Solbrig 1 972 ) , environmental stress (Grime 1 977 ) and predat ion (Wilbur et al . 1 974) . Thus , the explanation for the evolution of all observed combinations of life history characteristics on the basis of just one of these dimensions ( e . g . r- and K-selection ) will probably be an over- simplification and this has prevented the formulation of precise predictions which may be empiri cally tested ; the theory therefore has become imprecise . 1 . 2 Successional trends in reproductive effort Harper ( 1 967 ) questioned whether the proportion of a plant' s output that is devo ted to reproducti on is higher in colonizing species than in those of more mature habitats . It has been suggested that ecological succession invo lves a shift from r- to K-selection with increasing age or maturity of the community (Wilbur et al . 1 974) . r-K-selection theory predicts that r-selection will produce a specific combination of life history traits (a specific r-strategy ) and K-selection another (a K-strategy ) (Gadgil and Solbrig 1 972 ) . Thus ; seral species are predicted to possess more rapid growth to reproductive maturity, higher reproductive effort and more copious smaller ; often wind dispersed seed ( typical of production of r�strategists ) than species which replace them as the sere progresses (K-strategists ) . Attempts to answer Harper's question conform to a common pattern reproductive allocation decreases with increasing maturi ty of the sere (Gadgil and Solbrig 1 972 ; Abrahamson 1 975b ; Gaines et al . 1 974 ; Hickman 1 975 ; Newell and Tramer 1 978 ; McNamara and Quinn 1 977 ; Roos and Quinn 1 977 ; Abrahamson 1 979 ) . For example ; in a study with goldenrods ( Solidago species ) Abrahamson and Gadgil ( 1 973 ) followed the distribution of dry matter between reproductive and vegetative tissues in four perennial species from a series of sites that they ranked ac cording to successional maturi ty an immature , dry open and heavily disturbed site ; a wet site of int ermediate maturity ; and a mature hardwood site that was relatively undis turbed . Although not all species occurred at all sites , the appropriate comparisons be tween sites confirmed their prediction that reproductive effort would decline with increasing successional maturity of the community . However, Bradbury and Hofstra ( 1 975 ) were unable to show that the difference in reproductive effort between two Solidago canadensis populations from sites of different successional maturity was significant . In such studies with field populations there is no way of knowing whether the observed differences were environmentally-induced plastic differences or genetically based . Only where comparison of diffferent populations in a greenhouse or common site has been carried out have Page 9 these components of variation been distinguished . Hickman ( 1 975) found that for the annual ; ?o \�9or'1UY>'l Polyenum cascadense ; the trend of increasingly greater reproductive effort in populations from successively harsher; more open habitats found in the field ; was not found in the greenhouse . Hickman concluded that the field differences were directly attributable to environmental effects rather than to genetic differentiation between the populations . A similar result was obtained by Raynal (1979 ) in studying field and greenhouse population of Hieracium florentinum : field differences in reproductive effort disappeared when populations were grown under identical glasshouse conditions . Roos and Quinn ( 1 977) showed field differences in reproductive effort in the predicted direction between populations of the grass Andropogon scoparius at sites of varying age . In the greenhouse differences in the same direc tion were found although these were small and often not significant . They concluded that there appeared to be at least a possibi lity that genetic differentiation existed between young and old field populations in terms of reproductive effort . Thus the results to date are no more than suggestive that ecotypic differentiation in terms of resource allocation to rep roduction has occurred between populations of a single species at various seral stages . The adaptive strategy in early plant succession therefore appears to be high reproductive effort and copious production of small highly vagile offspring (Horn 1 974 ) . A similar pattern was des cribed for birds colonizing is lands ( Diamond 1 975 ) . Diamond termed the pioneers "supertramps" which were not particular about their choice of habitat and were prolific and dispersive . Page 1 0 The perennial rhizomatous growth habit is frequently encountered in plans of pioneer and early seral communities particularly where continual renewal of the substrate is occurring such as on sand dunes ( Ogden 1 974a) . In the harsh unstable environments of dune slacks in coastal New Zealand , native herbs possessing the rhizomatous habit and tolerance of prolonged inundation with water predominate (Esler 1 978 ) . The strategy of these plants which usually allows for both long distance dispersal by seeds and for rapid extension by rhizomes into newly exposed adjacent areas which may be unfavourable for seed germination , does not conform to the suite of life history characteristics predicted under r-selection for colonizing species . The strategy of rhizomatous perennials differs from that of annual f>rec.\u4e r-strategists in that the strategy does not � competition with o ther species which occurs in seral habitats . The rhizomatous perennial may produce a dense stand of ramets which resists invasion and so persists as a long-lived seral mono-culture . Such an efficient use of bio logical space by clonal perennials has been attributed to the lack of self-thinning within the stand ( Hutchings 1 979 ) . Hutchings found that shoot numbers remained relatively constant throughout the growing season and an increase in shoot size s topped when the "thinning line " (Yoda et al. 1 963 )* was reached . However � when shoots subsequently died Hutchings observed no further *Footnote : The thinning line , with a gradient of -3/2 � represents the changes through time of mean plant weight and plant density of populations of genets of a wide variety of species undergoing thinning . Page 1 1 accumulation of biomass . This lack of self-thinning was considered an important aspect of the efficient system of utilization and re-dist ribution of resources within the clone . It is this aspect which results in the development of closed stands of clonal perennial plants . Alternatively , the rhi zomatous perennial may escape to new areas through seed production. This choice "depends in part on decisions made by natural selection governing the changes in pattern of resource allocation in response to shoot density and in particular the relative proportions devoted to seeds and rhizomes " (Ogden 1 974a ) . Allocation has been shown to favour seed production as opposed to vegetative spread as the density of the clone increases in both Tussilago farfara (Ogden 1 974a ) and Rubus spp (Abrahamson 1 975 ) ; an adaptive response to low fertility conditions in open habitats (S-selec tion of Grime 1 977 ) . This view that allocation to seeds ( sexual reproduction ) and to rhizomes ( clonal growth ) may be alternatives has also been suggested by the work at Tripathi and Harper ( 1 973 ) who compared the number of seeds and rhizome buds produced per plant for two cogeners � Agropyron repens and A. caninum of contrasting growth habits . A . repens produces an ext�nsive rhizome system whereas A . caninum forms a closed tussock by production of intravaginal tillers . Although the estimates of seeds and buds were widely different between the two species , the sum of seeds plus rhizome buds per plant were remarkably similar. Seral perennials are also found in ruderal or disturbed habitats . For example , Plantago major in ruderal habitats has been found to allocate from 10 to 28% of its total biomass to seeds (Kawano 1 975 ; Hawthorn and Cavers 1 978) � whereas other perennials from less Page 1 2 disturbed habitats generally have a lower seed reproductive effort clona.l growlti ( Kawano 1 975 ) . The dry weight allocation to eloaapowth in Plantago major was found to overlap the range found in the rhizomatous perennial Tussilago farfara of seral habitats namely 0�22% (Ogden 1 974a ) and was similar to that ( 1 8% ) found after 10 weeks growth in the stoloniferous perennial Trifolium repens of relatively open , disturbed habitats ( Turkington and Cavers 1 978) . Trifolium repens is adversely affected by low light levels . It has thus been suggested that stolons in this species may have evolved permitting the species to escape the effects of interspecific interference ( Turkington and Cavers 1 978 ) . Such a growth habit would allow the species rapid local spread so effectively monopolizing available resources and thereby excluding other species ( K-strategy of MacArthur and Wilson 1 967 ) . A similar explanation can be given for the rhizomatous growth habit of seral perennials of low ferti lity sand dune and sand slack habitats . A characteristic feature of seral rhizomatous perennial species of nutrient deficient habitats is the linear clonal growth habit ( e . g . Carex acutiformis of marshlands in the British Isles , Grime 1 979 ; and Ammophila arenaria in sand dunes ; Raunkaier 1 934 ) . The adaptive significance of linear clonal growth in such habitats may be related either to ( 1 ) the efficient capture of nutrients by the roots of extending rhizomes and the efficient transport of these nutrients to other parts of the clone or ( 2 ) the recycling of nut rients from the older to the more juvenile parts of the clone ( Noble and Marshall 1 983 ) . Noble and Marshall found little evidence of basipetal movement of nitrogen and phosphorus , in one such species , Carex arenaria , in either field or glasshouse experiments ; even when older basipetal shoots were severely nutrient-stressed . On the contrary; preferential movement of nutrients taken up from the soil and from older parts of Page 1 3 the clone was found towards the younger , developing parts at the front edge of the clone . Such acropetal movement also applied to assimilate re-distribution in both this and other studies ( e . g . Forde 1 966 � in Agropyron repens ) . Thus � the linear rhizomatous habit of Carex arenaria allows for the continual extension into previous ly uncolonized sands and the exploration of scarce nutrient res ources (Noble and Marshall 1 983 ) . It is apparent , however , that there is only limited redistribution of water , nutrients or assimilate in Carex arenaria . Tietima ( 1 979 ) found no long distance water or mineral nutrient translocation along the rhizome in this species . He concluded that redistribution via the phloem is limited t o the extent that the plant could not use this supply to maintain a high growth rate in those plant parts not direct ly supplied by an external source of mineral ��Atr ients. The explanation of life history strategies in seral species will involve estimation of environmental parameters , variabilities in adult and juvenile mortality schedules , the degree of density dependent versus density independent mortality and the degree of environmental stress (Stearns 1 977) . 1 . 3 The quantification of life history strategies In section 1 . 1 , the idea was developed that the life history of an organism represents a series of compromises to a set of physical and biological environmental conditions . A life history has many components . The battery of tactics that makes up a life history strategy enables a plant to survive and reproduce ; that is , it implies a set of adaptive responses that have been put together over Page 1 4 evolutionary time (Wilbur 1 976 ) . The important components of a life history strategy according to Solbrig ( 1 980 ) include : 1 . soil seed pool � seedling and adult mortality. 2. age at first reproduction 3. reproductive life span 4 . proportion of fert ile to vegetative individuals at any one time . 5 . fecundity, seed production which depends upon flower number and pollination rat es . 6 . fecundity-age regression 1. reproductive effort . The quantification of a life history strategy of one plant population is thus a formidable task. In any one study only a sample of life history parameters can be measured . The choice is for the worker to decide the most efficient means by which to quantify a strategy. As noted previously , reproductive effort has been the life his tory parameter most commonly used to quantify the life history strategy of a population for the purpose of populations living in other environments . Measurement of reproductive effort comparison with Harper and Ogden ( 1 970) have indicated that the ideal measurement of reproductive effort involved the calculation of detailed energy budgets , expressed in energy units over the entire life of the plant : Page 1 5 Gross reproductive effort = Total energy as propagules Total energy of seed from plus Gross assimilation which plant grew Because of the difficulties in measuring plant respiration in the field , especially that of subterranean parts , the budget is usually of the energy remaining after the respiratory demands both for growth and maintenance of the organisms are met . Thus Net reproductive effort Total energy as propagules Total energy of original seed plus Net production If the metabolic efficiencies of different plant organs were similar then net reproductive effort would be a simple overestimate of gross reproductive effort . This however is probably an unreasonable assumption . In one of the few studies involving the direct measurement of both plant growth and maintenance respiration, Hansen and Jensen ( 1 977 ) showed that for Lolium multiflorum swards , the respiratory demands of roots were higher than those of the aerial fra ction ( leaves ) . Further, where per gram energy values are more or less similar in all plant organs , es timates of dry weight would at least be proportional to net production (Harper and Ogden 1 970) . Hickman and Pitelka ( 1 975 ) showed that for plants with largely Page 1 6 carbohydrate seed reserves � and therefore an approximate equivalence of energy values between vegetative and reproductive structures ; significant allocation pattern differences between populations could be reflected equally well by dry weight as by energy measures . They concluded therefore that time�consuming energy determinations were unwarranted for such studies . Thus , by taking into consideration the loss of plant parts and by ignoring the original seed weight as negligible compared to total dry weight , reproductive effort can be estimated by the ratio of total weight of seeds produced to total plant weight at time of maturity. This ratio should include underground parts and thereby will differ from measurements of "harvest index" ( sensu Donald 1 962) (Harper and Ogden 1 970 ) . No other technique for estimating rep roductive effort has proved superior for field s tudies requiring large sample size ( Primack 1 979 ) . Problems have been encountered however in the quantification of reproductive effort in rhizomatous perennial species . There has been a dichotomy of views over whether the allocation to rhi zomes is more properly viewed as allocation to ( c lonal ) growth ( Harper 1 977 ) or as allocation to ( asexual ) reproduction ( Abrahamson 1 979 ) . At a more practical level problems are encountered with parts continually dying ( and thus being liable to be lost ) redistribution of assimilate ( o ther resources ) within a plant between different organs and shedding of parts such as seed prior to harvest . I t has often been necessary to estimate the loss of seed ( and other pa�ts ) � for incorporation into the budget , by the regression of seed weight ( or that of other part ) on the size of a plant part not liable to be lost ( e . g . receptacle ; Harper and Ogden 1 970 ) . Page 1 7 I t is through viewing reproductive effort as a measure of the ' cost of reproduction ' � that is , a measure of a plants resource allocation to reproduction as a whole � that many of the attempts to measure reproductive effort have differed from that o f Harper and Ogden ( 1 970) . They measured that proportion of resources allocated to seeds and � in fact � adopted the loose definition of a seed as the dispersal unit and so the whole achene of the composi te Senecio vulgaris was taken as the " seed" in their study . Similarly, in many other s tudies , the estimation of reproductive effort has involved the calculat ion of the allocation to d issemules or diaspores ( the units of dispersal ) rather than that to seeds alone ( e . g . Handel 1 978 ) . The use of the ratio of weight of "seed" output to total plant weight at maturity allows a simple comparison of the efficiencies of plants with diverse morphologies as seed producers ( Harper and Ogden 1 970 ) . The redefinition of reproductive effort and the failure of workers to ac tually measure seed weight in the s tudy of dry weight allocation in plant populations ( e . g . Hawthorn and Cavers 1 978 ) precludes the possibility of such comparisons . The proportion of resources allocated to seeds � the " reproductive effort" of Harper and Ogden is a fundamentally different quantity from that devot ed to flowers � ( " reproductive effort" of Abrahamson and Gadgil 1 973 ) or to flowers and flower stalks � ( "reproductive effort" of Hawthorn and Cavers 1 978) . No consensus has developed as to what should b e included in this ratio . Trivedi and Tripathi ( 1 982 ) for example calculated three different " reproductive effort" values ( the allocation to seeds , to flowers and to flowers plus ancillary reproductive structures ) in order to make comparisons with previous studies . Page 1 8 Thompson and Stewart ( 1 981 ) argued that allocation t o all reproductive structures � that is " total reproduct ive effort " � is a more realistic estimate of the resources actually committed to reproduction than is seed output . I t becomes a matter for the worker to decide therefore whether or not a s truc ture is reproductive , or ancillary to rep roduction . Thompson and S tewart stated that they had no wish to be dogmatic on this issue . By measuring allocation to flower stems and/or flowers workers are no longer measuring the efficiency of a plant as a seed producer, ( reproductive effort s ensu Harper and Ogden ) : rather they are attempting to measure a far more elusive quantity , namely the resource c ost to a plant of seed reproduction . The cost to the plant of the production of structures anci llary to reproduction are quite rightly incorporated into this cost analysis . If a plant is viewed as the means by which one seed produces more seeds then the whole plant body could be argued to be ancillary to reproduction and total reproductive effort becomes 100% for all plants and thus meaningless as a concept . A less extreme point of view has been to include flowers and supporting structures not present in the vegetative plant body in the measurement of the ' effort ' involved in reproduction. A difficulty arises in the measurement of total rep roductive effort because many reproductive structures , including flowers in many species and even the dissemules themselves are photosynthetic and therefore can be expected to contribute t o their own increase in dry matter or energy (Bazzaz and Carlson 1 979 ) . For example , in wheat , es timates of the contribution of ear photosynthesis to final grain weigh��be as much as 60% ( Evans et al , 1 97 5 ) . In such instances , the production of those reproductive s tructures is not a cost to the Page 1 9 plant in the sense that Cody ( 1 966 ) , in developing the concept of energy allocation� visualized the production of eggs being a cost or drain on the birds limited energy pool . Where reproduc tive photosynthetic assimilat ion is occurring the drain of carbon to reproductive s tructures is being met by the increase in overall size of the pool of that res ource . From an evolutionary point of view, the budget of a resource the very act of allocation of which leads to an increase in the size of the pool of that resource , is of lit t le value since the allocation of such non-limiting resources will not allow estimates of the true costs of particular structures or functions and thus cannot be expected to shed light on the value natural selection has placed on those structures or functions ( Calow 1 979 ) . Harper and Ogden recogni zed that their selection of the net energy budget as the basis for describing the plants resource allocation was arbitrary acknowledged that the allocation of some o ther limited resource , for example , nitrogen might be more important in the evolution of life his tory or reproductive strategies . As Harper ( 1 977 ) stated the cost of reproduction cannot be measured without knowing the appropriate currency . It is to be expected that the relevant currency in which to measure res ource allocation in plants may vary between species and between habitats since the resource that limits population growth in one set of circumstances is not likely to be the same in all situations . Thompson and Stewart ( 1 981 ) opine that the measurement of energy or dry matter allocation in plants is inherently wrong . They suggest that the appropriate currency for allocation studies in plants is mineral nutrients since minerals represent a plant resource that forms a limited pool which is not increased in size by the process of Page 20 allocation itself . The only justification they see for the s tudy of energy or biomass allocation in p lants is where it can be shown that mineral nutrient concentrations in vegetative and reproductive parts do not substantially differ . No equivalence has emerged between vegetative amd reproductive structures for mineral nut rient concentrations . For example , Fagus sylvatica seeds contain six times as much minerals per gram dry weight as beech wood ( Harper 1 977 ) . In an allocation s tudy � Lovett Doust ( 1 980b ) showed for Smyrnium olusatrum � patterns of allocation of dry matter and phosphorus , a mineral nutrient with a reputedly crucial role as a storage element in seeds , to be qui te different . Similarly , van Andel and Vera ( 1 977) found that neither nitrogen , phosphorus , nor potassium allocation patterns reflected the dry matter allocation in either Senecio sylvaticus or Chamaernerion augustifolium , but by taking the three nutrients together , a good approximation was obtained . I t is reasonable , however , to suggest that the allocation of � for example , limited nitrogen to plant roots which would allow fo r further growth of this fraction and thus the further uptake of nitrogen so increasing the poo l of this resource , is the allocation of dry matter ( energy) that same as that involving the find irrelevant to the study strategies in plants . o f Thompson and the evolution Stewart ( 1 98 1 ) of life history Page 2 1 1 . 4 Sand plains and waterlogged soils Sand plains Very litt le work has been carried out on the ecology of sand plain systems other than in the British Isles and Holland (Ranwell 1 972 ) . Sand plains elsewhere in America � Eurasia and Australasia remain largely undescribed and in New Zealand at least the possibility of des cribing these unique vegetation stands and ass ociated fauna is rapidly diminishing due to their annihilation by the invasion of exotic plants ( Esler 1 978 ) principally as a result of the efforts to crop these waterlogged lands ( see for example ' Round Bush Management Plan ' 1 982 ) . Esler ' s studies ( 1 969 , 1 970) des cribed the interplay of sand ; plants , mois ture , wind and topography in shaping the sand dune system of coastal Manawatu . In this harsh , unstable environment � influenced by salt , seasonal submergence and drought , the stTong persistent west to north-west winds constantly move dry sand . The characteristic parabolic high dunes of the area form a complex pattern with o ften extensive low-lying plains enclosed between them and the foredune . When bare of vegetation, the parabolic rear dune is blown inland as a result of the removal of dry sand at the base of its windward s lope (windsweep ) down to the level of moist sand just above the water table . The damp hollow so formed at the base of the windsweep is colonized by several small rhizomatous plant species : Limosella lineata , Eleocharis neozelandica , Ranunculus acaulis , Selliera radicans ( the Manawatu sand plain form as opposed to the long-leaved salt marsh form, Ogden 1 974b ) and Myriophyllum votschii . Around the edge of this hollow Esler describes Carex £Umila forming a continuous fringe. This species of Carex has the ability to accumulate sand and Page 22 so form a low dune approximately 0. 3 metres higher than the hollow. The hollows so formed are characteristically bow-shaped . A number of sand dune species including the introduced Ammophila arenaria and the native Desmoschoenus spiralis germinate in the warm , damp sand and soon appear on the low dune ; building it higher and eventually replacing Carex pumila. As this is occurring , the rear dune will have moved further inland exposing more damp sand a t the the base of the windsweep � allowing the coloni zation and development of a new damp hollow . The process is thus repeated producing a sequence , both in space and time of small arcuate damp hollows separated by similarly shaped low hummocks within the more extensive sand plain. The youngest hollow of the chronotoposequence will be closest to the base of the windsweep of the inland moving rear dune � with progressively older hollows of greater distances seawards . Esler has des cribed the development of vegetation in the older hollows with Leptocarpus similis becoming dominant , and Schoenus nitens and other species present . Further vegetational development in these hollows may involve Cortaderia toetoe , Cordyline australis and , where deposition of sand causes the plain to become dry , Scirpus nodosus . On the older , higher , drier parts of the plain Lupinus arboreus � Hypochaeris radicata ; Leontodon taraxacoides and annual grasses and legumes form a more or less sparse cover ( Esler 1 978 ) . Waterlogged soils Frequent , seasonal or permanent waterlogging as a resul t of high rainfall , topogenic water accumulation or poor drainage leads to the formation of hydromorphic soils ( Armstrong 1 975 ) . Hokio series soils of the Manawatu sand country ( Cowie et al. 1 967 ) fall into this Page 23 category . Waterlogging and the ability to cope with these conditions will determine the nature and development of the vegetation in wet lands ( see for example Wil lis et al . 1 959 ; Jones 1 972 ) and thus the continuing development of the soi l . I n response to soil flooding � respiring aerobic micro-organisms will reduce oxygen concentrat ion to zero within a few hours ( Scott and Evans 1 95 5 ) and subsequently will be replaced by a population of anaerobic organisms . The diffusion rate of oxygen into oxygen-depleted soil is usually insufficient to maintain aerobic organisms except within a few centimetres of the surface of the flooded soil . For example , Ranwell ( 1 959) found that where the soil was waterlogged at the surface , but not flooded , completely anaerobic conditions were found only 3cm below the soil surface . Where the water table drops below the surface , anaerobic conditions can be expected to prevail even above the ground water level due to capillary action ( Boggie 1 972) . The sequence of events that might be expected in a mineral soil as a result of flooding ( Patri ck and Turner 1 968) � s tarts with a decline in oxygen concentration followed by the chemical reduction o f nitrat e , then manganes e , then iron , then sulphate . Where organic matter is present , the concentrations of the diverse organic products of anaerobic microbial metabolism and of carbon dioxide will also increase , al though because carbon dioxide has a very high solubi lity in water it will be dispersed more rapidly from its site of production than the opposite movement of oxygen . Thus � the adverse effects of waterlogging on plant growth and survival are rarely the result of carbon dioxide toxicity . Litt le is known of the effect s on plants in waterlogged habitats of the p roducts of anaerobic microbial metabolism al though ethylene has been implicated in the reduction of Page 24 transpiration, wil ting and deaf abscission that accompanies flooding ( Jones and Etherington 1 970) . At positive redox potential certain facultative anaerobic micro-organisms can use nitrate as an electron acceptor in respiration resulting in denitrification by the evolution of gaseous nitrogen or di-nit rogen oxide . Patrick and Tusneen ( 1 972 ) found large losses of nitrogen from continuously flooded soi ls in which a narrow surface aerobic layer remained overlying an anaerobic layer . They explained these losses as the result of the downward movement of nitrate formed in the surface layer by nitrification into the underlying layer where denitrification took place . In addition ; the loss of nitrogen from alternately flooded and drained soils as a result of denit rification has long been recognized (Russel 1 96 1 ) . In waterlogged soils nitrogen usually reaches the roots of plants as ammonium ions derived from decaying organic matter , al though nitrogen-fixation by free-living microo rganisms may also be important for example in rice paddies where both blue-green algae and bac teria have been demonstrated to fix nitrogen (McRae and Cas tro 1 967) . Nitrogen fixation by autotrophic blue-green algae ( e . g . Nostoc and Anabaena species ) is widespread (Mulder et al . 1 969) and important in raw soils as they pioneer the sere of N-fixing microorganisms (Etherington 1 97 5 ) . Azotobacter bacteria may be ecologically important N-fixers in highly deficient habitats such as on sand dunes where it is found in the rhizosphere zone of Ammophila arenaria which supplies organic metabolites to the bacterium and is provided with survival levels of nitrogen (Etherington 1 975 ) . - I Page 25 At increasingly lower redox potentials manganese and then iron will be reduced causing an increase in the concentration in the soil of exchangeable/soluble ions of both these e lements . Since soils usually contain more iron than manganese � the dominant redox couple in soils is that of i ron ( Feiii/Feii ) . High concentrations of both divalent iron and manganese ( soluble reduced forms ) are toxic to plants . Thus , plants in regularly or permanently waterlogged habitats where high concentrations of these ions p revai l mus t possess some means of protection against these effects . For example � the differential ability of sand slack species � Carex nigra and Agrostis stolonifera, to survive otherwise toxic concentrations of ferrous and manganous ions which occur in the slack habitat during periods o f maximum growth of the vegetation in spring and early summer appears t o be related to their abi lity t o exclude these ions ( Jones 1 972a , b ) . The deposition of ferric ions adjacent to roots in a waterlogged soil in an otherwise reduced ( ferrous ) state ( Armstrong 1 967 ) supports the view that plants may protect themselves from this hostile soil environment by leaking ROL ) . The formation of oxygen from the roots ( radial oxygen loss � an oxygenated rhizosphere sheath by ROL requires a gas exchange pathway between roo t cells through the plant to the leaves and s tomata. This pathway wil l be enhanced by the provision of air-space tissue which in wet land species may be up to 60% by volume of the plant body compared with 2-7% pore space typical of mes ophytes ( Armst rong 1 975 ) . The formation of large air spaces ( lacunae) in various tissue ( aerenchyma ) in grasses and sedges may resul t from cell collapse . Page 26 Leaf and stem xeromorphy is a common feature of many species of waterlogged habitats ( e . g . Carex pumila � Pegg 1 9 1 4 ; Juncus and Erica spp � Armstrong 1 975 ) despite what might be expected of plants growing in aquatic conditions . Armstrong ( 1 975 ) sugges ted that " it seems likely that the primary function of xeromorphism in these plants is not [ that ] of water conservation per se but rather to reduce the velocity of water movement to the roo t surface • • • [thus ] the time available for oxidation of phytotoxins in the rhizosphere zone of oxygenation will be increased" . The observation by Jones ( 1 97 1 ) that antitranspirants improved survival under waterlogging and limited iron uptake in normally waterlogging sensitive Erica cinerea , supports this view. Other adaptations to waterlogging include those involving the root system. Under permanent waterlogging adventitious roots may become orientated horizontally close to and even above the surface ( e . g . Burgess 1 974 ) presumably within a tolerable redox zone . Where the soil water table fluctuates � a shaving brush effect in the roo t system has been observed as rising water kills the root apices but not the bases from which new laterals grow . This flux of death and recruitment of root apices produces a mass of brush-like roots ( Armstrong 1 975 ) . 1 . 5 Carex pumila Thunb The genus Carex has 73 species in New Zealand falling into 1 7 sections within three subgenera (Moore and Edgar 1 970) . Carex pumila Thunb . F l . jap . 1 784 , 39 . C . li t torea Labill . �· 2 � 1 806 , 69 , t . 2 1 9 . C . pumila Thunb . subsp . Kuk . in Engl . Bot . Jb . 27 , 1 899 , 55 1 , is one Nov . Holl . Pl . littorea ( Labill ) of the two New Page 27 Zealand members of the section Paludosae of the subgenus Carex . Moore and Edgar ( 1 970) describe Carex pumila as quite unlike other members of the genus in New Zealand having rather coarse tufts of keeled , blue-green leaves up to 40cm long growing from a long� creeping, slender ( c . 2mm diameter) rhizome and large ( c7 x 3 mm) smooth � corky � turgid utricles enveloping each fruit . with a terminal cluster of male Carex flowers pumila is monoecious with usually several predominant ly female flower clusters lower on the culm or stem which may extend up to 20cm above ground (Moore and Edgar 1 970 ) . Carse ( 1 9 1 6 ) found specimens with leaves and culms up to double the lengths described above . In Carex , the flower clusters are morphologically spikes and each "flower" is a spikelet . In New Zealand , Carex pumila is found on coastal sands throughout the North Island (but rarely between Kawhia and Wanganui on the west and Tokomaru Bay and Porangahau on the east ) and South Is land (although rare on the west coast ) and Chatham is lands . Beyond New Zealand the species is recorded from Chile , Korea ; China , Japan ( type locality ) , Lord Howe is land , Australia and Tasmania (Moore and Edgar 1 970) . In Japan, i t is an element of the warm temperate region found in the littoral zone , in sandy places along the seacoast where it is common (Flora of Japan ) . In Australia , i t grows in southern and eastern coastal regions and , an interesting feature � at inland locations in the F linders Ranges and Southern Lofty regions of South Aust ralia (Black 1 978 ) . Carex pumila has received only brief mention in the New Zealand ecological literature principally as a sand-binding species confined to the sand plains of coastal dunelands ( Cockayne 1 9 1 1 ) where it has great abi lity to collect sand ( Es ler 1 969 ) . I ts appearance "contrary to what might be expected from the fact that the plant grows in moist Page 28 hollows" is that of a xerophyte ( Pegg 1 9 1 4 ) . However Cockayne commented that i ts water requirements ( a factor that Esler cites as having has a " critical influence " on species-distribution on the sand plain ) may be greater than those o f the sand�binding dune species ; Spinifex hirsutus and Desmoschoenus spiralis . This view is supported by Esler ' s figure 2 which shows that the local distribution of Carex pumila on a Manawatu sand plain extends below the level of the mean summer water table , unlike the dis tribution of dune species , such as Desmoschoenus , which when found close to the mean summer water table occur mainly as seedlings . Adult plants of these dune species are also found on the sand plains but only as re lics of vegetation of the dunes which covered the area before deflation ( sand removal by wind ) ( Esler 1 969 ) . Morpho logical variation within and between stands of Carex pumila in New Zealand undoubtedly exists ( Moore and Edgar 1 970 ; Carse 1 9 1 6 ; Esler 1 969 ) . Variation in both plant size and seed production occurs within a single patch which suggest ed to Esler ( 1 969 ) responses to differences in micro-habitat . The ability of Carex pumila to stabilize sand is crucial in its role in the development of the sand plain . The series of arcuate hummocks that are found on the sand plain separating the chronosequence of similarly-shade damp hollows in the wake of the receding rear dune are formed as a result of the sand-collecting ability of this pioneer species ( see section 1 . 3 and Esler 1 978 ) . Page 29 1 . 6 Aims of the s tudy The aim of the present study was to investigate the behaviour of Carex pumila from sites representative of the early stages of seral development on a sand plain within the Manawatu coastal dune system. The species plays a pivotal role in the process whereby a catenary sys tem of damp deflation hollows and low-lying dune ridges of increasing age at increasing distance from the rear dune is initiated on these sand plains (Esler 1 978) . In chapter 2 , the growth habit of the species is outlined ( section 2 . 1 ) along with observations on seed and clonal material collected in the field and grown at Palmerston North under a series of conditions . Attempts were made to germinate seed ( section 2 . 2 ) , to observe shoot phenology ( section 2 . 3 ) , to induce flowering ( section 2 . 4 ) , to tes t compatibility ( section 2 . 5 ) and to estimate reproductive assimilation , ie the photosynthetic contribution of female inflorescences to seed weight ( section 2 . 6 ) . Initial field studies ( chapter 3 ) involved a des cription o f the sand plain habitat ( section 3 . 1 ) and of aspects of the behaviour of naturally-occurring Carex pumila topodemes (henceforth called populations ) on the sand plain at closely adj acent sites at increasing distances from the rear dune ( section 3 . 2 ) . Phasic development of Carex pumila on the sand plain is des cribed ( 3 . 2 . 1 ) and the rates of clonal spread ( 3 . 2 . 2 ) and of leaf litter decomposition ( 3 . 2 . 3 ) estimated . The major thrust of the investigation involved an attempt to influence the behaviour of the species in situ by applying to them a series of perturbation treatments ( chapter 4) . It was considered that through such manipulations causal factors in the development of the Page 30 sere could be investigated . The approach taken in the perturbation experiments was to describe the life history strategies of the Carex pumila populations at the various sites under the contrasting experimental regimes in terms of the the distribution patterns of both dry weight and crude total nitrogen to component shoots and organs over time . This was done over four successive summers . During these three years successional trends were expected to become apparent at each site as the resident population aged . Thus , the development of the sere was monitored both in space , across the chronotopose�uence of low dunes and hollows on the sand plain, and in time , over the four growing seasons . The specific allo cation to seeds ( reproductive effort sensu Harper and Ogden 1 970) ; to all reproductive structures ( total rep roductive effort sensu Thompson and Stewart 1 98 1 ) and to rhizomatous and aerial spread was determined at each site for comparison one with another and against that for othe r species in seral communities . The allocation of ni trogen in this pioneer herb was determined as it was considered that this nut rient would be likely to limit plant growth on these recently exposed raw sandy soils . I ts allocation was therefore considered an appropriate currency for the determination of the life history strategy of this species . Energy and percentage lipid determinations were made on various organs including see d . The values obtained gave val idity t o the assumption that biomass measurement would be a good approximation of that of net production . Since the inflores cence of Carex pumila including the dissemules ( morphologically ' utricles ' ) themselves , remain green for an extended period during their development , an es timate was made of reproductive assimilat ion ( see above ) to enable a more complete estimation of the cost of sexual reproduction in this species . Further; since Carex pumila grows in a sandy substrata , the underground parts including both roo ts and rhizomes were readily Page 31 excavated for inclusion in the resource allocation budgets: Page 32 CHAPTER TWO Aspects of the biology of Carex pumila 2. 1 Growth habit 2 . 2 Germination 2 . 3 Seedling morphology 2 . 4 Shoot phenology 2 . 5 Self incompatibility 2 . 6 The contribution o f rep roductive assimilation to seed production .. Page 33 2 . 1 Growth habit The concept of plant architecture and the modular construction of plants is essential to an understanding of all branched structures in plants ( White 1 980 ) . A module ( ' article ' in the original � French) is an axis whose meristem creates all the differentiated structures of a shoot from inception to flowering ( Halle and Oldeman 1 970) . In rhizomatous species , the modular unit is a single aerial shoot and its ass ociated rhi zome and adventitious roots . Growth of such clonal species is usually horizontal and involves the continual serial addition of new modules or ' ramets ' . Collectively , the ramets make up the genetically distinct individual or ' genet ' which is ultimately derived from a single zygote . The adaptive significance of differences in growth forms of perennial plants between those clonal species that expand laterally and often fragment , and those that attain height and the physical integrity is usually maintained ( most trees ) , has been discussed by Harper ( 1 977 ) . Light is seen to play the dominant role in selec tion for �� vertical s t rategy whereas grazing or shortage of water or nutrients ( as in sand dune communities ) are significant in selec tion for the horizontal s trategy. The genetically-determined growth plan of a plant which determines its form is called the architectural model . Over twenty such models have been recognised in tree species (White 1 980 ) whereas in rhizomatous plants Bell and Tomblinson ( 1 980) recognized three basic types : octagonal , hexagonal and linear . These various , yet geometrically precise configurations , determined by characteristic angles of branching and patterns of bud growth are believed to confer certain advantages in terms of substrate exploration and exploitation Page 34 ( Be ll and Tomblinson 1 980) . The adaptive significance of differences in architecture between species within each of the three rhizomatous strategies may be more difficult to explain . For example � the rhizome system of the sand sedge Carex arenaria , which Noble ( 1 982 ) showed to allow for continual extension into bare sand and exploration for scarce nutrients � departs from strict lineari ty by branching at infrequent and irregular intervals at an average angle of 1 5 degrees . The adaptive significance o f this departure from linearity may be appropriate to a mobi le substrata which forms narrow bands ( Bell and Tomblinson 1 980 ) . Variation of the geometric pattern of rhizomatous growth within a clonal species also occurs . Branching pattern may be altered direct ly by alteration of environmental condi tions such as changes in amount of sand accretion (Marshall 1 965 ) or may be seen to differ genetically between population within a species ( Harris 1 970 ) . Further, variation within a clone may exist as seen in the various phases of clonal development , which reflects variation in age of rhizomes (Watt 1 947 ; Noble et al . 1 979 ) . Thus � rhizomes at the leading edge of a clone behave differently from those in older parts of the clone . In this sec tion , the growth habit of Carex pumila is des cribed and compared and contrasted with that of carices growing elsewhere . Observations Carex pumila is a rhizome geophyte in the sense of Raunkiaer ( 1 934 ) . I t possesses a more or less elongated hori zontal axis , called a rhizome , which superficially has a monopodial character. In fact , the rhi zome is a sympodium , an axis made up by the serial addition of the basal portions of successive individual branches or shoot modules . Page 35 Normally , the rhizome is transversely geotropic , growing at right angles to gravity� and at some later stage of its development it becomes negatively geo tropic . Thus , the apical portion of each module grows vertically � pro jecting into the air bearing foliage leaves and in some cases becoming reproductive ( figure 2 . 1 ) . A feature of the sympodial rhizome system of Carex pumila is its ability to change direc tion of growth in the vertical plane � so that its buds cont in(}.o.U� a characteristic depth immediately below are continua at the soil surface . Thus , the rhizome may grow more or less vertically upwards for some distance ( figure 2 . 2 ) and when the shoot apex arrives c lose to the surface the rhizome will revert to a horizontal direction of growth . This behaviour is characteristic of many rhizome geophytes (Raunkiaer 1 934 ) particularly those of sand dunes , peat bogs and other habitats where rapid accretion of the substrate occurs . I t is probably this ability which allows Carex pumila to tolerate moderate levels of sand accretion but which ultimately leads to its demise in the presence of more prolific sand accumulators such as Ammophila arenaria and Desmoschoenus spiralis . Should the rhi zome o f Carex pumila become exposed to the atmosphere � it will grow obliquely downwards until the apex becomes buried in the substrata where it resumes normal horizontal growth ( figure 2 . 3 ) • Through such changes in direc tion of growth a loop formation may occur . Two types of shoo t module can be distinguished in Carex pumila : one associated with long , the o ther with short or dwarf rhizome branches ( figure 2 . 2 ) . The adnation o f long rhizome modules makes up the basic linear pattern of rhizome architecture . Branching of the long rhi zome occurs at irregular intervals along this linear structure A d r . 05 ) differences were found between any of the four treatment X genet combinations for mean number of seeds per shoot ( table 2. 2 ) , this reduction in utricle weight in response to bagging resulted from the reduction in mean dry weight per utricle ( table 2. 3 ) . Each mean in table 2 . 3 ( dry weight per utricle ) was found to be significant ly (P < . 05 ) different from each of the other means . - I Page 47 Table 2 . 1 Dry weight of utr icles per shoot (g ) Mean SD n ... "'!S Self-pollinated shoots . 21447 . 0962 11 Open-pollinated shoots . 28934 . 0943 25 Means di ffer , P < . 05 . Table 2 . 2 Number of seeds per shoot Mean SD n Gl Self-pollinated 9 8 . 5 21 . 73 4 Gl Open-poll inated 96 . 0 20 . 57 14 G2 Self-pollinated 80 . 0 27 . 99 7 G2 Open-poll inated 7 2 . 82 22 . 07 11 Means not d ifferent . p > . 05 Table 2 . 3 Dry weight per utr icle (rrg) Mean SD n Self-pollinated 2 . 34 487 . 57 8 5 11 Open-poll inated 3 . 38102 . 5700 ?..--:::> Means di ffer , P < . 01 Although dry weights of ind ividual utr icles • . .;ere not measured , it was apparent that the reduction in mean utr icle weight brought about by bagging resulted from the fa ilure of a s ignificant ( P < . 001) proportion of utricles on each shoot in this treatment to produce mature seed ( table 2 . 4 ) . Utricles of Carex pumila enclose a s ing le pistil containing a sol itary ovule � When the ovule is ferti l ized and the seed mature , the enveloping utr icle is thick , corky and turgid, whereas unfertil i zed pist ils remain thin and flaccid . Page 48 Table 2 . 4 Percentage of utricles containing mature nuts . Mean SD n 1./. 6 Self-pollinated shoots 60. 84 �og » 1 1 5 . 22 Open-pollinated shoots 94 .00 � 25 Discussion The resul t of this experiment suggest that Carex pumila is part ly self-compatible . The percentage of utricles with mature seed and consequently the mean dry weight per utricle was reduced by about 30% on self-pollinated shoots in the glasshouse . If self-incompatibility is the total fai lure to set seed on selfing then Carex pumila cannot be considered self-incompatible . However � such a strict definition of self-incompatibility may no longer be warranted considering the accumulating evidence of mul tigenic systems determining incompatibi lity in plants ( Lewis 1 979 ) . Certainly , the strength of the self-incompatibility reaction varies in different taxa ( Ganders 1 979 ) and in some it is very weak with a high percentage of seed set with selfing , as in Nymphoides ( Ornduff 1 966) . The results of the present experiment suggesting only weak self-incompatibi lity in this New Zealand member of the section Paludosae are at variance with those of Faulkner ( 1 973 ) for north west European species of section Acutae of the same subgenus , Carex. Faulkner, by enclosing the female inflorescences on test plants with a glass ine s leeve so that only by artificially applying pollen were any seed set � found that self-pollination resulted in extremely low seed set � between 0-20% in all but one plant tested . The higher seed set in self-pollinated shoots of Carex pumila than in other sel f-incompatible members of the subgenus may be the result of apomixis ; although this is unlikely considering the observation that isolated fertile shoots of Carex Page 49 pumila on the sand plain at Tangimoana set few seed as reflec ted by their reduced mean seed weights ( see Ch . 4 ) . Apomixis is unknown elsewhere in the genus ( Handel 1 976 ) . Self-pollination may not be avoidable in Carex pumila in the field despite features , both structural and phenological ; that are normally associated with outcrossing : monoecy , protogyny and the cons picuous locat ion of the male spike on the terminal portion of the culm . Although stigmas are exserted before anthers there is overlap between pollen release and receptivity of the stigmas on the same shoot ( section 2 . 4 ) . Even the transfer of pollen between neighbouring Carex pumila shoo ts in the field may not invo lve cross-pollination since such shoots a re likely to be part of the same genet in this clonal species . In Anthoxanthum odoratum , a pro togynous , self-incompatible wind-pollinated grass species , Snaydon and Davies ( 1 976 ) showed that pollen flow rapidly declined t o a low level at 2 metres . Protogyny and protandry in self-incompatible species is considered by Connor ( 1 979 ) � unnecessary pheno logical action, except perhaps as legacies of the past or to reduce pseudo-compatibility. The latter may be the explanation of the combination of pro togyny and weak self-incompatibility in Carex pumila , a species that must be described as essentially allogamous . Page 50 2 . 6 The contribution of reproductive assimilation to propagule production The photosynthetic contribution of reproductive structures ( reproductive assimilation) to seed- or grain-filling has been the subject of numerous investigations particularly in economic species including wheat , barley , maize and rice. Quantitative estimates of this contribution vary conside rably not only according to genotype but also to the methods employed in their determination. In wheat ; for example , shading , excision, L 1 4-C jC02 feeding and infra�red gas analysis techniques have all been used (Thorne 1 966) with estimates of the contribution of ear photosynthesis to final grain weight ranging from 1 O% to 60% (Evans et al . 1 975 ) . Although shading and excision continue to be used as means of es timating the contribution of plant parts to propagule production, they may give somewhat biased estimates since these treatments are likely to have effects besides reducing or blocking photosynthesis in the shaded or excised tissues . For example , shading and excision may result in ( 1 ) a change in temperature , ( 2 ) a reduction in transpiration , ( 3 ) an increase in respiration or ( 4 ) the induction of compensatory changes in assimilate dis tribution within the plant. Thus propagule weight may be affected through treatment effects on translocation and photosynthesis in othe r organs . Using this technique Archbold ( 1 945 ) concluded that grain growth in wheat is largely based on concurrent pho tosynthesis rather than previously accumulated reserves . Thus , shading or defoliation after anthesis usually reduces wheat grain yeild ( e . g . Bremner 1 972 ) although this may not always be the case ( e . g. Fischer 1 975 ) . Evans et al . ( 1 97 5 ) suggest that 90-95% of the carbohydrate in wheat grains Page 5 1 comes from photosynthesis after anthesi s . In Proctor barley on the other hand Briscoe et al . ( 1 975 ) have estimated only 47% of final grain weight is contributed by post-anthesis photosynthesis ( 34% was attributed to flag leaf photosynthesis and 1 3% to reproductive assimilation i . e . photosynthesis in the ears ) . In Ambrosia trifida , an annual colonizer of disturbed habitats in Illinois , Bazzaz and Carlson ( 1 979 ) have estimated , by means of excision experiments and measurements of C02 flux , that reproductive assimilation accounted for 4 1 % and 57% , respectively, of the carbohydrate required to produce male and female inflorescences on intact plants . Thus , for the accurate analysis of reproductive effort , that is the cost to a plant of producing reproductive st ructures , reproductive assimilation should be taken into consideration (Bazzaz and Carlson 1 979 ) . Reproductive assimilation is likely to be signi ficant in sedges in which the sexual reproductive propagules contain chlo rophyll and remain green for an extended period during their growth and development . The following is an experiment designed to es timate the photosynthetic contribution of female inflorescences to propagule filling and reproductive effort in Carex pumila. The experiment followed the elaboration of "seed" (nut enclosed in a green pe ricarp or utric le ) by pollinated female inflo rescences on intact shoots under various shading and defoliation treatments , imposed after anthesis at the time rapid increase in seed weight was expected . Methods and materials Nine 30 x 30 cm turves were removed on 1 3 October , 1 979 from an area on which Carex pumila was growing vigorously in monoculture on the sand plains near Tangimoana ( 1 km north of the study area ; see Page 52 section 3 . 