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. Aspe2cts of the Accumulation of Cobalt, C opper and N ickal by Plants A thesis presented i n parti al ful filment of the requirements for the degree of Doctor of Philosophy in Chemistry Massey University Richard Stephen Morrison 1 980 ti"aao_J� Abs t r ac t Hyperaccumulation of heavy metal s was studied with the intention of elucidating the mechanisms of tolerance of hyperaccumul ator plant species. Two main areas are covered; cobalt and copper accumul ation by plants from Shaba Province , Z aire , and nickel accumul ation by species of the genus Alyssum. i i I n surveys of vegetation of metalliferous soils of Shab a , nine or ten new hyperaccumulators of cobalt were discovered along with eight or nine very strong accumulators. For copper , seven hyperaccumul ators and fi ve or six very strong accumulators were di scovered. Same families contained a higher frequency of hyperaccumul ators than others. There i s also a di fference i n superarder cl assification of cobal t and copper hyperaccumul ators on one hand and nickel hyperaccumulators on the o ther. Surveys o f the genera Aeol anthus , Ipomoea and Pandiak a were made but only one new copper hyperaccumulator was found : no new cobalt hyperaccumul ators were found. Several species had their ab ilities to accumul ate confi rmed. Pot trial s on three hyperaccumulators Aeal anthus biformifalius , Haumani astrum katangense and �· rabertii , showed accumul ation o f cobalt but not the expected accumulation of copper. The uptake curve was o f the exclusion-breakdown form . The limit of b re akdown , far each metal , was similar from speci es to species . Cob alt was less readily excluded than copper . The tolerance tests showed that some species have individuals wi th greatly enhanced abilities to survive higher metal cancentrations than i s normal far that specie s , while other species have more uni form tolerances . There appears to b e no requi rement fo r l arge metal concentrations at germination and seeds germinate more readily in the absence rather than the presence of the metal s . The di stribution of cobal t and copper within leaf tissue s , of five specie s , appears to be parallel within each species . For each metal , the distribution is parall el between di fferent species with the exception of Buchnera metallorum . More detailed studies on cob alt in H . robertii showed the di stribution to b e even over the leaf area but with small anomal ous regions of high concentration . The possibi l i ty that some of the cobalt was preci pitated as oxal ate crystal s is considered. The water-soluble cobal t fraction l igand could not be i dentified but was not protein­ aceous . I t has a mass of 5 , 200 g per mol e of cobal t . i i i A survey of t he genus Alyssum revealed thi rty-four taxa as hyperaccumul ators to add to the fourteen previously known . All the taxa are from section Odontarrhena. The geographical distribution of the hyperaccumulators is discussed a s is the possible evolution of hyperaccumul ators in subsections Compressa and Samari fera from non-accumul ators within them . Studies of nickel accumul ation by el even Alyssum species and the closely related Bornmuellera tymphaea showed similar characteri stics for all hyperaccumulators but two non-accumulators differed. A ri se-to-saturation uptake fo rm was noted. In the absence of nickel , cob al t could be accumul ated with a similar uptake fo rm. Cob alt accumul ation i n the presence o f nickel i s unknown . The rate_of uptake i s rel ativel y rap id . The tolerance of hyperaccumul ators to high nick el concentrations was confirmed in two types of tolerance tests; a substrata medium test and a solution test . The resul ts from the two tests are compared. The distribution of nickel between the plant organs is discussed. The analysis of mineral elements in l eaf materi al showed interesting di fferences b etween hyperaccumul ators and non-accumul ators for calcium , magnesium and manganese content but these could not be rel ated to differing nickel concentrations . A similar find was made for glucosinol ates . An organic aci d survey was restricted by the non-identi fi cation o f many aci ds. Separation of the nickel complexes was made . I dentification of li gands involved in nickel compl exation was attempted but few positive resul ts were found . Two ligands were common in signi ficant quantities for all species studi ed . The resul ts of these experiments were used to discuss possible evolution of hyperaccumulator speci es both in terms of thei r superorder distribution and their method o f metal ion uptake . An equilibrium mechani sm of uptake i s proposed which involve s a multipl i city of compl exes for the ion absorbed . The mechani sm di ffers from that which is commonly proposed for micronutrient elemental uptak e . i v Ac knowle2d gczmcznts I would like to thank all those peopl e who assisted or advi sed in any way during thi s projec t . I n parti cul ar thanks go to my supervi sors Drs . R. R . B rooks and R. D. Reeves for thei r constant advice and o versi ght. Special thanks also go to Prof. F. Malaisse and Monsieur J. Gregoi re (Universi te Nationals du Z aire ) and the ir eo-workers for the samples from the Shab an metalli ferous areas. Thanks al so go to the director and Monsieur A. Taton of the Jardin Bo tanique N ational de Belgique herbarium , and the institutions and persons l i sted in Appendix I for the suppl y of sampl es or seeds. To Prof. R. Hedges , Drs . I. G. Andrews , K. G. Couchman , Mr. C . Bi shop ( all Department of Chemi stry/Biochemi stry , Massey Uni versi ty) , Dr. G.J. Shaw (DSIR) , Prof. P. Horowitz ( Harvard Universi ty) , Dr. G. Merriam ( NH I , Bethesda) and o thers go my thank s fo r certain analyses ( mass spectral , proton microprobe) and advice on technique s . V Finally to �rs M. Wade ( typist ) , Mrs J. Trow ( illustrator) and Monsieur L. Lemaire ( Shaban photographs) my thanks for tasks well done . ... Table of Con tent s PAGE Abstract ii Acknowledgements V Table of Contents List of Figures List of Tables Chapter 1 . Introduction Part One: Cobalt and Copper in Vegetation of Shaba Province, Za'ire vi ix xii 1 Chapter 2 . Surveys of Vegetation of Metalliferous Soils 22 2 .1 Introduction 23 2 . 2 Analytical Methods 32 2 . 3 Cobalt and Copper in Vegetation of Fungurume 34 2.4 Other rlcral Surveys 55 2 .5 General Discussion 61 Chapter 3. Surveys of the Genera Aeclanthus, Ipomoea and Pandiaka 65 3 .1 Introduction 66 3 . 2 Analytical Methods 68 3 . 3 Aeclanthus Mart. (Lamiaceae) 69 3 . 4 Ipomoea L. (Ccnvclvulaceae) 74 3.5 Pandiaka (Mcq. ) Heck. f. (Amaranthaceae) 77 3. 6 General Discussion 79 vi Chapter 4 . Biogeochemical Studies on Three Metallophytes from Shaba 82 4 . 1 Introduction 83 4 .2 Experimental Methods 85 4 . 3 Cobalt and Copper Uptake 88 4 .4 Tolerance Tests 95 4 .5 Germination Tests 96 4 .6 Copper Accumulation by Aeolanthus biformifolius 98 4 .7 General Discussion 99 Chapter 5. Phytochemical Studies on Some Metallophytes from Shaba 101 5. 1 Introduction 102 5.2 Experimental Methods 104 5. 3 Distribution of Heavy Metals in Plant Tissue Extracts 110 5.4 Proton Microprobe Studies on Haumaniastrum robertii 114 5.5 Cobalt Complexes in Haumaniastrum robertii 119 Part Two: Nickel Accumulation by the Genus Alyssum L. Chapter 6. Survey of the Genus Alyssum 124 6 .1 Introduction 125 6 . 2 Analytical Methods 127 6 . 3 Results and Discussion 128 Chapter 7. Biogeochemical Studies on some Alyssum Species 150 7. 1 Introduction 151 7.2 Experimental Methods 154 vii 7 . 3 Nickel Uptake in Alyssum Species 1 57 7 . 4 "Trigger-Point" Theory Test 1 65 7 . 5 Rate of Uptake 1 67 7 . 6 Nickel Tolerance Studies 169 7 .7 Cobalt Uptake Studies 1 72 7 . 8 Bornmuellera tymphaea Studies 176 Chapter 8. Phytochemical Studies on Some Alyssum Species 179 8 . 1 Introduction 180 8 .2 Experimental Methods 182 8 . 3 Nickel Distribution Within the Plant Organs 1 87 8 .4 Mineral and Organic Component Analyses 187 8 . 5 Nickel Complexation 1 96 Chapter 9 . Concluding Discussion 224 References Cited 244 Appendices· 268 I. Coopera�ing Institutions and Persons 269 II. Hyperaccumulators Recorded in the Literature 271 III. Hyperaccumulators Discovered in this Work 280 IV. Some Terminology of Accumulation 285 Publications Arising from this Thesis 287 viii Li st of Fig u r Q s and Plate s FIGURE/PLATE 1 . 1 Cumulative frequency plot for cobalt and copper in plants from mineralized areas of Shaba Province, Zaire. 2. 1 Distribution of metalliferous rocks in Shaba Province, Za'!re. 2 . 2 The major cobalt and copper mining areas of PAGE 15 26 Shaba Province, Zaire. 28 2 . 3 View of the cupriferous hillocks of Fungurume. 35 2 . 4 The contrast between non-cupriferous and cupriferous hillocks . 37 2 . 5 View of the cellular siliceous rocks. 39 2 . 6 Location map of the cupriferous hillocks of Fungurume. 41 2 . 7 Representation of vegetation on metalliferous hillocks at Fungurume. 47 3.1 Transverse sections through the laminae of three Aeolanthus species. 72 3. 2 Foliar indumenta for three metal-tolerant species of Aeolanthus. 73 4 .1 Accumulation seedlings. curve for Haumaniastrum katangense 4 .2 Accumulation curve for Haumaniastrum robertii seedlings. 4 . 3 Accumulation curve for Aeolanthus biformifolius seedlings. 4 .4 Accumulation curve for Aeolanthus biformifolius grown from corms. 5.1 Proton microprobe micrographs showing the distribution of cobalt, calcium and potassium in a leaf. 89 90 91 94 115 ix 5.2 Line scan across leaf of Haumaniastrum robertii showing elemental concentrations. 11? 5. 3 X-ray spectra from proton microprobe. 118 5.4 Electrophoretogram of water-soluble cobal t complex extracted from Haumaniastrum roberti i . 12 1 6.1 Histograms of nickel concentration in Alyssum species . 13? 6. 2 Distribution , by vilayets , of Turkish species of section Ddontarrhena. 139 6. 3 Geographical distribution of species of section Odontarrhena i n the western Irano-Turanian region. 140 6.4 Geographical distribution of outlying specimens of species of section Ddontarrhena. 141 6 .5 The distribution of Alyssum species: section Ddontarrhena, subsection Compressa, series Integra. 143 6. 6 The distribution of Alyssum species: section Odontarrhena , subsection Compressa , series Crenulata. 144 6.? The distribution of Alyssum species: section Ddontarrhena, subsection Samarifera. 146 ?. 1 Nickel accumulation in Alyssum species as a function of substrata content. 158 ?. 2 Growth of some Alyssum species in nickel uptake trials . 161 ?. 3 Plot of the concentration index at low substrata concentrations of nickel . 166 ?.4 The rate of uptake of nickel by Alyssum euboeum. 168 ?.5 Tolerance tests involving new-root lengths of excised seedlings of Alyssum species grown in varying ni ckel solutions. 1?0 ?. 6 Cobalt uptake by Alyssum· species. 1?3 ?.? Nickel accumulation inBornmuellera tymphaea, as a function of substrata content, in comparison with the Alyssum species in figure ?. 1. 1?8 X xi 8.1 Plant organs analysed for nickel . 188 8 .2 Gas-liquid chromatographs of the methylated derivatives of the nickel-complexing ligands from Alyssum species. 1 9? 8. 3 Mass spectra of the methylated derivatives of the nickel-complexing ligands. 208 9 . 1 Proposed mechanism of nickel uptake by Alyssum species. 231 9 .2 Proposed sequence for development of various uptake mechani sms. 240 Li s t of Ta biQs TABLE PAGE - 2 .1 Total and extractable soil cob al t and copper in some Shaban soils . 42 2 .2 Plant communities on mineralized hillocks at Fungurume , Z alre . 45 2 . 3 Cobal t and copper concentrations in plants from mineralized hillocks at Fungurume , Zalre. 49 2 .4 Cobalt and copper concentrations in plants from Lupoto. 56 2 . 5 Cob alt and copper in vegetation from Ruashi . 57 2 . 6 Cobal t and copper in vegetation from Lubumbashi . 59 2 .7 Cobalt and copper in vegetation from Mindingi . 60 2 . 8 Distribution and advancement of cob alt and copper hyperaccumul ators. 63 3 . 1 Cobalt.and copper concentrations in speci es of 3 . 2 3 . 3 4 . 1 4 . 2 Aeolanthus . ?0 Cobalt and copper concentrations in some IE!omoea species . 75 Cobalt and copper concentrations in Pandi aka species . ?8 Germination of Haumaniastrum robertii seeds . 9? Germination of Aeolanthus bi formifolius seeds. 9? 4 . 3 Copper accumulation in Aeolantt.lus b iformifolius during the growing s��son . 99 5 . 1 The fractionation of cob�lt and copper in plant tissue extracts . 111 5 . 2 Statistical data for cobalt and copper associ ations i n plant fractions . 1 12 xii 6. 1 Nickel concentrations in Alyssum L. species. 1 29 7.1 Nickel uptake by Alyssum species growing in serpentine soils . 164 7.2 Tolerance l evels of Alyssum species tested by the soil culture method. 171 8.1 Nickel distribution within Alyssum heldreichii . 189 8 . 2 Nickel distribution within Alyssum serpyllifolium ssp. lusitanicum. 190 8 . 3 Nickel distribution within Alyssum murale . 