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. POPULATION DYNAMICS OF THE SEEDFLY, PEGOHYLEMYIA JACOBAEAE (HARDY) (DIPTERA: ANTHOMYIIDAE) f AND ITS POTENTIAL AS A BIOLOGICAL CONTROL AGENT OF RAGWORT, SENOCIO JACOBAEA L. A thesis submitted for the degree of Doctor of Philosophy at Massey University Jennifer Jane Dymock, May, 1985 M A1lllffillilli mlill YII Iffflif Y 1061938450 P la t e 1 Ragwort s e edfly adult i . ABSTRACT The aim of this study was to assess the effectiveness of a seedfly, Pegohylemyia jacobaeae (Diptera: Anthomyiidae) as a biological control agent of ragwort, Senecio jacobaea. Seedfly populations were sampled from October to April in 1982/83 and 1983/84 at two sites in the central North Islarrl, New Zealarrl. Supplementary laboratory expenments were corrlucted to clarify aspects of the seedfly/ragwort interaction. In the field a six week pre-oviposition period was recorded before ragwort flowered. The absence of ragwort flowers had no effect on ovary development in laboratory reared flies but the pre-oviposition period predicted by a day degree summation model was shorter than that observed in the field. The exterrled pre-oviposition period resulted in competition for oviposition sites when ragwort first flowered, with consequent low fecundity and multiple oviposition. This competition was higher in the secorrl year due to a tenfold increase in the number of flies emerging compared wi th the prev ious year. The total nunber of ragwort seedheads was similar in both years but inflorescence development was faster in 19 83/84. A combination of increased seedfly population, increased competition for oviposition sites and improved synchrony between II. emergence and ragwort flowering in 1983/84 resulted in a twofold increase in infestation levels compared to 1982/83. Estimates of seedfly mortality within the seedhead were 33% and 60% in 1982/83 arrl 1983/84 respectively but were affected by the variable length of the third larval instar. Laboratory and field data indicated that pupating third instar larvae leave the seed head in corrlitions of high surface moisture and prolonged dry conditions resulted in high mortality. Pupal mortality ranged fram 1403 to 57% but this was also affected by estimates of third instar larvae. Pupal diapause is initiated at temperatures above lS-20C and in the central North Island seedfly rarely diapause in winter® Seedfly infestation levels ranged fran 10% (1982/83) to 20% (1983/84) and up to 200 seedheads/m2 escaped predation. Ragwort seeds are long-lived,have a high genninating capacity, and uneaten seeds from seedfly infested seedheads were viable. It was concluded that at the study sites the seedfly's linpact on the ragwort population was negligible. ---------- iii ACKNOWLEJ::X;EMENTS I would like to thank my supervisors , Professor Brian springett , Dr Ian Stringer and Ms Pauline Syrett for their direction and constant advice throughout this study and their help in the drafting of the thesis . I am also indebted to Dr Nigel Barlow for his valuable advice on the seedfly model and Dr Oliver Sutherland for much useful discussion. Throughout the study I was supported by a Leonard Condell Farming Scholarship and funds for equipment, travel and thesis costs were provided by the Entomology Division of the D .S.I . R . The provision of the field site and accamodation at Tokoroa by New Zealand Forest Products Limited was much appreciated as was the cooperation of their Investigations Forester , Mr Barry Poole. I am grateful for the use of the controlled environment rooms at Plant Physiology Division , D .S .l.R . and the Agriculture and Horticulture Department , Massey University and thank the staff of the respective departments for their help during the experiments . Clare Hannan , Denise Harding , Mike Moffat and especially Alison Campbell also provided t imely assistance in potting ragwort . I would also like to thank S imon King, who welded the emergence traps; Barry Campbell, Liz Halligan, Hugh Neilsen and Dr Al Rowlands, who took photographs; Heidi Lowry, who typed the tables and graphs; Vaughan Hunt and Anne Meredith for help with computer graphics and printing of the final copy of the thesis . I am also grateful for the f riendship and interest shown by the senior students and staff of the Botany and Zoology Department, Massey University. Finally, special thanks are due to my family for their field assistance and encouraganent throughout this project. Table of Contents Abstract Acknowledgements List of Plates List of Figures List of Tables List of Appeooices CHAPTER ONE : I NTRODUCT I ON 1.1 INTRODUCTION 1. 2 THE HOST PLANT 1.2.1 1.2.2 1.2.3 1.3 1. 3.1 1.3.2 1.3.3 1.4 1.4.1 1.4.2 1.4.3 Distribution Biology and life history Toxicity aoo weed status RAGWORT CONTROL MEASURES Herbicide treatment Mechanical removal Insects as biological control agents of ragwort RAGWORT SEEDFLY Description Liberation and establishment in New Zealaoo Life history 1.4.4 Effectiveness as a biological control agent of ragwort 1.5 Objectives of the present study iv. Page i iii vii viii x xi 1 2 2 7 8 9 10 12 12 14 16 17 CHAPTER TWO : FIELD STUDIES 2.1 I NTRODUcr ION 2.2 THE FIELD SITES 2.2.1 The Redwoods Forest site 2.2.2 The Desert Rd s ite 2.3 METHODS AND EQUIPMENT 2.3.1 Sampling frequency 2.3.2 Adults 2.3.3 Eggs and larvae 2.3.4 Pupae 2.3.5 Ragwort populations 2.3.6 Ragwort seed germination 2.4 RESULTS AND DISCUSSION FROM THE REDWOODS FOREST SITE 2.4.1 Emergence 2.4.2 Sticky trap results 2.4.3 Seedfly infestation levels of ragwort 2.4.4 Multiple infestation of ragwort seedheads 2.4.5 Distribution of seedfly eggs 2.4.6 Pupae 2.4.7 Germination 2.5 RESULTS FROM THE DESERT ROAD SITE CHAPTER THREE : LABORATORY STUDIES 3.1 I NT RODUcr ION 3.2 METHODS AND EQUIPMENT 3.2.1 OVary development I Experiments at D.S.I.R. II Experiments at Massey University v. 19 19 22 23 27 28 29 32 32 33 34 37 39 45 46 53 53 57 57 58 60 3.2.2 Egg and larval development 3.2.3 The larval d ropping response 3.2.4 D iapause 3.3 RESULTS AND D ISCUSSION v i . 61 64 67 3.3.1 Ovary development 67 3.3.2 Development of eggs and larvae w ith in the seed head 70 3.3.3 Larval dropping 8 0 3.3.4 Pupal surv ivorsh ip and terminat ion of diapause 81 CHAPTER FOUR : MODELL ING THE SEEDFLY POPULAT ION 4.1 INTRODUCT ION 8 6 4.2 THE PRE-OV IPOSIT ION PERIOD 8 6 4.3 SEEDFLY OV IPOSIT ION 88 CHAPTER F IVE : D I SCUSSION 5.1 INTRODUCTION 94 5.2 RAGWORT CONTROL AT REDWOODS FOREST 95 5.3 RAGWORT SEEDFLY AND THE B IOLCX:;ICAL CONTROL OF RAGWORT 103 5.4 THEOR IES AND PRACT ICE OF B IOLOGICAL CONTROL OF WEEDS 107 5.5 CONCLUSIONS 114 SUMMARY 116 REFERENCES APPEND IX 1 APPEND IX 2 APPEND IX 3 119 127 129 130 L ist of Plates plate 1 Ragwort seedfly adult * 2 Redwoods Forest f ield s ite 3 Desert Road f ield s ite 4 St icky trap at Redwoods Forest vii. 5 Arrangement at Redwoods Forest for collect ing th i rd instar larvae 6 Cages for rear ing adults in the controlled env ironment roams at D. S.l.R. ** 7 Collect ing funnel for th i rd instar larvae leav ing ragwort in the controlled env i ronment rooms at D.S. I.R. ** 8 Reproduct ive organs fram a 14 day old female seedfly *** 9 Seedily spermatheca squashed to show sperm **** Photos taken by : * Barry Campbell ** *** **** L iz Hall igan Hugh Ne ilson Dr.R.E. Rowland vi i i. List of Figures Figure Page 1.1 See:lfly distribution in 1984 15 2.1 Location of the field sites 20 2.2 Area map REdwoexjs Forest field site 24 2.3 Map of Redwoods Forest si te 25 2.4 A rea map of Desert Rd site 26 2.5 Emergence trap 30 2.6 Number of seedfly adults caught in emergence traps at Redwoods Forest a) 1982/83 35 b) 1983/84 35 2.7 Total numbers of flies emerging at Redwoexjs Forest 36 2.8 Number of seedfly adults caught on sticky traps at Redwoods Forest 36 2.9 Calculated numbers of eggs and larvae in ragwort seedheads per hectare at Redwoods Forest a} colour graph - 1982/83 41 b) overlay - 1983/84 2.10 Calculated numbers of ragwort buds and capitulae per hectare at Redwoods Forest a} 1982/83 42 b) overlay - 1983/84 2.11 Cunulative percentage curves of seedfly and ragwort see:lhead developmental stages at Redwoexjs Forest a) 1982/83 43 b) overlay - 1983/84 2.12 Percentage of ragwort seedheads infested by see:lfly at Redwoods Forest 44 2.13 Incidence of multiple infestations by seedfly at Redwoexjs Forest a) 1982/83 50 b) 1983/84 50 2.14 Changes in the incidence of mult iple seedfly infestations w ith t ime 51 2.15 Estimates of the number o f seedfly pupae and third instar larvae per hectare at Redwoods Forest in 1983/84 51 2.16 The number of seedfly third instar larvae and rainfall at Redwoods Forest i x. a) third instar larvae within seedheads in 1982/83 52 b) third instar larvae within seedheads and leaving the plant in 1983/84 52 2.17 CUmulative percentage curves of seedfly and ragwort seedhead development stages at the Desert Rd a) 1982/83 55 b) overlay - 1983/84 2.18 Percentage of ragwort seedheads infested by seedfly at the Desert Rd site 56 3.1 Temperature fluctuations in relation to photoperiod in the controlled environment roam at Plant Physiology Division D.S . I . R., Pa�erston North 65 3.2 Emergence of seedfly adults in the laboratory 71 3.3 The relationship between ovary size, age diet of seedfly females 72 3.4 The rate of seedfly ovary development in relation to temperature 73 3.5 Duration of egg and larval stages of ragwort seedfly at different temperatures 77 3.6 The rate of seedfly development within the seedhead in relation to temperature 78 3.7 The rate of seedfly development fram oviposition to pupation in relation to temperature 78 3.8 Survivorship of seedfly at Redwoods Forest 79 3.9 Proportion of third instar larvae dropping at different humidities in the controlled temperature rooms at D.S.I.R. 3.10 The dropping response of seedfly third instar 83 larvae in relation to watering seedheads 84 3.11 Relationship between pupal development at 20C and previous exposure to cold temperatures 4.1 a) prev ious exposure to 5C 85 b) previous exposure to -lSC 85 Factors affecting seed fly infestation of ragwort seedheads at Redwooos Forest 87 4.2 Cumulative percentage curves of seedfly and ragwort seed head development stages at Redwoods Forest a) 1982/83 89 b) 1983/84 90 x. List of Tables Table page 2.1 Area sampled with emergence traps 31 2.2 Frequency distribution of ragwort sterns per plant at Redwoods Forest 31 2.3 Determination of seedfly larval instars from measurements of mouthpart lengths 40 2.4 Characteristics of seedheads containing seedfly eggs 0-24 hours old 40 2.5 Number of ragwort stems at Redwoods Forest 40 2.6 Number of eggs in female seedflies collected at Redwoods Forest 47 2.7 Relationship between the number of buds containing seedfly per stern and multiple infestations 47 2.8 Distribution of seedfly eggs in ragwort seedheads 48 2.9 Distribution of seedfly eggs in relation to presence of Nezara viridula 48 2.10 Estlinates of fly density from soil sampling and emergence traps 49 2.11 Number of ragwort seeds and their viability after seedfly attack 54 3.1 Allocation of seedfly eggs to the controlled environment rooms at D.S.l.R. 66 3.2 Total number of seedfly eggs from females reared in the laboratory 74 3.3 Mated state of seedfly females reared in laboratory conditions at D.S.l.R. 75 4.1 Ragwort buds per ovipositing female at the Redwoods Forest field site 91 4.2 Development tlines for ragwort seedfly reared in the controlled environment rooms at plant Physiology Division , D.S.I.R., Pabnerston North 91 5.1 Effectiveness of seed predators as biological control agents of weeds (From Julien (1982 ) ) 108 Appendix 1 2 3 L ist of Appendices D ist2 ibut ion map of ragwort rosettes in two 300m subplots of the Poles plot a) 1982/83 b) 1983/84 Fortn ightly mean da ily max imum and m in imum temperatures at the Redwoods Forest and Desert Rd f ield s ites dur ing the sampl ing per iod Number of th ird instar seedfly larvae dropp ing from cut sterns in the laboratory xi. Page 127 128 129 130 1. CHAPTER ONE: INTRODUCTION 1.1 Intrcduction Ragwort, Senecio jacobaea L., is a canmon wee::l throughout New Zealand and is a problem in waste areas, along roadsides, railway lines, on derelict agricultural land, forest margins and on over-grazed, poorly managed fastures. It was classified as noxious plant soon after its discovery in New Zealand because of its toxicity to stock. A prolific seeding plant it also reprcduces vegetatively in response to mechanical damage, herbicide treatment or environmental stress. A biological control programme for ragwort was initiated in the 1920s and 1930s with the introduction of two seed fly (Pegohylanyia) species and the cinnabar moth (TYria j acobaeaeL.) . HOY (1964), however, deferred further research on this when he stated that "the advent of hormone sprays offers a simple alternative to biocontrol of ragwort and no further efforts in this field are at present contemplated." In 1982, the D.S. I.R. again considered the intrcduction of new biological control agents of ragwort and redistribution of existing ones. The importance of - - -- - ------------- - - -- ��- 2. obta in ing bas ic informat ion on the effect iveness and d istr ibut ion of present populat ions was also real ised. Th is study was therefore in it iated to determ ine how the interact ion between R. jacobaeae and its host affected the seed fly' s potent ial as a control agent of ragwort. The remainder of th is chapter descr ibes the l ife h istory of ragwort and the var iety of control measures currently in use in New Zealand. The two ragwort seedfly species are descr ibed and publ ished informat ion on the ir l ife cycles and effect iveness as b iolog ical control agents of ragwort is p resented • 1.2 The host plant 1.2.1 D istr ibut ion Ragwort, Senec io j acobaea L. is a mEmber of the largest genus of the fam ily canpositae. MEmbers of th is genus range fran herbs to shrubs and are found th roughout the world. Few are of econon ic importance but a nunber are �s. Ragwort was f i rst noted in New Zealand near Dunedin in 1874 (Thomson, 1922). A nat ive dune plant of Europe and As ia, it was p robably introduced as a contam inant of crop seed. It spread qu ickly throughout !'Jew Zealand and is abundant on l ight d isturbed calcareous so ils, organ ic r ich alluvium, l ight loarns and clays. It can be found in most areas receiving more than 800 rom ra in per annum (Poole and Cairns, 1940). 1 .2.2 Biology and life history Although ragwort is canmonly considered a biennial which d ies 3. after flowering it frequently behaves as an annual or a perennial. Plants can be either single-stemmed or much branched from the base. The stans are 0'5-2 metres high, branching above the middle to give a flat topped dense compound cor� with as many as 2 500 capitulae (Poole and Cairns, 1940) . A capitulum contains on average 70 achenes (cameron, 1935), each producing a single seoo. When in bud the involucral bracts are folded over the flowers. Each bract is dark tipped and the massed tips make a dark spot at the centre of the flat topPEd bud. The stignas anerge before the anthers. The flowers have a faint odour and are visi too by a large number of insects. Nectar is present and the honey bee, � mellifera L., collects this but not the pollen (Harper and WOod, 1957). Pollen presentation occurs fram Bam to 5pm with peak periods at lOam and noon for ray and disc florets respectively (Harper and WOOd , 1957). In a single capitulum anthers continue to dehisce over a period of 4-9 days. Fertilisation may also occur on cut flowers that are still fresh (Poole and Cairns, 1940) • Attached to the top of the disc achenes is a pappus approxlinately 5mm in length. Its effectiveness is increased in conditions of high wind velocity and low relative humidity (cameron, 1935) . The general consensus is, however, that the pappus does little to aid wind d ispersal and the majority of the seeds fall near the plant (Sheldon and Burrows, 1973 and Smallfield, 1970). Poole and Cairns (1940) found that the seed shadow was of a negative exponential form and they estimated that the highest amount of 4 . seed to becane wirrlborne was 0 . 5%. No viable seej was voided by sparrows am Zebra finches when fro ragwort seeds ( Poole and Cairns , 1940 ) but achenes eaten by sheep pass through the d igestive systen umamaged am geminate in dung (Eadie and Robinson , 1953 ) . Animal s frequently have seed tangled in hair or wool am ragwort dispersal can often be traced to stock movenents . Infestations also follow water courses . Poole ( 1938b) found that ragwort seros read ily geminatro in water . Although 't::he seros in itially sank when the per icarp spl it they rose to the surface when the cotyledons fully opened . Ragwort seeds have a high germinating capacity under suitable cond itions . Cameron ( 1935) reported 80% v iabil ity of seros geminated on damp filter paper at 15C . Poole and Cairns (1940 ) working with New Zealarrl mater ial found gemination in the laboratory var ied fram 50-86% am Kelsey (1955) records an average germination of 70% ( range 58-96% ) • High tenperature , low so il moisture and low relative air humid ity reduced germination success in exper iments conducted by Van der Meijden and Van der Waals-Kooi ( 1979) . They foum that achenes possess innate dormancy but dormancy may be induced by adverse cond itions such as frost and drought . Green ( 1937) observro that after the bracts of the capi tulum have rolled back to allow d i spersal of the d i sc achenes they recurve upwards to fom a cup retaining the ray achenes . This can be linked to the d imorphism of achenes descr ibed by McEvoy (1984 ) . He found that d isc and ray 5. florets yield achenes of different dormancies: disc achenes germinate quickly with high germination success; ray achenes exhibi t low percentage and delayed germination. This, he suggests, sprecrls germination in space arrl time as those seeds carried away may take crlvantage of chance exogenous disturbances by germinating quickly. These risks are balanced by those seeds which germinate at the comparably safer rosette openings and also explains how ragwort populations persist in the absence of environmental disturbances. Light is required for germination and a soil covering greater than 4rnm in thickness results in enforced dormancy (van der Meijden and van der Waals-Kooi, 1979) . Absence of soil covering retards germination for a time but Poole arrl Cairns (1940) fourrl that seed sown on the soil surface gave the same ultimate germination as seErl sown urXier a very thin sprinkling of soil. Long term seed viability trials are currently underway at Ruakura Agriculture Research Centre arrl preliminary results show that viability increases with the depth that the seeds have been buried (Alex Thompson pers.camm, 1984) . From initial results Thompson and Makepeace (1983) estimate seed viability to decline to 1% in 4-5 years within 0-2 em of the surface layer and 10-16 years below 4crn soil depth. In the field germination success is also affected by availability of germination sites. Cameron (1935) showed that 2% of seed germinated when sown on cut grass compared to 53% on bare soil. McEvoy ( 1984) recordErl higher seedling densities in rosette 6. openirgs than in the vegetation irnnediately surrourrlirg the opening. He attr ibutes this either to the establishment probability beirg possibly higher inside the rosettes perimeter and/or to differences in sowing density. Maximun germination occurs in autunn followirg the main seEdfall (Poole, 1938a). The first true leaf appears about a month after germination but only a few leaves develop in the winter months. By mid Decenber suall rosettes have formed. These overwinter as low ve;Jetati ve rosettes, 5-15cm in diameter, with broad, horizontally orientated leaves. These generally bol t and flower during the early surrmer of the secorrl year of growth. The probability that a plant will flower increases with the diameter of the rosette at the start of the flowerirg season (van der Meijden arrl van der Waals-Kooi, 1979) • There are many known variations of the biennial life cycle described above. When Schnidl (1972) grew .§..jacobaea plants fran seed and planted them in ungrazed grassland in Victoria, Australia, 2% of those that flowered were annuals, 52% biennials and 46% perennials. Forbes (1977) constructed a model of population flux in a hypothetical population in which germination, maturation and death were constant fran year to year. The model predicted that of the plants that flowered 8% were annuals, 39% biennials and 53% perennials. Re;Jeneration occurs from root buds, from buds in the axils just above ground level or from the crown (Radcliffe, 1969). plants may regenerate within two months from root fragments less than 2an 7. in length but r oots of r osettes fonn buds more readily ( 37% of r oot fragments) than those of flo�ring plants ( lO%) (Poole and Cairns, 1940). Cairns ( 1938) attributes the high regenerative capacity of ragwort roots to a non-functional endodermis and development of pericyclic phellogen which results in rapid formation of new tissues. Perennial growth can be inducro by other forms of stress but this is not readily detectable as se€rllings are easily confused with rosettes that have arisen vegetatively. The latter are difficult to identify because the connection with the parent plant soon decays. Thompson ( 1977) concludro that infestations which persisted after herbicide treatment were composed largely of regrowth plants. Islam and Crawley ( 1983) found that over 75% of plants that had flowered produced shoots in the following spr ing but 42% of these had suffered sever e defoliation by cinnabar moth. Delayed 9rowth in rosettes due to drought in the secorrl sumner prevented flo�ring in a population of ragwort observed by Poole arrl Cairns ( 1940). Radcliffe ( l969) states that r osettes may persist for many years before flo�ring when strongly repressed by competition and both Cairns ( 1938) and Radcliffe ( 1969) warn that plants dying down after flowering may produce vegetative shoots fram the ir roots. 