1 ) to an unheated glasshouse in Palmerston North . Each turf included between 9 and 52 fertile shoots , all at a stage of development around male spike emergence but prior to stigma protrusion and anthesis . All vegetative shoots were cut at ground level and removed and any that appeared subsequently were similarly removed . As anthesis was reached , shoots were tagged and assigned to one of four groups for subsequent treatment . An attempt was made to include all four treatment groups on each turf . Turves were then removed to a controlled environment room at the DSIR Climate Laboratory at Plant Physiology Division , Palmerston North . Conditions in the room were : Temperature 1 8 °C � 0 . 5 °C constant day and night Relative humidity 52/81 + 5% day/night with corresponding Vapour pressure deficit 1 0/4 mb day/night Photosynthetic irradiance 1 55 � 5 w m-2 ( equivalent to a photosynthetic photon flux dens ity of 630 � 20 uE m-2 sec - 1 ) Day length 1 4 hours ( full lights ) Lamp combination 4 x 1 000W Sylvania "Metalarc " high pressure discharge lamps plus 4 x 1 000W Phi lips tungsten halogen lamps The sandy substrate was kept permanently above field capacity by maintaining the turves in water-filled trays . Page 53 Between seven and fourteen days after anthesis at the time rapid increase in seed weight was expected , the following shading and defoliation t reatments were imposed : ES : Terminal section of the culm including the female inflorescence was shaded SS : Green leaf laminae were removed at the junction between sheath and lamina and the shoot , excluding the terminal ferti le section with attached female and male spikes , was shaded D Green leaf laminae were removed as above and the entire shoot was shaded C Shoots were left untreated ( control treatment ) . These treatments allow two estimates to be made of the photosynthetic contribution of the inflorescence to firstly by the comparison of treatments the ES increase in seed weight and C and secondly by comparison of treatments D and SS . No attempt was made to prevent the translocation of stored assimilate from rhizomes to seeds . Shades were made from black cardboard folded to form a 1 00mm diameter cylinder enclosed at one end and covered with aluminium foil ( figure 2 . 1 2 ) . Cylinders were attached to stakes pushed into the turves . A maximum of four shoots were enclosed within any one shade . An aluminium foil covered black cardboard disc with holes cut to accomodate the culms was used to cover the basal end of each cylinder in treatment ES . 2 . 1 2 !::€ foliation and shad ing treatrrents . Left to right : Tre3. tment C, control ; treatment 3S , shoot defol iated and stem shaded ; treatment �) , i nflorescence shaded ; completely shaded . trea trr:ent D, shoot de foliated and 53q Page 54 At the beginning of the experiment between 1 6-22 November 1 979 all plant material removed as part of the treatment was weighed along with a set of control shoots from each turf . These shoots ( aerial parts only) were divided into their consti tuent parts � dried at 80oC for 24 hours and weighed � and numbers of female spikes per culm and seed per spike counted . The experiment was terminated on January 29-30 1 980 by which time most shoots were dead . The sandy substrate was removed by washing and complete shoot modules ( including the underground fractions ) were separated , divided into constituent parts ; dried as above and weighed . Numbers of seeds per inflorescence were counted . Statis tical Analysis Since a variable number of fertile shoo ts were found on each of the nine turves collected , equal numbers of shoots per treatment could not be ensured. Apart from the controls most trea tment groups possessed three shoots per turf ( average 3 . 74 , range 2 to 7 shoots ) . Means were obtained for each treatment on each turf for ( 1 ) dry weight per module for each shoot component , ( 2 ) number of seeds per shoot � ( 3 ) 1 000 seed weight and ( 4) the proportion of the total branch o r total aerial shoot weight as propagules . Comparisons o f these means were carried out by analysis of variance or Students t-test . Where significant treatment differences were revealed by use of the Students t-tes t , no more sophisticated statistical pro�edure was necessary. Results and discussion deemed Figure 2 . 1 3 shows the effect of post-anthesis shading and defoliation on the distribution of dry weight in aerial shoots of Dry v�e i ght (g ) p e r s h o o t 1 . 5 i F i gure 2 . 1 3 . D i str i but i on o f the a e r i a l d r y we i ght (grams/shoot ) o f f e rt i l e shoots o f Carex p um i l a t o c o mp o nent s tructures �nder contrast ing d e f o l i a t i o n shad ing treatments (P = pr e-exper i ment contra l ; C = untreated contra l ; ES = i n f l orescence shaded, i ntact p l ant ; ss = v e g etat i v e p o rt i o n s h a d ed ; de f o 1 i ated p l a nt ; 0 � who l e s ho o t shaded , d e f o l i a t e d p l ant ) . Vert i c a l b a r d e n o t e s + v a r i a n c e o f me a n s . p Seeds Anc i l l ury reproduct i ve structures Lam i na e Cu l m a n d l e a f s h e aths C ES SS Treatment 0 Page 55 Carex pumila . The mean aerial dry weight per shoot increased over the duration of the experiment in all four treatments compared to the pre-experiment control ( treatment P) . This increase was shared between the fertile and vegetative components of these reproductive shoots . The increases in each vegetative component ( that is , in culm and sheaths and in leaf laminae ) and in the major fertile component ( that is , in seeds ) were highly significant in all treatments ( P < . 01 ; table 2 . 5 ) . Figure 2 . 1 3 shows a greater proportional increase occurred in the fertile portion than in the vegetative portion of these shoots (P < . 00 1 ; table 2 . 6 ) . Figure 2 . 1 3 suggests that at the termination of the experiment only defoliation and shading of at leas t the vegetative portion of the aerial shoo t ( treatments SS and D ) had any effect on reducing the mean dry weight of shoots . Inflorescence shading ( treatment ES ) had no effect on this parameter. When aerial shoot components were divided into their constituent parts , however , the only significant treatment effec t was a reduction of the dry weight of seeds per shoot in response to defoliation and complete shoot shading ( treatment D, P < . 05 , figure 2 . 1 4 and table 2 . 7 ) . This effect was seen to be due to the reduction in mean dry weight per seed ( figure 2 . 1 5 ) ; the number of seeds per inflorescence was similar in all treatments ( table 2 . 7 ) . The uniformity in seed number per shoot was expected since the treatments were no t imposed until one to two weeks after anthesis at a time that seed numbers would have been determined . Earlier imposition of the shading/defoliation treatments was avoided since this has been seen to affect seed number ( Bremner 1 972 ; for wheat ) . Table 2 . 5 Analysis of variance Analysis of variance Culm and sheaths d . w . /shoot Leaf laminae d . w . /shoot Seeds d . w . /shoot Reproductive effort Seed number/shoot Seed weight ns = not significant F 4 . 62 5 - 276 8 . 355 6 . 6704 0 . 4466 22 . 809 Page 56 Level of significance P< . 01 P< . 01 P< . 00 1 P< . 001 ns P< . 001 Table 2 . 6 Reproductive effort : the proportion of dry weight of (a ) aerial shoot and (b ) total (aerial plus underground ) branch modules as propagules ( "seeds" ) . Reproductive effort ( % ) Treatment (a )Aerial (b )Total Control ( C ) 32 . 98 � 1 . 40 27 . 53 � 1 . 32 Inflorescence shaded (ES ) 32 . 94 + 1 . 59 24 . 27 � 1 . 95 Vegetative portion shaded ( SS ) 34 . 52 � 1 . 9 1 29 . 48 + 2 . 42 Complete shoot shaded (D ) 29 . 23 + 1 . 72 25 . 59 + 2 . 1 1 Pre-experiment control (P ) 2 1 • 94 + 1 . 89 Page 57 Figure 2 . 1 4 Female spikes at the end of the defoliation and shading experiment . c ss , ES f D ' I Increase in temperature from 1 5/ 10 ( day/night ) through 21 / 1 6 °C to 30/35 ° C has been shown to decrease both the duration of propagule fi lling and final propagule weight in wheat ( Sofield et al . 1 974 ) . The main temperature e ffect on duration of propagule filling in this species is attributed to day as opposed to night temperature (Evans et al . 1 975 ) . Phytotron experiments have also shown that temperature rather than radiation has the predominant effect on duration of grain filling in wheat (Evans et al . 1 975 ) . The temperature regime in this 0 experiment , namely 1 8 C constant day and night , was chosen to ensure a relatively prolonged duration of propagule filling and so maximal final propagule weight . Air temperature (and C02 concentration ) may have been increased above ambient within the enclosed cylinders ( figure 2 . 1 2 ) , especially in treatment D , despite the covered openings which were designed to a llow the passage of air but only a minimum of light . This may accoun"t at least in part for the reduceu seed ·,reight and shortened duration of propagule filling (as seen in numbers of 3 - Dr y We i g h t (m g ) p e r 2 seed 1 0 F i gure 2 . 15 Mean dry ��e ight (mg) per seed o f Carex p um i l a under contras t i ng d e f o l i a t ion a n d s h a d i n g treatments (see f i gure 2 . 13 f o r key t o treatments) . Ver t i c a l b a r i n d i c a tes + var i an ce o f means . I .,-...L. p T I I I 1 T 1 C ES SS Treatments T 1 0 51� Page )d shed s eed ) in treatment D at the end of the experiment compared to the other treatments . Table 2 . 7 Parameters of shoots under contrasting defoliation and shading treatments . ( a ) Treatment Contro l ( C ) Spikes shaded (ES ) Culm shaded (SS ) Shoot shaded (D ) Pre-expt control ( P) ( b ) Treatment Control ( C ) Spikes shaded ( ES ) Culm shaded (SS ) Shoot shaded (D ) Pre-expt control (P ) Mean (� SE ) No . of seeds Seed weight Seed weight per shoot ( mg ) per shoot ( g) 94 . 7 � 8 . 2 a 3 . 6408 � • 245 1 a • 347 � . 0439a 87 . 7 + 9 . 9 a 3 . 56 1 5 � 89 . 1 + 1 1 • 6a 3 . 407 1 + - . 1 48 1 a . 21 6 1 a . 332 � . 0352a . 293 + . 02 93a 9 1 . 7 � 8 . 7 a 2 . 441 4 + . 1 842b . 2 1 5 + . 0226b 76 . 9 + 8 . 6 a 1 . 2572 + . 1 0 1 2d . 1 01 + • 01 66d Mean (� SE ) Culm and sheath Rhizome DM per DW per shoot ( g ) dwarf segment (g ) - 3439 � . 0366a . 0663 � . 0095a . 2907 � . 05 1 8ab . 0554 + . 0075ab . 2900 � . 02 1 3ab . 046 1 + . 004 b . 2378 � . 023 1 b . 0445 + . 0052b . 1 36 5 � . 023 1 c Within each column, letters indicate Duncan ' s multiple range test ; a>b , P< . 05 ; b>c , P< . 0 1 ; b>d , P< . 001 . Page 59 Maximal seed weights achieved in this experiment were found on control shoots ( fi gure 2 . 1 5 ; mean 3 . 6408 � 0 . 245 1 mg per seed ) . Est imates of the various contributions to this final seed weight which treatrnent have been calculated by comparisons of the treatm' means from figure 2 . 1 5 � are shown in table 2 . 8 . At the time of imposition of the treatments � mean dry weight per seed was 1 . 2572 � . 1 0 1 2 mg � a lit t le more than one-third of the final weight achieved in the control shoots ten weeks later. Seed weight increased over the duration of the experiment in all treatments � including that in which the shoots were defoliated and the entire shoot shaded ( treatment D ) . This latter increase which corresponds to about one-third of final seed weight in the control shoots ( C ) must be att ributed to translocation of pre-formed assimilate � since photosynthesis in treatment D shoots may be assumed to be negligible. Table 2 . 8 Contributions to propagule weight in Carex pumila Proportion of control mean seed weight ( % ) Untreated control ( C ) 1 00 . 00 Inf lorescence photosynthesis ( a ) ( C - ES ) 2 . 1 8 ( b ) ( SS - D ) 26 . 52 Vegetative pho tosynthesis ( C - SS ) 6 . 42 Whole shoot photosynthesis ( C - D ) 32 . 94 Translocated/pre-formed assimilate ( D - p ) 32 . 52 Pre-experiment control ( P ) 34 . 53 �Gge 60 The source of pre-formed assimilate in this experiment mus t ultimately be the underground fraction rather than the stem� as suggest ed by Briscoe et al . ( 1 975) for barley. The mean dry weight of stems ( and sheaths ) actually increased over the duration of the experiment in all t reatments compared with the pre-experiment control ( treatment P , figure 2. 1 3 ) . The only possible source of assimilate leading to such an increase in treatment D mus t be translocation from the rhi zomes . Unfortunately , no data are available for pre�experiment rhizome weights for comparison with those at the end of the experiment . However , at that time , dwarf rhizome branches was obtained for the lowest mean dry weight of treatment D � in which the drain of rhizomatous reserves was expectedly greatest ( table 2 . 7 ) . The remaining ( one-third ) contribution to final seed weight is that of concurrent photosynthesis during the post-anthesis period . Two es timates of the photosynthetic contribution of the inflorescence to propagule production were obtained by calculating the reduction in seed weight with inflorescence shading , firstly , in intact shoots ( comparison of treatments ES and C ) and , secondly , in defoliated shoots ( comparison of treatments SS and D ) . The discrepancy between these two estimates ( table 2 . 8) indicates the ability of ferti le shoot modules of this c lonal species to c ompensate for reductions in photosynthesis in various component organs in the development of seeds . This ability to compensate for reductions in photosynthesis would suggest that , in an intact plant � the estimate o f the contribution of pre-formed assimilate to final seed weight of about one-third may be an over estimation and therefore the estimation of concurrent photosynthesis to be correspondingly conservative . Likewise , the estimate of reproductive assimilation over this post-anthesis period to final seed weight ( of 26%) may be an Page 6 1 underestimation . The similarity of final seed weight in treatments C and SS could be interpreted as an almost complete reliance by the seeds on reproductive assimilate with lit t le demand for leaf or culm assimilate . A simi lar conclusion was reported by Evans and Dunstone ( 1 970) for low-yielding primitive wheats . However� the acceptance o f a value greater than 26% for the contribution of post-anthesis reproductive assimilation to propagule filling would require further experimentation involving more direct measures of reproductive assimilation and inflorescence respiration (by infra red gas analysis ) and by the use of [ 1 4-C ]C02 to trace the fate of assimilates formed in different parts of the sedge shoot . CHAPTER THREE Initial field studies 3 . 1 Description of the habitat 3 . 1 . 1 The study area 3 . 1 . 2 C limate 3 . 1 . 3 Wind speeds 3 . 1 . 4 Vegetation cover 3 . 1 . 5 The water table 3 . 1 . 6 Soil water profile 3 . 1 . 7 Soil nutrient and organic matter status 3 . 1 . 8 Nitrogen fixation 3 . 2 Carex pumila on the sand plains 3 . 2 . 1 Phasic development 3 . 2 . 2 Rates o f clonal spread 3 . 2 . 3 Leaf litter decomposition Page 62 Page 63 3 . 1 Description of � habitat 3 . 1 . 1 The . study � The study area was established on a sand plain near the northern end of the Manawatu sand dune system ; 3 km south o f Tangimoana beach on the west coast of the north is land of New Zealand (40 degrees 1 9 ' 38"S , 1 75 degrees 1 4 ' E ; figure 3 . 1 ) . The area was situated at the landward end of the sand plain within 500 m of the coastal foredune ; at the base of the windsweep of a landward moving parabolic rear dune and sandwiched between two parallel longitudinal dunes ( the trailing arms of the parabolic rear dune ) ( figure 3 . 2 ) . The dunes framing the study area rose 6-8 m above the level of the sand plain . The s tudy area straddled a putative chronotopological series of damp deflation hollows ( dune slacks ) separated from each other by low arcuate dune ridges ( figure 3 . 2 ) presumed to have been formed in the wake of the receding rear dune ( Es ler 1 978 and section 1 . 4 ) . The youngest hollow of this catenary system was c losest to the rear dune with successively older hollows at progressively greater distances moving seawards . Five field sites were chosen in December 1 977 at progressively greater distances from the youngest ( terminal ) hollow ( figure 3 . 3 ) to represent sites of successively increasing age or seral maturity. The sites , which were numbered according to their relative distance from the terminal hollow , were : Site 0 (SO ) in the embryonic terminal deflation hollow was bare of vegetation in December 1 977 and remained so until the end of the following summer ( 1 978-79 ) . Site 1 ( S 1 ) on the edge of the adjacent low dune ridge F i gu r e. 3 . i L o ca t i o n o f r he s tu d� a re a \ I / I . \ l · ' , , � Ta ..... g imo � na Fa rm Se.H ie.rne.nt . .... ) / I . •/ . \ I . ) . l'- . . ) t�t STUDY AREA . \ . I ; ,.. 1 / I · \ / ( • 1 . I . · / ( \ . I / ... _ ' / - ; Lake Pvkepuke Un s t a b le � and .L ..... .. 0 z " L 1.!5 I . ( ' I of the. M a naw<3tu S a n d clune s'jste.tYI 0 2 .3 Sea I e of I< i I o me t re s I N S E T : L o c a I 1 t y rna p 5 3 . 2 View across the study area from the longitudiml dure , looking north. Ccast on extreme left. -- -- ------ �,. 0' � F i g ... H'e 3 .3 P l o r'\ v 1 ew of the sh-ld:J area shovv i ng 1-he c..-h r- o n o to p ose'\V\eVlc e. o f l, o l \ o ws and l ow d u.f'les Rea r d une so 0 / .S1 Te. rm 1 1'"\ 3.\ def' \ at- 1 or. . h o l l ow Y o ung S 2 0 \ ow ouf\e O l d � e f-1 � t i ov-1 ho \ l ow · s4 0 o lo 10 3o 4D 5o rv>e-tres o l d e r d e.f \ a t 1 o�" · ho l \ ow 0 1 c) er I o vv d uf'l8 · L�-te:.ra l ho\ I ow \ No r-th \ o---- � Page 64 included the front of Carex pumila rhizomes growing into the uncolonized terminal hollow . By the summer of 1 978-79 S 1 was about 5m behind this advancing front and sand had accumulated at the site so that the surface was approximately 30 cm higher than that of the terminal hollow. In this way � the low dune continued to encroach upon the terminal hollow ( figure 3 . 4) . By October 1 979 SO had been incorporated into the low dune and rep resented a similar juvenile phase of development ( see section 3 . 2 . 1 ) to S 1 nearly two years previously. Sites 2 , 3 and 4 ( S2 , S3 and S4 ) were situated on more stable parts of the sand plain, within the two successively more distant ( older ) damp def lation hollows which became ponds during the winter ( figure 3 . 4 ) . Within these hollows extreme variation in plant vigour and seeding capaci ty was found , typical of Carex pumila on the sand plains of the Manawatu coastal dune sys tem (Es ler 1 969 ) . The sites were chosen to represent the extremes in this variation range ( figure 3 . 5 ) . S2 and S3 were found within 50m of each o ther on the old hollow behind the low dune ridge on which S 1 was situated (figure 3 . 3 ) . S 4 was found in the older hollow further from the rear dune and separated from S2 and S3 by another , older , low dune ri dge ( figure 3 . 3 ) . 3 . Lf ?anorarra of the study area over tirrB . Top eo bot tom : 1 977 ; 1'13.y 1 979 ; Cc tober 1 980 ; August 1 98 1 . - _r 3 . 5 The field sites i n December 1 977 . Top right : S1 on edge of the low dure . Bare moist sand of termim.l hollow ( SO ) in backgroum . Top left : S2 in old deflation hollow, showing dense stand of_ Carex pumila . Bottom right : S3 in old deflation hollow. Bottom left : S4 in older deflation hollow. Note depauperation of Carex pumila on older deflation hollow s ites . Page 65 3 . 1 . 2 Clic:Jate The cl imate of the Manawatu coastal sand country is characterised by warm summers and mi ld winters , a low mean annual rainfall dis t ributed througho�t predominant ly west to the year and frequently northwest winds ( Robertson 1 9 59 ) . gale force Figure 3 . 6 shows the mean monthly rainfall ( 1 9 72- 1 9 79 ) at the Tangimoana Farm Settlement , a site approximately 2 . 5 km north-east of the s tudy area ( figure 3 . 1 ) . February ( late summed was the driest month wi th an increase in rainfal l in April and May ( autumn ) . July (mid-winter ) was the wettest month . The sub sequent decrease in mean monthly rainfall th rough spring to the late s ummer low in February shows fluctuations in both October ( spring ) and December ( summer ) . The mean annual rainfall recor ·ded at this sit e ( 1 9 7 2- 7 9 ) was 899 + 1 5 5 nm , typ:i.cal for thi s coastal region . Air and soil temperatu t·es reach a maximum in late summer ( January-February ) and a minimu� in mid winter ( July ) , although there is no great variation in mean monthly temperature ( figure 3 . 6 and CO\·Jie et al . 1 9 67 ) . On the sand p lains , a wider range of soil temperatures can be expected upon the sparsely covered l ow dunes than in the vegetated hollows where Cockayne ( 1 9 1 1 ) noted surface temperatures in excess of 38 degrees C as the surface sand dried out during summer . The winds in the region are variable . The predominant often �trong and gusty winds of the Manawatu coast ( figure 3 . 7 ) correspond to the direction of movement of the rear d une and the orientat ion 0f the longitudinal dunes ( namely , from west of north-west to eas t of sout�-eas t ; figure 3 . 2 ) . Figure 3 . 6 shows spring and summer to be the windiest ' period o f the year with mean dai ly windspeeds , based on F i g u r e 3 . 6 ��e a n month l y ( a ) r a i n f a l l , (b ) t emp era ture and (c ) w indspeed , at s i tes on the l�anawatu p la ins . 140 ea ) R A I NFALL 120 [ 1972- 1979 mean at Tang imoana Farm Sett lement .§ 100 � 80 r--1 � so L c ·rl m a: 40 20 ---- / · - . / 0�---------------------------------------------- 20 (b ) Tclv�PERA TURE 1978- 1980 mean at Aorang i u (f) ID ID - · - c... 01 ID 15 '0 / ID c... ::J / 4-l CO c... 10 ID D. / E ID 4-l c... ·rl • IJ'1 ---1 -r- � Page 75 Table 3 . 3 Mean summer water table levels at six sites on the sand plain , November 1 978 to February r979 . Site so Terminal hollow S 1 Low dune ridge S2 Old deflation hollow S3 " " " Older low dune ridge (between S4 Older deflation ho llow Distance below surface (mm ) 466 . 1 748 . 9 393 . 4 366 . 5 S3 and S4 ) 746 . 6 360 . 7 The elevation o f the old low dune and the young low dune at S1 was remarkably similar , despite the variable heights of these low dunes above the adj acent deflation hollows . .�ophila arenaria which has the ability to build the low dunes higher (Esler 1 978 ) was absent from bo th the old low dune and from S1 • This species however was present elsewhere on the young low dune forming an arcuate band ( figures 3 . 2 and 3 . 4 ) where by summer 1 978-79 it had accumulated sand to a greater height than at S 1 at which Carex pumila was the only species present . Ranwell ( 1 972 ) showed that the overall shape of the water table in a large isolated dune system is dome-shaped as a result of the more rapid drainage on the periphery of the dune system. Should the water table in the Manawatu sand dune system conform to this shape , the scale of variation over the study area is likely to be only small since the field sites were all within 1 00m of each other. Page 76 Generalised seasonal changes in water table levels on the study area were evident ( figure 3 . 1 0 ) . The water table levels were high during the winter often resulting in surface flooding in the hollows � steadily dropped during spring and summer with some fluctuation to a low in February-March , and rose more sharply in the autumn . This autumn rise in water table levels coincided with the death of the fertile shoot cohort of Carex pumila. Each winter, with the exception of 1 978 when the first half of the year to mid winter was uncharacteristica lly dry ( rainfall was 2 1 8mm compared with the 1 972-79 mean of 409mm ) , the water table in the hollows rose to the surface so that there was either surface puddling ( 1 980 ) or actual flooding ( 1 979 and 1 981 ) . The annual range in water table levels was between 600 and 700mm , approaching that described by Ranwell ( 1 959 ) for slacks in a dune system on Anglesey in North Wales where the annual rainfall was also similar to that on the Manawatu coas t . Short term fluctuations in water table levels were super-imposed upon these seasonal trends and show a simi lar pattern to that described by distribution o f rainfall ( figure 3 . 1 1 ) . For example between 1 0 and 1 4 November , 1 978 , the water table rose 1 30- 1 35mm in the two more mature o ld deflation hollow sites . The water table level rise over this period was not as evident on the low dune and on the immediately adjacent old hollow ( 32 ) where sand accretion occurred concurrent ly on these parts of the sand plain . The sharp rise in the water table level over this period coincided with the thunder storm of 1 1 - 1 2 November 1 978 when 27 . 4mm of rain fell . Between the 10 and 1 4 November 30 . 8mm of rain was recorded . Given a pore space of 40% for the soil in the older hollows ( see section 3 . 1 . 6 ) this rainfall accounted for only about one-half the observed rise of the water table in the older hollows . The remaining rise mus t have been the result of .. Cl) 0 CO - L :J Ul E 0 L - ,.-... E E Cl) 0 c CO +J Ul ·rl 0 E E .......... r-i r-i CO 100 0 - 100 -200 -300 -400 -500 -800 180 180 140 120 100 - 80 c ·rl CO a: 60 40 20 l6o, F i g u r e 3 . 1 0 W a t e r t a b l e l e v e l s i n t h e o l d d e f l a t i on h o l l o ws o v e r t i me (a v e r a g e d o v e r 52, 53 a n d S4) WATER 1979 1978 '\ s TABLE \ I I I \ \ \ \ \ � I ' LEVEL I I I \ I \ I \ \ \ \ \ I I I I I . - I '\ / '. I I \ ............. ............. 0 N 0 J F I I I I I I I I M A M J J A '\ . - · 1979 " .. " ... " '-, · 1980 s 0 F i gu r e 3 . 1 1 Tot a l month l y r a i n f a l l a t T ang imo a n a F a r m S e t t l e m e n t # , 1978- 1 980 . ( #2 km f ur ther i n l a n d and 1 . 6 km n o rth o f t he s tudy a rea ) R A I NFALL ' I I \ I \ I \ I \ I I I I I I I \ I \ I \ I \ I \ I \ I \ : \ / \ I \ I \ I \ I \ I \ I \ 1878 . -r- . . I / '\ I I 1878 . 1 \ I \ I \ I I I \ I I .. .... . 1979 . --- . 1980 0 �--------------------------------------------------------�----- s 0 N 0 J F M A M J J A s 0 topogenic water accumulation on the sand plain from high parabolic dune . 3 . 1 . 6 Soil water pro files the Page 77 surrounding SOI:Utdiftg At the end of the winter ( September ) 1 978 ten replicate soil samples were taken from each of five sites on the sand plain to determine the soil moisture profile . Sample cores ( 20mm diameter ) were taken at successive 1 0mm depths until the water table was reached . Each 1 00mm-long sample core was removed , placed into a plastic bag , sealed and labelled and taken to the laboratory for weighing, drying at 1 05 degrees C for 1 6 hours and reweighing. The volume of dry sand in each sample was also measured . The dry sand from each of the three old deflation hollow sites was placed in a 500 degree C oven for 24 hours to determine the organic matter fraction and then passed through a series of seven sieves ( 63-4000 pm ) to determine the various partic le size fractions . A hand magnet was used to remove the ferreous fract ion . The water content of the sand down the profile at each site , expressed as a percentage of the volume of wet sand , is shown in figure 3 . 1 2 . At each of the three sites within the old deflation hollows , only one 1 00mm-deep soil core sample was required to reach the water table . At these sites where the soil from the surface to the free water-table level was completely satu rated , water made up more than one-third of the volume of wet sand . This volume is called the pore space of this sandy soil . The water holding capacity of the soils across the study area was expectedly low as a resul t of the extremely low organic matter contents ( table 3 . 4 ) of these raw or relatively undeveloped soil s . 7l a. F igure 3 . 1Z So i l mo isture pro f i l es at f ive s i tes on the sand p l a in, in September 1979 . Page 78 Table 3 . 4 Percentage of total oven dry weight of various particle size classes of sand at 0- 1 00 mm depth at three deflation hollow sites on the sand plain near Tangimoana . Percentage ( % ) of total dry weight Particle size class Cum ) S2 S3 S4 >4000 1 000-4000 . 82 . 44 . 3 500- 1 000 1 . 1 7 1 . 4 1 . 4 250-500 37 . 72 1 1 . 34 1 4 . 6 1 25 -250 55 . 1 4 79 . 2 74 . 3 6 3- 1 25 4 . 79 ( 1 . 5 1 ) 6 . 39 ( 1 • 65 ) 8 . 2 ( 2 . 4 ) <63 • 22 . 6 . 6 organic matter • 1 6 . 6 . 7 ( Percentage of ferreous fraction in brackets ) The difference in water content of the saturated soils in the def lation hollows between S2 and S3 and S4 ( figure 3 . 1 2 ) was att ributed to the difference in mechanical composition of the sand ( table 3 . 4 ) . At S3 and S4 where there was a greater percentage of smaller sand particles than at S2 , there was also a reduced pore space as indicated by the reduced water content of the saturated soil J.e+ \o.tiol'! ( figure 3 . 1 2 ) . The ferreous fraction of the deflon hollow soils was confined to that fraction of "fine " sand between 63 and 1 25 pm diameter and accounted for from 1 . 5% to 2 . 4% of the total soil dry weight . This is comparable to the 3% "opaques " quoted by Gibb ( 1 977) for beach sample at Tangimoana Beach � a value considerably less than . I the 80-90% opaques in beach sand further north on the west coast of the north island at Patea . 3 . 1 . 7 Soil nutrient and organic matter status Frequent , seasonal or permanent wat erlogging on the sand plains of the Manawatu coastal dune system; largely as a result of topogenic water accumulation has led to the formation of hydromorphic soi ls on the low-lying areas c lassified by Cowie et al . ( 1 967) as Hokio series . Soi ls on the associated dunes derived from the same wind-blown parent material are mapped as Waitarere sand . Soils of the plains in which a more diverse and vigorous vegetation develops ( Es ler 1 969 , 1 970 , 1 978) might be expected t o b e enriched by the leaching of minerals from the surrounding dune ridges ( Salisbury 1 952 ) . Within the s tudy area , the soil was still very young with li ttle profile development except for a darkening of the upper 1 -2cm by organic matter in the older hollows . The soil in the originally bare terminal def lation hollow was raw sand and could be taken as " time-zero " in the chronotoposequence of soil development on the sand plains . The soi l nut ri ent , pH and organic matter status at different s tages in the sere were determined by sampling at five sites ( at increasing distances from the terminal hollow) in spring (September ) 1 978 . Fifteen replicate soil cores ( 20mm diameter x 1 00mm depth ) were taken at each si te , bulked , oven dried for 48 hours at 70 degrees C and analysed for calcium ; potassium , phosphorus and magnesium available for plant growth ; and percentage carbon. From the latter , percentage organic matter ( carbon x 1 . 7 approximately , Von Bemmeln ' s factor ; Ranwell 1 959 ) and percentage nitrogen (C/N ratio 20 ; Cowie et al . 1 967 ; Noble and Marshall 1 983 ) were determined . Two years Page 79 - I I Page 80 later in spring (September) 1 980 , the pH of the soil water was determined visually using a soi l indicator solution ( "BDH soi l testing reagent " ) . Twenty five ml samp les of soil water were taken at each si te , two drops of indicator solution added to each and the colour compared to a standard chart . The pH determination in the field were confirmed on return to the laboratory using a pH meter ( "Radiometer Copenhagen type 28" ) . The alkal ine soil pH was relatively invariable across the sites studied ( table 3 . 5 ) . The pH varied minimally between sites of at increasing successional maturity at increasing distances from the terminal hollow and by 1 980 had changed li ttle since the de terminations two years previous ly ( table 3 . 5 ) . Weathering of shell fragments evident throughout the sand in the study area together with .froW'I t\-.,e.. the lack of organic aci ds in the soil resul ting � lack of soil organic matter ( see below ) accounts for the basic nature of the soil solution. The continued weathering of shell fragments especially in the o lder deflation hollows could explain the s light apparent shifts upwards in pH over the two year period to September , 1 980 . Table 3 . 5 Soil pH at five sites on the sand plain in spring . Site SO Terminal hollow S 1 Low dune S2 Old deflation hollow S3 11 11 11 S4 Older deflation hollow Sept 1 978 Sept 1 980 8 . 1 8 . 0 8 . 0 8 . 0 8 . 1 8 . 3 8 . 1 8 . 6 8 . 0 8 . 2 Page 81 Shell fragments mixed with the sand provide a rich source of calcium available for plant growth which was found to be high in the older deflation hollow soils ( table 3 . 6 ) . These had undergone more weathering than those younger soi ls closer to the receding rear dune (SO and S 1 ) . Similarly� available potassium and magnesium were higher in the more aged soi ls of the older hollows � although like phosphorus which did not vary between sites � the supplies of these minerals avai lable for plant growth were relatively low ( table 3 . 6 ) . The sands of the Manawatu coastal dune system contain only small quanti ties of micaceous minerals ( Cowie et al . 1 967 ) which accounts for the limited amounts of available po tassium found . The organic matter content of the soi ls on the study area was extremely low , although it increased at si tes of increas ing seral maturity ( table 3 . 6 ) . Thus , in the oldest deflation hollow at S4 , percentage organic matter of the soil ( OM ) was three times that of the soil of both the young terminal deflation hollow ( SO) and the young low dune ( S 1 ) . In spring 1 978, soils at both SO and S 1 were raw. The organic matter content of the soils in the older hol lows was comparable to that described by Cowie et al . ( 1 967 ) for the youngest dune phase soils (Waitarere sand of the Manawatu coast ) and to that found in a young coastal s lack soil in a dune system on Anglesey , northwest Wales by Ranwell ( 1 959 ) . These OM values are however lower by at least an order of magnitude then those found in more mature sandy soils under Carex pu�ila arenaria on the same dune system on Anglesey (Noble and Marshall 1 983 ) and on the Manawatu sand country (Cowie et al . 1 967 , table 9 ) . P.T:Je 82 Table 3 . 6 Ava i lable nutr ient status and percentage carbon , organic matter and ni trogen of the soi l at five sites on the study area , in September 1978 . Available Ca K P Mg ( ppn soi l ) Young terminal hol l ow (SO) 375 40 1 35 Young low dune (Sl ) 375 40 l 35 Old deflation hol low (S2) 875 80 l 50 Old deflation hol l ow ( S3) 1000 80 l 75 Old deflation hol low ( S4 ) 750 80 1 65 Percentage N c OM . 0015 . 03 . 051 . 0015 . 0 3 . 0 51 . 003 . 06 . 102 . 004 . 08 . 136 . 0045 . 09 . 153 Percentage n i trogen in the soi ls sampled in September 1978 whi ch was ca lculated on the bas i s of a C/N ratio of 20 , was extremely low. These estimated values were lower than the bottom end of the range of soi l n itrogen l evels found by (Noble and Marshall 1983) using di rect measurements of tota l soi l kj eldahl nitrogen on a sand dune soi l under Carex arenar i a ( range 0 . 0005 to 0 . 02% ) . These latter so i l n itrogen levels of Noble and Mar shall coupled with the directly-estimated OM levels allow the ca lculati on of C/N rat i os for these so i l s by use of Bemmeln ' s factor . The values obtai ned (range 53 . 5 to 134 . 5) were between three and ten-fold greater than the high value ( 15 ) , ind icat ing l i tt le decompos i tion and a raw type of organic matter , quoted by Cowie et a l . ( 1967 ) for Wai tarere and in the Manawatu sand country . Page 83 The presence of nitrogen in the soil on the developing sand p lain near Tangimoana was attributed to the activity o f nitrogen-fixing bacteria and blue-green algae . Nostoc spp . which were observed to form characteristic ball-shaped colonies of variable diamet er (up to 1 0mm on the study area ) and Anabaena spp . which form a gelatinous bloom were found in the older moist deflation hollows throughout the study . At the end of winter 1 981 , large quantities of Anabaena bloom were found in the flooded terminal hollow ( figure 3 . 1 3 ) . The presence of nitrogen on the study area may also have resulted from leaching from the surrounding area . Nitrogen fixing Lupinus arboreus was plentiful across the sand plain on which the study area was situated ( figure 3 . 2 ) . Further , non-symbiotic nitrogen fixing bacteria of the rhizosphere zone of Ammophila arenaria , a species abundant both on the low dune ridges on the study area and on the surrounding high dunes , have been implicated in the presence of nitrogen in sand dune communities elsewhere in New Zealand ( Gadgil 1 969) . Figure 3 . 1 3 Anabaena bloom in the flooded terminal deflation hollow, in August 1 98 1 3 . 1 . 8 Es timation of nitrogen-fixation � acetylene reduction Stevens and Walker ( 1 970) state that in most chronosequence studies � atmospheric nitrogen-fixing plants will be present at some early stage in the vegetation succession � although not always at the pioneer stage , and tha t � with increasing successional time ; these p lants will be e liminated by those that utilize the accumulated nitrogen . Thus , the predicted pattern of nitrogen fixation rates in primary succession will show an early peak and subsequent decline ( fi gure 3 . 1 4 ) . Figu re 3 . 1 4 Predicted nitrogen fixation succession ( redrawn from Gorham et al . 1 979 ) . w � "' z 0 � X ii: I z +'•>wav-{ YhrzoVYle � _I .l L.;:.�J J J so Si S2 S3 S4 P ) N 0 0 . . t 1 n · ern odes per mod u l e - I T .l .� .l T I 1 so S i S2 S3 S4 (c) Mean i n ternode l e ngth T .L T __I_ T .i. .i. l _,.. 1 ..... so S i S2 S3 S4 \OO q . Page 1 01 reflec ts the rear the greater age of the populations at greater distances from dune and also the more prolonged periods of winter waterlogging in the hollows compared with the edge of the low dune . In the old hollow at S2 aerial looping of rhizomes was found � a phenomenon earlier linked with soil oxygen deficits under waterlogging (Raunkiaer 1 934 ) . Variation between sites in both number and length of internodes appears to contribute to the variation between sites in total length of rhi zome per sympodial module ( fi gure 3 . 20 ) . Measurements made on rhizome systems excavated from within and from the edge of the terminal hollow allowed the calculation of mean dry weight per unit length , for individual rhizome segments ( sympodial units ) in winter 1 980 ( table 3 . 8) . The mean dry weight per unit length was lower in the terminal rhizome segments than in the penultimate segments . This difference was especially apparent on rhizomes growing into the terminal hollow in May following the reproductive period . Given the correspondance between sympodial modules in I�y and July ( see arrows , table 3 . 8) it is apparent that a rhizome segment does not necessarily maintain a high dry weight ( carbohydrate content ) per unit length where the clone is continually expanding. · The generally low values obtained for all rhi zome segments for the genet excavated from the terminal hollow in July may reflect the observation that the rhi zome axes developed by this plant were in an active expansion phase unrestricted by o ther plants on the relatively bare terminal hollow. Further since few aerial shoot modules had developed on this genet the ability of this plant to generate large supplies of carbohydrate for storage was limited . By contrast � the distal rhizome segments on the edge of the low dune were connected to an ample carbohydrate generating capacity higher on the low dune . Page 1 02 Measurements of the diameter of rhizomes were also made in spring 1 979 . Figure 2 . 4 showed that the diameter of long rhizome branches was greater than that of dwarf rhizome branches and figure 3 . 2 1 although confirming this observation for branches at S 1 shows that there was considerable overlap for this character. Similarly wide variability was found both within and between sites representing various phases of development for diameter of long rhizome branches ( figure 3 . 21 ) . Rhizome diameter was greates t at the juvenile site (SO ) where the mean value was almost three times that given by Moore and Edgar ( 1 970 ) for the species in New Zealand . The diameter of long rhizome branches at S4 ( at the lower end of the range for rhizomes on the study area ) was still greater than that given by Moore and Edgar . This low value at S4 and those low values for long rhizomes at S 1 probably reflec t the observation that many long rhizome branches at thes e sites have resulted from the continued growth of ( generally lower diameter) branches . short In August 1 979 , measurements were also made on the length of dwarf rhizome branches at S1 . At this sit e , dwarf branches oft en became extended , and so the distinction between long and dwarf rhizome branches was not as clear-cut as at other sites . However ; dwarf branches were on average still signi ficantly shorter than long branches ( figure 3 . 20 ) . 3 . 2 . 3 Leaf litter decomposition Most aquatic decomposition studies have invo lved allochthonous organic material , from neighbouring terrest rial site s ; principally in streams and rivers where the action of moving water causes fragmentation of the litter and an apparent inc rease in breakdown -E E .._.. L Ci.) +l w E ro •rl 0 F i gure 3 . 2 1 Rh i zome 6 0 � 8 5 0 8 0 0 8 4 0 0 8 8 3 8 g 11 2 f! 1 d i ameter , Co.1e.� -- o Long branches c Short branches 0 0 0 0 pui'Y'i la. 0 1 02 C\ 0�------�------�--------�------�------�---- so S1 S 2 Site ss Page 1 03 rates . Only rec ently have decomposition rates of wetland species in relatively stationary aquatic conditions been measured ( e . g . Mason and Bryant 1 975 ; Danell and Sjoberg 1 979 ) . While Carex pumila on the sand plain in the Manawatu sand dune system cannot be described as a hydrophyte � neither is it a truely terres trial species since its habitat may be flooded for extended periods throughout the year. The aim of the present study was to measure the decomposition rate of Carex leaf litt er within a mois t sand slack using nylon mesh lit ter bags to enable comparisons with similar studies in both terrestrial and aquatic ecosystems and to enable more accurate es timates of net production of Carex pumila on the sand plain . Methods and materials The decomposition rate of Carex pumila litter was measured using 1 . 3 x 1 . 3mm-mesh nylon litter bags ( figure 3 . 22 ) measuring 1 5 cm x 1 0cm and containing between 7 and 8g dry weight of leaf material ( table 3 . 1 1 ) . The leaves were collected in autumn (April from the recently dead standing reproductive shoot cohort which had set seed during the previous summer . The leaves were air dried to a constant weight , chopped into 8cm lengths and accurately weighed amounts enclosed in a nylon mesh lit ter bag, along with a stainless steel numbered dis c . The edges of the bag were secured with nylon fabric (bias-binding) and nylon thread , and the total weight ( bag plus content s ) measured . The 33 bags were soaked in water for several days to reduce fragmentation losses , then placed on the soil surface in the old hollow adjacent to S2 , at 45cm spacing within an area 3m x 3m, on 9 May, 1 980 . The leaf material in each litter bag replaced an approximately equal quantity of litter which was removed from a 1 5 cm x 1 0cm patch on the ground at which each bag was placed . Page 1 04 Figure 3 . 22 Litter bag on an old deflation hollow Table 3 . 1 1 Outline of the procedure to determine the decomposition rate of Carex pumila leaves using litter bags in a deflation hollow on the sand plain . Starting date 9 May 1 980 Lit ter type dead leaves Total Initial leaf number dry weight of bags ( g I bag) n 33 Mean SD 7 . 6 1 6 5 • 2009 Number of bag removed after interval of 79 1 44 222 463 days 3 3 3 4 Page 105 At each sampling on subsequent v is i ts to the study area , at least three l itter bags were chosen at random, placed into a plastic bag for transport to the laboratory where the li tter bag and i ts contents were air dr ied to a constant weight , debris removed and the weight loss of the plant material calculated. When the experiment was terminated 20 bags remained in the field . The model expressing the loss of dry weight due to decomposition in terrestr ial ecosystems is given by Wt = Wo exp (-kt) where t 1s time , wt is the weight after time t, Wo is the initial weight , exp is the base of natural logar i thms and k is a decomposition coefficient for the specific l itter type (Olson 1963) . Minderman ( 1968) claimed that the model does not hold for whole l i tter , although it may apply to individual l itter constituents such as cellulose � To test whether Wt = Wo exp (-kt) holds for the Carex pumi la data , the regress ion of ln (Wt/Wo) on time t was calculated and goodness of fit of the data to the model was tested by use of the F-ratio . The slope (k) of the regression of ln (Wt/Wo) on time will be an approximation, at the given level of probabi l ity, of the rate of decay of the matG� ial . The model , which assumes a rate of decay proportional to the amount of material decaying over time , can be applied to predict the times of 50% and 95% leaf breakdown respectively) . Results and d iscuss ion (0 . 693/k and 3/k , Figure 3 . 23 shows the changes in dry weight of Carex pumila leaves resulting from decomposition in the moist sand slack habitat over time� The weight loss can be attributed to initial leaching which probably accounts for the apparent initially rapid decline in 0 3: '-...... +J 3: 0 3: '-...... +J 3: c r--1 Figure 3 . 23 Regress ion o f dry we ight o f rema in ing l itter of Carex pum i la o n t ime (t) . 1 � Wt I Wo �8 0 . 8 0 . 8 0 . 4 0 . 2 0 0 0 0 6 (:) 0�------�--------�--------�------��------� 0 100 200 300 400 500 0 l n (Wt I Wo ) 8 In (Wt I Wo) = - o . o o 1 9 7 t 0 l O�a. -0 . 2 0 1'" Wt :: Wo e_>(f' (-O . OO \C\1 -t) 0 0 -0 . 4 -0 . 8 -0 . 8 0 0 0 0 0 - 1--------�---------L--------�------��------� 0 100 200 300 T i me ( d a ys ) 400 500 Page 1 06 dry weight . This is followed by microbial decomposition which generally results in ni trogen-enrichment . However ; Hodkinson ( 1 975 ) found continual net loss of N from both Juncus and Deschampsia throughout his study . Predators such as nematodes soon appear which devour the fungi and bacteria. Larger invertebrates which normally; remove strips of plant material were probably excluded by the fine mesh of the nylon litter bags . Larger mesh sizes admitting larger fauna were not used to minimize inevitable losses through the mesh . The deviant observation at t = 463 days ( figure 3 . 23 ) may reflect such losses . The decomposition parameters obtained by applying the observed data to the model are given in table 3 . 1 2 . The decay coefficient of -0 . 002/day ( . 72/year ) was equivalent to a ' slow ' litter decomposition rate according to a classification proposed by Petersen and Cummins (Hodkins on 1 975 ) . However , this decay rate , which co rresponds to predicted half life and 95% life values of . 962 years and 4 . 326 years , respectively, was comparable to values ob tained for litter of other monocotyledonous species in ponds and reed swamps where current action was not a significant factor causing physical fragmentation of the lit ter . For example ; allochthonous lit ter of Deschampsia in beaver ponds ( Hodkinson 1 975 ) and Typha litter in a reed swamp (Mason and Bryant 1 975 ) produced similar hal f-life values ; al though the decomposition of Phragmites in the reed swamp (Mason and Bryant 1 975 ) and of Equisetum and Carex fluviatile in a lake in northern Sweden ( Danell and Sjoberg , 1 979) was measured at a faster rate . Thus ; despite the apparent initially rapid decline in dry weight of Carex pumila ( figure 3 . 23 ) approximately 80% of the litter remained after five months decomposition compared with only 50-60% of the Carex fluviatile lit ter in the Swedish study. n 13 Page 107 Table 3 . 12 Decomposition parameters for Carex pumila leaf l i tter Decay coefficient SD of decay * 1/2-life 95%-l i fe k (/day) -0 . 001973 coefficient F-ratio (days) 0 . 000136 210 351 (days) 1579 p < . 001 * F-ratio = mean square of regression I mean square of the residual Deviations of observed data from the model Wt = Wo exp (-kt) which are better observed on a log plot ( fi gure 3 . 23) are apparent although they were not seen to be s ignif icant on the basis of the stat istics used (P < . 001) . However , the four observations made at the end of the winter 1981 (t = 463 days) have a large i nfluence on the predicted value of the decay coefficient ( slope of the regression) ( table 3 . 13 ) . I f the observations made throughout the study period accurately reflect the state of decay of Carex l itter at those times then the rate of dec2y �3S not linear but changed over time� Both seasonal environmental effects and the stage in the decomposition process are probably involved . Decomposition involves a complex of factors such as leaching ana the activities of bacter ial and fungal decomposers and animsl a�J plant predators (Bryant and Mason 1975) • Environmental i nfluences on the microfauna related to the degree of aeration of the l itter which would have been detennined largely by water table levels in the deflation hollow most probably affected decomposition rates . At no stage during the first winter (days 0 to lOO) were the bags submerged since the water table did not rise above the surface over this period. However , dur ing the fol lowing winter ( 1981 , days 3 50 to 450 , approximately) extensive f looding in the old def lation hollow occurred � Subterranean decomposition rates (of roots Page 1 08 and rhizomes ) were not measured directly in this s tudy � but are predicted to be lower than that estimated for leaves on the surface since below the surface , anaerobic conditions apply for a greater proportion of the year. Table 3 . 1 3 Observed and predicted values of Y ( ln Wt/Wo ) � standard deviations of Y, residuals and s tandard residuals for Carex pumila leaf litter. Time Y ( days ) ( ln Wt/Wo ) 463 463 463 463 -1 . 2929 -0. 9753 -0 . 8387 -0 . 7607 Predicted Y value -0 . 9 1 36 -0 . 9 1 36 -0 . 91 36 -0 . 9 1 36 SD of predicted Y Residual 0 . 0630 0 . 0630 0 . 0630 0 . 0630 -0 . 3793 -0 . 06 1 7 0 . 0749 0 . 1 529 R denotes an observation with a large standard residual Std Residual -2 . 98 RX -0 . 49 X 0 . 59 X 1 . 20 X X denotes an observation whose X-value gives a large influence Page 1 09 CHAPTER FOUR Field perturbations 4 . 1 Introduction 4 . 2 Methods and materials 4 . 3 Results and discussion 4 . 3 . 1 Aerial shoot densities 4 . 3 . 2 Leaf area 4 . 3 . 3 Age and size distributions 4 . 3 . 4 Dry weight , energy, elemental contents 1 . Dry weight 2 . Net biomass accumulation rates 3. Energy and elemental concentrations 4. Total nitrogen content per unit area 4 . 3 . 5 Flowering and seed production 4 . 3 . 6 Allocation of dry matter and total nitrogen 4 . 3 . 7 The effect of nitrogen fertilizer addition Page 1 1 0 4 . 1 Introduction The sand plains of the dune system of coastal Manawatu provide a harsh environment for plant growth . External factors that are likely to limit growth in this habitat include ( 1 ) wind and shortages of both mineral nutrients and soil oxygen which restrict production and ( 2 ) the activities o f wind , herbivores and pathogens which c� the partial or total destruction of plant biomass . This combination of stress and disturbance factors ( sensu Grime 1 977 ) precludes all but a small set of rhizomatous perennial herbs ( Esler 1 969) . In the field it is unlikely that causal relationships between organisms and their environment can be recognized without deliberate perturbation of the system. The field investigation outlined in this chapter involved three different perturbation treatments involving external factors considered to be of significance in the development of pioneer plant populations on the embryonic sand plain system : ( 1 ) the removal of seed and seedlings of Carex pumila , ( 2 ) the construction of wind breaks ( shelters ) and , ( 3 ) the application of nitrogenous fertilizer . The aims of the study were twofold . Firstly� I wished to examine the behaviour and reproduction of Carex pumila and � secondly� to determine if within a particular seral stage the Carex pumila population could respond to a deliberate perturbation of the habitat in a way that would optimize reproductive output . The selective pressures of the physical and biotic environment � the habit of the species and its life history pattern influence the nature and pattern of resource allocation by a plant to vegetative and reproductive structures ( Snell and Burch 1 975 ) . Thus ; the way in which a plant allocates its limited resources is of significance in pa3e l l \ its ecological and evolutionary history as this will affect its survival and contribution to future generations (Abul-Fatih et al . 