191 8 .4 Concentrations of mineral elements in three Alyssum taxa. 192 8 .5 Organic acids and glucosinolates in thFee Alyssum taxa. 194 xiii CHAPTER 1 Introduc tion The occurrence of characte ristic plants on metal­ contaminated soil s has been ob served for centuri es . Thalius ( 1 588) noted that Minuartia verna (L . ) Hiern. was an i ndicator of metal contamination whi le Caesalpino ( 1 583 ) noted a species of plant ( prob ably Alyssum bertolonii Desv . ) restricted to serpentine rocks in I taly. Agricola ( 1 556 ) had previously described the anomalous appearance of plants growing over veins and metal ore outcrops. Interest i n the pl ants of these anomalous so i l s has continued for several reasons : ( 1 ) to identify the methods by which the plant i s able to adapt to the toxic envi ronment and associated with this ; ( 2 ) how to make fertile or productive those areas covered by these soils; ( 3 ) the use of these plants for geobotanical and bi ogeochemical pro specti ng; (4) the interest in very high concentrations of me tal s in pl ants and ( 5 ) the associ ated possibi l ity af l ow-energy extraction of me tals from thei r ore s by imi tati on of nature . Those pl ants which are restri cted to growing on these soil s are called me tall ophytes. However , many variations o n the basic description o f thi s term have been used by di fferent autho rs. Antonovics et al . ( 1 97 1 ) have di scussed this range of vari ations. Fo r the purposes of this thesi s , a metallophyte is taken as a species of pl ant with a h igh tol erance to elevated heavy metal contents in the soi l . Such a species i s often endemic to tho se s o i l s which contain elevated quantities of the heavy metal s . Most metallophytes survive on metal-rich soils by excluding the metal from entry into the plant . However , there exists a range of species , called accumul ato rs , which accumulate the metal within their tissue s . These species mus t have developed a form of physiological tolerance to the metal ab sorbed . Metals which have been found in high concentrations in pl ant ti ssues but which are normall y toxic to plants , at these level s , include aluminium ( Chenery , 1 948 , Moomaw � al. , 1 959 , Jones , 1 96 1 , 2 Chenery & Sporne , 1 976 ) , chromium (Lyon e t al . , 1 968 , Wil d , 1 974 , Jaffre , 1 980 ) , lead ( Nicolls et al . , 1 965 , Lag et al . , 1 969 , Shimwell & Lauri e , 1 972 , Johnston & Proctor , 1 977 , Simon , 1 978 , B arry & Clark , 1 978 , Crook s , 1 979 ) , manganese ( Denaeyer-de-Smet , 1 966 , Jaffre , 1 977 , 1 97 9 , 1 980 ) , zinc ( Nicolls et al . , 1 965 , E rnst , 1 96 5 , Shimwell & Laurie , 1 972 , Johnston & Procto r , 1 977 , B arry & Cl ark , 1 978 , Simo n , 1 978 , Crook s , 1 979 ) , cobal t , copper and nickel ( see detailed di scussion b elow ) . The discovery of thi s metal accumulation by plants has led to b iogeochemical and phytochemical investigations of thi s property . In general these studies have investi gated plant-soil rel ationships , the tol erance of the species to the metal s , the distribution and nature ( ioni c , soluble complex , bound complex , etc . ) of the metal in the pl ant , inter-elemental rel ationships within the plant and the use of these species for biogeochemical prospecting . Biogeochemical prospecting was developed by Tk al ich ( 1 938) and Brundin ( 1 939) and involves analysing pl ant materi al far the mineral el ements to get an indication of the el emental content of the sub strata in which it grew . Reviews of this method can b e found in Cannon ( 1 960 ) , M alyuga ( 1 964 ) and Brook s ( 1 972 ) . General reviews of heavy metals in plants and of metallophytes in general can b e found in Boll ard and Butler ( 1 966 ) , Antonovic s et al . , ( 1 97 1 ) , �rector and Woodell ( 1 975 ) , and Fay � �. , ( 1 978) . ( a ) Nickel The toxici ty of nickel to plants has been known since the work of Haselhoff ( 1893) . The main symptoms of nickel poisoning are chlorosis o r yellowing of the leaves followed by necrosi s . O ther symptoms recorded include stunting of g rowth , unusual spottings , g rowth deformi ties and ei ther stunti ng or expansion of the root system (Mishra & Kar, 1 974 ) . I n extreme cases death of the pl ant may follow. 3 Despi te thi s known toxici ty of nickel , the cause of toxici ty of serpentine soils is far from known . As well as nick el , cobalt and chromium are present i n high levels and both are al so toxic to plants in high concentration . The cause of toxicity of these metal s i s believed to b e interference with , a nd poisoning o f , enzymes ( Bowen , 1 966 ) . The calci um/magnesium ratio has been cited as a possible cause of toxi city because of the low calcium , high magnesium content (Loew & May , 1 90 1 ) . However low l evel s o f magnesium have been recorded for tropical serpentine soil s ( Bi rrell & Wri gh t , 1 945 ) s o that the ratio hypothesis h a s been modified ( Vlamis & Jenny , 1 948) to one of l ow soil saturation by calcium . The low soil saturation makes al ternative cati ons more readily available for uptake . Low l evel s o f major nutrients , ni trogen , phosphorus and potassium, have also been observed in serpentine soi l s . Agri cul tural practices of ferti l i z ing serpent ine soi ls w ith these el ements have however failed t o make t he so il fertile . A compari son of these latter two possib i l i ties as cause s of the infert il i ty can be found i n Raven et al . , ( 1 976) . O ther possibl e toxic i ty causes have also been recorded : Go rdon and Lipman (1926) suggested soil alkal inity but although thei r Cal i fornian serpentine soils are alkal i ne , most are not and this appears unl ikely to be a general cause; Walker ( 1 948) suggested low molybdenum l evel s but l ater information suggested that it was unl ikely to be a dominant cause (Walker , 1 954 ) . Physical characteristics o f these soil s do not appear to b e a cause of an unfavourable growth environment (Robinson et al . , 1 935 ) . Given the compl exi ty and variety of serpentine envi ronment s , it is hi ghly unl ikely that any single cause for the toxici ty will be found . Biogeochemical p rospecting for nickel has b een studied by Lyon e t al . , ( 1 968) , Timperley et al . , ( 1 97Da , 1 972a , b) , S everne ( 1 972 ) , Severne and Brooks ( 1 972 ) , Cole ( 1 973 ) , Lee e t al . , ( 1 977a) , Brooks and Wi ther ( 1 977 ) and·Wi ther and Brooks ( 1 977) . 4 I n 1 977 , Brooks , Lee et al . proposed the term hyperaccumulator for those plants which contain ni ckel concentrations greater than 1 , 000 �g N i /g of dry leaf materi al . Pl ants having concentrations of 1 00- 1 , 000 pg/g were termed strong accumulators. I t should be noted that these terms are a statement of a concentration level rel ative to ••normal " ( non-enriched) levels i n pl ant materi al and bear no relationship to nickel concentra­ tions in the sub st rata. "Normal i ty" is however rather difficult to defi ne : on " normal " ( l ow-ni ckel) soils the concentration of ni ckel i n the plant (on a dry wei ght b asi s) seldom exceeds 5 �g/g (mean approx . 0 . 5 �g/g) but on nickel-enri ched ( generi cally termed " serpentine") soi l s eg . of serpentinitic and peri dotit ic parent materi al , " no rmal " pl ant concentrations are 25-50 �g/g and may easily reach 1 00 �g/g . To add further to the confusion many "no rmal11 ( non- tolerant) pl ants will not grow on ni ckel-enri ched substrates. Indeed i t i s common to be able to recogni ze these enriched areas by the sharp discontinuity between the vegetation on and off them. Hyperaccumulation of ni ckel was first reported i n the Crucifer Alyssum bertoloni i Desv . i n 1 948 (Minguzzi & Vergnano) during a survey of the vegetation of serpentine outcrops at Impruneta , near Fl orence , I tal y . The specimen collected had a ni ckel concentration of 12,000 �g/g ( 1 . 2%) i n its leaves. The second recording of a hyperaccumul ator was made for another membsr of this ge�us; a. murals Waldst. & Kit . ( Doksopul o , 1 96 1 ) . Al though he reported the nickel concentration on an ash weight basis ( o ver 1 0%) the species is clearly a hyperaccumulator. In 1 969 a thi rd taxon within this genus was discovered : A . serpyllifol ium Desf . ssp. lusitanicum T . R . Dudley & P. Silva (Menezes de Sequeira , 1 969 ) . Thi s taxon , confined to serpentine soils around Braganca , nort�eastern Portugal , i s a subspecies of a widespread southwestern European and North African species . The first non-Alyssum to be discovered as a hyperaccumulator was Dicoma niccol i fera Wild (Wi l d , 1 97 1 , earl ier as Dicoma macrocephala ssp . , Wild , 1 970 ) . The highest ni ckel concentration recorded for this Zimbabwean species was 2 , 1 00 pg/g , dry weight . 5 S ince the mid-seventies , the l i st of hyperaccumulators has g rown rapidly through the work of R . R . B rooks and his ea-workers. Their first discovery o f a nickel hyperaccumulator was Hybanthus floribundus (Lindl . ) F . Muell . (Severne & Brook s , 1 972) . Although this was the first pub l i shed reference to this taxon , Cole ( 1 973 ) had made the di scovery earlier but had delayed publ i shin�. The nickel content of this species often exceeds 1% . All three subspecies of this species ( described by Bennett , 1 969) are hyperaccumulatars (Severne , 1 972) . In 1 974 the number of hyperaccumulators more than doubled. In Z imbabwe ( then Rhodesia) , Wild ( 1 974 ) reported very high nickel l evel s i n the ash of Pearsonia metallifera Wild . Lee ( 1 977) g ives a dry leaf concentration o f 1 . 06% far thi s species. The vast serpentine soil compl exes of New Caledonia suppl ied the other species : Gei sso i s pruinosa Brangn. & Gri s . , Homal ium guillainii (Vieill . ) B riq. , Hybanthus austrocal edonicus Sch in z . & Guill ., �· caledonicus Tu rcz . , Psychot ria douarrei (G. Beauv. ) Daniker (all Jaffre & Schmid , 1 974 ) and Homali um kanaliense (Vieill . ) Briq. (Brooks , Lee. & Jaffre , 1 974 ) . Following up thi s work by surveying the genera which contain known hyperaccumulators was to prove fruitful . B rook s , Lee e t al . ( 1 9??).in a survey o f the genera Homal ium and Hybanthus not only re-i denti fied the previously known hyper­ accumulatars but added a further five to the l i st . All fi ve are of the Homalium genus ( see Appendix I I (a) for species ) . Jaffr� , Brook s and Traw ( 1 979) surveyed the Geissois genus . Of the seventeen species sampled , seven were t o p rove hyperaccumulators . One , g • . pruinosa, had al ready been reported but the remaining six were all new . Earl ier (Jaffr� , Brooks e t al . , 1 9?6) the New Caledonian studies had resulted i n the di scovery of the hyperaccumulator Sebertia acuminata Pierre ex Baill . Thi s plant has an extremely 6 unusual l atex containing up to 25% nickel on a dry weight basis ( 1 1 . 2% wet weight ) . The tree i s known to the locals as seve-bleue (blue-sap ) from the colour of this l atex . S ince Homal ium and Hybanthus are in the closely related families of Fl acourtiaceae and Viol aceae respectively , Brooks and W ither ( 1 977 ) surveyed these families in South East Asia ( see al so W ither , 1 977 ) . Thi s resul ted in the identification of Rinorea bengalensi s (Wall . ) O . K . (Violaceae ) as a hyperaccumulator . Following on from thi s , t h e genus Rinorea was surveyed ( Brook s , Wither & Zepernick , 1 977 ) . Seventy o f the approx . 250 recognized species were sampled and a further hyperaccumul ator of nickel was found : �· javanica (Bl . ) O .K . The Rinorea hy�eraccumul ators are unusual in one respect; both �· bengalensis a nd �· javani ca also accumulate cobalt ( 545 �g/g and 670 �g/g respectively maximum values ) . Al though serpentine soils are enriched in cobalt no previous nick el hyperaccumulators had simultaneously accumul ated cobalt to thi s degree . Also in 1 977 , Wither and Brook s sampl ed vegetation collected at Jikodolong , Obi Island , Indonesia , a known ultrab asi c area , to try and identify fu rther hyperaccumul ators . Three species Myristica laurifolia Spruce ex DC var. bifurcata, Planchonella oxyedra Dubard and Trichospermum kjellbergii Burret were found . I· kjellbergi i al so had an elevated cob al t concentration (350 �g/g) . Anal ysis of further sampl e� of Planchonell a and Trichospermum revealed no further hyperaccumul ators i n these genera . However one specimen of�· oxyedra had an elevated cobalt content ( 240 �g/g) , thus joining the other Indonesian hyperaccumulators i n having both abno rmally-high nickel and elevated cobalt concentrations. Continuing the Flacourtiaceae l ine of development , a survey was made of other members of thi s fami ly in New Caledoni a (Jaffre , Kersten � al . , 1 979 ) . Thi s survey resul ted in the re-identificati on of the seven Homal ium hyperaccumulating species previously known and the di scovery of twelve more i n the genera Casearia ( 1 ) , L asiochl amys ( 1) and Xylosma ( 1 0 ) . 