1.2.3 Toxicity and weed status. Ragwort contains pyrrolizidine alkaloids responsible for chronic liver damage in cattle and horses. Dickonson and King (1978) noted the highest alkaloid content in flowers and roots while 8. White (1969) reported 002-003% alkaloid content in above ground parts and 0 -03% in the roots of New Zealand plants. Toxicity may also increase up to flowering time (Connor, 1977). Rabbits do not eat ragwort, (Harper and Wbod, 1957), but sheep and goats are often use::l to control the wee::l. Sheep were observe::l by poole and cairns (1940) to show a preference for the plant after they had acquire::l a taste for it. Many of the ill effects suffere::l subsequently by sheep are not always attributed to alkaloid poisoning which has similar symptoms to facial eczema (Mortimer and White, 1975). The toxic alkaloids are not lost on drying and deaths of cattle and horses may result after eating hay or silage containing ragwort (DEmpster, 1982) _ Honey from ragwort infested areas 15 reported to be dark coloured and tainted (Thomson, 1922) and can contain alkaloids (Deinzer et al., 1977). Milk from cows and goats fed ragwort also containe::l alkaloids (Dickonson and King, 1978). The effects on humans are not known. By 1900 ragwort had been placed on the schedule of noxious wee::ls in New Zealand (Radcliffe, 1969) primarily on the basis of its toxicity but its conspicuousness may also contribute to its weed status. Economic damage has never been calculated directly in terms of lost pasture and animal production but by costs of control measures (Swarbrick, 19B3). 1.3 Ragwort control measures 1.3.1 Herbicide treatment Application of herbicides seldom, if ever, gives permanent control of established ragwort infestations. 2,4-0 is the most widely �----�--���----� 9. used spray in New Zealand for its control and although is not as effective as dicamba or picloram it is less damaging to pasture (Thompson, 1974, 1977) . Ragwort palatibility can be increased by herbicide application resulting in further stock losses (cameron, 1935 and Thompson, 1974) • 1.3.2 Mechanical removal Mowing ragwort plants checks seeding but does not prevent regrowth. Cutting plants in flower pramotes strong regrowth fram the crown and roots regardless of the cutting height above ground (Cairns, 1938). Pulling out flowering plants results in same regeneration from root fragments and while fire destroys seed and many plants some regeneration from root fragments may occur. There is considerable regeneration from seed and root fragments after ploughing but quick establishment of a dense sward reduces recolonization. It has been suggested that fertilising pasture promotes a dense sward and decreases the chances of seedling establishment (Bill Makepeace, pers.comm.1982). Farmers must balance the costs of improving pasture against the costs of removing conspicuot5 plants which they are required by law to do. This often results in cheaper short term measures such as mowing or ploughing being resorted to. The application of costly herbicides which weaken pastures does little to solve the problem. Sheep grazing prevents flowering and reduces ragwort density but high sheep numbers are necessary to ensure ragwort plants are 10. grazed and sane stock losses occur as mentioned above. 1 . 3 . 3 Insects as biological control agents of ragwort Over 6 5 insect species from five different orders have been recorded on ragwort (Harper am Wocrl, 1957). When it was introduced into North America, Argentina, Australia and New Zealam, ragwort was not accompanied by its native fauna but it was expected that native Senecio species may provide a source of endemic insects which would take advantage of a new host. Miller (1970) cites six species with potential to control ragwort in New Zealand but none of them are native. These incl ude three Lepidopteran species, two Diptera am an aphid. The cut worm Ariathisa comma (Walk.) (Noctuidae) damages juvenile ragwort but is a P2st of cuI ti vated plants. The pyral id stem borer Hom eo san a vagella Zell e also infests a high proportion of plants but does not prevent flowering (Cottier, 1931). The noctuid �ctemera annulata Ed. cannonly known as the magpie moth, is the best known insect found on ragwort in New Zealand. The larva, which is often confused with the cinnabar moth, feeds on ragwort foliage but seldom reaches high densities. It also feeds on native species of Senecio, Brachyglottis, groundsel and cineraria (Miller, 1970). There a.fe no published data on the life history of the magpie moth al though it was foum to sequester ragwort alkaloids (Benn et al., 1979). It is attacked by three parasites; the ichneumonid Echthromorpha intricatoria (Fabr.) which oviposits in the pupae and two tachinids, Cerosomya �ctemeriana (Huds.) and �.�asta(Hutt) which parasitize the larvae. The Dipteran stem 11. borer, Agromyza aeneiventris Fln., is cammon on ragwort (Miller, 1970) but has little effect. A Dipteran leaf miner, Phytomyza atricornis Mg., rarely damages ragwort plants in the field but high densities occur in the sheltered conditions of the glasshouse and insectary (Kelsey, 1937 and pers.obs.) where it may kill the plants. The aphid, Brachycaudus helichyrysi (Kltb.) which clunps flower s together wi th its honey dew may have some effect on seed dispersal (Miller, 1970) . Several insects have been released in New Zealand specifically as biological control agents of ragwort. The most conspicuous of these is the cinnabar moth, � jacobaeae L. (Lepidoptera: Arctiidae). It is univoltine throughout its range with an obligatory d iapause in the pupal stage. Al trough liberated throughout New Zealand from 1929 to 1932 populations persisted only in the south of the North Island (Miller, 1970) . As a control agent of ragwort the cinnabar moth has had mixed success overseas. In the Netherlands there have been local extinctions of both the moth and ragwort (van der Meijden, 1976 ) while at Weeting Heath in England both plant and insect populations experience high amplitude fluctuations with no control at low plant densities (Dempster, 1982) . In Oregon, USA, the weed has been maintained at relatively low densities by cinnabar moth over a five year period (Stlinac and Isaacson, 1976) . but in British Columbia, Canada, moth populations are stabilising at levels below that needed for complete defoliation (Harris et al., 1976) . 12. The rag\\Ort flea beetle, Long i tarsus j acobaeae Wat. (Coleoptera: Curcu1ionidae) , was liberated in New Zealand during 1983 with further releases planned (P.Syrett, pers.comm.l984). There are t\\O biotypes. The SWiss strain has a facultative egg diapause and the Italian strain has a facultative adult aestivation period (Syrett, 1983). The beetle attacks the roots of ragwort duriO) winter am sprir:g am it is hoped that it may canplanent the attack of Tyria jacobaeae which is active in summer. 1.4 Ragwort seedfly 1.4.1 Description P�ohylemyia jacobaeae (Hardy) and R.seneciella (�ade) are t\\O similar European anthcmyi id fl ies. Both are cannonl y called ragwort seedflies because their larvae develop in flower heads of ragwort. Miller (1970) describes R.seneciella as a small, dull greyish- pubescent insect; darker in the male, and 4-Smm long. Collin (1936) states that R.jacobaeae adults are slightly larger averaging 6rrm in length compared to 4-5mm for R.seneciella. The t\\O species are more reliably distinguished by differences in shape and hair arranganent of the ovipositor in females and the hypopygium am fifth abdominal sternite in males (Holloway, 1983). The underside of the costal vein in R.jacobaeae also has an additional row of evenly spaced hairs (Holloway, 1983). 1.4.2 Liberation and establishment in New Zealand Both species of pegohylanria were imported into New Zealarrl fran England between 1928 and 1939 (Miller, 1970) but field recoveries of PegohY��l�c: throughout New Zea1arrl in 1981 ard 1982 have all 13. been of �.jacobaeae (Holloway, 1983 ) . Seedfly were transported to New zealand as pupae and released as adults. In February 1936 , 612 insects were released at the Cawthron Institute, Nelson 1n the South Island and 1 , 378 near Putaruru in the North Islam. Durin:J autunn am fran late winter to early summer 1937 , approx 92 , 470 flies were liberated, 80 ,000 of them at Putaruru am the remaimer at Coranandel, Matamata, Tirau, Opotiki, Ohinemuri, Paeroa, and Te Awamutu and again in Nelson. These came fran the 7th to lOth consigrments fran Englam and only �.seneciella specimens were preservErl fran these (Holloway, 1983 ) . Unfortunately no record of the proportions of the two species releasErl was kept am no distinction was made between them in follow up observations. SeErlfly populations were initially out of synchrony with conditions in the Southern Hemisphere and liberation at a site at Ngongotaha near Rotorua in autunn 1940 was followErl by an observErl winter infestation in June 1940 and a further generation in the next summer. Infestations persisted at the Cawthron Institute, at Tirau and Putaruru until 1944/45 am at Ngongotaha until at least 1942 but no flies were found there in 1950 (Kelsey, 1955 ) . He found, however, a thriving population at RErlwoods Forest, Ngatira, the offspring of 36 adul ts releasErl there in February 1937. In 1984 seedfly was present onl y 1n the central North Island , Bay 14. of Plenty and the southern Waikato (Fig.l.l). 1.4.3 Life history The life cycle of the ragwort seedflies has been described by Cameron (1935), Miller (1970) and Frick and Andres (1967). The seedfly pupa overwinters in the soil and aJults energe in spring as ragwort boos are developing. The eggs are laid down among the florets or alongside the green bracts. Only one larva develops per seedhead even though several eggs may be laid. First instar larvae are found inside individual developing florets but may move from one floret to another and can be detected by chewed windows in the wall of the floret (pers.obs.). Neither the egg nor the early first instar are visible externally. Damage becomes visible as browning of the florets caused by late first instar and second instar larvae feeding. Later the florets are displaced by the feeding larva and the pappus becomes prenaturely extrooed and matted together by larval secretion. Signs of larval damage differ in the two species. The early sign of a dark spot in the centre of the disc is more common in Oregon where only P.seneciella have established and the pappus, although not extruded as far, is covered with a white, sticky exudate not seen in New Zealand (P.Syrett, pers.comm.1984). Third instar larvae of E.jacobaeae observed leaving the plant during this study W2re not attached to a thread but dropped freely or crawled down the stem. Some information on seedfly life history was obtained in New Zealand from flies imported for biological control (Miller, 1970) but no distinction was rnade between the two species. The longest Fisi tion to larvae leavin:J the flowers ranged fran 36 days to 94 days depending on the temperature and larvae pupated within five days of leaving the seedhead . The pupal stage ran:Jed fran 130-240 days . 1.4.4 Effectiveness as a biological control agent of ragwort Cameron (1935) rep::>rted that Pegohylemyia infested approximately 8-9% of the capitula in southern Englarrl and 33-34% in northern Scotland with 75% of the seed in the capitulun destroye:3 in both areas. He also notErl that the northern range of p.jacobae� extendErl much further north than Tyria in the British Isles arrl suggests that this advantage may be useful in higher and colder reg ions, where it may be imp::>ssible to establish � . In New Zealarrl, Kelsey (1955) reporte:3 that 91% of the florets of early flowers and 43% of those in the main flowering period were infested by seErlfly at RErlwoads Forest in 1949/50 while 60-98% of the early crop and 8-71% of the main crop was infestErl in 1953/54. Kelsey (1955) also found that the average number of seErl surviving from infested capitula was 12, (range 0-22) but these failErl to germinate. SOme information on larval and pupal mortality was obtainErl from flies col lected from Redwoods Forest for shipment to Australia (Hoy, 1958). Of 33 dead larvae examined 32 were infected by nenatodes of the genus Rhabditis and one larva had been inva::Jed by a fungus. Fourteen out of 21 pupae examinErl were parasitize:3 by a fUnJus, 2 by Rhatdit�� species aoo in 2 the cause of death was 17. unknown. Most of the nematode species were saprophytic so that flies infecte:::l had diEd fran other causes. Accordinj to HOY (1960) the flies arrivEd in Australia in good condition and three quarters of them emerge:::l successfully. He concludEd that little internal damage had occurre:::l during the of retrieval of pupae by siev inj because the nuuber of defonnEd fl ies that emergEd was low. Seedfly released in Australia in the 1930s and 1950s faile:::l to establish (Waterhouse, 1967) but those releasEd by Frick in California in 1967 were well established by 1968 and 1969 (Frick, 1969). Andres and Davis (1973) found that this release site had been inadvertently mown but the fly was recovered close by in 1976 am more releases were planne:::l (Andres�, 1976). R._s_e_n_ec_i_e _l_la_ is now also established in Oregon (Isaacson and Ehrensing, 1977) after initial difficulties. Seedfly is hard to locate in the years immediately following release but usually reappears in the vicinity several years later (P.Syrett, pers.comm.l984) • 1.5 objectives of the present study Field populations of see:::lfly and ragwort were studied during the summers of 1982/83 and 1983/84. The field sites are described am the sampling methods and data collected are presented in the following chapter. The field observations were supplementEd with laboratory studies which are described in Chapter 3. This involvEd experimental manipUlation of the pre-oviposition period to determine the relationship between oviposition site availability am infestation levels. In addition, egg and larval development rates obtained under controlled conditions were used -- -- - --- - - - -- 18. to construct survivorship curves. Chapter 4 draws on the information obtained in the laboratory to construct a model of the seedfly pre-oviposition period and development within the seedhead. This model was compare:] wi th the field population to check the synchrony of various stages of seedfly life history with its host. 'Ihroughout the chapters the resul ts are canpared with published informa Hon on other anthanyi id fl ies. However, information on seed eating anthomyiids is scant, with research concentrate:] mainly on those flies of econanic importance. The majority of these are root fee:]ers which have differing life history strategies. Some are less host specific than seedflies and others have more than one generation per year. However, their taxonanic relationship to ragwort seedfly warrants their inclusion for comparative purposes. The final chapter is a discussion of the factors which influenced seedfly infestation levels of ragwort during the study per ied and the seed fly's role in the biolog ical control of ragwort. 19. QiAPTER '!WO: FIELD STUDIES 2.1 Intrcx3uction Two seedfly populations were sampled in the field over two successive summers to obtain information on their life history and infestation levels of ragwort. The field sites chosen (Fig.2.l) were representative of the clilnatic conditions experienced by the seedfly in New Zealand. Research effort was concentrated at the lower altitude, warmer Redwoods Forest site while a more exposed area at the Desert Ed was the site of less intense data collection. The latter was usErl for canpadrg the phenology of the popUlations at the two si tes. The si tes selected were protectErl fram interference by grazirg anilnals, other weed control operations and curious people. 2.2 The field sites 2.2.1 The Redwoods Forest si te This was a 6-2 hectare site situated 10km SE of Putaruru, North Island, New Zealand (N.Z.M.S.I, N75, Grid Reference Area 350120). It is part of RErlwoods Forest owned by New Zealand Forest Products Limited Kinleith. The area was formerly unproductive for dairy ?i re 2 . 1 Loca t ion of t�e field s ites North Island 20. North t For.at D x Deaert Rd ____ .--.. , 21. fanning due to cobalt deficiency and was originally the site of a trial plantation of redwoods (Sequoia species). stands of Pinus radiata Don., Eucalyptus regnans F.Muell. and .E!.delegatensis R.T.Bak. were planted in the 1940s and 1950s (Fig.2.2). The site was leased for grazing from 1/10/59 to 1/10/79 and pines were planted at the "Pines Plot" and the steeper "Road Plot" in 1981/82. The "poles Plot" was a grasslarrl strip providing clearance for double pole power lines (Fig.2.3 and plate 2). Geology, soil type and topography The area is part of the Mamaku Plateau and was fonned in the Pleistocene when large ignimbrite floods from the Rotorua area poured northwest am north to Tirau and Tauranga. Subsequent erosion has partially stripped away the soft upper layers of the ignimbrite sheets exposing the harder rocks below, forming the rolling country and rocky outcrops typical of the region. The yellow-brown pumice soils of the area are derived from the Taupo rhyolitic ash showers of 1800 B.P.(Gibbs, 1965). The site is 226m above sea level and forms part of the catchment of the Waihou River which flows north to the Hauraki plains. Climate Meteorological data from 1931-1983 were obtained from Kinleith, lOKrns southwest of the Redwo0:3s Forest f ield site. The mean monthly temperature for thi s period ranged from 17·SC in February to 6.9C in July. The mean annual ra infall was 1544mm and there was an average of 172 raindays per year. Winds are predominantly from 22. the west and southwest. Vegetation The dominant pasture species in both areas of recent pine plantation and the Poles Plot were Yorkshire fog (Holcus lanatus L.) and cocksfoot (Dactylis glamerata L.). The other grasses present included brown top (Agrostis tenuis Sibth.) , perennial ryegrass (Lolium perenne L.) and prairie grass (Bramus catharticus L.). Lotus f2Erlun:::ulatus L. was comnon as were creeping and giant buttercup (Ranun:::ulus repens L.) and (R.acris L.) . Dock (Rumex obtusifolius L.), broad and narrow leaved plantain (Plantago major L. and P.lanceolata L.) , blackberry (Rubus fruticosus L.) and catsear (Hypochoeris radicata L.) were also present. 2.2.2 The Desert Road site This 100m2 site is also part of a grassland strip providing clearance for power pylons. It is situated in rnanuka shrubland in Tongariro National Park, North Island, N.Z., O·Skrns north of the oturere trig on State Highway 1 (Desert Rd) (N.Z.M.S.I, Nl12, Grid Reference Area 255785) (Fig.2.4 and plate 3). Geology, soil type and topography Pyroclastic deposits of the Desert Rd area consist of Tongariro ash overlain by dark Mangatawai andesitic ash deposited by Mt. Ngauruhoe during its growth in the early Pleistocene (Gibbs, 1965). The area, at an altitude of 850m, is drained to the east by the 23. Makahikitoa Stream. Climate Meteorolog ical data for the area was obtained from Waiouru 800m a.s.I.(Fig.214> (N.Z. Meteorological Service, 1983). For the period 1966-1980 the mean annual rainfall was 1096 rom with an average of 145 raindays per year. The mean daily maximun was 13-7 C in February and 3·6 C in July. There was on average 96 frost days and 19-5 snow days per year. vegetation Brown top (Agrostis tenuis) and prairie grass (Bramus catharticus) together with catsear (Hypochoeris rad icata L.) and hawksbeard (Crepis capillaris L.) were the dominant plants. The surrounding vegetation was predaninantly manuka (Leptospermun scoparitml Forst.) and the low, herbaceous shrubs, Cassinia fulvida Hook., Coriaria pteridoides Oliver and Gaultheria antipoda Forst. were also present. 2.3 Methods and Fquipuent 2.3.1 SampliIl9 frequency Fly populations were sampled fortnightly at both sites from October to April. This sampliIl9 frequency was constrained by available travel furrls . Tanperature was monitored hourly at Redwooos Forest by a Grant tanperature recorder. The tanperature sensitive probe was plaCed among the seedheads one metre above the grourrl. At the Desert R::3 si te a thermohygrograph was placed at 2 P R edwo o d s Fo r e s t f i e l d s i t e s how i n g P in e s P l o t i n t h e f o r g round a n d d o u b l - po e e l e c t r i c i ty w i r s o f t h e P o l e s P l ot in t h e d i s tan c e D e s er t R o ad f i e l d s i t e w i t h e m r g n e e t rap s , s u r r o u n d i n g man u ka s c ru b and ov e rh d p ow e r \v i r e s / F i g u r e 2 . 