1 979) . Selection then may be thought of as having optimized ; amongst other things ; the allocation of resources between vegetative and reproductive functions in such a way that the organism ' s fitness is maximized . In the present study the patterns of allocation of the resources of dry weight and total nitrogen by Carex pumila populations over time were s tudied . Answers to the following questions based on those originally posed by Harper ( 1 967) were sought : ( 1 ) Is the proportion of the output of Carex pumila that is devoted to reproduction similar in the younger populations to those older populations further from the terminal hollow? ( 2 ) Is the proportion of dry weight and crude total nitrogen devo ted to aerial growth greater in the more crowded populations? ( 3 ) Can the proportion of the output of Carex pumila that is devoted to seeds be al tered by deliberate perturbation treatments? (4) Are the processes of clonal growth competitive with these involved in producing seeds? What is the relative expenditure of resources on seeds and clonal growth in the various populations? ( 5 ) What is the expenditure on organs ancillary to the seed? ( 6 ) Do the patterns of allocation of dry matter and of crude total nitrogen to component plant parts differ? thie+ie!d It was recognised that thiield approach would be insufficient to distinguish between plastic and genetic population differences as the sere progressed in either space or time . To enable genetic differentiation to be demonstrated , populations raised from seed or shoots would need to be compared at a common site ( Turesson 1 922 ; Davies and Snaydon 1 973 ) . Perturbation 1 : Seed removal In predictably changing habitats such as those of a primary succession, the increasing interference between plants may be a good cue for the decreasing favourability of the site for the pioneer species . In such habitats when densi ty was low� c lonal expansion would be possible . Clonal growth would be most advantageous as it would facilitate local spread and occupation of a site by a favoured genet . However , when population density became high due to crowding� dispersal through seed production would become advantageous as it would allow escape to new , perhaps more favourable sites ( Abrahamson 1 975 ) . Thus , genets in seral habitats are predicted to exhibit a plastic reproductive response to increasing seral maturity , allowing them to switch from clonal growth to seed output depending upon the conditions . Hawthorn and Cavers ( 1 976 ) found that removal of spikes from fertile shoo ts of the perennial Plantago rugelii for two years in permanent quadrats in both old pasture and recently-disturbed pasture near London , Ontario , lead to reduced densities of this species . The degree of reduction of plant densi ty in response to this perturbation can be considered to be direct ly proportional to the rate of new genet recruitment from the seed bank in the soil , relative to other means of maintenance of the population. roge l l3 Carex pumila is a species with an apparent duality in the means by which it perpetuates itself. Under certain circumstances � it is a prolific seed producer� yet at the same time tenure at a site is maintained by the extension of rhizomes through the sandy substrate and the production of new aerial shoots . In clonal species such as Carex pumila � shoot density reduction through the removal of the source of new genets germinating in the soil seed bank may be offset by the vegetative production of new daughter shoots ( ramets ) . In the experiment outlined below � both current and previous years seeds of Carex pumila were removed to investigate the role of sexual reproduction versus clonal growth in the establishment and/or maintenance of populations of this pioneer species at different stages of the sand plain sere . Perturbation 2 : Shelter Wind has long been recognised by agriculturalists and more recently by ecologists as a factor affecting plant growth . Leonard Cockayne ( 1 91 1 ) was the first to suggest that wind was an important ecological factor in the New Zealand environment noting that· in windy places compact prostrate bushes of a similar growth habit to that predomi()ateJ . resulting from divarication pominat�. Wind exposure is a term used by ecologists to describe the stresses and disturbances experienced by plants growing in windy places . Exposure may be manifest as mechanical damage (Daubenmire 1 959) , anatomical and morphological changes or reduced plant yield ( Grace 1 977 ) . Wind tunnel studies have shown high winds ( 7- 1 2 m/sec ) reduce relative growth rates , rate of leaf elongation and leaf area ratio , effects which have been attributed to reductions in leaf water potential ( Grace 1 974 ; Grace and Russell 1 978) . Similarly , the Page 1 1 4 anatomical and morphological changes ( including smaller thicker leaves ; shorter internodes ; increased root/shoot ratios ; increased amounts of vascular tissue ; and increased number of stomata per unit leaf area ) are like those features that develop in response to soil water deficits . Such responses to wind are primarily moisture stress effects (Hollows 1 978 ) . The reduction of leaf elongation and the increased incidence of sclerophylly brought about by increased wind speeds in the grass Festuca arundinacea have been shown to be less extreme when phosphorus supply to the plant was adequate ( Pitcairn and Grace 1 982 ) . The possible role of nutrient stress in causing sclerophylly is further suggested by the observation of this feature in phosphorus deficient soils of Australia (Beadle 1 954 ) . Amelioration o f the effec ts of wind on plants by the use of wind breaks and shelter be lts have been a pract ical means of increasing plant yie ld ( Grace 1 977 ) . Wind breaks however , affect variables other than wind speed , such as temperature . Thus , there remains an uncertainty about the interpre tation of results of shelter experiments in the field that are not complemented with controlled environment studies in a wind tunnel . In coastal habitats wind � salt spray and nutrient and soil moisture defici ts are likely to be important factors affec ting plant growth . The problem of distinguishing between the effects of each is exacerbated by the observation that similar results may be achieved in response to all of these factors ( Grace 1 977 ) . Thus , there remains an uncertainty about the sugges tion of Turesson ( 1 922 � 1 925 ) that ecotypic differentiation of dwarf coastal populations in several plant species was the result of the lesser ability of such plants to trap salt carried by the wind . Page 1 1 5 An outstanding characteristic of the west coast of the north island of New Zealand is the large number of days per year with high wind gusts and in coastal Manawatu in particular the intensity and persistence of the highly directional winds which abound for most of the year ( section 3 . 1 ) . It is therefore possible that the magnitude of wind gusts which was seen to vary across the study area ( section 3 . 1 . 8 ) could confer a degree o f adaptive significance to the variation in vigour of Carex pumila across the same area. Thus ; wind breaks were constructed across the sere to tes t this hypothesis . Perturbation � Nitrogen fertilizer addition It is well established that sand dune communities are under stress from low nutrient s tatus of the soil . Rapid increases in productivity and turnover of ramets and individual plant parts have been observed when fertilizers are applied (Yemm and Willis 1 96 1 ; Willis 1 963 ; Huiskes 1 979 ; Noble et al . 1 979 ) . Nitrogen and phosphorus are the elements most commonly found to be deficient in sand dune soils . For example , Smith ( 1 977 ) reported "excellent growth responses to applications o f nitrogen" and less marked responses to phosphorus ; on sand country 7km north-east of the study area . It is probable that soils of low-lying dune hollows are enriched by leaching from dune ridges that surround them ( Salisbury 1 95 2 ) and in this respect the ho llows can be considered more favourable habitats for plants than the ridges . This view is confirmed by the experimental work of Jones ( 1 975 ) who noted amongst the beneficial effects o f waterlogging in relation t o the development of dune slack vegetation , the increased availability of phosphorus and potassium . Thus ; one might expect a diminished response to added fertilizer by vegetation in the less impoverished dune hollows compared with that of Page 1 1 6 vegetation on the more nutrient-stressed dune ridges . However ; responses of sand dune communities to fertilizer apply to vegetation on both dune ridges and dune hollows , although these responses are not uniform to all species within these communities ( e . g. Watkinson et al . 1 979 ) . The application of ni trogenous fertilizer was the final perturbation to the sand plain system. The objective was to determine the effect of this element ; whose deficiency was suspec ted to be the major cause of stress in this pioneer sand plain community , on populations of Carex pumila , the major species present . 4 . 2 Methods and materials Experiment 1 : Seed removal On December 26 and 27 , 1 977 all female spikes on fertile shoots and seed of Carex pumila on the ground were removed from two 2 x 2 m plots at each of four sites in the study area described above ( section 3 . 1 ) . Two similarly sized undis turbed control plots were also established at each site . The four plots at each of the old deflation hollow sites (namely , S2 , S3 and S4) formed a square , replicate treatments diagonally opposite each other, with a 30cm wide pathway running each way between the plots . Since , on the edge of the low dune , the band of Carex pumila shoots invading the embryonic terminal hollow formed a relatively narrow band , the four plots at this site were established in a line ; again with 30cm pathways be tween plot s . Carex pumila seeds ( each a nut encased in a corky utricle ) which are buoyant were redistributed on the study area during the winter 1 978 as a result of flooding and wind disturbance . Thus , seed was removed again from the ground on the treatment plots , in early spring 1 978 . Harvesting programme and procedure The harvesting programme for this experiment is given below : Harvest number 2 3 4 5 6 7 8 9 1 0 1 1 1 2 26/27 December 25 January 3 April 1 0 August 1 8 October November 1 4 November 29 November 1 3 December 27 December 1 0 January 1 9 February 7 September 1 977 - Treatment established 1 978 1 978 1 978 1 978 1 978 1 978 1 978 1 978 1 978 1 979 1 979 1 979 At each harvest , on each of the 1 6 ( 4 sites x 2 treatments x 2 replicates ) plots , one 30 x 30cm quadrat was harvested , initially towards the periphery and then with subsequent harvests progressively closer towards the centre of the 2 x 2m plot in order that trampling damage to unharvested plants be avoided . The aerial portion of all shoots , both living and dead � within each quadrat were removed at ground level and placed into a labelled bag . Below ground parts were sampled from the same areas used for the above-ground sampling by digging up all plant material in the quadrat to a depth of 30cm (up to 1 m at S 1 where sand accumulation occurred) . A spade was used to sever the rhizomes at the edge of the quadrat . At S2 , S3 and S4 in the old deflation hollows where the interweaving of rhizomes and matting of roo ts had caused a turf to form � the below-ground samples were removed more or less intact for later separation . A t S 1 where a turf had not formed � the sand fell away from the roots and rhizomes . Underground plant parts were removed from the sandy substrate by washing. On return to the laboratory� all plant material from each quadrat ( both aerial and underground fractions ) was thoroughly washed in several changes of water and divided into constituent parts � dried at 80 degrees C in a forced venti lation oven for 24 hours and then weighed . Numbers of dead , vegetative and fertile shoots and of seed were counted . Roots may have been underestimated through loss in the washing and separating procedure ; roots were also observed to penetrate below the sample depth . These errors are likely to be small ( < 5% ) since firstly, at S2 , S3 and S4 � the bulk of root material was associated with the mat of rhi zomes in the upper 0- 1 5 cm of soi l , and secondly , Carex pumila roots do not penetrate the grains of the substrate ( cf roots of species growing in clay soils ) . They were therefore easily separated from the sandy substrata , with little root damage or loss . 5 X S3 ) Experiment 2 : Shelter wind breaks On 1 October 1 979 , a 1 m high wind break was constructed around a 5 m area at each of four sites on the study area (SO � S 1 , S2 and using 50mm mesh wire netting and hessian sack-cloth . All shelters , with the exception of that at S2 on the less exposed part of the study area ( section 3 . 1 . 8) , required renovation within two months when 6mm pore diameter Netlon wind break material was used . This too required continual maintenance during the latter part of the summer 1 979-80. Page 119 Harvesting programme and procedure The harvesting programme for the shelter experiment is g iven below: Harvest number 1 October 1979 - Windbreaks erected 13 17 October 1979 14 30 October 1979 15 6 Novanber 1979 16 14 Novanber 1979 17 27 November 1979 18 4 December 1979 19 11 December 1979 20 17 Decanber 1979 21 8 January 1980 22 5 February 1980 23 5 Apri l 1980 24 28 July 1980 The harvesting procedure di ffered from that outl ined above for harvests 1 to 12 . The 30 x 30 cm turves with aer ial shoots attached were l ifted and carried to the laboratory where the sandy substrate was removed by washing . This procedure facil itateJ the separation of intact br.anch modules , including the rhizome and adventit ious roots . Vegetative, fert i le and dead branch modules were counted and further separated into constituent parts . They were dried and weighed as above� In addition to the counts of seed and of shoots taken previously each living shoot was placed in a size category according to the height to the l igule of the youngest fully-expanded leaf (collar Page 120 height) • The number of both l iving and dead leaves per shoot were counted and the area of green laminae per shoot measured us ing an � electronic scanner . On ferti le shoots , the number of female spikes per culm and seeds per spike were counted, and the stage of development estimated . Plant samples were collected from two contiguous 30 x 30 cm quadrats ( ladder quadrats) within each treatment at S2 and S3 . On the controls on the low dune ( SO and Sl) four contiguous 30 x 30 cm samples were harvested . No samples were harvested from wi th in the shelters at SO or Sl , although during this period measurements and counts of the aerial shoot population on these plots were made . Ladder quadrats (Pearsall and Gorham 1956) were used to avoid overestimation of biamass owing to clumping of the vegetation and edge effects . The l imited area of the Carex pumila stands at each of the stages of development across the study area precluded replicate ladder sampling . Experiment 3 : Nitrogen ferti l i zer add ition In the spring of 1980 ( September 17) ammonium sulphate ferti l i zer was applied by hand at a rate equivalent to 50kg of elemental nitrogen per ha to two replicate plots at each of three sites representing di fferent stages of the sand plain sere. Two control plots (no added N) per site were establ ished adjacent to the fertilized plots in such a position that i t was unlikely ferti l i zer could be blown onto them� Plots were established in the old hollow where the Carex pumila population was senile and formed only a small proportion of the total l iving biomass ( S3) and at two contiguous sites on the low dune where the Carex pumila populations were in a mature (Sl) or adolescent (SO) phase of development . The ferti l i zed plots were not sampled at the subsequent harvest on 17 October , but were left undisturbed until the harvest on 17 December , · 3 months after the application of nitrogen Page 121 when a single 30 x 30 cm quadrat was sampled from each plot . In add ition two contiguous 30 x 30 cm quadrats were sampled from the old hollow at S2 to complete the temporal sequence at this site . Harvesting programme and procedure The harvesting programme for the ni trogen fertili zer perturbation experiment is g iven below : Harvest number 25 26 17 September 1980 N-ferti l i zer added 17 October 1980 17 December 1980 The turf from each 30 x 30cm quadrat was lifted intact , or where this was not feas ible care was taken not to sever below-ground connections between aerial shoots within the sample and removed to the laboratory and treated as for harvests 12-24 allowing the separation of whole modules of both dwarf and long s�oots into vegetative , reproductive and dead categor ies , for counting a�� further separation into components . Numbers of green and of dead leaves per living shoot , numbers of seed per 30 x 30 cm and nunbers of l ive shoots within var ious s ize classes according to collar height were also counted. Energy and l ipid determination Lipid determinations were made by ether extraction and energy determinations by use of a Gallenkamp adiabatic Bomb calorimeter . Energy determinations were carr ied out on constituent organs of carex pumila from S2 at harvest 1 and on seeds from S2 at harvests 7 to 10 and from each of four s ites ( Sl , S2 ,S3 and S4 ) collected on 25 January 1979 . Page 1 22 Crude total ni trogen The determination of crude total nitrogen (% ) of constituents organs of Carex pumila was carried out using the following reagents (Haselmore , pers comm) : 1 . Digestion mix . 1 00 g of potassium sulphate (AR ) plus 1 g of selenium powder were added to 1 litre of concentrated sulphuric acid (AR ) in a large conical flask . This mixture was heated in a fume cupboard at c 300 degrees C for c 2 hours until a clear solution formed . 2 . Hydrogen peroxide (AR) , 30% (W/V) . 3 . Phenol reagent . 50g phenol (AR) plus 0 . 25g sodium nitroprusside ( AR ) was made up to 1 litre with distilled water in an amber bot t le and kept in the refrigerator . 4 . Hypochlori te reagent . 25g sodium hydroxide (AR) plus 56ml of 4% sodium hypochlorite was made up to 1 litre with distilled water and stored as for phenol . 5 . Nitrogen standard . 2 . 360g ammonium sulphate (AR) was mixed with 1 00ml of distilled water to give a nitrogen concentration of 5 mg N per ml . The procedure involved taking approximately 50mg of oven dried plant material passed through a 1mm mesh in a micro hammer mill , accurately weighed into digestion tubes ( 1 8 x 1 50mm, glass test tubes ) to which ml of digestion mix then O . Sml of hydrogen peroxide was carefully added . The tubes were gently shaken t o aid solubi lisation. The samples , usually in batches of 48 , were digested on an aluminium heat block initially set at 1 50 degrees c . They were then heated to Page 1 23 320 degrees C in about 2 hours and held at this temperature for a further 2 1 /2 hours ( 4 1 /2 hours total ) . The solubilisation and digestion were carried out in a fume cupboard . Safety glasses and gloves were worn when the digestion mix and hydrogen peroxide were added and the tubes placed on the heated diges tion block . At the end of the 4 1 /2 hour digestion period , the samples were transferred to a metal test tube holding block and left to coo l � then 8 1 /2 ml of distilled water added , mixed thoroughly and again left to coo l . 5 0 pl aliquots of this solution were transferred to a second set of test tubes then 1 ml of phenol reagent and 1 ml of hypochlorite reagent were added , mixing thoroughly using a vortex mixer on the addition of each reagent . Samples were left for 1 hour at room temperature ( or 20 minutes at 37 degrees C ) then 8 1 /2 ml of distilled water added , mixed thoroughly and read for absorbance at 630nm using a Hitachi spectrophotometer. A set of ni trogen standards containing 0 , 0 . 5 , 1 . 0 , 1 . 5 , 2 . 0 and 2 . 5mg N were taken through the entire procedure , with each batch of plant samples . Absorbance at 630nm was linear over the range of nitrogen concentrations (0-2 . 5 mgN/ml ) in the standards . Thus , a straight line was fitted to the plot of absorbance versus nitrogen concentration for the standards in each sample batch . The slope of this curve ( straight line ) was used to calculate the crude total nitrogen concentration ( % ) o f each sample in that batch . Component organs of Carex pumila shoots from harvests to 26 were thus tested for percent crude total nitrogen . The total nitrogen content of this plant material in grams ni trogen per unit area of ground was then calculated from the dry weight data , so allowing the dete rmination of the allocation of total nitrogen , as well as that of dry weight , to component organs and shoot types . Page 1 24 Other elements Oven dried subsamples of constituent organs of Carex pumila were passed through a 1mm mesh in a micro-hammer mill . Accurately weighed samples (approximately 0 . 1 g) of this finely milled material were ashed in a muffle furnace at 500 degrees C for more than two hours . The ash , approximately 0 . 01 g per sample , was dissolved in 2 ml of constant B . P . 2 molar hydrochloric acid . This solution was decanted to remove any residue , resulting from incomplete combustion in the furnace . Determinations were made on these samples of K and Na by flame photometry and of Cu , Zn , Fe , Ni , Mn, Mg and Ca by atomic absorption spectrophotometry . Standard solutions and blanks were run throughout the procedures along with the plant samples . Raw data for dry weights , leaf areas , leaf , shoot and seed numbers and elemental concentrations have been placeJ on a magnetic tape held by the author. Statistical analysis At each harvest , the data were analysed by analysis of variance by use of Teddybear (Wilson 1 979 ) , a stat istical package held at Massey University for use on the Burroughs B6700 Computer. For each variate analysed , means and within-treatment variances were calculated and tested for residual heterogeneity of variance (Bartlett ' s tes t ) , skewness ( G1 ) , kurtosis ( G2 and A) and normality. Duncan ' s Multiple Range Tes t was performed to dete rmine the significance of differences between level means within each factor� and a normality plot was calculated and printed . Where deviations froo no rmality were indicated by these tests , transformations of the data were carried out . Log transformations were most frequently applied where the Page 1 25 frequency distribution was skewed to the right ( G1 positive ) . Bart lett ' s test for heterocedasticity is extremely sensitive to deviations from normality and thus a significant test result may merely indicate non�normality rather than heteroscedasticity (Sokal and Rohlf 1 967) . Thus � transformations to make the data normally distributed often resulted in a non-significant Bartlett ' s test (non-heteroscedasticity) where heteroscedasticity had previously been indicated . When transformation of the data did not show non-heteroscedasticity no further manipulation of the data ( Sokal and Rohlf 1 967 , p . 376 ) was attempted . Variates analysed included dry weight and total nitrogen content per unit area for individual and combined plant parts � ratios of various combinations of these , and dry weight and total nitrogen content of plant parts per branch module. At the final harvest ( H26 ) � the unbalanced experimental design (control treatment only) at S2 ) required that the analysis of variance involve the partitioning of the sums of squares and degrees of freedom into arbitrary contrasts . The data were analysed by treating each site x experimental treatment combination as a different level of a single factor, "site " . The sums of squares and degrees of freedom for this factor ( site ) were then partitioned into single-degree-of-freedom contrasts . The contrast coefficients are shown below : Levels of fac tor "site" 3C 3N 1 C 1 N OC ON 2C 8ontrast name A 0 - 1 0 - 1 0 B 0 2 0 - 1 0 - 1 0 D 0 0 - 1 - 1 0 E 0 0 - 1 - 1 0 F 0 0 - 1 - 1 0 I -2 -2 0 Page 1 26 December 1980 harvest will therefore appear twice , fi rst as a point 1n the temporal sequence ( sections 4 . 3 . 1 - 4 . 3 . 6) and second , as a contol for the� ni trogen experiment ( 4 . 3 . 7) . 4 . 3 . 1 Aerial shoot densities Figure 4 . 1 shows the sum of the densities of vegetative � ferti le and dead shoots of Carex pumi la at four sites on the study area in January 1 978. At sites of increasing distance from the terminal deflation hollow � the densities of living and of all shoots form n-shaped plots � reminiscent of those des cribed for rhi zomatous perennial species displaying phasic development ( see references and further results in section 3 . 2 . 1 ) . As a result of the nature of the formation o f the sand plain, these sites form a chronoseries ( section 3 . 1 ) . The populations at the youngest and oldest sites , S1 and S4 respec tively , had similar and reduced total living aerial shoot densities compared with the populations at sites of intermediate age , at S2 and S3 . The populations at S1 and S4 could be distinguished by the proportion of dead shoots present in the total shoot population and by the ratio of fertile to vegetative shoots ( figure 4 . 1 ) . Dead and fertile shoots were absent from the juvenile Carex pumila shoot population at S 1 on the edge of the terminal hollow . Vegetative shoot densities were remarkably similar at all four sites ( figure 4 . 1 ) . Figure 4 . 2 shows the changes in living shoot densities of Carex pumila at each site monitored on the study area over time . In winter, spring and summer 1 980 , these shoots were divided into those attached E CJ 0 140 120 (T) 100 X 0 (T) c.. Q) a. (J) +I 0 0 .c U1 4- 0 80 60 c.. 40 Q) .n E ::::1 z 20 Figure 4 . 1 The sum of th e density o f vegetat ive , ferti le and dead shoots o f Carex P.Umila at four s ites on the sand p l a in near Tang i moana, in January 1978 . Vertical bar denotes +- var iance of l ive shoot means . S i S2 Site sa Dead Shoots Fert i l e shoots Vegetat ive shoots S4 F igttre. 4- 2 n.e Su"" o.f �e. dens et.:� of ve.ge..t-otive or)d. -fert . le shoo-Is of C are>' pu.l'll, l a over bme at f1ve. s1+e..s On -ItA e. C.'t:lv-d p iO:r>"� . -- duril')5 \ql S , , q , q and. Jqw 40 o tota l shoots (a) s ite 1 30 �· • vegetative shoots 20 • • 10 0 (b) s ite 2 60 50 0 E 0 40 0 (TJ 30 X 0 20 (TJ 10 (_ • Q) ()_ 0 (c ) s ite 3 (_ Q) 30 0 _o E 20 ::J z 10 0 (d) s ite 4 (e) s ite 0 30 20 � 0 .... .... ...... 10 ... 0 Jan R..b M ar Apv- MQ� Jut) Ju l Aug Sep Oct Nov Oec Jan Feb Mar Apr May Jun Jul Aug Sep act Nov Dec Jan Feb Mar Apr May Jun Ju i Aug Sep Oct Nov 1 878 1 979 1 980 Page 1 27 to dwarf and to long rhizome branches ( figure 4 . 3 ) . From the changes between successive harves ts , new vegetative shoot recruitment and sho o t mortality may be inferred. These data show two major trends p firs tly of seasonal fluctuations and second ly� of longer term changes that could be predicted on the basis of population age , ie distance from the terminal hollow. In the younger populations on the edge of the low dune , shoot densities increased from year to year in contras t to the trend in the older populations , both on the deflation ho llows and on the low dune , of dec lining shoot densities be tween years . The seasonal fluctuat ions in density of Carex pumila shoots are not in phase for the populations on the l ow dune and on the deflation hollows . In conjunction with data on the dis tributions of shoot age at consecutive harves ts ( section 4 . 3 . 3 ) , seas onal density fluctuations show shoot mortality to be most apparent in winter/early spring and again in mid-summer on the deflation hol lows ( S2 , S3 and S4 ) , and in bo th spring and summer on the low dune (SO and S 1 ) . Shoot mortality also occurred in autumn at all sites on the study area , as the culmination of the events leading to flowering and seed production in these monocarpic shoo ts . Major pe riods of shoot recrui tment occurred in autumn and spring on the deflation hollows , periods that coincided with more mesic soi l conditions compared with those in summer (dry ) and winter ( flooded ) . On the low dune , where the winter water table levels remained below the surface , net shoot recrui tment continued throughout winter . Thus , whereas a seasonal trough in shoot densi ty was found in winter on the deflation hollows , a seasonal peak was found at this time of the year or early spring on the low dune . 25 20 15 1 0 5 0 20 E 0 15 0 (f) 10 X 0 5 (f) 0 50 45 (_ Q) o_ 40 >- +-' 35 ·rl UJ c 30 Q) "0 +-' 25 0 0 ..c 20 (f) 15 10 5 0 10 5 0 F igur'e 4 . 3 The sum of the dens i ty o f d�1ar f and l ong shoot �2l G\ popu l at ions o f Carex Qumi la over t ime at f our s i tes on the sand p la i n bctl�een Ju l y (l� i nter) and December' (summer ) 1980 (a ) s i te ( b ) s i te (c ) s i te (d ) s i te Ju l 0 o�mrf f ert i l e shoots 01�ar f vegetat ive shoots Long fert i le shoots 1 2 3 ==== Au g Sep Oc t 1 98 0 Nov Dec Page 1 28 The lack of appearance of new shoots above the surface at S 1 during autumn 1 978 was associated with a marked accumulation of about 300mm of sand at this site . Continuing growth of this juvenile Carex pumila population was however � not prevented ( see section 4 . 3 . 4 ) . When the sites on the study area were established in December 1 977 , S 1 was at the edge of the low dune and included the front of Carex pumila rhizomes advancing into the terminal hollow. By the end of March 1 978 , the rhizome front had extended into the terminal ho llow a further O .Sm and sand accretion had occurred around shoots at S1 which were then further behind the rhizome front ( section 3 . 2 . 2 ) . Continuing vegetative shoot recrui tment and the lack of shoot mortality over the winter at S1 lead to an increased total living shoot density by August 1 978 ( figure 4 . 2 ) . Further building of the low dune , as a result of the deposi tion of dry wind-blown sand around Carex pumila sho ots in spring 1 978 , resulted in a massive decline ( P< . 01 ) in the dens ity of living shoots and a concomitant increase in that of dead shoots at this site between August and mid-October 1 978 ( figure 4 . 2 ) . Figure 4. 2 shows that the recovery of the living shoot density at S1 to the late winter 1 978 peak took the whole of the subsequent 1 978-79 summer � to February 1 979 . Few shoots at S 1 were seen to be fertile in the summer of 1 978-79 ( figure 4 . 2 , difference between total living and vegetative shoot densities ) , although their density ( 24 . 6 / m2 averaged over harves ts 5 to 1 1 ) represented almos t 50% of the spring nadir of total living shoots . The steady increase in vegetative shoo t density at S 1 over summer 1 978-79 continued the following autumn and winter, so that by Page 1 29 spring ( September) 1 979 mean aerial living shoot density was the greatest yet seen at this maturing site ( figure 4 . 2 ) . In spring 1 979 ; shoot density at S 1 decreased . This decline was not as massive as that in 1 978 ( figure 4 . 2 ) , when the site was closer to the edge of the low dune and so more vulnerable to sand deposition. However ; the movement of sand by the strong seasonal winds continued across the si te in summer 1 979-80 . In the face of this continuing site disturbance and increased moisture stress as a result of the lowered summer water table level , the ageing Carex pumila population at S 1 showed a reduced ability to maintain itself ; net vegetative shoot densities remained low over this period . The density of fertile shoots at S 1 i n the summer 1 979-80 was significantly greater than that of in the previous year , although as a proportion of the late spring 1 979 living shoot density, it was similar to that proportion of the late spring 1 978 shoot density found to be fertile in summer 1 978-79 , namely almost 50% ( figure 4 . 2 ) . Recruitment of new vegetative shoots on the low dune at S 1 later in the summer 1 979-80 and the following autumn resulted in a net increase in vegetative shoo t density to Apri l 1 980 , although this increase was considerably smalle r than that observed twelve months earlier. The increasing senility of this low dune population was reflected by the net decrease in living shoot density over the winter 1 980 , whereas during the two previous winters net shoot recruitment occurred . However , between mid-winter ( July) and late spring ( Oc tober ) 1 980 , shoo t density at S1 increased , whereas in previous years , the deposition of dry wind-blown sand on the low dune in spring was associated with net shoot mortality. Page 1 30 By December 1 980 � fertile shoo t dens ity on the control plots represented 1 00% of the living shoot population present at the site five months previously. This density was similar to that for fertile shoots observed at this site during summer 1 979�80 . The reduced spring/early summer shoot mortality on the low dune in 1 980 was also reflected by the density of vegetative shoots at S1 which in December 1 980 was no less than that observed in mid-summer twelve months previously. The senescence of the Carex pumila population on the low dune at S1 was retarded by the increased stability of the sand plain brought about by the artificial planting of Ammophila arenaria on the surrounding high dunes and lateral sand plains in spring between 20 August and 1 7 September , 1 980 ( figure 3 . 4 ) . Thus , during the windy period of spring 1 980 the supply of mobile sand was drastically reduced and the putative major cause of spring/early summer shoot mortality on the low dune was removed . From spring 1 979 , a further site ( SO ) , · where Carex pumila rhizomes were extending into the terminal deflation hollow � was monitored. so was situated on the edge of �e low dune in a comparable position to that of S 1 when this site was laid down in summer 1 977-78 . The age distribution of the shoot population at SO in mid-October 1 979 ( section 4 . 3 . 3 ) shows predominantly young shoots , with a few aged shoots , which were likely to have been recruited the previous autumn. This conc lusion was supported by the observation in mid-summer 1 979-80 of a small proportion o f fertile shoots in the total shoot population. Aerial shoot density at SO increased during spring 1 979/early summer 1 979-80 ( figure 4 . 2 ) , despite the acc retion of wind-blown sand around Carex pumila shoo ts on this part of the low dune ( figure 4 . 4 ) . A large decline in total living shoot density was observed at SO in midsummer 1 979-80 between 1 7 December and 8 January. Page 1 3 1 At this latter dat e , dead vegetative shoots first appeared in the budget a t SO . Continuing shoo t rec ruitment at SO in late summer 1 979-80 resulted in a return to the earlier vegetative shoot density peak ( fi gure 4. 2 ) . Figure 4 . 4 Views of the low dune showing newly deposited ( dry ) sand� in ( a ) spring 1 979 ( left ) and (b ) autumn 1 980 ( right ) . Subsequent deposition of large quantities on sand on the low dune by high winds during February 1 980 resulted in shoots at SO being buried by 21 February 1 980 ( figure 4 . 4b) . At subsequent harvests on the edge of the low dune in autumn and winter 1 980 , the density of living shoo ts was reduced ( figure 4 . 2) . The surviving population in Page 1 32 July showed a more mature age struc ture than that at the site in April which was skewed to the right ( section 4 . 3 . 3 ) . Over the subsequent spring and early summer, the effect of recruitment of both dwarf and long shoots was evident with the total living shoot density reaching a plateau in December 1 980 ( figure 4 . 3 ) . Of these shoots , more than 50% were fertile . The densities of both dwarf and long fertile shoots at this site in summer 1 980-81 was significantly ( P< . 05 ) greater than that of the winter dwarf and long shoot site in July . They were however, populations present at the no different from that of the combined shoot plus shoot bud populations present at this time . The implication of this observation is that both long and dwarf underground shoots of Carex pumila that are present in winter and that have undergone little or no orthotropic development can develop during the subsequent spring and summer into fully fertile shoots , capable of producing ripe seed . This was confirmed by the monitoring of tagged shoots . An emerging shoot in spring was observed to have produced seed by the end of summer , four months later. In the old hollows at S2 , S3 and S4 , the pattern of change in shoo t densities observed over the duration of the study showed differences to those seen on the low dune site . Sand accretion which was associated with aerial shoot mortality on the low dune was not seen to be significant in the old hollows . Periods of shoot mortality in the old hollows other than that in autumn associated with flowering in these monocarpic shoots were identified during the winter and spring when the water table was close to or above the surface bringing about anaerobiosis/reducing conditions in the surface layers of the soi l , and again in mid-summer . There were also two periods during the year that new vegetative ramet recruitment caused net increases in vegetative shoot densities on the old hollows ; in spring-early summer Page 1 33 and again in late-summer autumn , as on the low dune at S 1 . Mean vegeta tive shoo t densities inc reased at the the three sites in the older deflation hollows over late summer/autumn to the end o f March 1 978 ( figure 4 . 2 ) . This net autumn recrui tment of new vegetative shoots decreased in significance at sites of increasing age (distance from the terminal ho llow ) . These site differences resulted par t ly from the difference in absolute dens ity of total living shoots from which daughter tillers may be produced ( figure 4 . 1 ) , and partly by the difference in the proport ion of these living shoo ts that , on average , gave rise to a new daughter ramet ( table 4 . 1 ) . Both of these factors showed decreases with increasing age of si te . The reduction in vegeta t ive shoo t densities at the three old deflation hollow sites over winter (Apri l to August ) 1 978 mirrored the autumn inc reases , with decreased morta lity both in terms of absolute numbers per uni t area ( fi gure 4 . 2 ) and relative rate ( table 4 . 1 ) at s i tes of increasing senility . This result was similar to that of Noble et al . ( 1 979 ) who found the greatest death rates of Carex arenaria shoots at sites whe re birth rates of shoots were also greatest . Page 1 34 Table 4 . 1 Change in vegetative shoot density of Carex pumila in autumn and in winter 1 978 at four si tes on the study area � relative to the combined fertile plus vegetative shoot densi ty at these sites in January 1 978 . Relative change in sho o t density January-April April-August Nett S 1 - • 05 1 ns + . 544 *** + ** S2 + . 887 *** - . 696 *** + * S3 + • 397 * - . 483 * - ns S4 + . 094 ns - • 1 25 ns - ns Level of significance of the changes at each site are indicated : ns = no t significantly different from zero ; * = P< . 05 ; ** = P< . 01 ; *** = P< . 00 1 . The effect of spring vegetative shoot recruitment was seen at S2 between Augus t and mid-October 1 978 when total living shoo t density inc reased by an amount equivalent to the subsequent midsummer mean vegetative shoot density ( figure 4 . 2 ) . Similar increases in to ta l living shoot densities a t S 3 and S4 later in spring , in October-November 1 978 , were smaller and due to large within-site variability were not statistical ly significant . During the late spring 1 978 and early summer 1 978-79 , mortality of larger and o lder , non-flowering ramets continued at S2 , S3 and S4 in the older hollows , indicated by the decline in density of vegetative shoo ts ( figure 4 . 2 ) . The midsummer mean densi ty of vegetative shoo ts at each of these sites was remarkably similar to that observed twelve months earlier. The rec ruitment of vegetative shoo ts in late summer 1 978-79 was evident at each of the old hollow si tes . Page 1 35 Between February and September 1 979 , a net increase in vegetative shoo t density was observed at S2 , in contrast to a net decline in density of these shoots at the more senile sites � S3 and S4 . Since this was the net effect of autumn recruitment and winter mortality on the old hol lows in 1 978 ( table 4 . 1 ) , it was assumed that a similar pat tern was followed in 1 979 . Death of vegetative shoots on the old hollow continued into spring 1 979 , as in 1 978 . Shoot mortality at S2 and S3 between early September and mid October 1 979 occurred mainly in the older , larger shoo t cohort ( section 4 . 3 . 3 ) . The pattern of change in vegetative shoot dens ity in the old hollow during summer 1 979-80 was similar to but more accentuated than that s een during the previous summer . Recruitment of new vegetative shoo ts resulted in a net increase in vegetative shoo t densi ty leading up to mid-summer at bo th S2 and S3 ( figure 4 . 2 ) . This was followed by a net loss in vegetative shoo t density in late December-January as the vegetative shoot populations aged ( section 4 . 3 . 3 ) . Late summer and autumn recrui tment more than doubled vegetative shoo t density at S2 between January and July 1 980 , following the pattern seen in the two previous years . In the more senile old hol low population at S3 , a very much smaller ne t vegetative shoot density increase occurred between February and April 1 980 , although by mid-winter ( July) , shoot mortality had resulted in a net decline in density of living shoo ts . This net result of autumn recruitment and winter mortality at S3 in 1 980 was observed in both of the two previous years . This accounts for the pattern of declining shoot density with increasing senility of this old population . Page 1 36 Spring shoot mortality at S2 saw the loss of most of the long shoots by October 1 980 (figure 4 . 3 ) . Those long shoo ts surviving were in the main reproductive . In the spring and early summer 1 980�81 ; the decline in dens ity of the combined dwarf plus long living shoo t popu lation at S2 was similar t o that seen in 1 979 . The combined density of the surviving dwarf and long tillers ; which were seen to be fertile was not significantly different from fertile shoo t densi ty found at this site in summer 1 979-80 . At the final harves t in December 1 980 , the dens ity of the surviving vegetative shoot coho rt , of predominantly dwarf tillers , was less than 80 / m2 ; the lowest seen on this part o f the old ho llow at any time during the study . Fluctuations in shoot dens ity in the more senile o ld hollow popu lation at S3 in spring and early summer 1 980-8 1 were small and not statistically signi ficant . As at S2 , the dens ity of the surviving vegetative shoo t cohort of exc lusively dwarf tillers in December 1 980 was the lowest seen here during the entire study . The overall decline in living shoo t densi ties at the two old hollow si tes over the three years o f the s tudy is evident from figure 4 . 2 . The seasonal maximum dens ity of ferti le shoo ts of Carex pumila ob tained at each of the old ho llow si tes moni tored , which may be used as a crude measure of the fecundity of the population , was obse rved in 1 977-78 , the first summer of the study ( table 4 . 2 ) . In each successive year , as the resident population at each of these si tes became progressively more seni le , the mean densi ty of fertile shoo ts averaged over the late spring and summer harves ts declined asymptotically. Thus , whe reas the differences at S2 and S3 between 1 977-78 and 1 978-79 and at S2 between 1 978-79 and 1 979-80 were large and significant (P< . 05 ) , later differences where the popu lations were in more advanced stages of senility were smaller in magnitude ( table Page 1 37 4 . 2 ) . Table 4 . 2 Mean (� standard deviation ) densi ty of fertile shoo ts of Carex pumila at five sites on the study area � in four consecutive summers . 1 977-78 1 978-79 1 979-80 1 980-8 1 so ne ne 0 . 25+ . 08 1 2 . 00+ . 71 ne = not es timated Density / 30 x 30 cm S1 S2 0 2 . 2+ • 3 1 1 • 7+ 1 • 28 9 . 5�1 . 77 30 . 1�1 • 5 20 . 2+1 . 32 1 1 • 6+ 1 • 32 1 7 . 7+6 . S3 1 3 . 4+ 1 . 5 5 . 8+ • 62 1 . 9+ • 48 . 8+ . 42 S4 2 . 5�1 . 77 1 . 9+ • 38 ne ne On the low dune at bo th SO and S1 , younger phases in the development of the resident populations were represent ed . Ferti le shoo·t density at these sites increased be tween years . At S 1 in 1 977-78 and SO in 1 979-80 , the shoot populat ions were almost exclusively vegetative . These represented the juveni le phase of the succession at these si tes . In subsequent years � mean summer fe rtile shoo t density increased to maxima in 1 979-80 at S 1 and in 1 980-8 1 at SO ( table 4 . 2 ) . These represented mature phases of the seral development on the low dune . The decline in fertile shoo t density between summer 1 979-80 and summer 1 980-8 1 was only small . It was assumed however , to represent the beginning of the steeper decline in this parameter between years similar to that seen with the increas ing senil ity of populations in the old deflat ion hollows ( table 4 . 2 ) . Page 1 38 Ogden ( 1 974a ) predicted that for the rhizomatous perennial Tussilago farfara , the dens ity of flowering shoo ts in one season would be closely related to the number of vegetative shoots in the preceding year . His data showed this was approximately the case � at least in some of his experimental treatments . The data for Carex pumila were equally equivocal . Where the populations were mature or early senile (as at S2 in each year of the study , S 1 in 1 979-80 and 1 980-81 and SO in 1 980-81 ) this relationship was approximated . However , where the Carex pumila populations were in a more advanced stage of senility (S3 in 1 980-81 and S4 in 1 978-79 ) , the dens ity of flowering shoots was significantly less than the potential shown by vegetative shoot density in the previous summer . Net mortality of the following years potentially reproductive shoots over later summer, autumn , winter and following spring in the more senile populations accounts for this dec line . The reduced densi ty of fertile shoots in the adolescent populations on the low dune at S 1 in 1 978-79 was accounted for by the massive spring mortality of shoo ts assoc iated with sand accretion on the low dune between August and October 1 978. This fertile shoot density did however represent almost 50% of the surviving October population . The green laminar area per unit ground area ( leaf area index, LAI ) of vegetative shoot populations moni tored in summer 1 979-80 ranged between 0 . 01 and 0 . 364 at S3 and S2 , respec tively ( figure 4 . 5 ) . The LAI of vegetative sho ot s at S3 , which remained less than 0 . 022 throughout the summer , was significantly lower than that at each of the other sites , at which LAI values were remarkably similar in '0 c :J 0 L m 4- 0 (\I E ........ ,..... (\j E .._. ro Q) L CO ..... ro Q.) rl c Q.) Q} L t!l F i gure 4 . 5 Green l e a f area o f vegetat i v e shoots o f Car�x pum i l a per u n i t area o f ground a t four s i tes on the study area in summer 1 979-80 0 . 3 - 0 . 2 - 0 . 1 - . --- . / · - · --- . --- -- · 53 o L_:::-_ · - -Nov Dec Jan Feb 1 980 1 979 Page 1 39 early November ( figure 4. 5 ) . The LAI of vegetative shoots generally increased over the summer at these three si tes , paralleling the changes in shoot density ( figure 4 . 2 ) . As fertile shoo ts mature � their green leaf area declines ( section 4 . 3 . 5 ) . At the beginning of summer 1 979-80 ( in November ) , the LAI of fertile shoots was es timated at 0 . 033 and 0 . 1 23 at S3 and S2 respectively . By early February 1 980 , the LAI of fertile shoo ts had declined to less than 0 . 01 3 at both sites . 4 . 3 . 3 Age and size of shoot populations 1 . Shoot size dis tributions The frequency distribution of the to tal number of leaves per dead Carex pumila shoot at four sites in September 1 979 ( figure 4 . 6 ) shows that shoot mortality occurred in the progressively more senile popu lations when shoots pos sessed fewer leaves . This indicates lesser longevity of shoots and/or a less rapid rate of leaf production in the more senile populations . Since variable numbers of leaves were produced by shoots by the time of shoo t death , both within and be tween si tes (figure 4 . 6 ) , the absolute numbers of leaves per living shoo t could not been taken as a satisfactory guide to shoo t age . A better approximation was the ratio of the number of dead leaves to the total number of leaves per shoot . Shoots with few or no dead leaves sugges ted a young shoot with long life expectancy , produ ced a low dead/total leaf ratio value and so fell into the younger age classes . Shoo ts with predominantly dead leaves , sugges ted an older shoot with a shorter life expec tancy � produced a high dead/total leaf ratio value and so fell into the o lder � 0 c Q) :J 0" Q) (_ l.L F i gure 4 . 6 o f l ea v es per 15 l S i te 1 10 l I : J 15 J S i t e 2 10 . 15 S i te 3 10 5 15 S i t e 4 10 1 2 The f r equency d i str i but ion o f the n umber dead shoot a t f our s i tes in Septemb e r 1979 DoDo DD 3 4 5 6 7 8 9 1 0 Nu mb e r o f l e a v e s Page 1 40 age classes . As aerial shoots of Carex pumila emerged above the surface of the soil they were enclosed by scale-like leaves ( bracts ) . These shoo ts were placed into age class 0 distinct from those older shoo ts which possessed expanded green leaf laminae , but no dead leaves ( dead leaf to total leaf ratio = 0 , age class 1 ) . 2 . Shoot � distributions Figure 4 . 7 shows the age distributions of populations at each of five si tes on the study area between September 1 979 and December 1 980 . Notable features of this figure are : 1 . I n autumn/winter : the skewed ( to the right ) shape of the juveni le populations at SO on the edge of the low dune in both years ; the distinctly bimodal age structure of the more mature populations at S 1 and S2 in both years ; and the flattened (platykurtic ) age dis tribution in the senile populations at S3 in both years and at S4 in 1 979 . 2 . In the early spring: the re juvenes cence ( skewing to the right ) in all populations sugges ting new vegetative shoo t rec rui tment and/or mortality of old shoo ts . 3 . Later in spring/early summer : the identification of fertile shoots within the more aged shoot classes . 4 . During the summer ( 1 979-80 ) : the movement of the fert ile shoot cohort at each site into the older age classes . A similar ageing of the vegetative shoot populations at each si te was also apparent . The living shoot population in the old hollow at S2 in July 1 980 showed a bimodal age distribution of predominantly old , long shoots and predominantly young , dwarf shoots ( figure 4 . 7 ) . The young dwarf shoot age class in July , which included many shoots with up to four f i gure 4 . 7 Frequency d i str ibut i on o f age c l asses o f Carcx illJril i l a shoots at f o ur s i tes on the sand p l a i n �80 ]H l ti­ c no Q) g. 40 U) I S4 G:- 3J �.J 20 '1 1 0 J 10. lrl,� 0 _J, . l-� 60 50 40 30 20 10 0 �iil....kl��- 60 50 40 0 1 2 3 4 5 Si S2 0 1 2 3 4 5 0 i 2 3 4 5 Age c l ass S3 nOn C1L fl_ 0 1 2 3 4 5 � i gure 4 . 7 Age c l ass d i str ibut i ons cont inued . so S 1 S2 83 H l7 60 1 50 40 30 c 40 30 20 10 0 H22.. 50 40 30 20 10 0 � 0 1 2 3 4 5 0 1 2 3 4 5 Age c l ass F i gu re 4 . 7 Age c l as s d i str i but i ons o f s hoots c o n t i nu e d l�c so so H24 Dwa r f 50 40 30 20 10 0 ..&.......�-....�.0..._ 20 L Long 10 0 . ,.Ll_CL__ 40 H25 Owar f 30 20 10 0 Oct 20 L Long 1 0 o :10.:.-o..L...-_ 100 H26 N Dwar f 90 80 70 so 50 40 30 20 10 0 ...L-�""'"��- 30 20 10 Long 0 ��__,..�...,_-r- 0 1 2 3 4 5 S 1 S2 S3 o._ --.-6c_-,.____� Ono__ oe1nnO __:_o ....... Do� C ; control plo15 N = h ttrOJeV\ -fer+dtuJ plots Oom;u:::z· , 0Qrzp , 9Zl . � . 0 1 2 3 4 5 0 1 2 3 4 5 Age c l ass Page 1 4 1 green leaves � accounted for the older vegetative dwarf shoot classes and the dwarf fertile shoot category three months later , in spring. The young age class of dwarf vegetative tillers in October 1 980 included many new recrui ts since winter showing no leaf lamina expansion . In midwinter 1 980 , no significant differences in mean age were observed between site populations averaged over all shoots ( table 4 . 3 ) . However , the data sugges t that the population at S3 was more senile than the others . This difference became significant by October (P< . 05 , table 4 . 3 ) . In October 1 980 , it was possible to identify fertile shoots at S2 , which were significantly older than vegetative shoo ts (P< . 001 , mean age = 0 . 429 and 0 . 26 5 , respectively ) . Table 4 . 3 Mean age ( ratio of dead to total number of leaves per shoot ) at four sites on the sand plain , in winter and spring 1 980 . Mean (� SE ) July October Site 0 0. 222 � 0 . 034b 0 . 1 93 � 0 . 02 b Site 0. 292 � o . 042b 0. 287 � 0 . 03 1 b Site 2 0 . 235 � 0 . 045b 0 . 326 � 0 . 025ab Site 3 0 . 37 1 � 0 . 059ab 0 . 408 � 0 . 043a a > b , p < . 05 In July, more than 40% of shoots at S3 fell into the two oldest age classes ( figure 4 . 7 ) . Living shoots with fewer than five leaves canno t be included in the oldest age class 5 . Thus , populations such as that at S3 , where the total number of leaves pe r shoot was only infrequently greater than five � are likely to be under-represented in this oldest age class . Page 1 42 Dwarf shoo ts fell into lower age classes than long shoots at all sites in midwinter ( July ) 1 980 ( figure 4 . 