7 All fifty-three species of Fl acourti aceae listed by Sleumer ( 1974 ) were anal ysed. The New Caledoni an surveys continued with the·gen�s Phyll anthus following the results of spot tests (using dimethylglyoxime) done on specimens at the Noumea herbarium. The spot tests showed the possibi l i ty of there being several hyperaccumul ators in this genus (Kersten , 1979 ) . The results of the survey ( Kersten et al . , 1979) revealed ten further hyperaccumulators three of which also had elevated cobalt 8 level s similar to those of the Indonesi an nickel hyperaccumulators . A fu rther seven species of hyperaccumul ators from New Caledoni a have been reported by J affre ( 1980 ) . These brought to fi fty-eight the number of nickel hyperaccumulators reported from this island . When one considers that the first three hyperaccumul ators were all of the genus Alyssum it i s surprising that no systematic studies in this genus were reported until 1978 when Brooks and Radford (1978a) su rveyed all species listed in Fl ora Europaea by Ball and Dudley ( 1 964) . In addi tion to the two species and one subspecies previously known , eleven new hyperaccumulators were found. I t was noted that all the hyperaccumulators in Alyssum bel onged to section Odontarrhena ( C . A . Meyer ) w. Koch . The five other sections were devoid of hyperaccumulators even for individual s growing on ultrab asic rock s . O ther hyperaccumul ators within the Cruciferae (Alyssum is a member of this famil y ) have since been discovered. Vergnano Gambi and Gabbrielli ( 1979) in a survey of vegetation on Italian ophiolitic outcrops di scovered two Cruciferous hyperaccumulators in the Valle d1Aosta : Cardamine resedifolia L . and Thlaspi rotundifolium (L . ) Gaud. A survey of genera in Tribus Alysseae has reveal�d that only one further genus contai ns hyperaccumul ators (Reeves , Brooks & Dudley , 1981 ) . This i s the genus Bornmuellera with hyperaccumulation in three species and the hybrid �· x petri Greuter , Charpin & Di ttrich . In o ther Cruciferous genera , nickel hyperaccumul ation has been reported for Peltaria emarginata (Boi ss . ) Hausskn . (Reeves , B rooks & Press , 1980 ) , Streptanthus polygaloides Gray ( Reeves , B rooks & MacFarlane , 1 981 ) ( the fi rst American nickel hyperaccumul ator di scovered) and various Thl aspi speci e s . Owing to a certain amount of confusion i n the taxonomy of Thl aspi and rel ated genera in Europe , no firm number has been assigned to the l i st of hyperaccumul ators but i t appears that approx . thi rty taxa have thi s character. In addi tion Thl aspi montanum L . ( three varieties ) i n North America and I• japonicum Boiss . in Japan are hyperaccumulators (Reeves & Brooks , unpubli shed data ) . Among these many species which hyperaccumul ate nickel , the majority h ave maximum concentrations of less than 1% . N ineteen taxa have maximum concentrations from 1 -2% : Alyssum argenteum All . ( 1 . 08%) , �· bertolonii ( 1 . 22%) , �· heldreichii Hausskn . ( 1 . 25%) , �· markgrafi i O .E . Schul z ( 1 . 37%) , �· roberti anum Be rnard ex Gren. & Godron ( 1 . 25%) , Bornmuellera b aldaccii (Degen ) Heywood ssp . rech ingeri Greuter ( 1 . 20%) , B . glabrescens (Boi ss� & Bal . ) Cull en & Dudl ey ( 1 . 92%) , B . x petri ( 1 . 1 4%) , Gei sso is pruinosa (1 . 36% ) , Homal ium franci i Guill . ( 1 . 4 5%) Hybanthus austrocaledonicus (1.85%), �· floribundus ( 1 . 42%) , Lasiochlamys pel tata Sleumer ( 1 . 1 0%) , Pl anchonella oxyedra (1.96%), Rinorea bengal ensis ( 1 . 7 5%) , Seberti a acuminata ( 1 . 1 7%) , Streptanthus polygaloides ( 1 . 48%) and Thlaspi montanum vars . montanum ( 1 . 7 1 %) and cal ifprnicum (Watson ) P . Holmgren ( 1 . 1 6%) . Only nine taxa have exceeded a concentration of 2%. These are Barnmuellera b aldaccii ssp . b aldaccii ( 2 . 13%) , B . b aldaccii ssp . markgrafii ( 2 . 73%) , g. tymphaea ( Hausskn . ) Hausskn . (3 . 1 2% ) , Geisso i s intermedia Vieil l . ex Pampan ( 2 . 29%) , Hamali um guillainii ( 2 .90%) , Peltaria emarginata (3 . 44%) , Phyll anthus serpentinu s Maare (3 .8 1%) , tap of the table Psychotria douarrei ( 4 . 70%) and Thlaspi mantanum var . sisk iyauense P. Holmgren ( 2 . 46%) . Many b iogeachemical and phytochemical studies of nickel hyperaccumulators have been reported . These i nclude pl ant-sail relationships , i ntraplant di stribution of nickel 9 and complexes of nickel within the pl ant. Brooks ( 1980 ) has revi ewed these studi es . The most striking plant-soil elemental rel ationships for ni�kel hyperaccumul ators are their abil i ty to accumulate nickel rel ative to the soil s , the ab i l ity to a ccumul ate calcium and potassium to adequate ( i f low) physiologi c levels from nutrient-deficient soils , and the abil ity to restrict magnesium uptake to an " acceptable" l evel . Studies by Lee et �· ( 1977a ) showed leaf-nickel l evel s in Homal ium 1 0 k analiense to be strongly rel ated to only manganese and extractable-nickel soil level s and in Hybanthus austrocaledonicus to b e strongly rel ated to both total and extractable-nickel soil l evel s . They used these resul ts to suggest that hyperaccumul ation may therefore b e controlled by plant organic rather than soi l inorgani c constituents . Calcium uptake appeared unaffected by other soil element concentrations. Po tassium uptake vari ed : H . austrocaledonicus on a richer soil had no signi fi cant potassium-soil element rela tionships whereas �· kanal iense on a poorer soil had several antagoni sts to potassium uptake including both calcium and magnesium . The di stribution and nature of nickel in hyper­ accumul ators has been more widely studied . In Alyssum bertolonii nickel has been shown to b e in the epidermi s and scl erenchyma�c ti ssues of stems ( Vergnano Gambi , 1967 ) . Thi s was done by staining the ti ssues with dimethylglyoxime . Pelosi et al . ( 1974 ) studying the nickel compl ex of this species deduced that it was an organic acid complex . These workers later concluded that mali c and malonic acids along with a third unidenti fied acid were i nvolved ( Pelosi et al . , 1976 ) . Pancaro et al . ( 1977 ) confi rmed the i nvol vement of these two bi carboxyl ic acids i n �· bertoloni i . They extended their study to i nclude �· serpyllifolium ssp. lusitani cum and concluded that mal i c acid was al so involved here though malonic acid was not. This latter study was essenti all y confi rmed by Lee � al . ( 1 978 ) . Farago � al . ( 1 975 ) showed that ni ck el i n Hybanthus floribundus was located in large epidermal cel l s by using two stains ( dithioxami de and dimethylgl yoxime) . They also stained cells to detect pectin (using ruthenium red ) . Thi s was also found i n high concentration i n the large epidermal cells . Kelly et al . , ( 1 975 ) found no evi dence of nickel accumulation i n specific ti ssues o f some N ew Caledonian Hybanthus species . This work was done by di fferenti al centri fugation and supported by electron microprobe studies which showed a uni form nickel di stribution within the leaves . Sequenti al extraction series, (Farago e t al . , 1 97 5 , Lee , 1 977 ) have shown that water and dilute acid will extract over 80% of nickel from l eaves of hyperaccumul ators . Thi s shows that nickel in hyperaccumulators tends to b e as a h ighly soluble polar complex or i s easily exchangeable . ·severne ( 1 972 ) reported a low molecul ar weight , water-soluble complex from Hybanthus floribundus but could not compl ete the i dentification . Further work (Kelly et al . , 1 975 ) found this ni ckel as both aquo-Ni 2+ and as a�o:-molecul ar weight complex . No amino acid rel ationships could be detected . Farago et al . , ( 1 975 ) showed , by paper chromatography , that n ickel pectinate was a possib il ity . They also i nferred some nickel association with a spart ic aci d and cysteine . Kersten ( 1 979 ) separated and i dentified a ci tratonickelate ( I I ) compl ex from this specie s . Work ing on the New Caledonian hyperaccumul ators Kelly � al . ( 1 975 ) , using Hybanthus austrocaledonicus , tl• c aledonicus and Psychotri a douarrei , s howed no association b etween nick el and amino acids . N i ckel was also found to be 2� present as both aquo-Ni and a compl ex of a molar mass of approx . 200 . Further work (Lee , 1 977 , Lee e t al . , 1 977b , Lee e t al . , 1 978) has shown the presence of a negatively charged ci tratonickelate ( I I ) complex i n eleven N ew Cal edonian hyperaccumulating and strongly accumulating species (Sebertia acuminata , Psychotri a douarrei , Geissois pruinosa , Hybanthus caledonicus , H . austrocaledonicus and six Homalium species ) _ . 1 1 Seventeen species (the el even above plus six fu rther Homal ium speci es ) showed a l inear relationship fo r nickel-citri c aci d concentrations . Three hyperaccumulators from outside New Cal edonia ( Alyssum bertolonii , �· serpyllifoliuT ssp. lusitanicum and Pearsonia metall ifera) d id not fit thi s relationship. The two Alyssum species have complexes with mal i c and/or malonic aci ds while Pearsoni a metall i fera has a complex with an uni denti fied trihydroxycarboxylic acid of formul a c5H 1 0o5 (Lee , 1 977 , Lee et al . , 1 978) . While Psychotria douarrei had a similar nickel-citric acid relati onship to other New Cal edonian speci es , Kelly et al . ( 1 975) had shown that the bulk of the extracted nickel i n 2+ thi s species appeared to be aquo-Ni ( 94% o f total n ickel ) . Lee ( 1 977 ) has shown the presence o f a c itratonickelate ( I I ) compl ex but di d not i nvesti gate the ve ry large non-citrato­ nickel ate ( I I ) portion . Kersten (1979) further investi gated thi s speci es and showed that most of the supposedly aquo-Ni 2+ i s in fact a malatonickelate ( I I ) complex. �ersten (1979) also investi gated the ni ckel complexes of seven other hyperaccumulators . I ncluded in these wera Homalium k anal iense , which had been sho�n by Lee et �· (1977b) to have a c itratonick el ate ( I I ) compl ex , and Hybanthus floribundus ( see prior di scussion ) . He confirmed the presence of citratonickelate ( I I ) in the Homalium specie s . Four further species ( Caseari� silvanae ( J . R . & G. Forster) Sl eume r , Lasiochl amys peltata Sleumer , Xylosma vincentii Guill . from New Cal edoni a and Ri norea bengalensi s from N ew Guinea) had citratonickelate ( I I ) compl exe s . Phyll anthus serpentinus however had both citrato- and mal ato-nickelate ( I I ) compl exes . 1 2 It i s worth noting that the two mixed-complex species (� . douarrei and f• serpentinus ) h ad much higher nick el levels than the other species . Had they s aturated one system before starting another? Studies of i nter-el emental rel ati onships invol ving nickel in leaves have shown few si gnificant relationships . Lee ( 1 977 ) studi ed three New Cal edonian species; Hybanthus austrocaledonicus had a positive ni ckel-phosphorus relati onsh ip , Homalium guillaini i a positive nickel-magnesium relationship and Homal ium k analiense had positive nickel­ copper and nick el-z inc relationshi ps . Lee relates thi s low number of rel ationships to the probabil ity that a di fferent uptake mechani sm i s operating for nickel compared to the other elements . Work ing on Hybanthus flo ribundu s , Farago � al . , ( 1 975 ) associated a high nick el content with a h igh calcium content. They al so suggested that a relationship between nickel and cal cium - magnesium existed . A nickel - cal cium - magnesium rel ationship has al so been suggested for Alyssum bertolon i i ( Vergnano Gambi � al . , 1 977 ) . Kersten ( 1 979 ) study ing Rino re a bengalensi s showed positi ve nickel­ cob alt, nick el-copper and nick el-sodium relati onships . �· bengalensis is al so a very strong accumul ator of cob alt. As the nick el-cobalt rel ationship in serpenti ne soils is also generally strong and positive it appears that �· bengalensi s refl ects thi s soil rel ationsh i p . Thi s i s 2t variance with most other studied hyperaccumul ators . �· bengalensis appe ars more sensiti ve to the envi ronment than the others which , conversely , appear to exercise a greater degree o f control over whi ch elements are accumulated . Thi s sensitivity does however mak e �· bengalensis a b etter biogeochemi cal prospecting species ( c� Brooks & W ither , 1 977 , Kersten , 1 979 ) . ( b ) Cobalt and Copper 1 3 �No rmal " pl ant levels for cobalt and copper are < 1 �g/g and 3- 1 5 �g/g (mean 8 �g/g ) respecti vely on a dry weight basis . These concentrations occur for soils with concentrations of 1 -8 �g Co/g and 25-75 �g Cu/g . Po i soning of pl ants by cobalt and copper results in chloros i s and possibly necros is . Various spottings may al so occur , as can stunting and reduced root growth . Hunter and Vergnano ( 1 953 ) rated these two metal s as less toxi c than nickel with copper more toxic than cobalt. On mineraliz ed soi ls it becomes more di fficult to speak o f a "normal" concentration but most pl ants contain 20- 1 00 �g Co/g and 20-70 �g Cu/g . It is o f i nterest to note the much greater i ncrease in cob alt relati ve to normal (40- fol d ) than that for copper ( 5-7 fold ) . Thi s i s to be expected 1 4 as copper being an essenti al el ement tends to b e more tightl y controlled i n uptake than the non-essential cobalt ( Timperley et al . , 1 970b ) . The content of cobalt and copper in mineralized soils can b e very high . Duvigneaud and Denaeyer-de-Smet ( 1 963) give values for some samples from metalliferous soils in Shaba : cobalt, low 600 - 900 �g/g , h igh> 40 , 000 pg/g; copper , low 900 - 2 , 000 pg/g , high ) 90 , 000 �g/g . Few plants will tolerate such metal l evel s . Thu s , as fo r nickel , cobalt and copper mineral i z ation can b e detected b y the changes in flcra which accompany them. Indeed a transect across mi neral ized h ills shows regular changes in flora with different metal concentrations ( see Duvi gneau d , 1 9 58 pp . 