2 A r ea �laD North t Rai lway l ine D ubie pole­o line Road Stream Pines Eucalypts o 24 . o f Fo r e s t f i e l d s i t e 100 200 300 metres Figur e 2 . 3 Map of Redw o o d s For e s t s i t e ,----------------- - ------- - -- -- North t Poles Plot -:'/ <5P Plo Pines 226 In o 25 Key me t r e s =:= Double pole- line Road Sticky trap Contour line Scrub SP Sub plot 25 . 50 '--------------�------�----- --- ----�------------- ---- --- ---- Figure 2 . 4 Area Man o f the D e s e rt Road s i t e 26. North t ,/ Turangl 2 5 kms /� / K ey Tongariro Nationa l Park Park Boundary Il 850m Oturere Trig x Field S ite State Highway 1 (Desert Rd) l--________ .�. __ , I J r I I I I I I I /' , /.W . �/ a louru -1 2 5 k m s ,/ / ,; / ,; / ,; / ,/ / K a imanawa 1 . State Forest 2 , k ms 27 . ground level . This recorded temperature cont inuously for ten days following sampling . Tanperature data for the final four days of the sampl ing interval was obtained from a meteorolog ical station at the Rang ipo Power station 2kms west of the site . 2 . 3 . 2 Adults Adult seedfly were sampled with emergence traps and sticky traps . The former gave an estimate of absolute fly numbers and the latter an estimate of longevity. Emergence traps These were 30cm high and sampled an area of 0 -2 m2 (Fig . 2 . 5) . The frames were constructed of No . 8 fencing wire and covered with green plast ic mesh . 125ml "T .V. L . " collecting j ars with their l ids sl it were inverted over a Scm section of plastic tubing glued into the apex of each trap . Ten centimetre wire projections at the base of the frame were pushed into the ground to anchor the trap . The number and posi tion of traps used is sumnar ised in Table 2 . 1 . The traps were d istr ibuted randomly by marking out a gr id at each plot and selecting the coord inates for the traps from a table of random numbers . Sticky traps These consisted of an aluminium cylindr ical drum lOcm in d iameter and 30cm long (area of 942crn2 ) , supported on a ver tical 1m steel rod attached to a wooden block inside the cylinder (Plate 4 ) . A sheet of clear plastic , wrapped around the drum and held in place 28 . by rubber bams, was covered wi th an OOhesi ve , "Tack trap 1 " • Fi ve of these sticky traps , ( two white and three yellow) , 'Were placed at Redwoods Forest (Fig . 2 . 3) am two traps ( one white am one yellow) at the Desert Rj field si te . The plastic sheets were replaced at each sampl ing date after the number of fl ies caught hOO been recorded . 2 . 3 . 3 Eggs and larvae Because of the variabil ity o f plant s i ze (Table 2 . 2) the fly population wi thin seEdheOOs was sampled on a stan basis . Fi fty stems were collected at each sampl ing date from Redwoods Forest am ten fran the Desert R:1 s ite . This number o f stans was the maxirmm that could be analysed between sampl ing dates . Stems were collected at random by mov ing from plant to plant and taking every nth stan encountered . N was prov ided by a table of random numbers between 1 and 15 . Samples 'Were frozen as soon as possible to prevent any further development of the fl ies with in the capitula . When thawed , the seedheads W'ere d issected urrler a microscope and the number of eggs am larvae per stan were recorded . Larval instars 'Were determined by measuring the max imum length of· their sclerotised mouthparts . These fell into three, clearly defined size groups ( Table 2.3) representing the three larval instars . The number of seEdhea:1s per stan was al so recorded . Seed head s expected to conta in eggs W'ere des ignated as "buds" and those capable of supportilXJ larvae were termed "capi tulae" . When eggs and larvae cohab i ted a seed head i t was al so recorded as a I . A n ima l Repe l l en t s I n c . G eorg i a USA 29 . capitulun. 2.3.4 Pupae Pupae were sampled at the Redwoods Forest si te using the fol lowing two methcrls : ( i ) Soi l samples : SOil samples covering the same area as emergence traps and lOam deep were obtained randomly using the method for selecting the posi tion of emergence traps . These samples were considered to be deep enough to collect all the pupae present because an ini tial pilot tr ial of ten 0_ 3m2 samples had shown that seedfly pupate in the top 0-5am layer of soil . Soil from each sample was placed in water so that the orange pupae could be collected as they floated to the surface . ( i i ) Collection of pupae at the stem base ( autumn 1984 only) : Stems wereselected randomly as in 2.3.3. Each stem was t iErl to an aluninium stake to prevent wind movement . All the additional stems were removed when a stem was selectErl fran a multiple-stemmed plant . A plastic tray, 30am in d iameter and 120m deep , was sl it to centre am f ittErl aroum the base of the stem. The sl it was then wired together and the tray was filled with moistened sam ( Plate 5) . The sam was s ieved to a grain size of less than 2-5m:n so that when sievErl again any pupae present could be easily col lected. The sam was covered with cut grass to reduce evaporation and "tack trap " was spread arourrl the r im of the tray to prevent any larvae escapinj. When the pupae were col lected at each sampling date any third instar larvae found were returned to the ir respective trays . P l a t e 4 P lat e 5 S t i c t r a t R ed wo o d s o r e s t A r ran g ement a t R e dwo o d s F o r s t f o r c o l c t in t hi r d in s ta r larvae . The plant s s taked a n d a t ra y f i l l ed w i th s an d w a s a c e d a t t h e b a G .. F i gu r e 2 . 5 plastic mesh \ Emerq;eCl c e Trap TN.L. collecting jar ___ 1 0cm 30. -plastic tubing b 1 2 . 1 A r e a i n 2 ( R e f e r m 31 . t o F i g 2 . 3 f o r p l o t l o c a t i o n s P o l e s P l o t P i n e s P l o t Eoad P l o t D e s e r t r� o a d Spr i n 1 4 . 2 1 9 8 2 ( 7 0 S p r i n g 4 . 0 5 1 9 8 3 ( 2 0 - - 4 . 0 5 t ra p s ) - - ( 2 0 t raps ) 7 . 1 4 . 0 5 4 . 0 5 t ra p s ) ( 3 5 tra p s ) ( 2 0 t rap s ) ( 2 0 t ra p s ) - F r e q u e n c y d i s t r i OU t i on 0 f s t e m s � p I a n t a t R edwo o d s Fo r e s t Y ea r v.,rh e n s 8 'J D l e d S t e m s / I) an t 1 9 8 / 8 3 1 9 8 3 / 8 4 1 1 0 5 9 4 2 3 3 3 8 3 3 4 3 6 I 1 4 2 1 4 5 Q 1 1 / 6 7 7 7 5 7 8 2 5 9 1 4 1 () i ' u '3 1 1 1 T () t 'l ; 2 1 () ;) 7 32. 2 . 3 . 5 Ragwort populations Rosettes larger than lSam in diameter in two 300m2 subplots of the Poles plot (Fig . 2. 3) were mapped in spring 1982 and 1983 and autumn 1983 and 1984 (Append ix 1 ) . No distinction was male between regrowth plants and those which hal developed fran . seedlings. The mrnber of stems flowering in these two subplots was also recordErl at each sampling date . At the end of the sumner the number of stems flowering as a percentage of the total which f lowerErl was calculated for each sampling date at the two subplots . In crldition, the total number that had flowered over the entire site was determined in autumn from 9m2 quadrats (Greig-Sui th , 1964) selectEd at randan fran the mapped grids at each plot (section 2. 3. 2) . Absolute estimates of the total number of stems f lowering at each sampling date over the entire site could then be determined from the cumulative percentage curves of f lower ing at the two subplots. This method was preferrEd to collecting stems on an area basis as much larger sample sizes would have been requirErl to allow for differing ragwort densities in the three plots. 2. 3. 6 Ragwort seed germination Samples of seedheads were collected fran Whakamaru and the Desert Rj in autunn 1982 am fran Redwoods Forest in autumn 1983( F ig . 2 . 1 .>. The seeds \>Jere removed fran the seec1heads and placed in petr i dishes on moist f i l ter paper at 20e with a 14 hour l ight and 10 hour dark photorer iod . Percentage germination was recorded after two weeks . 2 . 4 Results and discussion fram the Redwoods Forest s i te 2 . 4 . 1 Emerg ence Emersence trap efficiency 33 . 119 flies (54 males and 6 5 females) were released into traps as they became available fram the laboratory population . 50±8% of the males and 20±5% of the females released were found dead in the collecting j ars. Although the trapperl flies often decayed over the two week per iod of each tr ial the wing s al ways remained intact. Emergence trap results Flies emerged over a per iod of 2 months in 1982 and 4 months in 1983 ( Fig . 2 . 6a+b) . Males emerged sl ightly before females but the sex of a proportion of flies caught ( 7% in 1982/82 and 37% in 1983/84) could not be determined. The female :male rat io of seedfly adults that emerged at RedWO. 7 6 5 4 3 2 � 2 5 0 ·n +-' [f) l(\ 2 0 0 :;::: o +' �o 1 5 0 ;::l (D () H 1 0 0 Ql ,.0 E ;::l c rl cd +-' o b 5 0 3 1 O c t - - - 1 4 2 9 N o v / / , / " / " / " 1 2 2 - - - - D e c J a n S a mpl ing d a t e 1 9 8 2/ 8 3 1 9 8 3 / 8 4 F e b M a r F i gu r e 2 . 8 N u mb e r o f S e e d f ly a d u l t s c a ugh t o n s t i c ky t ra D s a t R edwo o d s F o r e s t 1 7 3 1 D c : . J 1 C, �? 9 .1 ' D e C' / ' 9 2 3 J :1 n " " , 7 2 CJ F b 1 9 8 2 / 8 3 1 9 8 3 / 8 4 5 1 9 10 r 37. Del ia antigua Meigen. Neither of these two species are host specific . Adult seedflies were caught on the sticky traps for a period of five months in summer 1983/84 . Sticky trap catches dropped by 50% in the six weeks after they peaked in November 1983 and by a further 50% in the next 12 days ( F ig . 2 . 8 ) . Longevity, however , was measured in the wheat bulb fly in the field . It was 55 days for males and sl ightly longer for females ( Jones , 1970 ) . Female beet fl ies , Pe9crn:tia hyoscyami Panz . also l ived longer than males ( 70 and 50 days respectively) but longevity decreased with increasEd temperature . MatEd males l ived longer than urInatEd ones but the reverse occurrEd in females (Hafez et al . , 1970 ) . 2 . 4 . 3 SeEdfly infestation levels of ragwort All stems were infested by seed fly at the s ite . OViposition coincidEd with the f irst appearance of buds in the field in both years . Character ist ics o f seedheads containing 83gs were measured within 24 hours of ov iposition by females on pottEd ragwort plants placed in the field . E. jacobaeae ov iposited on buds with a d iameter rang ing fran 3 * 8 to 5 - 8 mn and a disc floret length of 2 -7 to 5 ' 6mn (Table 2 . 4 ) . The number of ray florets was constant at 13 . Z immerman et a1 . ( 1984 ) found a positive correlation between 839 deposition and inflorescence si ze in the seEd predator Hy1em:ta (Del ia) species and Fr ick ( 1970 ) found bud si ze preferences in E. seneciel la . It is l ikely that si ze is ind icative of the stage of development of the bud and this is an 3 8 . linportant consideration for the ovipositing female since f irst instar larvae must complete their development within unopened florets . Depth of the bud measura:3 as the length of the d isc florets was also considered as seedfly eggs which are approxlinately lmn in length are laid upr ight. contact chemost imulation and colour is also l ikely to be linportant in the choice of ov iposi tion s i tes but was not investigata:3 in this study. Frick ( 1970 ) found P.seneciella females preferred to oviposi t on yellow and orange models but that odour was requira:3 to stlinulate v iposit ion . The mrnbers of seedfly wi thi n seedheads at Redwoods Forest throughout both summers are presented in Figures 2.9a+b and Figures 2.10a+b g ive estimates of the number of seedheads at each sampl ing date . Ragwort densi ty was 28 , 788 stems per hectare in 1982/83 and 30 , 132 in 1983/84 . Table 2. 5 summar i zes the stem densi ties at the three plots in both years. The var ious stages of both seedfly and ragwort l i fe history are representa:3 as a cumulative percentage of the total of the stage present in the season ( Figs.2.1la+b). The ov iposit ion curve is calculated fram the sum of the eggs , f irst and second instar larvae since the sampl ing interval is less than the total duration of these stages . A pre-ov iposi tion per iod of approxlinately six weeks was observed in 198 2/83 and 1983/84. The timing of seed fly emergence arrl oviposition was slinilar in both seasons . Although ragwort began to flower i n early December in both years ragwort 3 9 . inflorescence development was faster in 1983/84 . In this year the availability of oviposition sites (represented as the hatched area in Figures 2 . lla+b) was increased. No oviposition took place on buds of ragwort plants transplanted at the site fran. the 14th of November to 28th of November 1983 . Unfortunately there were no other flowering plants available to continue this experlinent . Oviposition did , however , occur on transplanted plants 12 days before the field plants flowered in 1984 . When the crop contents of 30 adults collected durirg the pre-oviposition period in spring 1984 were examined no pollen was found but the crops of 6 of the flies contained a yellow liquid. The number of eggs in females collected in the field in 1982 rarged from 0-29 with a mean of 7 ·3 (Table 2 .6) . Percentage infestation declined throughout the summer fran. an initial peak of 19% in 1982 and 39% in 1983 (Fig . 2 . 12) . The effective levels of infestion, in late February, when tllird instar larvae peak and all the ragwort has flowered, was approximately 10% in 1982/83 and 20% in 1983/84. 2 . 4 . 4 Multiple infestation of ragwort seedheads The majority of seedfly eggs are deposi ted singly in ragwort capitulae . Larvae develop exclusively in the capitulae and have no abil i ty to move from one capitulum to another . Mul tiple infestation does occur but only one fly per seedhead surv ives to 40 . T a b l e J . 3 D e t e r m in a t i o n o f e e d f l v l a r v a l i n s ta l' s J2.y mou thpa r t l engt h s r- - r--- 1 s t I n s ta r 2n d I n s ta l' 3 rd I n s t a r l a rv a e l a r v a e l a r va e - x 0 . 2 7 m m 0 . 4 6 rn m 0 . 7 1 rnm S E 0 . 0 0 20 m m 0 . 0 0 2 3 rn m 0 . 0 0 2 9 rn m R a n g e 0 . 2 2 - 0 . 3 1 m m 0 . 4 - 0 . 5 4 m m 0 . 6 4 - 0 . 8 0 m m n 1 0 0 1 0 0 1 0 0 T a b l e 2 . 4 � ha ra c t er i s t i c s o f s e e dh e ad s c on t a in ing s e e d f ly e ggs 0 - 2 4 h o u r s o l d - D i a me t e r N o . o f d i s c L en g t h o f d i s c o f s e e d h ead f l o r e t s f l o r e t - x 4 . 5 m m 4 8 . 8 3 . 7 m m S '" L, 0 . 0 4 :r, m 0 . 5 4 0 . 0 7 m m qan e; e 3 . 8 - 5 . 8 m m 3 8 - 6 3 2 . 7 - 5 . 6 m m n 1 0 0 1 0 0 1 0 0 Table 2 . 5 N u mb e r o f r agw o r t s t e rn s a t R edwo o d s Fo r e s t N u m b e r s 2 ( x o r" ) (l e r m ± U L, P o ] s P l o t P in e s P l o t Road P lo t i-----. l\ u tu mn O . S 6 :t 0 . 1 4 .3 . L::'± O . 3 8 4 . 9 1: 0 . 1 3 1 9 8 3 n 6 0 n 1 0 n :: 1 0 A i l t u mn D . 6 5 ± 0 . 0 C) 5 . 9 7 ± () . 3 0 3 . 8 1 ± 0 . 2 1 1 9 (8 4 11 6 D n _. 2 0 n = 2 0 ., 2 2 3 , 1 2 5 m 2 T o t a l 2 5 , O O CJ rn ,:: 1 ) , 7 5 [) m a r e rl \) r p i (J t __ L.. . -- -- 4 1 . Figu r e 2 . 9 E s t i ma t i on o f t h e nu mbe r s o f s e edfl and la rva e at in ragwo rt s e edheads Ba r s ind i c a t e R edwo o d s Fo r e s t a ) b ) x 1 0 5 5 aJ H cd 4 +-' () aJ ...c H 3 aJ P, CfJ H aJ .D 2 E ;::l Z C o l our graph - 1 9 8 2 / 8 3 O v e rlay - 1 9 8 3 / 8 4 / 1 3 2 7 D e c ,,1 , / , / / , 1 0 2 4 J an 7 2 1 F e b S a mpl ing dC" t e - - - -eggs la r vae , -- 1 s t ins tar --- 2nd i n s tar --- 3rd i nstar 7 2 1 4 Mar Apr [ \)\<, I TA L <"' c?"i = I h\ � PA<;c wti'-1 C VCa.. L'I"\'I . - .'II\ o"t 5SIt-"{ L t BR A- Il.,( ) 4 1 . F igu r e 2 . 9 E s t i mat ion o f t h e nu mbe r s o f s e edfl and larva e at in ragwo rt s e ed head s Ba rs ind i c a t e R edwo o d s Fo r e s t a ) b ) x 1 0 5 5 ill H (1j 4 +' C) ill -c H 3 ill p., U) H ill ,D 2 E ;:l Z C o l o u r graph - 1 9 8 2/ 8 3 O v erlay - 1 9 8 3 / 8 4 1 3 2 7 D e c Samp l in g de. t e _ = = = eggs la rvae , � ........... 1 s t i ns tar --- 2nd i n s tar ��:::: 3rd i nstar 4 2 , M ea n no . o f cap i t u i ae per stem D a t e 1 9 8 2 1 8 3 1 9 83 /84 1 3 / 1 2 8 '9 0'0 5 2 7 / 1 2 1 0 '6 8·2 1 0 1 1 1 3 ·2 3 3'9 24 1 1 2 2 ·8 5 3 ·7 7 / 2 4 4 ·1 86 ·9 2 1 / 2 6 8 ·9 1 1 0 ' 5 7/ 3 8 3 · 2 8 5 '3 42 . _ _ S ?A-L r_ N Ill'! C{Cf2-. LA:-'-( - .v..A � �E--f L- I � (L ��( J C. O P '- ! i 1-1 1 'I_I e F i g u r e 2 . 1 0 a ) 1 9 8 2/ 8 3 E s t i mat ion o f t h e nu mber o f ra wo rt bud s and ca n i tulae Bars in d i c a t e s tandard e r r o r s} at Redwoods Fo r e s t b ) o v e r l a y 1 9 8 3 / 8 4 x 1 06r---.---.---.---.---.---.---.---.---.---.---.--. Q) H qj +' C) Q) ..c: H Q) P, Ul '"d qj Q) ..c: H Q) ;3 o rl '+-I '+-I o H Q) .D E ;:l :z: 3 2 1 2 4 J an 7 2 1 F eb Mar \ \ \ \ Sa mpl ing date M ea n no . o f cap i t u l ae per s t em D a t e 1 9 82 / 8 3 1 9 83 / 84 1 3 / 1 2 8 '9 0 '05 2 7 1 1 2 1 0 '6 8·2 1 0 / 1 1 3 ·2 3 3·9 24 / 1 2 2 ·8 5 3 ·7 7 / 2 4 4 ·1 8 6 ·9 2 1 / 2 6 8 ·9 1 1 0 '5 71 3 8 3 · 2 8 5 '3 \ \ bu d s cap itu l a e \ \ \ \ \ 4 1 5 2 A p r May �, ,'" 4 � �� '� C / 2 0 �" " 1 0 � � F i r e 2 . 1 1 C u m u la t i v e p e r c en t age c u r v e s o f s e e d f ly a nd r agw o r t s e e d h e a d d ev e l opment §.,�ge s o f Redwo o d s F o r e s t T c u r v e s ha v e b e en f i t t e d by e y e a n d t h e ha t c h e d a r ea r e pr e s en t s t h e r e g i o wh e r e s e e d h e a d s a r e a v a i l a b l e f o r o v i po s i t i on b y s e e d f l y � 9 Q r'J / S " I u tC:._ c .... J b ) v c l a y 1 9 8 3 / 8 4 1 5 2 9 �) ov 1 3 2 7 D e c 1 0 2 /+ J an 7 2 1 F e b 7 2 1 M a r /+ A p r 9 0 r-' -::: ....J 8 0 0 .,...) G) 7 0 ,- ....J :.- 6 0 :::: (. 5 0 .... , r ,- ---' r:: 4 0 G) 0 � c 3 0 0 , > 2 0 'ri .w X 1 0 -1 :J E ::c - , f i �u r e 2 . 1 1 C u mu la t i v e pe r c en t age c u r v e s o f s e e d f ly a n d ragwo r t s e e d h ea d d ev e l opme n t s tage s o f R edwo o d s Fo r e s t T h e c u r v e s ha v e b e en f i t t ed by e y e and t h e h a t c h ed a r ea r ep r e s en t s t h e r e g i o n V1 h e r e s e e d h ea d s a r e ava i l a b l e f o r o v ip o s i t i on by s e ed f l y a ) 1 9 8 2/ 8 3 b ) O v e r l a y 1 9 8 3 / 8 4 1 5 2 9 .J o v 1 3 2 7 D e c 1 0 2 4 J an S a mpl ing d a t e 7 2 1 F e b 7 2 1 tvl a r 4 A p r j I 1 0 -- .f' - � (' g -.;:, -' � - v> -c) � s· (7\ L i 0 4- � � ( -. I � ..:l> v' (/. r)' ...( r - � � � w � ...r U 44 . F i gu r e 2 . 1 2 P e r c en ta � e o f r a wo r t s e edhead s in f e s t ed b v s e ed f l v a t R e dwo o d s F o r e s t B a r s r ep r e s en t s tanda rd e r ro r s c: 0 5 0 -r<. � qj +" (fJ 4 0 QJ 4-; c: or< QJ bG 3 0 qj +.) c: QJ () 2 0 H QJ P.-, I 0 1 3 2 7 D e c 1 9 8 2/ 8 3 - - - - � 1 0 2 4 J an 7 2 1 F e b S a mpl ing d a t e 1 9 8 3 / 8 � 7 2 1 �1 a r 4 Ap r 45 . pupate . In thi s study mul t iple infestat i on fol lowed the general trend in egg laying ( Fi gs . 2 . 9a+b ) ,'l i th the lower i nc idence of mUlt iple larval i nfestat ion ( Fi gs . 2 . 13a+b ) due to the high mortal i ty of older larvae as cond i t ions become crowded in L�e seedhead . As a percentage of the total seedheads i nfested , multiple infestation dec l i ned from a peak of 22% a t the beginning of the season in 1982/83 and 1 6% i n 1 983/84 ( Fi g . 2 . 14 ) . F'ubl i shed informati on on other anthomyi id seed predators i nd icates that mul t iple ovirosi t ion i s not unusual . Z i mmerman ( 1979 ) and ( 1980 ) recorded 1 . 6% mUl tiple i nfestati on in a Hylemya speci es that ovifXJsi ts i n the seedheads of Poli monium fol i ossi si mum Gray and 25% i n a speci es oviros i t i ng i n IfXJmopsi s aggreqata Pursh seedheads . 2 . 4 . 5 Distribution of seedf ly eggs The mean crowding of seedfly eggs , def i ned as the mean number of other i ndividuals per sample uni t per i ndi vidual ( IJloyd , 1967 ) , was calculated for seed f ly at each sampling date . I t i s not dependent on the number of buds considered. as fXJtent ial ovifXJsi t i on s i tes for seed f ly , whi ch was subjec t i vely est imated in thi s study, and was therefore pre ferred to a calclliat ion of the i ndex of d ispersion (Table 2 . 7 ) ( Southwood, 1978 ) . Accord i ng to Lloyd ' s methcd the d i stribut ion of seedfly eggs was uni form ( the ratio of mean crowding to mean dens i ty is less than one ) i n early December 46 . and again in mid to late January (Table 2 . 8 ) . Unfortunately estimates of mean densi ty are decreased by an overest imation of buds suitable as ov iposi tion sites for seedfly which increases thi s ratio of mean crowding to mean densi ty . Eggs o f Nezara v ir idula L. ( Heteroptera : Pentatomidae) which are found in ragwort seedheads do not appear to deter ov iposi tion by seed fly • They were found to)ether with seedfly eggs and larvae more often than expected i f the d istr ibution was random (Table 2 . 9 ) • 2 . 4 . 6 Pupae Table 2 . 10 shows that est imates of pupal densi ty frc:m anergence traps at the "Pines" and "Road" plots in 1983 were three times higher than those from the soil samples . Add i t ional soil samples in autumn 1983 gave an estimate of 11 ' 7 pupae/m2 compared to the average pupal density for the site in spr ing 1983 of 11 · 4 . Figure 2 . 