7 ) . This difference between long and dwarf shoot populations continued to be observed in both spring (October ) and summer (December) 1 980 . However � the ultimate size that could be obtained by a shoot was no t restricted by the type of rhizome segment to which it was attached. Shoots with the maximum number of leaves counted ( 1 1 ) and in the larges t size catego ry ( collar height >20cm) were found attached to both long and dwarf rhizome branches . As the populations aged at each of the four si tes monitored during the second half of 1 980 , site differenc es emerged . New shoot recruitment between July and October at SO and S 1 on the low dune resulted in the mean age of these two populations averaged over all shoots in October remaining similar to that seen in July . By cont rast on the old ho llow at S2 and more markedly at S3 , mean age increased over this period . At S2 in October , shoots that were seen to be fertile were on average older than the vegetative shoot populat ion . However , the most senescent shoots (age class 5 ) were predominantly vegetative , in both dwarf and long shoot catego ries ( figure 4 . 7 ) . Aged vegetative shoots were absent from the living shoot population in December 1 980 . By December 1 980 , the only shoots in the two oldest age classes were fertile except in the most senile populat ion at S3 . At all sites , vegetative shoots were predominantly those attached to dwarf as opposed to long rhizome modules and � with the exception of those at S3 , predominantly young ( figure 4 . 7 ) . The increasing contribution of dwarf shoots to the to tal with increasing age was also shown in table 3 . 8 with the increasing age of the clone back along a single rhizome axis from the tip. Page 1 43 4 . 3 . 4 � weight , energy and crude total nitrogen � unit area of ground 1 • Dry weight As a pioneer of a primary succession, Carex pumila forms more o r less monotypic patches . I have coined the term sward � to express the dry weight of these swards at a given time . Sward mass includes both above- and below-ground parts of both living and dead material of the total vegetation. Figure 4 .8 shows the sward mass obtained at four sites on the study area in January 1 978. Although the bulk of the vegetation at all of these si tes was Carex pumila , other species progressively increased proportionately at sites of greater putative age , at increasing distances from the terminal ho llow . The estimated age of the vegetation at S1 in January 1 978 was 0-5 months ( section 3 . 2 . 2 ) . The sward mass at this site , whi ch enti rely comprised living shoots of Carex pumila , was low. The difference between this and each of the other three older sites could be attributed to the greater opportunity ( time ) for accumulation of dry matter . The decreasing sui tability of the older defla tion hollow sites for growth of the pioneer , Carex pumila � was reflected in ( 1 ) the decreased sward mass , ( 2 ) the increased proportion in other spec ies ( figure 4 . 8 ) , (3) the decreased aerial biomass of Carex pumila and ( 4) the increased proportion of herbage mass of this species contributed by dead shoots ( figu re 4 . 9 ) . Dead aerial shoo t modules of Carex pumila are continually being lost through decomposition and removal by wind and other disturbance facto rs . Thus , their herbage mass can be expected to be reduced at the older deflation hollow sites � since this will tend to reflect the herbage mass of living shoots of the species ( figure 4 . 9 ) . A I I I - I Figure 4 . 8 The sum of the sward mass o f Carex pumila and other spec ies, at each o f four s ites on the sand p l a in near Tang imoana, i n January 1978 . S i S2 S3 S4 S i te 100 80 E (.) 0 (T) X 60 0 (T) L Q) a. :::E 40 0 (/) E CO L Cl 20 Figure 4 . 9 The sum of the herbage mass o f vegetative, ferti le and dead shoots of Carex pumila at f our s ites on the sand p lain, in January 1978 . Si S2 Site S3 Dead Shoots Fert i l e shoots Vegetat ive shoots S4 Page 1 44 relatively slower rate of loss of underground parts than of aerial parts , through slower decomposition rates and/or the absence of removal by wind as a factor , would result in the greater accumulation of roots and rhizomes re lative to herbage . This accounts for the difference in shape of the plots of sward mass and herbage mass for this species across the sand plain ( figure 4 . 8 and 4 . 9 respectively) . The shape of these plots , like that of shoot densities ( figure 4 . 1 ) ; are sugges tive of the n-shaped curve described by Watt ( 1 947 ) expressing " the general course of change in total production" for vegetation displaying phasic development . Living shoo t populations of Carex pumila at these four si tes , monitored at the beginning of the study , were also c learly distinguished by the mean size of aerial shoo t modules ( figure 4 . 1 0 ) . Figure 4 . 1 0a des cribed an n-shaped plot similar to that seen for herbage mass of these same live shoot populations ( figure 4 . 9 ) . Fertile sho ot modules were on average larger than vegetative modules ( figures 4 . 1 0a and b) , indicating the greater average age of the former. This however , masks the diversity of both size and age of individuals within these populations ( figure 4 . 7 ) . Inclusion of roots and rhizomes in the comparison of mean size of branch modules results in the disappearance of the distinction between the juvenile ( S 1 ) and mature (S2 ) populations ( figure 4 . 1 0b ) . Figure 4 . 1 1 shows the standing herbage mass of living shoots of Carex pumila ( aerial biomass , AB ) at sequential harves ts at each of four sites on the study area during 1 978 and 1 979 . From aerial biomass and shoot density, the dry weight of aerial shoo t modules at each site was calculated at each harves t and also plot ted over time ( figure 4. 1 2 ) . In spring and summer , fertile shoots were shown +.l 0 0 .c. en ........ - Cl - +.l .c. Cl ·rl Cl) � >- '- 0 F igure 4 . 10 (a) The mean dry we i ght p er aer ia l shoot module o f Carex pum i l a at f ou r s i tes o n the s a nd p la i n in January 1978; f o r ( i ) v e ge­ tat ive , ( i i ) f er t i le and ( i i i ) d ead shoots (Ver t i ca l bars d enote +- s tandard dev i a t ion ) 2 ( i ) vegetat i ve shoots 1 . 6 1 . 2 0 . 8 0 . 4 2 ( · · ) f ert · l e shoots 1 1 1 1 . 6 - T 1 . 2 - .l 0 . 8 - 0 . 4 - 0 2 ( i i i ) dead shoots 1 . 6 1 . 2 0 . 8 0 . 4 S i 52 Site -r-� I I 53 54 Figure 4 . 10 �) The mean dry we i ght per branch module (inc lud ing undergr'ound par'tsl of Carex pum i l a at four s i tes on the sand p l a i n i n January 1978; for (i ) vegetat ive and ( i i ) fert i le shoots . 1 ( i ) vegetative ! : 1 . 6 - 1 . 2 - 0 . 8 - ,-- - --l ! I I I I i 0 . 4 - I J ! 0 - --- ___ [ - l - - ! ,- -- -- �---] j _ l _ _ _ _ I _ _ _ _ _ _ __ I I I ( i i ) fert i le shoots 2 ____; 1 . 8 _; 1 . 2 - I i o . s � I r-- ___ l . : I : . : ! o . 4 -; ; 1 I 1 � - - - � o L _ _ ___ _ _ _ j _____ LL ___ U __ __ L_· S i S2 S3 S4 S i te E 0 0 (T) X 0 (T) ' ::::E 0 Ul E ro '- Cl Figure 4 . 1 1 The sum of t h e aer ia l b iomass o f fert i le and vegetative shoots on the sand p la in, dur ing 1978 and 1979 20 i (a) s ite I 1: m I I n 8 ,,Q flU I � m 80 (b) s ite 2 70 . 60 � • Fertile shoots � 50 � Vegetative shoots � � 40 30 20 10 � � � . � � � � . � � � I � � � � I I I 0 20 i (c) s ite 3 1 : _ � � A D a ij R li a 9 " m !: ] {d) s:te 4 • a a a a • m " a A • Jan Feb Mar Apr May Jun Ju l 1978 Aug Sep Oct Nov Oec Jan Feb Mar Apr May 1979 Jun Ju l Aug Sep t ("\ CD r-i :J D 0 E 4-l 0 0 .c. (f) -........ 2: 0 (f) E ro c... Ol F i gu re 4 . 1 2 (a ) D i str i b u t i o n o f the a er i a l b i o ma s s p e r f er t i1 e s h o o t o f Carex p u_m i l a to c omponent p arts over t i me at f o ur s i tes dur i n g summer 1 978-79 � ( i ) s i te 1 0 . 5 J :J 0 . 2 0 . 1 • . 0 . 8 ( i i ) s i t e 2 0 . 7 - 0 . 6 . 0 . � . 0 . 4 � - 0 . 3 - 0 . 2 - 0 . 1 - 0 ( i i i ) s i te 3 0 . 2 0 . 1 � 0 n ( i v ) s i te 4 "j 0 0 Oct s eeds sp ikes vegeta p �_n_l lJu � � n Nov Dec Jan Feb 1 978-79 t i v e arts Figure 4 . 12 (b) D istr ibut ion o f the b i omass per fert i l e shoot o f Care� pumi la to component parts over t i me at three s i tes dur ing summer 1979-80 0 6 ( i ) s i t e 1 . l 0 . 5 --- 0 . 4 __; 0 . 3 .c � . I 0 . 2 - i t 0 . 1 � �: . � � 0 j_ ______ . 1 . 5 l ( i i ) s i t e 2 1 . 4 __: 1 . 3 _: 1 . 2 - w 1 . 1 - r-1 :-, "0 1 ....: 0 E .p 0 0 . 8 - 0 0 . 8 -.c Ul '-.. o . 7 --: 6 0 . 8 - Ul E 0 . 5 ....: 10 c... 01 0 . 4 - 0 . 3 � I 0 . 2 � ' 0 . 1 � 0 J ( i i i ) s i t e 3 Oct Nov Oec 1 879-80 R roots [ ! rhi zomes n 1 i seeds I . , I l l � sp i kes I . i l I , i ! dead leaves 1 1 ! I I ! j j ! · I ! I green sheaths sca le leaves I green laminae - . . - - . . . . ----- Jan Feb Page 1 45 separately from vegetative shoots in both figures 4 . 1 1 and 4 . 1 2 . For those intervals that standing ae rial biomass increased over time ( figure 4 . 1 1 ) , the mean dry weight per aerial module did no t necessarily also increase ( figure 4 . 1 2 ) . The relative increases in shoot recruitment and growth of newly recruited shoots will account for this discrepancy. At harves ts from September 1 979 � rhizome segments of Carex pumila � with at tached adventi tious roots � were separated with their aerial shoo t portions intac t . Thus , total (aerial plus underground ) biomass of Carex pumila populations was subsequently followed over time ( figure 4. 1 3 ) . Figures 4 . 1 1 and 4 . 1 3 show seasonal fluctuations in biomass of Carex pumila . Longer term trends shown in figures 4 . 1 1 and 4 . 1 3 are similar to the changes in shoot dens ities of these same populations with time ( figure 4 . 2 ) and of biomass differences be tween sites of increasing age at increasing dis tances from the terminal hollow ( figure 4 . 9 ) . Standing biomass of vegetative shoo t populations of Carex pumila increased over autumn at all si tes except where the population was in an advanced stage of senility . Increases were also seen in spring and early summer in the total shoot population . The latter increases can be attributed to vegetative shoot recruitment and growth , and to the growth in size of shoo ts which were fertile ( figure 4 . 1 2 ) . Net losses of biomass occurred in winter at all sites , except in the most juvenile populations . Aerial biomass of vegetative shoots increased over autumn 1 978 , between January and Apri l at each site , except S4 where the Carex pumila population was already in an advanced senile s tage of development ( figure 4 . 1 1 ) . The average size of shoots at S2 also increased (figure 4 . 1 2 ) . At S1 where the population was in a juveni le stage of development , the increase in aerial biomass continued to the E 0 0 20 ('I) 15 X 0 ('I) " 10 L 0 Ul � 5 '­ Cl (a) s ite 0 Figure 4 . 13 The sum of th e b iomass o f fert i le and vegetative shoots o f Carex pum i la at f our s ites on the sand p la in, dur ing 1979 and 1980 c===J fert i le shoots vegetat ive shoots 0��----��������--�--��--------��--------------------�--------------�----------�-- 15 ., (b) s ite 1 10 5 0 I W m mm m W M � M U W W W M Sep Oct Nov 1 979 Dec Jan Feb Mar Apr May Jun Ju l 1 980 Aug Sep Oct Nov Dec � SI E 0 0 (Y) X 0 (Y) .......... :::E 0 m E 10 (_ Cl Figure 4 . 13 cont inued (c) s ite 2 30 25 c===J fertile shoots 20 15 10 5 � � vegetative shoots � � � � � � � � � � � � � I � � � �� � � � �� � � � � �� � � � � � � � I �� � � � I �� � � � � �� � � � � � � � � � �� � I � � � �� � � � � � � � I � � � � 0 fd) s1te a : � � � � ra � m � � I Fa ra = Sep Oct Nov Dec 1979 Jan Feb Mar Apr May 1980 Jun Jul Aug Sep act Nov Dec � (51 IJ I I . I Page 1 46 end of winter 1 978 , whereas at the o lder sites in the damp deflation hollows , a net loss over winter was observed . A net loss of aerial biomass was subsequently observed in early spring 1 978 at S1 associated with the inundation of the site by sand . The reduction of aerial biomass of Carex pumila in the deflation hollows ( S2 , S3 and S4 ) during the winter and early spring and on the low dune (S 1 ) in early spring 1 978 coincided with the mortality of principally old vegetative shoots ( sections 4 . 3 . 1 and 4 . 3 . 3 ) . Many of the larger , olde r , living shoots present in the populations in early spring 1 978 subsequently proved to be fertile . Figure 4 . 1 2 shows the changes in the average size of fert i le shoots and of each component organ at each site over time , in summer 1 978-79 . The increases in mean aerial dry wei ght per fertile shoot over time can be attributed to increases in both vegetative and ferti le fractions ( fi gure 4 . 1 2 ) . This may result from the mobilization of resources found in the rhizomatous fraction and/or the as similation of carbon in bo th vegetative and fertile aerial shoot parts ( section 2 . 6 ) . From mid -summer , decreases in the mean dry wei ght of both vegetative and fertile shoot components were observed ( figure 4 . 1 2 ) . Decreases in the green leaf and stem frac tions could be assumed to result from firs tly , the senescence of these parts and concomitant increase in the dead leaf fraction and secondly , from the redis tribution of resources to seeds . Decreases in dead leaves and in seeds could only be explained by loss . The standing mid-summer 1 978-79 aerial ( vegetative was greate r than shoot ) biomass at S1 in that at either S3 or S4 and remained so during autumn and winter 1 979 despite the ne t loss of herbage over this pe riod at all of these sites ( figure 4 . 1 1 ) . The standing aerial biomass of Carex pumila was greater at S2 than at any Page 1 47 of the others , throughout the entire 20-month period shown in figure 4 . 1 1 . This can be attributed firstly, to the larger net accumulation of herbage at this si te in autumn 1 978 and to a lesser extent in autumn 1 979 despite the large losses also recorded � and secondly� to the greater proportionate allocation of biomass to aerial parts at S2 compared with other sites . Over the period shown in figure 4 . 1 1 , roots and rhizomes o f Carex pumila were bulked for all shoot modules . Estimates of the biomass of roots and rhizomes at tached to living shoot modules during this period were made by use of the ratio of the biomass of living to total living plus dead shoot modules ( figure 4. 1 4 ) . At those sites where Carex pumila was a recent colonist all branch modules were living and the ratio equalled unity ( S 1 in January 1 978 and SO in October 1 979 ) . As the populations at these sites aged and Carex pumila modules died , an asymptotic decline in the ratio occurred . The ratio showed deviations from strict curvilinearity reflecting the balance of shoot births and deaths , and of decomposition and biomass accumulation. At the late winter 1 979 harvest � the standing biomass values of the Carex pumila populations at S 1 and at S2 were similar ( fi gure 4 . 1 3 ) , despite the difference seen in figure 4. 1 1 for standing ae rial biomass . Total biomass at S1 and S2 was significantly ( P< . 001 ) greater than those values ob tained elsewhere on the sand plain � at S3 and S4. Since biomass estimates in the two most aged Carex pumila populations were low , and reduced compared with the previous year , it was decided to discontinue harves ts at one of these � namely S4 . 1 1 . 7 sr- I .� I ;o . 5 0 1- � . 2st- so figwe 4. 14 The ratio of the dry weight of l iv ing to total ( l iv ing plus dead) shoots of Carex pumi la over t ime, at four sites on the sand plain 5 .0 Page 1 48 The net losses of biomass observed at all sites over the winter of 1 979 in the aerial shoot fraction (fi gure 4. 1 1 ) ; continued into the spring to mid-November 1 979 . These reductions seen in to tal biomass of living shoots between harves ts at S1 � S2 and S3 ( figure 4 . 1 3 ) can be attributed to the mortali ty of principally older ; larger vegetative shoots ( see section 4 . 3 . 1 and 4 . 3 . 3 ) and mirror those described for aerial biomass during winter/early spring 1 978 ( figure 4 . 1 1 ) . The subsequent spring 1 979 and summer 1 979-80 net increases in biomass of the vegetative shoot population at each site monitored ( figure 4 . 1 3 ) also mirro red those seen for aerial biomass twelve months previous ly ( fi gu re 4 • 1 1 ) • At SO not previously monitored , the net increases in biomass over spring and early summer 1 979-80 resulted in the summer mean standing biomass of this juvenile population being similar to that found in the old hollow at S2 ( figure 4 . 1 3 ) . The S2 population was still distinguished however , by the standing crop of fert ile shoo ts ( figure 4 . 1 3 ) . Fertile shoots were found only in some plots and at some of the summer 1 979-80 harvests in vegetative shoot biomass at this young S1 and S3 population at SO . The in mid-summer 1 979-80 was significantly reduced compared with that at bo th SO and S2 . The populations at S 1 and S3 could also be distinguished from each other by the biomass of fertile shoots ( figure 4 . 1 3 ) . During 1 980 , further net increases in biomass by vegetative shoots was seen over autumn at all but the newly monitored SO on the edge of the low dune ( figure 4 . 1 3 ) . Deposition of wind-blown sand on the edge of the low dune during February 1 980 (figure 4 . 4b ) resulted in net loss of biomass at SO between the late summer (February) and autumn ( April ) harvests . The Carex pumila population subsequently Page 1 49 observed at this site included numerous new young shoots . Net loss of biomass in early winter 1 980 (April-July ) was also evident higher on the low dune at S 1 and on the waterlogged deflation hollow at both S2 and S3 . In winter and spring 1 980 , the living shoot population of Carex pumila at each site was divided into those attached to dwarf and long rhizome modules , and in the subsequent summer � each of these two types of shoot was further divided into fertile and vegetative module s . The mean size of the long shoo t type was consistently higher than that of the dwarf type both on the basis of to tal (aerial plus underground ) and aerial dry weight pe r branch module , throughout the period July to December 1 980 ( figure 4 . 1 6 ) . This difference can be relat ed to the relative ages of these populations ( figure 4 . 7 ) . The maximum size that could be obtained by an individual shoot at tached to either a dwarf or long rhizome module was the same ( section 4 . 3 . 3 ) . In July 1 980 , the biomass of dwarf and long shoots per uni t area of ground was greater at S2 than at any of the other si tes ( figure 4 . 1 5 ) , as a result of the densi ty of each of these two types of shoo ts. The mean dry weight per shoot module for both dwarf and long shoots was no different between S2 and any of the other si tes ( figure 4 . 1 6 ) , despite differences in age structure ( figure 4 . 7 ) . In 1 980 , the now famil iar spring/summer net inc rease in population biomass of Carex pumila began earlie r on the low dune at SO and S1 than on the deflation ho llow ( figures 4 . 1 3 and 4 . 1 5 ) . The increased stability of the low dune in 1 980 as a result of the planting of Ammophila arenaria on the surrounding dunes ( figure 3 . 4 ) accounted in part for this difference . The increasing senility of the populations on the deflation ho llows and the high water table in late Figure 4 . 15 The sum of th e b iomass o f dwar f and long shoot populat ions of Carex pumila (grams OM I 30 x 30 cm) over time at four s i tes on the sand p la in between Ju ly (winter) and December (summer) 1980 20 (a) s i te 0 Dwar f fert i le shoots E 15 0 0 Cl) X 0 Cl) L Q) D. 2 0 (f) E m L 0) 10 5 0 5 0 15 10 5 (b) s i te 1 (c ) s i te 2 0 ""'-- 3 Ju l t,.,,. Au g Dwarf vegetat ive shoots Long ferti le shoots Long vegetat ive shoots Se p Oc t 1 9 8 0 No v De c -+-l 0 0 .c (f) "-... 2.: 0 (f) E ro (_ (J) F i gure 4 . 1 6 The d i s tr ibut i o n o f b i omass o f c omponent p arts o f ( a ) dwar f a n d ( b ) l o ng branch modu l e s o f Carex pum i l a a t f our s i tes o n the sand p l a i n i n Ju l y , October and December 1980 (a ) Dv.t a r f b r a n c �1 e s ( i ) s i t e 0 1 . 4 1 . 2 1 0 . 8 0 . 6 0 . 4 0 . 2 . '·' 0 ( i i ) s i te 1 0 . 4 0 . 4 l ( i i i ) s i t e 2 0 . 2 0 � 0 . 4 ( i v ) s i te 3 0 . 2 Roots F f e r t i l e s h o o t s Rh i z omes V v e g e t a t i v e s h o ots Seeds Anc i l lary f er t i le structure s Dead leaves Green l a m i nae Green s t1 eat t1s F Bracts 0 �----�------·------���·---------- -���---- J u l Au g Sep O c t No v De c 1 9 8 0 1 . 6 1 . 4 1 . 2 1 0 . 8 0 . 6 0 . 4 0 . 2 .c 0 0 . 8 c m L ..0 0 . 6 ......___ 2:: 0 . 4 0 (f) 0 . 2 E m L Ol 0 0 . 8 0 . 6 0 . 4 0 . 2 0 0 . 4 0 . 2 0 F i gu r e 4 . 1 6 c o n t i n u e d ( b ) L o n g s h o o ts ( i ) s i te 0 ( i i ) s i t e 1 ( i i i ) s i t e 2 ( i v ) s i te 3 Ju l Au g Roots Rh i zomes Seeds Anc i l l ary structures Dead l e aves Gre e n l am inae Green sheaths Bracts j Sep O ct 1 9 8 0 14-9 (.., . Nov D ec winter also contributed to this difference . The net biomass increases in the two older populations at S 1 and S2 were small in both the dwarf and the long shoot components ( figure 4 . 1 5 ) . By contrast , the increases in the younger population at SO occurred at a higher rate than elsewhere on the sand plain � especially over the late spring/summer period from October to December 1 980 ( figure 4 . 1 5 ) . The increases at SO between July and October 1 980 in both long and dwarf shoo t components ( figure 4 . 1 5 ) resulted from an increase in densit ies of these two shoot types ( figure 4 . 3 ) and also from an increase in the average size of dwarf shoots ( figure 4 . 1 6 ) . The October-December 1 980 increases ( figure 4 . 1 5 ) were largely at tributable to the increase in mean size of bo th dwarf and long shoots ( figure 4 . 1 6 ) . Thus , the standing biomass of Carex pumila at S2 was surpassed for the first time in the present study by that at another site , namely SO ( figures 4 . 1 3 and 4 . 1 5 ) . In December 1 980 , the mean dry weight per aerial shoot module was affected not only by the type of shoot ( dwarf versus long) but also by their sexual state ( vegetative versus fertile ) , by site ( figure 4 . 1 7 ) and by nitrogen fertilizer addition ( section 4 . 3 . 7 ) . Averaged over the four sites , mean size of fertile shoo ts was greater than that of vegetative shoo ts (P < . 01 ) and long shoots greater than dwarf shoo ts (P < . 01 ) , effec ts attributable to the relative ages of these shoot populations ( figure 4 . 7 ) . On the contro l plots (no added ferti lizer ) at the four sites studied , the mean size of aerial shoo ts was significantly greater at SO than at al l other si tes , for each shoot catego ry represented at SO (P< . 001 ; figure 4 . 1 7 ) . These differences may also be a difference in the mean age of shoo ts . The age distribution of dwarf vegetative shoots at SO in December 1 980 indicated an older population than that at other sites ( figure 4 . 7 ) . 1 F igure 4 . 17 The mean aer i a l dry we ight per shoot o f dwar f (D) and l ong (L) , f ert i le (F) and vegetat ive (V) shoot modules o f Carex pum i l a at four s ites on the sand p la in in December 1980 Leiiers (o. -f) re+e.,.-- -lu Duii\Co<"\'5 Y>Iu.. l+-,p le.- ·��e... t-est. DV OF LV LF a +-' 0 . 8 i ,...... a 0 0 .r:. 1/) ' 0 . 8 � b ::::E Cl 1/) E � 0 . 4 � I I I I I I c. Cl 0 . 2 de ef e.f e.t 0 I I I I I I 11 I I I I I I I I I I I I I I 11 I I I I I SO S 1 S2 S3 SO S 1 S2 S3 SO S 1 52 S3 SO S i 52 S3 01 0 p Page 1 5 1 In summer 1 979-80 , no direct evidence was obtained for the redis tribution of resources stored in rhizomes at tached to fertile shoots during flowering and seed production. Over this peri od , no change was observed in mean dry weight of rhizomes per fertile shoo t above.. b ( see be±Gw, figure 4 . 1 2 ) . Further , the mean dry weight of the rhizome fraction per branch module was no different between rhizomes at tached to vegetative and to fert ile shoo ts ( table 4 . 4 ) . These observat ions support the view that the rhizome of Carex pumila ac ts not so much as a storage organ releasing accumulated carbon to other shoo t part s , but more as a migratory organ like that of other " creeping" sedges ( Salisbury 1 942 ) allowing local c oloni zation and resource capture . Table 4 . 4 Mean dry weight per rhizome segment at tached t o ferti le , vegetative and dead shoots at four si tes on the sand plain in summer 1 979-80 . grams DM / branch Shoot category so S 1 S2 S3 Vegetative . 446 a . 1 57 b • 1 39 b . 063 b Ferti le . 1 56 b . 1 29 b Dead . 42 a . 564 a . 42 a . 565 a a > b ; P < . 001 ; Duncan ' s multiple range tes t . However , the es timation o f mean dry weight o f rhizomes per branch module during this summer was complicated , first ly by dwarf and long rhizome segments attached to live aerial shoo ts not being separated , and secondly by long rhizome sympodial segments being longer than the Page 1 52 dimensions of the quadrat laid down for sampling (figure 3 . 20 ) . In summer 1 980-8 1 when dwarf and long rhizome segments were separated from each other , the evidence shows that the dry weight of rhizomes of fertile shoot modules was reduced compared with that of vegetative shoots ( table 4 . 5 ) . In Carex arenaria ; a sand sedge of similar rhi zome archi tecture to Carex pumila , up to two-thirds of the seasonal variation in d ry weight of rhizomes could be accounted for by the variation in carbohydrate content ( from figure 7 , Noble 1 982 ) . Thus , given a similar determination of dry weight of rhizomes of Carex pumila by carbohydrate content , differences shown in table 4 . 5 might also be attributed to differences in stored carbohydrate . Likely major sinks for such res erves in o lder rhi zomes inc lude seeds and younger rhizomes in more dis tal parts of the clone . Table 4 . 5 Mean dry wei ght per rhizome segment , December 1 980 . C = control , N = nitrogen fertilizer treatment . Dwarf Long vegetative ferti le vegetative fertile so c . 1 80 . 07 1 • 530 . 638 N • 09 1 . 079 1 . 446 . 79 S 1 c . 022 . 022 . 5 1 5 . 234 N . 040 . 039 • 526 • 1 45 S2 c . 025 . 01 9 . 440 . 296 Differences obtained in spring and summer 1 979-80 in te rms of the dry weight of rhizomes (per uni t length and per sympodial segmen t � tables 4 . 6 and 4 . 4 respectively ) can be related to the prop9rtion o f long as opposed t o dwarf rhizome segments in the populations . A s the Page 1 53 populations aged , an inc reasing proportion of the live shoot population can be expected to be made up by dwarf as opposed to long shoots ( table 3 . 8 and section 4 . 3 . 6 ) . The diameter and length of dwarf or short rhizome segments were si gnificantly less than that o f long rhizome segments ( section 3 . 2 . 2 ) and therefore the former can b e expec ted t o store less carbohydrate per uni t length and per rhizome segmen t . Thus , the mean dry weight per rhizome segment attached to vegetative shoots de creased at sites of increasing age ( SO to S3 , table 4 . 4 ) . At SO , where the Car ex Eumila popu lation was in a adolescent stage of development ( made up by predominantly long branch modules ) , the mean dry weight of rhizome segments attached to vegetative shoo ts was highly significantly greater than at each of the other three sites ( P< .001 ; table 4 .4 ) and comparable with the values obtained for long branch segments in December 1 980 ( table 4 . 5 ) . In contras t in the o ldest population at S3 , rhi zome weight was similar to that found for dwarf rhizome segments in December 1 980 ( table 4 . 5 ) . Also , the mean dry wei ght per rhizome segment attached to dead shoo ts was greater than that of rhizomes at tached to live shoo ts ( table 4 . 4 ) . At a given si te , the dead shoot populat ion can be expec ted to be made up by a greater proportion of long branch modules than the current live shoo t population. The mean dry weight pe r rhizome segment at tached to dead shoots at the three o lder sites studied was similar to that at SO where the shoo t population was made up of predominantly long branch modules . In September 1 979 , rhizome segments attached to dead aerial sho ots were found to be significant ly heavier per uni t length than those attached to green shoot modules at all si tes except at S2 ( table 4 . 6 ) . Noble ( 1 982 ) noted that in Carex arenaria , rhizomes may remain Page 1 54 alive and contain substantial amounts of starch for several years after losing their aerial shoots through death and subsequent decay . This may also apply to Carex pumila in which live dwarf branch modules were seen to be growing from old long rhizome segments � the aerial portions of which were decayed ( figure 2 . 4 ) . The mean dry weight pe r unit length o f rhizome at S2 reflects the larger capacity of the shoo ts at this compared with other sites to generate supplies of carbohydrate for storage ( see figures 4 . 1 0 and 4 . 1 2 for mean shoo t size ) . The range of values for the dry weight per uni t length o f rhizomes obtained a t S 2 ( namely 2-4 g/m) was similar t o the maximum values reported by Noble ( 1 982 ) for Carex arenaria . Table 4. 6 Mean dry weight per unit length ( g/m) of rhizome segments attached to living and dead shoot modules at five sites on the sand plain , in spring 1 979 . Dry weight of rhi zome segments ( g/m ) Site 0 Site Site 2 Site 3 Site 4 Live 1 . 22 ab 2 . 1 1 de 2 . 72 f 1 . 1 3 ab 0 . 85 a Dead 2 . 55 f 2 . 55 ef 1 . 73 cd 1 . 56 be Letters indicate Duncan ' s mul tiple range tes t , P < . 05 2 . Net biomass accumulation rates Net herbage accumulation rate (NHAR) over a particular sampling interval may be calculated in many ways from standing herbage mass data obtained at sequential harves ts . Trough-peak analysis of live Page 1 55 herbage material which involves the summation o f posi tive increments in aerial biomass per unit area of ground (AB ) between harves ts (method 6a of Singh et al . 1 975 ) may be expressed : k * NHAR = � (AB n - AB n-l ) / ( t K-t0 ) • • • • • • • • • • • • • • • • • • • • . ( 1 ) h=- 1 where harves ts are at times tO to tk ; * indicates that only positive increments of aerial biomass are included . Where standing herbage mass is divided into living and dead shoo t components , then positive increments in dead material may be included in this calculation where they coincide with posi tive increments in live material ; the rationale is that where live and dead materials increased over the same sampling interval then new production mus t have been rapidly transferred from the live to the dead component (Method 7 of Singh et al . 1 975 ) . This may be written : where DH is standing herbage mass o f dead shoo ts and � indicates that positive increments of dead material will only be included where this coincides with posi tive .aerial biomass increments . This calculation does not allow for losses by physical removal by disturbance fac tors such as grazing animals and wind , decay or translocation to underground o rgans . NHAR of a single shoot cohort may be calculated similarly by summing positive increments in mean aerial biomass per shoo t (ABS ) between harves ts and multiplying this by the mean shoot density of this shoot cohort ( SD ) : k NHAR = 2_ ( ABSn r'\: 1 * ABS n_ 1 ) SD I ( tk- t0) • • • • • • • • • • • • • ( 3 ) Page 1 56 Equation 3 will be useful in those situations where spatial variability in shoo t density , results in the biomass per unit area showing a variability in space that is at least as large as the variations over t ime . The summation of positive increments in biomass per uni t area be tween consecutive harves ts ( equation 1 ) will run the risk of inflating estimates of net herbage accumulation by creating a mythical popu lation in which apparent increments in biomass be tween harves ts , which may at leas t in part result from spatial variation , will appear to occur on a common uni t area of ground . Estimates of net herbage accumulation by vegetative shoo t populations at each site were made by between consecutive harves ts in figure 4 . 1 1 application of equation 1 ( table 4 . 7a ) . Positive increments in dry weight of dead shoots per unit area that coincided with positive increments in aerial biomass of vegetative shoots ( equation 2 ) were also calculated in 1 978 ( table 4 . 7b ) . By substi tuting total standing biomass ( SB ) for aerial biomass (AB ) , equat ion was also applied to the total biomass of vegetative shoo t populat ions during 1 979 and 1 980 , using figures 4 . 1 3 and 4 . 1 5 ( table 4 . 7c ) . From July 1 980 dwarf and long branch modules were treated separately for this exercise . In summer 1 978-79 and 1 979-80 , the analysis on ferti le shoot populations was based on both equation and equation 3 ( table 4 . 7d ) . The latter estimate was the more conservative . Page 1 57 Table 4 . 7 Net accumulation of biomass of Carex pumila at sites ( SO-S4 ) between successive harvests . - indicates ne t biomass loss . ( a ) Herbage estimates for vegetative shoots based on equation grams DM I m2 I day S 1 S2 S3 S4 H1 -H2 . 64 9 - 79 . 6 H2-H3 . 28 H3-H4 H4-H5 . 56 • 1 6 H5-H6 . 03 3 . 1 9 • 1 9 H6-H7 1 . 55 3 . 87 H7-H8 . 46 1 . 04 - 33 H8-H9 . 24 . 08 H9-H1 0 3 - 92 1 . 39 . 38 . 02 m ?=Hl � • 01 1 . 64 . 06 . 06 - ( b ) Positive inc rements in dead shoots pe r uni t area that coincided with posi tive increments in ( a ) . grams DM per m2 S 1 S2 S3 S4 H1 -H2 1 . 67 H2-H3 53 . 89 H3-H4 H4-H5 H5-H6 1 6 . 24 H6-H7 H7-H8 47 . 22 H8-H9 1 1 • 36 5 . 06 H9-H1 0 92 . 44 H1 0-H1 1 22 . 89 H1 1 -H1 2 Page 1 58 ( c ) Total biomass increments for vegetative shoots based on equation during 1 979 and 1 980 . grams DM I m2 I day so S 1 S2 S3 H1 O-H1 1 2 . 32 0 . 1 36 H 1 1 -H1 2 H 1 2 -H1 3 ( 0 . 32 ) 0 . 1 58 H1 3 -H1 4 4 . 07 4 . 01 5 . 82 0 . 94 H1 4-H1 5 2 . 63 2 . 25 H1 5-H1 6 H1 6-H1 7 1 . 24 1 . 1 3 2 . 09 H1 7-H1 8 1 4 . 74 1 5 . 48 1 . 86 H1 8-H1 9 ������� 8 : �� o . 1 1 2:71 6 : �� a : 1 � H23-H24 H24-H25 Dwarf 0 . 57 0 . 07 0 . 1 1 0 . 002 Long 0 . 28 H25-H26 Dwarf 1 . 20 0 . 02 o . 1 6 Long o . 1 1 0 . 07 0 . 1 6 ( d ) Seas onal herbage increments for fe rtile shoo ts based equation 1 , ( ii ) equation 3 ( fi gure 4 . 1 2 ) and mean shoot and ( iii ) equation 1 ( fi gure 4 . 1 5 ) . ne = not estimated . on ( i ) density so ( i ) 1 978-79 ne ( ii ) 1 978-79 ne 1 979-80 1 . 1 7 ( iii ) 1 980-8 1 44 . 00 S 1 1 7 . 67 7 . 61 29 . 77 grams DM I m2 S2 S3 268 . 9 31 . 33 96 . 85 9 . 1 2 85 . 23 5 . 70 1 7 . 22 S4 7 . 45 0 . 50 ne ne The greatest rates of net biomass accumulation that were es timated at any site during the study were found at S2 during autumn 1 978 and spring 1 979 . The autumn 1 978 increase measure d at 9 . 79 g DM I m2 I day which represented a five-fold increase in standing herbage ( figure 4 . 1 1 ) compares with rates of production of green mate rial in sedge dominated wetlands in North America ( Bernard and Macdonald 1 974 ) . The spring 1 979 increase was greater ( 1 5 . 5 g DM I m2 I day ) , but inc luded unde rground organs . Page 1 59 The estimated daily rates o f net biomass accumulation for long and dwarf shoots of Carex pumila from July to October and from October to December 1 980 are given in table 4 . 8 . No significant differences were found between the two shoot types . The maximum rate estimated over these periods which were found at SO to be 2 . 04 g DM I m2 I day for the combined dwarf and long, vegetative and fertile shoo t population was still considerably less than that es timated for aerial shoot growth in the old ho llow at S2 in autumn 1 978 , or for to tal biomass at S2 in spring 1 979 . Table 4 . 8 Estimates of aerial shoot growth rates of long and dwarf shoots at sites on the sand plain during sp ring and early summer 1 980 . SO dwarf shoots long shoot s Aerial shoot rate o f increase in size (mg DM I shoot I day ) July-October 2 . 62 2 . 72 October-December 7 . 1 3 6 . 34 During 1 980 biomass accumulation at SO was reduced by the effect of grazing by rabbits . This form of habitat disturbance was evident on young shoo ts , especially across the fringe of the low dune throughout the study . Figure 3 . 5 shows evidence of rabbits in mid-summer 1 977-78 . In late August 1 980 , almost one month after the winter harves t was taken , young shoo ts on the fringe of the low dune were extensively damaged by grazing ( fi gure 4 . 1 8 ) . At the spring 1 980 harvest previous ly grazed shoots which could readily be identified were included in the sample at this si te . The number of leaves per shoot to have elongated since grazing, and the inc rease in length of the longes t leaf per grazed shoot are shown in table 4 . 9 . No Page 1 60 differences between dwarf and long shoots both of which were grazed were apparent . Table 4 . 9 Rates of leaf emergence and increase in leaf length at so ; between August and October 1 980 Rate of leaf Rate of increase emergence of leaf length (number/60 days ) (mm / day ) Dwarf shoots 3 . 0 + . 223 5 . 05 + . 483 Long shoots 2 . 95 + • 25 5 - 33 + . 296 On the low dune , the shelters erec ted in spring 1 979 excluded mammalian grazers . Measurements of numbers of leaves per shoot and length of leaf laminae were made in October 1 980 . Where grazing had been prevented , shoots possessed up to eight leaves , cf five where grazing had previous ly occurred . The frequency distribution o f length of leaf laminae indicated many more shorter (younger ) leaves where grazing was prevented ( fi gure 4 . 1 8) . Lit t le difference in maximum leaf length was apparent between the two treatments . The estimated annual net biomass ac cumulation of Carex pumila shoot populations at the four sites on the sand plain from January 1 978 are given in table 4 . 1 0 . The estimates were based o n ( i ) equat ion and ( ii ) equation 2 for vegetative shoots , and equation 3 for fertile shoots . 50 40 >- (J c 30 Q) ::J g 20 c:_ l.L 10 0 50 40 >- (J c 30 Q) ::J cr 20 Q) c:_ l.L 10 0 Figure 4 . 18 The e f f ect o f rabbits graz ing on the low dune (b ) The f r equency d is t r ibut i o n o f l e a f l e n gth in s pr i n g 1980 , o n s h e l tered a n d contra l p l o t s - ( ' ) c o n tro l 1 - - - - I I - ( " ) h l t 1 1 s e er - - - - r I r 1 0 -5 5- 10 t o - 1 5 t S -.20 20 -25 25 -30 3o - �5 Le n g t h o f l e a f l a m i n a ( cm ) Page 1 6 1 Table 4 . 1 0 Estimated rates of annual net biomass accumulation of Carex pumila (g DM I m2 I year ) at five sites on the sand plain � during 1 978 � 1 979 and 1 980 . grams DM I m2 I year so S1 S2 S3 S4 1 978 ( i ) 409 . 71 1 1 45 - 9 1 57 . 1 21 . 84 ( ii ) 781 . 42 1 2 1 1 . 1 3 1 1 . 95 33 - 98 1 979 250. 5 1 35 . 77 405 . 57 42 . 69 ne 1 980 268 . 01 64 . 1 1 1 1 9 . 89 20 . 5 ne ( i ) aerial and ( ii ) total biomass estimates for 1 978 The standing biomass ( SB ) in midsummer was found to be posi tively related to the annual net biomass accumulation (ANBA) that occurred both subsequently and previously ( figure 4 . 1 9 ) . these relationships are : The equations for ANBA � -22 + and ANBA = - 1 9 . 9 + 1 • 1 6 SB 1 • • • • • • • • • • • • • ( 4) 1 . 9 1 SB2 • • • • • • • • • • • • • ( 5 ) res pectively ; where SB1 is the standing biomass at the beginning and SB2 the standing biomass at the end of a twelve month period of net biomass accumulation. The estimated intercept on the ANBA axis ( SB = 0) in each case is small and not found to be significantly different from zero . Further � the estimated slope coefficients for the ANBAISB regression lines are significantly different from zero in both cases ( P< . 01 ) . The simple ANBA - SB correlation coefficient ( r ) values sugges t that the standing biomass in mid-summer is a better predictor of subsequent than of previous annual net biomass accumulation. However � comparison of the residual mean squares from the two analysis Page 1 62 of variance tables giving the sums of squares due to regression ( table 4 . 1 1 ) shows that there was no significant difference between these two parameters as predictors o f annual net biomass accumulation (F-ratio for 1 0 and 9 degrees of freedom was not significant ) . Table 4 . 1 1 Analysis of variance of the regression of standing biomass ( SB ) on annual net biomass accumulation (ANBA ) . ( SB 1 and SB2 = SB before and after a period of biomass accumulation� respectively) . ANALYSIS OF VARIANCE ANBA vs SB 1 DUE TO DF ss MS=SS/DF REGRESSION 41 3836 4 1 3836 RESIDUAL 9 1 26489 1 4054 TOTAL 1 0 540325 ANALYSIS OF VARIANCE ANBA vs SB2 DUE TO DF ss MS=SS/DF REGRESSION 376822 376822 RESIDUAL 1 0 1 691 34 1 69 1 3 TOTAL 1 1 545957 Page 1 63 Figure 4 . 1 9 The relationship between annual net biomass accumulation and standing biomass of Carex pumila before a 1 2-month period of biomass accumulation . ANBA ( g/m2/yr ) 900+ * R-SQUARED = 76 . 6 % 600+ 300+ * * * * *** 0+ *2 +---------+---------+---------+---------+---------+ o . oo 1 50 .00 300.00 450.00 SB 1 ( g / m2 ) 600 .00 750 . 00 Cumulative losses of biomass in the resident Carex pumila populations during 1 978 , 1 979 and 1 980 were greater than cumulative gains (net biomass accumulation ) over these same annual periods , at all si tes monitored on the study area � with the exception of S 1 in 1 978 and SO in 1 980 . The standing biomass estimates of both fertile Page 1 64 and vegetative shoot populations at each of the former sites therefore were lower than those obtained at the same site twelve months previously ( figures 4 . 1 1 and 4 . 1 3 ) . This trend was also seen in tables 4 . 1 2 and 4 . 1 3 which show the mean dry weight per shoot module in each of the four summers of the study . The trend reflects the increasing age/senility of these seral populations . The reductions in standing biomass ( figures 4 . 1 1 and 4 . 1 3 ) were mos t marked in absolute terms at those sites at which net biomass ( or herbage ) accumulation had previous ly been greates t (as at S2 during 1 978) . The observation of Noble et al. ( 1 979 ) that populations of Carex arenaria that have the highest rates of shoot births also had the highes t death rates probably reflects a similar phenomenon. In less productive populations at other sites � the decline between years although less marked in absolute terms still represented large relative losses . For example � at S3 between 1 978 and 1 979 , a five-fold decline in standing biomass occurred ( figure 4 . 1 1 ) . The exceptions to the trend of overall net losses of biomass between years apply to those populations which were in a juvenile phase of development when firs t monitored . At SO � annual net biomass accumulation rate for 1 980 � which was probably underestimated due to site disturbance by rabbit-grazing in winter and sand deposition in autumn, was 50 per cent higher than that estimated for 1 979 ( table 4 . 1 0 ) . However the expected mortality of older shoots and suppression of growth as a result of the movement of large quantities of sand onto the site did not eventuate in spring 1 980 ( see above ) . This resulted in a similarly increased standing biomass at this site in December 1 980 compared with that twelve months previously ( figure 4 . 1 3 ) . The inundation of S1 by sand in spring 1 978 resulted in the loss of about 95 per cent of live herbage in this juveni le/adolescent population Page 1 65 over this August-October period . Recovery of of this population later in the spring/early summer was reflected in the mid-summer 1 978-79 aerial biomass value which surpassed that of January 1 978 � but did not show a continuing increase over the autumn and winter 1 979 as in 1 978 (figure 4 . 1 1 ) . Thus , from autumn 1 979 , standing biomass values at this site were less than those recorded twelve months previously. The declining biomass per uni t area of the population at S 1 over time between summer 1 978-79 and 1 980-81 was not paralleled by shoot density ( section 4 . 3 . 1 ) • . The declining biomass was attributable to the depauperation of shoot modules ( tables 4 . 1 2 and 4 . 1 3 ) . The accretion of sand at S1 over the duration of the study not only caused disturbance to the plants , but also increased nutrient- and water-stress by removing them from ready access to the water table . Table 4 . 1 2 Mean seasonal maximum aerial dry weight per shoot module of reproductive shoot populations of Carex pumila at five sites on the sand plain in each of four summers . grams DW I aerial shoot module so S1 S2 S3 S4 1 977-78 ne 0 1 . 24 . 56 . 1 85 1 978-79 ne . 5 1 2 . 777 . 202 . 1 20 1 979-80 . 932 . 402 1 . 1 9 . 1 26 ne 1 980-81 Control . 762 . 1 38 . 25 1 0 ne Table 4 . 1 3 Winter minimum and summer maximum mean dry weights per aerial shoot module of the vegetative shoot populations of Carex pumila at five sites on the sand plain in successive years , 1 977-78 to 1 980-81 . grams DW I aerial shoot module so S 1 S2 S3 S4 ( a ) Winter minima 1 978 ne 0 . 1 9c 0 . 46a 0 .08d 0 .06d 1 979 0. 1 7c 0 . 1 8c 0 . 27b 0 . 03d 0 .03d 1 980 0. 1 2d 0. 1 2d 0. 1 2d 0 . 06d ne (b ) sv�7::7�axima 0. 292 0 .82 0 . 29 0 . 1 4 1 978-79 0 . 335 0 .45 0 . 1 2 0 . 075 1 979-80 0. 482 0 . 295 0. 61 6 0 . 1 2 ne 1 980-8 1 0. 65 0 . 1 2 0 . 2 1 0 . 075 ne a > b ; P< . 05 a > c ; P< . 01 b > c ; P< . 01 c > d ; P< . 05 Page 1 66 As shown in figure 4 . 8 , species other than Carex pumila made up only a small proportion of the total sward mass in January 1 978 � although at sites of increasing distance from the terminal hollow this proportion increased. As the resident swards at each site on the sand plain aged over the three years of the study� the standing mass of other species in the sward increased ( table 4 . 1 4 ) . These increases in other species with time cont rast with the decline in biomass of Carex pumila seen between consecutive years . The differences observed between sites on the sand plain at a given point in time then� parallel the successional changes with time and are consistent with the view that the populations at these sites form a successional series . Table 4 . 1 4 Sward mass of species other than Carex pumila in 1 979-80 and 1 980-81 at two sites on the sand plain . S2 S3 1 979-80 29 . 1 + 42 1 42 . 1 + 57 . 1 grams DM / m2 3 . Energy and elemental concentrations ( i ) Energy 1 980-81 352 . 9 � 252 . 1 1 96 + 55 . 6 The energy values determined for component organs of Carex pumila are shown in table 4 . 1 5 . At S2 , the mean energy value obtained for seeds was higher ( P< . 05 ) than that for each of the o ther plant parts assayed ( except roots , table 4 . 1 5a ) . The energy content of seeds increased over the season suggesting the elaboration of high energy Page 1 67 containing compounds as seed weight increased . At S2 � mature seed was seen to have a significantly higher ( P< . 05 ) mean energy value than that of seeds collected earlier in the season before grain filling had commenced ( table 4 . 1 5b) . There were also highly significant ( P< .001 ) site differences for the energy value of seeds collected at the end of the season ( table 4 . 1 5c ) . The lower energy values of seeds found at S1 and S4 compared with S2 and S3 were similar to that for immature seeds at S2 ( table 4. 1 5b) . These reduced energy values for seed sugges t incomplete seed development . This was confirmed by the reduced mean seed weights found at these sites compared with S2 and S3 ( section 4 . 3 . 5 ) . Page 1 68 Table 4 . 1 5 Mean (� standard error ) energy values ( joules / gram DM ) of Carex pumila ( a ) for component parts from S2 in January 1 978� (b ) for seeds from S2 at two different stages of fertile shoot development (November � 1 978 and January � 1 979 ) and ( c ) for seeds from four different sites in January� 1 979 . ( a ) Plant part ( b ) Seeds ( c ) S eeds Roots Rhizomes Leaves Glumes/rhachillae Male Seed inflorescences Stage of development 7-8 ( seed filling) 1 0 ( maturity ) Site S4 S3 S2 S 1 j oules/gram 1 9334 . 4 � 56 . 4 1 8426 . 1 + 22 . 4 1 87 6 1 • 0 � 277 . 3 1 5023 . 0 1 5600 . 7 20234 . 4 + 1 66 . 6 j oules/g 1 8007 . 5 � 26 . 0 20305 . 6 � 483 . 9 j oules/g 1 6986. 2 20334 - 9 � 1 42 - 9 20443 - 7 � 98. 9 1 8032 . 6 � 31 2 . 3 # Energy values of stored material dried at 80 degrees C and corrected for % dry weight . The energy values for seeds ( and other organs) of Carex pumila suggest that the lipid content of these organs is low. This was confirmed by a direct de termination of percent lipid content on seeds from mature fertile shoots co llected in January 1 978 ; a value of 3 . 6 1 5 � 0. 21 6% fat by weight was obtained. These values contrast with Page 1 69 the energy content of seeds of other plant species which store much larger concentrations of lipids ( Levin 1 974) . I t is unlikely that proportionate allocation patterns in Carex pumila to component shoots and organs based on joules would differ significantly from those based on dry weight . Working with Lupinus species , which store up to 1 5% lipids in their seeds , Hickman and Pitelka ( 1 975 ) confirmed that the proportionate energy allocation pat terns using both grams dry weight and stored energy units as measures of energy investment did not differ significantly. The pre liminary energy values obtained for Carex pumila seeds (both percent lipid and j oules/gram) , the small differences between plant organs in terms of energy content/gram DM ( table 4 . 1 5a ) and the relatively high within-site variance for proportional plant part dry weights ( section 4 . 3 . 6 ) suggest that a similar result would be obtained in this study . More extensive energy determinations were not the refore carried out . ( ii ) Elemental concentrations The elemental concentrations of Cu , Zn, Fe , Ni ; Mn, Mg; Ca , Na and K of constituent organs of Carex pumila sampled from the mature population in the old hollow at S2 in October 1 978 are shown in table 4 . 1 6 . Variation between constituent organs was evident from this table . Further , compared with values quoted by Brooks for plant material generally , Carex pumila on the study area contained reduced amounts of Zn, Fe and Mn, and elevated amounts of Mg and Na . High concentrations of these latter two elements in this coastal habitat accounts for the massive plant concentrations . The prevailing west to north-west winds bring sea spray onto the coastal dunelands . The low plant concentration of Fe is interesting in that it reflects the low Page 1 70 concentrations in the subst rata ( table 3 . 6 ) in contrast to the sand immediately north of the study area , south of the Rangitikei River mouth which contains massive amounts of magnetic minerals ( Gibb 1 977 ) . On this prograding coast , sand at inland sites can be expected to differ little in mineral composition from that on the beach ( Oliver 1 948 ) . Table 4 . 1 6 Mean ( + standard deviation) elemental concentrations of constituent organs of Carex pumila in October 1 978 . Concentration (ppm ) Cu Zn Fe Ni Mn Roots 9 - 5 + 3 - 5 1 3 . 0 + 2 . 1 59 - 5 + 1 4 . 3 3 . 8 + 1 . 9 61 . 8 + 1 3 .8 Rhizomes 7 . 0 + 1 . 4 20 . 0 + 7 . 2 1 9 . 2 +" 5 . 1 3 . 8 + 1 . 8 n.o + 6 . 2 Green lvs 1 0 . 5 + 3 . 5 25 . 5 + 8. 1 1 1 . 0 +" 6 . 4 3 - 3 + 0 .8 91 . 2 + 28 . 8 Dead lvs 6 . 5 + 0 . 5 8 . 5 .:. 1 . 5 1 4 . 5 .:. 4 . 5 1 6 . 5 + 1 3 . 5 87 . 0-+ 3 . 0 Bracts ### Roots �hizomes reanl.leaves f�ctseaves ### ### Means . for 7 . 0 1 0 . 0 38 . 0 1 . 0 39 . 0 9 . 0 70 . 0 500 . 0 3 . 0 400 . 0 Mg Ca Na K 2550 + 1 0 . 30 6394 + 1 030 7052 + 1 733 5546 + 1 547 1���2i ���7 ����2i �� ��fiai ��26 ����J ;��4 700 5000 200 3000 plant material generally, from Brooks ( 1 972 , table 1 1 . 2 ) The allocation of limited minerals to different plant parts may be of greater significance in the evo lution of plant strategies on these nutrient-stressed sand plains than that of dry weight . These data suggest that allocation patterns based on dry weight and minerals in Carex pumila will differ significantly. This was tested for crude total nitrogen . (iii ) Crude total nit rogen In Carex pumila , percent crude total nitrogen ( %TN) values for seeds and green leaf laminae were higher than that for green leaf sheaths , which in turn was higher than those of all other parts of Page 1 7 1 live branch modules ( table 4 . 1 7 ) . These differences � averaged over the four sites monitored during 1 979-80 � were highly significant (P< . 001 ) . Taking each component organ in turn� it was found that the %TN values for seeds and for green leaf laminae were similar at all four sites across the study area ( table 4 . 1 8) . By contras t � %TN of rhizomes � roots and bracts were significantly greater at S2 than at other sites (P< . 001 � P< . 001 and P< . 05 � respectively) . This obse rvation probably reflects the greater concentration of available nitrogen in the soil in the old hollow than on the low dune , as a result of the activity of nitrogen-fixing blue-green algae (section 3 . 1 . 5 ) . Despite the reduced amounts of nitrogen available for uptake on the low dune , the concentrations of this element in green leaf laminae and seeds were not reduced on these sites . The low percent nitrogen values found for roots , rhizomes and bracts on the old hollow at S3 � where blue-green algae were also evident � sugges t that the nitrogen available for plant growth was not taken up by Carex pumila . Other later seral species � where they are present as at S3 in 1 979-80 , are likely to preferentially absorb soil nitrogen ( see also section 4 . 3 . 7 ) . Table 4 . 1 7 Percent crude total nitrogen content ( %TN ) of component organs of live and dead branch modules of Carex pumila � averaged over four sites , in summer 1 979-80 . Roots Rhizomes Bracts Green sheaths Green laminae Dead leaves L Seeds . D , etters �d�cate uncan s Live shoots 0 . 47 b 0 . 53 b 0 . 45 b 1 . 09 c 1 . 73 d 0. 56 b % TN Dead shoots 0 .425 b 0. 284 b 0. 568 b mu1 !.7p9led t · t · P< 001 l1<. range es , • Page 1 72 Table 4 . 