263 & 265, Malaisse & Grego ire , 1 978 p . 254 ) . No concentration criteria have yet been determined for hyperaccumulati on of cobalt or copper in pl ants . In declaring 1 , 000 �g/g as the hyperaccumulation l evel fo r nickel Brook s , Lee et �· ( 1 977 ) had tak en a value a n order of m·agni tu de hi gher than the ni ckel levels found in "no rmal" pl ants g rowing in nick el-rich substrates . No such easily determi ned level has been found for cobalt or copper . Work cited in thi s section and the results within thi s thesi s h ave been analysed to determine a level for hyperaccumul ation of these metal s . A cumulative frequency plot against concentration i n leaf material shows a di scontinuity slightly b elow 1 , 000 �g/g for both metals ( see fig . 1 . 1 ) . Thi s has been rounded to 1 , 000 pg/g for simpl icity and set as the level above which hyperaccumulati on occurs . In studies of the "copperbelt" i n Shaba ( ex-Katanga) Province , Zaire ( ex-Belgi an Congo ) , Duvi gneaud ( 1 959) had shown very strong accumulati on ( i e . greater than 500 �g/g ) of cobalt i n Crotal ari a cobalticola Duvi gn. & Pl ancke 1 0,000,..---------------� - � 0 E � s.... ""'0 en ' en :::l... - c 0 ....... a s.... ....... c � u c 8 1 000 1 00 1 5 o Cobalt • Copper 10 2 0 LD 60 80 90 95 98 99 99·5 t Cumulat ive 0/o Figun2 1 · 1 C umula t ive frequ ency p lot for coba l t and copper i n p lan ts from mineral iz ed areas of Shaba Prov in ce, Zai"re . ( 530 �g/g ) whi l e Duvi gneaud and Denaeyer-de-Smet ( 1 963) showed both hyperaccumulation and ve ry strong accumul ati on of copper ( Ascolepi s metal lorum Duvi gn . & Leonard , 1 , 200 �g/g , Silene cobalticol a Duvign. & Pl ancke , 1 , 660 �g/g , Haumani astrum roberti i ( Robyns ) Duvign. & Pl anck e , 1 , 960 �g/g and Pandi ak a metallorum Duvi gn. & van Back . , 740 pg/g ) . In 1 960 , Kubota et al . found very strong accumulati on of cob alt in Nyssa sylvati ca Marsh. var. b i flora (Walt .) Sarg . At 845 pg/g thi s was the highest cob alt leaf concentration then known . In 1 966 , Dyk eman and De Sousa reported high copper level s i n pl ants growing over the Tantramar copper swamp i n Canada . The copper swamp had a concentration of 7% copper . The acidic conditions made copper readi ly available but chelation by organi c material may have lowered the availab ility and made conditi ons less toxic. The hi ghest pl ant levels were 1 , 120 pg/g in Abies bal samea (L.) Mill., 726 pg/g in Larix l ar icina (Du Roi) Koch and 500 �g/g in Ledum groenlandi cum Oedr . In 1972, Ernst repo rted a concentration of 890 �g Cu/g in Indigofer8 dyeri Oritt . from the Copper King anomaly in northern Zinbab�e. In 1977 , Malai sse , Brooks and their coworkers started systemati c studies of two types : f iel d su rveys over specific "copper clearings" i n Shab a and genera surveys for those genera i n whi ch field surveys show hyperaccumulation. Starting from the genera cited by Duvigneaud ( 1 959) and Duvi gneaud and Denaeyer-de-Smet ( 1 963 ) , Brook s ( 1 977 ) surveyed the genus Haumani astrum and Brook s , McCl eave and Mal aisse ( 1 977 ) the African members of the genus Crotal ari a . The results with respect to copper showed no further hyperaccumul ator in e ithe r genus although £• peschiana Duvi gn . & Temp . at 705 �g/g is a very strong accumul ator . The s�veys d id , howeve r , reidenti fy �· roberti i as a hyperaccumul ator . The cobalt results i n Crotal ari a failed to show any very strong accumul ation . No£. cobalti cola specimens were anal ysed. The survey o f Haumaniastrum however l e d to the fi rst i denti fi cati on o f cobalt 1 6 hyperaccumulation with a concentration of 1 0 , 220 �g/g ( 1 . 02%) in H . roberti i . Thi s exceeded by over one order of magnitude , the previously recorded h i ghest cobalt levels in dry leaf materi al . Also in 1 977 , Brook s , McCleave and Schofi eld surveyed the Nyssaceae for cobalt . As well as re-i dentifying very strong accumul ation by Nyssa sylvati ca var. b i flora they showed very strong accumul ation by Nyssa sylvatica var . syl vati ca ( 530 �g/g) . I n 1 975 , Agrosti s stoloni fera L . was found to hyperaccumulate copper on contaminated soi l s near a metal refinery at Prescot , U . K . (Wu et al . , 1 975 ) whi l e in 1 977 , Agrosti s gigantea Roth . on mine-waste sites near Sudbury , Canada , showed very strong accumulation o f thi s element (Hogan et al . , 1 977 ) . The next major report of hyperaccumulation came from Shaba. Malaisse and Gregoi re ( 1 978) surveyed the vegetation at the Mine de l'Etoile near Lubumbashi . From thi s study , faur copper and three cob alt hyperaccumul ators were discovered (see Appendix II ( b) for speci es) . Al so di scovered were Four copper and five cob alt very strong accumul ators . Many serpentine flora surveys have tested for cobalt as well as n i ckel since serpentine soils tend to b e enriched in thi s element �!so . Despite thi s , few "serpentine" plants have elevated cobalt levels and onl y three have ever reached very strong accumul ation l evel s : Rinorea bengalensi s and �· javani ca from I ndonesi a ( Brook s , W ither & Zepernick , 1 977) and Phyll anthus ngoyensis Schlect . from New Cal edoni a (Kersten �al . , 1 979 ) . Work done on copper-lead-z i nc enriched areas of Europe have measured copper level s in the vegetation but no copper levels reported have reached those o f very strong accumul ation . As for ni ckel , cobalt and copper accumul ating pl ants 1 7 can be used for b i ogeochemical prospecting (Warren � al . , 1 94 9 , Warren & Del avault, 1 950 , Ri ddell , 1 952 , Brook s , W ither & Westra , 1 978 ) . I n the i r di scovery o f a copper b i ogeochemi cal anomaly on S a l ajar I s l and , Indonesi a , B rook s , W i ther and Westra ( 1978 ) found ve ry s trong accumu l ation in three pl ant speci e s . Hype raccumul ation o f cob a l t and copper appears to b e much rarer than hype raccumul ation o f ni ckel . L evel s reached b y those p l ants whi ch do hyperaccumul ate cob a l t and/or copper tend to be lower than tho se reached by " nickel plant s " . Thus only H aumani astrum robertii (1 . 02%) h as surpassed the 1% Co level and Aeol anthus b i formi fol ius ( 1 . 37%) the 1% Cu level . I t shou l d b e noted that all the hype raccumul ators o f cob a l t and a l l except o ne of copper come from the S h ab an " copperb e l t� 1 8 Very l i t tl e information on the b i ogeochemi s try o r phytochemi stry o f accumulators or hyperaccumul ators of copper i s avai l ab l e . However some i nformation does exi s t for the very s t rong accumu l a tor I ndigofera dyeri ( E rnst , 1972 ) and the s trong accumu l a to r (m ax . 324 �g/g , Re i l l y, 1967 ) Becium homblei ( De Wild.) Duvign . & Planck e . I n wo rk on l• dyeri and l• seti flora Bak . (a tol e rant but non-accumulating member of thi s g enu s ) E rnst showed both th a t popu l a t i ons from mineralized �re as h a d greater tol erance than those from non-mineral i z e d a re a s a n d that thi s tol erance was speci fi c to the minerals present in the o rig inal so i l s rather than a general he avy metal tol erance . Tol erance was tested by the method of compara tive protopl a smat­ ology ( Repp , 1963 , ·Gri es , 1966 ) . A sequenti al extraction seri e s done on !· dyeri showed that almost hal f the coppe r was extracted b y water suggesting the location o f thi s copper wi thin the cell vacuo l e . Very l i ttl e copper was extracted b y a n o rganic sol vent ( bu tanol ) . Most copper not extracted b y w a ter coul d b e extracted b y exchange processes u s ing sodium chloride , ci t ri c ac id and d ilute hydrochlori c aci d . L es s than 2% remained i n the residue . The exchangeab l e copper was beli eved to have been bound to the cell wal l . N o at temp t was made to i dent i fy the nature o f the copper in any fraction . T he " copper fl ower" , Becium homble i h as al so been stud i ed ( Reil l y , 1 96 9 , R e i l l y e t al . , 1970 , Howard-Wi l l i ams , 196 9 , 1970 ) . Re i l ly and eo-wo rk ers have suggested that copper i n this speci e s is complexed to amino aci d s , parti cularl y cysteine . Evi dence ( Re i l l y , 1 969 ) that 50% of the copper i s ex tractabl e i n to o rganic sol vents ( dioxan , butanol , methano l ) i s , they sugge s t , indi cat ive of an org an ic compl ex of coppe r . A fu rther 20% o f the copper i s solub l e i n water and d ilute aci d ( ionic and exchangeab l e copper ) and 30% i s insolub le ( tightly bound copper ) . R e i l ly e t �· ( 1 970 ) found 1 7% o f the copper a ssoci ated with the cell wall a n d suggested that the copper was bound to the polysacchari des , l i gnins or a ssoci ated proteins o f the wal l . By d ialysis against water they deduced that 20-25% o f the copper was ei ther i onic or i n a water-solub l e compl e x . D i al y si s agai nst a tartaric a c i d solution showed a further 1 0- 1 5% o f t h e copper was o n ly l i ghtly complexed ( stab i l i ty l ower than tartari c aci d-copper complex ) . By paper chromatog raphy on water-ex tracted copper they showed tha t no i no rganic Cu 2+ was presen t . Howard-W i l l i �ms ( 1 969) tested the tol erance of B . homb l e i b y the roo ting technique and showed that speci e s g row i n� an coppe r-rich so i l s had a greater copper res i s tanc� 1 .h � n tho s e grow ing on nick e l- ri ch o r normal so i l s . He al so showed i n po t t ri a l s that B . homb l e i can survi ve o n so i l s w i th onl y a trace of copper . Howard­ Wi l l i ams ( 1 970) gives the fi el d tole rance range a s trace to 1 0 , 000 �g Cu/g dry so il . Testing o f seed germi nation showed that seeds readily germinated i n d i s t i l l e d wate r ; a finding which contradicted that o f Horscroft ( 1 9 6 1 ) who found that seeds required solutions o f 50-600 �g Cu/cm 3 for germination . S ome s tudies on the Canadian accumul ators h ave b een made . Dykeman and De Sousa ( 1 966 ) concl uded that the copper content o f the pl ants was not rel ated to the total copper content of the soil . They based thi s conclusion on the fact that vascu l ar p l ants grew on . sub strates w i th a h i gher total copper than su rround i ng sub strates but whose seedlings died in these surrounding sub strates . Hogan e t al . ( 1 977 ) compared el emental con tents of tolerant and non-tol erant clones o f Agros t i s gigante a . The tol erant clones h a d h i gher copper , 1 9 i ron , nick el and z inc content but manganese , potassium , cal cium and magnesium showed no di fferences . Despite thi s , rooti ng tolerance tests showed greater tolerance for only two of five clones growing on the copper-rich sub strate compared to the clones g rowing on surrounding soi l s . It was shown that these two tolerant clones came from areas with no higher metal content than the other three clones , but that the i r s ites had a signifi cantly lower pH l evel . At l ower pH level s , the metals tend to be more soluble and hence pl ant-available so that thi s may have controlled the development of tolerance . No detailed studies on accumulators o r hyperaccumulators of cobalt have yet been made . The work discussed i n the succeeding chapters covers i nvesti gations i nto several aspects of hyperaccumul ation . Fo r cobalt and copper hyperaccumul ation , f ield and genera surveys are reported . These l ead i nto b iog�ochemi cal and phytochemi cal studies on some of these pl ant species . Biogeochemical and phytochemi cal data a re a lso presented on species of the Alyssum genus which containa ,B l arge numb er of hyperaccumul atars of nickel . Thi s wo rk attempts to increase our unde rstanding of the hyperaccumulation of these metal s in pl ants . 'l' " 2 0 PA R T ONE C o b al t a n d Co ppe r in Veg etat ion of S h a ba P rovince, z a·t"r e I CHAPTER 2 S u r ve y s of Vege t a t io n of M et a l l i fero u s Soi ls 2 . 