15 gives the estimates of third instar larvae and pupae collected in the trays in autumn 1 9 84 at each sampling date . Sampl ing of pupae by record ing those found in the trays was cumulative and the density estimated for the s ite for 1984 was 670 , 44 6 per hectare . Sampl ing in this year coincided with or was immediately preceded by per iods of ra in whi le in 1983 a dry per iod in early March appeared to delay the dropping response of third instar larvae ( Figs . 2 . l6a+b) . T a b l e 2 . 6 N u m b e r o f e ggs i n f e ma l e s c o l l e c t e d a t R e d w o o d s F o r e s t 4 7 . Da t e c o l l e c t e d N u m b e r o f e g g s 1 3 / 1 2 / 8 2 2 7 / 1 2 / 8 2 1 0 / 1 / 8 3 2 4 / 1 / 8 3 7 / 2 / 8 3 2 1 / 2 / 8 3 T a b l e 2 . 7 Da t e 1 3 / 1 2 / 8 2 2 7 / 1 2 / 8 2 1 0 / 1 / 8 3 2 4 / 1 / 8 3 7 / 2 / 8 3 1 2 / 1 2 / 8 3 2 6 / 1 2 / 8 3 9 / 1 / 8 !;. , 2 3 / 1 / 8 4 / 6 / 2 / 8 4 1 6 , 1 2 , 0 , 2 , 0 , 1 , 0 , 2 , 0 , 0 2 9 2 5 , 2 , 1 6 , 0 , 0 1 5 . 1 9 , 1 5 , 0 , 0 0 , 2 9 , 1 6 , 1 7 0 , 0 , 0 , 0 , 2 G o o d n e s s o f f i t o f s e e d f l y e gg d i s t r i b u t i o n w i th t h e P o i s s o n d i s t r i bu t i o n ( S o u t hwo o d , 1 9 7 8 ) X 2 =: §J� F J2 X 2 � " d ::: -- x 4 4 2 . 1 4 8 5 - 1 . 4 2 0 9 4 . 6 2 3 4 7 - 3 . 8 * 2 7 9 5 . 6 3 0 0 3 - 2 . 7 1t 2 0 3 0 . 1 2 1 3 8 - 1 . 7 1 8 2 9 . 8 1 7 8 0 0 . 9 1 6 9 . 3 1 7 4 - 0 . 2 2 0 9 0 . 8 2 1 4 7 - 0 . 8 2 2 8 4 . 3 2 3 2 2 - 0 . 5 1 9 4 7 . 1 1 9 7 0 - 0 . 3 1 5 8 9 . 2 1 6 4 5 - 0 . 9 * i nd i c a t e s n o n - ra n d o m d i s t r i bu t i o n s T a b l e 2 . 8 Da t e 1 3 / 1 2 / 8 2 2 7 / 1 2 / 8 2 1 0 / 1 / 8 3 1 2 / 1 2 / 8 3 2 6 / 1 2 / 8 3 9 / 1 / 8 4 2 3 / 1 / 8 4 T a b l e 2 . 9 48 . D i s t r i bu t i o n of s e e ct f ly e ggs i n ragwo r t s e e d h ea d s ;\1 ean in G >1 e a n " 2 m E m e rg en c e Trap s n u m b e r f l i e s / N u mb e r s a mp l e s 7 . 8 7 0 6 . 7 2 0 4 1 . 9 3 5 4 8 . 6 2 0 o f 50 . F' i gu 1' e ? 1 3 I n c i d en c e o f mu l t i o l e in f e s ta t i o n s by s e e d f l v a t R e d wo o d s Fo r e s t (fJ E QJ +' (fJ 0 li\ h QJ 0., >, u ,:::: QJ ;:J 0" QJ h S'''' (/J E QJ +' (fJ 0 li\ h QJ 0, >;, () ,:::: (l) ;:J CT C) ):. ; '�L, 8 0 a ) 1 9 8 2 / 8 3 ') e g g s "- 7 0 2 l a r v a e 6 0 5 0 4 0 3 0 2 0 1 0 1 e g g t 3 e gg s �� ----- - � - ........... _ . . �. - ' . ' - ' - . . --- - 1 3 2 7 D e c 1 0 J an ? ' � 4 Sa mpl ing Da t e 1 l a rvae 7 F e b b ) 1 9 8 3 / 8 4 8 0 7 0 6 0 5 0 4 0 3 0 2 0 1 () , ) : j --- 2 e g g s 2 1 s t i n s t ar l arvae egg and 1 s t i ns ta r l a r vae - - 3 eg g s 1 1 s t i n s t a r - - - - a n d 1 2 nd i ns tar l a rva _ _ _ _ _ 2 2 nd i n s ta r l a r v a e 5 1 . i g u r e 2 . 1 4 C hang e s i n t h e i n c i d e n c e o f m u l t i o l e � s e e d f l y i n f e s ta t i o n s w i t h t i m e c: o TI ',1 11) +) ,:< ('j o +' cD (j) +' (j) +) c.-, cd c: (j) 'L! (j) cd ri (j) � .-e; ·rl 'tj +' Q) ri (j) :::l Ul E c.-, .-e; O +' ·rl (j) ): bD cd � +' ri c:: (-+-4 (j) 'L! o (j) H (j) (j) Ul n .. 2 0 1 5 1 0 5 \ \ \ \ \ \ \ \ \ \ \ 1 2 D e c - - - - - 1 9 8 / 8 3 1 9 8 3 / 8 4 \ \ \ "- "- "- "- "- " " .... .... .... .... .... " " 2 6 9 2 3 7 J an F e b Sa mI21 i n 0: da t e F i gu r e 2 . 1 5 E s t i ma t e s o f t h e n u m b e r o f s e ed f l v pu na e and 3 2 E 3 r d i n s ta r l a r v a e D o r h e c ta r e a t R e dw o o d s F o r e s t i n 1 9 8 3 / 8 4 3 rd I n s ta r l a r vae in a n t 3 r I n s t a r l a rvae in t ra y s P u pa e i n t rays r ) r, I ! \ 6 \ \ \ \ h cD � (f) ):; H QJ H co cD H � r<\ 0 QJ '+-! ..c o H H QJ ill 0.J .D E 5 2 . F i ;::u r e . 1 T h e n u m b e r o f s e ed f ly 3rd i n s ta r la r v a e a n d ra i n fa l l a t R e dwo o d s F o r e s t a ) 3 rd i n s ta r l a r v a e w i t h i n s e ed h ead s i n 1 9 8 2/ 8 3 5 4 3 2 26 9 2 3 D e c ,J a n 6 20 F e b Sa mpl ing da t e 3 rd I n s t a r la rva e w i t h i n s e ed h ea d s 6 1 9 �'l a r 3 A p r 4 5 3 6 2 7 1 8 9 E E rl rl cD <+-< ;;::: .,-j �\j cr' l� b ) 3 r d i n s ta r l a r va e w i t h i n s e ed h e a d s and l e a v i n t h p l a n t in 1 9 8 3 / 8 4 5 x 1 0 5 QJ cD > H 4 cD rl H cD � 3 (f) ;;::: .,-j QJ �" '(j Cd h +, 2 r<\ () Q.\ CH �c: 0 �. h OJ QJ '-1 , .D E� ;:J Z 3 r d I n s t a r l a rv a e w i t h i n s e e d h e a d s _ _ 3 I n s t a r l a rva e l ea v i n fJ' ,) t h e p l a n t , 1 4 5 3 6 2 7 1 8 9 A n r ,s::; � .,-j OJ rl � � cD E E rl rl cD <+-< ;;::: .,-j Cd 53 . 2 . 4 . 7 Germination The nunber of seeds ranain ing in infested seedheads var ied between each sampl ing site (Table 2 . 11) and the mean germination success rangerl from 28-43% . In crld ition , the mean mrnber of seeds (±S E ) i n uneaten seedheads of the same plants was 29±2 · S ( n=38 ) and rangerl fran 2-58 . The mean percentage germination of these seerls (±S E ) was 72 - 7±2 · 8% . As all plants were infested , germination of ragwort seerls could not be canparerl wi th See:lS fran plants not subjected to seed predation . 2 . 5 Resul ts fran the Desert R3 si te Only 16 fl ies were caught in emergence traps in 1982/83 and 5 in 1983/84 am no fl ies were caught on sticky traps . Consequently l i ttle is known about the pre-ov iposi tion period at this site . Ragwort density at the s ite was 0 - 35 stems/m2 in 1982/83 and 0-23 stems/m2 in 1983/84 and all plants were single stemmerl . The seerlfly l i fe h istory and ragwort flower ing was delayed in the colder cond i tions at the Desert RJ . A canparison of the temperatures recorded at the two sites is presenterl in Appemix 2 . OV iposition and flower ing occurred five weeks later than at the Redwoods Forest s i te (Figs . 2 . l7a+b) . The overal l infestation level s at the s i te were 1% in 1982/83 and 3% in 1983/84 (Fig . 2 . 18) but the small sample s i zes resul ted in large errors . Therefore any apparent trends in the infestation level s at th i s site may not reflect the true si tuation . Tabl e 2 . 1 1 Whakamaru A u tu mn 1 9 8 2 D e s e r t R o ad A u t u mn 1 9 8 2 R e o o d s Fo r e s t A u t u mn 1 9 8 3 N u m b e r o f ragwor t s e e d s and th e i r v i a b i l i ty a f t er s e ed f ly a t t a ck N o . o f s e e d s r e ma in in g p e r s e e dh ead x ::: 1 7 . 6 S E ::;: 1 . 4 r an g e ::;: 0 - 5 5 n ::: 1 0 1 - 7 . 6 x ::: SE :::: 0 . 7 ran g e ::;: 0 - 4 2 n ::: 1 2 3 - x ::: 1 3 . 2 SE 0 . 5 ran e ::: 1 - 3 3 n - 3 9 P e r c en ta g e g e rm i n a t i o n - A SE n - X SE n - X S ]:;' ' LJ n ::: 4 3 . 3 ::: 2 . 8 ::: 9 5 ::: 28 . 0 ::: 3 . 0 == 1 0 2 ::: 3 0 . 1 :: 4 . 4 -- 3 9 5 4 . , i gu r e 2 . 1 7 C u mu l a t i v e p e r c en t ag e c u rv e s of s e ed f l y a n d ragwo r t s e edh e a d d e v e l opment s t ag e s a t t h e D e s er t R o a d . . u r v e s f i t t e d b y e y e ; h a t c h e d a r ea r ep r e s en t s r e g i on o f wh e r e s e e d h e a d s a r e a v a i l a bl e f o r o v ipo s i t ion by s e ed f l y a ) 90 C\l 80 0 dJ 70 .c � 60 0 dJ OJ 50 C\l C dJ 40 u >- Q) Q 30 dJ > � 20 C\l :::J E 10 :::J U 1 9 8 2 / 8 3 b ) Ov e rlay 1 9 8 3 / 8 4 1 0 2 4 7 2 1 7 2 1 4 J a n F e b i1 a r A p r S a m21 ing da t e In U1 . F i gu r e 2 . 1 7 C u mu la t i v e pe r c en tage c u r v e s o f s e ed fly d ev e l opmen t s tage s a t t h e D e s e r t R o ad C u r v e s f i t t e d b y e y e ; ha t c h ed a r ea r e p r e s en t s r e g i on a r e a v a i l a bl e f or o v ipo s it i o n by s e edfly a ) 1 9 8 2 / 8 3 b ) O v e r l a y 1 9 8 3 / 8 4 <1l 0 Q) £ � 0 Q) OJ <1l C Q) () � Q) Q Q) > - <1l ::l E ::l U 30 20 10 1 0 2 4 J an 7 2 1 F eb 7 2 1 l� a r S a mpl ing da t e 4 A p r a n d ragw o r t s e e d h e a d ./' Q o f wh e r e s e e d h e a d s .J'l =, .> r r· � -£. 1 - V> -J � ..r R l 3 ..l. 0 c:.. � s: -( S: � V\ -l r (jJ � t VI V1 -..( . \...J F i gu r e 2 . 1 8 c 5 0 � ·c-l +' (Ij � (!} 4 0 QJ 4-; c ·rl QJ 3 0 Qj) cD � C QJ () ") Ii ,� �" ,. \1' ,-:w 1 0 56 . P e r c e n t a e o f ra wo r t s e edhead s i n f e s t e d b s e e d f ly a t t h e D e s e rt R oad s i t e Ba r s r epr e s ent s tandard e r r o r s ) 1 3 2 7 D e c 1 0 2 4 J an - -I - "- " "- "' 7 2 1 F eb Sampl ing d a t e - - - - - 1 9 8 2/ 8 3 1 9 8 3 / 8 4 7 2 1 4 A p r M a r 3 . 1 Introduction - - - - -- ---�-- �---- - CHAPTER THREE : LABORATORY srUDIES Field data from spr ing 1982/83 and 1983/84 indicated a 57. pre-oviposi tion period of approximately six weeks . However , there was an absence of ragwort flowers during thi s period which suggested that there was no nutr i tional requirement for ovary development associated with ragwort flowers and that ragwort rosettes and/or other f ield plants prov ided food for seedfly adults . There was also a possibi l i ty that ovary development could be slowed or delayed by the absence of ragwort f lowers . To determine the factors affecting seedfly ovary development rates fl ies were reared at d i fferent temperatures in the absence and presence of var ious d iets and stages of ragwort development . Surv i vorship curves for eggs am larvae w i thin seedhecds were constructed from their development rates measured in laboratory condit ions . Factors affecting the stage when third instar larvae leave the plant to pupate were also investigated in v iew of the var i able durat ion of the thi rd larval instar observed in the f ield . Further estimates of pupal surv ival over winter are presented with some information on pupal d iapause . 3 . 2 Methods and Equipnent 3 . 2 . 1 Ovary development Fl ies were reared in cages in controlled cond it ions to determ ine the effect of d i et and presence of the host plant on ovary development . The cages were 1m high by 70cm x 70cm and stood on 58 . moveable trolleys 700m above the ground (Plate 6) . The frames were constructed of 6mn gauge al unini urn rod f i tted into ( 20m) 3 perspex corner blocks . The frame was canpletely enclose] in fly proof nylon mesh and access was provided through a flap sealed with velcro . D iets were provided in 10 ml bottles which protruded through rectangular polystyrene trays pos itioned at one of the top corners of each cage . Cotton wicks were inserted through the screw tops of the bottles to d ispense the diets . Female flies recaptured from the cages were preserved in Carnoy ' s solution (Humason , 1967) . When the females were d issected the lengths and widths of thei r ovar ies were measured us ing an occular m icrometer am the number of eggs with a chor ion were r ecorded. These were eggs which remained intact when placed for a few minutes in a solution of 50% acet ic acid . Squashes were made of the spermathecae , examined umer a phase contrast m icroscope and the presence of sperm recordEd . I . Exper iments at D .S. I . R . Two thousand seedfly pupae were col lected fran Redwocx:1s Forest i n February and August , 1983 by the flotation method descr ibEd in Chapter 2 . The pupae were divided into lots of 50 and kept in 200ml T .V .L col lect ing j ars in moi st , ster i le vermicul i te . These were kept at lOC until the end of September and then transferrEd to 20C . Many of the fl ies which emerged had Wlng deform it ies . Only 700 adults were perfectly formed and 4 24 of these were females . 375 59 , o f the females surv ived anaestheti zation with 002 and were released into the cages wi th the males . They were allocated to the d ifferent treatment accord ing to the number emerg ing dai ly and the total number expected to hatch . Because daily emergence was low, individual females were marked so that they could be aged to the nearest day . This was done by awlyirg quick dryirg Modelair dope paint to the dorsal surface of the thorax with a fine needle . Markirg was avoided where possible and no males were marked . The flies were then released into the cages in the controlled env ironment roans at plant Phys iology Div ision , D . S . I . R. , palmerston North . Temperature treatments for these exper iments were maintained constant over the l ight and dark per iods at 10 , 15 , 20 and 25 (±O o S) C . The vapour pressure def icit was also mainta ined constant , at 5mb, resulting in relative humid it ies of 35 , 53 , 66 and 75 (±5 ) % respectively. The photoper iod was 14 hours l ight to 10 hours dark and the irradiance ( 400-700nm 2 waveband ) measured at plant he ight was between 146 and 149 W/m • The l ighting system used consisted of four lOOOW Sylvania "Metal-arc" high-pressure d ischarge lamps, and four Phi l ips tungsten iodide lamps (Warr ing ton et al . , 1978 ) . Carbon d iox ide levels measured with an infrared gas analyser dur ing the course o f the exper iments were 330�20ppm . The fl ies were prov ided wi th f lower ing ragwort potted in a so i l/pumice mix with ferti l i ser . These plants were watered da i ly ( 200mls) v ia an automated m icrotubule system . The fol lowing d iets based on the anthomyi id d iets dev i sed by S ingh ( 1977) were 60 . provided • 1) 70mls water , 4gms yeast , lOgms sugar , 20mls skim mil k , 2mls casein peptone (4gms/l) . 2 ) 5gms sugar , 250mls water , t� egg whites . 3 } Golden syrup : skim m ilk : water , 1 : 1 : 2 ratio by volume . 4 ) 30ml s honey, 30ml s condensed milk, 300ml s water . Plants were also lightly watered every three days to provide surface water for the fl ies . Additional experiments were conducted at 20e where fl ies were provided with ; a ) ragwort rosettes only b) flowering ragwort only I I . Exper iments at Massey Uni vers i ty Five hundred pupae obtained fram Entomology Divis ion , D . S . T . R. in November 1984 were used for exper iments in the controlled env ironment rooms of the Department of Agr icul ture and Horticulture , Massey Univers ity . These had been obta ined in autunn 1984 by inverting the seedhecrls over moi st sand . The pupae were subsequently collected us i ng the flotation method . They were stored in mo ist verm icul i te at lOe and transferred to 20e in early June to ini tiate development . As the wings became visible through the pupar ium the pupae were r eturned to LOCo When all the pupae had reached th i s stage they were transferred to 20e to complete development . Th is ensured that large numbers of f l ies emerged together so these d id not have to be mar ked . The f l ies were 61 . allowed to hatch d irectly into the cages to further reduce handl ing • Temperatures in the controlled environment rooms were maintained at 12 , 15 , l8C and the humidi ty ranged from 65 to 75% . The photoperiod was 14 hours l ight to 10 hours dark and the l ight sources consi sted of 2 Phi l ips 1000W tungsten halogen lamps, 6 Phi l ips 375W mercury iodide lamps and 3 lOOW incandescent bulbs illuminating an area of 4m2 • A range of cut flower ing f ield plants found at the field site dur ing the pre-ov iposi ti on per iod were prov ided in add i tion to ragwort rosettes and the d iets descr ibed above . These included cocksfoot , Yorkshi re fog , dandel ion , buttercup ( Ranunculus acr is L.) and paspalum ( Paspalum d ilatum Poir) . Fi fty of the 70 females that emerged in per fect condition from this consignment were reared in these cond i t ions and the remaining 20 were used to compare the ovary size of females reared on add itional d iet and ragwort flowers with those reared on ragwort rosettes and d iet . 3 . 2 . 2 Egg and larval development Seedfly eggs and larvae were reared in the controlled env ironment rooms at the D . S. l . R . to measure the ir development rates at d i f ferent temperatures . prov is ion of host plants Five hundred ragwort rosettes were dug from the f ield in spr ing 1983 am potted in two 1 i tre pots in a so i l/pun ice m i x w i th 6 2 . fert il iser . Th is ensured that a supply of buds su itable for ov ipos it ion by seedfly would be ava ilable when gravid females were present at the f ield s ite in December and January. The plants were kept in an open glasshouse and watered regularly. They were sprayed two months before the start of the exper iments with a non-systemic insect ic ide, ' Attack ' , (47 0 5% pirliniphos-methyl and 2 · 5% permethr in) and the fung ic ide ' Manzate 500 ' Dupont (act ive ingredient 80% mancozeb). Slug k iller was d istr ibuted throughout the pots. On several occas ions plants were temporar ily installed in the controlled environment rooms at 20 and 25C to accelerate bud development . Cond it ions in the controlled env ironment rooms These exper iments followed those on ovary development at plant Physiology Div is ion , D .S . l .R , Palmerston North and the conditions descr ibed above were continued . Development was completed quickly at 25C so this room was converted to a temperature regline which fluctuated between 25C and l 5C for comparison with the development rates at constant 20C. There was a six hour changeover from day/night and night/day cond itions , the l ights go ing off two hours before the complet ion of the changeover to night cond itions ( 15C) and coming on two hours after the start of the changeover to day ( 25C) (Fig . 3 . l) . The relative hum id i ty was 75/53% . Obtaining eggs for the exper iments Potted plants with buds sui table for oviposition were transported to the field site at Redwoods Forest . Several tr ips were made transporting 60 plants each time to m inimize the number of plants 6 3 . d i scarded when insuff icient eggs were laid . The plants were left in the field for 24 hours starting at midday. They were watered in the field the fol lowing morning to prevent wilting . Samel ing procedures in the controlled env ironment roams After the six hour return tr ip to palmerston North the plants were placed in the controlled env ironment roams and connected to the automated watering system. They were not watered externally dur ing the experiments . The following day a sample o f boos was d issected to determine the infestation level of seedfly and the total number of eggs in the sixty plants was estimate:] . When infestation level s were below 3% the plants were d iscarded because they d id not provide suffic ient animals to sample destructively through to pupation . SUbsequent samples were taken every 12 hours . Sampl ing continue:] each time unt il at least 20 animals had been found . Fl ies were allocated to the controlled env ironment roams as summarized in Table 3 . 1 . One see:3fly population was placed at 25C first . Al though there were only suff ic ient numbers of fl ies for 12 hourly sampl ing at the egg arrl f irst larval instar stages this gave an est imate of the shortest duration of the l i fe history that could be expected and reduced the numbers of fl ies sample:] from populations developing at cooler temperatures . Fl ies placed at lOC were d i scarded when i t was real i sed that the projected duration of the egg and larval stages was 143 days and thi s exceeded the time allocated for these exper iments in the controlled env ironment rooms . 64 . Third instar larvae that dropped to pupate were collected dai ly fram l25rnl potties at the base of cardboard cones f itted around several stems ( usually 4-5 ) bound together ( Plate 7 ) . Each pottle was covered in black ta{:e am f i lled with moi st vermicul ite . I t was unscrewe::1 fram the base of the cone each day so that the larvae which had dropped could be transferred to a separate container to pupate . Pupae in thi s conta iner were measured and transferred daily to controlled tem{:erature cabinets to continue development at the ir respective temperatures . 3 . 2 . 3 The larval dropping response Time of day of larval dropping Infested stems collected from the field were suspended in the laboratory over trays l ined with black polythene. The temperature was maintai ned at constant 20C with a photoperiod of 14 hours l ight am 10 hours dark . The daily dropping of larvae in response to simulated rain Third instar larvae reared in fluctuating temperatures that rEmained in the seedheads after the peak in droppirg were subjected to following treatments . 1 ) Single drench , one l i tre of water , us ing a water irg can . 2) Light spray, 500ml s of water , us ing an atomiser . 3 ) Light spray daily, 500mls of water , using an atamiser , followed by daily drench , one l i tre of water , us ing a water ing can . The number of thi rd instar larvae d roppirg da ily was recorded for each treatment . P l ate 6 P l a te 7 C a ge for rearin g adu l ts i n t he c o n tro l led en v iron men t ro o m a t D . S . l . R . T he arrangemen t f or d ispensing the d iets is a t t he t o p r i g h t ­ hand c orner o f t he c a ge . C o l lec t i ng funnel f or t h ird ins t ar l arvae lea v i n g ragwort i n t h e c ontro l l ed en v iron men t ro o ms a t D . S . l . R . 65 . F i gu r e 3 . 1 T e mpe r a tu r e f l u c tu a t i o n s i n r el a t i on t o £ho t ope r iod i n t h e c o nt ro l l ed en v i r on m en t r o o m a t P lant P hy s i o l ogy D ivi s ion D . S . I . R . 2 5 u 2 0 Q) H ;::l 1 5 +> m H Q) 0, 1 0 os CD E-< 5 o 6 . 0 0 Dark 1 2 . 0 0 L i h t T i m e ( H o u I S ) 1 8 . 0 0 2 4 . 0 0 D a rk T a b e 3 . 