1 8 Mean ( + standard deviation) percent crude total nitrogen content (%TN ) of component organs of Carex pumila . Percent total nitrogen (%TN ) (a ) Living branch modules , 1 979-80 Roots Rhizomes Bracts so - 497 + . 1 24 • 449 + . 1 59 . 403 + . 1 00 Green sheaths 1 . 327 + . 446 Green laminae Dead leaves Seeds 1 . 920 + - 474 . 62 1 + . 1 87 1 . 935 + . 827 S1 . 41 9 + . 1 1 4 • 505 + . 1 77 . 426 + . 088 1 . 1 07 + • 328 1 . 963 + . 340 • 508 + . 1 36 1 - 497 + • 342 (b ) Dead branch modules , 1 979-80 SO S1 Roots . 378 + . 030 . 472 + . 07 1 Rhizomes . 289 + . 044 . 404 + . 076 Leaves . 5 1 4 + . 1 36 . 565 + . 089 ( c ) 1 978-79 S1 S2 # Roots # . 326 + . 049 Rhizomes # • 250 + . 037 Green leaves 1 . 625 + . 236 Dead leaves • 865 + . 237 Underground parts attached - 456 + . 043 . 21 2 + . 043 1 . 277 + . 1 98 • 689 + • 1 1 8 to living and S2 . 61 2 + . 1 23 • 758 + . 1 88 . 52 1 + . 1 00 1 . 03 1 + - 336 1 . 582 + - 430 • 598 + . 1 45 1 . 6 1 8 + . 46 1 S2 . 380 + . 1 54 . 224 + . 049 . 609 + . 080 S3 - 356 + . 1 24 . 1 57 +" . 05 1 1 . 422 + . 238 • 744 + . 1 94 dead shoots S3 - 350 + . 062 - 393 + . 085 - 437 + . 067 . 882 + • 239 1 . 447 + . 1 5 5 . 520 + . 1 1 1 2 . 1 1 0 + . 488 S3 . 468 + . 036 . 220 + . 008 ( • 582 + • 444 ) S4 - 390 + . 1 37 . 1 33 + . 047 1 . 327 + . 28 1 .7 1 4 + . 1 64 I ! In the harvest of 1 978-79 , the underground portions of individual - j shoot modules were not separated so that values of dry weight and consequently percent total nitrogen es timates were obtained for the total root and total rhizome fractions only. No difference between sites were found for percent nitrogen content of roots . However , there was a progressive increase in percent nitrogen values for total rhizomes from the oldest to the youngest site ( table 4 . 1 8c ) . The greater percent total nitrogen values of total rhizomes at the younger si tes may reflec t a greater proportion of living as opposed to dead rhizomes in this total underground fraction. In summer 1 979-80 , the percent total nitrogen content of rhizome modules attached to live shoots was twice that of rhizomes at tached to dead shoots ( table 4 . 1 7 ) . The depletion of nitrogen from live shoot modules can be seen by comparing values for green leaf laminae and sheaths with values for dead leaves (on both living and dead shoots ; tables 4 . 1 7 and 4 . 1 8) . I Page 1 73 Figure 4 . 20 shows the effect of season ( time ) on the percent nitrogen content of seeds of Carex pumila at each of the sites monitored on the plain. The positive slopes of these curves reflect the increasing availability of ni trogen on the sand plain at the sites matured between 1 977-78 and 1 980�8 1 . Free�living nitrogen�fixing blue�green algae ( Nostoc and Anabaena species ) found on the plain were probably responsible for the increases of nitrogen in the developing sand plain ecosystem. The subsequent decline in percent total nitrogen of seeds at S2 in the 1 980-8 1 season ( figure 4 . 20 ) reflects the decreasing suitabili ty of the habitat for Carex pumila � as other species became abundant , interfering with the ability of the pioneer to accumulate resources . The fertile shoot population of Carex pumila at this site in summer 1 980-8 1 was represented by a few depauperate shoots in contrast to the dense sward found in December 1 977 . Seed filling by this senescent shoot population was reduced compared with previous seasons ( seed weight , section 4 . 3 . 5 ) . In summer 1 980-8 1 , the percent nitrogen content of seeds at S2 was also significantly lower (P< .001 ) than those values at the two younger sites on the low dune . The percent total nitrogen values for Carex pumila shoo ts in December 1 980 differed substantially� not only between component organs and according to the maturity of the population ( site ) , but also according to the vegetative or flowering state of the shoot and further whether the shoot was derived from a dwarf or a long branch . These differences are probably related to the depletion of nitrogen from shoots with age. Table 4 . 1 9 shows the relative depletion of nitrogen from vegetative organs of fertile shoots of Carex pumila compared with vegetative shoots in December 1 980 . This observation reflects the greater maturity of the fertile shoot cohort and the onset of 2 . 5 2 ,-... � .....__.. Q) 1 . 5 Ol ro � c Q) 0 - (__ Q) 0... 1 0 . 5 Figure 4 . 20 Percent crude tota l n itrogen content of seeds of Carex �um i l a in four summers at three s ites on the sand p la i n . Vertical bars ind i cate 85% er for s ite means based on poo led standard dev i at ions S3 S 1 I so 01 r� � I 52 0�-----------------------------------77-78 78-79 79-80 80-8 1 senescence in this part of the clone . Accompanying this senescence of vegetative parts is the growth and development of seeds which act as a store for nitrogen and other nutrient elements ( Jefferies et al . 1 979 ; Lovett Doust 1 980) . The chanelling of nitrogen to seeds of Carex pumila resulted in a substantially greater concentration of this element in dissemules than in any other organ (approximately 5x that of rhizomes and 7x that of roots of fertile shoots � table 4 . 1 9) . Structures already low in nitrogen in vegetative shoots (namely � bracts and roots ) were not further reduced in fertile shoots . Table 4 . 1 9 Mean percent crude total nitrogen content ( %TN ) of component organs of vegetative and fertile shoots of Carex pumila in December 1 980 . Percent total nitrogen ( %TN ) Vegetative shoots Reproductive shoots Rhizomes 0 . 67 0 . 44 Rhizome tips 1 . 28 ne Roots 0 . 38 0 . 29 Bracts 0 . 35 0 . 38 Green leaves Sheaths 0 . 99 0 . 43 Laminae 1 . 85 1 . 55 Dead leaves 0 . 68 0. 50 Male spikes 0 . 79 Seeds 2 . 1 3 Although the percent nitrogen of the vegetative portion of fertile shoots was significantly ( P< .001 ) different ( lower ) than that of vegetative shoots , the difference was not the same for all • component organs ( see significant ' state x organ ' interaction term in Page 1 75 the analysis of variance , P< .001 ; table 4 . 20) . Further; the significant ' state x site x organ ' interaction (P< . 05 ; table 4 . 20 ) indicates that the relative depletion of nitrogen from the component organs of fertile shoots compared with vegetative shoots was greater at SO than at S1 . Page 1 76 Table 4 . 20 Analysis of variance of percent total nitrogen content of component organs (or) of Carex pumila shoots at two sites ( si ) under two treatments ( tr , + nitrogen fertilizer) in December 1 980 . Shoots were divided into -vegetative and fertile states (st ) and dwarf and long types ( ty) . SOURCE D .F . (M. V. ) st ty st*ty si st*si ty*si st*ty*si tr st*tr ty*tr st*ty*tr si*tr st*si*tr ty*si*tr s t*ty*si*tr or st*or ty*or st*ty*or s i*or st*si*or ty*si*or s t*ty*si*or tr*or st*tr*or ty*tr*or st*ty*tr*or g�:�r:�?*or ty*si*tr*or e rror 1 1 . 1 1 1 1 1 1 1 . 1 1 1 1 1 1 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 Estimated treatment means : st L . S . D . 5% L . S . D . 1 % L . S . D • • 1 % Veg 0 .82 Fert 0 . 60 0 .04 0 . 06 0 .09 ty Dw o . 74 Lo 0. 68 0 . 04 0 .06 0 . 09 s . s . 1 . 1 5 o. 1 1 0 . 05 0 . 98 0. 64 0 .08 0 . 1 3 0 . 02 5 . 1 04E-03 0. 1 4 0 . 1 5 0 . 04 4 . 578E-05 7 . 01 9E-04 0 .03 20 . 46 0 .83 0. 1 1 0 . 01 0. 33 0 . 25 0. 44 0. 06 0 . 24 0 .01 0 . 26 0. 1 2 8 : 6§ M . S . S . 1 . 1 5 0 . 1 1 0 .05 0 . 98 0. 64 0 .08 0. 1 3 0 . 02 5 . 1 04E-03 0. 1 4 0 . 1 5 0 . 04 4 . 578E-05 7 . 01 9E-04 0 . 03 4 . 09 0. 1 7 0 . 02 2 . 1 1 8E.:.03 0 . 07 0 . 05 0. 09 0 . 01 0 . 05 2 . 834E-03 0 . 05 0 .02 8 : 8� 0. 05 0. 01 0 . 02 4 . 585E-03 si so 0 .8 1 S1 0 . 6 1 0 . 04 0. 06 0 . 09 tr N 0 . 70 c 0 . 72 0. 04 0 . 06 0 . 09 F RATIO 250. 47 *** 22 . 97 ** 1 0 . 99 * 21 4 . 63 *** 1 38. 92 *** 1 8. 06 ** 28. 1 5 ** 3 . 72 ns 1 . 1 1 ns 3 1 . 1 0 ** 33 - 49 ** 7 . 70 * 9. 983E-03 ns 0 . 1 5 ns 6 . 1 1 ns 892 . 40 *** 36 .00 *** 4 . 90 ns 0 . 46 ns 1 4 . 27 ** 1 o . 75 * 1 9 . 35 ** 2 . 45 ns 1 0 . 38 * 0 . 62 ns 1 1 . 1 7 ** 5 . 42 * 1 3 : 5§ : 2 . 32 ns or Rhiz 0 . 56 Root 0 . 33 Brac 0. 37 ·shea 0. 71 Lami 1 . 70 Dead 0. 59 0 . 06 0 . 1 0 0 . 1 6 Page 1 77 The fate of nitrogen in the vegetative plant parts lost during senescence ( compare both dead with green leaves and vegetative components of fertile with vegetative shoots � table 4 . 1 9) was not directly monitored in the present study. However� the differential dist ribution of nitrogen amongst the component organs of Carex pumila shoots ( table 4 . 1 9) suggests an acropetal direction of movement of this element through the plant and its accumulation in the distal parts , in the case of vegetative shoots in the green leaves and rhizome tips and in the case of fertile shoots ultimately in the seeds . Previous studies on other rhizomatous perennial species suggest only a very limited redis tribution of resources between immediately adjacent shoot modules within the clone and a largely acropetal direction of movement of nutrients along the rhizome axis (Forde 1 966 ; Tietima 1 979 ; Noble and Marshall 1 983 ) . Even when older basipetal shoots were severely nutrient stressed , Noble and Marshall ( 1 983 ) found little evidence of basipetal movement of ni trogen or phosphorus in Carex arenaria , a clonal species � in either fie ld or glasshouse experiments . In December 1 980 , the %TN averaged over all vegetative organs was significantly less at SO than at S1 (P< . 001 � table 4 . 20 ) . The %TN values at SO were less than those at S1 for each component organ except bracts , which had uniformly low %TN values at both sites ( si gnificant ' site x organ ' interaction� P< . 01 � in the analysis of variance , table 4 . 20 ) . The effect of site was largely attributable to the low %TN values for reproductive shoots at SO ( significant ' state x site ' interaction, P< .001 , table 4 . 20 ) al though the reduction of %TN of dwarf shoots between the two sites was greater (P< . 05 ) than that of long shoots (significant ' type x site ' interaction� table 4 . 20 ) . These differences between the two sites reflects the more rapid Page 1 78 development and onset of senescence of dwarf and reproductive shoo ts at the younger site . At the older; drier low dune s1 ; development of these shoots was more prolonged . The differences between sites for %TN for the various organs , were not all similar ( significant ' state x site x organ ' ; P< . 05 ; and ' type x site x organ ' ; P< .001 ; interactions ; table 4 . 20) . The overall reduction of nitrogen content of component organs at SO was not evident for green leaf laminae of vegetative shoots (averaged over both shoot types and both treatments ) nor for dwarf shoots (averaged over bo th states of shoot and both treatments ) . Further ; the %TN of rhizomes of long shoots (averaged over both states and bo th treatments ) was greater at the younger of the two low dune sites . In Carex pumila ; the percent crude total nitrogen values showed significant differences not only between plant parts and between sites ; but also over time ; between types of shoot (whether at tached to dwarf or long rhizome segments ) , the sexual state of the shoot and in response to nitrogen fertilizer addition (section 4 . 3 . 7 ) . These differences will alter es timates and comparisons of total nitrogen allocation from those based on dry weight alone . Since the percent nitrogen values differed by up to 7X between plant parts ; it seems reasonable to expec t that the allocation patterns based on total nitrogen and dry weight will differ significantly. This was tested directly ( section 4 . 3 . 6 ) . The percent crude total nitrogen values for plant species other than Carex pumila obtained in the harvested samples were also determined ( table 4 . 21 ) . Since this sward mass was not divided into constituent species no� into aerial and underground parts ; the values given in this table should not be compared with those obtained for Carex pumila which apply to constituent plant parts . Page 1 79 Table 4 . 21 Mean (� standard deviation) percent crude total nitrogen content (%TN ) of plant species other than Carex pumila found on the study area in December 1 980 . Site S2 S3 Treatment Control Control Fertilizer 4. Total nitrogen content � unit � %� 1 . 06 + 0. 1 6 1 . 05 0. 99 At the beginning of the study the total nitrogen content of the sward mass was significantly (P< . 01 ) greater at S2 in the old deflation hollow than that at any other site (figure 4 . 21 ) mainly as a result of the considerably larger sward mass at the former site (figure 4 .8) . Further contributing factors were ( 1 ) the proportion of sward mass in herbage , for which total nitrogen concentration was greater than other plant parts ( table 4 . 1 7 ) , was also greater at S2 than at other sites (section 4 . 3 . 6 ) and (2) percent total nitrogen values for several live shoot components were greater at S2 than at other sites ( table 4 . 1 8) . Dead aerial shoots of Carex pumila contributed a large and significant proportion of the total amount of nitrogen in the sward mass at the three old deflation hollow si tes . This nitrogen would become available to growing plants only as a result of decomposition of this dead material. The time for 50% decomposition of dead Carex pumila leaves was shown in section 3 . 2 . 3 to be almost a year ( 35 1 days ) . 14 - 12 - 10 - (\] E "-.. 8 . z (f) E m 6 - (_ CJ) 4 - 2 - 0 F\ gure 4 . 2 1 The d i str i but ion o f the tota l n i trogen content o f the sward mass to Carex gum i l a and other spec i e s at four s ites i n thre e succes s i v e years 1 = 1977-78, 2 = 1978-79, 3 = 1979-80 1 3 ;=.;.: Other spec ies Car e x pumila r:-:- c-'-- ;:-: � � r:- ...:..._,_ r:-: r-r-- r-=- ,.-- r---=-r-- r-- s i te 1 s i te 2 s i te 3 s i te 4 Page 1 80 The fluctuations observed in terms of the total nitrogen content of the standing biomass of Carex pumila over time followed the changes already outlined for standing biomass on which the former values were based . The magnitude of the relative increases of the former parameter were however greater than those of the latter . At S2 � the to tal nitrogen content of vegetative shoots increased 7-fold during autumn 1 978 ( table 4 . 22 ) whereas the dry weight increase although large ( 5-fold ; figure 4 . 1 1 ) was not as great . Similarly at S1 , total nitrogen more than trebled between January and August 1 978 whereas the dry weight increase over this same period was less than double . No such dis crepancy was found in the magnitude of the relative decline in dry weight August 1 978 . and total nitrogen at for example , S2 over the winter to The explanation for this difference lies in the percent nitrogen content of the plant tissues that were undergoing changes in standing dry matter . The initial increases in aerial biomass during 1 978 at both S1 and S2 resulted from increases in live (green ) herbage which possessed relatively high total nitrogen concentrations ( table 4 . 1 7 ) . The decreased standing aerial biomass over winter 1 978 at S2 , by contrast , resulted from leaf and shoot senescence . The percent total nitrogen content of dead leaves was considerably less that that of green leaves ( table 4 . 1 7 ) . The larger relative decrease of total nitrogen compared with that of dry weight at S 1 during early spring 1 978 can be similarly explained . Inundation of the site by sand removed green shoots that would otherwise have continued growing, rather than senescent shoots with a high proportion of dead tissue and lower N-concentrations . Page 1 81 Table 4 . 22 Total nitrogen content of the aerial biomass of Carex pumila during 1 978. grams N I m2 S1 S2 Summer 1 977-78 . 58 1 . 47 Autumn 1 978 • 96 1 0 . 20 Winter 1 978 2 . 05 3 . 60 Spring 1 978 • 1 3 3 . 50 Summer 1 978-79 . 66 3 . 82 The increase in total nitrogen content of Carex pumila shoots during biomass accumulation depends upon nitrogen inputs from the surrounding area , even at those sites where nitrogen-fixing blue�green algae were apparent . The increase in total nitrogen by vegetative shoots at S2 during autumn 1 978 , equivalent to 8. 73 g N I m2 � was somewhat greater than that that could be supplied by nitrogen-fixation in the soil (section 3. 1 . 6 ) . It was estimated that N-fixation in the surface layer of the soil to 25mm depth could have provided only 1 . 36 g Nlm2 over this autumn period ( table 3 . 5 ) . The balance can be attributed firstly to decomposition of Carex pumila and other plants in situ and secondly to topogenic accumulation from the surrounding high dunes and low dunes on the sand plain . Nitrogen fixation can be expected in the rhizosphere zone of Ammophila arenaria which provides a favourable environment for nitrogen�fixing bacteria (Hassouna and Waring 1 964 ; Wahab 1 975 ) and in Lupinus arboreus through its symbiotic association with Rhi zobium bacteria . Both plant species were abundant on the plain and surrounding high dunes . Fixed nitrogen will subsequently be made available to other plants in the ecosystem through plant exudates and decomposition ( Gadgil ; 1 971 ) . Page 1 82 The differences in total nitrogen concentrations of the various plant parts resulted in changes in the levels of significance of differences between site populations for total nitrogen content compared with those seen for dry weight . Table 4 . 23 shows such a comparison for fertile shoot populations at S 1 and S2 in December 1 979 . For those organs with low percent nitrogen values ( roots � spikes and bracts ) , no significant differences were found between S 1 and S2 for total nitrogen content � whereas these sites had been seen to differ when the comparison was based on dry weights ( table 4. 23 ) . Conversely � for leaves and seeds ( organs with significant ly higher concentrations of nitrogen ) the level of significance between the two sites for dry weight was increased when total nitrogen contents were considered ( table 4. 23 ) . Page 1 83 Table 4 . 23 The dry weight and total nitrogen content of component organs of fertile shoots per unit area ; at S 1 and S2 in December 1 979 . Roots Rhizomes grams DW I m2 S1 S2 2 . 33 ** 9 . oo 9 .00 ns 1 7 . 67 Seed 6. 67 ** 42 . 22 * 8. 67 Spikes 1 . 33 Dead leaves 6 . 33 Green laminae 3 . 89 Culms & sheaths 3 . 00 Bracts 1 .00 ** 1 53 . 89 * 23 . 78 * 39. 56 * 7. 44 A er Total 22. 22 275 - 50 33 . 44 302 . 1 1 . 25 grams N I m2 S 1 S2 . 01 0 ns . 054 . 046 ns . 1 34 • 099 *** . 683 . 006 ns . 034 . 032 . 076 . 033 .004 . 306 *** ** *** ns . 923 . 376 . 408 . 039 2 . 46 2 . 65 * P< . 05 ; ** P< . 01 ; *** P< . 001 ; ns not significant The difference between S2 and the other sites in terms of the total amount of nitrogen in the sward mass was maintained over the duration of the study . Over these three years , the Carex pumila component of sward mass or the living fraction of this at each of the old deflation hollow sites declined ; both in terms of grams dry matter and total amount of nitrogen ( table 4. 24 and figure 4 . 22 ) . These losses which reflected the increasing senility of these populations ; were only partially balanced by the increases in both dry weight and total nitrogen of the ' other species ' fraction ( table 4 . 25 ) . By contrast , a t the two low dune sites ; sward and live herbage mass and the total nitrogen contents of these parameters showed no such decline 5 4 C\.1 E .......... z 3 ({) E m (_ Ol 2 2 . 5 2 1 . 5 1 0 . 5 F i gure 4 . 22 Di str i b u t i o n o f the tota l n i trogen content o f Carex P-Um i l a b i omass to c omponent organs at f our s i tes in summer 1 979-80 and 1 980-8 1 ( i ) 1 979-80 roots rhiz omes seeds sp ikes dead leaves green lam inae green sheaths sca l e leaves ( i i ) 1 980-8 1 s i te I s \ -t e 2 Page 1 84 between the first and second summers in which these sites were monitored . Carex pumila was the only species included in the harvests at these two sites . Table 4 . 24 Total nitrogen content (grams N per unit area ) of the aerial biomass of Carex pumila at five sites on the sand plain between 1 977-78 and 1 980-81 ( summer means ) • grams N / m2 so S 1 S2 S3 S4 1 977-78 ne 0. 58 5 . 6 1 1 . 22 0 . 23 1 978-79 ne 0. 65 3 . 82 0 . 29 0 . 1 4 1 979.;_80 0 . 97 o. 51 4 . 1 8 0 . 1 ne 1 980-81 1 . 36 0. 24 0 . 55 0 . 03 ne In summer 1 980-81 , as in the previous summer � the total nitrogen content per unit area of living shoot modules (underground parts included ) of Carex pumila was greater (P< . 05 ) at the younger site towards the edge of the low dune than at the older site higher on the low dune ( figure 4 . 22 ) . This greater quantity of plant nitrogen per unit area can be accounted for not by greater percentage levels of N , which were actually lower ( P< . 05 ) in shoots at the more adolescent site in 1 980-81 ( table 4 . 20) but simply by the greater living biomass at the younger of the two sites . By summer 1 980-81 � the total nitrogen content of the living shoot population at S2 in the old deflation hollow had declined substantially from that in the previous season ( figure 4 . 22 ) as a result of the general senescence of the Carex Page 1 85 population at this site . Thus � although the total nitrogen content of live Carex shoots at the adolescent site on the edge of the low dune was greater than that at any of the other sites monitored in 1 980�81 it was less than half that at S2 in the deflation hollow in summer 1 979-80 � the former value . The extremely low value obtained at S3 in the old deflation hollow in both 1 979�80 and 1 980�81 reflects the advanced stage of senescence of the Carex population on this part of the sand plain in both these years . Table 4 . 25 Total nitrogen content of species other than Carex pumila grams N / m2 S 1 S2 S3 S4 1 977-78 0 0 . 22 0. 99 1 . 78 1 978-79 0 0 . 26 1 . 1 1 1 . 67 1 979-80 0 0 . 29 1 . 42 1 . 70 1 980-81 0 3 - 53 1 . 96 ne 4 . 3 . 5 Flowering and seed production At those sites at which fert ile shoots of Carex pumila were present in the summers of 1 978-79 � 1 979-80 and 1 980-81 � heading ( the emergence of the terminal male spike from the mouth of the sheath of the youngest leaf ; stage 5 ) occurred during October . In spring 1 978 � heading on the earliest fertile shoots at S 1 � S2 and S3 occurred be tween visits to the study area on 3 and 1 8 October whereas in 1 979 earliest heading was between 6 and 1 3 October (at S 1 and S2 ) . Page 1 86 From spring 1 979 the stage of shoot development was sco red and leaf laminar area was measured on individual fertile shoots within the harves ted samples . The changes in these parameters over time at S 1 and S2 are shown in figure 4 . 23 and 4 . 24 . Figure 4 . 23 may be compared with the phenology of flowering determined for fertile shoots removed to the laboratory in spring 1 979 and monitored daily ( section 2 . 3 ) . Heading was reached between 30 October and 4 November. The uppermost female spike emerged ( stage 5 . 5 ) as a result of continuing culm elongation six to eight days later. Within a further six days culm elongation ceased and stigmas were seen to have emerged and anthes is begun ( stage 6 ) . Anthesis was completed ( stage 7) within another seven to nine days . Noticeable swelling of utricles ( stage 8) occurred within another seven to nine days . The period of seed filling stages 7 to 9) extended for 5 to 6 weeks between the end of November and early January. Figure 4 . 23 b indicates the variability within a site population for stage of fertile shoot development . The development of fert ile shoots of Carex pumila over the summer 1 979-80 was accompanied by the gradual decline in the green leaf area per shoot (figure 4. 24 ) and thus the ability of vegetative shoot parts to assimilate carbon. Since green laminar area was measured on individual shoots , this dec line gives a more accurate picture of the onset of maturity in this coherent shoot cohort than the changes in dry weight of green leaf laminae per shoot � which were based on bulk lea f samples . Changes in green laminar area per fertile shoot be tween November 1 979 and February 1 980 mirrors the changes in stage of shoo t development these shoots ( figure 4 . 23 ) . 10 9 .j.J c QJ 8 E 0.. 0 r-t QJ > QJ 7 "'0 .._ 0 QJ C) ro 6 .j.J CJ) 5 4 Figure 4 . 23 (a) Mean (± standard error) stage o f development o f fertile shoots o f Carex pum i la over time at two s ites on the sand p l a in, dur ing summer 1979-80 . r I I I I I I I I I I I I I I I I ' · I / OCT NOV 1978 -- -::1: I------ __.. I r-- ---- :r---- /I/ DEC S ite 1 Site 2 JAN 1880 4-23b 2ert ile shoots on the low dure , 27 November 1 979 . Left to right : stage 8-9 ; stage 7 ; stage 5 ; stage 6 . F igure 4 . 24 Mean area o f green lea f l aminae per ferti l e shoot over t ime a t two s i tes on the sand p la in, dur ing summer 1979-80 10 ,..--._ \ [\j E 0 8 ..j.J \/ " 0 0 .c (f) . \ (._ 6 Q) o_ ro ~ Q) (._ ro (._ 4 ro c ·rl E / '-......._ -ro � -------rl c 2 Q) Q) (._ (.!) 52 . S i 0 Nov Dec Jan Feb The observation of the timing and duration of seed�filling determined from the visual score of stage of development was confirmed by direct measurement of the dry weight of seeds which when expressed on a per seed basis was seen to increase rapidly over December to maxima in early January ( figure 4 . 25 ) . The more precocious development o f the fertile shoot population at S1 compared with that at S2 was evident throughout the spring/early�mid summer of 1 979-80 ( figure 4 . 23 ) . Thus , by early January 1 980 , mean stage of development of fertile shoots at S2 in the old deflation hollow was similar to that found three to four weeks previously on the low dune site S1 ( figure 4 . 23 ) . The rapid increase in mean dry weight of the inflorescence ( fi gure 4 . 26 ) could be attributed to the increase in mean dry weight of seeds ( figure 4 . 25 ) rather than the mean number of seeds . In both 1 978-79 and 1 979-80 mean number of seeds per inflorescence was observed to have reached a maximum at each site monitored by late November ( figure 4 . 27 ) . Earlier reduced estimates resulted from the incomplete elongation of the culm. More than one-third of the fertile shoots 1 979 . at S 1 and S2 had not developed beyond stage 5 by mid-November Thus , female spikes on these reproductive shoots were still to emerge , accounting for the reduced mean number of female spikes per inflorescence at S 1 and S2 at the mid-November harvest ( figure 4 . 28 ) . Further� the most distal female spike was observed to possess fewer spikelets than later-to-emerge spikes lower on the culm ( table 4 . 26 ) . Thus � although the full complement of spikelets was likely to have been present in early summer , the mean number of spikelets counted on emerged female spikes was lower in mid-November 1 979 than that observed later in the summer ( figure 4 . 29 ) . F i gure 4 . 2o Mean s eed we i ght o f Carex pum i l a over -- ·'---- t i me at three s it�s on t h e sand p l a i n dur i n g summer ( a ) 1978-79 and ( b ) 1979-80 . Vert i c a l bars d enote ± standard errrors o f mea n s 5 . 5 (a ) 1 978-79 5 4 . 5 4 3 . 5 3 ,-.... 2 . 5 01 E "----"' 2 D 1 . 5 Q) Q) U1 1 (_ 0 . 5 Q) 0.. s i t e 2 s i te 1 s i t e 3 0�----------------------------------------- -t-J 4 . 5 ( b ) 1 979-80 ..c 01 ·rl 4 Q) 3 . 5 3: >- 3 (_ 0 2 . 5 2 1 . 5 1 0 . 5 /r! ,�;/I X s i t e 1 s i t e 2 �/ /T �± o ----��o�v�------_,�e�c�--------�a�n=----- G.) u c G.) u UJ Figure 4 . 26 Mean dry we ight o f seeds per in f lorescence o f Carex pumila over time at three sites on the sand p l a in dur ing summer (a) 1978-79 and (b) 1979-80 . 0 . 25 r (a ) 1 978-79 0 . 2 0 . 15 . -�-- · s ite 2 G.) r_ 0 . 1 / I .,. .,. ... s ite 1 0 r-t 4- c ·rl 0 . 05 - 0 -------- 0 ------- ------- ---- .,. ...... .,. ... .,. ------- s ite 3 r_ G.) 0.. o�-------------------------------------------- 0 0 . 25 (b ) 1 878-80 ....--.. s ite 2 0) � ...__... 0 . 2 ...j..J / ..c 0) ·rl 0 . 15 G.) 3: ------- s ite 1 --->- ---, r_ 0 . 1 I I 0 I I I I I I I 0 . 05 ; --- / 0 D V e c a n I <61 b Q) 0 c Q) 50 40 0 30 U1 Q) L 0 20 r-1 4-- c ·rl L Q) 0. iO F i gure 4 . 27 Mean (± standard error) number of s p i k e l e ts t
    1 J . (!) � ·rl n (J) c... Q) b. (J) "0 (!) (!) (J) 4- 0 c... Q) .n E :J z 25 20 15 10 5 F igure 4 . 29 Mean (± standard error) number o f seeds (sp i k e lets) per f ema l e sp ike o ver time at three s i tes on the sand p l a i n dur ing summer 1979-80 . o Site 1 • Site 2 i� A Site 3 · f 1�I ( �I !\ ! - I .- -1- X - - - - - - _ _ , · · · . . . I .. - · · t "" ..-. .. ,' 0�----------------------------�------------- NOV OEC JAN FEB 1878 1 880 . - . . . . , . Considerable variation was found both within and between sites in the size of inflorescences produced by Carex pumila � as a result of the number of female spikes that developed on each fertile shoot and the number of spikelets produced per spike. In mid�summer 1 979�80 � mean number of spikelets per inflorescence found on fertile shoots at S2 was greater than at S 1 . Both the number of spikelets per spike ( figure 4 . 29 ) and number of female spikes per inflorescence ( figure 4 . 28) contributed to this difference . The terminal spike of Carex pumila is male ; with occasionally �sually one very small partly female spike at its base ( figure 2 . 1 0 ) . Such spikes produced few ( 0 to 5 ) seed . The remaining one to four spikes on shoots observed in the present study were female , although o ccasionally the most distal flowers on these spikes were male ( figure 2 . 1 0) . Female spikes produced up to 60 seeds each . Table 4 . 26 shows that the mean number of seeds per female spike varies very little for inflorescences with two , three or four female spikes . Where inflorescences possessed a single female spike � the number of s pikelets was reduced . The average number of seeds produced by the larges t inflorescences in the old hollow at S2 was comparable with that produced by Carex arenaria shoots with five well-developed spikes per inflorescence ( Noble 1 982 ) . Page 1 89 Table 4 . 26 Mean number of spikelets (seeds ) per female spike on fertile shoots of Carex pumila at four si tes on the sand plain ; in December 1 979 , ( a ) on the most distal and most basal female spikes on the culm and ( b ) on inflorescences with different numbers of spikes . ( a ) so Most distal 1 8. 0 Most basal 25 . 0 ( b ) Number o f spikes per inflorescence 4 3 2 Number of seeds per spike S1 1 3 . 0 .:. 7 - 4 1 6 . 2 .:. 5 . 2 S2 22 . 2 + 1 1 . 2 27 • 4 .:. 1 0 • 5 S3 3 . 8 .:. 4. 2 8 . 7 .:. 3 - 1 Number of seeds per spike so S1 23 1 7 . 5 1 6 . 62 1 2 . 8 S2 24 .0 26 .48 25 . 56 1 9 . 0 S3 8 . 33 6 . 00 The maturation of fertile shoots of Carex pumila during the late summer/autumn 1 977-78 , 1 978-79 and 1 979-80 was accompanied by a decline in the mean dry weight of seeds per inflorescence ( figure 4 . 26 ) . This decline was largely attributable to the shedding of mature seed seen in the decrease in mean number of seeds per inflorescence over this late summer period ( figure 4 . 27 ) . However , even where no seed was seen to be shed , as from fertile shoots at S2 between January and April 1 978 , mean seed dry weight still declined ( figure 4 . 25 ) . This decline was attributed to the decrease in dry wei ght of the utricle which during stages 7 and 8 was photosynthetic and turgid but turned brown during the latter stages of fertile shoo t development ( stages 9 and 1 0 ) . Where reduction of both mean number and dry weight of seeds per inflorescence over the late summer periods was evident ; the mean dry weight per shed seed was es timated . Table 4 . 27 shows that shed seed was , on average , heavier than that retained by fertile shoo ts at each site monitored both in 1 977�78 and 1 978�79 ( compare with figure 4 . 25 ) . The estimated mean dry weight per shed seed at S3 in autumn 1 978 was 7 . 436 mg ( table 4. 27 ) � a value comparable with the maximum mean dry weight of seeds found in summer 1 977-78 at S2 ( figure 4 . 25 ) . Dissemules shed at S4 in both 1 977-78 and 1 978-79 and at S3 in 1 978-79 � from the more senile old deflation hollow populations , al though heavier than those remaining on shoots ; were so light ( table 4 . 27 ) that they were unlikely to contain viable seed . Page 1 91 Table 4 . 27 Estimated parameters of seed shed from time of seasonal maximum mean weight per seed (a ) in 1 977-78 to harvest 2 and (b ) in 1 978-79 to harvest 1 1 at various sites on the sand plain. Seed shed ------------------------------------------------ Number per inflorescence Dry weight ---------------------------------- mg/seed g/infl . g/30x30cm (a ) 1 977-78 Site 2 0 S ite 3 20. 7 1 7 . 4360 0 . 1 54 2 . 06 Site 4 1 6 . 29 2 . 1 486 0 . 035 0. 087 (b ) 1 978-79 Site 6 . 32 6 . 9620 0 . 044 0. 0972 Site 2 1 4 . 60 6 . 0274 0. 088 1 . 7776 Site 3 9 . 1 0 2 . 2333 0 . 020 0. 1 1 83 Site 4 4 . 00 1 . 7500 0. 007 0. 0 1 32 In summer 1 979-80 , the number of spikelets (seeds ) on each female spike , and number of spikes per inflorescence were counted and length of each spike was measured providing more direct evidence for the shedding of seed from midsummer . The linear regression of spike length on number of seeds per spike in mid-December 1 979 at each site monitored allowed the estimation of the number of seeds subsequently shed from those spikes on which it was apparent seeds were missing when harvested later in the summer . The linear regression equations use d to estimate numbers of seeds shed are given in table 4 . 28. Page 1 92 Table 4 . 28 Regression equations used to estimate seed number per female spike (y = seed number per spike ; X = spike length ) � in 1 979.:080 Site y = 1 • 21 1 X + 2 . 769 r2 = - 55 1 Site 2 y - 1 . 43 1 :x: - 0 .033 r2 '"' . 873 Site 2 Sheltered y • 1 . 1 1 7x + 4 . 62 r2 .. . 742 Site 3 y .. 1 . 288x - 0 . 62 1 r2 = . 894 In early January 1 980 � more than one half of the spikes counted at S 1 were found to have shed some of their seed . On average these spikes had lost 1 0 . 1 + 1 . 49 seeds each ( range 1 - 21 seeds / spike ) � which when averaged over all spikes in the sample was a mean loss of 5 . 82 � 1 . 78 seeds/spike ( table 4 . 29 ) . This decrease from mid-December to early January was sufficiently large to be seen as significant when viewed indirectly through the mean number of seeds remaining on spikes ( figure 4 . 29 ) . Figure 4 . 29 also suggests a decline in seed number / spike at S2 in the old hollow , although this dec line was not found to be statistically significant . The decline however � was real . At S2 � a small proportion ( c 5% ) of the spikes sampled were observed to have lost some seed ( range 3-37 seeds / spike ) � although when averaged over all spikes present in the sample the mean number of seeds lost per spike was small ( table 4 . 29 ) . The decline in seed number pe r spike average over all spikes present at S3 over this same period was similarly small ( 0 . 64 + 1 . 1 1 seeds/spike ) despite one.:Othird of the spikes having lost seed. The mean loss of dry weight per spike� per inflorescence and per uni t area as a result of the shedding of seeds was estimated on the basis that shed seed was on average equivalent to the mean dry weight/seed at the harvest preceding loss . The mean dry weight of seeds , per seed and per inflorescence continued to increase Page 1 93 at all sites monitored over this December-January period ( figures 4 . 25 and 4 . 26 ) when mature seed was being shed . The dry weight of seeds los t will therefore be underestimated since shed seed tended to be the larger � heavier ones on the inflorescence . This was confirmed by the direct observation of the small proportion of dropped seed that was incorporated in the samples in January and February 1 980 . This proportion of recaptured seed gives an estimate of the vagility of Carex pumila dissemules . Table 4 . 29 Mean loss of seeds (number and dry weight ) from December 1 979 to ( a ) early January 1 980 and (b ) early February 1 980 (� SD � n=4) . ( a ) Site Site 2 Site 3 Site Site 2 Site 3 ( b ) Site Site 2 Site Site 2 /30 x 30 cm 1 36 • 75 � 1 7 . 77 25 . 75 � 31 . 78 4 . 5 + 7 . 79 Number of seeds shed /inflorescence 1 3 . 34 � 3 . 74 3 . 0 � 3 . 1 2 1 . 1 3 � 1 . 95 Dry weight of seeds shed g/30x30cm g/inflorescence g/spike 0 . 45 0 . 08 1 0 . 006 /30 x 30 cm 1 77 . 5 � 53 . 39 86 . 0 � 57 . 6 0 . 044 0 . 01 9 0 .009 0 . 003 0 . 002 0 .001 Number of seeds shed /inflorescence 1 2 . 41 + 2 . 28 9 . 6 1 .:. 7 . 72 Dry weight of seeds shed g/30x30cm g/inflorescence g/spike 0 . 99 0. 299 0 . 07 0 .033 0 .027 0 . 01 4 /spike 5 . 82 � 1 . 78 1 • 06 + 1 . 1 6 0 . 64 + 1 . 1 1 mg/seed 3 . 2904 3 . 1 462 1 . 3469 /spike 7 . 035 � 1 . 02 4 . 07 � 3 . 1 8 mg/seed 3 . 8286 3 . 4766 Page 1 94 The decline in mean number of seeds per spike estimated either directly from the regression of seed number on spike length ( table 4 . 29) or indirectly from the mean number of seeds remaining on spikes ( figure 4 . 29 ) continued over January to early February 1 980 . There is also evidence that at S 1 whole female spikes were shed by fertile shoots over this same period . The mean number o f spikes per inflorescence in early February 1 980 was significantly ( P< . 05 ) lower than that observed at the previous harves ts ( figure 4 . 28) . Assuming shed spikes at S 1 each possessed 9. 9 seeds ( figure 4 . 29 ) then the loss of both number and dry weight of seeds from fertile shoots at site 1 over the late summer to February 1 980 is shown in table 4 . 29 The loss of seeds from fertile shoots at S2 and SO over this same late summer period was at tributed solely to the shedding of mature seed from spikes and not to the loss of whole spikes ( figure 4 . 28 ) . Seasonal maxima for mean seed dry weight and mean seed number were observed in December/January 1 978-79 and 1 979-80 ( figures 4 . 25 , 4 . 26 � 4 . 27 and 4 . 29 ) . Thus , the best estimates of these parameters in summer 1 977-78 and 1 980-8 1 would have been obtained at the harvests in January 1 978 and December 1 980 . The seasonal maxima for mean dry weight of seeds per inflo rescence and per seed and mean number of seeds per inflorescence at each of the old deflation hollow si tes dec reased from year-to-year as the resident Carex pumila populations aged ( figures 4 . 30 , 4 . 31 and 4. 32 re3pectively ) , reflecting the increasing senility of these populations . At SO and S 1 at which the Carex pumila populations were juvenile when first monitored ( in spring 1 979 and summer 1 977-78 , respectively ) � seasonal maxima for mean dry 0 . 1 - 0 Q) 0 . 4 - 0 c Q) 0 Ul Q) c.. 0 0 . 3 - r-t '+- c ..... ......... Ul 0 . 2 - "'0 Q) Q) Ul - !?) 0 . 1 - � 0 0 0 . 2 - 0 . 1 - Figure 4 .30 Seasona l max imum mean dry we i ght o f seeds per in f lorescence in four consecutive s ummers at var io u s s ites on the s a n d p l a in Ver t i ca l bars denote +- s tandard error (a) s ite 1 T 1 (b) s ite 2 T 1 T 1 I I .L J (c) s ite 3 T T o ------�---- --��� -----� -�� ������·---- -- -- -- ---- 0 . 2 - (d) s ite 4 (e) s i te 0 0 . 1 1977-78 1978-79 1979-80 1980-8 1 4 - 3 � 2 - 1 - 0 8 . 7 - 8 - 5 . 4 . "0 Q.l Q.l 3 Ul . c... Q.l 2 D. Ol E � .L 0 7 - +J .r::. 8 - Ol -rl Q.l 3::: 5 - >- c... 0 4 - 3 - 2 - 1 - 0 2 - 1 . 0 Figure 4 . 3 1 Seasona l max imum mean dry we i ght per seed of Carex pumi la in four consecut ive summers at f i ve s ites on the study area Vert ica l bars denote ± standard error of means ( ) ' t 1 a s 1 e (b ) s i t e 2 T j_ � ( ) . t 3 c 5 1 e (d) s · t e 4 1 . _I_ j_ ,... ...... I .L (e) s · t e 0 1 _I j_ --y-- 1 877-78 1 878-78 1 878-80 1 880-8 1 30 20 10 0 Q) 0 60 c Q) 0 50 (f) Q) L 40 0 rl 4- c 30 ·rl -.......... 20 (f) TI 10 Q) Q) (f) 0 Figure 4 . 32 Seasona l max imum mean number o f seeds per in f lorescence of Carex pumi la in f our summers at f ive s ites on the study area - (a) s i te 1 - - I I (b) 5 , tP 2 1 � - . . . . 4- 40 • (c ) s i te 3 o - L 30 • Q) .0 E :J z 20 - 10 - 0 70 - 60 50 - 40 - 30 . 20 . 10 . 0 (d ) s i te 4 (e) s i t e 0 I J J I 1977-78 1 978-79 1 979-80 1 980-8 1 Page 1 95 weight of seeds per fertile shoot and per seed initially increased between years as the populations matured (between 1 977�78 � 1 978�79 and 1 979-80 at S 1 � and between 1 979-80 and 1 980-8 1 at SO ; ( figures 4 . 30 and 4 . 31 ) . The maximum values obtained for these parameters on the low dune ( at S 1 in 1 979�80 and SO in 1 980�8 1 ) were lower than local maxima obtained in the deflation hollows ( at either S2 in 1 977�78 and 1 978-79 or S3 in 1 977-78 ) , reflecting the less favourable conditions for Carex pumila growth in the former habitat . As the population at S1 aged further in 1 980-81 � seed weight and number were considerably lower than that achieved in previous years ( figures 4 . 30 � 4 . 31 � and 4. 32 ) reflecting the senility of this population. Within any one season , the maximum values obtained for these various seed dry weight and seed number parameters in the old hollow populations reflected the proximity of the site to the terminal hollow and therefore the putative age the population. With increasing distance from the terminal hollow , maximum mean seed dry weight per inflorescence and per seed and mean seed number per inflorescence and per female spike decreased in a manner similar to that seen between seasons at each of these sites . Younger development on a deflation hollow si te were not phases of Carex pumila represented on the sand plain at any stage during the four summers of the present study. On the low dune however � such phases of development were represented . In 1 977-78 and 1 979-80 at S 1 and SO respectively� the Carex pumila populations were juvenile , producing no or few fertile shoots . Comparisons be tween SO and S1 for seed parameters was possible in 1 980-8 1 , when the three seed parameters were reduced at the older ( S1 ) compared with the younger ( SO ) population (figures 4 . 30 , 4 . 3 1 and 4 . 32 ) . Page 1 96 In summer 1 979-80 , shelter affected the development of fertile shoots of Carex pumila on the low dune at S1 differently from that on the deflation hollow at S2 . At S 1 shelter slowed the onset of maturity of fertile shoots , whereas the effect was reversed at S2 . At S 1 , the effect was significant ( P< . 01 ) by mid�summer ( table 4 . 30) . In early February 1 980 � mean stage of development of . fertile shoots on the control plots was 9 . 74 , with all shoots scoring 9 . 5 or 1 0 . A similar mean score was achieved on the sheltered plots in late February (nearly three weeks later ) and even then 25% of fertile shoots scored 9 . 0 (figure 4 . 33 ) . Further mean length of female spikes at S1 was increased by shelter ( table 4 . 31 ) and consequently mean seed output of these spikes was also increased. Shelter had no effect on the number of female spikes found on fertile shoots at this or any other site. Table 4 . 30 The effect of shelter on the stage of fertile shoot development Mean stage of development S 1 S2 Control Shelter Control Shelter December 8 . 78 ** 8 . 41 8 . 39 ns 8. 25 January 8 . 93 ** 8 . 60 8 . 64 ns 8 . 77 February 9 - 74 ( 9 . 73 ) 9 - 69 ns 9 - 85 ( Estimate in brackets was obtained 3 weeks after that for control plots ) By contrast , shelter hastened the onset of maturity of fertile shoots on the old deflation hollow at S2 . By early February , most fertile shoots on the sheltered plots at S2 had reached stage 1 0 , whereas in the controls more than half were at stage 9 . 5· or 9 (figure 30 >- 0 20 c Q) ::::J cr Q) (_ lL 10 0 . . - F i gure 4 . 33 The f requency d istr ibution o f the stage of deve lopment of f ert i le shoots i n February 1980, on sheltered and control p l ots S i te 1 S ite 2 Contro l She lter Contro l She lter r-- r-- r-- r-- ,...... - ,...... r-- In n n q q.s to q q.s 10 q q.5 to q q.s 10 Stag e o f d e v e l opment Page 1 97 4 . 33 ) . The mean difference however � was small and barely significant ( P= . 1 0) . The main expression of the more advanced development of fertile shoots on the sheltered plots at S2 � seen both in early January and early February 1 980 � was the significantly increased seed shedding. In early February 1 980 � the mean number of seeds remaining on spikes � which did not differ between treatments in December 1 979 , was highly significantly lower in the sheltered plots than in the controls . Further� at S2 shelter affected both the slope and the intercept on the y-axis of the linear regression of spike length on number of seeds per spike. The effect was a more uniform number of seeds per spike ; ie seed number for the larger than average spikes tended to be reduced , whereas for the shorter than average spikes � seed number pe r spike was increased by shelter. Page 1 98 Table 4 .3 1 The effect of shelter on (a ) mean spike length and (b ) mean number of s eeds/spike in December 1 979 and ( c ) mean number of seeds/spike at S2 over time . ( a ) Mean spike length (mm ) Control Shelter S1 1 1 . 52 a 1 7 . 32 b S2 1 7 . 48 b 1 5 . 73 b (b ) Mean number of seeds / spike Control Shelter S 1 1 3 . 57 a 23 . 7 b S2 23 . 67 b 25 . 83 b ( c ) Mean number of seeds I spike Control Shelter December 23 . 67 25 . 83 January 24 . 02 20 . 98 February 25 . 97 1 1 . 27 a < b ; P< . 05 ; Duncan ' s multiple range tes t . The effect o f shelter on fertile shoots a t S2 was similar t o the effect of increased temperatures on wheat development ; namely� to shorten the duration of grain..:.filling by hastening the onset of maturity and therefore reduce maximum seed yields ( Sofield et al . 1 974 ) . I t is likely that an effect of increased shelter at S2 was to increase the a·lready-elevated temperatures at this relatively calm old deflation hollow site. On the more exposed low dune ( S1 ) , the result of the reduced wind speeds within the shelters � to allow fertile shoots to continue growing into late summer � was probably related to the amelioration of the water-stress conditions . 4 . 3 . 6 Allocation of dry weight and total nitrogen 1 . Sward mass In January 1 978 � the pattern of distribution of the dry weight of the Carex pumila sward to roots � rhizomes and herbage was similar at all sites monitored , with the exception of S2 . At this latter site the resident population of Carex pumila was characterised by a large proportion ( 63%) of the total sward mass in herbage and a correspondingly small proportion ( 20%) in rhizomes . At the other three sites , S 1 , S3 and S4 � the balance of herbage and rhizome mass was reversed , with less than 30% of the total sward mass in herbage � and between 45% and 60% in rhizomes ( figure 4 . 34 ) . The balance was found in roots . This difference can be related to soi l fertility. It is well established that plants respond to low soil nutrient regimes by a decreased aerial : underground ratio ( eg Lovett Doust 1 980b ) . This response seen in rhizomatous perennial species as an increased proportional allocation to rhizomes allows the clone to remain and expand in situ in areas unfavourable to tall competito rs ( Ogden 1 974a ) . Plant density remained high at the expense of extensive rhizome growth on the relatively fertile site � S2 . The juveni le population at S1 could be dis tinguished from the other three populations by the proportion of herbage associated with live shoo ts - 1 00% at S 1 in contrast to S2 , S3 and S4 where at least 50% of the herbage was made up by dead shoots . The proportion of the sward attributable to live shoots (aerial biomass ) was however still greater at S2 than at S1 . At S3 and S4 , at increasing distances from the terminal hollow, aerial biomass made up a decreasingly small +> .c. Ol .,..., Cl.) X >­ '- 0 E c Cl.) t.> '­ Cl.l a. c C1.l Ol 0 '­ +> ...... :z: � 10 +> 0 I- Figure 4 . 34 Proport ional a l location o f dry weight and tota l nitrogen o f the sward mass o f Carex pumi l a t o aerial and underground c omponents, a t each o f four s ites (Si , S2, S3 and 54) , in summer (a ) 1977-78 and �) 1978-79 (a) 1977- 1978 1 0 0 l ' i 80 � 80 � i i 40 � I i I 20 -; � -0 I ' ! r ' I 20 � 40 ! I I ' ' 60 --; I ll � Dead shoots __ __ _ _ __ I -- --- -- ---- � Green shoots , -- - - - ·- I I Rh izomes �-- - I le . . - l �-- . 1-. ·--- --r� l l - I I I I Roots . 80 1 I 100 �� - - l=d �- - -LJ - - - --- S i S 2 S3 S4 S i te 100 80 60 40 20 0 20 40 80 80 100 (b) 1978-79 l I i � I ! -i I i I ! ' I _j I ! I I I � I l I _. I I .... I L __ _ _ F 1 1 I I I -- - �- - �--- I I S i S2 I - . �- -l - - - - I 53 54 S ite _D .J) .0 Page 200 pro portion of the sward � reflecting the increasing senility of these populations . The proportional allocation of the total nitrogen content of the sward mass of Carex pumila to aerial and underground components is also shown in figure 4 . 34 . At all four sites , green shoots made up a greater proportion of the total than they did as dry weight while roo ts , rhizomes and dead shoots made up a correspondingly smaller proportion. This difference was the result of the higher concentration of total nitrogen in green and seed tissues . Throughout 1 978 , the proportion of the Carex pumila sward found in herbage , measured both in dry weight and total ni trogen � continued to be greater at S2 than at all other sites . For instance ; in April 1 978 , 80% of the total nitrogen of the Carex pumila sward at S2 was found in aerial shoots - 72% in live shoots . However � the proportion of the sward mass of Carex pumila found in herbage ; and in particular aerial biomass , decreased at the three older sites between summer 1 977-78 and 1 978-79 ( figure 4 . 34 ) . As the maturity of the vegetation at S 1 increased during 1 978 � herbage allocation increased ( from 45% to 62% of total nitrogen between January and August 1 978) . The proportion of the sward mass at each site found in rhizomes in December 1 978 was similar to that found eleven months previously , whereas that proportion found in roots had increased ( figure 4 . 34 ) . Aerial biomass as a proportion of sward mass was only 1 1 . 1 % and 1 0 . 3% at S1 and S2 respectively ( 40% and 21 % for total nitrogen ) . These values were significantly greate r than those at either S3 or S4 ( figure 4 . 34 ) . Page 201 The decline in the proportional allocation of the sward to live shoots as the maturity of the population increased was also seen across the study area at sites at increasing distance from the terminal hollow in summer 1 979�80 (figure 4 . 35 ) . At sites of increasing age the proportion of the sward in live branches decreased . At S 1 where in summer 1 977-78 all shoots were living and vegetative � figure 4 . 35 shows that the total nitrogen apportioned to live branches was just 55% of the total and that 50% of this was apportioned to fertile shoots . 2 . Biomass The proportional allocation of biomass ( live branches ) of Carex pumila to rhizomes was greatest at those young sites towards the edge of the low dune where this species was spreading into the terminal deflation hollow. At the rhizome front 1 00% of the biomass of Carex pumila was underground , and almost entirely in the form of rhizomes . Adventitious root development usually follows the orthotropic development of branch modules ( section 2 . 1 ) . At S 1 during 1 978 and 1 979 and at SO during 1 979 and 1 980 , more than 50% of biomass was allocated to rhizomes . At older sites both on the deflation hollows ( S2 , S3 , and S4 ) and on the low dune ( S1 in summers 1 979-80 and 1 980-81 , and SO in summer 1 980-8 1 ) the proportion of biomass in rhizomes was considerably smaller ( table 4 . 32 ) . c: 0 ·rl +J (... 1 0 . 75 0 0 . 5 0. 0 (... n.. 0 . 25 Figure 4 . 35 The proportion o f the total n itrogen content of the sr�ard mass of Carex pumila found in vegetative, fertile and dead shoots, at four sites in summer i979-80 . Vegetati ve Fert i le Dead so S i .52 53 Page 202 Table 4 . 32 Proportional allocation of standing biomass of Carex :eumila to component parts � over time . Summer Winter Summer Winter Summer Winter Summer 1 977-8 1 978 1 978-9 1 979 1 979-80 1 980 1 980�1 Site 1 Roots . 21 3 . 265 . 265 . 065 . 05 1 . 023 . 09 Rhizomes . 579 . 462 . 52 - 475 . 320 . 5 1 6 - 372 Seeds 0 .002 • 1 41 . 028 Aerial . 209 . 273 . 21 8 . 46 . 488 . 46 1 . 5 1 0 Site 2 Roots . 1 5 . 1 62 . 222 . 06 .061 . 1 1 3 . 047 Rhizomes . 1 86 . 21 4 . 1 87 . 237 . 1 87 . 499 . 31 6 Seeds . 1 44 . 089 . 065 . 1 00 Aerial • 521 . 624 . 494 . 703 . 687 . 388 - 537 Site 3 Roots . 1 06 . 1 70 . 289 . 255 . 089 . 1 95 . 1 72 Rhizomes . 1 82 . 31 4 • 341 . 36 1 . 236 . 448 • 1 1 2 Seeds • 1 6 . 025 .047 0 Aerial . 552 . 5 1 6 . 345 . 484 . 628 . 357 . 71 6 Site 4 Roots . 29 . 339 • 361 . 1 60 ne ne ne Rhizomes . 377 . 406 - 359 - 473 ne ne ne Seeds . 01 1 . 006 ne ne ne Aerial . 322 . 255 . 274 . 367 ne ne ne Site 0 Roots ne ne ne . 050# . 1 23 . 032 . 026 Rhizomes ne ne ne . 524# - 542 . 737 . 257 Seeds ne ne ne 0 . 1 00 Aerial ne ne ne . 426# . 335 • 23 1 . 61 7 # Spring 1 979 Rhizome allocation in Carex :eumila fluctuated seasonally being greater in winter/spring than in summer when aerial allocation reached a seasonal peak ( table 4 . 32 ) . This trend was obscured in the juvenile population at S1 where the initially high allocation to rhizomes was not superceded the following winter . As this population aged , the seasonal fluctuation became apparent . At S2 � the proportional allocation of biomass to rhizomes � which was remarkably uniform in each of three successive summers ( 1 977-78 � 1 978-79 and 1 979-80 ) despite the winter increases , was significantly lower than that at other sites � and the aerial allocation significantly higher . Page 203 The proportion of dry weight devoted to aerial growth was greater in the more crowded populations compared with the less crowded populations . This was the case at S2 where aerial shoot densities were greater than at other sites . Aerial allocation also increased as populations increased in density with time on the low dune ( table 4 . 32 ) . Since aerial shoot parts contain higher concentrations of total nitrogen per gram dry weight than either roots or rhizomes ( table 4 . 