1 I NTRODUCTION The " co pperbel t " of Shab a Province , Z al re and northern Zamb i a fo rms one o f the worl d ' s greatest metal l o genic p rovinces . I t contains some one hundred m ine ral i z ed areas w i th a total area o f 20km2 di spersed over 22 , 000km2 • The mineral i z e d areas often cover several square k i lometre s . The principal metal s of the mineral ize d areas a re copper , cob a l t and n ick el al though others , uranium , z i nc , l ea d , manganese , may a l so b e present ( Duvi gneaud , 1 9 58) . Such mineral i zed a reas are h i ghly toxic to " no rmal " pl ants but tol erant popul at ions h ave evolved o n them . Al though D e W i l deman ( 1 92 1 ) mentions thi s " copper flo ra " , i t was Robyns ( 1 9 32) who b eg an the detai l ed studi e s . He noted a herb aceous veget ation zone on rich copper so i l s surrounded by a t rans i t ional zone of stunted woody vegetation extending out to the " no rmal " open woodl and . Duvigneaud ( 1 958 , 1 959) and Duvigneaud and Denaeyer-de-Sme t ( 1 963) g ive a more detailed study o f this flo ra . They al so report concentra tions o f cob al t ( Duvi gneaud , 1 959) and copper ( Duvigneaud and Denaeyer-de-Smet , 1 963) i n pl ant materi al , g ivi ng recognition to the extremely h i g h l evel s some speci es of thi s tolerant flora cou l d accumul ate . 2 . 1 . 1 Physical Envi ronment ( a ) Geology , Bowen and Gunati l ak a ( 1976) have recentl y reviewed the geology of the " copper b e l t " . The geologi c al components of the region have b een classified as b elonging to e i ther a b asement compl ex o r the K atangan sequence . The b asement compl ex consists o f the Lufubu system schi s t s , gneisses a n d quart z i te s overl ai n , unconfo rmab l y , b y the Muva system sediment s , and o l d gran i t es from a post-Lufubu orogenic phas e . There is an absence o f l arge scal e mineral i zation i n th is complex . The Muva sedimentation was ended b y an i ntense erogeni c deformation which compressed the sediments i nto long north-east trend ing fol ds . 2 3 2 4 Subsequent movement now g i ves the fol d s a no rth-west t rend . The sediments o f the Lower Roan period are bel ieved to h ave come from thi s compl ex . The Katangan sequence was l a i d on t o t h i s b asemen t complex . At thi s time the compl ex had a rel ief o f some 30Dm . The sequence is o f marine sedimentary o ri g in from a series o f marine t ransgressions . The i n i t i al t ransgression l eft a l ayer o f sandstones and conglomerates o n the fl ank s o f the paleo ri dg e s . Above thi s , t h e Lower Roan c l a s t i c sediments form a l ayer o f 800-2000m depth , covering t he pal eori dges . T he Lower R o an deposi ts g radu al l y change to dolomi te-rich depos its o f the Upper Roan peri o d . These , in turn , g rade i nto carbonaceous shales o f t h e Mwashi a group . T he Mwash i a group i s topped b y an e rosional contact with the Kundel ungu group . Thi s group consists o f t i l l i t e and marine sedimentary ho rizons l a i d i n Upper Pre-Camb ri an time s . At least two majo r ti l l i te hori zons exist a s testimony to fluctuations in the paleocl imate o f Cent ral Afri ca . Fol l owing the Kundelungu sedimentation , the Kundel unguan o rogeny rap i dl y raised the are a . Thi s ri se w a s s trong er in the north . I n Shaba i t formed a l arge arc with a succession o f anti c l ines and syncl ines . The o re bodies o f the " copperbel t " are confi ned to the Lower Roan sediments , particul arl y some 150-200m of sediments near the m i ddl e o f the s eries . I n rel ation to t h e paleoridge s , t h e h igh­ grade ore is found on the slopes wi th l i t tl e on ei ther the ri dge tops or val l e y floors . The mineral s are be l ieved to h ave been deposited a s sul phides from a reducing , shal low marine envi ronmen t . This woul d however g ive only a low grade o r e . Fu rther enri chment by di agene s i s appears to have o ccu rred but the exact nature of the pro cess remains unknown . Also unknown are the o ri gins of the metal s depo sited al though some may h ave come from the b asement compl ex ( see B owen & Gunati l ak a , 1 976 ) . S i nce the final emergence from the sea , Central Afri ca has been subj ected to t hree penepl anations ( Duvigneau d , 1958 ) . The first peneplanati on , fini shing in the l ate Cretaceous , • was followed b y a general ri sing o f the African continen t . The second penepl anation fini shed in the mi d-Te rti ary . Today th is i s the main e ro sional surface o f Shab a . I t now h a s a general a l t itude o f appro x . 1 , 3DDm . The third peneplanation was more l ocal i z e d , i n southern Shab a , and today has an a l t i tude of 1 , 2DD-1 , 30Dm . ( b ) Topograph y . The l andscape o f Shab a today i s o f a series o f pl ateaux , o f approx . 1 , 300m a l t i tude , with few h i l l s o r vall e y s . T h e highest a l t itude i s appro x . 1 , 700m (Kundelungu Plateau ) whi l e the l owest i s l e ss than 700m i n the l arger river val leys ( e . g . Lufira V al l ey ) . E ro sion has uncovered many under­ l ying rocks w i th the h arder rock s , including the metal l i ferous conglomerate , o utcropping as h i l locks o r crest s . The l o c al i zation o f the outcrops is a result o f fracturing o f the metall i fe rous belt during the rotation o f the belt from a north-east to a south-east trend ( fi g 2 . 1 ) . Many of these h il locks are o r have been mined o ver the years and fur ther mining is pl anned in many areas ( fig 2 . 2 ) . The current l argest mines are at Kolwezi ( Shewry et al . , 1979 ) . ( c ) Climate . The Shab an cl imate i s characteri zed b y a hot rainy season and a coo l e r dry season . The r a iny season occu rs i n summer , l ast ing from October to Apri l . Rainfall genera l l y exceeds125mm/month from Novemb e r t o M arch with J anuary having a mean p reci pi tation o f 200-300mm . The J anuary mean temperatu re i s 22-24°C with a m�an d a i ly maximum o f 26-28°C . The dry season in winter is very dry wi th no measurab l e precip i tation between June and Augu st . The July mean rainfall i s l e s s than D . 1mm . The Ju ly mean temperature i s 15-17°C with a mean d a i l y minimum o f 7 . 5-10°C . Frosts are absent o r r are i n north Shaba but may occu r , infrequentl y , Lubumb ashi . The dryness o f threat t o p l ant l i fe . D a t a 2 .1 . 2 S o i l s a n d Vege t ation i n the south around and below / th� dry season makes f i re a major here are from Schul z e and McGee ( 1978 ) . Phytogeographi c al l y Shab a b elong� to the Z amb e s i an . Domain of the Sudano-Z ambes i an Region (Lebrun , 1947 , Duvigneaud , 1 958 , W erger & Coetze e , 1978 ) . Duvigneaud ( 1 958 ) recogni z ed three 2 5 1f 26. _______ ,..., ,., -. - ,"'&:. . � .,.. , , . .. . . . . "-. ""('\ . \ · · � { ....... . , ' - ... �) \ 0 �· "'ba. ,.... � - ' , __ Zambia ', ow. ' . . ___ ... .... ..... ...... ,. .... 27 . 2s · ..._.metal l i ferous rocks 0 20 40 60 I I ( \ I \ I \ km ' I I ) I I I I I l ) .J I / Zambia I ' l \ \ "'- ,_ F i g ura 2 · 1 = D i str i bu t i on of ma ta l l i ferous ro c k s in Shaba Prov ince2.. za·i re . 11 . .K@Y- to f i gure 2 · 2 = , 1 . Lubumbash i l L ' Etoi le2 2. Ruash i 3 . Kasanka north and N iamume2nda 4. Lu i swi sh i 5 . Lupoto 6 . K ipush i 7. Kamwa l i 8 . Lu i sh ia 9. Kamatanda 1 0. Ka konge2 1 1 . L i kas i 12. Kambove 1 3 . Fungurume 1 4 . M ind ing i 1 5 . Swambo 1 6 . Te n kQ 1 7. C ha bar a 1 8. Me2nda 1 9. Kasompi 2 0. Ti lwiz e2mbe2 2 1 . Ko lwez i 22 . D i ku luwe 23 . Ka lon gwe 1 f � 26 ° 1 7e e1 6 w13 -- 22e•21 20� e23 - � e1a .... "'"' __ ,. { ' � r \ ' ' \ \ , .._ \ -) \ _ , .. -'? ·�., 0 -- � � . ....� " Zambta ' ---'\ .... .... \ 1 4._ 15 27 ° e1o eg -12 e 11 � ( ea t1 · � Lu'" R. 0 25 km et. r- ----' a5e e3 e e I I 2 1 I) '• 6 I ' \ .... , ,� Zambi a ' , . ' Figure 2·2 = The major coba lt and coppe r m i n ing araas of Shaba Province ., za·ir cz . � 1 f sectors in Shab a ; "Lunda" ( renamed Central Angol an b y Werger & Coetzee , 1 978 ) , Lower K atangan and K a t ango-Z amb i an . The metall i ferous areas of thi s study are within the K a tango-Zamb i � n sector . Thi s sector i s character i zed b y "miombo " woodl ands . Miombo h as various species o f B rachystegi a , Julb ernar d i a and I soberl i ni a a s the major c anopy components . On mesic and ferti l e l oamy so i l s w ith rainfall above 1 , 1 00mm/year , the miombo develops a canopy up to 2 5m high wi th a wel l-devel oped understo rey , a t 5- 1 0m height , o f shrub s and smal l t ree s . A tall and dense herbaceous l ayer develops b e tween the l arger species and this p rovi des good fuel fo r the dry season fi res . Too much heavy firing may however degrade the miombo to g r as s l and (Werger & Coetzee , 1 978 ) . On d ri e r , poo rer soi l s the miombo has l e s s hei ght , appro x . 17m , and l e s s herbaceous under­ growth . On the poorest soi l s the unde rgrowth may b e repl aced by Cryptosepalum maraviense O l i v . 2 9 On deep soi l s w i th l i ttle compaction B rachystegi a longi fo l i a B ent h . a n d E rythrophl eum africanum (Welw . ) Harms . domi nate . On l i g h t s andy soi l s o f thi s natu re , they are joined b y Guibourti a coleosperma ( B enth.) J . L �onard and Burkea afri cana Hook . whi l e on the rich heavy red earths B rachystegi a spicifo rmis B enth . i s common . On shallower , g ravel l y soi l s w ith greater compaction and y el low or grey colourat ion , these species g ive way to B rachystegi a u t i l i s Burt t , D avy & Hutc h . These cond i tions a re o ften found on slope s . The areas o f very shal low o r skeletal soi l s , general l y on crests o f rocky massifs , are domi nated b y B rachystegi a mi crophyl l a Harm s . and B . bussei Harm s . On l at eri t i c soi l s , waterlogged in t h e rainy s e a so n , often gravell y and compacted , with grey or whi t e colouratio n , the d omi nant speci e s are B rachystegi a stipul ata D e W i l d . and I soberl i n i a tomento s a ( Harms . ) Crai b . & Stapf . W i th in the o pen forest (miomb o ) are areas o f termit e h i l l s ( " te rmi teri a " ) . A special flora h a s become associ ated wi th these termi teri a (Mala i sse , 1 978 ) . The fl ora tends to be scl e rophyllous i n nature wi th a w i de ly vari abl e species rang e . The " so i l " tends t o b e abnormal l y dry b u t nu trient ri c h . On alluvi al so i l s along the ma jor river val leys , ri verine fo rests are found . R i verine fo rest is a r ich and vari ed vegetational group i ng with speci e s o f o ther phytogeographi c al areas ( part icularly Gui neo-Congo l i an ) appearing . The soil s are general l y moi s t and r i ch . On b adly d r a ined o r i nundated s o i l s i n valleys , t he miombo vegetation g i ve s way to savannas : thorny , m icrophyllous savanna , b ro adleaf savanna o r mixed savann a . .The most important speci e s are herb aceous Hyparrhen i a spp . and woo dy Acaci a spp . Grasslands may develo p on sandy or shallow rocky soil s , d i l ungus and d ambo s . D i lungus are areas o f poor , waterlogged or wet soi l s w i t h l i ttle o r no aera t i on . Such aeration i s i nsuffi c i ent fo r woody spec i e s to g row s o t h a t these area s h ave steppes of various grasses and sedges as the i r p rincipal vege t ation el ement s . D ilungus are most frequently found o� old penepl a in surfac e s . D ambo s are fl a t , wide oval depressions character i zed b y poo r , b adly drained or impermeable clay soi l s . The vegetation i s again dominated b y g rasses and sedges al though some d ambo s may have woody spec i e s p re sent . D ambo s with woody spec ies are " b u shy dambos " whi l e those wi thout are " herb aceous damb o s " . A spec i al i z e d grassland community h as develo ped on metal l i ferous soi l s ( se e section 2 . 1 . 3 ) . A recent review o f the phytogeography of S h ab a c a n b e found i n Werger a n d Coetzee ( 1 978 ) whi l e a more detailed account can b e found in Duvi gneaud ( 1 958) . 2 . 1 . 3 Metal l i ferous Soi l s and Vegetation The metal l i ferous soi l s o f Shab a occur, general l y , on h i llocks o r c re s t s . The tops of these s tructures have skeletal gravell y soil s with outcrops of the underly ing rock s . The slopes are frequent l y composed o f colluvial depos i t s of the crest soil . Where the slopes are of non-mineral i zed mother rock the heavy metal content of the colluvion is d i luted . I l luvial depo s i t s may , however , enr ich such non-mineral i ze d are a s , parti cu l a rl y during , the rainy season . A contamination halo ( or " po i soned d ambo " ) , formed b y i lluvi al outwash , 3 0 general l y surrounds the crest o r h i l l . The halo i s surrounded by a corona until " no rmal " heavy metal content is reached i n the fo rest so i l s . 