1 A l l o c a t i on o f s e e d f ly eggs t o the c o n t ro l l e d e n v i r o n m en t r o o m s a t D . S . l . R . J a t '" h en e ;z: s E s t i ma t ed T e mp e ra t u r e e r e i d n in f e s ta t i o n o f r o o m e g g s p l a n t p l a c e d i r1 l e v e l s '"J e r e a l l o c a t e d t e f j J t o \,1 . ' on 1 / 1 2 / 8 3 3 % 2 5 c t ''J n o o n 1 1 / 1 2 / 8 3 0 0 1 7 / 1 2 / .8 3 < 3 % D i s c ard e d t o n o o n 1 8 / 1 2 / 8 ] o o n 2 6 / 1 2 / 8 ] 1 0 % 1 0 c t c n o o n 2 7 / 1 2 / 8 3 ;1 0 0 n 2 / 1 / 8 4 2 0 % 1 5 t n o o n 3 / 1 / 8 4 o o n 9 / 1 / 8 4 2 5 % 2 0 c t o n o o n 1 0 / 1 / 8 4 o o n 2 3 / 1 / 8 4 2 5 % 1 5 - 2 5 c ( a v r a e t o n o o n 2 1+ / 1 / 8 4 2 0 C ) '--- N u m b e r s o f ;l u mb r o f s e e d f l y s e e d f l y c a u 2: "-<- d e s t ru c t i v e l y fu n n e l s s a mp l ed 66 1 5 N o t c o mp l e t e d b e c a u s e o f t i m e r e s t r i c t i o n s 4 8 0 i, 5 6 7 1 9 4 5 3 0 3 5 1 p r i c t o d r a p - i n g e xp e r i m e n t s 6 7 . 3 . 2 . 4 Diapause Pupae col lected fran Wai rakei and the Desert R::l in March 1982 ( F i g . 2 . 1 ) using the flotation method were kept in lots o f 10 and 4 respectively in moist sand within 250ml T.V. L specimen j ar s . They were subjectEd to -15C and 5C for v aryifB lengths of time and then placed at 20C to continue development . 3 . 3 Results and Discussion 3 . 3 . 1 OVary development Mortal i ty of caged fl ies El'nergence of perfectly formEd adults from the pupae collected in winter 1983 is shown in Figure 3 . 2 . This has the same scale as Figure 2 . 7 to compare the spread in emergence in the laboratory and f ield populations and shows that males emergEd sl ightly before females . Mortal ity in the cages could not be accurately determined as few of the dead fl ies could be found among the potted plants . It is not known whether reduced handl ing increased survival as no d irect compari son of mortal ity can be made between the experiments at D . S. I .R and at Massey because of the d iffer ing ways in which the populations were samplEd . Fi fty of the 375 females released into the cages at D . S . I .R . were culled at var ious ages for d issection while 13 of the 50 fl ies reared at Massey University surv ived to maturity . Twelve of the 20 females used to determine the effects of ragwort flowers on ovary development were available for d issection after 14 days at 12C. The only publ ished information on mortal ity in anthanyi ids is prov ided by Theunissen ( 1974 ) who found a higher mortal i ty in caged on ion fly males ( 59% mortal ity at l5C after 20 days) 68 . compared to females ( 25% ) . Maturation of seed fly ovar ies All fl ies reared on ragwort rosettes without additional food d ied within five days. Figure 3 . 3 shows the effects of temperatures and d iet on ovary si ze in the controlled temperature rooms at Plant Physiology Div i sion , D .S . l . R. The si ze of ovaries from flies which had 'been marked were 70% that of mrnarked ones and the measurements which have been adj usted to allow for marking are ind icated by arrows . Females could not be reared to maturi ty on ragwort flowers alone and after 14 days at 20e ovary size was 40% that of those reared with add i t ional d iet . There was no significant d i fference in ovary si ze after 14 days at 12e between flies reared with flowers and diet and those reared with rosettes and d iet ( t= o 39 P= · 72 d f=lO) • The effect of d iet on ovary developnent has been invest igatoo for several anthamyi id spec ies . Hafez et al . ( l970) found that d iet was an important determinant of the len:lth of the pre-ov ipos ition period in laboratory rearoo beet fl ies while add itional protein was requirEd to mature more than one batch of eggs in the cabbage root fly (Finch , 1971 ) . In subsequent stud ies (Finch , 197 4 ) found that the surfaces of pollen of Yorkshire fog and cocks foot together with tal l oat grass , Arrhenatherum elatius L . , prov ided suffic ient nutr i tive carbohydrates for maturation of the f irst batch of eggs in the cabbage root fly. Plantain , blackberry , dandel ion , white clover , w i ld cherry, marsh mar igold and sting ing nettle ( Ur t ica d io ica L . ) were al so foed sources for the cabbage 69. root fly but the last three were not available for the ragwort seedfly at e i ther of the f ield si tes , nor were the Umbell i fers , Anthri scus sylvestr is L . ( cow parsley) and Heracleum �hond1 ium L. (hogweed) . The latter prov ided the most nutr i tious nectar for cabbage root fly adults w ith 94 and 86% sugar solutions respectively ( Finch , 1974 ) . Studies on ovary development of the cabbage root fly ( Finch , 1971 ) and the onion f ly ( Theunissen, 1974) have shown that eggs are laid in batches and that all eggs in one batch r ipen together . OVary development rates for R .jacobaeae were therefore calculated as the reciprocal of the time fran energence to the first appearance of an egg in the med ian ov iduct ( Plate 8 ) . The development rates of fl ies reared with add i tional d iet both in the presence and absence o f ragwort flowers are g iven in Figure 3 . 4 . These resul ts do not include marked fl ies . The duration of the pre-oviposi tion per iod decreased with increasing temperature and was approximately 39 days at 10C and 14 days at 20C. In canparison , the pre-ov iposi t ion period of the beet fly was 8 days at l5C and 3 · 5 days a t 30C ( Hafez et al . , 1970) ; 7 days at roam tenperature for R.seneciella ( Fr ick , 1969 ) and five weeks in the field for the wheat bulb fly (Jones , 1970 ) . The numbers of eggs in R . j acobaeae fenales reared in the controlled env ironment roans ranged fran 2-30 (Table 3.2) . There were 16 ovar ioles in each ovary and thi s was the same for the wheat bulb fly (Gough , 1946 ) . The max imum number of eggs laid by the '#heat bulb fly in the laboratory was 180 (Long , 1958b) and Raw 70. et al. (l968) found that females of th is species la id an average of 40 e:;1gs which represented those of the first batch and part of the secorrl. Females were mated one day after emergence at 25C (Table 3. 3 ) but it is not known how often th is �urs. Although examination of spermathecae squashes ( Plate 9) showed if the fl ies had been mated dur in:J the ir l ife time the ir age when mated could not be determ ined. A female was fourrl to have been mated w i th in 2 days of emergence at lOC so it is unl ikely that cooler temperatu res extend the pre-mating per iod in �. jacobaeae as observed by Hafe z et al. ( 1970) for the beet fly. They found that male beet fl ies were capable of mating four females and females copulated more than once. The only other relevant data publ ished on anthany i id mating behaviour is fram Jones ( 1970 ) who reported that wheat bulb fly males each served several females and by the time ovar ies were half develop=rl the spermathecae con ta ined spermato zoa. 3. 3. 2 Developuent of eggs am larvae w i th in the seed head The duration of the egg, and first and second larval instars was taken as the time between 50% of the population enter ing a s tage and 50% leav ing the stage (F ig. 3. S ) . At lSC the durations of the egg s tage , f irst and second larval instars were 5 - 5 , 7 arrl 8 days respectively. The duration of the third larval instar was calculated as the mean time fran ov iposi tion to larval dropping m inus the mean time fran ov iposi t ion to when 50% of the population had entered the thi rd instar . This was estimated as 31· 9 days at 15C and 23 - 5 days at 20C. Fi gure 3 . 6 shows the development rates 7 1. F i gu r e 3 . 2 Emerg en c e o f s e e d f ly adu l t s i n t h e l a b o r a t o ry Pupae w e r e c o l l e c t ed i n A u gu s t 1 9 8 3 and kept a t 1 0 C un t i l O c t o b e r 1 9 8 3 when t h ey w e r e t ran s f e rred t o 2 0 C b.O ;;::: ·rl 4 0 � 3 0 r- Q) e Q) U) Q) ·rl H Q) ,D E ;::l z \ t II II 1 1 1 1 �I 1 , I I \ \ \ ma l e s f e ma l e s - F i gu r e 3 . 3 T h e r e l a t i o n s h i n b e t w e en o v a ry s i z e , age and d i e t o f s e ed fly f e ma l e s P o i n t s i nd i c a t ed b y a r ro w s h a v e b e en i n c r ea s e d b y 4 0 % t o c o r r e c t f o r m a r k i n g s . �, �,� +' 0 � 7 (,-; 0 ff). 6 ..c: ,, ) \j ' ' � c:. ?:---. � (lJ E \j E '" c:� rh cD Ul i. >O Ul Q) H ..c: '" cD +' H > bn cD 0 c: > 3 Q) 0 rl Q) cC: 2 +' (,-; 0 E ;::l Ul r I l- I l I I I f- � I ----r------ r ---- -,---� <> o o a a )( 0 days 0-- ---tIl - - - IOC f lowe rs 1- d i e t 15C - - + 20C Ao 6> 20C f lowers only +- - - - + 25C f l owe rs 1- d i e t j _L __ � �___ � _ __ -1... __ _ _ __ 1. ____ "--___ -' 2 0 2 5 3 0 3 5 A g e o f f l i e s ( d a y s ) 73 . F i gu r e 3 . 4 The r a t e o f s e edfly o v a ry d e v e l opmen t in r e l a t io n t o t empe ra tu r e +' +' [J} 0 >-, :::l .r! '(j c.-, .r! > o 0 +'l :::: ill cO o ·ri e re ill Q) bJ' E >-, � � n - � ;:: ,C U • l) 6 + :. t,-4 bi ' he Q) Q) E " r-1 :::: Q) cd u \.) c o CD i--; H Q. cO ·ri Q) U 0-, Q) Q. 0:: cO Y = -0. 01 78 + 0.0043 x 5 1 0 1 5 T e mp e r a tu r e /' /' /' /' - - - - F i t t ed b y + R ea r e d w i t h r a g ­ w o r t f l o w e r s & d i e t p r e s en t 6. R ea r e d w i t h ra g ­ 'wo r t r o s e t t e s & d i e t & a d d i t i o n a l f i e l d p l a n t s p r e s en t 2 0 2 5 ( C ) 8 r e d u c t i v e o r s f r o m a 1 s e e d fl y . e a r e d a t 2 0 C i th and d i t i on a l d i e t . (Mag n i f i ca t i on 40 x ) o ov a ry s p s p a t h e c m . o m ed ian o v i d u d Y o l d rag lvo r t f e ma l e f l o \v e r s S e e d f l y s p e r ma t h e c a s q ua h d t o s how s p e r m . Da r k a r as i n d i e a t e fr m en s o f t e s p e r mat e e a . ( M agn i f i c a t ion 320 x ) Tabl e 3 . 2 Total number o f s e edfly eggs fro m s e edfly f e ma l e s r ea r e d in the l aborat o ry Cond i t i on s i n A g e ( days ) N o . eggs c on t r o l l ed en v i ro n men t when with a r o o m s d i s s e c t ed c hor ion Ro s et t e s & d i et - 1 2°C 3 0 6* 1 S o C 1 6 2 2 1 1 6 �� 2 1 S 2 1 1 2 2 1 2 1 8 ° C 1 6 1 8 1 6 3 �( 1 6 2 0 1 6 1 0 1 6 6 1 8 " 9 x 1 8 8 Fl ow e r s & d i e t 2 0 ° C 1 3 2 1 3 1 9 1 3 2 1 3 3 0 1 4 1 9 �( 2 5 ° C 1 4 1 8 '" 1 6 6 " " * i n c l u d e s e g g f o u n d i n t h e m e d i an o v i d u c t . 74 . 75 . Tabl e 3 . 3 Mated s t a t e o f f e mal e s reared in labo rat o ry c on d i t i o n s a t D . S . I . R . S o me f e ma l e s \.J ere ma rked w i t h m odel a eroplan e " do p e " t o i n d i c a t e when t hey e m erged . T reat ment A e, e ( Days ) Marked ( + ) P r e s en c e Unma rked ( - ) o f s p e rm 1 0 C 2 + + Flowers & D i et 3 + - 4 + - 8 + + 8 + + 3 5 + - 1 5 c 2 - - Flowers & D i et 4 + - 8 + - 8 + - 1 2 - - 24 - - 20 C 2 + + Flowers only 2 + - 3 + - 3 + - l± - + 2 0 C 3 + - D i e t only 3 + - 3 + - 2 0 C 2 - - Flowers & D i e t 2 + - 2 + - 3 + - 4 + - 7 + - 8 + - 8 + - 1 2 - + 1 3 -. + 2 5 C 1 + - F l o w e rs & D i et 1 + + 1 + - 1 + + 2 + + 2 + - 2 + - 2 + - 2 + + 3 + - 3 + - 3 + - 5 + + 5 + + 7 - + 7 - - 1 0 + - 1 0 + - 76 . of seedfly within the seedhead fram oviposition to larval dropping in relation to temperature . Developnent rates fram CN iposi t ion to pupation were al so measured (Fig . 3 . 7 ) . The mean duration of the third larval instar after it had left the seErlhead calculated fram these two graphs was 3 - 2 days at 15C. Developnent rates of seedfly at constant 20C were similar to those under fluctuating temperatures with a mean of 20C. Developnent rates under fluctuating and constant temperatures have been also measured for other Diptera with var iable results . Wilkinson and Daugherty ( 1970) found that developnent tlines for eggs of Bradysia impatiens (Johannsen) (Diptera : Sc iaridae) were slinilar under constant and fluctuating temperatures but larval development was shorter under fluctuating temperatures . Larval mortal i ty , however , was higher under the var iable temperatures . Sidd iqui and Barlow ( 1972) found that developnent tlines of the immature stages of Drosophila melanogaster (Meigen) (Diptera : Drosophil idae) tended to be shorter at al ternating temperatures than at the corresponding constant mean temperature . Surv ivorship curves (Fig . 3 . 8 ) were constructed by integrating the curves for each stage in Figures 2 . 9a+b and d iv id ing by the duration of the stage at 15C ( Southwood , 1978 ) . This was the average temperature exper ienced at -Redwoods Forest dur ing the sumner . It was estimated that 1 , 966 , 358 eggs survivErl to halfway through the egg stage at Redwoods Forest in 1982/83 and 5 , 761 , 447 in 1983/84 . Therefore , assuming no mortal ity in the first hal f of the egg stage , 12 - 9 eggs were laid per female that emerged at the il) C\l > '- C\l '- C\l il) til bJj .� Cti +> u CI) C N .c: 0 CJ Cti il) Q) C\l > c.-. '- 0 !E '- Q) C\l to - Cti � +> .- c Q) til () h >< Q) 0-. til OJ OJ il) , aj Ct-i '0 0 0 . 0 4 t\D 0 . 0 3 Q) s:: S ·r! .r! p, +' 0.. o Q) H ..c: 'O +' 0 . 0 2 o c.-. +' o s:: ri O aj .r! O . 0 1 U +' O ·r! H (f) Q) P-< O aj .r! P-< > U ·r! H Q) > aj O:: O ri e r r o r s ) 5 1 0 I y = - 0 . 0 0 4 9 7 + 0 . 0 0 1 5 9 x 1 5 20 2 5 T e mp e ratu r e ( C ) F i gu r e 3 . 7 The rat e o f s e e dfl 78 . to u pa t i o n in r e la t i on to Ba r s E o H c.-. s::: Q) 0 E .r! .r! +' +' cO P., Q) ::l ..c: o.. +' o c.-. +' o s::: ri 0 C\J ·rl 0 . 0 4 0 . 0 3 0 . 0 2 u +) 0 . 0 1 O 'rl H (f) 0. 0 .r! P-< U ·r! Q) > � o ind i ca t e s tanda rd e r r o r s I y = - 0 . 0 1 3 7 + 0 . 0 0 2 0 7 x r ., 1 0 1 ') 2 0 r C 1: 1 p e r .1 t i l r 8 ( C ) 2 5 H Q) A r rows i nd ica t e 0 . unde r e s t i ma t e s o f QJj s::: 1 s t i n s t a r l a r vae .r! > See pag e 80 .r! > H ;:l (f) CfJ Q) .,-1 rl Q) c.-. H cd c.-. � o u Q) H ..s::: Q) ..0 E ;:l ?: F i gu r e 3 . 8 Su r v i v o r s h i o o f s e e d f ly a t R edwoods Fo r e s t T h e p o i n t s w e r e c a l c u l a t ed u s in g t h e m e t h od o f S o u t hw o o d ( 1 9 78 ) and t h e cu r v e s a r e f i t t ed by e y e . \ 8 6 4 + l' 1 9 8 3 / 8 4 x... ... ... 2 -- .... -- -- - - - -lit-- _ - - - - - - - - - - - --- 1 9 8 2/ 8 3 5 1 0 1 5 20 2 5 3 0 3 5 4 0 4 5 A g e ( Day s ) 80 . site in 1982/83 and 3 - 9 in 1983/84 . First instar larvae were d ifficult to locate in the d iscolourerl , thawerl seerlhead samples and as a result they were urrler-estimated . The surv ivorship curves show that an estimated 88% of the fl ies surviv ing to hal fway through the egg stage surv i verl to hal fway through the secorrl larval instar at Rerlwooos Forest in 1982/83 canparerl to 69% in 1983/84 . Mortal i ty fran hal fway through the secorrl larval instar to halfway through the third larval instar was 4% in 1982/83 and 31% in 1983/84 . In autumn 1984 , the number of thi rd instar larvae within seerlheads is estimaterl as 370 , 000/ha (Fig . 3 . 8 ) comparerl to 1 , 580 , 500 third instar larvae per hectare estimaterl fran the trays in this year . This can be canparerl to 670 , 450 pupae/ha estimated from the trays (Chapter 2 , Section 2 . 4 . 6 ) • There was no significant d ifference in the si ze of pupae from larvae rearerl at constant 15e arrl constant 20e ( t=- l o 53 P=O o13 df=51- 9 ) but larvae r,eared at constant 20e producErl significantly larger pupae than those rearerl in fluctuating temperatures with a mean of 20e ( t=2 095 P=0 ·0036 df=204 - 7 ) • 3 . 3 . 3 Larval dropping No daily rhythm associated with photoper iod was detected in the numbers of third instar larvae leav ing cut stems to pupate ( Append ix 3) . The possibil ity ex ists that humid ity , which was closely l inked to the temperature ma inta ined in the controlled env ironment roans , 8 1 . i nfluenced the larval dropping response . Fi gure 3 . 9 shows that the number of larvae dropping increased wi th increasing humidity and temperature in the rooms . Drenching plants w i th water from a watering can resulted in a marked increase In larvae dropping ( Fig . 3 . 10 ) . Da i ly drenching resulted in continued high numbers dropping unti l all larvae had dropped . Light spraying wi th an atomiser increased the number droppi ng but not to the same extent as heavy watering . 3 . 3 . 4 Pupal survi vorship and termination of diapause Overwintering survival could be calculated for the 1983 winter usi ng the information on survi vorship of larvae w i thin ragwort seedheads . 35 . 3 thi rd instar larvae/m2 wi thin seedheads were estimated in autumn 1983 compared to 30 adults/m2 est imated to have emerged in the following spr i ng ( Chapter 2 ) . Summer diapause in �. jacobaeae was ini t iated when the temperatures were between 1 5 and 20C . I t i s therefore unlikely that seed fly pupae diapause in the f ield as temperatures in the soi l during autumn and winter drop below 1 5C. Adults emerged from pupae subjected to constant 15C after 209 days but did not emerge from pupae placed at constant 2OC . Temperatures requi red to terminate diapause were one week or less at 5C and one minute at -15C . Development , however , was not arrested at 5C . When the t ime spent at 5C was increased the time spent at 20C unt i l development was completed decreased but i t is not clear whether thi s occurred in pupae placed at -15C ( Fig . 3 . 11a+b) . There was no s igni ficant di fference in development time at 20C between pupae collected at Wai rakei and 82 . fran the Desert R:l ( t=-1 · S2 P=0 · 09 df=14 · S for �pae subjected to -lSC and t=- 1 · 60 P=0 . 12 d f=34.S for those placed at SC) . Pupal developnent times have been measured for other anthanyiids but all of them have more than one generation per year . For pupae , of the seed corn maggot , Delia platura (�igen) the day degrees above the threshold of 3 ' 9C for developnent was 140 (Fuooerburk et al . , 1974 ) . This species has two generations per year canpared to the cabbage root fly which has two generations in the north of England and three in the south . Coll ier am Finch ( 1983) found that a temperature of 0-6C was required for the latter to canplete diapause developnent with a further 14 days at 20C elapsing before the crlul ts started energ ing • This is canparable to earl ier stud ies on the cabbage root fly . Coaker and Wright ( 1 963) fouOO that once d iapause developnent in the cabbage root fly has been canpleted by a suitable cold per iod an accumulation of about 20S day degrees above a threshold of S . 6C was required for Plpae to complete morphogenesis . 83 . F i gu r e 3 . 9 P ropo r t i on o f t h i rd i n s t a r l a rva e d ropD ing at d i f f e r en t hu mi d i t i e s i n th e c o nt r o l l ed t e mperatur e r o o m s a t D . S . l o R . +> C;...j cO ..c +> H cO CI) rl 'D cO ( 25C ) H '0 CI) C;...j ( 1 5 C ) 0 10;-""- 4 5 5 3 6 1 6 9 7 7 8 5 R e l a t i v e hu m i d i t y ,---, Q) s:: Q) s.:: :\! � > � h CD bC rl C ·rl ;.... � Cd u � Q) UJ rl C rl ·rl 0 U 'D h h (V\ Q) p. � o t>rJ C h ·rl Q) p; .D o... E 0 � h Z 'D F i gu r e 3 . 1 0 T h e d ro2pin� r e spon s e of s e e d f ly 3 r d i n s ta r l a r v a e i n r e l a t i on t o w a t e r i ng s e ed h ea d s a ) n o t wa t e r e d b ) l i g h t l y s p ray ed w i t h 5 0 0 m l s wa t e r d a i l y c ) s in g l e d r en c h w i t h 1 0 0 0 m l s wa t e r a t t h e s ta r t d ) l ig h t l y s pray e d w i t h 1 0 0 m l s o f w a t e r on d a y s 2 a nd 3 t h en d r en c h ed da i l y f r o m day 4 2 0 a ) 1 0 2 0 b ) 1 0 2 0 c ) 1 0 2 0 d ) 1 0 0 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3 Day s f ro m s ta r t o f e x p e r i m en t s co ::.. 8 5 . F i g u r e 3 . 1 1 R e l a t i on s h ip b e t w e en nupa l d e v e l opm en t t i me a t 2 0 C an d pr e v i ou s e xpo s u r e t o c o l d t e mpe ra tu re s (.) o N Q) E ·rl +' C/J » ctl '0 0 0 0 N +' ctl Q) E ·rl +' +> s:: Q) E 0.. 0 rl Q) :> Q) '0 s:: ctl Q) ...... ...-- a ) 2 2 0 1 9 0 1 6 0 • 1 3 0 1 0 0 b ) 2 5 0 2 4 0 2 3 0 2 2 0 2 1 0 2 0 0 1 9 0 1 8 0 1 7 0 1 6 0 • ... p r e v i o u s e xp o s u r e t o 5 C ... ... • • • • • • • • ..... .... ..... . . ..... Reg re s s i o n l i n e s lo r : .. - - - . P u ')a e c o l l ec t ­ ed ' f ro m Wa i rake i P u pa e c o l l e c t ­ e d f r o m the D e s ert Road 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 26 28 • ... P e r i o d o f e xpo s u r e t o 5 C ( week s ) pr evi o u s expo s u r e t o - 1 5 C Regre s s i o n l i nes lo r : • I P u pa e fro m Wa i rake i y= 2 4 0 - 4 ' S x .. _ _ _ .. P u pa e D e s ert .... .... • .... + .... • . .... • .... .... • ... + .... ... + ..... - ... ..... .... y = 1 8 6 + 1 '4 x 2 4 6 10 ... ... • 100 P e r i o d of e XDo s u r e to - 1 5 C ( l o g m in u t e s ) ... 1 000 fro m Road • t h e 86. CHAPTER FOUR: MODELLING THE SEEDFLY POPULATION 4 . 1 Introduction Seedfly development rates in response to temperature obtained in laboratory conditions are used to construct a model of the pre-oviposition period and oviposition at Redwoods Forest . The models are compared with the data collected dur ing the sampling programme at the field site during the study. The degree of synchrony between seedfly emergence and ragwort flower ing which is seen as an important factor affecting seedfly infestation levels of ragwort (Fig . 4 . 1) is also predicted . 4 . 2 The pre-oviposition period OVary development in ragwort seedfly requires 232 day degrees above the threshold temperature for development of 4 . l4C. This was calculated from the regression of ovary development rate against temperature (Fig . 3 . 4 ) Day degree summations were calculated from the field temperature data accord ing to the following formula . [max _t_em�p_:_m_in_t_em--A..p]- 4 . 