1 7 ) , the increased proportional allocation to aerial parts in the more crowded populations � based on dry weights � was expectedly greater when based on total nitrogen . The reduced aerial allocation at the two older deflation hollow si tes ( S3 and S4 ) compared with S2 can be attributed at least in part to the greater senility of the former sites . Likewise as the population at S2 became senile by December 1 980 , aerial allocation was decreased compared with that found at this site in previous summers . The winter/spring maximum proportional allocation to rhizomes at S2 in 1 980 ( 55% of total DW) was more than double the previous site maximum obtained in winter 1 979 . Subsequently aerial growth did no t occur to the extent observed in previous years . The tendency towards a decreased aerial allocation as populations became increasingly senile was not apparent in all populations . As Carex pumila populations aged � shoot dens ity become increasingly reduced . Whether an aerial shoot in such a population was attached to a dwarf or a long rhizome branch markedly affected the proportional allocation ratio . In the extremely depauperate population at S3 in summer 1 980-8 1 , all shoots on the control plots were attached to dwarf rhizome branches . Proportional rhizome allocation was 1 1 % of biomass and aerial shoot allocation By contrast ; on nitrogen-fertilized plots at this site where the Carex population was Page 204 equally depauperate , but one-third of the aerial shoots were attached to long rhizome branches ( section 4 . 3 . 7 ) � the proportional allocation of biomass to rhizomes was markedly higher ( 24%) � and that to aerial shoot parts lower ( 63%) � than the controls . Reproductive effort (RE ) � the proportional allocation o f biomass to seeds ( sexual reproductive effort� sensu Ogden 1 974a ) � was minimal in young populations . At S1 in 1 978-79 , the Carex pumila population was sufficiently mature (up to 2 years old ) that fertile shoots appeared in the budgets on most plots at all harvests � although at an extremely low density. Population RE (RE as a proportion of total biomass ) at S 1 at the time of maximum seed weight was only 0 . 2% ( table 4 . 32 ) . At S 1 in 1 977-78 and SO in 1 979-80 where the Carex pumila populations were less than one year old � fertile shoots were encountered only infrequently . At most harves ts at these sites , all shoots sampled were vegetative . Thus , population RE was even smaller than that at S 1 in 1 978-79 . Maximum population reproductive effort found on the sand plain during the entire study ( 1 6% of biomass ) was obtained in summer 1 977-78 at S3 in the old deflation hollow. This value was approached only at S2 in the same summer ( 1 4 . 4% of biomass ) , and on the low dune as the populations matured � in 1 979-80 at S 1 ( 1 4 . 1 %) and in 1 980-81 at SO ( 1 0% ) . In older populations , reproductive effort was reduced . These data sugges t an overall trend of increasing reproductive effort with population age over the first years from the time of colonization of a site followed by a period of declining reproductive effort as the population aged further� becoming more senile . At SO and S 1 this pattern was shown with increases over the first two and three years respectively followed by a decline at S 1 in the fourth Page 205 summer. Only the declining phase of this pattern was shown in the deflation hollows at S2 , S3 and S4 . At the beginning of the study the resident populations at these sites were al ready mature showing maximal reproductive effort . The pattern between years (of an increase followed by a decline in reproductive effort with increasing seral maturity ) was also seen in January 1 978 , by comparing sites at increasing distances from the terminal hollow. An apparent compensation was seen between allocation to seeds and to new rhizomes in summer 1 979-80 . In December 1 979 , when seed weight was maximal , the sum of the proportional allocation of the total nitrogen content of Carex pumila biomass to seeds ( dispersal ) and to rhizomes of vegetative shoots ( current local colonization ) was closely similar at all four sites ( table 4 - 33 ) . These data apply to si tes where the allocation of total nitrogen to dispersal and local colonization can be accounted for exc lusively by rhizomes ( at the juvenile si te , SO) , predominantly by seeds (at the mature sites , S 1 and S2 ) or predominantly by rhizomes (at the senile site , S3 ) . Page 206 Table 4 . 33 Proportional allocation of total nitrogen of Carex pumila biomass to seeds , rhizomes of vegetative shoots and the sum of the two , at four si tes on the sand plain ( a ) in December 1 979 and ( b ) in December 1 980 with (N ) and without ( C ) N-fertilizer addition . ( a ) December 1 979 Proportion of total nitrogen so S1 S2 S3 Seeds 0 . 1 43 . 1 48 .047 Rhizomes (veg. shoots ) . 249 . 075 . 066 . 1 77 Sum . 249 . 21 9 . 21 5 . 223 ( b ) December 1 980 so S1 S2 S3 c N c N c c Seeds . 252 . 1 20 . 072 . 099 . 248 0 Rhizomes (veg. ) . 052 . 209 . 093 . 1 58 . 039 . 072 Sum . 304 . 329 . 1 65 . 256 . 287 . 072 The conclusion that rhizomatous extension and seed reproduction in Carex pumila are alternative processes is also supported by the summer 1 980-8 1 data . In December 1 980 , the total allocation of total nitrogen to dispersal ( seeds ) and local colonization ( rhizomes of vegetative shoots ) was similar at those sites where the Carex population was not in an extremely impoverished condition (namely, at SO , S1 nitrogen-fertilized plots and at S2 ; table 4 . 33 ) . In the old hollow at S3 and to a lesser extent on the control plots at S 1 , the allocation of total nitrogen to seeds plus rhizomes of vegetative shoots was somewhat reduced . At these two latter sites , nutrients were low and/or unavailable to Carex pumila as a result soil of the presence of later seral species . These data strongly sugges t a soil-nitrogen limitation to clonal growth and seed reproduction in Page 207 Carex pumila . Population reproductive effort in Carex pumila will be determined both by the proportion of fertile shoots in the total population and by individual reproductive effort � the proportion of fertile shoot biomass that is allocated to reproductive tissues � and finally to seeds . 3 . Fertile shoot biomass Figure 4 . 36 shows the proportional allocation of the standing aerial biomass of fertile shoots to component vegetative and fertile parts over time during 1 978-79 and of the standing total biomass of fertile branches over time during 1 979-80 . In each season , reproductive allocation increased to a maximum in December I early January at all sites . Total individual reproductive effort (allocation to all reproductive structures , inc luding seeds ) reached a maximum somewhat earlier at S 1 than at the other sites in the old hollows ( figure 4 . 36 ) . The more promiscuous development did not result in a continuing increase in reproductive allocation as at the other sites . In summer 1 978-79 , total RE (as a proportion of the aerial biomass of fertile shoots ) was similar at S 1 , S2 and S3 (about 30% of dry weight ) and higher than at S4 ( figure 4 . 36 ) . Given that maximum individual reproductive effort occurred in December I early January, the January 1 978 and December 1 980 estimates of reproductive effort can be expected to approximate seasonal maxima and may be compared with the 1 978-79 and 1 979-80 maxima. Figure 4 . 37 shows this comparison . (f) (f) rn E 0 ·rl .0 rl rn ·rl (_ Q) rn 4- Figure 4 . 36 �) Proport i ona l a l locat i o n o f aer i a l b iomass of fert i le shoots of Carex pumi l a t o component parts over t ime in summer 1978-79 . ( i ) s ite 1 1 r----------------------------------- 0 . 8 0 . 6 0 . 4 0 . 2 0 ""' Anci l la��P=-:�· � / Seeds ----- ----- Cu lm, leaves and bracts 0 ( i i ) s ite 2 c 0 1 r----------------------------------- ·rl 4-J (_ 0 n. 0 . 8 0 (_ 0.. 0 . 6 0 . 4 0 . 2 . ------ ----- � · -=::::::::::::: : Last seeds ---­ · ---- o �--�N�o-v----------�D�e-c--------�J�a-n ________ __ 1 878 1879 .. (/) (/) CO E 0 ·rl ..Cl rl CO ·rl '- Q) CO � 0 c 0 ·rl 4-.J '- 0 Figure 4 . 36 (a) cont inued : Al location o f aerial b iomass ( i i i ) s ite 3 1 �--------------------------------- ----- - � : ----- - ----- ��- · � 0 . 8 0 . 6 0 . 4 0 . 2 . . 0 Nov Dec Jan ( i i i ) s i te 4 1 ��-------==-�------------------- � . D. 0 . 8 0 ·, £. �-V '- 0.. 0 . 6 0 . 4 0 . 2 o �--�N�o-v----------�D�e --c--------�J � a - n ---------- 1 978 1 979 ( \ ) old hol low - S2 ( ii ) low dune - S 1 1.001;:::=;:::::==::::::::::::::::====���� r-==========:=::::===:=:-1 .8 c: 6 o · :;: .... 0 a. 0 .... c.. NOV 1 9 7 9 D EC loat l eawet dead • ••••• • •••• JAN 1 9 8 0 N figure 4.36b Pattern of allocation of fertile shoot biomass of Carex pumi la to component structures over time in summer 1 979-80, at S 1 and S2. 1 980 1 0 . 5 0 1 c 0 ·:-1 4J (._ 0 . 5 0 0. 0 (._ 0... 0 1 Figure 4 . 37 Proportional a l locat ion o f aer ial b iomass o f fert i l e shoots of Carex pumi la to component parts in mid-summer in four consecutive seasons (a ) S i t e 1 (b ) s (c) S i t e 3 77-79 78-79 (e) S i te 79-80 80-81 Seeds Sp ikes Vegetative --- 2o7 ct Page 208 The greatest individual reproductive effort was seen at that site where population reproductive effort was also greatest , ie at S3 in summer 1 977�78 when 49% of the aerial biomass of fertile shoots was apportioned to reproductive parts ; 42% to seeds alone ( figure 4 . 37 ) . Seed RE � and as a consequence total RE � in January 1 978 were highly significantly greater (P< . 001 ) at S3 than at S2 . further reduced . No ferti le shoots were present population at S1 in this first summer. Values at S4 were in the juvenile Individual RE was seen to increase between years at the two sites on the low dune that were juvenile when first monitored (namely� S 1 and SO ; figure 4 . 37 ) . By contrast � at the old hollow sites where the resident Carex pumila populations had been mature (S2) or more or less senile ( S3 and S4 ) at the beginning of the study; seasonal maximum individual RE progressively declined between years as the populations became more senile . Similarly at S1 between 1 979-80 and 1 980-81 where the Carex pumila population was entering a senile phase of development , individual reproductive effort declined ( figure 4 . 37 ) . This decline was a consequence of the inability of these depauperate shoots to channel resources into the development of sexual repro ductive parts . This response reflects the demise of the species on these sites . Figure 4 . 36b shows the proportional allocation of biomass ( including roo ts and rhizomes ) of fertile shoots to component parts over time during summer 1 979-80 . At both S1 and S2 seeds increased to a maximum in December when the allocation to rhizomes reached a low. Green leaf laminae and to a lesser extent stems and sheaths decreased over the entire duration ( November to February ) . Rhizomes contributed a considerably larger proportion at S 1 than at s2 ; and ( dead ) leaves contributed considerably more at S2 than S 1 . As a consequence of the Page 209 large contribution of underground parts to the total dry weight at S 1 ( 25- 35%) � seeds made up a much smaller proportion of total biomass than they did as a proportion of the aerial biomass alone ( 1 8% and 30% respectively in mid�December ) . At S2 where aerial parts made up 85-90% of total biomass of fertile shoots � this discrepancy was not so apparent ( seed RE � 1 4% and 1 6% as a proportion of total and aerial biomass of fertile shoots respectively in mid-December ) . Thus � whereas seed RE at S1 was seen to be highly significantly ( P< . 001 ) greater than at S2 when based on aerial biomass , the difference was not significant when based on total biomass . At S1 and S2 individual reproductive effort as a proportion of the total biomass of fertile shoots was lower in December 1 980 than in the previous summer. At the youngest site colonized 1 8 months previously, individual reproductive effort was similar to that at the mature site S 1 in summer 1 979-80 . The development of fertile shoots in the younger population ( SO ) in December 1 980 was delayed compared with that in the older population ( S2 ) . At SO the proportion of dry weight of fertile shoots associated with green leaves and culms ( 29 . 3%) was significantly greater ( P< . 05 ) than that of S2 ( 1 3% ) whereas the proportions were reversed for dead leaves ( 1 8 . 2% and 31 % for SO and S2 respectively ; table 4 . 34 ) . Despite the tardiness of development of fertile shoots at SO � it was unlikely that individual RE would have been substantially increased by subsequent seed filling. Page 2 1 0 Table 4 . 34 Proportional allocation of biomass of fertile shoots of Carex pumila to component organs in December 1 980 . (Controls only ) . Proportion of total biomass of fertile shoots so S1 S2 Roots ' . 03 . 066 . 073 Rhizomes . 273 • 31 . 287 Seeds • . 1 4 1 . 046 • 1 1 1 Ancillary fertile parts . 048 . 036 . 01 3 Dead leaves . 1 82 . 295 • 31 Green leaf laminae . 1 42 . 1 07 . 047 Green sheaths . 1 5 1 . 093 . 083 Scale leaves . 034 . 044 . 049 The pattern of dry weight partitioning between component parts of fertile shoots was significantly different from that based on total nitrogen content . In January 1 978 � individual seed RE as a proportion of aerial biomass was highly significantly less than that based on the total nitrogen content of these shoots ( table 4 . 35a) . Twelve months later in early January 1 979 seed RE based on dry weight and total nitrogen showed similar discrepancies . Individual seed RE based on total nitrogen was 1 0% greater than that based on dry weight . At this harvest � the vegetative portion of fertile shoots was divided into green and brown ( dead) fractions . The proportion of total nitrogen found in · dead leaves was 1 0% less than that based on dry weight ( table 4 - 35b) suggesting that the sink for labile nitrogen moving from leaves during their senescence was the seeds . fa Be. Table 4. 35 Proportional allocation of total nitrogen ( TN ) and dry weight ( DW) of the aerial biomass of fertile shoots of Carex Eumila to component parts � in summer ( a ) 1 977-79 and (b ) 1 978-79 . Proportion of total ( a ) 1 977-78 Site 2 Site 3 DW TN DW TN Seeds • 31 ** . 38 . 42 ** . so Ancillary fertile parts . 03 ns . 02 . 07 ns . 05 Vegetative parts . 66 ns . 60 • 5 1 ns - 45 ( b ) 1 978-79 Site Site 2 Site 3 Site 4 DW TN DW TN DW TN DW TN Seeds • 22 *** . 32 . 23 *** - 33 . 20 *** . 30 . 08 ** . 1 3 Anci llary . 1 0 ns . 07 . 04 ns . 03 . 07 ns . 05 . 04 ns . 03 Dead leaves . 65 *** - 55 . 64 *** • 5 1 • 71 *** . 60 . 83 ** . 76 Green leaves . 04 ns . 06 . 09 ns • 1 3 . 03 ns . 05 . 05 ns . 09 Again in summer 1 979-80 , the pat tern of allocation to component organs of the fertile shoot population based on total nitrogen differed significantly from that based on dry weight ( figure 4 . 38 ) . Not only the allocation to seeds , which was between 8% and 1 5% greater when based on total nitrogen depending upon the site , but also the proportional allocation to green leaf laminae , dead leaves and rhizomes differed when based on tota l nitrogen from that based on dry weight ( figure 4 . 38 ) . ?.tl 1 0 . 75 0 . 5 0 . 25 0 Figure 4 . 38 Proport iona l a l location o f dry we ight and tota l n itrogen of fert i le shoot populations o f Carex pumila to component parts a t 5 1, 52 and 53 in December 1979 DW TN - (a) S i te 1 Roots Rhizomes - Seeds - Sp ikes Dead leaves - Green lamina e culms Bracts 1 1 (b) S i te 2 c 0 . 75 0 ·rl +J L o a. 0 L 0 . 5 (L 0 . 25 0 �--------_J=====L�===-L_ ________ _ 1 0 . 75 0 . 5 0 . 25 0 - (c) S i te - - - 3 DW TN 2 \ \ C\ Page 21 2 4 . Fertile to vegetative shoot rat io The ratio of fertile to vegetative shoots has been used as a measure of reproductive effort of natural populationa of herbaceous species in sera! habitats ( Abrahamaon 1 975 ) . Such an estimate can be made on the basis of shoot density ( figure 4 . 39 ) or some res ource such as dry matter ( figure 4 . 40 ) or nitrogen . Figure 4 . 39 shows this ratio in four Carex pumila populationa in January 1 978 . The ratio was . greatest in the mature population at S2 and progressively decreased at sites further from the terminal hollow. These data suggest that the relative expenditure on reproduction ( cf vegetative production of new tillers ) was greater at the former si te . A similar trend was shown in figure 4 . 40 for the proportional allocation of aerial biomass to vegetative and fertile shoots . By contrast , table 4 . 32 showed that population reproductive effort was greater at S3 than S2 , in summer 1 977-78 . Figure 4 . 40 shows the proportional allocation of standing aerial biomass at the time of maximum seed reproductive effort ( figure 4 . 36 ) to vegetative and fertile shoots at various sites on the sand plain in each of four summers . It is apparent from figures 4 . 1 1 and 4 . 1 3 that maximum investment in fertile and vegetative shoot biomasa occurred at different times of the year ( namely� early summer and autumn, respectively ) and , therefore , a simple ferti le to vegetative shoot ratio at any one time of the year ( such as figure 4 . 40 ) may not provide the moat accurate measure of the relative importance of sexual reproduction and clonal growth in these populations . The juvenile populations at S 1 in summer 1 977-78 and at SO in summer 1 979-80 stand out since the entire aerial biomass was apportioned to vegetative shoo ts . As these two populations m�tured 0 ·r-1 2 . 4 2 1 . 8 +' 1 . 2 10 a: 0 . 8 0 . 4 0 - - - - - Figure 4 . 39 The rat io o f the dens ity of fert i le shoots to vegetative shoots at four s ites on the sand pla in, in January 1978 . Vertical bar denotes +- variance of the mean . I 1 T I - S i S 2 S3 S4 Site 21'2 a Figure 4 . 40 Proportiona l a l location o f aer ia l . b iomass o f Carex Qumila to vegetative shoots (ABV) and to vegetat ive (AYR) and ferti le fractions of reproductive shoots in four consecut ive seasons 0 . 5 __: i I i 0 l_·---���'1'·- 0 . 5 - 0 j_ ___ - 77-79 78-79 79-80 80-81 �eeds :Jp ikes AYR ABV 2 1 2 1::> Page 2 1 3 between years , the proportion of aerial biomass found in fertile shoots progressively increased ( figure 4 . 40 ) . This may be interpreted as an adaptive phenotypic response to increasing density and concomitant maturity by these populations . It is important to note that this shift from vegetative to fertile shoot growth as the populations aged � is not likely to have involved a change in the genetic structure of the resident populations : Carex pumila genets are long-lived , although individual ramets ( branch modules ) on the sand plain have a life expectancy of no more than 1 2 months . Such a phenotypic response by populations to increasing stand density and seral maturity have been no ted elsewhere by Ogden ( 1 974a ; for Tussilago farfara ) , Thomas ( 1 974 ; ( 1 975 ; for Rubus spp ) . for Hieracium) and Abrahamson The evidence in the old ho llows at S2 � S3 and S4 shows a different trend . In January 1 978 the contribution of fertile shoots to total standing aerial biomass of Carex pumila decreased from S2 to S3 to S4 both in absolute terms ( g DW / unit area ; figure 4 . 9 ) and as a proportion ( figure 4 . 40 ) . The same ranking of these sites for allocation of biomass to fertile shoots was also seen in each of the three subsequent summers ( figures 4 . 1 1 , 4 . 1 3 and 4 . 40 ) . S2 � S3 and S4 � which were sited at progressively increasing distances from the terminal hollow� represented sites of progressively increasing age . A similar t rend o f decreasing reproductive effort over time was sugges ted at each of these sites by figure 4 . 37 . Confirming this trend , figure 4 . 40 shows there was a decreasing proportion of aerial biomass in fertile shoots at both S3 and S4 between years . However � at S2 this proportion remained more or less similar for three seasons before balancing in favour of fertile shoots in summer 1 980-8 1 . This shift towards an increased "reproductive allocation" at S2 was however Page 2 1 4 more apparent than real . The elevated fertile to total aerial biomass ratio in December 1 980 at S2 resulted from the lack of recruitment and growth of new vegetative shoots during the spring and early summer . The absolute size per fertile shoot and biomass per unit area of fertile shoots at S2 in December 1 980 had decreased substantially compared with previous seasons ( table 4 . 9 and figure 4 . 1 3 respectively ) , whereas the proportional allocation of aerial biomass of fertile shoots to seeds was no different from the low value found the previous year ( figure 4 . 37 ) . The trend of decreasing reproductive allocation with time on the old hollows (between years at each site and across sites of increasing putative age in any one season ) cannot be considered adaptive , but reflects the increasing senility of these Carex pumila populations which were being replaced on the sand plain by later seral spec ies . In the senile Carex pumila populations ( S1 in 1 980-81 ; S2 in 1 980-8 1 ; S3 in 1 978-79 , 1 979-80 and 1 980-8 1 ; and S4 in 1 977-78 and 1 978-79 ) fe.w shoots showed signs of sexual reproductive development and on those that did , few seeds were produced ( section 4. 3 . 5 ) . Shoots in these populations were depauperate compared with those in previous years ( table 4 . 9 ) with a large proportion of their biomass associated with dead leaves . The young populations at which little or no fertile development occurred , by contrast , had a considerably larger proportion of their aerial biomass in green leaves . 5 . Dwarf and long shoots (a ) Dwarf/total ratio The proportion of the biomass of Carex pumila associated with dwarf as opposed to long branches varied between sites and over time between July (winter ) , October ( spring) and December ( summer ) 1 980 I Page 21 5 ( table 4 . 36 ) . I t was shown above ( section 3 . 2 . 1 ) that the ratio of the biomass of dwarf to total ( dwarf plus long) branches increased with increasing age of the clone ( table 3 . 8 ) . At the July 1 980 harvest the low BD/BMS ratio estimated at SO ( namely 1 5% ; table 4 . 36 ) suggests the adolescence o f the resident population. Rabbits which preferentially grazed young and therefore mainly dwarf aerial shoot modules ( figure 4 . 1 8) contributed to the reduced BD/BMS value at this site. At S 1 , the BD/BMS ratio ( 65 . 9% ) reflected the greater age and contribution of dwarf shoots to this population. Similarly , at S3 where the Carex population was senile , the BD/BMS ratio was high , although also more variable ( larger standard deviation values ) than at SO . The BD/BMS ratio was also low ( 1 5 . 7% ) at S2 . The S2 value reflected the relatively large contribution of old large long branch modules to the standing biomass . This population was not young. Table 4 . 36 Ratio of the biomass of dwarf to total ( dwarf plus long) branches of Carex pumila at four si tes in July , October and December 1 980 . July October December Proportion of total biomass in dwarf branches ( BD / BMS ) so . 1 5 . 497 . 607 S 1 . 659 . 455 . 432 S2 . 1 57 . 401 . 38 1 S3 . 689 . 301 .492 . Page 21 6 At both SO and S2 where the BD/BMS ratio had been low in July 1 980 , BD/BMS increased over the spring ( table 4 . 36 ) . At SO this reflected increasing contribution of dwarf shoots to biomass increase Q A t S2 , the increased BD/BMS ratio to October came about through the death of old large long shoots . Dwarf shoo t biomass remained cons tant over this spring period ( figure 4 . 1 5 ) . BD/BMS remained more or less unchanged at S2 between October and December when both dwarf and long shoot biomass per unit area increased ( figure 4 . 1 5 ) . The proportional contribution o f dwarf shoots to total nitrogen content of biomass of Carex pumila in December 1 980 ( table 4 . 37 ) was greater than that based on dry weight ( table 4 . 36 ) . This effect was attributable to the greater proportional contribution of green leaves and seeds to biomass in dwarf branches compared with long branches ( see part ( b ) below ) . These organs were found to have higher total nitrogen concentrations than all other plant parts ( table 4 . 1 7 ) . At SO where the clone was vigorous and expanding, dwarf shoots accounted for a larger proportion of the total nitrogen content of biomass than at the three more senile sites ( 51 , S2 and S3 ) . At sites of increasing distance from the terminal ho llow� an increasing proportion of total nitrogen was found in fertile shoo ts . The proportion of the total nitrogen content of biomass accounted for by dwarf and by long fertile shoots increased at sites of increasing age ( table 4 . 37 ) . Dwarf shoots at S3 do not fit this trend . Dwarf shoots at 53 that did not produce inflorescenc es formed an age structure similar to that for fertile shoots elsewhere . It is likely that the aged dwarf shoots of this most senile population were fertile but lacked the resources to express this condition. Page 2 1 7 Table 4 - 37 Proportion of the total nitrogen content of the biomass of Car ex: pumila allocated to dwarf and to long branches at four sites in December 1 980 . Proportion of total nitrogen content of biomass so S1 S2 S3 Dwarf vegetative - 456 . 1 92 . 098 • 521 fertile . 275 . 324 . 406 . 071 Long vegetative . 01 7 . 1 92 . 057 . 048 fertile . 253 . 292 - 440 . 360 (b ) Allocation to component organs Dwarf branches were found to be significantly smaller than long branches (figure 4. 1 6 ) . This difference was attributable to the smaller mean dry weight per branch module of both aerial and underground fractions . The proportional distribution of the biomass of each shoot type to component parts also differed . Not only was the proportional allocation of biomass of each branch type to rhizomes notably greater for long branch populations but also the pattern of allocation within the aerial fractions differed (figure 4 . 41 ) . Averaged over all four sites � dwarf shoots allocated a significantly greater proportion of their aerial biomass to both scale leaves and sheaths of green leaves , and a correspondingly smaller proportion to dead leaves , than long shoots � sugges ting the younger average age of the former population. This was confirmed by the age distributions of these populations ( section 4 . 3 . 3 ) . c 0 ·r-1 .._; [.. 0 a. 0 [.. a... 0 . 8 F igure 4 . 4 1 Proportional a l location o f aer i a l b iomass o f dwar f and o f long shoots o f Carex �um i la to component organs, averaged over tour s ites in Ju ly 1980 0 . 6 � Dead leaves I I ' I i I I 0 . 4 � 0 . 2 0 -'------- Dwarf Long Green l am inae Green sheaths Scale leaves Page 21 8 ( i ) Long shoots In July 1 980 , the proportional allocation of biomass to component parts by long and by dwarf branch populations was not consistent between all sites (highly significant site x type interaction� P< .001 table 4 . 38 ) . At SO , S2 and S3 the proportional allocation of long branch population biomass to rhizomes was highly significantly ( P< . 01 ) greater than rhizome allocation by dwarf branch populations whereas at S 1 , the two branch types did not differ significantly ( table 4 . 38 ) . Despite this similari ty, the mean dry weight per long rhizome module at S 1 was still greater ( P< . 01 ) than that per dwarf rhizome module ( figure 4 . 1 6 ) • Table 4 . 38 Analysis of variance of the proportion of biomass of dwarf and of long branch modules in rhizomes , in July 1 980 SOURCE D . F . (M. V . ) s . s . M . s . s . site 3 0 . 1 2 0 .04 rep 3 4 . 905E-03 1 . 635E-03 type 1 · o . 53 0. 53 site*type 3 0 . 43 0. 1 4 error 1 7 ( 4 ) 0. 22 0 .01 Estimated treatment means : S3 S2 S1 so site 0 . 50 0 . 38 0 . 48 0 . 55 rep 0 . 49 0 . 46 0 . 47 0. 50 type Dwarf 0 . 33 Long 0 . 62 *** so ) S2, P<.01 Long > Dwarf ; P< . 001 Estimated means : The proportion of biomass of dwarf and branch modules in rhizomes in July 1 980 Dwarf Long Site 0 0 . 30a 0 .80c Site 1 0 . 57b 0 . 39a Site 2 0 . 21 a 0 . 56ab Site 3 0 . 25a 0 . 75bc a > b , P< . 01 ; a > c , P< . 001 , b > c , P< . 05 F RATIO 3 . 01 ns 0 . 1 2 ns 40 . 54 *** 1 1 • 01 *** of long Figure 4 . 42 shows that rhizomes contributed a greater proportion of long branch biomass than any other structure ; at each of the sites monitored . As a proportion of the long branch population biomass ; rhizome allocation was significantly greater ( P< . 01 ) at SO than at S2 and S 1 . At S3 rhizome allocation was also significantly greater (P< . 01 ) than at S1 . The value at S3 was intermediate between and not significantly different from those at SO and S2 ( table 4 . 38 ) . Both mean dry weight of the aerial fraction per long shoot and the proportional allocation of long branch biomass to aerial parts was greater at S1 than at all other sites in July 1 980 ( figures 4 . 42 and 4 . 1 6 ) . This was attributable to the greater proportional contribution of green and dead leaves to total long branch biomass at this site compared with other sites. Dead leaves at S2 also made up a significantly greater proportion ( P< . 01 ) of total long branch biomass than at SO ( figure 4 . 42 ) . At all sites the mean dry weight per long branch module declined between July and October 1 980 although at SO the mean aerial dry weight per long shoot ( figure 4 . 1 6 ) and the proportional allocation of biomass of long branches to aerial parts ( figure 4 . 42 ) were seen to increase over this period. At s 1 ; S2 and S3 ; the proportional allocation of long branch biomass to component parts remained more or less constant ( figure 4 .42 ) . By December 1 980 no living long shoots were found in the control plots at S3 . At the other sites many of the long ( and dwarf) shoots present in the winter and spring survived as fertile shoots . Long fertile shoots on the control plots allocated on average over SO , S1 and S2 1 1 . 4% of their aerial biomass to seed s ; a significantly smaller proportion than dwarf shoots (P< . 05 ; table 4 . 39) . Differences c 0 ·rl +J c... 0 0. 0 c... 0... 21C!Cl . F igure 4 . 42 (a) Proportional a l locat ion o f b iomass o f dwar f branches o f Care x pum i l a to component organs over t i me i n Ju ly , October and December 1980 . F = fert i l e shoots : V = vegetat ive shoots i n December . 1 : ( i ) s i t e 0 0 . 75 - 0 . 5 ·- 0 . 25 - 0 _j . . 1 --, ( i ) s i te I � 0 . 75 -0 . 5 ·- � 0 . 25 -' 0 _j _ ______ 1 --, o . 75 - 0 . 5 - 0 . 25 - 0 .l._ _ _ - - · 1 -l ( i ) s i te 0 . 75 ·- 0 . 5 - I 0 . 25 - I 0 ..l . Ju l 1 2 3 - - - Aug ·- Sep Oct 1 980 seeds sp ikes . -- . - - -- - - No v V roots rh izomes dead leaves Dec green laminae green sheaths sca le leaves Figure 4 . 42 (b) Proport iona l a l l ocat ion o f b iomass o f long branches o f Carex pumila t o component organs over t ime in Ju ly , October and December 1980 . F :::: fert i l e shoots; V == vegetat ive shoots in December . ( i ) s i te 0 i l n 0 . 75 -- I I 1 l ( i i ) s te 1 I c ' -� 0 . 75 J .j.J (_ 0 0 . 5 ....! 0. 0 fl 0 . ·25 � ' 0 j _ ____ _ ( i i i ) s i t e 2 1 l I 0 . 75 __: 0 . 5 ...; ; 0 . 25 _, I 0 j ___ _ ( i v ) s i te 3 I ; I o l. _ _ Ju l Aug Sep nn roots 1 , , rhizomes §Br��s �� I I . I dead leaves �: green laminae ). ::; green �heath: . . _ _ __ _ _ . _ _ _ . scale l.eaves Oct Nov 1 980 Dec Page 220 r between the sites with regard to the allocation of aerial biomass of long shoots to sexual reproductive parts however were found ( table 4 . 38 ) ; with that at SO being greater ( P< . 05 ) than at either S1 or S2 . There was no difference between long and dwarf shoots at S 1 in their proportional allocation to sexual reproductive parts ( table 4 . 40 ) . ��8pSrti6� Agf1Y���i�t Kfo��Rgeof0ter�rig sh88£gd�£t!�ftedet�0a�artsan� to long rhizome segments , in December 1 980 . ' SOURCE D . F . (M. V . ) s . s . M. s . s . F RATIO 5It9 ____________________ 2 __ ____________ o:o3 _______ o:o2 ______ 297o7-** rep 1 1 • 556E-03 1 • 556E-03 2 . 77 ns type 1 0 . 01 0 . 01 24 . 86 ** site*type 2 8. 753E-03 4 . 376E-03 7 . 79 * error 5 2 . 809E-03 5 . 6 1 8E-04 Estimated treatment site S1 0. 08 so 0. 20 S2 0 . 1 7 means : rep 0 . 1 4 0 . 1 6 L . S . D . 5% L . S . D . 1 % L . S . D • • 1 % 0 . 04 0 . 07 o . 1 1 0 . 04 0 . 06 0 . 09 Estimated means : Rhizome type Site 0 Site 1 Site 2 Proportion of Dwarf 0 . 24a 0 . 07c 0 . 23a type Dwarf 0. 1 8 Long 0. 1 1 0 . 04 0 . 06 0 . 09 aerial biomass Long 0 . 1 5b 0 . 08c 0 . 1 1 bc a > b , P< . 05 ; a > c , P< . 00 1 ; b > c , P< . 05 A further difference between long fertile shoot populations ( sites ) in December 1 980 was found in the proportional allocation of aerial biomass to dead leaves ; which was significantly less at SO compared with S1 and S2 ( table 4 . 41 ) . Page 221 Table 4 . 40 Total reproductive effort as a proportion of aerial biomass of fert ile shoots attached to dwarf and to long rhizome segments � in December 1 980 . Proportion of aerial biomass Rhizome type Dwarf Long Site 0 0 . 32a 0 . 21 b Site 0 . 1 4c 0 . 1 3c S ite 2 0 . 32a 0 . 1 5c a > b , P< . 01 ; a > c , P< .001 ; b > c , P< . 05 Table 4 . 41 Analysis of variance of the proportion of aerial biomass of fertile shoots attached to dwarf and of long rhizome segments in dead leaves , in December 1 980 . 1 9 SOURCE site rep type site*type error D .F . (M. V . ) 2 1 1 2 5 Estimated treatment means : site rep S 1 0 . 46 0 . 40 so 0. 27 0 . 4 1 S2 0 . 49 L. S . D . 5% 0 . 1 2 0 . 1 0 L. S .D . 1 % 0 . 1 8 0 . 1 5 L. S . D . . 1 % 0 . 31 0 . 25 Estimated means : Dwarf Site 0 0 . 23a Site 1 0. 53bc Site 2 0 . 39b c > b > a , P< . 05 ; c > a , s . s . 0. 1 2 1 . 605E-04 5 - 54 1 E-03 0 . 06 0 . 02 type Dwarf Long 0 . 1 0 0 . 1 5 0 . 25 Long 0 . 31 ab 0 . 39b 0 . 58c P< . 01 M. s . s . 0 . 06 1 . 605E-04 5 - 541 E-03 0 . 03 4. 1 64E-03 0 . 38 0 . 43 F RATIO 1 3 . 87 ** 0 . 04 ns 1 . 33 ns 6 . 75 * The mean December values for aerial and total dry weight per long fertile shoot at all sites which were greater than the corresponding values per long shoot in October ( figure 4. 1 6 ) indicate growth in individual shoot size over this period. At SO and S 1 � the ?age 222 distribution of biomass of these shoots to aerial versus underground parts was similar to that in October ( figure 4 . 42 ) suggesting that sexual reproductive development resulted from the reallocation of aerial resources . The proportion of the biomass of long fertile branches in sexual reproductive parts (namely 7- 1 1 % ; 4-8% in seeds alone ) was similar to the early summer decrease in proportional allocation of biomass to green leaf laminae (figure 4 . 42 ) . By contrast � at S2 the proportional allocation o f long branch biomass to aerial parts increased between October and December at the expense of the rhizomatous fraction ( figure 4 . 42 ) . At S 1 and S2 where long vegetative branch modules were present in December 1 980 , the proportional allocation of biomass to rhizomes and the mean dry weight per long rhizome module were greater in long vegetative than in long fertile shoo ts . A further difference between these two long shoot populations � averaged over both sites � was that the former possessed a significantly reduced proportion and mean dry weight per branch module in dead leaves compared to the latter, confirming the difference in age of the two populations seen in section 4 . 3 . 3 at both sites . ( ii ) Dwarf shoots In July 1 980 � the mean dry weight per dwarf branch module was similar at all four sites monitored on the sand plain ( figure 4 . 1 6 ) . However� the distribution of biomass of these dwarf shoot populations differed between the sites . Dwarf branches at S 1 possessed a significantly greater (P< . 01 ) proportion of their to tal biomass in rhizomes ; and consequently a smaller proportion in aerial parts ; than comparable branches at so ; S2 and S3 ( figure 4 . 42 ) . Page 223 The greater s enility of the dwarf shoot population at S3 compared with those at the other three sites was indicated by the greater proportional allocation of aerial biomass to dead leaves (more than one half aerial biomass at S3 ; cf less than one�third at each of the other three sites ; figure 4 . 42 ) . The proportional allocation of total biomass to roots was also larger at S3 ( 1 6% ) than at other sites ( 0-3%; figure 4 . 42 ) further indicating the advanced age of this shoot population . No adventitious roots were found attached to dwarf rhizome branches at SO where the population was young and in an adolescent phase of development . It was not surprising; therefore ; to find that at s3 ; the mean dry weight per dwarf branch module declined over the early spring to October ; while at other sites the mean size of dwarf shoots increased over this period ( figure 4 . 1 6 ) . The increase in mean dry weight per dwarf shoot between July and October 1 980 at S2 was only small and occurred more in the rhizome fraction than in other shoot parts . However ; this increased proportional allocation of biomass to rhizomes at S2 portended the subsequent increase in mean aerial (and total) dry weight per dwarf shoot to December at this site ( figure 4. 1 6 ) . At SO and s 1 ; where the July-October increases in mean dry weight per dwarf branch module were greater than at s2 ; the proportion of the total dwarf shoot biomass allocated to green leaf laminae increased from 21 % and 7% to 41 % and 33% respectively , significantly (P< . 05 ) greater percentages than at either S2 or S3 ( figure 4 . 42 ) . This increased allocation to leaf laminae at SO and S 1 was balanced by a reduced proportional allocation to rhizomes and bracts ( figure 4 . 42 ) , although , at least at SO , the mean dry weight per dwarf ·rhizome branch increased over this period ( figure 4 . 1 6 ) . Page 224 By December 1 980 , reallocation of biomass within dwarf shoots in response to flowering and seed production was evident on the control plots at SO , S1 and S2 but not at S3 ( figure 4 . 42 ) . The mean dry weight per dwarf fertile branch module had further increased from the October values at SO and to a lesser extent at S2 , whereas at S1 dwarf fertile shoot mean size was less than that of the dwarf shoot population in October ( figure 4 . 1 6 ) . The proportion of the aerial biomass of dwarf fertile shoots found in sexual reproductive parts , including seeds , was highly significantly (P< .001 ) greater at SO and S2 than at S 1 ( tables 4 . 37 and 4 . 38 ; see also section 4 . 3 . 7 figure 4 .48 ) . This total reproductive effort is seen in figure 4 . 42 to have occurred between October and December more as a drain on the rhizome than on the vegetative aerial fraction although at SO the proportional allocation to vegetative aerial parts by dwarf fertile shoots was also somewhat lower than that seen in October. The dead leaf fraction of total biomass of dwarf fertile shoo ts in December was seen to have increased from the corresponding values in October , both as a proportion (figure 4 . 42 ) and in terms of mean dry weight per dwarf branch module ( figure 4. 1 6 ) . In December � dead leaves made up a significantly great er proportion of aerial biomass of dwarf fertile shoots at S1 than at either SO or S2 ( tQb \e 4 . 4 \ ) . In contrast to their fertile counterparts , dwarf vegetative shoots showed a proportionately smaller allocation to dead leaves and a greater allocation both to green leaves and to rhizomes . This confirms the adolescence of the latter shoot cohort compared with the more mature fertile shoot population . This was also shown in figure 4 . 7 , where at all sites , vegetative shoots were seen to belong to the Page 225 younger age classes , compared with fertile shoots . Differences were also found between sites , in December 1 980 , in terms of the proportional allocation of dwarf vegetative shoo t biomass to component parts (figure 4 . 42 ) . At SO green leaf laminae made up a significantly greater proportion of dwarf vegetative shoot biomass than at S 1 ( P< . 05 ) , whereas dead leaves and scale leaves together made up the balance at the latter site. Figure 4 . 7 indicates a greater range of age classes within the dwarf vegetative shoot population at S 1 compared with SO in December 1 980 , on the control plots . The greater mean proportional contribution of scale leaves and dead leaves to aerial biomass of dwarf vegetative shoots at S1 compared with SO may reflect the contribution of very young (age class 0) and older (age class 3) shoots respectively in the population at S1 . Dwarf vegetative shoots in these age classes were absent from SO ( control plots ) at this harvest . 4 . 3 . 7 The effect of nitrogen fertilizer addition 1 . Results (a ) Biomass Nitrogen fertilizer addition ( equivalent to 50 kg N / ha ) in spring 1 980 had by December of that year increased the sward mass at all site s relative to the controls ( figure 4 - 43 ) . However , significant differences were found between sites representing different stages in the sere , for components of these populations . 750 700 650 600 550 C\1 E 500 -......... 450 3: 0 400 Ul 350 E m L 300 (J) 250 200 150 100 50 0 350 300 250 200 150 100 50 ) F i gure 4 . 43 The d i str i b ut i o n o f b i omass o f t he sward to Carex pum i l a and to other spec i e s w i th a n d w i thout n i trogen f e rt i l i z er i n December 1 980 (a ) Con tr o l D Other s p e c i e s � Car ex pum i l a (b ) N-f e r t i l i z e r s ite 0 s ite 1 s ite 2 s i te 3 22.5a Page 226 In the old hollow at S3 , the depauperate Carex pumila population� which constituted only a minor fraction of the sward mass � showed litt le or no response to nitrogen ferti lizer. The differences between the fertilized and control plots both in terms of biomass and shoot and shoot bud densities of Carex pumila were small and not statistically significant ( figure 4 . 43 and table 4 . 42 ) . At this site � species other than Carex pumila � notably Selliera radicans and Hypochaeris glabra , responded to fertilizer addition by increasing biomass ( figure 4 . 43 ) . However � the difference in amount of total nitrogen in the sward mass of the vegetation as a whole between control and fertilized plots at this site was equivalent to 9 kg N / ha , less than one-fifth of the nitrogen added . Table 4 . 42 Shoot and shoot bud densities at SO, S1 , S2 and S3 on the sand plain, in December 1 980 , with (N) and without ( C ) nitrogen fertilizer addition . Number of shoots or buds per 30 X 30 cm so S1 S2 S3 Shoot type c N c N c c N Dwarf vegetative 7 . 0 2 1 . 0 5 . 5 24 . 0 7 . 0 2 . 5 3 . 0 fertile 7 . 0 6 . 0 7 . 0 1 1 . 0 1 7 . 0 o. o 0 . 5 Long vegetative o. 5 6 . 0 1 . 0 4 . 0 1 . 5 o . o 0 . 5 fertile 5 . 0 2 . 0 2 . 5 1 . 0 8 . 5 o . o 1 . 0 Expanding buds 1 6 . 0 37 . 0 1 0 . 0 36 . 5 1 0 . 5 o . o o . o Dead shoots 1 . 5 0 . 5 1 1 . 0 4 . 5 Page 227 On the low dune at SO and S1 � . nitrogen fertilizer increased biomass of Carex pumila ( figure 4 . 43 ) . At these sites other species were not found in the harvested samples . The magnitude of the increase in the more adolescent population at SO � which was 50% greater than that at S 1 , was equivalent to 1 560 kg dry matter I ha or 20 . 7 kg nitrogen I ha ( figure 4 . 44 ) . The increase in total nitrogen represented almost half of the additional nitrogen applied three months previously . The increases in dry matter and total nitrogen at S1 ( namely 1 1 20 kg DM I ha and 1 3 . 1 kg N I ha ) were more simi lar to those at S3 than at SO . At SO , significant increases in dry weight ( DW) and total nitrogen (TN) per unit area occurred in response to fertilizer in root (P< . 05 for DW ; P< . 01 for TN ) , rhizome (P< . 05 for DW and TN ) � culm including sheath (P< .05 for TN) , and green laminar (P< . 05 for TN) fractions . Both DW and TN of seeds were similar under the two treatments whereas dead leaves were reduced ( figure 4 . 44 ) . At S1 , increases in DW and TN per unit area occurred in all branch components (figure 4 - 44) . These increases were relatively small (cf SO) and only for TN of roots was the increase found to be statistically significant (P< . 05 ) . e magnitude of the increases in rhizomes , leaves and seeds were however , greater than those in roots . (b ) Total nitrogen concentrations , %TN By bulking component shoot populations at each site , significant effects of fertilizer addition were masked . For example , nitrogen fertilizer did not affect the total nitrogen concentration ( %TN) of Carex pumila shoots , averaged over all vegetative organs at both SO and S 1 , and over both shoot types ( those attached to dwarf and to long rhizome modules ) and both states of shoot (vegetative and flowering ; 350 300 (\J E "-... 250 3: 0 200 (f) E CO 150 (_ Ol 100 50 3 . 5 3 (\J E 2 . 5 "-... z (J) E 2 � 1 . 5 Ol 1 0 . 5 F i gure 4 . 44 E f f ect o f n it r o g en f ert i l izer o n the d i str i b ut i on o f (a) dry we i gh t a n d (b) tota l n itrog en content o f Carex Qum i l a b i omass t o component o r gans a t two s it e s i n summer 1 980-8 1 (a ) Dry we i ght (b ) Tot a 1 n itrogen c N S ITE 0 1"---t �� c roots rhizomes seeds sp ikes dead leaves green laminae green sheaths scal e leaves N S I TE 1 227 0, Page 228 table 4 . 20) . However, there was a differential effect of fertilizer on component organs which differed between sites and between dwarf and long shoots (see significant ' site x treatment x organ ' and ' type x treatment x organ ' interactions � P< . 05 � table 4 . 20 ) . At the more mature site (S 1 ) , additional fertilizer increased the %TN of green leaves ( laminae and sheaths ) of dwarf vegetative shoots � and of underground organs , particularly rhizomes , attached to all but dwarf fertile shoots ( table 4 . 43 ) . At the younger site (SO) � increases were seen in green leaves (both sheaths and laminae ) of dwarf shoots . %TN values for long shoots remained constant or were decreased , as in the case of rhizomes attached to vegetative shoots at SO , and green leaf laminae and dead leaves at S1 • %TN of seeds on shoots attached to both types of rhizome module were increased by the nitrogen treatment at S1 ( table 4 . 43 ) . No such effect occurred at SO. Table 4 . 43 The effect of nitrogen fertilizer on the %TN of component parts of Carex pumila shoots at two sites on the sand plain. Rhiz Root Bracts Green Green Dead Seeds sheath laminae leaves ( a ) Dwarf vegetative Site 0 Control · Fertilizer Site 1 Control Fertilizer ( b ) Long vegetative Site 0 Control Fertilizer Site 1 Control Fertilizer ( c ) Dwarf fertile Site 0 Control Fertilizer Site 1 Control ( d ) Long fe�ti!�lizer Site 0 Control Fertilizer 0 . 49 0 . 6 1 o . 61 1 . 20 0 . 94 0 . 62 0 . 35 0 . 5 2 0 . 24 0 . 32 0 . 63 0 . 65 0 . 29 0 . 37 0 . 27 0 . 44 0 . 32 0 . 5 5 0 . 40 0 . 42 0 . 29 0 . 3 1 o . 1 1 0 . 1 3 0 . 44 0 . 42 0 . 1 9 0 . 1 0 Site Control 0 . 38 0 . 39 Fertilizer 0 . 66 0 . 5 5 a > b , P< . 05 ; a > c , P< . 01 0 . 34 0 . 73 1 . 82 0. 46 0 . 27 1 • 1 7 2 . 33 0 . 53 0 . 41 1 . 04 1 . 66 1 . 1 5 0 . 44 1 . 37 2 . 09 0 . 68 0 . 39 0 . 95 1 . 76 0 . 56 0 . 43 0 . 9 1 1 . 74 0 . 58 0 . 22 1 . 02 2 .00 0 . 96 0 . 29 0 . 7 1 1 . 40 0 . 52 0 . 25 0 . 1 6 1 . 35 0 . 23 2 . 09a 0 . 4 1 0 . 24 1 . 62 0 . 1 5 2 . 1 6a 0 . 48 0 . 81 1 . 59 0 . 97 1 . 96a 0 . 40 0 . 7 1 1 . 8 1 0 . 5 1 2 . 32bc 0 . 42 0 . 1 2 0 . 99 0 . 34 1 . 84a 0 . 35 0 . 20 1 . 07 0 . 3 5 2 . 09a 0 . 45 0 . 55 2 . 1 0 0 . 76 2 . 07a 0 . 3 1 0 . 64 1 . 87 0 . 69 2 . 53c Page 229 ( c ) Density Nitrogen fertilizer also affed the density and biomass o f dwarf shoots differently from that o f long shoots . Further� these responses differed between SO and S 1 . This site x shoot type x treatment interaction can be summarized by the effect of fertilizer addition on the ratio of dwarf to total ( long plus dwarf) shoots . On the basis o f shoot density , the dwarf to total shoot ratio was increased by the nitrogen treatment at both SO and S 1 ( table 4 - 44 ) , suggesting that fertilizer application had a greater effect on dwarf than on long shoot densities . Table 4 . 44 The effect of nitrogen fertilizer addition on the ratio of dwarf shoots to total ( long plus dwarf ) shoots ; (a ) density , (b ) total biomass and ( c ) aerial biomass . RATIO Dwarf / total shoot Control N-ferti lizer ( a ) Shoot density so 0 . 72 0 . 77 S 1 0 . 78 0 . 89 ( b ) Total biomass so 0 . 58 0 . 43 S 1 0 . 43 0 . 65 ( c ) Aerial biomass so 0 . 69 0 . 66 S 1 0 . 56 0 . 79 S2 0 . 53 S3 0 . 76 Page 230 At both SO and S 1 , fertilizer addition resulted in large increases in dwarf shoot and shoot bud densities in Carex pumila ( table 4 . 42 ) . By December , the bulk of the increased density of dwarf shoots was vegetative , . although at S1 dwarf fertile shoot density was also increased by this treatment (P< . 05 ) . This latter increase is of interest in that to be fertile , these shoots must have been present before the application of fertilizer in spring . The density of ferti le shoots at SO and of shoots attached to long rhizome segments at both sites ( SO and S 1 ) were not affected by _ this treatment . However , averaged over SO and S1 , the density of long shoots in which elongation of the culm and inflorescence development occurred was reduced by the nitrogen treatment (P . 056 ) with a concomitant increase in the density of long vegetative shoots (P = . 054 ; table 4 - 42 ) . The density of dead shoots observed three months after nitrogen fertilizer application at S 1 was reduced compared with the controls , whereas at SO and S3 no such effect was seen ( table 4 . 42 ) . At SO , the ratio of dwarf to total shoots based on biomass was decreased by nitrogen fertilizer , in contrast to that based on shoot density . At S1 , the ratio was increased by the fertilizer treatment ( site x treatment interaction significant , P< . 05 ; table 4 - 44 ) . This indicates a greater biomass response to nitrogen fertilizer addition by long relative to the dwarf shoots at SO , whereas at S 1 , the greater response was by dwarf shoots . This was confirmed when the biomass responses of component shoot populations to fertilizer addition were examined . Page 231 (d ) B!l weight and total nitrogen of component populations and organs Biomass of component shoot populations of Carex pumila was increased by the nitrogen treatment only where shoot densities were also increased. Significant effects were found in dwarf vegetative shoots at both SO and S 1 , dwarf fertile shoots at S 1 , and long vegetative shoots at SO ( tables 4 . 42 and 4. 45 ) . The size of the increases in dwarf vegetative shoot biomass at the two sites were similar . Despite the increase of almost one order of magnitude at S 1 , which was clearly evident in the field (figure 4 . 45 ) , the standing biomass of dwarf vegetative shoots on the fertilized plots was s till less than that on the control plots at SO . The increase in biomass of the long vegetative shoot population at SO (P< . 01 ; table 4 . 45 ) was largely attributable to the increase by more than 30-fold of the rhizome fraction (P< . 01 ; table 4 . 46 ) . The fertilizer-induced increase seen in the aerial fraction of these shoots was smaller , but nonetheless significant (P< . 01 , table 4 .45 ) . The increased total and aerial biomass of dwarf fertile shoots at S1 in response to the fertilizer perturbation was attributable to large relative increases in rhizome , green leaf, dead leaf and seed fractions ( table 4 . 46 ) . Page 232 Figure 4 . 45 The effect of nitrogen fertilizer on the vegetation on the low dune . Table 4 . 45 The effect of nitrogen fertilizer addition on (a ) total and ( b ) aerial biomass of component shoot populations of Carex pumila on the sand plain � in December 1 980 . ( a ) Total Dwarf vegetative fertile Long vegetative fertile ( b ) Aerial Dwarf vegetative fertile Long vegetative fertile so C N 5 . 08 * 9 . 2 5 . 56 ns 4 . 56 0 . 27 ** 1 4 • 5 1 7 . 55 ns 4 . 21 3 - 95 * 6 . 41 5 . 04 ns 4 . 1 9 0 ** 3 - 59 4 . 1 1 ns 2 . 06 grams dry weight I 30 x 30 cm S 1 S2 S3 C N C C N 0 . 56 * 4 - 99 o . gs * 3 - 45 0 .84 ns 4 . 6 1 1 . 1 S ns 0 . 57 o. 37 * 3 . 50 o. 72 * 2 . S7 0 . 23 ns 1 . 59 0 . 60 ns 0 . 41 O .S6 3 . 60 o .s6 6 . 37 0. 6 1 3 . 20 1 0. 1 5 3 . 1 8 0 . 1 7 ns 0 . 37 0 ns 0 . 06 0 ns 0 . 09 0 ns 0 . 21 0 . 1 2 0 ns 0. 25 ns 0 . 05 0 ns O . OS 0 ns 0 Table 4 . 46 Mean dry weight per uni t area of component organs of Carex pumila populations in table 4 . 45 . grams dry weight I 30 x 30 cm SO S1 S2 S3 C N C N C C N ( a ) Dwarf vegetative shoots 0 . 465 2 . 335 0 . 06 4 . 31 1 . 645 0 . 39 Roots 0 . 1 4 Rhizomes 0985 Dead leaves 0 . 625 Green laminae 2 . 1 75 Green sheaths O . S9 Bracts 0 . 26 ( b ) Dwarf Roots Rhizomes Seeds fertile shoots 0 . 02 0 . 01 0 . 5 0 . 36 1 • 21 3 1 . 27S Female spikes Male spikes Dead leaves Green laminae Culmslsheaths Bracts 0 . 31 2 0 . 342 0 . 085 0 .07 . 1 . 1 35 0 . 47 1 • 01 1 • 05 1 . 1 O . S? 0 . 1 S 0 . 1 1 ( c ) Long vegetative shoots Roots 0 . 005 2 . 25 Rhizomes 0 . 265 S . 675 Bracts 0 0 . 295 Sheaths 0 0 . 795 Green laminae 0 2 . 31 5 Dead leaves 0 0 . 1 S ( d ) Long fertile shoots Roots 0 . 355 0 . 255 Rhizomes 3 . 0S 1 . S9 ���gfe spikes 8 : �6§ 8: ?2� ���a ������ ? : 9g 8 : 9� 5 Green laminae Culmslsheaths Bracts 0 . 855 0 . 505 O . SS5 0. 375 o. 26 0. 1 6 . os • 1 1 5 . 065 • 1 6 . OS5 . 055 • 1 . 1 55 . 056 . 01 9 . 03 . 3S5 • 1 . 095 . 04 . 095 . 5 1 5 . 04 . 045 • 1 1 . 035 . 045 - 53 :Mg : 955 • 1 1 • 1 . 055 o . 61 0 . 885 0 . 2S 2 . 29 0 . 7 0 . 225 0 . 1 6 0 . 425 0 . 57 0 . 1 56 0 . 075 1 . 23 0 . 41 0 . 35 0 . 075 0 . 92 2 . 1 05 0 . 075 0 . 2S5 0. 955 0 . 27 . 01 . 1 45 : 8f� : ?M . os . 055 . 03 .075 . 1 75 . 07 . 25 . 1 75 • 1 1 5 . 075 . 32 . 757 . 1 69 • 1 1 1 . 25 . 23 . 45 . 235 . 055 . 66 . 025 . 045 . 075 0 . 66 2 . 35 : �5� 1 :g�5 . 235 . 375 • 25 . 03 . 02 . 05 . 04 . 02 • 01 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 8 0 0 0 . 08 . 04 . 065 • 1 . 055 . 