3 1 The presence i n the soi l of l arge quanti ti e s o f h eavy metal s severely restricts the g rowth of many plant specie s . I n particu l a r woody speci e s a re inhi b i te d . The vegetation o f these regions is thus herb aceous or , a t most , weak l y bushy . Duvi gneaud and Denaey e r-de-Smet ( 19 6 3 ) divi de the vegetational groups into five fo rms . E ach form exists on di fferent zones of each metal l i ferous are a . The five forms are : ( 1 ) the sw�rds , usual l y open , composed of grasses ( Sporobolu s , Eragrost i s , Monocymb ium , e tc . spp . ) , sedges (Bulbostyl i s , Asco l epi s , e t c . spp ) , annual d i co tyl edones ( H aumaniastrum , Cro tal ari a , e t c . s pp . ) , monocotyl edones w i th co rms o r bulbs (Gl adio l u s , L apeyrous ia , Eriospermum , D asystachi s , e tc . spp . ) and perennial d i co ty ledones wi th thick underground stock s ( Icomum , e tc . spp . ) ; ( 2 ) the s teppes , c losed fo rmation , wi th grass es (Loudeti a , Tri s tachya , Andropogon , etc . spp . ) predominat ing a n d a characteri s t i c geofrute (Cryptosepalum sp . ) associ ated w i t h vari ous perenni a l s w i th bulbs o r co rms ( Haumani astrum , Commel ina , e t c . spp . ) o r woody stock s ( Becium, Acalypha , e t c . spp. ) ; ( 3 ) t h ickets o f U apac a robynsii De W i l d ; ( 4 ) b ush s avanna of • transi tion to the surrounding open forest ; and ( 5 ) b rush on rocky ground with smal l tree s , bushes and shrubs . The flora of these metall i ferous soi l s i s generall y derived from the areas surrounding them . Duvi gneaud ( 1 958) consi ders that it may be deri ved from : (1 ) open forest and pioneer plants of rocky h i l l s ; (2) open forest on compacted yellow e a rth s ; ( 3 ) bushy d ambos ; ( 4 ) herbaceous dambos ; ( 5 ) U apaca robyns i i b e l t around d i l ungu s ; ( 6 ) dry steppes w i th geofrutes o f di lungu s ; and ( 7 ) mo i s t , grassy s teppes o f d i lungus . Duvigneaud ( 1 958) further cons i dered that the flora i s not derived from : ( 1 ) open forest on fertil e , deep red e arths ; ( 2 ) scle rophyllous vegetation o f termi terl a ; ( 3 ) riverine forests ; o r (4 ) a l luvial s avanna of Acaci a spp . The metal l i ferous flora may also be divi ded on an ecologi cal b as is ( Duvigneaud and Denaeyer-de-Smet , 1 963 ) . They recogni z e four major d ivisions : metallophytes , metallophi l e s , metalloresi stants a n d metal l i tuge s . Metallophytes g row on the richest metal l i ferous soi l s and, i ndee d , may be confined to such soil s . Sucn species include Hauman iast rum spp . , Bulbostyl i s spp . , I comum spp . , and Gramineae spec ies . Metal lophiles a re found o n soi l s o f l ower heavy metal content and are not confined to metal l i ferous so i l s . Exampl es i nclude Becium homb l ei , O lax ob tusi fol i a D e W i ld . , and U apaca robynsi i . Metal l o resi stants are ubiqui tous speci e s whi ch are i ndi fferent to the p resence of heavy metal s in the soil . They i nclude Loude t i a simpl ex ( N ees ) C . E . Hub�, Crotalaria corneti i Taub . � Dewevre , A ndropogon f i l i folius ( Nees ) S teud . , Xerophyt a spp . , Aeol anthus spp . , and Ti thon i a spp . Metal l i fuge speci es do not o ccur on metal l i ferous soi l s . Thei r p resence i ndicates a l ack of heavy metals in the soi l . Hyparrheni a spp . are in t h i s category . For a revi ew o f these ecolog i cal d ivisions s e e W i l d ( 1978 ) . 2 . 2 ANALYTICAL METHODS Vegetation sampl es from metal l i ferous h i l locks in Shab a were collected b y P ro f . F . Mal ai sse and Monsi eur J . Grego i re o f the Universi t� N�tionale du Z a1 re a t L ubumb ashi . These sampl es were washed and a i r-drie d . So il s amples were al so col l ected and dri ed . A summary of the pl ant communit i es a t Fungurume was al so p rovi ded b y these workers . Upon arri val the vegetation sample s , mainly leaf materi al , were analysed for the i r cob alt and copper content . S ampl es o f 0 . 02-0 . 04 g , dry weigh t , wera accurately wei ghed a n d placed i n 5cm3borosil i cate test-tub e s . These samples were ashed i n a muffl e furnace at 500°C for 2-3 hour s . The ash was d issolved 3 in 1 cm o f 2M hydrochlor ic aci d prepared from redi stilled constant-boi l i ng hydrochloric aci d . I f necessary sl ight warming was used to ensure dissolution of the ash . 3 2 S ampl es contaminated b y so il had a red , insolub l e sediment left after heating . These sampl e s w ere reject e d . Thi s method o f anal y s i s has b een found to b e sati sfactory for heavy met al determi nations in l eaf s ampl es (Wi ther , 1 977 ) . The sol utions were then analysed fo r coba l t and copper us ing a Vari an-Techtron model AAS atomi c ab sorption spec trophotometer . Automat ic background correction was made b y coupl ing a V ari an-Techtron model BC6 background corrector uni t to the spectropho tometer . Anal y s i s was carri ed out using cob al t l i ne s at 240 . 8nm and 304 . 4nm . and copper l i nes at 324 . 8nm and 2 1 8 . 2nm . Fo r each elemen t , low metal contents were determined at the mos t sensi t ive l i ne s , 240 . 8nm and 324 . 8nm respectivel y , and h igh metal contents at the l e s s sensi t ive l i nes , 304 . 4nm and 2 1 8 . 2nm respecti vel y . S t andards containing both cobal t and copper were prepared from BDH 1 000 �g/g anal yti cal g rade stock solutions by di lution w i th 2M hydrochloric aci d . The s tandard concentrations ranged from 5 �g/g t o 1 00 �g/g . Soil s amples were dri ed at 1 1 0°C overnight and si eved to 3 3 3 -80 mesh s i z e . Samples of 0 . 1 g were placed in 1 50cm polypropyl ene squat b eak ers and 20cm3of a 1 : 1 n i t ri c : hydrofluo ric aci d mi xture added . The s ampl e s were then taken to drynes s over a wate r-b ath and redi ssolved i n 1 0 cm3 o f 2M hydrochloric a ci d , prepared as above . Thi s volume was diluted as appropriate ( 1 0- 1 00cm3 ) for the expected concentrati ons . The solutions were then anal y sed for cob alt a nd copper a s outlined above . Extractab l e cob alt and copper soi l concentrations were determ ined b y sh�k ing O . Sg o f dry soi l , s i eved to -80 mesh si z e , in 5cm� o f an am�onium oxal ate-oxal i c a c id buffer , pH5 (Grigg ' s reagen t , Gr igg , 1 953 ) for 1 hour . The samples were l e ft overnight to equ i l ibrate b efore centr i fugat ion . The supernatant was then decanted o ff for anal y s is . Anal y s i s fo r cobal t and copper was c arri e d out a s for the other sampl e s . 2 . 3 COBALT AND COPPER IN VEGETATION OF FUNGURUME 2 . 3 . 1 S tudy Area The metal l i ferous h illocks o f Fungu rume are s i tu ated appro x . 1 80km northwest o f Lubumb ashi ( fi g 2 . 2 ) . The terrain i s i n marked rel ief to the plateaux which dominate the Shab an a rea . The herb aceous vegetation o f the metal l i ferous h illock s is in contrast with the open forest o f adjacent non­ metal l i ferous h i llocks ( pl ates 2 . 3 & 2 . 4 ) . The anomalous appearance o f the h illocks is accentuated by i rregul ar r idge s , composed o f cel lul ar s i l i ceous rocks , upon them ( p late 2 . 5 ) . A location map ( fi g . 2 . 6 ) indi cates the po sition o f the s i x metal l i fe rous hill ock s ( named I -VI ) in t h e study are a . Work was however confi ned to hillocks I I I - IV and V-V I . Malachite ( Cuco 3 . Cu (O H ) 2 ) i s the princi pal mineral i n the superfi cial so i l . The total copper content i n t h e soi l can reach 3 . 5% , dry wei ght b as i s , with assoc iated cob al t at 0 . 7% ( t abl e 2 . 1 ) . The extractable metal levels reach 3 . 2% copper and 0 . 3% cobal t . Percentage extraction b y the oxalate buffer i s h i g h , i nd i cating a ready avai l abi l i ty o f the metal s to the pl an t s . Copper extraction general l y exceeds 80% wh i l e cobal t has a lower extractab i l i ty , 45-80%. Thus copper i s mo re readi l y available than cobal t . 2 . 3 . 2 Pl ant Communi t i e s T h e plant communities at Fungu rume have been summari z ed b y specie s present i n t ab l e 2 . 2 . Schematic repre sentations o f these communi t i es and thei r di stribution ove r the h illocks are shown i n fig . 2 . 7 . Further b ri ef comments o n the communi t i es a r e g iven b elow. A-Open forest (miombo ) . The o pen forest su rrounding the mineral i z e d h i llocks i nt ru de only upon the l i ghtl y mine ral i z ed areas . The composi tion varies according to i t s posi tion . A calcicolous open forest dominated b y B rachystegi a bussei extends along the foot o f h i l lock s V and V I . 3 4 Pl a t<2 2 · 3 Vi <2w o f th<2 cupr if<2ro u s h i l locks of Fung u ru m<2 . I n thcz r i g ht for<2 gro und 1 s a blo c k of c e2 l l u l ar s i l i c e o u s roc k s } in t h <2 c e2 n t r<2 i s the s t e2 p p <2 s a vanna of the conta m ina t i on h a l o w h i l <2 in the b ac k gro undJ a c r o s s the2 va l l e2y o f t h<2 O i p e ta R iv<2 r J t h<2 p ro f i l <2 of h i l lo c k I i s s <2 <2n . N Pt a tc2 2 . 4 T hQ c o n t ra s t b e t WQQn n o n - c u p r i f e2ro u s a nd c u p r i fe2rous h i l loc k s . To t he2 l e f t J a n on-cupri fe r o u s h i ll o ck cove r Q d i n o pC2 n fo rQs t a n d i n th e c e n t r e J c up r i f e r o u s h i l l o c k ill cove rQd i n sh o r t he2 r ba c e o u s veg etat i o n. � -+- 0 - a... P ta t Q 2 . 5 V i e w of c e l l u l a r s i l i ce o u s roc k s . O u t cr o p s of c e l l u tar s i l i c e o u s ro c ks a s S Q e n o n t he l owe r s l o pes of h i l l o c k Tl . T h e r o ck s c ar r y a s p e2 c i a l f l o ra ; t h Q Ve llo z i a c eae s teppe i s s een be t w ee n the r o c k y ou t c ro p s . P i a t e. 2 . 5 Cl .c. (/) ..., .:X '+- u 0 0 � -· - E a. .c. 0 :J lO E (/) '-. ::J :J N 0 m c '- c � .o C:)l :J '- ....... '+- LL. ::J 0 '- m u 0.. 0 ::J '+-iLl _J u 0 < 0 w 0 L Ln � 0::: � E z � 0 LL. TABLE 2 . 1 To tal and extractab l e soil cobalt and copper i n some S h aban so i l s . COBALT � Soil Source T · E %E T Fungurume Lupo to Ruashi Mindingi A 9 23 780 85 1 892 B 9 3 6 1 6 6 280 c 67 5 1 7 6 333 D 387 205 53 5469 E 1 289 862 67 9206 F 6867 2657 39 35200 G 1 4 1 6 49 1 3 5 3 6 1 H 99 1 769 78 3 1 1 9 I 334 1 6 3 49 4 6 1 9 J 1 4 22 680 48 1 1 9 39 A 1 6 9 1 6 3 9 6 52 1 5 B 1 63 1 08 66 4498 c 1 24 7 1 57 475 1 A 467 340 73 6667 B 1 6.4 82 50 3284 c 1 0309 2 588 2 5 2 1 556 A 583 499 86 3 1 07 B 1 98 1 1 594 80 39394 c 2549 1 846 72 1 5496 D 4800 3984 8 3 44748 �Al l concentrations as pg/g dry soi l . T = To tal concentration . E = Extractab l e concentration . %E = Percentage of metal extracted . COPPER � E 1 830 235 3 1 4 2827 7088 32480 285 2 564 3077 1 0728 2885 3550 1 885 4669 1 799 1 6 1 1 3 2642 30876 1 4 970 36853 4 2 - %E 97 84 94 52 77 92 79 82 67 90 55 79 40 70 55 7 5 8 5 78 97 8 2 On h i ll o ck I V the open forest thins and change s , progressivel y , t o a n arbo raceous s avann a . Alb i z z i a ad i anth i fol i a , B rachystegi a spi ci fo rmi s , Dchna schwei nfurthi ana and V i tex madiens i s ssp . mi l anjens i s occur i n t h e l e ss vegetated stages . The copper content o f t h e so i l reaches 280 �g/g which is seven times the normal l evel for copper i n forested areas but simil a r to . that of an open forest , dominated by Brachystegi a m i c rophyl l a , a t L i k asi (Duvi gneaud & Denaeyer-de-Smet , 1963 ) . The o pen forest near h i l lock I l l is the ri chest flori s t i c al l y w i t h all the u su a l e l ements o f a m iomb o . 8-U apaca robyns i i t h ickets o n s teep s lopes . These th i ck ets occur a t the base o f the h i l locks . The presence of Loude t i a superb a D . N . s ignal s the t rans i tion t o an arboraceous savanna . C-S teppe-savanna w i thin the di spersion h al o . As i s common fo r ore depo s i t s , a d i spersion h al o o f heavy metal contami nat ion exi sts . As copper is mo re mob i l e than cob al t , a t Fungurume , thi s halo has a much greater content o f copper than cobal t . D-Vel loz i aceae steppe . I n Shab a , V e l loz i aceae popu l at ions are usual l y confined to rocky slopes o r , more rarel y , quart z it ic slabs (Mal a i sse , 1 97 5 ) . A t Fungurume they are found on so i l s ri ch i n cob a l t and copp er . X e rophyt a spp . a r e t h e domi nant speci e s of Vel l o z i ace ae . E-Commel inaceae and Convolvul aceae stepp e . These occur on so i l s w i t h ve ry high heavy me t al concentrations . The prime representat ives o f the fam i l i e s are Comme l i n a z igz ag and Ipomoea alpina respect ivel y . Smal l shru b s wi th woody s tems are common . F-Rendl i a cupri col a swa rd s . The access tracks a re bordered by a t h i ck sward o f R endl i a cup r i co l a exi st ing i n the fo rm of bell-shaped cushions . These swards a re mono spec i f i c and of only a few metres dimension . 