14C Thi s method , which is val id over intermediate temperatures (Wagner et al . , 1984 ) , was cons idered sui table for th i s model s ince the daily minimum temperature at the f ield s i te was always above the lower threshold for development . The emergence curves at Redwoods Forest were transfo rmed into 8 7. F igu r e 4 . 1 Fa c t o rs a f f e c t i ng s e ed f ly i n f e s t a t i o n o f ragwort s e ed h e a d s a t R edwoo d s Fo res t N o . o f p l an t s i n t h e p re s en t y e ar .. - S yn c hrony b e tween a d u l t e m e r g en c e and ragwort f l owe r i n g N o . p l an t s i n p r e v i ou s y ea r N o . f l i e s e m e r g i n g I n f e s ta t i on l ev e l s in � t h e p r e s en t y ea r S e e d f l y m o rt a l i t y 1 ) e g g s & l a r v a e 2 ) pupa e I n f e s t a t i o n l ev e l s in t h e p r e v i o u s y e a r Adu l t m o r ta l i t y I S i z e o f p l a n t s i n t h e p r e s ent y e a r 88 . curves for gravid females (Fig . 4 . 2a+b) by summation of the day degrees from female emergence to 233 . From this the number of buds per female at each sampl ing date could be calculated fran the data collected at the field s ite ( Table 4 . 1) . This represents the number of buds available at the tline of sampl ing and assumes no mortal i ty and balanced emigration and ilnnigration in the adult seedfly population . 4 . 3 Seedfly ov iposition Dur ing the field studies fl ies were able to pass fran egg to second larval instar or f i rst to third larval instar with in the 14 day sampl ing interval . As the total duration of the egg and first and second larval instars is 20 days the oviposition curve was calculated from the sum of these stages obtained fran each fortnightly sample . Accord ing to this method oviposition coincided with the start of ragwort flower ing in both 1982 and 1983 (Fig 2 . l la+b) . The val id ity of this method can , however , be checked by summing the day degrees for development backwards fran the third larval instar stage . The day degree summation fram egg to the third larval instar leaving the plant was 629 . It was calculated fram the regression equation for development rates against temperature (Fig . 3 . 6 ) . This was then reduced by the mean ratio of 0 . 383 ( the time fram ov ipos i t ion to the beginning of the third larval instar d ivided by the time to the end of the th ird instar ( Table 4 . 2) ) . This gave an est imate of the day degrees fram ov iposition to the start of the th i rd larval instar of 24 1 . The only other publ i shed " ' G .,..:. r � -' :.-- G t>�� - > or' �� -� -' ::0 E ::0 (..; � i 2 u r e 4 . 2 a Cu mu la t i v e u e r c en ta � e c u r v e s o f s e ed f l a n d ra �wo r t s e edh ead d ev e l o u m ent s tage s a t R edwo o d s F o r e s t i n 1 9 8 2 8 3 9 0 8 0 7 0 6 0 5 0 4 0 3 0 2 0 1 0 1 1 5 2 9 Nov ", _ / / / 1 3 I I �I col � I - , �I �, 0,1 I / / I I I Dec I I / / / / 1 0 J an S a mp l i n g d a t e Feb f ie l d data - -- - m odel M a r Apr co -0 o - - � - " � -+.:: "-" s ""'- � > +', ;;J - - :::l " - � L ; u r e 1. . 2 b C u mu la t i v e e r c en t a a e c u r v e s o f s e ed f l d e v e l o p m e n t s t age s a t R e dwo o d s F o r e s t 9 0 8 0 7 0 6 0 5 0 4 0 3 0 2 0 f i e l d da ta - --- mode I I G 1 5 N ov 29 1 3 2 7 1 0 Dec J a n 24 S a m p l i n g d a t e 7 2 1 Feb 7 21 Ma r s e ed h e a d '<> o • 9 1 • [' ;.; b I ( I; . 1 l l a ,.;w o c t bu d s pe r o v ipo s i t i ng J" l; ma l e a t t h e R e d wo o d s F o r e s t f i e l d s i t e D il t ( � 1 2 / 1 2 2 6 / 1 2 9 / 1 2 3 / 1 6 / 2 2 0 / 2 Bu d s / f e ma l e 1 9 8 2 / 8 3 0 . 1 4 0 . 2 2 2 . 9 8 2 6 . 3 8 6 5 . 1 7 1 0 6 . 7 3 B u d s / f e ma l e 1 9 8 3 / 8 4 0 . 0 3 0 . 9 4 2 . 7 0 6 . 0 3 9 . 7 4 1 0 . 6 3 Ta hl E: 4 . 2 D e v e l opm en t t i m e s f o r ragw o r t s e ed f ly r e a r e d i n t h e c on t r o l l ed e n v i r o n m e n t r o o ms a t P l a n t P hys i o l ogy D i v i s i on , D . S . l . R . , Pa l m e r s t o n N o r t h R e a r e d a t R e a r e d a t 1 5 C 2 0 C �1 ean d e v e l o p m en t t i m e ( d a y s ) f ro m o v i po s i t i o n t o en d o f t h e 2 n d la r va l i n s ta r 2 0 . 5 1 3 . 9 M ea n d e v e l o p m e n t t i m e ( d a y s ) f r o m <.) v i po s i t i o n t o 3 r d j. n s ta r l a r v a e 1 e a v i n r.; t h e r l a n t 1) 0 . "3 3 7 . 4 11. < 1 t i 0 <) r t i m e' \' 0 ,� n d o f 2n d i n s t a r / t i m c t o 3 rd i n s t a r 1 en v (� t i l e p l a n t ( M C: i i ll - O . ) �� 3 ) o . -� 9 � 0 . 3 7 3 9 2 . infonnation on egg and larval development rates in the Anthamyi idae is for the seed corn maggot , Del ia platura Meigen , which has two generations per year . The number of day degrees above the threshold of 3 . 9C to complete development was 230 for eggs and larvae (Funderburk et al . , 1984) . The day degrees above the threshold for the seedfly were summed to 241 fram when the population at Redwoods Forest entered the third larval instar . The latter was taken as the first part of the graph of third instar larvae in Figures 2 . 9a+b When the number of third instar larvae was increasing and not under the influence of the larval dropping response . The calculation is therefore not affected by the extended duration of the third larval ins tar measured under laboratory conditions (Chapter 3 ) . The day degrees were calculated as in Section 4 . 2 as the threshold temperature for development of 3 . l3C was always above the minimum temperatures exper ienced by seedfly in the f ield . Figures 4 . 2a+b show the oviposition curves pred icted by this model in 1982 and 1983 . Although the model showed that in both years ov iposit ion commenced as soon as ragwort oviposition sites became available it also predicted that the oviposit ion curve in early spr ing 1983 was not as steep as that estimated fram the fortnightly samples . The ov iposit ion period of ragwort seedfly in the field was two weeks . This was calculated as the time between the curve of grav id females and the ov ipos it ion curve obta ined from the model . This can be compared with data on the ov iposition per iod of the beet fly ( Hafez �al . , 1970) . In this species the 9 3 . egg laying per iod decrease1 with increasing tenperature . It was measured in the laboratory as 42 days at l6C and 11 days at 30C. 94 . CHAPTER FIVE: DISCUSSION 5 . 1 Introduction The concept of biological control of weeds by herbivores is based on the premise that herbivorous anlinals affect the abundance of plant populations . I ts opponents , however , argue that because there are few instances of depletion of plant populations , herbivores are rarely food l imited . Hairston et al e ( 1960) , for example , deduced fram this that "producers are neither herbivore limited nor catastrophe l imited and must therefore be l imited by their own exhaustion of a resource . " They conclude that herbivores are controlled by the depredations of natural enemies in a density dependent fashion . Andrewartha and Birch ( 1954) , on the other hand , bel ieve that density independent factors such as weather and other features of the physical environment control the abundance of animal populations . Depletion of green plants is not , however , a necessary corollary of food l im itation in a herbivore population . Not all plant parts are edible (Bernays , 1981 ) and herbivory may be reduced by the production of deterrents or escape in t ime and/or space . plant populations may also pers ist despi te high levels of herbivory because of reproductive strategies involv ing high seed production or vegetative propagation . While acknowledg ing that herbivores can be food l imited but not l imit their host ( s) Crawley ( 1983) suggests that this food l imitat ion is an essential cr i ter ion for weed management . He states that " i f herbivores are not food l imited , it is most unlikely that they could increase to suf ficiently high densities to br i ng about the required dramatic 95. reductions in the numbers of weeds. " Huffaker ( 1964) also states that "an (biological control) insect must be capable of decisive destruction of its host plant , thus determining the latter ' s abundance . " However , there are many who bel ieve that herbivores have a greater effect on plant populations than that indicated by the low level of consumption . Mattson and Addy ( 1975) suggest that herbivores control the function of Whole ecosystems by regulating nutr ient cycling and stimulating the redistribution of nutr ients wi thin the plant fram stored reserves . Smiley ( 1985) suggests that chemical defence is only one of a number of possible elements that mediate plant-insect coevolution While Janzen ( 1970) argues that seed predation can be l inked to the species composition of plants in tropical forests . The more obvious ways in which ragwort reduces the impact of herbivores include alkaloid production , vegetative reproduction and the production of numerous seeds of high germinating capacity. This study, however , has focussed on the escape in time of a signi f icant proportion of ragwort seeds from predation by the ragwort seedfly. The contr ibution of the ragwort seedfly to an integrated control programme for ragwort is d iscussed and examples of successes us ing this approach are g iven . The value of seed predators as b iological control agents of weeds in general is also d iscussed . 5 . 2 Ragwort control at Redwoods Forest The lack of synchrony between seed fly emergence and flower ing of ragwort dur ing this study is one of the major factors reducing the 96 . impact of the seedfly on ragwort seed production . The pre-oviposi tion period predicted by the model was shorter than that observed in the field . However , there are two factors which may affect the val id ity of the model . Firstly, the number of females reared to matur i ty was low. Mortal i ty was high despite efforts to reduce handl ing . Hoy ( 1960) considered pupal damage responsible for emergence of deformed fl ies and many fl ies emerged in this study with wing deformities probably as a result of the collecting methods . This may also account for mortal ity i n the fl ies released into the cages . Humid ity in the controlled environment rooms may have affected mortality but this was not accurately measured . Copulation was unl ikely to have been an l imiting factor in ovary maturation since females were mated soon after emergence and those not mated reached matur ity at a sim ilar age . The second factor influencing the rate of ovary development in the laboratory was diet . The add it ional d iet g iven to adults reared in the controlled temperature roams could have resulted in accelerated ovary development . However , when it was not provided females failed to reach matur i ty and this has occurred dur ing previous attempts to rear seedfly adults ( P . Syrett , pers .comm . 1983 ) and also Hylemyia species ( Zimmerman et al . , 1984 ) . The d iets prov ided represented a range of proteins and carbohydrates s im i lar to that recommended for anthomyi ids by Singh ( 1977 ) and those found to be necessary for development by researchers on the beet fly ( Hafez et al . , 1970) , the cabbage root fly ( Finch , 1971 ) , the wheat bulb fly and the on ion fly (Coaker 9 7. and Finch , 1973 ) . There is no requirement for ragwort flowers for ovary development and no signi ficant d ifference between development times in the presence or the absence of ragwort flowers . A pre-oviposition period in the absence of ragwort flowers is therefore not l ikely to affect ovary development in the f ield . However , in this case , the period appears to be extended in the f ield by the absence of oviposition sites and this is supported by the fact that females were able to lay eggs on transplanted plants which flowered before the f ield plants . The pre-ov iposition model predicted that competition for ov iposition sites was high when ragwort first flowered and data on multiple infestation and fecundity supported this . Fecundity was low in both sumners . 12 · 9 eggs were laid per female that emerged at the site in 1982/83 and 3 · 9 in 1983/84 . There was a tenfold increase in the number of fl ies emerg ing from 1982/83 to 1983/84 which increased competition for ov iposition s i tes . Density and size of ragwort was similar in both years but a combination of the increase in the number of fl ies and lower fecundity in 1983/84 resulted in a doubl ing of infestation levels from 10% in 1982/83 to 20% in 1983/84 . In 1982/83 the carrying capacity of ragwort far exceeded the egg laying potential of seedfly ( 32 eggs per batch) with 404 seedheads produced by ragwort per female emerged . In the following year , however , there were 39 ragwort seedheads per seedfly female that emerged at Redwoods Forest s i te . OViposit ion behav iour may also have been affected by temperature , weather cond it ions and mating 98 . activity . Hafez et al . ( 1970) found that fewer eggs were laid by beet fly females at cooler temperatures . I t is also likely that the retr icted number of oviposition sites may have affected egg production . Coaker and Finch ( 1973 ) found that ovary maturat ion in the cabbage root fly does not depend on chemostimulation from the host plant but the nearer to matur ity females are when they receive a chemical stimulus the more eggs they lay. The high percentage of multiple . infestation at the beginning of spring also suggests competition for ov iposition sites . A max imum of 22% of the seedheads containing seed fly were multiple infestations in 1982/83 and 16% in 1983/84 . The highest recorded loss of eggs due to multiple infestation was 9 -4% and 14 ·3% in the early December samples in 1982 and 1983 respectively. A uniform egg distr ibution was recorded in �. j acobaeae in early December and January . This is the optimal strategy for the seedfly since only one egg per seedhead surv ives to pupation . Z immerman ( 1979 , 1980) saw the uni form egg distr ibution observed for Hylemyia in P . fol iossisimum at least , as ev idence for the ex istence of an ov iposition deterr ing pheromone (ODP) associated with the egg although the pheromone has not yet been isolated . Certainly, egg recognition has been demonstrated to be fac i l i tated by the presence of an ODP in at least 24 species of phytophagous insects ( Prokopy , 198 1 ) and usually occurs in situations where there is larval competition for resources . The cost of pheromone product ion 1 S high ( Prokopy , 198 1 ) and fo r 9 9 . this reason Z immennan ( 1982) argued that OOP production in Hylemya species was facultative . He postulated that an ooP was not deposited when the female was ovipositing on l.aggregata since high egg mortal ity due to dessication was reducing competition for food reserves within the seedhead . As a result , the incidence of multiple infestation was high ( 25%) . Although a predominantly uni fonn egg d istr ibution was recorded for R.jacobaeae and females were observed wiping their abdomens on seedheads after ov iposit ion the ev idence for an ODP is equally scant . The high levels of multiple infestation recorded in this study could have resulted from lowered thresholds for oviposition so that females oviposited in spi te of an OOP or the presence of another egg . prokopy ( 1972) found that the physiolog ical state of female apple maggot fl ies ( Rhagoletis pomonella) has a strong influence on the degree of responsiveness to OOPs . Females deprived of ov iposition sites for two days were more l ikely to ov iposit in OOP�arked fruit than non-deprived females . Whether this occurs in the ragwort seed fly is , at present , largely speculative . It is l ikely that the number of buds suitable as ov iposition s ites for seedfly was overestimated in this study . The subjectivity involved in this estimation was reduced by determining the s i ze of buds recently infested but it was too time-consuming to measure every bud dur ing the data collection and the chemical cues involved in the cho ice of ov iposit ion s ites were not invest igated . Such an overestimat i on affects the calculation of the type of 1 00 . distr ibution o f seedfly eggs . An overestimation of buds also reduces the estlinates of infestation levels in early spr ing which would be expected to be higher than the 18% and 36% observed in 1982 and 1983 respectively if competition for oviposition sites was high at this time . Other factors affecting seedfly fecundity and infestation levels of ragwort are adult mortal ity , emigration and llnriigration but no information was obtained about these during this study. No d irect comparison could be made between emergence and sticky trap catches since the latter does not provide an absolute estimate of adults and is biased toward males . In addition nothing is known about the dispersal capabil ities of �.jacobaeae but Finch and Skinner ( 1975) found that the d ispersal range of the cabbage root fly was within a 2-3 kilometre rad ius of the release s ite . This represented a dispersal rate of 100m per day al though only 38% of the males and 15% of the females released were recaptured . Mortal i ty at the egg ,larval and pupal stages wil l also contr ibute to the potential rate of increase of the seedfly population . The cause of death of larvae in seedheads with multiple infestations is unknown but cann ibal ism cannot be ruled out . Inoue (1983) found that 64% of fi rst nymphal feed ings in the assass in bug , Agr iosphodru� dohrn i Signoret , were cannibal i st ic and this was reduced to 0 ·04% in the th ird instar . Although his comments relate to a predaceous bug , Inoue bel ieves that most cases of cann ibal ism are attr ibutable to resource l im itat ion . He suggests that i t may funct ion as a means of prov id ing an eas i ly access ible 1 0 1 • nutr itional source for the linmature stages which have low locomotive abil ity . The assassin bug nymphs are slinilar to seedfly larvae in that they are largely confined to the sites where oviposition took place . Cannibal ism was reduced in this species by modi fication of the female oviposition behaviour . More than half the overlapping ovipositions took place on the same day which resulted in 95% of the larvae escaping cannibal ism. Zimmennan ( 1980) found that single Hylemya females repeatedly oviposit in the same seedhead and bel ieves that the first hatched larva does not destroy the others since it would be detroying its own sibl ings . I t is not known whether R.jacobaeae females oviposit in this way but it is l ikely that availabil ity of resources in the seedhead is 1 imi ted . Feed ing and excretory products of older larvae damage the unopened florets essential for development of f irst instar larvae . Larger seedheads resul ting fram fasciation were recorded supporting three third instar larvae . Survivorship was calculated fram the area under graph of numbers of each stage at each sampl ing date . The area is affected by the duration of the sampl ing interval . Increasing the frequency of sampl ing would have more accurately pinpointed the peaks of each stage. The surv ivorship curves presented are al so steepened by the probable under-estimat ion of f i rst instar larvae because of the d i fficulty of find ing the larvae within the florets when the frozen seed head had thawed . The d ispar ity between estimates of pupae col lected in the trays and th i rd instar larvae within 1 0 2 • seedheads in 1984 suggests that third ins tar larvae are also I.ll"rler-estimated . The duration of the third larval instar used to calculate the number surv iv ing to this stage was artific ially extended by the dry conditions in the controlled env ironment roans . There is certainly same evidence , both fran the field studies and experimental work , that the duration of the third larval instar within the seed head is var iable . Third instar larvae were experimentally shown to leave the plant in conditions of high surface moisture and this would be an advantage to a small animal at r isk fram dessication . Larvae have been observed dropping fram the plant and , also on one occasion when fine rain was fall ing , using surface moisture on the leaves and stem of the plant to sl ide to the ground . Thi s behaviour was responsible for the high estimate of third ins tar larvae at Redwoods Forest which have left the seedhead as sampl ing co incided with per iods of rain . I t is expected that high mortal ity in th ird instar larvae similar to that observed in the controlled env ironment roans could occur i f there was a prolonged dry period in late summer and autumn . This is not nonnally the case within the seedfly' s present range in New Zealard • Informat ion on pupal surv ival over winter in 1983 are contrad ictory . Est imates o f pupal density in autumn ard spr ing in 1983 by so i l sampl ing were 66% lower than the correspording estima tes o f th ird instar l a rvae w i thin seedheads pr ior to pupa t ion and the number o f ad ul ts caught by emergence traps . The 1 03 . pupal mortal ity calculated from the third instar larvae within seedheads to those emerging the following spr ing was 14 . 