025 . 005 . 005 0 0 . 005 . 02 . 01 . 01 0 . 005 . os 0 0 0 0 0 . 1 5 • 1 1 5 8 . 835 . 02 . 025 0 Page 234 The effect of nitrogen fertilizer addition in increasing biomass of various component shoot populations of Carex pumila could not be attributed solely to the increases in density of these shoots . Increased growth of shoots also occurred in response to the fertilizer treatment . Table 4 . 47 shows the effect of nitrogen fertilizer on the mean size of both dwarf and long and fertile and vegetative shoots at three si tes on the study area . Where nitrogen fertilizer did not affec t shoot density or biomass per unit area , mean branch size also was not affected ( table 4 . 47 ) . Nitrogen fertilizer did however affect the dis tribution of biomass within all component shoot populations ( see be low ) . A significant site x treatment interaction (P< . 05 ) was found for mean size of dwarf vegetative shoots at SO and S1 . At S 1 � the mean size of these aerial shoots under the fertilizer regime was not significantly greater than that of the controls ( table 4 . 47 ) . By contras t , the mean size of dwarf vegetative shoots was decreased by the fertilizer treatment , at SO . At S 1 , the mean dry weight of green leaf laminae and mean number of green leaves per dwarf vegetative shoo t module were significantly increased by the nitrogen treatment (P< . 05 , table 4 . 47 and figure 4 . 46 respectively ) . Figure 4 . 46 shows that most shoots at S1 including the new recruits arising during the period since nitrogen application had produced as many ( 3-6 ) or more ( 7-9) green leaves as the control shoots . The fertilizer-induce increase in aerial growth of dwarf vegetative shoots at SO (where dwarf shoot recruitment was also increased ) was insufficient for the average shoot to reach or surpass the mean size of the controls ( table 4 . 47 ) . The number and proportion of smaller than average shoots in the population was increased (figure 4 . 46 ) . However � the mean size of dwarf vegetative shoots in the N-plots at SO was still larger than F i gure 4 . 46 The f requency d istr i but i ons o f the number o f green (G) and o f dead (D) l e aves per shoot of Car ex pumi la on (a) contr o l and (b) N­ f e r t i l i z e d p l o ts , in December 1980 ( a ) CONTROL (b ) N-FERT ILIZER 14 12 10 8 6 4 2 ( i ) S i t e 0 Green leaves 14 12 10 8 6 4 2 Green l eaves 0 ... ferti I e shoots � = ve_se..+o.trve 5hooh o -.:..-.=.�.__.._._.u,c,"""'"'�"---- a ���������- Dead leaves Dead leaves 14 • 14 12 12 10 10 8 8 6 6 4 4 2 2 0 0 ..A,....=I.-j,J..�-4.------ ·I 0 1 2 3 4 5 8 7 8 9 10 0 1 2 3 4 5 8 7 8 9 10 14 12 10 8 6 4 2 ( i i ) S i t e 1 Green leaves Dead leaves 0 ...... """"""...._._.'-'-"UJ.J-_.._ ............ .__ _ 0 1 2 3 4 5 8 7 8 9 10 ( i i i ) S i te 3 : � Green l eaves 0 j wwwa� 0 1 2 3 4 5 8 7 8 9 10 14 12 10 8 6 4 2 a ���======���- 14 12 10 8 6 4 2 Dead leaves a -�������_.- 0 1 2 3 4 5 6 7 8 9 10 4 � Green leaves � �oWl� 4 � Dead leaves � �gow 0 1 2 3 4 5 6 7 8 9 10 Number o f l eaves p er s h o o t Page 235 that of dwarf vegetative shoots in the N-plots at S1 ( table 4 . 47 ) . Table 4 . 47 The effect of nitrogen fertilizer on mean dry weight per branch module of component parts of Carex pumila shoots � in December 1 980 . grams dry weight / branch module so S1 S2 S3 c N c N c c N ( a ) Dwarf vegetative shoots Total - 739 . 475 . 1 01 • 21 . 1 23 . 099 • 1 1 1 Aerial . 58 - 33 . 067 . 1 44 . 087 . 075 . 075 Roots . 01 9 . 029 . 01 3 . 026 • 01 . 01 5 . 024 Rhizomes . 1 44 • 1 1 3 . 022 . 04 . 025 . 009 . 01 3 Dead leaves . 088 .003 . 01 3 • 01 1 . 009 . 03 1 . 025 Green laminae . 31 8 . 224 . 03 . 095 . 037 . 025 . 026 Green sheaths . 1 32 . 086 . 01 4 . 029 . 025 . 01 3 . 01 5 Bracts . 036 . 02 .009 . 009 . 01 7 . 006 . 009 (b ) Dwarf fertile shoots Total - 794 - 772 . 1 38 • 31 1 • 21 1 0 • 1 1 Aerial • 71 9 . 69 1 . 1 02 . 257 . 1 88 0 . 09 Roots . 004 .002 . 01 4 . 01 5 .004 0 . 01 Rhizomes . 07 1 . 079 . 022 . 039 .01 9 0 . 01 Seeds . 1 73 • 21 4 . 008 . 05 . 045 0 0 Female spikes . 045 . 056 .003 . 01 4 • 01 0 0 Male spikes . 01 2 • 01 1 .004 .007 .006 0 . 01 Dead leaves . 1 62 . 072 . 054 • 1 1 1 .074 0 . 04 Green laminae . 1 44 . 1 78 . 01 4 . 037 .01 4 0 . 02 Culms/sheaths . 1 57 . 1 43 . 01 4 • 03 1 . 026 0 . 02 Bracts . 026 . 01 7 . 006 . 009 . 01 4 0 0 ( c ) Long vegetative shoots Total - 54 2 . 352 . 83 1 . 1 36 . 557 0 . 1 7 Aerial 0 . 565 • 22 . 38 . 1 00 0 0 Roots . 01 - 34 1 . 095 . 23 . 035 0 . 01 Rhizomes • 53 1 . 446 . 5 1 5 . 526 . 422 0 . 1 6 Bracts 0 . 049 . 04 . 01 9 . 01 8 0 0 Sheaths 0 . 1 33 . 045 . 07 1 . 03 0 0 Green laminae 0 - 353 • 1 1 . 239 . 052 0 0 Dead leaves 0 . 03 . 035 . 068 0 0 0 ( d ) Long fertile shoots Total 1 . 548 1 . 985 - 549 . 565 - 75 1 0 • 2 1 Aerial . 826 1 . 01 . 285 • 41 - 378 0 . 08 Roots . 084 . 1 85 . 03 . 01 . 077 0 . 01 5 Rhizomes . 638 - 79 . 234 . 1 45 . 296 0 • 1 1 5 Seeds . 1 26 . 287 . 023 . 028 . 043 0 0 Female spikes . 034 . 064 .009 . 01 7 . 01 0 0 Male spikes . 01 2 . 01 2 . 008 . 01 5 . 006 0 0 Dead leaves . 255 . 1 93 . 1 05 . 1 85 . 21 7 0 . 035 Green laminae • 1 7 . 208 . 065 . 08 . 028 0 . 02 Culms/sheaths . 1 75 . 1 68 . 048 . 055 . 045 0 . 025 Bracts . 052 . 08 . 023 . 03 . 029 0 0 Page 236 At SO , nitrogen was also seen to decrease both the mean number of dead leaves per dwarf vegetative shoot and the frequency of these shoots with two or more dead leaves each ( fi gure 4 .46 ) . This latter effect cannot be attributed to the increased recruitment of dwarf shoots s ince the time of nitrogen application ; but to a delay in the onset of leaf senescence by nitrogen . This conclusion is supported by the appearance in the treated plots ( c f the · controls ) of dwarf vegetative shoots with more than seven green leaves each ( fi gure 4 . 46 ) . The change in response to nitrogen in the balance of · dead and green leaves within dwarf vegetative shoots was also seen in the altered proportional dis tribution of the total dry weight of these shoots to green laminae and dead leaves ( see below ) . At SO the mean size o f long vegetative branches was increased over four-fold by the nitrogen treatment (P< . 01 ; table 4 . 47 ) , largely though an increase in the mean size of the rhizome fraction (P< . 01 ) , but also through the simple occurrence of aerial parts which were absent on the control plots . The samples taken from the control plots at SO contained only a single 30cm portion of one long rhizome branch module . The dry weight of this rhizome portion was similar to that estimated for long rhizome segments attached to fertile shoots . No significant fertilizer effect was seen at this site for this latter parameter. At S1 , the mean size of dwarf fertile shoots was greater on the treated plots , than on the controls . This difference was attributed to increases in the rhizome fraction and all aerial shoot components . These increases did not occur equally in all branch components ( see proportional allocation of biomass to component organs below ) . At SO , the mean size of fertile shoots attached to both dwarf and long rhizome branches showed no significant response to the nitrogen Page 237 treatment ( table 4 . 47 ) , although redistribution of dry weight within these shoots did occur ( see below ) . ( e ) Flowering and seed production Nitrogen fertilizer application affected flowering and seed production at both low dune sites . These effects were due to in part changes in fertile shoot densi ties and in part the redistribution of resources within these shoots . The changes in fertile shoot densities were outlined in section ( b ) above . Despite the reduced density of fertile shoots at SO in response to the nitrogen treatment , there was no de crease in mean dry weight of seeds per unit area . This resulted from an increased mean number of seeds per reproductive culm (P< . 01 for long fertile shoots ; P< . 05 for dwarf fertile shoo ts , table 4 . 48 ) . At SO , mean size and nitrogen concentration of seeds on shoots attached to both dwarf and long rhizome branches were not significantly altered by the fertilizer treatment ( table 4 . 48 and 4 . 43 , respectively ) . Thus � the to tal nit rogen content per seed at this site also was not found to differ significantly between control and fertilized plots ( table 4 . 48 ) . Page 238 Table 4 . 48 The effect of nitrogen fertilizer addition on ( a ) mean seed weight � (b) mean number of seeds per inflorescence and ( c ) total nitrogen content of seeds of Carex pumila on shoots attached to dwarf and long rhizome branches � in December 1 980 . ( a ) Mean seed weight gramslculm mglseed Control N-fertilizer Control N-fertilizer so Dwarf 0 . 1 73 ** 0 . 21 4 2 . 267 ns 2 . 436 wg�f 8 : 66§ ** 8 : 6�7 f : 36t �s � = 1�� S 1 ** Long 0 . 023 ns 0 . 028 1 . 929 ns 2 . 667 S2 Dwarf 0 . 045 2 . 542 Long 0 . 043 2 . 21 9 ( b ) Mean seed numberlculm Control N-fertilizer so Dwarf 76 . 64 * 87 . 81 Long 54 . 1 7 ** 1 09 . 5 S 1 Dwarf 5 . 68 ** 22 . 69 Long 1 1 • 5 ns 1 0 . 5 S2 Dwarf 1 7 . 24 Long 1 8 . 9 1 ( c ) Total nitrogen content grams N I m2 mg N I seed Control N-fertilizer Control N-fertilizer so Dwarf 0 . 282 0 . 307 0 . 0474 ns 0 . 0526 Long 0 . 1 29 0 . 1 1 8 0 . 0428 ns 0 . 0541 S 1 Dwarf 0 . 01 2 0 . 1 47 0 . 0255 * 0 . 0499 Long 0 . 01 1 0 . 008 0 . 0399 ns 0 . 0675 At S 1 , the density of fertile shoo ts and their seed output per unit area of ground were extremely reduced and significantly lower than at SO ( tables 4 . 42 · and 4 . 48 ) . Nitrogen fertilizer addition increased both density of fertile shoots attached to dwarf rhizome branches and seed output per unit area , measured in both dry weight and total nitrogen . This was achieved through increases in both mean number of seeds per inflorescence and mean seed size on dwarf fertile shoots . The similarly sized increase in mean dry weight per seed on sho ots attached to long rhizome branches at this site was not statistically significant . �age �;� At S 1 , %TN of seeds was significantly increased by the nitrogen treatment (P< . 05 and P< . 01 for dwarf and long fertile shoots ; respectively , table 4 . 43 ) . Thus ; with the increased mean dry weight of seeds (P< . 05 for dwarf shoots , table 4 . 48 ) , their total nitrogen content was also significantly increased ( P< . 05 ) and was at least as great as that at SO ( for seeds borne on shoots attached to both types of rhizome module ; table 4 . 48 ) . Thus ; the mean dry weight and total nitrogen content of seeds per inflorescence and per unit area were increased 5- 1 0 fold by the nitrogen perturbation ( tables 4 . 46 and 4 . 48 ) . However ; the seed output per unit area (grams DW and grams TN ) at S 1 , even with the large relative increases in response to ni trogen fertilizer , was less than that at SO by a factor of two ( tables 4 . 46 and 4 . 48) . ( f ) Proportional allocation of biomass and to tal nitrogen ( i ) Distribution of biomass (DW and TN ) between component shoots and organs Nitrogen fertilizer addition decreased the ratio of fertile shoot biomass to total biomass , based on both dry weight ( DW ) and total nitrogen (TN ) , at SO and S1 ( table 4 . 49 ) . This reflects the greater increase in vegetative shoot biomass relative to that of fertile shoots , in response to this perturbation, at both these low dune sites . This response occurred largely as an increase in the proportional allocation to green leaf laminae and to rhizomes of vegetative shoots ( table 4 . 50 ) . As a proportion of the to tal population, green leaves and dwarf vegetative shoots ; which are composed in a large proportion by green leaves (see below ) , constitute a considerably higher proportion of total nit rogen than dry weight ( tables 4 . 50 and 4 - 49 , respectively) . At SO , the proportional Page 240 allocation to seeds (population seed reproductive effort ) was lower on the fertilized plots compared with allocation of total nitrogen of biomass the to controls . dispersal The sum of the ( seeds ) plus local colonization ( rhi zomes of vegetative shoots ) was similar on both treated and untreated plots at SO and at S2 . On the controls at S 1 � this sum was reduced , although it was increased in response to additional nitrogen . In the senile population at S3 � this allocation to seeds and to rhizomes of vegetative shoots was further reduced ( cf controls at S 1 ; table 4 . 31 ) . Table 4 . 49 The effect of nitrogen fertilizer ( N ) on the proportional allocation of biomass based on (a ) dry weight and (b ) total nitrogen to c omponent shoot populations , in December 1 980 . C = control Proportion of total so S 1 c N c N (a ) Dry weight Dwarf vegetative shoo ts • 275 . 283 . 1 57 . 366 Long vegetative shoots . 01 5 - 447 . 236 - 338 Dwarf fertile shoots • 301 . 1 4 . 275 . 253 Long fertile shoots . 409 • 1 3 - 33 1 . 042 (b ) Total nitrogen Dwarf vegetative shoo ts - 456 - 357 . 1 92 . 529 Long vegetative shoots . 01 7 - • 378 . 1 92 . 1 96 �arfffeftile gho�ts ong er � e s oo s = �� � : 6�t : �§� : 6�6 Table 4 . 50 Effect of nitrogen fertilizer (N) on the proportional allocation of biomass based on ( a ) dry weight and (b ) total nitrogen of Carex pumila to component organs , in December 1 980 . Proportion of total SO S1 c N c N (a ) Dry weight Roots . 028 . 092 . 09 . 1 25 Rhizomes Fert . 1 92 . 063 . 1 89 . 035 Veg . 07 de - 345 . 1 8 hS . 226 Seeds . 1 0 .t" . 055 . 028 ns . 044 Green laminae • 21 9 . 252 . 1 35 . 275 (b ) Total nitrogen Roots . 007 . 035 . 039 . 048 Rhizomes Fert . 078 . 01 9 . 1 09 . 01 5 Veg . 048 .,s . 21 5 . 086 hs . 1 57 Seeds • 252 i:b' . 1 1 5 . 072 1\S . 099 Green laminae .44 . 498 . 309 . 48 1 Page 241 ( ii ) Allocation of DW and TN within component shoot populations In December 1 980 � the proportional allocation of dwarf vegetative shoot biomass to component organs differed significantly from that of long vegetative shoots � at all sites and under both treatments ( figure 4 . 47 ) . The difference was attributable to rhizome allocation which was highly significantly greater in long than in dwarf vegetative shoot populations ( table 4 . 5 1 ) . Nitrogen fertilizer addition reduced the proportional allocation of vegetative shoot biomass to rhizomes in the long shoot population at SO , despite the large increase in the absolute allocation of biomass to this fraction ( tables 4 . 46 and 4 . 47 ) . In the control plots at this site , long vegetative shoots were represented by a portion of a single rhizome segment , whereas on the treated plots , aerial shoot modules attached to this branch type were found . As a consequence , the proportion of biomass of these branches in rhizomes was reduced . Table 4 . 5 1 The effect of nitrogen fertilizer on the proportional allocation of biomass of (a ) dwarf and (b) long vegetative shoot populations of Carex pumila to rhizomes � in December 1 980 . Proportional allocation to rhizomes Site Control N-fertilizer (a ) dwarf shoots so . 1 9 c . 25 c S 1 . 22 c . 1 9 c (b ) long shoots so . 95 a . 60 b S 1 . 58 b . 49 b a > b , P< . 00 1 ; b > c , P< . 01 c 0 F i gure 4 . 47 The e f f e c t o f n it rogen f ert i l izer o n the p r opor t i onate a l l o c a t i o n o f b i omass of f er t i l e (F ) a n d vegetat ive (V) shoots attached to dwar f (D) a n d to l o ng (L) rhizome segments at SO, S 1 a nd S3; i n December 1980 S i te 0 DV OF LV LF 1.0Q C N C N C N C N .80 .60 .40 .20 0 C N S i t e 1 DV 1.00 .so C N OF C N LV C N LF :8 .60 (_ 0 g .40 (_ 0... .20 C N S i t e 3 DV 1.00 .80 .60 .40 .20 0 C N DV C N OF C N OF C N LV C N LV C N L F C N LF �cots Page 242 Long and dwarf vegetative shoots did not differ in the proportional allocation of their aerial biomass to individual aerial parts ( bracts � sheaths of green leaves � green leaf laminae or dead leaves ) at either SO or S1 on either control or ferti lized plots ( figure 4 . 48) . However� nitrogen fertilizer addition did increase the pro portion of aerial biomass found in green leaf laminae � at both SO and S 1 and in both dwarf and long vegetative shoot populations ( table 4 . 52 ) . The increase in long vegetative shoots at SO was small and not significant . On the nitrogen plots at both low dune sites , the proportion of the aerial biomass of long and of dwarf vegetative shoots found in green laminae ranged from 60 - 67%� values that were not significantly different from each other ( table 4 . 52 ) . The increased allocation of aerial biomass to green leaf laminae in response to the nitrogen treatment in the vegetative shoot populations occurred at the expense of the allocation to bracts at S 1 , and to dead leaves of dwarf shoots at SO ( figure 4 . 48 and table 4 . 52 ) . The proportional allocation of aerial biomass to sheaths of green leaves was remarkably similar in all vegetative shoot populations in December 1 980 on the low dune ( figure 4 . 48 ) . (/) (/) 10 E 0 •r-t .0 rl 10 •r-t t.. Q.l 10 4-· 0 c 0 •r-t � t.. 0 c. 0 t.. 0.. Figure 4 . 48 (a) The e f fect o f n itrogen fert i l izer on the pro­ port iona l a l l ocat ion o f aeria l b iomass of fertile shoots attach­ ed t o dwarf (D) and t o long (L) rhizome segments, December 1980 . ( i ) S i te 0 Dwar f Lon g C N C ...--:-N'-----. 1 0 . 8 0 . 6 0 . 4 0 . 2 0 ( i i ) S i t e 1 1 c N C N Dwar f C N C N Lo n g Seeds Anc.1\l� re-pro chvt:.. i'Ot"ts Dead leo.ves .. U) U) m E 0 ·rl ..0 r-1 m ....-! c... Q.) m � 0 c 0 ·rl +I c... 0 n 0 c... 0.. Figure 4 . 48 fu) The e f fec� o f n itrogen fertilizer on the pro­ p ortional a l location of aerial b iomass of vegetative shoots at­ tached to dwarf Wl and to long (L) rhizome segments, Dec 1980 . ( i ) S i te 0 1 0 . 8 0 . 6 0 . 4 0 . 2 0 ( i i ) S i t e 1 1 c N C N Qwar f C N C N Long Brac.tG �. / Page 243 Table 4 . 52 The effect of nitrogen fertilizer (N ) on the proportional allocation of aerial biomass of (a ) dwarf and (b ) long vegetative shoots of Carex pumila to component organs , in December 1 980 . C = control . (a ) Dwarf Green laminae Green sheaths Bracts Dead leaves (b ) Long Green laminae Green sheaths Bracts Dead leaves Proportion of total dry weight c • 55 . 23 . 07 • 1 6 so ** ns ns * N . 67 . 26 . 06 • 01 • 62 ns • 65 • 22 ns • 22 . 1 4 ns . 08 . 03 ns . 05 c . 45 . 22 • 1 4 • 1 9 S 1 N *** . 66 ns . 20 ** . 06 ns . 08 • 47 * . 60 • 21 ns • 1 6 . 1 8 ** . 05 • 1 4 ns • 1 9 Significance of differences are : * P< . 05 , ** P< . 01 ; *** P< . 001 At SO and S 1 , nitrogen fertilizer addition effected a redistribution of resources within fertile shoots ( figures 4 . 47 and 4 . 48) . This resulted in an increased total reproductive effort and seed reproductive effort , in fertile shoots attached to long and dwarf rhizome modules at SO and to dwarf rhizome modules at S 1 . At so; the dry weight and total nitrogen content of seeds per unit area was maintained on the fertilized plots ( tables 4 . 46 and 4 . 48 ) , despite the reduced proportion of the total population in fertile shoots ( table 4 . 49 ) . This redistribution of resources to reproductive structures at SO , which occurred within rhizome branches , appeared to parts (figure 4 . 47 ) . The Page 244 shoots attached to both long and dwarf be at the expense of other aerial shoot greates t reduction in the proportional allocation of aerial biomass of fertile shoots at SO was in the dead leaf fraction , for both shoot types ( figure 4 . 48 ) . At this site, the proportional allocation of dwarf fertile shoot biomass to green leaves was also increased by the fertilizer application . This juvenes cence was associated with an increased number of green leaves per shoot ( figure 4 . 46 ) . The increase in the proportion of dwarf fertile shoots at stage 8 and a reduction of those at stage 9 at this site also suggest that the development of these shoots to maturi ty was retarded by the fertilizer treatment ( figure 4 . 49 ) . At S 1 , nitrogen fertil!zer increased the proportional allocation of dwarf fertile shoot biomass to sexual reproductive parts at the expense of the rhizome fraction ( figure 4 . 47 ) . This redistribution of res ources was also seen as an increased mean number of seeds per inflorescence and an increased mean DW and TN per seed ( table 4 . 48 ) . At S1 , reallocation of resources within long reproductive shoo ts also occurred in response to nitrogen addition. This also was manifest as a decreased allocation to rhizomes and an increased aerial shoot allocation ( figures 4 . 47 and 4 . 48 ) . However, unlike that in dwarf fertile shoots at this site , reproductive allocation within the aerial fraction was unchanged . The proportion found in dead leaves of long fertile shoots at S1 was increased . At S3 in the old hollow, fe rtile shoots on the nitrogen fertilized plots had not developed beyond stage 5 by December 1 980 ( figure 4 . 49 ) . Nitrogen may have delayed the senescence of these shoots . An equally plausible and not mutually exclusive explanation is that the stunted fertile shoots which were found only on the 15 10 5 0 25 20 >- 0 c 15 Cl) :J cr Cl) 10 c... LL 5 0 F i gure 4 . 49 The e f f ect o f n itrogen f e rt i l izer on the f requency d istr i b ut i on o f the stage o f deve l opment o f f e rt i le shoots o f Carex pum i l a , i n December 1980 (a) CONTROL ( i ) S i te o 1 2 3 4 5 6 7 8 9 10 ( i i ) S i te 1 1 2 3 4 5 6 7 8 9 10 ( i i i ) S i te 2 1 2 3 4 5 6 7 8 9 !O 15 10 5 0 25 20 15 10 5 0 (b ) N-FERTI L I ZER 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 ( i v ) S i t e 3 1 2 3 4 5 6 7 8 9 10 Stage o f d eve l opment . . Page 245 N-treated plots at this site simply lacked the vigour or resources of either stored reserves or assimilatory power for further elongation of the culm and female spike and seed development . Table 4 . 53 compares the proportional allocation of the aerial biomass of component fertile shoot populations of Carex pumila to seeds based on grams dry weight with that based on grams total ni trogen as measures of resource investment . Since seeds were found to possess higher concentrations of total nitrogen than all other plant parts , with the exception of green leaf laminae , differences were expected between the two estimates . Seeds generally made up a higher proportion of the total nitrogen than of the dry weight of these fertile shoots . Table 4 . 53 Proportional allocation of dry weight and total nitrogen of the aerial biomass of fertile shoots of Carex pumila to seeds , in December 1 980 . (a ) Means Site so S 1 Shoot type Dwarf Long Dwarf Long Proportion of total in seeds Treatment Control N-fertilizer Control N-fertilizer Control N-fertilizer Control N-fertilizer Dry weight 0. 24 0. 3 1 • 1 5 0. 29 0. 07 0. 1 9 0. 08 0 .04 *** *** *** *** ns ** ns *** Total nitrogen 0 . 53 0 . 54 0 . 44 0. 62 0 . 1 4 0 . 43 0 . 1 5 0. 1 4 Significance o f differences are : ** P< . 01 , *** P< . 001 Table 4 . 53 (b ) Analysis of variance SOURCE D .F . (M. V . ) s . s . type 0. 03 site 0. 39 type*site 9 . 878E.:.03 treat 0 .06 type*treat 5 . 625E-03 site*treat 5 . 629E-05 type*site*treat 0. 05 rep 0 .0 1 res 0. 29 type*res 3 . 1 80E-04 site*res 0. 05 type*site*res 6 . 809E-03 treat*:r;-es 3 . 65QE ... Q3 type*treat*res 3 . 872E-04 site*treat*res 5 . 1 4 1 E-03 type*site*trea*res 6 . 325E-03 error 1 3 ( 2 ) 0 .04 M. s . s . 0 . 03 0 . 39 9 . 878E-03 0 .06 5 . 625E-03 5 . 629E-05 0. 05 0 . 0 1 0 . 29 3 . 1 80E-04 0. 05 6 . 809E-03 3 . 650E-03 3 . 872E-04 5 . 1 4 1 E-03 6 . 325E-03 2 . 798E-03 Page 246 F RATIO 1 1 • 94 1 38. 42 3 . 53 22 . 38 2 . 01 0 . 02 1 7 . 48 ** *** ns *** ns ns ** 4 . 96 * 1 02 . 87 *** o. 1 1 ns 1 8. 1 7 *** 2 . 43 ns 1 . 30 ns 0 . 1 4 ns 1 . 84 ns 2 . 26 ns Estimated treatment means and significance of differences between means type site treat rep resource Dwarf 0. 31 S1 0. 1 6 c 0 . 23 0. 25 TN 0. 37 Long 0. 24 SO 0 . 39 N 0 . 32 0. 30 DW 0. 1 7 ** *** *** *** rage 4� / The analysis of variance showed that averaged over all site x shoot type x treatment combinations � the difference between dry weight and total nitrogen estimates was highly significant ( P< . 001 ) � and that the magnitude of the difference was greater at SO than at S1 averaged over the four shoot type x treatment combinations ( site x resource interaction highly significant � P< . 00 1 ) . Table 4 . 53 shows that the es timates for all site x shoot type x treatment combinations to differ by a factor of two . At SO � all individual comparisons of dry weight and total nitrogen were highly significant ( P< . 001 ) , and at S 1 where nitrogen fertilizer significantly increased the proportionate allocation to seeds ( P< . 05 and < .00 1 for DW and TN ; respec tively ) , the · difference was also highly significant (P< . 01 ) . For the other site x shoot type x treatment combinations at S 1 � the differences were smaller in magnitude and , although all in the expected direction , not found to be statistically significant ( table 4 - 53 ) . The site x shoot type x treatment interaction term in the analysis of variance was highly significant (P< . 01 ) indicating that the effect of fertilizer was not similar for both shoot types at both sites . Nitrogen fertilizer significantly increased the proportion of both dry weight and total nitrogen of fertile shoots in seeds , only in dwarf shoots at S 1 and in long shoots at SO ( table 4 - 53 ) . In dwaif aAd loAg fePtile shoots at SO aAd S1, PespeGtively, no s�oh effeot of fePtili3eP wae obeePved (P>.05). No significant difference was observed between dwarf and long fertile shoots in the proportionate allocation of either dry weight or total nitrogen of the aerial biomass to seeds , on the control plots at either low dune si te . Only at S 1 , significantly increased the proportionate and < . 001 for DW and TN � respectively ) � where nitrogen fertilizer allocation to seeds ( P< . 05 did such a difference become Page 248 apparent ( P< . 001 ) . 2 . Discussion Nitrogen availability was found to be limiting plant growth in the study area . At all sites � each representing a different seral stage , nitrogen fertili zer addition increased both the sward mass and its total nitrogen content . The effects of nitrogen fertilizer application on the growth of populations of rhizomatous perennial species in various s tages of development on a sand dune sere have been reported elsewhere ( Huskies 1 979 ; Noble et al . 1 979) . On stands of both Carex arenaria and Ammophila arenaria , nitrogen-phosphorus-potassium (NPK ) fertilizer increased the density and flux of shoots (Carex) and leaves ( Ammophila) but only in situations where these pioneer species were the monopolist inhabitant . In well developed dune slack vegetation, other species were abundant and made the most visible response to the added fertilizer . The pioneers behaved as though little if any of the added nutrients were reaching their roots . Some of these observations were also made on Carex pumila in the sand plain sere in the present study. In the old hollow where Carex pumila made up only a tiny fraction of the sward , other species , principally the round-leaf Manawatu sand plain form of Selliera radicans ( Ogden 1 974b) and Hypochaeris glabra , made most response to added nitrogen fertilizer . The differences between the nitrogen and control plots in terms of Carex pumila were very small . On younger sites where this species was the sole or major inhabitant , nutrient application greatly increased the number of buds recruited into active growth . By December three months after Page 249 fertilizer application , many of these had given rise to an aerial sho ot . On the older site �where the population was entering a senile phase , these shoots had grown at leas t to the size of the controls . In the younger phase where Carex pumila was more vigorous , the nitrogen-induced cohort did not reach the size of their controls . Evidence o f the flux o f leaves and shoots within the population was circumstantial . If the flux of either leaves or shoot modules were increased by fertilizer addition � then it should be possible to observe the net effec t of this in a harves ted sample of the standing crop : namely, an increase in the number of new shoots and leaves and an increase in the number of dead leaves per shoot or dead ramets per unit area� provided these dead units were not lost . The evidence left no doubt that nitrogen fertilizer addition increased the birth rate of new aerial shoots on the low dune . However , the evidence for the increased death rate of leaves or ramets is equivocal . In the mature low dune phase additional nitrogen hastened senescence of reproductive shoots seen as an increase in the numbers of dead leaves per shoot and as an advance in the stage of shoot development (figure 4 - 49 ) . In vegetative shoots � nitrogen fertilizer increased the frequency of shoots with at least one dead leaf (figure 4 . 46 ) . However , the density of dead shoots at this site was reduced by the fertilizer application. This can be linked with the fertilizer-induced increase in density of reproductive shoots ( table 4 . 42 ) . These shoots predate the application of fertilizer . The response therefore can be attributed to the increased longevity of old dwarf shoots , that on the controls had died by December. This also accounts for the effect of ni trogen at this site of hastening the onset of maturi ty ( death ) in dwarf fe rtile shoots at this site , an effect not seen at SO ( figures 4 . 7 and 4 . 49 ) . Page 250 In the younger low dune phase there was no evidence of increased death rates of either single leaves or whole shoots in response to nitrogen . On the contra�, the evidence at this site pointed to nitrogen retarding senescence by slowing the maturation of fertile shoots concomitantly reducing the numbers of dead leaves on both fertile and vegetative tillers ( figures 4 . 46 and 4 . 49 ) . The conclusion that fertilizer increased the death rate of shoots of Carex arenaria (Noble et al . 1 979 ) was based on results five to seven months after fertilizer application. At three months when observations of the effect of a similar treatment was made in the present study , Noble et al . ( figure 8) show dead shoot densi ty to be no different between the fertilizer treatment and the controls . It is probable that the increased birth rate of Carex pumila shoots in response to nitrogen fertilizer application on the Manawatu sand plain , preludes an increased death rate that was not seen at the December 1 980 harvest , only three months after fertilizer application. On the sand plain , shoot mortality is associated with lowering water table levels in mid-late summer and flooding of the plain in winte r . The December harvest was therefore probably too early to reflect major changes in dead shoot densities on either the treated or untreated plots . This nitrogen effect at SO which was closer to the water table than S 1 is consistent with that often seen in crop species . Where ample soil water is available additional nitrogen leads to increased vegetative growth seen as a flush of new tillers and or an increased leaf area index (Scott et al . 1 977 ; for wheat ) . Page 25 1 The fertilizer-induced increase in densities of bo th expanding buds and dwarf vegetative shoots was likely the result of nitrogen breaking bud dormancy . This phenomenon is known from other rhizomatous perennial species (Mcintyre 1 96 5 ) and was said to be responsible for the decrease in number of dormant buds in Carex arenaria populations in the younger phases of development in response to NPK fertilizer application (Watkinson et al . 1 979 ) . Late spring dormancy of Agropyron repens buds was attributed by Waring ( 1 964 ) to the reduction of labile nitrogen in the rhizomes of this species as a result of the flush of aerial growth . Seeds are a major sink for elemental nitrogen in plants ( Jefferies et al . 1 979 ) . Thus where nitrogen is limiting plant growth , the number , weight or nitrogen concentration of seeds , or a comb ination of these , might be expec ted to be reduced . An increase in the availability of this nutrient ni trogen-fertilizer application under such ci rcumstances might therefore be expected to increase at least one of these paramete rs . On the low dune � nitrogen fertilizer increased the mean number of seeds per fertile culm. In the more senile of these populations at S 1 , fertile shoot density, mean seed weight and the nitrogen concentration of seeds were also increased . These effects strongly sugges t that on the older part of the low dune Carex pumila is normally under stress from a limited supply of nitrogen . As a consequence � the allocation to seeds � which behave as a sink for nitrogen , is reduced . A similar conclusion was reported for Salicornia europaea, an annual coastal halophyte in Norfolk, England ( Jefferies et al . 1 979 ) . In its coastal habitat , Salicornia is normally under stress brought about by a limited supply of nitrogen . In this situation seed numbers were reduced . Page 252 At SO , where the younger Carex pumila population was more vigorous , the evidence was not so unequivocal. Additional nitrogen did not affect the nitrogen concentration of seeds nor seed size. Despite the increased number of seeds per flowering shoo t � seed output per unit area was unchanged by the nitrogen treatment . In perennial sedges and grasses , the proportion of reproductive shoots in the total population is an important component of the reproductive effort . At SO , the density of fertile shoots was reduced on the fertilized plots compared to the controls . Thus , it could not be concluded that soil nit rogen on this young part of the sand plain was limiting seed output . However , nitrogen was un�oubtedly limiting vegetative growth in this young part of the clone . Some other factor would appear to be limiting seed output . This view is confirmed by the nitrogen concentrations of constituent organs of the species which were greater at SO than at other sites studied , inc luding those in the old hollow where nitrogen-fixing blue-green algae were evident. This suggests that absorbed nitrogen is transported acropetally to younger expanding plant parts . Reproduction is a lethal activity for Carex pumila shoots. At the same time vegetative growth , including rhizomatous spread � continues in adjacent parts of the clone . Some evidence was presented that within Carex pumila populations on the low dune , competition occurred for resources between vegetative expansion and sexual reproduction . At maturity , the biomass of fertile shoots bec omes concentrated in the seeds , and nitrogen more so . Seed production appears then to deplete the shoot of limited nit rogen that might otherwise be used for other life activities . At SO , especially on the fe rtilized plots , where the clone was growing most vigorously , the depletion of ni trogen from the vegetative parts of fertile shoots and Page 253 its proportionate allocation to seeds was greates t . Page 254 CHAPTER FIVE Discussion 5 . 1 The sand plains The sand plains of coastal Manawatu provide sites of particular ecological significance in that they are continuous ly in a state of succession. By virtue of the continuous nature of the formation of the sand plain (Esler 1 978 ) , sites suitable for colonization are a permanent feature of this habitat . Thus the youngest seral stages are readily encountered . It was on this basis that the study area was chosen. Carex pumila is the major colonist in this habitat . The aims of the present study were to describe the pattern of development of Carex pumila in populations of increasing seral maturity on the sand plain , and to determine if within a particular seral stage the population could respond to a deliberate perturbation treatment . Sequential harves ts were taken and the vegetation , which was dried and weighed , was divided into Carex pumila and other species . The former was further divided into component organs , including underground parts . Dead shoots were included and shed plant parts were estimated . In December 1 977 , an area was selected at the base of a landward-moving rear dune which was bare of vegetation. The newly exposed damp sand had not had suffici ent time or supply of germinable seed for plants to become es tablished . It remained essentially bare Page 255 of vegetation for two years des pite its appearance as an ideal environment for seedling establishment , especially in autumn and spring when the combination of mois ture and augmented temperatures prevailed . However , during the winter months the water table rose at least to the surface whereas in summer the lowered water table levels resulted in drying of the sand surface and further deflation during windy periods . Thus , seedling es tablishment was not without hazard on this area. In summer 1 978-79 , seedlings of several species failed to survive on the terminal hollow. They had the sand on which they were es tablished removed by wind . By winter 1 980 seedlings of several species had established and with the greater stability of the area as a result of marram planting on the surrounding dunes in spring 1 980 , these plants remained at the site over the ensuing summer . Thus , by winter 1 98 1 , the terminal hollow had the appearance of a dense plant cover . The essential patchiness of the young vegetation with locally pure stands of several pioneer species including Carex pumila , Limosella lineata , Selliera radicans , Eleocharis neozelandica and Scirpus nodosus was seen. The presence of Selliera radicans on the terminal hollow may not result from seedlings but the establishment of a plantlet from a thickened , bulbous � succulent leaf broken off from the parent plant . Such leaves were seen especially in autumn, and appear characteristic of the Manawatu sand plain form of the species . Thus , the colonization of the terminal hollow which was determined by chance factors of migration and establishment by seeds and plant fragments occurred too late during the study to allow anything but casual observation of the phenomena . Page 256 The field sites , chosen at increasing distances from the terminal hollow , represented sites of successively increasing age and seral maturity . Variation between sites in terms of quantifiable soil parameters was slight . A small degree of development of a humus layer on the surface in the old hollows � and nitrogen fixing blue-green algae were apparent in the old deflation hollows at the beginning of the study . However, it was not until winter 1 981 that similar quantities of Anabaena bloom were found in the terminal hollow. Plant available calcium , potassium , phosphorus and magnesium were greater in the older soils on the study area . Similarly percentage organic matter in the top 1 0cm increased with seral maturity. However, little decomposition had occurred and an extremely raw type of organic matter was present . The soil in the terminal deflation hollow and on the edge of the young low dune can be expected to have higher nutrient levels than raw beach sand , as a result of leaching of nutrients from the surrounding dunes ( cf Salisbury 1 952) , but otherwise these soils are little different . Variation in other physical environmental factors between sites was found that related to the topography of the sand plain . The relative shelter at S2 at the landward end of the old hollow adjacent to the young low dune was greater than at all other sites . Further , with the accretion o f sand by Carex pumila on the low dune � the sites S 1 and , in 1 979 and 1 980 , SO became less subject to fluctuations in water table levels . Sites in the old hollows ( S2 � S3 and S4) and in the terminal hollow ( SO in 1 978 and 1 979 ) were flooded in winter unlike low dune sites ( S 1 and , in 1 980 , SO) which were 300-500 mm higher, and out of reach of the water table. Page 257 The lowered water table levels in summer on the low dune would subject plants to long periods of moisture stress . Water table levels remained up to 1 000mm from the low dune surface at the end of summer. Given that the capillary fringe from a free water surface in very fine sand ( 30-50 microns particle size ) may be no more than 400mm (Ranwell 1 972 ) and that most roots attached to live shoot modules are found within 300mm of the soil surface , it is unlikely that plants on the low dune were able to absorb water from the water table or its capillary fringe during dry summer periods . The soil moisture profile on the low dune in spring with the water table at 580mm from the surface , shows that water makes up a little over 10% by volume of the top 1 00mm of sand . The primary source of this water comes from rainfall , being held in the surface layer by the low moisture-holding capaci ty of the sand . The vegetation analysis reflects the drought conditions that prevai l on the low dune . Other ecosystem parameters , similar to those sugges ted by McNaughton ( 1 975 ) to be important determinants of r- and K- selection , also showed wide variation between si tes at the beginning of the study : Interference (biomass , grams DW or TN / unit area ) Annual density variation (variance/mean shoo t density ) These vegetation parameters are wholly or largely determined by the performance of Carex pumila . 5 . 2 Carex pumila � the sand plain As a pioneer species , Carex pumila is doomed to extinction on the sites it colonizes . Thus , its success ultimately depends on its ability to put individuals into the early stages of the sand plain sere , elsewhere . Its strategy appears to be typical of that Page 258 encountered in plants of sand dune communities . It is successful in this highly disturbed , nutrient stressed habitat in that firstly � once present , it quickly expands to occupy the space available shutting-out other species , and secondly� it is continually escaping the interference from other more vigorous later seral species by either ( 1 ) spreading by rhizomes into adjacent uncolonized areas or ( 2 ) possessing a sufficiently large reproductive output that ensures the colonization by seedlings of newly exposed sites � elsewhere . 1 . Basic morphology The basic morphology and growth of Carex pumila on the sand plains of coastal Manawatu is remarkably similar to that of Carex arenaria , a sand sedge of coastal areas in Europe where it is abundant on mobile and semi-fixed dunes , but also found in moist dune slacks (Noble 1 982 ) . By contrast , Carex pumila is most abundant on the seasonally flooded slacks , but is also found on more elevated ground above the ground water table level . The modular construction of Carex pumila was readily apparent . The modular unit is a single aerial shoo t with its associated rhizome and adventitious roots which include both large diameter sinker roots , and finer more highly branched roots . Large diameter succulent roots are initiated in response to flooding in a variety of flood-tolerant plants (Keely 1 979 ) . The development of sinker roots preceded that of the finer root type, on young Carex pumila rhizomes invading the flooded terminal deflation hollow. The rhizome axis of Carex pumila is , then , a sympodium , made up by the serial addition of the basal portion of successive modules . Normally , the rhizome grows horizontally although , like other rhizome geophytes of habitats where rapid acc retion of the substrata occurs , it may grow vertically until the Page 259 apex arrives close to the surface when it reverts to horizontal growth (Raunkiaer 1 934 ) . Carex pumila is replaced � however � by dune species such as Desmoschoenus spiralis which have a greater ability to accumulate sand . Branching of the rhizome occurs at irregular intervals along its length , always at the base of the leafy orthotropic portion of a branch module . The length of each sympodial rhizome segment and the frequency of branching determine the density of the clone . On the edge of the low dune where Carex pumila was invading the terminal hollow , rhizome segments were often more than 1 m long with infrequent branching whereas in the old hollow at site S2 branching was more frequent and rhizome segments much shorter ( 20-60cm) . Contributing further to the density of the clone are short or dwarf modular uni ts which arise like extravaginal aerial shoot modules described grass tillers from the base of the above . Thus , where aerial shoot density is low as on the low dune tufts of loosely packed aerial shoots will form a patchwork . On the old deflation hollow aerial shoot density was sufficiently high to produce a dense ground cover obscuring the patchiness of these tufts . The linear rhizomatous archi tecture of Carex pumila like that of Carex arenaria ( Noble and Marshall 1 983 ) , allows for continual expansion into bare sand and the appropriation of scarce nutrients . The branching along the primary rhizome axis at an angle of 1 5 degrees allows for the more complete coverage of the moist bare sand of the bow-shaped terminal deflation hollow . As the bare sand is colonized by Carex pumila and other rhizomatous species , the process repeats itself as the rear dune moves further inland exposing more damp sand at its base (Esler 1 978) . Thus a catenary system of damp hollows separated by similar arcuate low dune ridges is formed on the sand Page 260 plain. The rhizome architecture and growth of Carex pumila allows for the rapid colonization and concomitant resource capture of the terminal hollow as the hollow is forming . The ability of this species to regenerate from seed allows for the more complete colonization of the growing area remaining unexploited by the rhizomatous spread of already existing clones at its edge . In winter 1 980 � a Carex pumila genet was excavated from the terminal deflation and measured . Two primary rhizome axes were evident , each with four successive long rhizome segments ( sympodial units ) . The mean length of each 4-unit axis was 2m, which , given the estimated seasonal rates of rhi zome extension, suggested a spring 1 979 germination and establishment of the excavated genet . Given time , clonal expansion may result in a single genet spreading over a considerable area , with the possibility that one or only a few genets may be responsible for the entire Carex pumila sward on the study area . This is not unreasonable given the infrequent observations of seedlings of this species on the sand plain during the study period . Excavated rhi zomes of Carex pumila on both the low dune and on the deflation hollows were seen to extend over 5m along a single rhizome axis , each of which may be branched one or more times over this distance. Fragmentation of the genet through decomposition of old buried rhizome segments prevented excavated rhizome axes being traced further back from the apex. The encroachment of the low dune by 1 8m onto the terminal hollow over the duration of the study , was the result of the contemporaneous extension of such branched rhizome axes onto the terminal hollow . The rate of extension of the rhizome Page 26 1 front , estimated at 5 . 08m per year , was greater than that determined for Carex arenaria growing into bare sand of either a dune face or a deflation hollow ( Noble 1 982 ) . Given this rate of encroachment of the vegetation onto the moist bare sand at the base of the windsweep of the rear dune , the ages of the field sites in December 1 977 were o ; 2 . 5 , 5 and 1 6 years for S 1 , S2 , S3 and S4 , respectively . These estimates are subject to the uncertainty of rates of extension of the sand plain in previous years . Over the study period marked seasonal and year-to-year variation was found in rates of extension of the rhizome front into the terminal hollow. Further, the estimates of site age , at least of the two most distal sites ( S3 and S4 ) ; may be inflated , given the manner in which the deflation ho llows are colonized : ie not exc lusively by the gradual systematic encroachment of rhizomes of Carex pumila , but also by chance migration and establishment of seeds and plant fragments , of this and other species . 2 . Flowering and seed production As a consequence of the pattern of formation of the sand plain new sites suitable for colonization by seedlings of Carex pumila are reliably close , either in space or time . At maturity, sexual reproductive dissemules of Carex pumila ; weighing up to 8mg each , possess a corky covering (utricle ) . Thus ; in winter they float on the ponds formed by the raised water table levels and are readily dispersed by wind . The direction of the prevailing west to northwest winds will tend to move dissemules in the direction of unexploited damp sand of a newly exposed or growing terminal deflation hollow. Floral development in Carex pumila appears to be under environmental control . Heading ( the emergence of the terminal male spike from the mouth of the sheath of the flag leaf) and subsequent Page 262 flowering phenomena occurred at the same time of the year in three successive years . From observations made on the sand plain in 1978 , 1979 and 1980 , heading began in the second week in October and continued into early Novenber � An atterpt to implicate photoperiod in the timing of flowering in Carex pumila by inducing floral ini tiation in seedlings less than one year old , was not successfuL Seedlings placed in long days (continuous illumination) with and without prior cold treatment ( 30 days at 5 degrees C) did not initiate floral primordia . This does not of course preclude some other combination of temperature and photoperiod from being responsible for the entrainment of this phenomenon. The changing soil moisture conditions in spring , resulting from the lowering water table levels may also be involved in the environmental entrainment of flowering phenophases in carex pumila , although it is difficult to see that the precision of , for example, heading within such a narrow �irne span in each of three years could be entrained by an environmental variable as open to the vaguaries of rainfall as water table levels � Despite this caution, the mean date of heading for the fertile shoot populations on the low dune and in the deflation hollow differed in the direction predicted on the basis of the hypothesi s that culrn elongation and, as a consequence, heading is determined by an environmental variable linked with the subsiding water table levels of spring � Soil moisture has been shown to be an important environmental variable; along with temperature and photoperiod, in the entrainment of the timing of flowering in natural communities in temperate zones (Evans 1971) . The timing of floral events is such that anthesis occurred in this wind pollinated species during the most windy period of the year-� Even at 30 to 60 an above ground level , at the height of the male spike wind speeds of up to 7rn/sec averaged over a five minute period were recorded on the sand plain . Thus , the chances of pollen transfer between fertile . shoots Page 263 and so the chances of cross pollination are greater than if flowering occurred at a different time of the year . Carex pumila must be described as essentially allogamous . I t is monoecious and possesses a conspicuous starninate spike terminally placed on the culrn, although it was found in a glasshouse experiment to be partially self-compatible (section 2 � 5) . Individual shoots showed limited protogyny although synchronous protogyny over local site populations was not demonstrated. Thus , self-poll ination may not be avoidable in the field , s ince neighbouring fertile shoots are likely to be part of the same genet of this clonal species � Further , pollen flow, despite anemophi ly , may be highly restricted (Handel 1976) . Griffiths ( 1950) found gene flow in Lolium perenne to be less than 1% of maximum at a d istance of 18 .an� The precocity of heading on the more exposed, drier low dune compared with the deflation hollow might also be attr ibutable to the poor soil fertil ity conditions at the former site� Although his data show no precocity in flowering in poor soil conditions ; Ogden ( 1974a) suggested that from casual inspection of Tussilago farfara plants in spring this might be incorrectly inferred since shoots on low fertil ity soi l possessed a greater proportion of empty receptacles than plants in more ferti le soi l . The divergence in timing o f heading of 14 days in 1979 between the low dune and the deflation hollow was not as great for subsequent floral events (emergence of stigmas , . anthesis and grain-fi lling) . Thus , maximum seed size was achieved at about the same time of the year in all fertile shoot populations (namely, early January) � The more rapid increase in seed weight over late November/ early December at the deflation hollow site compared with the low dune site was expected g iven the higher temperatures that prevailed at the former Page 264 site as a result of the reduced wind speeds experienced . This effect of shortening the duration of grain-filling in response to increased temperatures has been established in economic species including wheat ( Sofield et al . 1 974 ) . Further evidence that increased shelter hastens the onset of maturity of fertile shoots was found on the old deflation hollow in summer 1 979-80 . By early February most fertile shoots on the sheltered plots at this site were mature ( dead) and shedding seed whereas on the control plots less than half the fertile shoots were mature and fallen seed made up in insignificant proportion of the total seed output . Maximum seed output found on the sand plain was obtained in the old deflation hollow in summer 1 977-78 (namely , 1 7000 seeds or 1 30 . 5 grams DW/m2 ) . This is well in excess o f that reported by Noble ( 1 982 ) for Carex arenaria growing on sand dune soils in North Wales , although still somewhat less than that recorded for this same species in Breckland (7200 seeds or 7 . 92 grams DW/m2 and 1 96000 seeds or 21 5 . 