4 3 G-Oxytenanthera abyssinica ( A . R i c h ) Munro thickets . T h is fo rmation i s found on al luvi a o f the D i peta R iver and represents the transi tion from the fo rest ( o ften degraded to Park i a fi l i co i de s ) to the vege tation of the minera l i z ed h i l lock s . H-Haumaniastrum robertii carpet on mi ning wo rk s . The excavations resul ting from mining acti vities and the re si dual soi l s , r ich i n mineral s , a re favoured local i ti es fo r the occurrence o f a c arpet o f Haumani astrum rob e rti i . G rasses and sedges may accompany thi s speci e s . ! -Loose thi ckets of Euphorb i a i ngens . A steep face o ri ented towards the no rth and hence sub jected to sunshine during the cool dry season suppo rts a parti cul ar vege tation . Euphorb i a i ngens (a c acti form pl ant o f candel abra-like aspect ) and S arcostemma viminal e ( a l i ana w i th crassul escent stem ) characteri z e these thickets . Plants typical o f termi teri a are a lso present . The i r xerophi l i c and c rassul escent tendenci e s have been previ ously noted (W i l d , 1 9 52 , Aub rev i ll e , 1 9 57 , Colonval-El enk ov & Mal a i sse , 1 97 5 , Mal aisse , 1 978 ) . J-Denuded ve r t i cal f ace s . Several vertical faces denu de d of vegetation were ob served on the sou t h e a s t slopes of h i l lock s V and VI ( f i g . 2 . 7 ) . K-Vegetation o f ou tcrops o f cel lular s i l i ceous rock s . Thi s vegetation g rouping has been termed " chasmophyt i c " because o f the di ssected n a ture o f the sub s trate . Rocks receiving sunshine are covered b y two l i chens , one frut i culous and the o ther fol i ar . I n the cavi ties o f the s i l i ceous rocks i s found Eupho rb i a fanshawei , a cactus-l ike pl ant wi th napifo rm roo t s . The accumul ation o f a l i ttle s o i l i n the c avities resul t s i n the appearance o f Aeol anthus rosul i fo l iu s , Faroa acaul i s , Monadenium sp . and several Pt.e r idophytes . The verti c al wal l s o f the ro cks have cushions containing Aeol anthus s axati l i s on them . · The shaded cavi ties are covered w i th a c arpet o f Hepaticas .domi nated b y Pl agiochasma eximium. A number of o ther spec ies al so tolerate thi s envi ronment . , 4 4 4 5 TABLE 2 .2 Pl ant c ommuni t i es on mineral i z e d h i l locks at Fungurume , Zai re . H Community A-Open forest ( up to 280 tJQ/g Cu i n so i l) . B-Steep thickets (250 1-'Q/g Co and 350 1-lg/g Cu so i l ) . C-Steppe-savanna (up to 350 1-fQ/g Co and 5000 1-lg/g Cu in soi l ) . D-Vell o z i aceae steppe (up to 1 500 1-fQ/g Co and 12000 pg/g Cu in soi l ) . E-Commel i naceae and Convo l vul aceae steppe ( 700 fJQ/g C o and 25000 fJQ/g Cu in so i l ) . Dominant Species B rach�ste9i a bussei U apaca robynsi i X e rophyta retinervi s egu i setoi des x . demeesmaek e ri ana - Commel ina zigz ag Ipomoea alpina var . Associ ated Species Albi z z i a adianthi fo l i a A . antunesi ana Brachystegi a spic i fo rm i s Cussoni a arborea Dalbergi a boehmii Dipl o rynchus cond�locarpon K i rk i a acuminata Dchna schweinfu rthi ana Pseudo l a chnostyl i s maprounei fol i a Stercul i a guingueloba Stegano taeni a aral i acea Psycho t ri a sp. Vi tex madi ensi s ssp . mil anjensis Loudeti a superba Phragmanthera rufescens var . co rne ti i Haumani astrum rosu l atum Loudeti a simpl ex Polygala pet i t i ana P . usafuensis - Crassu l a alba Lapeyrousra-Brythranthra var. welwi tschi i Morae a carsoni i Pandiaka metall o rum Spu riodaucus marthoz i anus A lectra sess i l i flora Anisopappus davyi Bulbostyli s abo rtiva B. mucronata Crotal aria cornet i i Cryptosepalum dasycl a dum Cyanoti s longi fo l i a Hibi scus rhodanthus Sopub i a drege ana F-Rendl i a cuprico l a swards G-D xytenanthera abyssi n ic a thickets H-Haumaniastrum robertii carpet !-Loose thickets of Euphorb i a i ngens J-Denuded ve rti cal faces K-Vegetation of outcrops of cellul a r s i l i ceous rock s . Rendl i a cupri col a Dxytenanthera abyssi nica Haum an i astrum roberti i Euphorb i a i ngens S arco stemma vimi nale Various � See al so Figure 2 . 7 Park i a fi l i co i des Bulbostyl i s abort iva Eragrosti s bo ehmi i Rendl i a cupricol a Annona senegalensis Cusson i a arbo rea Dioscorea bulbi fera S el aginel l a abyssini c a S t eganotaeni a ara l i acea Tacca leontopetal o i des Aeol anthus rosu l i fo l iu s A . s axat i l i s Aneimi a ango l ens i s Anthoceros punctatus A . mandoni Euphorb i a fanshawe i Faroa acaul i s Fo ssomb ron i a s p . Gongylanthus eri cetorum Mohria c affrorum Monadenium s p . Pel l a e a pectin ifo rmis �· goudotii Plagi ochasma eximium R i cc i a sp . Targion ia hypophyl l a N TII 1 3 00 1 250 1 200 1 1 50 Co 93 271 6 81 0 Colbalt and Copper E Cu 280 361 2 5 200 values refer to dried Q) soi l and are in u n i t s ""0 of J.J9 /g . '::J - - n/nn fh, Js s pe2cie s . A . A. rosulifolius scales . simple g landular �� CQllS B. A saxa tilis C. A . biformifolius D. A . saxafilis st i ff 2 -·3- cQ llad . ha i rs w i th smooth ce ll wa l ls ( 0 ) o r g ra nu la r wall s ( E ) E. A. rosulifolius flgure 3 . 2 Fo l iar i d umenta for three mata l ­ t ol e r a n t s p a c i a s of Aeo!an thus . tropi cal Afri can speci e s . Several spec ies appear to b e double­ l i sted b y Index Kewen si s . A . b i fo rmi fol i u s i s one such speci e s , being l i sted a s I . b i fo rmifol ium D e W i l d . a s wel l . The original reference c i ted for both names is De W i l deman ( 1 927 ) . De W i l deman ( 1 927 ) formed �· b i fo rmi fol ius by spl i tt ing A . l i neari s ( also known as I . l i neare Burk . ) i nto three new spec i e s . The o ther two species formed b e i ng �· elongatus (!. elongatum De W i l d . ) and �· tubercul atus (!. tubercul atum De W i l d J . None o f �· l i neari s , A . elongatus or A . tub erculatus show l evel s of cobalt or copper i n l e a f t i s sues above tho se " normal " to mi ne ral i z ed a reas . T h e �· g roup l i neari s species a r e found on col luvion whi ch i s r ich i n copper o r i n crevi ces on cel lular s i l i ceous rock s . The o ther copper hyperaccumulato r , �· rosu l i fol ius , and the very strong cob a l t accumu l ato r , �· saxati l i s , are both found o n cel l ul ar s i l i ceous rock s . The metal l i ferous species o f Aeol anthus are general l y geophyte s wi th corms or hemicryptophytes w i t h woody s tems . The growth cycle o f these speci e s is o ften sho rt : at Mine de l ' t toi le , �· b i fo rmi fo l i u s has a three month " rainy season" growing cycle (Mal a i s se & Grego i re , 1 978 ) . 3 . 4 I POMOEA L . ( CONVOLVULACEAE ) Ipomoea L . ( Convolvul aceae ) i s a l a rge genus ( approx . 500 speci e s ) of tropical and warm temperate regions (Wi l l i s , 1 973 ) . The plants themselves are o ften vines o r creepers. The genus i s p rob abl y best known a s the genus o f the sweet potato , !• b atatas Po i r . The genus is wel l represented i n Shab a although d iff icul t i es i n speci f i c i denti fi cation do occur ( Duvi gneaud & Denaeyer-de-Smet , 1 96 3 ) . Thi rty-ei ght specimens covering twenty-seven taxa were anal ysed for the i r cobalt and copper content in this survey . All taxa are from the Central A frican reg ion . The resul ts of t h e anal yses a r e shown i n tab l e 3 . 2 . No hyperaccumulation o f either cobal t o r copper has been recorded for thi s genus e i ther b e fo re , o r during , t h is survey . However the fact that very s trong accumulation o f cobal t has b een shown ( Chapter 2 ) and that the di stribution of taxa over the various 7 4 7 5 TABLE 3 . 2 Cob alt and c opper c oncentrations i n some Ipomoea s peci e s . Species Location Coba l t M Copper M 1.- algina Rendle �i t ega , Burundi X X 40 Etoil e , Za i re 1 78 6 3 l· algi na Rend le ssp . hock i i Etoile , Z a i re 1 1 6 489 (De W i l d. ) Duvi g . et Dew i t . l · aguatica Fo rsk . K ambove , Z ai re XX 22 l· arachno sgerma Welw . Kasenga , Z a i re XX 10 1· b arteri B ak . Kasapa , Z a i re X X 1 7 Lubumbashi , Z ai re XX 29 1 · b l egharoph�l l a Hall . K antal a , Z ai re XX 20 l · cairica ( L . ) Sweet Eto il e , Z ai re 4 49 J . crassige s Hook . var . crassipe s Mwi nilunga , Z ambi a XX 7 J . cregi d i fo rm i s Hal l . f . var m i c rocephal a ( H a l l . f . ) Verdc . Aberco rn , Z ambi a XX 1 0 1 · dammeriana De W i l d . M anono , Z ai re 2 36 1 · debee r st i i De W i l d . K atuba , Z ai re XX 68 1 · digi tata L . Kakonkani a , Z ai re XX 28 1 · eriocarpa R . B r . Bunkeya , Z ai re XX 1 2 1 · ful vi caul i s (Cho i sy ) Hall . f . Muh i l a , Z ai re XX 1 5 var . a speri fol i a ( H all . f . ) Lubumbashi , Z ai re 4 1 2 Verdc . I . i nvolucrata B e auv . var . L akulu , Z a ire XX 1 7 i nvolucrata Keyberg , Z ai re XX 55 1 · l apath i fo l i a Hall . f . Rusi z i , Burundi XX 1 4 var l agathi fol i a I . - l i no sepal a H all . f . K afubu , Z aire 5 37 Lubumbashi , Z a i re 2 44 L u i swi shi (M ine ) , Z ai re 44 57 I . lukafuensi s D e W i l d . Lutshi puk a , Z a i re XX 1 5 K ibondj a , Z ai re XX 1 1 0 Lubumb ashi , Z a i re 6 1 1 9 I . o chracea ( L i ndl . ) G . Don . var . o ch racea L o fo i , Z ai re XX 22 I . pes-caprae ( L . ) R . B r . s s p . - 6ras!I!ensis C L . ) van Uvi ra , Z a i re 1 1 8 Oo s t s tr . j . pharbi t i fo rm i s B ak . Katshup a , Z a i re XX 6 · Mutombo-Mukulu , Z a'i re XX 1 9 Ki sanga , Z a i re XX 1 2 I . pes-tigr i di s L . v a r . Upemb a , Z a i re XX 4 ees-tigrid i s I . ei l eata Roxb . Keyb erg , Z a i re XX 2 1 I . erismatOS:::£Ehon W elw . M aseb a , Z a i re XX 5 Lungulungu , Z ai re 2 8 I . rub ens Choisv D ikuluwe , Z a ire XX 1 1 I . shueangensi s Baker Lubumb ashi , Z ai re 3 66 I . vernal i s R . E . F r . Kapiri , Z a i re 1 1 7 M Concentrations expressed a s pg/g dry wei ght . xx Cob al t concentration l ess than 1 pg/g . hillocks has been studied b y Duvi gneaud and Denaeyer-de-Smet ( 1 96 3 , pp . 1 36- 144 ) makes a study o f the cob a l t and copper content of i nterest . Many o f the taxa o n metal l i ferous soi l s have d i scontin­ uous ranges over the hi l lock s . The h i ghest cobalt content i n t h i s survey was found i n l· alpina ( 1 78 �g/g ) w ith l • alpin a s sp . hock i i having a s l i ghtly l ower conten t . l· l i nosepal a and I . luk afuensi s al so show anomal ous values ( > 5 pg/g ) . l• alpin a ssp . hock i i had t h e h ighest copper content (489 pg/g ) , j u s t below very strong accumul ation conten t . I . luk afuens i s had the second h i ghest copper content but b arel y passed the 1 00 �g/g l evel . O ther spec i e s to 7 7 record anomalous copper concentrations ( > 50 pg/g ) were l• debeersti i , l• shupangensi s , l· alpi na , l· l i no sepal a and l• i nvolucrata var . involucrata . I n the fi e l d surveys ( Chapter 2 ) , l• alpina had shown very strong accumulation o f cob a l t ( table 2 . 3 ) but the sub spec i e s fo r this specimen is not k nown . Duvigneau d and Denaeyer-de-Smet ( 1 96 3 ) have recorded l• alpina ssp . a rgyrophyl l a Duvi gn . & Dewi t . a t Fungurume . They found thi s taxa g rowi ng o n soi l s low i n copper wi th I . debee rsti i s sp . debeerst i i repl acing i t on so i l s rich i n coppe r . l· debeerst i i does not appe ar to b e a s trong accumul ator o f ei ther me tal al though i t c an reach anomalou s l eve l s fo r both . Anomalous leve l s are however " no rmal " for plants on metal l i ferous substrat e s . Further research comb in ing botanical studies o f me tal l i ferous-so i l taxa and biogeochemi cal studies of the i r cob alt and copper up take may help i n understanding the causes o f the di scontinuous ranges of the vari ous taxa and the nature o f the i r evolut ion . 3 . 5 PANDIAKA (MOQ. ) HOOK . f . ( AMARANTHACEAE ) Pandiaka (Moq . ) Hook . f . ( Amaranthaceae ) i s a g enus o f some twenty speci e s from tropical and southern Afri ca . Members of the genus tend to be herb aceous i n . g rowth . Fi fteen sampl e s covering e ight taxa o f Pandiaka were anal ysed fo r cobalt and copper . The resu l t s of these analy s e s are shown i n tab l e 3 . 3 . The most striking fea ture o f table 3 . 3 i s the hyperaccumulation o f copper b y f. metal l o rum . Thi s species has a l so shown copper hyperaccumulation i n the Fungurume fi e l d survey ( Chapter 2 ) . The cob a l t content i s a l so markedly e l evat ed , reachi ng the l evel of very �- E_. �- P . P . �- P . P . TABLE 3 . 3 Coba lt and copper co ncentrat ions in Pand i aka species . Species . andongensi s H i ern . carsoni ( B ak . ) Cl arke carsoni ( B ak . ) C l arke l i ne ar i fol i a Hauman gl abra (Schinz .) Hauman k as sneri Suessenguth var . metal lorum Duv i g. et Van Back . obovata Sue s s . �o l�s tach�a Sue s s . Location K as enga , Zai re M arungu , Z ai re M an i k a , Z ai re Lubumbashi , Z a i re K al ashi e , Z aire Lu iswi shi , Z ai re Mukuen , Z ai re · K eyberg , Z ai re K al a , Z ai re Fungurume , Z ai re Fungurume , Z a i re Fungurume , Z ai re K atub a , Z ai re Katema , Zal re K anseni a , Z a i re � Concentrations expre ssed as �g/g dry weig h t . x x Concentrat ion l ess than 3 �g/g . 