3% compared to 2 -5% for the d i fferences in so il sampling estlinates of pupae in autumn and spr ing 1983 . unfortunately the effic iency of soil sampl ing was not measured . There were twice as many pupae collected in trays as third instar larvae estimated within seedheads in autumn 1984 . I f the number of thi rd instar larvae within seedheads were under-estimated to the same extent in autumn 1983 the estlinate of pupal mortal ity dur ing the following winter would then be 57% . Pupae are able to withstand very cold temperatures in laboratory conditions but do not surv ive dess ication (Miller , 1970) . Other factors affecting pupal survival over winter were not tested but queen ants of Cheloner antarct icus found in the sand trays d id not prey on the thi rd instar larvae or pupae when g iven the opportunity . 5 . 3 Ragwort seedfly and the biolog ical control o f ragwort The lack of synchrony found in thi s study was also alluded to by Kelsey ( 1955 ) dur ing field observations 7 years after the release of seedfly at Redwoods Forest . He noted that although the late flower crop accounted for approx imately 12 · 5% of the total flower crop there were no adult fl ies to carry on infestation . A lack of synchrony was al so noticed by Hoy ( 1958) when 4000 pupae were collected from the s i te for shipment to Austral ia . H igh levels of infestation were recorded soon after seedfly release but a per iod of adj ustment to Southern Hemisphere seasons may have been 1 04 , necessary as adults were or ig inally released in autumn . Unfortunately it is not known whether a similar lag between emergence and oviposition is the norm in the fly' s country of or ig in . Infestation levels vary throughout England and are higher in the north . Infestation levels were also low at the Desert ad site but the degree of asynchrony at this site is unknown. It is possible that seedfly biotypes with later adult emergence may occur within the fly ' s native range. However , the timing of flower ing was var iable in this study and it is not known how cl imate , rainfall and soil type affect this . It is l ikely that ragwort b iotypes also ex ist , further compl icating the situation . The ragwort seed fly' s potential as a biocontrol agent of ragwort would be greatly enhanced if two generations per year were possible as it would enable the fly to infest some of the later developing flowers . The spread of ragwort flowering extends over 3-4 months but seedfly is not able to complete its l i fe cycle in less than 9 months at constant l 5C which represents the average summer temperatures in the field . Even if seedfly were able to infest every seedhead , uneaten seeds from seedheads infested by seedfly will germinate and it is not known if ragwort compensates for seed predation by reducing its allocation of resources to this form of reproduct ion . The contr ibution of f .iacobaeae to the integrated control of ragwort wil l depend on competi tion with other species for seedhead resources . In particular , competi t ion wi th the closely related f.�en��ella may be intense . These two spec ies coex ist in 1 0 5 . their country of or ig in but i t is not known in what proportions and to what extent the infestation levels are affected by the competition . Fr ick ( 1969) states that the pre-ov iposition per iod in P . seneciella is 7 to 10 days but does not elaborate under what conditions this was measured . If this is an estimate at average spr ing temperatures then the population may be expected to be even further out of synchrony with the ragwort population i f released into New Zealand and may expla in why it failed to establ ish here . However , in England the shorter pre-ov iposition per iod of �. seneciella may serve to reduce the overlap of the oviposition period of the two species . Seedfly may also be in d irect competition with cinnabar moth larvae which prefer to feed on the developing ragwort buds (Dempster , 1982 ) . Most biological control programmes at present follow the innoculative approach , that is , herbivore populations are released and left to increase to effective levels on their own . The alternative is the inundative approach where relatively short term control is achieved by releasing a biocontrol agent in heavy concentrations (Tisdell et al . , 1984) . Th is is not sel f perpetuating and may be costly . I t certainly would not be considered worthwhile for the ragwort seed fly • Although easy to collect and transport seedfly pupae can easily be damaged or dessicated . A high proportion of the initial releases , both in New Zealand and overseas , were unsuccessful and rear ing adults has proved to be d i ff icult . Further research in this area would be required before large scale releases could be considered feasible . --�----.----� 1 06 . There is evidence that biological control of ragwort may more effectively be accompl ished by a cinnabar moth/flea beetle combination . The ragwort flea beetle Longitarsus jacobaeae Which feeds on the roots over winter is expected to g ive good control in conjunction with either grazing or cinnabar moth defol iation (Hawkes and Johnson , 1976) . Ragwort ' s regenerative capacity is dependent on it root carbohydrate reserves so that root damage is l ikely to ser iously weaken the plant . Climatic stress may enhance the effect of the cinnabar moth � jacobaeae as a biological control agent of ragwort . On the east coast of the USA defol iation by the cinnabar moth led to a decl ine in plant numbers because regenerating rosettes were killed by frost at the cr itical stage of recovery (Harris et al . , 1976 ) . Autumn frosts damaged regrowth flowers after defol iation bY Tyria in a population stud ied by Islam ( 1981) in England . Cox and McEvoy ( 1983 ) state that the full potential of the cinnabar moth as a biocontrol agent will be apparent in years wi th below average summer rainfall as this decreases the capaci ty for ragwort to compensate for defol iation . Farming practices may also work against the spread of ragwort . Smith ( 1982) states that there are many farms where ragwort is kept in check with relat ively l i ttle annual effort in time or cost . Prov ided that a vigorous and competitive pasture sward can be developed and mainta ined , ragwort can be kept in check . He ci tes lack of finance as a major factor leading to neglect of persistent control of ragwort . 1 0 7. 5 . 4 Theories and practice of biological control of weeds The insects that have been highly effective in the biological control of weeds include a stem borer (Cactoblastis species , Lepidoptera : Pyral idae , on Opuntia) ; plant suckers ( 3 species of Dactylopius , Hemiptera : Dactylopi idae , on Opuntia species) ; leaf feeders (Chrysol ina quadrigemina Suffrian , Coleoptera : Chrysanel idae , on St John ' s Wort , Hypericun perforatun L . am Metrogaleruca obscura DeGeer . (Coleoptera : Chrysanel idae) on black sage , Cordia macrostachya (Jacquin» . A gall insect , Procecidochares uti l is Stone (Diptera : Tephr itidae) was effective against crofton weed , Eupator ium adenophorum Sprenge and Agasicles hygrophila Selman and Vogt (Coleoptera : Chrysomel idae) a stem and leaf feeder has substantially reduced all igator weed , Alternanthera philoxeroides (Martius) (Jul ien , 1982) . However , 60-75% of introductions recorded by Jul ien ( 1982 ) have been unsuccessful and DeBach ( 1974) cites that 75% of the 41 projects for which there was publ ished information achieved a measurable degree of success . Eight had been rated as a canplete success , 9 as substant ial successes and 14 as partial successes . Insect seed predators are often favoured biological agents because of thei r host specificity but their success as biological control agents of weeds is generally low . Table 5 . 1 shows the type of control obtained by seed predators as presented in the world catalogue of weed biocontrol programmes (Jul ien , 1982) . In addition , 20% of the 120 weed species ( representing 34 famil ies) which have been the targets of biolog ical control programmes (Jul ien , 198 2 ) belong to the family Asteraceae ( =Campositae) . Tabl e 5. 1 Eff e c t i v en e s s o f c ontrol agen t s Weed and i t s o r igin C o mp o s itae Carduu s a can thoid e s L . ( plumel e s s t h i s t l e ) Eura s ia , N . Afri ca , W . A s ia Carduu s nutan s L . ( nodding th i s t l e ) Europ e , A s ia Cardu�� pycno c ephal u s L . ( s l en d e r w in g e d t h i s t l e ) Euro p e , A s ia Carduu s t enu if lo ru s Cu r t i s (winged thi s t l e ) W . Eu rope S i lybu m mar ian u m L . ( milk t hi s t l e ) M ed i t erran ean , S . W . Europe C on t r o l A gent Rhin o cyllus c o n i c u s Fro el i c h ( C o l eopt era ; C ur cu l ion ida e ) " " " Rhino cyl l u s c o n i c u s Fro e l i c h ( C o l eopt e ra : Curcu l i on ida e ) 1 08 . Statu s and d egree o f c ontro l Canada ; thi s t l e s tands a r e report edly l e s s d en s e . USA : d e s tr o y s s o m e of t h e s e ed s but d o e s n o t redu c e t h i s t l e d en s it y . Canada : r edu c ed t hi s tl e t o l e s s t han 1 0 % of i t s f o r me r d en s i t Y . L e s s e f f e c t when t hi s t l e i s growin g w ithout c o mp e t i t ion . USA : t hi s t l e r edu c ed 9 0 - 9 9 % at s o m e r e l ea s e s it e s . N Z : e xpe ct g ood c ontrol USA : h i g h rat e s o f flowerhead infe s t ­ at ion and s e ed d e s t ru c t i on bu t l i t t l e e f f e c t on o v e rall w e ed d en s ity where we evil s hav e b e en l ong e s ta bl i shed N Z : exp e c t g ood c on t r o l USA : 9 0 % or m o r e o f fl owerheads a ttac ked in s o m e s it e s bu t l i t t l e d i r e c t s e e d d e s t ru c t i on o r e f f e c t o n plan t d en s i t y Tabl e c ont . . . 2 Weed an d i t s o r i gin C ontrol A g ent Xan thium spin o su m L . Eua r e s ta bu llans . ( Bathu r s t burr) W i e d e mann C o s mopol i tan ( D ip l era : T ep h r i t i da e ) XanthiuIIl s t ru marium L . ( noogo ora bu r r ) C o s mopol i t an L e gu mino c ea e U l e x europa eu s L . ( g o r s e , fu r z e) W . Eu r o p e V e rb enac ea e Lan tana ca mara L . t r o p i cal A me r i ca B oragina c ea e C ord ia ma c r o s ta c hya ( bl a ck s a g e) L i n s t i s dal matia L . L ina ria vu lga r i s L . ( t oad flax) Eu r o p e O roban c ha c ea e Oro ban che cu mana Wa l t er Eu ra s i a O roban che ra mosa L . Eua r e s ta a egu a l i s L o ew . ( D ip t e ra : T ephri t idae ) E xapi on u l i c i s ( Fo r st er) ( kn o wn a l s o a s Apion ) Epin o t ia lantana Bu s c k ( L ep idopt era : T o r t r i c i da e ) Oph i o myi a lan tana e Frogatt ( D ipt e ra : A gromy z ida e ) Eu ryt o ma a t i va Burks ( Hymenopt era : Euryt o mida e ) B ra c hypt er o l u s pu l i c a r iu s ( C ol eopt era : N i t i d u l i da e ) and Gymn a e t r o n ant i rrhin i ( C o l eopt era : C ur cu l io n i da e ) P hyt o myza o r o ban c h ia Kal t en ba c h ( D ipt e ra : A g r o my z idae ) " 1 09 . Statu s and d e g r e e o f c on tr o l S . A f ri ca : i nf e s t s u p t o 2 0 % o f bu rr s Au s t ral ia : g i ving no c on t r o l of t h e weed Hawa i i : e f f e c t n e g l i g ibl e USA : no d e t e c tabl e i mpact exc ept a t o n e int e r ior s it e Au s t ral ia : ha s n o t aff e c t ed s pread o f go r s e i n Ta s man ia or V i c t o r ia Au s t ral ia : m in i mal s e ed r ed u c t i on S . A f r i c a : c o nt r i bu t e , l i t t l e t o s e ed d e s t ru c t i on Mau r i t iu s : h i gh p e r c en ta g e o f s e e d s are a f f e c t ed bu t part p la y ed i s unclear I n c o mb inat i o n g i v e 8 0 - 9 0 % s e ed r edu c t ­ ion . A ppare n t ly t h e r ea s on f o r t h e d e c l in e i n s e ri o u sn e s s o f t he w e ed Yugo s lav i a : c an a c h i ev e c on s id erabl e c ontrol by d e s t r o y ing up t o 96% o f s e e d s USSR : g o o d c o n t r o l a chi eved bu t pro b l e m s w ith syn c hron i s in g P . o ro ban c h ia de vel opment o f t h e we ed Tabl e cont . . . 3 Weed and i t s o ri gin O r o ban che ramo sa L . P ro ta c ea e Hakea s erv i c ea Shrad e r ( s ilky hakea ) A u s tralia C ontrol A g ent " Eryt enna con sputa ( Pa s c o e ) ( C o l e o p t era : C u r cu l i on idae ) 1 1 0 . Statu s and d e g r e e o f c ontrol Yu go s lav ia : can achieve c on s id e ra bl e control by d e s t r o y ­ ing up t o 96% o f s e eds S , A f r i c a : s ign i f i c ant s e ed d e s t ru c t i on - ------------ ----- 1 1 1 . These character ist ically produce numerous small seeds which are eas i ly d ispersed so that the potent i al for predators wh ich feed on the seeds once they have left the plant to reduce plant populations is therefore also l lini ted . Burdon and Marshall ( 1981 ) have also recorded that asexual reproduct ion occurred in 60% of the 40 target species they surveyed . Poor synchroni zat ion of damage has also been recorded in a number of b iolog ical control programmes involv ing seed predators . Waterhouse ( 1967 ) suggested that l ack o f synchrony between the seedfly Euaresta aqual is Loew. and the occurrence of seeds of the Noogoora burr ( Xanth iurn pungens Wal l r . ) at the r ight stage of development is respons ible for the low level of infestation of thi s weed . For ster ( 1977 ) found that pod infestation of gorse , ylex europaeus L . by the seed weev il Exapion ul icus Forster ( formerly �pion) over twelve months was 19 · 5% . The weed flowered throughout the year but infestat ion decl ined from a max imum of 4 4 % in spr ing . popay �al . ( 1984 ) found that max imum egg laying of the nodd ing thi stle receptacle weev i l ( Rh inocyllus con icu� d id not coinc ide w i th max imum flower production in i ts host Carduus nutans . Th i s resul ted in abnost 100% infestation of the f irst flowers but decl ined to approx imately 5% in February 1983 and 20% in February 1984 . The few weev ils remain ing ( 0 · 2/m2 ) in March 1984 were able to infest a h igh proportion of the low n�er ( 0 0 05/m2 ) of f lowerheads ava i l able . The highest overal l 2 infestat ion recorded was 7 5 0 9% , wh ich al l owed 2000 seeds per m to escape predat ion . 1 1 2 • Successes in biolog ical control of weeds have often resul ted fram a combinat ion of phys ical and b iolog ical factors as has al ready been descr ibed for ragwor t . Var iations in so i l , water , d isturbance of the hab i tat and cropping practices all influence abundance of weeds ( Andres et al . , 1976 ) and the level of control is not always sim ilar throughout the range of the weed . For example , the t i ng id Teleonomia scrupulosa causes death of Lantana only i f heavy attack i s coupled wi th severe stress , i .e . prolonged drought ( Schroeder , 1983 ) . Diseases assoc iated with the cactus feed ing moth Cactoblast i s cactorum ass i sted in the successful control of pr ickly pear in Austral ia (Wi l son , 1964 ) . Defol iation of �r icum rosettes in the autumn and winter by the larvae of Chrysol ina quadr igem ina contr ibute to the death of the plant dur ing the dry summer season in Cal i fornia . The defol iated plants d id not have tline to redevelop an adequate root system before the summer drought ( Huffaker , 1957 ) . Explo i tat ion of stress factors involv ing interactions between cl imate , so i l , competing plants and natural enem ies i s seen as an important strategy in the b iolog ical control of weeds ( Harr is , 1980) . Stress , however , acted aga i nst pr ickly pea r control in parts of Queensland , Austral ia as i t reduced plant succulence and hence insect attack ( Schroeder , 198 3 ) . Determ ination of weed status i s a facto r often neglected pr ior to the i n i t ia t ion of a control programme . Often est imates of the sever i ty of a weed problem come from the costs of control measures employed ag a i nst the weed and these somet imes do not reflect the rea l i nc idence o f i n festa t i ons . Stud ies on weed cover i n New 1 1 3 . Zealand include a survey of scrubweed cover of South I sland agr icul tural and pastoral land from 1972-1976 by Bascard and Jowett ( 1981 ) . Measurements of i ncidence and cover of 1 2 herbaceous weed species at improved pasture s i tes ( 1/1000 hectares) chosen randomly from four areas in both the North and South Islands also began last summer ( Ian Popay and Dave Kel ly , pers .comm . 1 985 ) . Thi s will g ive some long overdue quant i tative data on the weed status of many spec ies includ ing ragwort . The loss in animal production as a result of weed infestat ions is often d i ff icult to determine . For ragwort , in part icular , alkaloid poisoning o f domest ic stock is an L�portant factor . Only 287 cases were d iagnosed in Animal Heal th Laboratories in New Zealand from 1973-1984 ( Peter Watson , pers .cam. 1985 ) which represents 0 . 05 % of the cases investigated . However , thi s is not l i kely to represent the true s i tuation as many cases are not reported by farmers or veter inar ians . Sub-lethal effects of ragwort poi son ing in stock such as poor g rowth , lowered m i lk production and qual i ty are also d i f f icult to quant i fy . Assessment o f the weed status o f the target spec ies is more compl icated where a confl ict of interests ar ises . Commerc ial blackberry product ion 1n Austral ia is threatened by the apparently i l legal introduction of the blackberry rust , Phragmidium v iolaceum , and Austral ian graziers and beekeepers have al so opposed the C . S . I . R . O . b iolog ical control programne for Echi� pl�nat:�g i ne '!12 L . ( Salva t ion Jane/Paterson ' s Curse) ( P . Syrett , pers . coom . 1984 ) . 1 1 4 . Quanti tat ive reviews of insect releases are also ilnportant . The level of reduction of the target species , more read ily precisely obtained for seed predators ( Table 5 . 1) , rarely determine the overall effect on the weed population since l ittle is known of the ecology of the plant . For ragwort , in part icular , nothing is known about the level of seed reduction required to ensure a significant reduction in the the population , the factors which govern flowering or compensatory growth in response to seed predation . 5 . 5 Conclusions This study has shown that assessment of insect releases is ilnportant for planning further research in the field of biological control . In add i tion to determining the potential of R. jacobaeae as a b iological control agent of ragwort it also highl ights the need for assessing the weed status of the target plant and research into the ecology of the weed with respect to maintenance of the weed population and responses of the plant to herbivore attack . The prospects for biological control of weeds using insects must also be considered in relation to current theor ies of how herbivores affect plant populations as outl ined at the beg inning of this chapter . The aims of weed control and consequently the ways In which control programmes are conducted depends on the expectat ions of the interested part ies . There are those who bel ieve that "control" of the weed populat ion is feas ible . That i s , plant numbers or biomass can be restra ined or reduced to an ----------_._--- 1 1 5 . acceptably low level . "Depletion" or erad ication of the weed may even be considered a real istic alin of such projects despi te the rar ity of such events ( Hairston et al . , 1960 ) . Herbicide appl ication , mechanical removal and augmentation of herbivore populations are examples of control measures employed to this end and generally require regular input of labour and money. The alternative is the establ ishment of a herbivore population that regulates or continually adapts and adj usts the weed population resul ting in "long term stabil isation of weed densi ty at a sub-econanic level" (Schroroer , 1983) . The research input into this approach is higher since the more subtle ways that herbivores and their hosts interact with each other and the physical environment must be investigated . Often the weed population may be regulated at a level that does not j ustify the cost of a biolog ical control programme . Herbivore releases must therefore be continually reviewed and the " sub-econanic level" of the weed appropr iate for cl imate and land use should be determined as soon as possible . 1 1 6 • SU�RY populations of ragwort seedfly were sampled fortnightly at two cl imatically d i fferent field s ites from October to April in 1982/83 and 1983/84 to determine the potential of the seedfly as a biolog ical control agent of ragwort . Data collection was more intense at the wanner Redwoods Forest site than at the higher al titude Desert Rj site . Seedfly adults emerged six weeks before ragwort flowered in both years at Redwoods Forest and oviposition coincided with the first appearance of ragwort flowers . The absence of ragwort flowers d id not affect ovary development in fl ies reared in the laboratory but the necessary nutr ients must be obtained from sources other than ragwort rosettes . A model based on the ovary development rates of fl ies prov ided with artif icial d iets and a range of flower ing field plants showed that the preoviposition per iod in the field was extended because of the absence of ov iposi tion si tes . Females also laid eggs on flower ing ragwort transplanted at the field site prior to the onset of flower ing in the field population . This lack of synchrony between seedfly emergence am ragwort flower ing resulted in competition for ov iposition sites when ragwort first flowered . The number of buds available for ov iposition per seedfly ranged from 0 · 14 to 106 - 7 in 1982/83 and 0 - 03 to 10 · 6 in 1983/84 . Multiple infestations occurred in 22% and 16% of the seedheads in 1982/83 and 1983/84 respectively. Fecundity was also low in both years . The nunber of eggs laid per female emerged at Redwooos Forest was 12- 9 in 1982/8 3 and 3 - 9 in 1983/84 . 