6 grams DW/m2 respectively) . By summer 1 980-8 1 , seed output of Carex pumila in the old deflation hollow was considerably reduced , especially in .terms of dry weight per unit area ( 5028 seeds or 1 2 . 4 grams DW/m2 ) . Maximum seed output per uni t area in summer 1 980-8 1 which was found on the younger part of the low dune (8956 seeds or 20 . 5 grams DW/m2) was not affected by nitrogen-fertilizer application . Only on the more senile low dune si te where seed output was comparatively depressed did additional ni trogen increase seed output ( to 2928 seeds or 6 . 64 grams DW/m2 ) . The marked increases in production of fertile shoots and seeds in response to nitrogen-phosphorus-potassium (NPK) fertilizer reported by Noble ( 1 982 ) apply to a spring fertilizer application, administered the previous year. Seed output of Carex Page 265 pumila on the unfertilized plots on the younger part of the low dune in summer 1 980-81 was still in excess of these NPK-fertilizer-augmented seed outputs reported by Noble ( 1 982 ) for Carex arenaria . 3 . Phasic development Clonal expansion in Carex pumila results in several phases of population development being expressed on the ground at any one time and , at any one site , the resident population passing through similar phases of development over time . Similar patterns of development have been described in rhizomatous perennial species by several workers (eg Watt 1 947 ; Thomas and Dale 1 974 ; Noble et al . 1 979 ) . The pattern in Carex pumila includes a juvenile phase of rhizomatous spread on uncolonized moist sand , followed by an adolescent phase in which aerial shoot density increased . These initial phases involving expansion of the rhizome system, initially by the development of long rhizome branch modules followed by that of dwarf rhizome modules� occur to the exclusion of seed production. A mature phase follows in which w� . a proportion of the local ly high density of shoots � fert�le . The allocation of resources to seed production did not however occur to the exclusion of continuing rhizomatous growth and vegetative shoot development . In the senile phase which follows both rhizomatous and aerial growth , and seed production, were diminished to the extent that at the oldest old deflation hollow site at the end of the study period , fertile shoots of Carex pumila set no seed and none of the surviving depauperate shoots were attached to long rhizome branches . Thus , rather than showing increased density and sward mass (K-selection ; McNaughton 1 975 ) with increasing age of the population , n-shaped plots were found ( cf Watt 1 947 ) . Page 266 The question was asked , what determines this pattern of response , and in particular the balance between rhizomatous spread , aerial growth and seed production? The response of Carex pumila to increasing plant-to�plant interference during increasing seral maturity can be viewed in terms of the population closing its ranks and persisting as a seral monoculture . Such a response has been sugges ted by Ogden ( 1 974a ) as an alternative to that of escape to new areas through seed production, or fragmentation of the genet and dispersal of these uni ts . The persistence of Carex pumila as an effective monoculture depends upon the extremely low nutrient resources normally available . The rhizome architecture and growth strategy of Carex pumila resulted in the spread of the species on the low dune to cover the available ground , appropriating the limited nutrients and effectively precluding the growth of other more vigorous species . On the low dune Carex pumila was the sole or major species present over the three years that harvests were taken . Thus , despite the relatively low plant cover afforded by Carex pumila , the species may be considered to have closed its ranks to other species as effectively as it had done on the deflation hollow at S2 in 1 978 where the clone was considerably more vigorous and ground cover by this species was almost complete . Addition of nutrients to the low dune in spring 1 980 resulted in increased density of shoots attached to dwarf rhizome branches and hence increased aerial growth . It also resulted in other more vigorous species that were capable of responding rapidly to the added nutrient ( characteristic of more competitive-strategy species , Grime 1 977 ) , and that were normally precluded by the nutrient stress conditions , becoming established . Increased nutrient status of the soil by fertilizer addition therefore hastened the seral progression . Page 267 Where nutrient resources were more plentiful � as on the old deflation hollow at S2 in December 1 977 � the dense sward formed by the relatively young exclusively other Carex pumila population also excluded almost species . However , the demise of this population over time was more rapid than that on the more nutrient-stressed ; younger sites on the low dune . The occupation of a site by Carex pumila may last several years ( S4 where Carex pumila was almost totally replaced by December 1 980 , was es timated to be 1 9 years old ) � but as with all seral species , the occupation is only temporary. Species other than Carex pumila contributed an increasing proportion of sward mass at increasingly older sites � both in time and space . Carex pumila is not completely replaced by these associated species . The demise of Carex pumila at the older deflation hollow sites then sugges ts the possibility that its decomposition may release compounds that are detrimental to the germination and/or growth of other species . The production of such compounds has been conside red to form an important part of the competitive strategy of plants ( Grime 1 979) . The leaching of inhibitory substances from live or dead and decaying parts of Carex pumila may also prevent the reinvasion of these soils by new rhizomes of this species (cf decomposition of Tussilago farfara ; Ogden 1 974a) . 4. � Allocation of resources � component parts In Carex pumila allocation of resources can be thought of as being partitioned between the ecologically important ends of reproduction ( dispersal to neighbouring sites by seeds ) � competition ( exploitation of a given site by aerial shoot growth ) and clonal spread ( exploration of immediately adjacent sites by long rhizomes ) . The short rhizome segments of dwarf shoots are probably best Page 268 considered part of the allocation to competition in that they enable the more complete exploitation of a given site . The distinction between the latter two categories ( namely, competition and clonal spread) is not absolute , as once colonization of an area by long rhizomes has been achieved , the subsequent aerial and adventitious roo t development of these branch modules leads to the capture of res ources at that particular site . Further , dwarf rhizome branches may occasionally become greatly extended , giving the appearance of a long rhizome branch , albei t reduced in diamete r ( section 2 . 1 ) . The patterns of biomass allocation by Carex pumila populations showed seasonal fluctuat ions with the proportionate allocation of biomass to rhizomes reaching maxima at each site in winter, in contrast to summer aerial allocation maxima . This pattern is associated with the patterns of shoot mortality and recruitment . Major episodes of shoot mortality in Carex pumila were found in late summer, with the demise of the fertile shoot cohort , and during winter/early spring when soil conditions were anaerobic due to flooding. The main periods of shoot rec ruitment of Carex pumila , namely spring and autumn, can be related to the conditions that prevail in the surface layers of the soil which in turn are largely determined by the seasonal variation in the water table levels . In spring with the lowering of the water table , aerobic conditions returned to the soil surface which had been waterlogged (anaerobic ) for several months during winter. In autumn the rising water table brought with it water-soluble nutrients back to the surface . The release of nut rients from the senescent fertile shoot cohort in autumn, may contribute to a flush of nutrients including nitrogen into the system at this time of the year , leading to the release of bud dormancy in other parts of the clone . Bud dormancy in summer was Page 269 likely to be caused by the depletion of labile nitrogen within wore in� rhi zomes as a result of aerial shoot demands on this nutrient (Waring 1964) . The proportionate allocation of biomass to rhi zomes decreased with increasing age of the population. This was most marked on the low dune where Carex pumila was invading the bare sand of the continually growing terminal deflation hollow. At the rhizome front the entire biomass was in the form of rhizomes . Further back from the advancing front along the primary rhizome axes , an increasing proportion of total biomass was found in aerial parts . This increase was attributable to the increasing contribution of dwarf shoot modules to total biomass . The extreme in this trend was seen on the old deflation at S3 where at the end of the study all live shoots on the control plots were attached to dwarf rhizome branches . At maturity , a dwarf shoot size was no different from that of long shoots . However , the underground biomass of the whole branch module of each of these two branch types differed markedly. The most extreme divergence in the relative size of mature short and long rhizome units was found on the edge of the low dune in winter 1 980 where long rhizome se�ents measured up to 1 . 4m. At this adolescent site 74% of population biomass was found in rhizomes . The high rhizome allocation in the more juvenile compared with more mature populations suggests the importance in the strategy of Carex pumila on the embryonic sand plain soils of plant parts devoted to the capture of below-ground resources . Where these resources were more plentiful , as on the old deflation hollow at S2 , and when nitrogen fertilizer was applied , aerial allocation was significantly increased . Page 270 The trend of decreasing rhizome allocation with age when comparing sites of increasing distance (putative age ) from the terminal hollow at any one time was obscurred by the behaviour of the Carex pumila population at the sheltered site on the downwind end of the old deflation hollow at S2 . The proportionate allocation of biomass at this site favoured aerial shoots whereas the balance was in the opposite direction at all other sites , including those older ones further from the terminal hollow. This increased aerial growth at S2 was attributable to the greater shelter and soil fertility at this site compared with others . At S2 the tactics of Carex pumila are more that of a C-strategist ( Grime 1 977 ) than at other sites • . Carex pumila was able to respond to the decreased stress/disturbance conditions by producing a dense cover of tall growing shoots . However , when the bulk of these had died by the end of summer 1 978-79 , recruitment of shoots of other later seral species was not prevented . These were more rapid than Carex pumila in occupying the space made available by the mortality of Carex pumila sho ots , and by summer 1 980-8 1 had largely replaced the latter ( figure 3 . 8 ) . Since the energy values of component parts of Carex pumila were dissimilar , the allocation patterns to these components based on dry weight and on energy can be expected to differ . However , ( 1 ) the energy value of mature seeds of this species ( namely , 20200 joules/gram /DW ) , the highest of all plant parts � was low relative to other species including those in which allocation patterns based on dry weight and energy were not seen to differ significantly (Hickman and Pitelka 1 975 ; for Lupinus ) and ( 2 ) the divergence in energy values between component organs in Carex pumila was low relative to the variances in proportional allocation to components between Page 271 replicates within treatments . Thus , dry weight and energy allocation pattern differences are unlikely to be significant . Further ; Hickman and Pitelka ( 1 975 ) predicted that allocation patterns based on dry weight and energy would not differ significantly in plants in which seeds stored only a moderate amount of lipid (up to about 1 5% ) . Seed lipid content of mature seeds of Carex pumila was 3 . 6% of dry weight . The same conclusion cannot be drawn for the comparison of allocation patterns based on dry weight and total nitrogen. Total nitrogen concentrations in seeds of Carex pumila were 7x that of roots and 5x that of rhizomes . Values for green leaf laminae and sheaths were intermediate between seeds and underground parts and significantly different from both. Further differences were found between sites , averaged over all shoots , between long and dwarf shoots , and between vegetative and fertile shoots , for comparable structures . These differences probably reflec t the declining percent total nitrogen values with shoot age . Thus , proportionate allocation patterns , within and between Carex pumila populations , based on dry weight and total nitrogen were expectedly different . These differences were most often highly significant . The question now to be asked is whether nitrogen is a limiting resource in this seral habitat and , if so , whether the allocation of this resource is of more importance in the evolution of life history strategies on the sand plain than allocation of carbon (dry weight ) . The evidence is unequivocal . Nitrogen does limit the growth of the vegetation on the sand plain . However , it is only on the earlier seral stages that additional nitrogen fertilizer increased the growth of Carex pumila . On the old deflation hollow where the Carex pumila population was senile , and run-down compared with the resident Carex population at this site three years previously ; other species Page 272 responded to the fertilizer addition. Carex pumila is then less of a C-strategist ( sensu Grime 1 977 ) than these other species � principally Hypochaeris glabra . However , Carex pumila was able to respond to high nutrient levels in the soil on the deflation hollows . At S2 during 1 978 � the standing biomass and the estimated rates of biomass accumulation of this species were greater than at any other site during the course of the study . The increase in the amounts of plant tota l nitrogen in Carex pumila shoots at S2 in autumn 1 978 was more than 3x that estimated that could be supplied by nitrogen fixation in situ . This confirms the view that the vigour of Carex pumila can be attributed to the accumulation of nutrients from elsewhere on the sand plain . Sources of nitrogen include ni trogen-fixation in the rhizosphere zone of Ammophila arenaria and by Lupinus arboreus ; species that were prominent on sites both within and immediately adj acent to the study area . The amounts of plant total nitrogen in Carex pumila on the low dune were remarkable in that there was no immediately obvious source of this nutrient . Nitrogen-fixing blue-green algae ; Anabaena and Nostoc spp . , were confined to the deflation hollows and nitrogen-fixation in the rhizosphere zone of Carex pumila itself can probably be discounted . Such associations of bacteria with members of the Cyperaceae are unknown (Forde pers comm) . The differential distribution of concentrations of total nitrogen between plant parts within Carex pumila indicates an acropetal movement of this nutrient , ie the efficient channelling of this scarce resource to the continuously growing rhizome front . However ; Noble Page 273 and Marshall ( 1 983 ) contend that cut off from a supply of a limiting nutrient � continued extension of rhizomes (of Carex arenaria ) would not be possible. Thus , the efficiency of redirecting nutrients within the clone from senescing shoots to new growing points may be high in stress tolerators such as those in sand dune communities but is not absolute . Nutrients may be absorbed directly from the soil by rhizomes , as in the absorption of phosphate by buried stolons of Trifolium repens ( Hay and Dunlop 1 982 ) . In winter 1 98 1 , a new source of nitrogen in the terminal hollow was apparent with the presence of the Anabaena. Prior to this seepage of nitrogen fixed in the rhizosphere zone of Ammophila arenaria on the adjacent low dune and surrounding high dune seems the most likely source of this nutrient used by Carex pumila on the edge of the terminal hollow. 5. � Seed reproductive effort Like many plant species , the sexual reproductive structures of Carex their weight pumila remain development . of seeds is green and photosynthetic for an Not until the major period of complete does the outer covering extended period of increase in dry (utricle ) lose its green colour, turning yellow or red-brown. The photosynthetic contribution of female spikes of Carex pumila to final seed weight was estimated at 26%. This value must be considered conservative due to the apparent maintenance of final seed weight in the face of post-anthesis defoliation , and shading of all but the inflorescence ( section 2 . 6 ) . In the field , Carex pumila shoots are often buried by sand . Thus , the elaboration of seed in such instances may occur as a result of the translocation of pre-formed assimilate from other parts of the shoot , or clone , or as sugges ted in the defoliation experiment � from reproductive assimilation . photosynthesis in female spikes to Whatever the post-anthesis contribution of increase in seed Page 274 weight ( 26-1 00%) , reproductive effort estimates in Carex pumila ( based on dry weight ) cannot be considered a drain on carbon resources of the plant , in the same way that the production of progeny by animal species draws on a finite pool of resources within the individual . As concluded by Thompson and Stewart ( 1 98 1 ) � the allocation of nutrients may be of more fundem�ntal significance in the evolution of life history strategies in plants than the allocation of carbon . Seed reproductive effort as a proportion of the total living population biomass ( SRE2 ) of Carex pumila was in the range 0- 1 6% of total dry weight ; 1 0- 1 6% for mature populations that were neither juvenile nor senile . This puts Carex pumila midway along the r-K-strategy continuum between monocarpic grain crops and wild annual species , in which seed reproductive effort is characteris ticlly 20-40% of annual net assimilation, and polycarpic herbaceous perennials and forest trees ( seed RE <5%; Harper et al . 1 970 ) . These values are greater than those found by Ogden ( 1 974a ) for another seral perennial , Tussilago farfara , in which 3-8% of annual net production was found in seeds . Such estimates of reproductive effort , based on total population biomass , will differ from those based on fertile shoots alone , according to the proportion of the stand made up by non-flowering individuals . Seed reproductive effort as a proportion of the biomass of fertile shoot populations of Carex pumila varied from 0 to 3 1 . 6%. This maximum was obtained in the old deflation hollow at S3 in January 1 978 where population seed reproductive effort was also maximal . At the more fertile and sheltered old deflation hollow site , reproductive shoots had undergone c onsiderably more shoot growth (mean size was 1 . 24g/aerial shoot cf 0. 56g/shoot at S3 ) and had allocated a larger proportion of this total to vegetative parts . Total reproductive Page 275 effort ( sensu Thompson and S tewart 1 98 1 ) expressed as a proportion of biomass of fertile shoots was 36 . 5% at that site where seed effort was maximal . This proportion includes the dry weight of the terminal male spike and rhachillae and glumes of female spikes . The culm was included with leaves as vegetative. These values of seed and total reproductive effort may be compared with those presented by Abrahamson ( 1 979 ) which apply to randomly chosen flowering or frui ting plants from populations of 50 species of wildflowers from earlier (field ) and later (woodland ) secondary successional communities . The greatest reproductive effort values for Carex pumila were within the upper end of the range found by Abrahamson ( 1 979 ) for field populations of perennial species . Reproductive effort (RE) of Carex pumila in the old deflation hollow at S3 , in January 1 978 Population seed RE (after Ogden 1 974a ) Individual seed RE (after Abrahamson 1 979) 11 total RE (after Thompson and Stewart Harvest index (seeds as a proportion of aerial .. " .. 11 " Proportion of total DW 11 ( 1 98 1 ) 11 parts ) " TN 1 6% 31 . 6% 36 . 5% 42% 50% Other estimates of the " reproductive effort " of this Carex pumila population have been made to enable comparison with other studies . Seed effort as a proportion of grams dry matter and total nitrogen content of aerial shoot biomass ( equivalent to harves t index ; Donald 1 962) was 42% and 50%, respectively . Page 276 Abrahamson ( 1 980 ) arguing from the behaviour of perennial rhizomatous species in sand dune communities predicted that with increasing shoot density there will be a shift from vegetative reproduction ( sic ) to seed production enabling dispersal to new favourable sites . The increasing interference increasing density was suggested to be a between reliable shoots cue to with the decreasing favourability of the site for seral perennials . In Carex pumila such interactions between shoots as the population passes from a juvenile to an adolescent and to a mature stage of development are intraspecific and most probably intraclonal . The shift involving an increased reproductive effort with increasing density , as predicted by Abrahamson , was observed by comparison ( 1 ) of shoots on the low dune at increasing distances back from the leading edge of the clone , ( 2 ) of sites at increasing distances from the terminal hollow in January 1 978 and ( 3 ) of populations at bo th low dune sites over time . In populations of increasing senility (as on each of the deflation hollow sites over time , and between old deflation hollow si tes in 2nd , 3rd and 4th summers of the study ) reproductive effort decreased . At these sites interspecific interference of Carex pumila populations was increased . The response by this seral species was to decline . The shifts in reproductive effort of Carex pumila as populations matured and subsequently became more senile must be considered phenotypic . Genetic diffe rentiation with respect to resource allocation between Carex pumila populations across the study area can be excluded . The oldest population studied was at most 1 6- 1 8 years and although rapid population differentiation has been demonstrated in outcrossing, wind-pollinated species within short time spans ( Bradshaw 1 959 ; Antonovics 1 97 1 ; Davies and Snaydon 1 973 ) � the low rate of Page 277 turnover of Carex pumila genets on the sand plain would preclude this phenomenon , given the infrequency that seedlings of this species were observed . The mode of growth of Carex pumila indicates that one or few genets can expand to occupy the space available and � despite the rapidity of turnover of individual ramets ( few emerging shoots tagged were living 1 2 months later) , are likely to remain until the demise of the population at that site . A single genet therefore must withstand a wide range of environmental conditions both at a given site on the sand plain and across a wide area . This ability coupled with the possibility of limited genetic recombination during sexual reproduction as a result of restricted pollen flow between genets , may bring about the evolutionary conservatism of this open pollinated species . A similar conclusion was reached by McWilliam et al . ( 1 97 1 ) for one such species , Phalaris tuberosa , in Australia . The ability to alter reproductive effort ( or any of character) in response to environmental changes is nonetheless under genetic control . In Carex pumila reproductive effort was altered by habitat perturbation . Nitrogen was shown to limit seed output on the more mature site on the low dune in summer 1 980-81 , but not on the younger low dune site where Carex pumila was in a more adolescent phase of development . At the more senile site , additional nitrogen increased mean seed size and total nitrogen content per seed ( partly through increasing the levels of total nitrogen in seeds ) on fertile shoots attached to dwarf rhizome branches . The density of these fertile shoots was also increased by the nitrogen perturbation , an effect attributable to the delay in senescence of older shoots which had died on the control plots by December 1 980 . The timing of the nitrogen application ( spring 1 980) precluded any newly recruited shoots undergoing floral development and seed production by the time of harvest in December. Page 278 The main effect of increased elemental nitrogen on the components of seed output was an increase in the number of mature seed set by a fertile shoot at both low dune sites . Again because of the timing of the nitrogen application� numbers of floral primodia were unlikely to be affected by this treatment . Thus the increased seed output � through an increase in numbers of seed counted per culm was an effec t of the increased nitrogen allowing the normal development of a greater number of preformed ovules . On the control plots where. nitrogen levels were low , normal development of many ovules would have been prevented , or arrested prematurely� so accounting for any reduction in seed weight . At the younger low dune site, additional nitrogen increased the seed number pe r fertile shoo t , without affecting seed output per unit area . At both low dune sites , additional nitrogen decreased the proportion of shoots attached to long rhizome branches that became reproductive , without altering the total density of long shoots . The nitrogen fertilizer perturbation affected vegetative shoot populations of Carex pumila more greatly than fertile shoot populations three months after nitrogen fertilizer application ; ie the greater response was in the younger portion of the clone at each site . The proportionate allocation of both dry weight and total nitrogen of the population as a whole to vegetative shoo ts was increased by this perturbation. Vegetative shoot leaves per shoot and shoot density all increased . siz e , numbers of The latter can be explained as a result of the increased amounts of nitrogen in the plant releasing the dormancy of shoot buds . The proportionate allocation of total population biomass to seeds was little affected by the treatment . Ogden ( 1 974a ) also observed an unresponsiveness of total population seed reproductive effort to soil fertility Page 279 differences by a seral rhizomatous perennial of low fertility habitats , namely Tussilago farfara . By contrast to Tussilago farfara under high and low soil fertility conditions ( Ogden 1 974a ) , Carex pumila was not found to decrease the allocation of biomass to rhizomes under the higher fertility regime in al l � population5. In the senile low dune population fertilizer addition did reduce the prportionate allocation to rhizomes , while increasing that to the aerial fraction , principally through increased dwarf vegetative shoot recruitment . In the younger low dune population, despite the increased dwarf vegetative shoot recrui tment and growth brought about by the additional fertilizer , a large increase in the proportionate allocation to rhizomes occurred . It i s likely that the timing of the fertilizer application and the measurement of the response in the present study accounts for this difference . Twelve months after fertilizer application casual observation of the study area showed that aerial shoot growth on the younger low dune site had been more greatly increased by the perturbation than that found at three months . It is possible that the allocation of annual net accumulation to rhizomes at this young site was decreased by the fertilizer treatment in contrast to that finding above , only three months from the time of nitrogen application. The pattern of an increased proportionate allocation to rhizomes in poorer soils , a response that maintains the plant in an area unfavourable to taller growing species (Ogden 1 974a ) � was confirmed in the present study by comparison of the Carex pumila populations on the more fertile part of the old deflation hollow at S2 and elsewhere . At S2 where soil fertility conditions were improved cf other si tes � density and both absolute and proportionate allocation of biomass to aerial shoots were greater and the proportionate allocation to Page 280 rhizomes considerably lower than at other sites . Despite the reduction in the proportion and density of fertile shoots in the total population in response to fertilizer application, seed output per uni t area was maintained. This came about through reallocation of biomass within fertile shoot populations ; a result manifes t in the increased number of seeds per fertile culm and , on the senile low dune site where mean seed weight was reduced on the controls , inc reased mean seed weight . The total expenditure of dry weight by Carex pumila on seed production and rhizomatous growth was similar to that apportioned to seeds alone by seral annuals ( cf Turkington and Cavers 1 979 ) . This total was achieved by a larger relative expenditure on rhizomes than on seeds . Since seeds were found to store a considerably greater concentrationp of plant total nitrogen , this balance between seeds and rhizomes in mature populations was reversed when based on total nitrogen . Thus although the processes of sexual reproduction and rhizomatous growth are fundamentally different , they may be considered to be competing for the same limited resources within the plant . REFERENCES Abdel-Wahab A .M . 1 975 . Nitrogen fixation by Bacillus strains isolated from the rhiz osphere of Ammophila arenaria . Plant and Soil 42 , 703-705 . Abrahamson W . G . 1 975a. Reproduction of Rubus hispidus 1 . in different habitats . Am. Midl . Nat . 93 , 471 �478 . Abrahamson W. G . 1 975b . Reproductive strategies in dewberries . Ecology 56 , 72 1 -726 . Page 281 Abrahamson W. G . 1 979�- Patterns of resource allocation in wildflower populations of fields and woods . Amer. J. Bot . 66 � 71 �79 . Abrahamson W . G . 1 979� A comment on vegetative and seed reproduction in plants . Evolution 33 , 5 1 7-5 1 9 . Abrahamson W. G . and M . Gadgil. 1 973 . Growth form and reproductive effort in golderods ( Solidago , Compositae ) . Amer. Nat . 1 07 , 65 1 -661 . Abul-Fatih H .A . and F . A . Bazzaz . 1 979 . The biology of Ambrosa trifida 1. III Growth and biomass allocation . New Phyto l . 83 , 829-838 . Andel J . van and F . Vera . 1 977 . Reproductive allocation in Senecio sylvaticus and Chamaenerion augustifolium in relation to mineral nutrition. J. Ecol . 65 � 747-758 . Antonovics J . 1 97 1 . The effects of a heterogeneous environment in the genetics of natural populations . Amer. Sci . 5 9 � 593�599 . Archbold . 1 945 . J . agric . Sci . 77 , 453-46 1 . Page 282 Armstrong W. 1 967 . The oxidizing activity of roots in waterlogged soils . Physiol . Pl . 20 , 920-926 . Armstrong W. 1 975 . Waterlogged soils . In Environment and plant ecology. ed J .R . Etherington . Wiley and sons � London. Bazzaz F . A. and R .W . Carlson . 1 979 . Photosynthetic contribution of flowers and seeds to reproductive effort of an annual colonizer . New Phytql . 82 , 223-232 . Beadle N . c .w. 1 954 . Soil phosphate and the delimination of plant communities in eastern Australia. Ecology 35 , 370-375 . Bell A . D . and P .B . Tomlinson . 1 980 . Adaptive architecture in rhizomatous plants . Bot . J. Linn . Soc . 80 , 1 25- 1 60 . Black J .M . 1 978 . Flora of South Australia ; Part 1 . Third edition revised and edited by J . P . Jessop. Govt . Printer, South Australia . Boggie R . 1 972 . Effect of water table height on root development of Pinus contorta on deep peat in Scotland . Oikos 23 , 304-31 2 . Bostock S . J . and R .A . Benton. 1 979 . The reproductive strategies of five perennial composite . J . Ecol . 67 ; 9 1 - 1 07 . Boyce . M. S . 1 979 . Seasonality and patterns of natural selection for life histories . Amer. Nat . 1 1 4 , 569-583 . Bradbury I . K . and G . Hofstra . 1 975 . The partitioning of net energy resources in two populatons of Solidago canadensis during a single developmental cycle in southern Ontario . Can . J . Bot . 54 , 2449-56 . Bradshaw A . D . 1 959 . Population differentiation in Agrostis tenuis . I . New Phytol . 58 , 208-227 . Page 283 Bremner P.M. 1 972 . The accumulation of dry matter and nitrogen by grains in different positions of the wheat ear as influenced by shading and defoliation. Aust . J . bio l . Sci . 25 ; 657-68 1 . Briscoe P .V . � J . N . Gallagher , E . J . Little john , J .L . Monteith and R . K . Scott . 1 975 . Barley and its environment . IV sources of assimilate for the grain . J . appl . Eco l . 1 2 , 295-31 8 . Brooks R .R . 1 972 . Geobotany and biogeochemistry in mineral exploration . Harper Row , New York . 292pp . Burgess R . E . 1 974 . Intraspecific variation in Yorkshire fog . M. S c . thesis . Massey University . Calow P. 1 979 . The cost of reproduction - a physiological approach . Bio l . Rev. 54 , 23-40 . Carse H. 1 91 6 . Carex pumila var macrocarpa . Trans . N . z. Instit . 48 , 241 . Cockayne L . 1 9 1 1 . The dune areas of New Zealand ; their geo logy, botony and reclamation . Append . J . House Repres . C 1 3 , 1 -76 . Cody M . L . 1 966 . A general theory of clutch size. Evolution 20 � 1 74-84. Connor H . E . 1 979 . Breeding systems in the grasses : A survey. N . z. Jl . Bot . 1 7 , 547-74 . Cowie J . D . , P . Fitzgerald and w. Owers . 1 967 . Soils of the Manawatu­ Rangitikei sand country. N . z. Soil bureau bulletin 29 , 58p . Danell K . and K . Sjoberg . 1 979 . Decomposition of Carex and Equisetum in a northern Swedish lake : Dry weight loss and colonization by macro-invertebrates . J . Ecol . 67 , 1 9 1 -200 . Page 284 Davies J. N . , L . G. Briarty and J . O . Rieley. 1 973 . Observations on the swollen lateral roots of the Cyperaceae . New Phytol . 72 � 1 67� 1 74 . Davies M . S . and R .W . Snaydon. 1 973 . Physiological differences among populations of Anthoxanthum odoratum. J. appl . Ecol . 1 0 � 33�45 . Diamond J . M. 1 975 . Incidence functions , assembly rules and resource coupling of New Guinea bird communities . eds J .M . Diamond and M . L. Cody. Harvard Univ. Press , Cambridge . Dobzansky T . 1 950. Evolution in the tropics . Am. Sci . 38 � 209-22 1 . Donald C .M . 1 962 . In search of yield . J. Aust . Inst . agric . Sci . 28 , 1 7 1 -8. Esler A . E . 1 969. Manawatu sand plain vegetation . Proc . N. z. Ecol . Soc . 1 6 , 32-35 . Esler A . E . 1 970 . Manawatu sand dune vegetation . Proc . N. z. Ecol . Soc . 1 7 , 41 -46 . Esler A . E . 1 978. Botany of the Manawatu district of New Zealand . Govt . Printer , Wellington. 206pp . Etherington J .R. 1 975 . Environment and plant ecology. Wiley and sons � London . Evans L . T . 1 971 . Flower induction and the florigen concept . Ann. Rev . Pl . Physiol . 22 , 365-394. Evans L. T . and R . L . Dunstone . 1 970 . Some physiological aspects of evolution in wheat. Aust . J. biol . Sci. 23 , 725-741 . Evans L . T . , I . F. Wardlow and R .A . Fisher. 1 975 . Wheat . In Crop Physiology . ed L. T . Evans . Cambridge Univ . Press � London . Page 285 Faulkner J . S . 1 973 . Chromosome studies on Carex section acutae in North-West Europe . Bot . J. Linn. Soc . 65 , 27 1 -30 1 . Fischer R . A . 1 975 . Yield potential in a dwarf spring wheat and the effects of shading. C rop Sci . 1 5 � 607- 1 3 . Forde B. J . 1 966 . Translocation in grasses . II . Perennial ryegrass and couch grass . N. z. Jl. Bot . 4 , 496-5 1 4 . Gadgil R . L . 1 969 . Nitrogen in the ecosystem during dune stabilisation as Woodhill S .F . 1 . The pre-radiata phase . N . Z . Forest Service � Fores t Research Institute , Soils and Nutrition Report 3 , 1 3p . Gadgil R . L . 1 971 . The nutritional role of Lupinus arboreus in coastal sand dune fores try. 3 . Nitrogen distribution in the ecosystem bef.ore t ree planting. Plant and Soi l 35 , 1 1 3-26 . Gadgil M. and Q . T . Solbrig. 1 972 . The concept of r- and K-selection : Evidence from wild flowers and some theoretical cons iderations . Amer. Nat . 1 06 , 1 4-31 . Gaines M. S . , K . J . Vogt , J . L . Hamrick and J. Caldwell . 1 974 . Reproductive strategies and growth patterns in sunflowers ( Helianthus ) . Amer . Nat . 1 08 , 889-894 . Ganders F . R . 1 979 . The biology of heterostyly. N . z. Jl . Bot . 1 7 , 607-636 . Gibb J . G . 1 977 . Heavy minerals of west coas t , North Island � beach sands . In Soil Groups of New Zealand � Part 2 Yellow-brown sands . ed V . E . Neall . N . z. Soc . Soi l Sci . Galley F . B . 1 96 1 . Energy values of ecological material. Ecology 42 , 581 -584 . Page 286 Gorham E . , P .M. Bitousek and W. A . Reiners . 1 967 . The regulation of chemical budgets over the course of terrestrial ecosystem succession. Ann. Rev. Ecol . Syst . 1 0 , 53-84 . Grace J . 1 974 . The effect of wind on grasses . I Cuticular and stomatal transpiration. J. exp . Bot . 25 , 542-51 . Grace J . 1 977 . Plant responses to wind . Academic Press , London. Grace J . and G . Russell . 1 977 . The effect of wind on grasses . III Influence of continuous drought or wind on anatomy and water relations of Festuca arundinacea Schreb . J. exp . Bot . 28 , 268-78 . Granhall U . 1 975 . Nitrogen fixation by blue-green algae in temperate soils . In Nitrogen fixation by free-living micro­ organisms . ed W. D . P . Stewart . IBP 6. Cambridge University Greig-Smith P. 1 957 . Quantitative plant ecology. Butterworths , London . Griffiths D . J . 1 950 . The liability of seed crops of perennial ryegrass to contamination by wind-borne pollen. J. agric . Sci . 40 , 1 9-38 . Grime J . P . 1 977 . Evidence for the exis tence of three primary strategies in plants and its relevance to ecological and evolutionary theory. Amer. Nat . 1 1 1 , 1 1 69- 1 1 94 . Grime J . P . 1 979 . Plant strategies and vegetation processes . Wiley, Chichester . Page 287 Halle F . and R .A .A . Oldeman. 1 970 . Essai sur l ' Architecture et la Dunamique de C roissance des Arbres Tropicaux. Collection de Monographies de Botanique et de Biologie Vegetale � No . 6 . Paris : Masson et Cie . English translation by B. C . Stone , 1 975 . Kuala Lumpur � Malaya : University Press . Hamrick J . L . , Y .B . Linhart and J .B . Mitton . 1 979 . Relationships between life history characteristics and electrophoretically detectable genetic variation in plants. Ann . Rev. Ecol . Syst . 1 0 , 1 73-200 . Handel S . N . 1 976a . Restricted pollen flow of two woodland herbs determined by neutron-activation analysis . Nature 260 , 422-423 . Handel S . N . 1 976b . Dispersal ecology of Carex pedunculata ( Cyperaceae ) , a new North American myrmecochore . Amer . J. Bot . 63 , 1 07 1 - 1079 . Handel S . N . 1 978 . The competi tive relationship of three woodland sedges and its bearing on the evolution of antdispersal of carex pedunculata . Evolution 32 � 1 5 1 - 1 63 . Hansen G . K . and C . R . Jensen . 1 977 . Growth and maintenance respiration in whole plants , tops and roots of Lolium multiflorum. Physiol . Plant . 39 , 1 55- 1 64 . Hardy R .W. F . , R . D . Holsten , E . K . Jackson and R . C . Burns . 1 968. The acetylene-ethylene assay for N-fixation : laboratory and field evaluation . Pl . Physiol . 43 , 1 1 85- 1 207 . Harper J . L . 1 967 . A Darwinian approach to plant ecology. J . Ecol . 55 � 247-270 . Harper J . L . 1 977 . Population biology of plants . Academic Press � London. Page 288 Harper J . L . , P. H . Lovell and K . G . Moore . 1 970 . The shapes and sizes of seeds . Ann . Rev . Ecol . Syst . 1 � 327�356 . Harper J. L . and J . Ogden. 1 970 . The reproductive strategy of higher plants . I The concept of strategy with special reference to Senecio vulgaris L. J. Ecol . 58, 681 -698. Harris w. 1 970 . Genecological aspects of flowering and vegetative reproduction in Rumex acetosella 1. N. z. Jl . Bot . 8 � 99- 1 1 3 . Hawthorn W .R . and P. B . Cavers . 1 978 . Resource allocation in young plants of two perennial species of Plantago . Can. J . Bot . 56 , 2533-2537 . Heslop-Harrison J. 1 964 . Forty years of genecology . Adv . Ecol . Res . 2 , 1 59-24 7 . Hickman J . C . 1 975 . Environmental unpredictability and plastic energy allocation strategies in the annual Polygonum cascadense ( Polygonaceae ) . J . Ecol . 63 � 689-702 . Hickman J . C . and L . F . Pitelka . 1 975 . Dry weight indicates energy allocation in ecological strategy analysis of plants . Oecologia 21 , 1 1 7- 1 2 1 . Hodkinson I . D . 1 975 . Dry weight loss and chemical changes in vascular plant litter of terrestial origin, occurring in a beaver pond ecosytem. J. Eco l . 63 , 1 3 1 - 1 42 . Hollows D. G. 1 978. A physiological investigation of the adaptive significance of juvenility in Pennantia corymbosa . M. Sc . thesis . Massey University. Page 289 Horn M. S . 1 974 . The ecology of secondary succession . Ann. Rev. Eco l . Sys t . 5 , 25-37 . Huiskes A . H . L . 1 979 . The population dynamics of Ammophila arenaria . In Ecological processes in coastal environments . eds R . L . Jefferies and A . J . Davey. Blackwells , Oxford . Hutchings M. J . 1 979 . Wei ght-density relationships in ramet populations of clonal perennial herbs , with special reference to the -3/2 power law . J. Ecol . 67 , 21 -33 . Jefferies R . L . , A . J . Davey and T . Rudmik . 1 979 . The growth strategies of coas tal halophytes . In Ecological processes in coastal environments . eds R . L . Jefferies and A . J . Davey. Blackwells , Oxford . Jermy A. C . and T . G. Tutin. 1 968 . British sedges . London Bot . Soc . Br. Isles . Jones H . E . 1 97 1 . Comparative studies of plant growth and distribution in relation to waterlogging. II . An experimental study of the relationship between transpiration and the uptake of iron in Erica cinerea L. and E . tetralix L. J. Eco l . 59 � 1 67- 1 78 . Jones H .E . and J .R . Etherington . 1 970 . Comparative studies of plant growth and distribution in relation to waterlogging. I . The survival of Erica cinerea L . and E . tetralix L. and its apparent relationship to iron and manganese uptake in waterlogged soils . J . Ecol . 58 , 487-496 . Jones R . 1 972a. Comparative studies of plant growth and dis tribution in relation to waterlogging. V . The uptake of iron and manganese by dune and slack plants . J. Ecol . 60 , 1 3 1 -1 40 . Page 290 Jones R . 1 972b . C omparative studies of plant growth and distribution in relation to waterlogging. VI . The effect of manganese on the growth of dune and slack plants . J . Ecol . 60 , 1 4 1 - 1 46 . Jones R . 1 975 . Comparative studies of plant growth and distribution in relation to waterlogging. IX The uptake of potassium by dune and dune slack plants . J . Eco l . 63 � 859-866 . Kalin Arroyo M. T . , J . J . Armesto and c . Villagran . 1 98 1 . Plant phenological patterns in the high andean Cordillera of central Chile . J . Eco l . 69 , 205-223 . Kawano S . 1 975 . The productive and reproductive biology of flowering plants . II The concept of life history strategy in plants . J . Coll . Lib . Arts , Toyama Univ. 8 , 5 1 -86 . Keely J . E . 1 979 . Population differentiation along a flood frequency gradient : physiological adaptations to flooding in Nyssa sylvatica. Ecol . monog . 49 , 89- 1 08 . Kershaw K .K . 1 964 . Quantitative and dynamic ecology. Edward Arnold , London . 1 83pp . Lawrie A . C . 1 98 1 . Nitrogen-fixation by native Australian legumes . Aust . J . Bot . 29 , 1 43- 1 5 7 . Levin D .A . 1 974 . The oil content of seeds : an ecological perspec tive . Amer . Nat . 1 08 , 1 93-206 . Lewis D . 1 979 . Genetic versatility of incompatibility in plants . N . z. Jl. Bot . 1 7 , 637-44 . Lovett Doust J . 1 980a . A comparative study of life history and resource allocation in selected Umbelliferae . Biol . J . Linn . Soc . 1 3 , 1 39- 1 54 . Page 291 MacArthur R .H . and E . O . Wilson . 1 967 . The theory of island biogeography. Princeton Univ .Press , Princeton , N . J . 203pp Mcintyre G . I . 1 965 . Some effects of the ni trogen supply on the growth and development of Agropyron repens ( L . ) Beauv. Weed research 5 , 1 - 1 2 . McNamara J . and J .A . Quinn. 1 977 . Resource allocation and reproduction in population of Amphicarpium purshii (Graminae ) . Amer. J . Bot . 64 , 1 7-23 . McNaughton S . J . 1 975 . r- and K-selection in Typha . Amer. Nat . 1 09 , 25 1 -26 1 . McRae I . e . and T . F . Castro 1 967 . Nitrogen fixation in some tropical rice soils . Soil Sci . 1 03 , 277-280 . McWilliam J .R . , H . E . Sc�roede r , D . R . Marshall and R . N . Oram. 1 97 1 . Genetic stability of Australian phalaris (Phalaris tuberosa ) under domestication. Aust . J . agric . Res . 22 , 895-908 . Marshall J . K . 1 965 . Corynephorus canescens (L . ) P. Beauv . as a model for the Ammophila problem. J . Ecol . 53 , 447-463 . Mason G . F . and R . J . Bryant . 1 975 . Production , nutrient contents and decomposition of Phragmites communis Trin. and Typha augustifolia L . J . Eco l . 63 , 71 -95 . Moore L . B . and E . Edgar . 1 970 . Flora of New Zealand . Vol . 2 . Govt . Printer , Wellington . 354pp . Mulder E . G . , T .A . Lie and J .W . Woldendorp . 1 969 . Biology and soi l fertility . In Soil Biology , 1 63-208 , UNESCO , Paris . Page 292 Newell S . J . and E . J . Tramer . 1 978 . Reproductive strategies in herbaceous plant communities during succession. Ecology 59 ; 228-234 . Noble J . C . 1 982 . Biological flora of the British Isles : Carex arenaria L . J . Eco l . 70 , 867-886 . Noble J . C . , A . D . Bell and J . L . Harper . 1 979 . The population biology of plants with clonal growth . J . Ecol . 67 , 983- 1 008 . Noble J . C . and C . Marshall . 1 983 . The population biology of plants with clonal growth . II The nutrient strategy and modular phjs tolo�':1 phyaiotheFapy of Carex arenaria. J . Ecol . (in press ) . Ogden J . 1 974a. The reproductive strategy of higher plants : II The reproductive strategy of Tussilago farfara L. J . Ecol . 62 , 291 . 324 . Ogden J . 1 974b . Observations on two coastal ecotypes of Selliera radicans Cav. (Goodeniaceae ) growing in the Manawatu District of New Zealand . N. z. Jl . Bot . 1 2 , 541 -50 . Olson J . S . 1 963 . Energy storage and the balance of producers and decomposers in ecological systems . Ecology 44 , 322-33 1 . Ornduff R . 1 966 . The origin of dioecism from heterostyly in Nymphoides (Menyanthaceae ) . Evolution 20 , 309 . 1 4 . Parry G . D . 1 98 1 . The meanings of r- and k-selection . Oecologia 48 , 260-264 . Patrick W. H . and M. E . Tusneen . 1 972 . Nitrogen loss from flooded soil . Ecology 53 , 735-737 . Page 293 Pearsall W . H . and E. Gorham. 1 956 . Production ecology. I. Standing crops of natural vegetation. Oikos 7 , 1 93-201 . Pegg E . J . 1 9 1 4 . An ecological study of some New Zealand sand dune plants . Trans . N . z. Ins tit . 46 � 1 50� 1 77 . Pickett S . T . A . 1 976 . Succession : An evolutionary interpretation. Amer . Nat . 1 1 0 , 1 07- 1 1 9 . Primack R .B . 1 979 . Reproductive effort in annual and perennial species of Plantago (Plantaginaceae ) . Amer . Nat . 1 1 4 , 5 1 -62 . Ranwell D . S . 1 959 . Newborough Warren , Anglesey. I . The dune system and dune slack habi tat . J . Ecol . 47 , 57 1 -601 . Ranwell D . S . 1 97 2 . Ecology of salt marshes and sand dunes . Chapman and Hall , London . Raunkaier C . 1 934 . The life form of plants and plant geography . Oxford . Raynal D . J . 1 979 . Population ecology of Hieracium florentinum ( Compositae ) in a central New York limestone quarry . J . appl. E col . 1 6 , 287-298 . Robertson N . G . 1 959 . The climate of New Zealand . In A descriptive atlas of New Zealand . Govt . Printer , Wellignton. 1 09pp . Roos F. H . and J . A . Quinn . 1 977 . Phenology and reproductive allocation in Andropogon scoparius ( Graminae ) populations in communities of different successional stages . Amer. J. Bot . 64 , 535-540 . Russel E .W. 1 96 1 . Soil conditions and plant growth . 9th Edn . Longmans , London. Salisbury E . J . 1 942 . The reproductive capacity of plants . Bell and Sons � London. 244pp Salisbury E . J . 1 952 . Downs and dunes . Bell , London . Schaffer W .M . 1 974 . Selection for optimal life histories : the effects of age structure . Ecology 55 , 29 1 -303 . Scott A . D . and D .D . Evans . 1 955 . Dissolved oxygen in saturated soil . Soil Sci . Soc . Amer. Proc . 1 9 , 7- 1 6 . Scott W .R . , C . T . Dougherty and R . H .M . Langer . 1 977 . Development and yeild components of high yielding wheat crops . N. z . J . agric . Res . 20 , 205- 1 2 . Singh J . S . , W. K. Laurenroth and R . K . Steinhorst . 1 975 . Review and assessment of various techniques for estimating net aerial primary production in grassland from harves t data . Bot . Review 41 , 1 8 1 -232 . Smith R . G . 1 977 . Fertilizer recommendatins in the Manawatu sand country . In soil groups of New Zealand , Part 2 . Yellow-brown sands � ed V . E . Neall . N . z. Soc . Soil Science . Snaydon R .W . and M. S . Davies . 1 976 . Rapid population differentiation in a mosaic environment . IV Populations of Anthoxanthum odoratum at sharp boundaries . Heredity 37 , 9-25 . Snell T .W. and D .G . Burch. 1 975 . The effects of density on resource partitioning in Chamaesyce hirta (Euphorbiaceae ) . Ecology 56 , 742-746 . Sofield I . , L . T . Evans and I . F . Wardlow. 1 974 . The effects of temperature and light on grain filling in wheat . In Mechanisms of regulation of plant growth . eds R .L . Bieleski , A .R . Fergusson and M.M . Cresswell. R. Soc . N . z. Bull . 1 2 . Page 295 Sokal R . R . and F . J . Rohlf. 1 969 . Biometry. Freeman � San Francisco . 776pp . Solbrig O . T . 1 980 . Genetic structure of plant populations . In Demography and dynamics of plant populations . ed O . T . Solbrig. Bot . Monographs Vol . 1 5 . Blackwells , Oxford . 222pp . Stearns S . C . 1 976 . Life his tory tactics : a review of the ideas . Quart . Rev. Biol . 5 1 , 3-47 . Stevens P . R . and T .W . Walker . 1 970 . The chronosequence concept and soil formation. Quart . Rev. Biol . 45 , 333-350 . Stewart W. D .P . 1 967 . Transfer of biologically fixed nitrogen in a sand dune slack region . Nature 21 4 , 603-604 . Thompson K . and A. J . A . Stewart . 1 98 1 . The measurement and meaning of reproductive effort in plants . Amer . Nat . 1 1 7 , 205-21 1 . Thorne G . N . 1 966 . Physiological aspects of grain yield in cereals . In the growth of cereals and grasses . eds F . L . Milthorpe and I . D . Ivins . Butterworths , London. Tietema T. 1 979 . The ecophysiology of sand sedge ( Carex arenaria ) : the response to local mineral sources . In Ecological processes in coastal environments . eds R . L . Jefferies and A . J . Davey. Blackwells , Oxford . Tripathi R . S . and J . L . Harper. 1 973 . The comparative biology of Agropyron repens ( L . ) Beauv . and A caninum ( L. ) Beauv . 1 . The growth of mixed populations established from tillers and from seeds . J . Ecol . 61 , 353-368 . Trivedi s . and R . S . Tripathi . 1 982 . Growth and reproductive strategies of two annual weeds as affected by soil nitrogen and density levels . New Phytol . 91 � 489�500. Page 296 Turesson G. 1 922 . The genotypical response of the plant species to the habitat. Hereditas 3 , 2 1 1 -350. Turesson G. 1 925 . The plant species in relation to habitat and climate . Contribution to the knowledge of genecological units . Hereditas 6 , 1 47-236 . Turkington R .A . and P .B . Cavers . 1 978 . Reproductive strategies and growth patterns in four legumes . Can. J. Bot . 56, 41 3-41 6 . Turner F. T. and W. H . Patrick 1 968 . Chemical changes in waterlogged soils as a result of oxygen depletion. ·Trans . 9th Int . Congr. Soil Sci . 4 , 53-6 5 . Waldren R . P . and R . D . Flowerday. 1 979 . Growth stages and dist­ ribution of dry matter� nitrogen, phosphorus and potassium in winter wheat . Agron. J. 71 � 39 1 . Wareing P .F . 1 964 . The developmental physiology of rhizomatous and creeping plants . Seventh British Weed Control Conferences 1 , 1 020- 1 030 . Watkinson A . R . , A. H .L . Huiskes and J . C . Noble . 1 979. The demography of sand dune species with contrasting life cycles . In Ecological processes in coastal environments . eds R . L . Jefferies and A . J . Davey. Blackwells , Oxford . Watt A. S . 1 947 . Pattern and process in the plant community. J. Ecol. 35 , 1 -22 Page 297 Waughman G . J . 1 97 1 . Field use of the acetylene reduction assay for nitrogen fixation. Oikos 22 , 1 1 1 - 1 1 3 . Waughman G . J . 1 972 . Acetylene reduction assay for nitrogen fixation in sand dunes . Oikos 23 , 206-21 2 . Waughman G . J . 1 977 . The effect o f temperature on nitrogenase activity . J . exp . Bot . 28 , 949-60 . White J . 1 980 . Demographic factors in populations of plants . In Demography and evo lution in plant populations . ed O . T . Solbrig. Bot . Monographs vo1 . 1 5 . Blackwells ; Oxford . 222pp . Wilbur H .M . 1 976 . Life history evolution in seven milkweeds of the genus Asclepias . J. Ecol . 64 , 223-240 . Wilbur H .M . 1 977 . Propagule size , number and dispersion pattern in Ambystoma and Asclepias . Amer. Nat . 1 1 1 , 43-68 . Wilbur H .M . , D .W . Tinkle and J. P . Collins . 1 974 . Environmental certainty, trophic level and resource availability in life history evolution . Amer . Nat . 1 08 , 805-8 1 7 . Willis A . J . 1 963 . Braunton Burrows : the effects on the vegetation of the addition of mineral nutrients to the dune soils . J . Ecol . 5 1 , 353-374 . Willis A . J . and E .W . Yemm. 1 96 1 . Braunton Burrows : mineral nutrient status of the dune soils. J . Ecol . 49 , 377-390 . Wilson J .B . 1 979 . Teddybear . Technical report T5 , Edition 2 . 5 . University o f Otago , Dunedin .