1 1 7 . OVerall infestation levels were 10% and 1% at Redwoods Forest and the Desert Rd respectively in 1982/831 and 20% and 3% in 1983/84. Ragwort density was constant at Redwoods Forest throughout the study per iod with an estimate of 28, 788 stems per hectare in 1982/83 and 30, 132 in 1983/84. The increase in infestation levels in the second year can be attr ibuted partly to the increase in the number of fl ies that emerged at this s ite . Approxbnately 1, 869, 000 fl ies emerged in 1982/83 canpared with 195, 360 in 1983/84. The number of ragwort seedhead s exceeded the egg laying capacity of the seedfly population in 1982/83 wi th 404 seedheads produced by ragwort dur ing the sumner per female that emerged at the s ite canpared to 39 seedhea::Is per female in 1 983/84. Seedfly females oviposited in seedheads with a diameter ranging fran 3 - 8 to 5 - 8 rrm and a disc floret len:Jth ranging fran 2 -7 to 5 - 6 rtm. The d istr ibution o f seedfly eggs wi thin seedheads was uniform in early December and January in both years. The mean number of seeds remaining in seedfly infested seedheads ranged fran 7- 6 to 17- 6 and the percentage germination ranged fran 28% to 43%. Seedfly mortality from halfway through the egg stage to halfway through the third larval instar was estbnated as 33% in 1982/83 and 60% in 1983/84. Third instar larvae about to pupate leave the seedheads in conditions of high surface moisture so that the duration of the third larval instar may have been extended in the dry conditions of the controlled environment roans. The number of third instar larvae within seedheads may have been under estimated as a result. High estimates of third instar larvae which had left ragwort to pupate were obtained because sampling coincided with conditions favourable for larval d ropping . Temperatures as low as - 15C for short periods do not haDTI 1 1 8 . seedfly pupae and temperatures must reach lS-20C to initiate d iapause . Estlinates of pupal mortal ity over winter ranged fram 14 - 3% to 57% . In view of the high numbers of seeds escaping predation , the high germinating capacity and longev ity of ragwort seed and ragwort ' s abil ity to reproduce vegetatively, seedfly ' s linpact on ragwort populations is considered neglig ible . 1 1 9 . References Andres , L .A. and Davis , C . J . ( 1973) . The biological control of weeds with insects in the united states . Proceedings of the 2nd International Symposium on the Biological Control of Weeds pages 11-28 . Andres , L .A . , Dunn , P . H . , Hawkes , R.B. and Maddox , D.M. ( 1976) . Current happenings in b iological control . proceedings of the 28th Annual cal ifornian Weed Conference pages 81-87 . Andrewartha , H .G. and Birch , L .C . ( 1954) . The d istr ibution and abundance of animals . University of Chicago Press , 782 pages . Bascand , L . D . and Jowett , G .H . ( 1981 ) . Scrubweed cover of South Island agr icultural and pastoral land . New Zealand Journal of Experimental Agr iculture 1: 307-327 . Benn , M . , DeGrave , J . , Gnanasunderam, C. and Hutchins , P . ( 1979) . Host-plant pyrrol iz idine alkaloids in Nyctemera annulata Boisduval : Their persistence through the l i fe cycle and transfer to a parasite . Separatum Experientia 35 : 731-732 . Bernays , E .A . ( 1981 ) . plant tannins and insect herbivores : appraisal . Ecological Entomology �(4 ) : 353-360 . an Burdon , J . J . and Marshall , D .O . reproductive mode of weeds . 649-658 . ( 1981) . Biological control and the Journal of AWlied Ecology 18 : Cairns , D. ( 1938) . Vegetative propagation of ragwort . New Zealand Journal of Science and Technology 20 : 173A-183A. Cameron , E . ( 1935) . A study of the natural control of ragwort ( Senecio jacobaea . L) . Journal of Ecology 23 : 265-322 . Coaker , T . H . , and Finch , S. ( 1973 ) . The association of the cabbage root fly with its food and host plants . Symposium of the Royal Entomological Society London �: 119-128 . Coaker , T . H . , and Wright , D .W. ( 1963) . Influence of temperature on emergence of the cabbage root fly from overwintering pupae . Annals of Appl ied BiOlogy 52 : 337-343 . Collier , R. H . and Finch , S . ( 1983 ) . Effect of intensity and duration of low temperatures in regulating d iapause development in the cabbage root fly Del ia rad icum. Entomologia Experimentalis et Appl icata 34 ( 2) : 193-200 . Coll in , J . E . Senecio . 53-54 . ( 1936) . A note on Anthomyidae reared from the flowers of Entomologists Record and Journal of Variation 48 ( 5) : Connor , H . E . ( 1977 ) . The poisonous plants in New Zealand Second - --- - ---� 1 2 0 . Edition . Government Pr inter Well ington , New Zealand , 247 pages . Cottier , w. ( 1931) . The blue stem-borer of ragwort . New Zealand Journal of Agr iculture 4 2 : 333-337 . Cox , C .S . and McEvoy, P .B. ( 1983 ) . Effect of summer moisture stress on the capacity of ragwort to compensate for defol iation by Tyria jacobaeae . Journal of Appl ied Ecology 20 ( 1 ) : 225-234 . Crawley, M .J . ( 1983) . Herbivory: the dynamics of Animal- Plant Interactions . Studies in Ecology : Vol 10 . Blackwell Scienti fic Publications , 437 pages . DeBach , P . ( 1974) . Biological control by natural enemies Cambridge Universi ty Press , 323 pages . Deinzer , M.L . , Thomson , P .A . , Burgett , D .M. and Isaacson , D . L . ( 1977) . Pyrrol i z id ine alkaloids : their occurrence in honey from tansy ragwort ( Senecio j acobaea L . ) . Science 195 : 497-499 . Dempster , J . P . ( 1982) . The ecology of the cinnabar moth , Tyria jacobaeae L . ( Lepidoptera : Arcti idae) . Advances in EcOlogical Research 13 : 1-36 . Dickonson , J .O . and King , R. P . ( 1978) . The transfer of pyrrol i z id ine alkaloids from Senecio jacobaea into the milk of lactating cows and goats . In Effects of poisonous plants on l ivestock Joint united StateS-Austral ia Symposium: 201-208 . Ead ie , I .McL. and Robinson , B .D . ( 1953) . Control of ragwort by hormone type weedicides . Journal of the Austral ian Institute of Agr icultural Science 19 : 192-196 . Finch , S . ( 1971 ) . The fecundity of the cabbage root fly Erioischia brassicae (Bouche) under f ield conditions . Entomologia Exper imental is et Applicata 14 : 147-160 . Finch , S . ( 1974) . Sugars available from flowers visited by the adult cabbage root fly Er ioischia brassicae (Bch) Diptera : Anthomyi idae . Bulletin of Entomological Research 64 : 257-263 . Finch , S. and Sk inner , G . ( 1975) . Dispersal of the cabbage root fly. 1-19 . Annals of �ed BiOlogy 81 : Forbes , J .C . ( 1977 ) . Population flux and mortal ity in a ragwort infestat ion . weed Research 17 : 387-391 . Forster , J .M. ( 1977) . Aspects of the biology of Apion ul icis ( Forster ) Coleoptera : Curcul ionidae . MSc Thesis , Auckland University, New Zealand . Fr ick , K . E. ( 1969) . Un i ted Sta tes • Attempt to establ ish the ragwort seedfly in the Journal of Economic �tomology 62 : 1135-1138 . Fr ick , K . E . ( 1970) . Behav iour of adult Hylemyia se�ec�ella , an 1 2 1 . Anthomyi id (Diptera) used for the biological control of tansy ragwort . Annals of the Entomological Society of America 63 ( 1) : 184-187 . Frick , K . E. and Andres , L .A . ( 1967) . Host specificity of the ragwort seed fly , Hylanya seneciella . Journal of Economic Entomology 60 ( 2) : 457-463 . Funderburk , J . E . , Higley , L .G . and Pedigo , L . P . ( 1984 ) . Seedcorn maggot Del ia platuraMeigen (Anthomyi idae) . Phenology in Central Iowa and examination of a thermal-un it system to predict development under field conditions . Environmental Entomology 13 ( 1) : 105-109 . Gibbs , H . S . ( 1965 ) . Volcanic ash soils in New Zealand . New Zealand D .S . I . R . Information Series No . 65 . Gough, H .C . ( 1946) . Studies on the wheat bulb fly Leptohylemyia coarctata Fall . Numbers in relation to crop damage . Bulletin of Entomological Research 37 : 439-4 54 . Green , H . E . ( 1937) . Dispersal of Senecio jacobaea . Journal of Ecology 25 : 569 . Greig-Smith , P . ( 1964 ) . Quanti tative plant Ecology Second Edition . Butterworth London , 198 pages . Hafez, M. , El-Ziaday , S . and Dimetry, V . Z . ( 1970 ) . Studies on the bionomics of the beetfly pegomyia hyoscyami . Bulletin of the Entomological Society of Egypt 54 : 433-450 . Hairston , N .G . , Smith , F . E . and Slobodkin , L .B. ( 1960) . Communi ty structure , population control and competition . American Natural ist 94 : 421-425 . Harper , J .L . Isles . and WOOd , W.A. ( 1957) . Biological flora of the Brit ish Journal of Ecology 4 5 : 617-637 . Harris , P . ( 1980) . Stress as a strategy in the biolog ical control of weeds ( abstract) . Proceedings of the 5th International symposiun on the B iological Control of Weeds page 47 . Harr is , M .O . and Miller , J . R . ( 1983) . Colour stimul i and oviposition behav iour of the onion fly Del ia ant iqua Meigen (Anthomyi idae) . Annals of the Entomological SOCiety of Amer ica 76 ( 4 ) : 766-771 . Harr i s , P . , Thompson , L . S . , Wi lkinson , A .T .S . and Neary , M. E . ( 1976) . Reproductive biology of tansy ragwort , cl imate and biological control by the cinnabar moth in Canada . Proceedings of the 4th International Symposium on the BiOlogical Control of Weeds pages 163-173 . Hawkes , C .S . ( 1975 ) . Phys iolog ical cond it ion of adult cabbage root fly ( Er io ischia bras�icae) attracted to host plants . Journal of 1 2 2 . Appl ied Ecology 12 : 497-506 . Hawkes , R .B . and Johnson , G .R . ( 1976 ) . Longitarsus jacobaeae aids moth in the biological control of tansy ragwort . proceedings of the 4th International Symposium on the Biological Control of Weeds pages 193-196 . Hawkes , C .S . , Patton , S . and Coaker , T . H . ( 1978 ) . Mechanisms of host plant finding in adult cabbage root fly, Del ia brassicae . Entanologia Experimental is et Applicata 1! : 219-227 . Holloway , B.A. ( 1983) . Species of ragwort seedfl ies imported into New Zealand (Diptera : Anthanyi idae) . New Zealand Journal of Agr icultural Research 26 : 245-249 . Hoy , J .M . ( 1958 ) . The collection of Hylemyia seneciella (Meade) (Diptera , Muscidae) for shipment to Austral ia . New Zealand Journal of Science ! ( 3 ) : 417-422. Hoy , J .M. ( 1960) . Collection of Hylemyia seneciella (Meade) in 1959 season . New zealand Journal of Sc ience 1 ( 1) : 100-102 . Hoy , J .M. ( 1964) . Present and Future prospects for biological control of weeds . New Zealand Science Rev iew 22 : 17-19 . Huffaker , C .B . ( 1957) . Fundamentals of biological control of weeds. Hilgard ia 27 : 101-157 . Huffaker , C .B . ( 1964 ) . Fundamentals of biological weed control . In Biological Control of insect pests and weeds , P. DeBach (ed) . London , Chapnan and Hall , 844 pages . Humason , G .L . ( 1967) . Francisco , W. H . Animal t issue techniques Second Edition . Freeman , 569 pages . rnoue , H . ( 1983 ) . Nymphal cannabalism in response to ov iposition behaviour of adults in the assassin bug Agr iosphodrus dohrni Signoret . Researches in Population Ecology 25 : 189-197 . San I saacson , D .L . and Ehrensing , D .T . ( 1977) . Biolog ical control o f tansy ragwort . Oregon Department of Agr iculture Weed Control Bul letin No . 1 . I s l am , Z , ( 1 9 81 ) , M S c Thes is U n i vers i ty of Lon don Islam , Z. and Crawley , M. J . ( 1983 ) . Canpensation and regrowth in . ragwort attacked by cinnabar moth . Journal of Ecology 71 ( 3 ) : 829-843 . Janzen , D . H . ( 1970 ) . tropical forests . Herb ivores and the number of tree species in Amer ican Natural ist 104 : 501-528 Jones , M. ( 1970) . Observations on feeding and egg development of the wheat bulb fly �tohylemyia coarctata ( Fal l . ) . Bulletin of Entomolog ical Research 60 : 199-207 . 1 2 3 . Jul ien , M.H. ( 1982) . Biocontrol of weeds : a world catalogue of agents and their target weeds . Ccmnonweal th Insti tute of Biological Control 108 pages . Kelsey , J .M. ( 1937) . The ragwort leaf-miner ( Phytomyza atr icornis Mj.) am its parasite (Dacnusa areolaris Nees . ) . New Zealam Journal of Science am Technology 18 : 762-763 . Kelsey, J .M. ( 1955) . Ragwort seed fly establ ishment in New Zealam . New Zealam Journal of Science am TechnologyA 36 ( 6 ) : 605-607 . Lloyd , M. ( 1967) . l'1ean cro�Hng . Journal of Animal Ecology 36 : 1-30 . Long , D .B . ( 1958b) . Field observations on adults of the Wheat bulb fly Leptohylemyia coarctata Fall . Bulletin of Entomological Research 49 : 77-94 . Mattson , W.J . am Addy , N .D . ( 1975) . Phytophagous insects as regulators of forest pr imary production . Science 190 : 515-522 . McEvoy, P .B . ( 1984) . Seedl ing d ispersion and the persistence of ragwort , Senecio jacobaea (Compositae) , in a grassland dominated by perennial species . Oikos 42 ( 2 ) : 138-143 . Meijden , E . van der . ( 1976 ) . Interactions between the cinnabar moth am tansy ragwort . ProceediNJs of the 4th International Symposiun on the Biological Control of Weeds pages 159-162 . Mei jden , E . van der am Waals-Kooi , R .E . van der . ( 1979) . Population ecology of Senecio jacobaea in a dune system. I . Reproducti ve strategy and biennial habi t . Journal of Ecology 67 : 131-153 . Mil ler , D . ( 1970 ) . Biolog ical control of weeds in New Zeal am 1927-1948 . New Zealand D .S . I .R . Information Ser ies No . 74 . Mortimer , P . H . and White , E . P . ( 1975) . Tox ic ity of same composi te ( Senecio) weoos . proceediNJs of the 28th New Zeal am Weed am Pest Control Conference pages 88-91 . New Zealand Meteorological Service ( 1983) . Summar ies of Cl imatOlogical observations to 1980 . Miscellaneous Publ ication No . 177 page 90 . Poole , A .L . ( 1938a) . Botanical stud ies on ragwort . New Zealand Journal of Agr icul ture 56 : 83-90 . Poole , A. L . ( 1938b) . Germination of ragwort in water . New Zealand Journal of Agr icul ture 57 : 95-96 . Poole , A .L . and Cairns , D . ( 1940 ) . ( Senecio jacobaeae .L) control . No . 82 , 62 pages . Botanical aspects of ragwort New Zealand D .S . I . R Bulletin 1 2 4 • Papay, A. I . , Lyttle , L .A. , Edmonds , O .K . and Phung , H .T . ( 1984 ) . Incidence of the nodd ing thistle receptacle weevil on nodd ing and slender winged thistle . Proceedings of the 37th New Zealand Weed and Pest Control Conference pages 28-32. Prokopy, R.J . ( 1972 ) . Evidence for a marking pheromone deterr ing repeated oviposition in the apple maggot fl ies . Environmental Entomology .!: 326-332 . Prokopy, R .J . ( 1981) . OViposition-deterr ing pheromone system of apple maggot fl ies . In Management of Insect Pests with ssmiochemicals , E .R . Mitchell (ed) . Plenum 1981 , pages 477-494 . Radcl i ffe , J . E . ( 1969) . Agriculture 119 ( 1 ) : Ragwort control . New Zealand Journal of 80-83 . Raw, F . , Jones , M.G. and Gregory, P . H . ( 1968 ) . The food of female wheat bulb fl ies ( Leptohylemyia coarctata Fall . ) . plant pathologY 17 : 23-25 . Schmidl , L . ( 1972) . Biology and control o f ragwort , Senecio jacobaea . L in Victor ia , Austral ia . Weed Research 12 ( 1 ) : 37-45 . Schroeder , D . ( 1983) . Biolog ical control of weeds . In Recent Advances in Weed Research Edi tor W.W.Fletcher , Commonwealth Agr icultural Bureau pages 41-77 . Sheldon , J .C. and Burrows , F.M. ( 1973) . The dispersal effectiveness of the achene pappus uni ts of selected Campositae in steady winds with convection . New Phytologist 72 : 665-675 . Sidd iqui , W .H . and Barlow, C.A. ( 1972) . Development under constant and var iable temperatures in Drosophila species . Annals of the Entomological Society of Amer ica 65 : 993-1001 . Singh , P . ( 1977) . Artificial d iets for insects , mites and spiders IF1/Plenum Press 594 pages . Smallfield , P .W. ( 1970 ) . The grasslands revolution In New Zealand Hodder and Stoughton , Auckland , 151 pages . Smiley , J .T . ( 1985) . Are chemical barr iers necessary for evolution of butterfly-plant associations? Oecologia 65 : 580-583 . Smith , B .A. J . ( 1982) . Ragwort biology and control . Advisory Serv ices Div ision booklet , New Zealand Min i stry of Agr icul ture and Fisher ies 20 pages . Southwood , T . R . E . ( 1978) . Ecological Methods; with particular reference to the study of insect populations Second Edition . Chapman and Hall , 524 pages . stimac , J . L . and I saacson , D . L . ( 1976 ) . C i nnabar moth a s a biolog ical control agent of tansy ragwort : comparison o f the 1 25 . population dynamics in Englarrl arrl oregon . proceedings of the 4th International Symposium on the Biological Control of weeds pages 155-158 . Swarbrick , J .T . ( 1983) . Economic j ustification of weed research and control at Government level . Austral ian weeds l ( 3 ) : 86 . Syrett , P . ( 1983) . Biological control of ragwort in New Zealarrl : a review. Austral ian weeds � ( 3 ) : 96-101 . Theunissen , J . ( 1974 ) . Effects of temperature on egg chamber development in the onion fly, H�lemyia antiqua (Dipt .Anthomyi idae) . Entomologlca Experilnental is et Appl icata 17 : 355-366 . Thompson , A. ( 1974 ) . Herbicide effects on ragwort and pasture . proceedings of the 27th New Zealarrl weed arrl Pest Control Conference pages 90-93 . Thompson , A. ( 1977) . Herbicides for the spot treatment of ragwort in pasture . ProceErlings of the 30th New Zealarrl Weed am Pest Control Conference pages 34-37 . Thompson , A . and Makepeace , W. ( 1983) . Longevity of bur iErl (Senecio jacobaea L . ) seErl . New Zealarrl Journal of Experilnental Agr iculture 1 1 : 89-90 . Thomson , G .M. Zealand • ( 1922) . The natural isation of anlinals and plants in New Cambr idge University Press , 607 pages . Tisdell , C .A . , Auld , B.A. and Menz , K.M. ( 1984 ) . On assessing the value of biocontrol of weErls . Protection Ecology �(2 ) : 169-179 . Traynier , R .M.M. ( 1967a) . Effect of host plant odour on the behav iour of the adult cabbage root fly Er ioischia brassicae . Entomologica Experilnental is et Appl icata 10 : 3 21-328 . Wagner , T . L . , Wu, H . , Sharpe, P .J .H . , Schoolfield , R.M. and Coulson , R .N . ( 1984) . Modell ing insect development rates : A l i terature rev iew and appl ication of a biophysical model . Annals of the Entomological Society of Amer ica 77 ( 2 ) : 208-220 . warrington , I . J . , Di xon , T . , Rothbotham, R.W. and Rook , D .A. ( 1978) . Light systems in major New Zealand control1Erl env ironment facil i ties . Journal of Agr icultural Engineer ing Research 23 : 23-36 . Waterhouse , D .F . ( 1967) . The entomolog ical control of weeds in Austral ia . Symposium of the 11th Paci f ic Sc ience Congress . Conta ined in Mushi 39 : 109-118 . Whi te , E .P . ( 1969 ) . Alkaloids of some herbaceous Senecio species in New Zealand . New Zealam Journal of Science 12 : 165-170 . 1 2 6 . Wilkinson , J .D . and Daugherty , D .M. ( 1970) . Comparative development (Diptera : Sciaridae) under constant and variable temperatures . Annals of the Entomological Society of �er ica 63 : 1078-1083 . Wilson , F . ( 1964) . Entomology 2.: The biological control of weeds. 225-244 . Annual Review of Zimmennan , M . ( 1979) . OViposition behaviour and the ex istance of an oviposition-deterr ing pheromone in Hylemya . Environmental Entomolesy 8 : 277-279 . Zimnerrnan , M. ( 1980) . Selective deposition of an oop by Hylemya Environmental Entomology 2.: 321-324 . Z imnennan , M. ( 1982 ) . Facultative deposition of an oviposition­ deterr ing pheromone by Hylemya . Env ironmental Entomology 11 : 519-522 . Zimmennan , M. , Cibula , D.A. and Schulte , B. ( 1984 ) . OViposition Behaviour of Hylemya (Del ia) sp : Suboptilnal Host Plant Choice? Environmental Entomology 13 : 696-700 . A ppen d i x 1 . a ) 1 9 8 2 / 8 3 + ++ + - + . 0 + - • + J '0 • I � �... .. • • • + .. " .� . .: : � . • • • I. •• � . :- . .. + :: , . .. . . � . . # � 0 Fe -+ + • .f. ++:1:+ -:t:+ V + • II. t-!- ... + • • .. f! 0 . • + + � + + -+t t:t: o� r- I-- + + + I-- I-- f.!. • 0 + + 0 D i s t r i bu t i�n map o f ragwort ro s e t t e s in two 3 0 0 m s ubpl o t s o f t h e P o l e s P l o t ... � . .f.+ : ++ + ... + + + 00 . . . ' + OW . T,,\ + : . . er· o _ . . .. • 0 · .. . ' -> :. . . ." . - I . . . • . . .. . . . . o • -+' • 0' - :'l I' - '" - + +-- .. . .- . o . .. - . . : o + + ow . . . . · . .\, + ' - o '. a. • • ,-li o . "'++ _ . . I + + 0+ - • - , .-; : .... . J • . . .. . -.... - 1 2 7 . S c a l e � = 2 m e t r e s + ro s et t e s p r e s en t in S p r i n g 1 9 8 2 wh i c h fa i l ed to f l ower • r o s et t e s wh i c h fl owered o v e r Summer T h e ma p on t he l e f t r ep resen t s t he mos t no r t he r n s u b p lo t ( se e F ig . 2 3 ) � ) 1 9 8 3 / 8 4 + ?- $ x • x + r- x x x x l x x xx e xx xx . f- X xx xx • x • + + . ,;.x + r- t + � e x x . x �t � + f±� + f- r- f- f-" f- r-x . r- · x r- r- e+- x x x x Xx � + � x· +". • + + . x + + + x+·t xx x • x x x .. x X � + + •• �. xx x · x x x+ + x + x .'If<. )OO()( x . ? .. � +x x + x . xx � ++ ++ + x + ++ x x x x x �) x x x x + x Xt >