7 8 Cob alt " Copper " XX 1 8 XX 1 5 XX 5 XX 79 XX 7 134 1 23 XX 39 4 1 20 XX 8 570 6270 1 6 1 1 1 30 1 0 1 629 25 27 XX 1 3 8 8 strong accumul at ion . The fi e l d survey s ample at 448 �g/g was only just below thi s l evel . O ther anomalous values were recorded fo r both cobal t and copper i n �· c arsoni var . l ineari fo l i a , for coba l t o n ly i n P . obovata a nd P . polys tachya , and fo r copper onl y i n �· carsoni and �· gl ab r a . A l l s ampl e s , anomalous o r not , were from Z ai re . �· carson i , �· carsoni var. l i neari fol i a and �· gl abra may be local l y u seful fo r i ndicating mineral i z at i on but none are unive rsal indicato r s . Thus for P. c a rsoni var . l i neari fol i a , a specimen from Lu i sw ish i Mine showed s i g n i fi cant heavy metal l evel s whi l e a specimen from the Kalashie sal t pan showed onl y trace 7 9 amounts . I t i s necessary to note that P . metallorum h ad previ ou sly been considered to b e a vari ant o f �· c arsoni ( Erns t , 1 974 ) . Duvi gneaud and Denaeyer-de-Smet ( 1 96 3 ) consi dered that P . metallo rum prob ably h a s many vari e t i e s , b as ing the i r suppo s i ti on on the w ide variab i l i ty in l e af forms and s i z e s . They a lso consi dered the spec ies to b e l i t tl e di fferent from other spec ies found in di lungus and dambo s and was prob abl y derived from them by ecol og i cal i so l a ti on . Unl i ke most o ther speci e s whi ch o n l y flower i n e i ther t h e rainy o r the dry se ason , �· metallorum fl owers throughout t h e y e a r . The speci e s is wi despread in Shab a , b e ing present on vi rtual l y all the cupri ferous outcrops . 3 . 6 GENERAL D I SCUSSION The surveys of Aeolanthu s , Ipomoea and Pandiaka have revealed only one new hype raccumulato r : �· rosul i fo l i u s showed hyper­ accumul ation of copper . The hyperaccumul ation o f cob alt and/o r copper i n �· b i fo rmi fol i u s and P . metallorum w a s confi rmed . Al so revealed was the abi l i t y o f Aeol anthus speci e s , in general , to accumulate s i gn ifi c ant quanti t i e s of cob alt even when not g rowing o ver notably r i ch sub strates . The use o f herbarium specimens h a s greatly a ided the rapi d i t y w i th w h i c h t h e systemat ic surveys o f genera fo r metal accumul ation can b e carried o u t . Thi s i nformation may b e useful i n helping wi th the del ineation o f specie s , o r lower taxonomic rank s , for plants found o n metal l i ferous soi l s . Undoubtedly the final del i neations will b e made o n flor i s t i c characters but how shou l d the e ffect o f physiologi cal stress from an adverse (metal-ri c h ) environment as a 8 0 cause of mo rpho log i cal di fferences be noted in the rank i ng? Such probl ems are no t unk nown among n i cke l hype raccumul ators : D i coma n iccol i fera was or ig inal l y described as a Di coma macrocephala subspecies (W i l d , 1 970 , 1 97 1 ) a nd t he Alys sum taxon , �· serpyl l i fol ium ssp . l us i tani cum shows hyperaccumu l a ti o n while the t axon �· serpyl l i fol ium s sp . serpyl l i fol ium does not ( B rook s & R adford , 1 978a ) . T h i s i s a ga in a form o f ecotyp i c differen ti at ion b u t b e tween taxa o f metal l i ferous and non-metal l i ferous soi l s . The e x i s tence of ecotyp i c di fferenti at ion b e tween t axa o f various met al l i ferous soi l s onl y serves to compound the i ssue . I t has been noted i n t h i s chapter that P . met a l l orum appears to have ecotyp ic di fferent i a ti on between t axa on metal l i ferous or non­ metal l i ferous so i l s and between the d ifferent me t al l i ferous outcro p s . Eco t yp i c di fferent i at ions w i t h i n Aeol anthus a n d Ipomo e a however are general l y b e tween taxa on the d i fferent metal l i ferous outcrops . Finding c auses for these ecotyp i c di fferenti ations w i l l have to i nvolve w i de spread sampl ings of specimens from both metal l i ferous and non-metal l i ferous soi l s and experimental work on thei r phy s i o l ogy and morphology . Shewry � �· ( 1 97 9 ) h ave sh own mo rphological . di fferences b e tween Cryptosepalum maravi ense specimens on a rock y , cupri ferou s so i l a t K a z inyanga and a non-cupri ferous s o i l near D ikuluwe . They have a l so investi gated Xe rophyta specimens , a t Dikul uwe , on a rock y , cupri ferous soi l , a rocky , non-cupriferous so i l and i n a cupriferous dambo and found that two b a s i c morpho logical forms exi s ted : form one on rocky s i t es w i t h or wi thou t copper enri chment ; form two bn deep , fine , dry soi l in the cupri ferous damb o . There were small e r di fferences between the specimens o n the rock y , cupri ferous s o i l and the corresponding non-cupri ferous soi l . Thus i t c an b e seen that i t i s not onl y the metal content which h as an effect on the pl ant mo rphology but that o ther soil factors are also involve d . Howard-Wi l l i am s ( 1 970 ) found s i gni ficant di fferences in l eaf shapes o f B ec ium homb l e i specimens col l ected on soi l s containing over 3,000 p g Cu/g compared w ith those g rowing o n s o i l s w i t h l e s s than 1 , 000 pg Cu/g . These resu l t s , from f i e ld s ample analysi s , coul d not b e reproduced i n shade-house experimental t r i al s . F rom thi s , Howard-W i l l i ams suspected that the di fferences were due to m i cro cl ima t i c factors rather than so i l-metal l evel s . Thus t vari a t ions b e tween p l ant specimens o f the same or closel y-rel ated spe c i e s can make s pec if ic i denti ficat ion d i ffi cul t b ecause of the wide range o f vari at ions po ss ib le as responses to many and vari ed envi ronmental facto r s . W i ld and Bradshaw ( 1 977 ) l i st ten speci e s whi ch show some deg ree o f morpholog i cal di fferenti ation between tol erant and non-tolerant popu l a tions on toxic ( copper-ri ch , n i ckel-rich o r serpentine ) so i l s i n Z imb abwe . I n the final anal y si s , ecotypic di fferenti at ion b ecomes a question as to whether the di fferences are " re a l " ( genet i cally i nheri ted ) or merely a re su l t of s tre s s and t h i s can onl y b e determined by fu rther experimental work . 8 1 CHA P TER 4 B i ogQo chC2m ical Stud ie2 s on Some Me2tal l ophy tes from Shaba 4 . 1 I NTRODUCTION 8 3 The e x i stence o f metallophytes l eads i nevitab ly to questions concerning the abi l i ti es o f these plants to tolerate the metal s . Fo r those metallophytes which are a l so hyperac cumul ators , a further question arises a s to the maximum amount which can b e accumulated wi thout k i l l ing the plan t . Few studies h ave been made o f cobalt and copper metallophytes however , and none have b een made of cob alt and coppe r hyperaccumul ators . Among the copper-tole rant spe c i e s stud i ed are several from the Dugal d R i ve r area o f Austral i a ( N i co l l s e t al . , 1 96 5 ) , B ecium homb l e i ( R ei l l y , 1 969 , Howard-Wi l l i ams , 1 970 ) , Mimulus guttatu s DC . ( Allen & Sheppar d , 1 97 1 ) , I ndigo fera spp . ( E rnst , 1 972 ) , Agrosti s stoloni fera L . (Wu e t al . , 1 97 5 ) , Agro s t i s gigantea ( Hogan � al . , 1 977 ) and two Fennoscandi an spec ies ( C rook s , 1 979 ) . Most of these s tud ies have inve stigated tolerance to metal s wi th a mo re l imited numbe r involving uptak e characteri st ics . I n s tudies o f uptake characteristi cs , three general fo rms o f uptake can b e di st ingui shed : the exclusion-break down fo rm ( N i co l l s et al . , 1 96 5 , Wu et al . , 1 97 5 , Crook s , 1 979 ) , the ri se­ to-saturat ion fo rm ( Rei l l y , 1 969 ) and the l inear form ( N i co l l s e t al . , 1 9 65 ) . The exclusion-b reakdown form , characterized b y entry restrictions at low soil metal l evel s with an apparent b reakdown o f the exclusion mechani sm w i th consequent accumul ation o f t h e metal on soi l s with a h igher metal conten t , can be found fo r almo s t any metal studi ed ( eg . n i ckel , copper , z inc , l e a d ) i rrespective o f e ssent i al i ty o r non-essential i ty . Thi s contrasts w i t h the work o f Timperley e t al . ( 1 970b ) who sugges t that d ifferent uptake pattern s coul d b e expected for e ssenti al a n d non-essenti al e l ements . I n thei r studi e s , e ssen t i al element s appeared to h ave the exclusion­ b re ak down fo rm and non-essenti al e l ements the l inear fo rm. Further contrast to Timperley e t al . ( 1 970b ) comes from a s tudy o f B ecium homble i ( R e i ll y , 1 96 9 ) i n which copper uptak e shows not the exclusion-b reakdown form but the ri se-to-saturation form. Thi s form i s characteri zed b y a l i near r i se i n the metal content o f t h e plant a t l ow soil metal l evel s w i t h a flattening o f thi s curve a t h igher soi l metal l evel s . Whether a real saturation l evel of some kind i s reached o r some o ther e ffect g ives an apparent saturation l evel is unknown . It shou l d b e no ted that i n t h e i r stud ies on Agrosti s stoloni fera , Wu e t al . ( 1 97 5 ) showed both these forms o f uptake : the exclusion-b reakdown form for l e aves and the ri se-to-satu ration form for root s . As coul d be expected the point o f exclusion-break down ( i e . the po int where the copper concentration begi n s to ris e ) i n the l eaves corresponded well with the saturation point in the root s . No other study has i nvolved both root and leaf uptake patterns . The l inear form of uptak e , character i z ed by a purely l i near cu rve , has been ob served 8 4 for z inc i n three Austral ian metal-tolerant species ( N i co l l s e t al . , 1 96 5 ) and for non-essential e l ements i n non-tolerant pl ants ( Timperley e t al . , 1 970b , B ecket t & D avi s , 1 977 ) . The tolerance o f a speci e s to a parti cular metal can b e tested i n various way s . I n 1 9 57 , W ilkins conce ived t h e solution tolerance test i n which the pl ant to b e tested was rated on root g rowth i n control and experimental solutions . From the root g rowt h , a tol erance i n dex , I , was calcul ated : I = l ength o f roo t i n experimental solution length o f roo t i n control solution X 1 00 For h i s control solut ion W i lk i ns ( 1 957 ) tested both d i st i ll ed water and a full nutri ent solut ion . He found that the concentration l evels requi red for lead studi es were too l ow i n aqueous solution to b e easi l y h andl ed but were h igher and , hence , more easily h andle d i n the ful l nutrien t soiut ion . B y testing the components o f the full nutri ent solution indivi duall y , he found that calcium amel i o rated the tox i c i ty o f the l ead . For h is subsequent work , and for most o ther workers (Jowett , 1 958 , 1 964 , McNeil l y , 1 968 , Allen & Sheppard , 1 97 1 , Wu & Antonovi c s , 1 97 6 , Crai g , 1 977 ) , calcium ni trate a t 0 . 5- 1 g/l h a s been added t o di st illed water to form t h e control solution . Thi s t est can not only show tol er ance to one metal but can a l so b e u sed to arrange o rders o f tol erance ( o r conversely o rders of tox i c i t y ) for vari ous metals within a g iven species (Jowett , 1 958 , Crai g , 1 977 ) . A second techniqu e , comparative pro topl asmatology , has been used b y Repp ( 1 963 ) , Gri es ( 1 96 6 ) and Ernst ( 1 972 ) . Thi s method uses cells , generally from shoot s , and places them i n g raduated metal solutions for a g i ven t ime b e fo re being tested for vitality (the ab il ity to pl asmol i se in sugar solution ) . This method gives only a l imit of tolerance . Allen and Sheppard ( 1 97 1 ) developed a soil rooti ng method as well as using the solution rooting method above . Thi s new method 8 5 involved potting cuttings i nto a copper-rich soil and an uncontaminated soi l , leaving them to develop roots and then cal cul ating a tol erance index as for the solution rooting technique . The soil rooti ng technique appeared to work as well as the solution rooting techni que. • Horscroft ( 1 96 1 ) reported germination tests on Becium homblei in a solution culture experiment and found that the species requi red 50-600 �g Cu/g for germination which suggests a physiologi cal requi rement for high copper concentrations . Howard-Williams (1 970 ) , however , found that the same species germinated readi ly in di stilled water . Allen and Sheppard ( 1 97 1 ) tested the germination and establishment of Mimulus guttatus seedl ings in contaminated and uncontaminated soi l s . They found that thi s test easily di sti ngui shed between the tolerant and non-tole rant popul ations . All tolerant popul ations germinated readi ly in uncontami nated so i l s . The studies reported here look at the uptake and accumulation of cob alt and copper in three Shaban hyperaccumulators : Haumaniastrum k atangense , tl• roberti i and Aeol anthus b i formi fol ius . The tolerance of two of these speci e s , tl• katangense and �· b iformi folius , was d�termined by a soi l culture method. Germination experiments were also carri ed out on two species ; H . roberti i and �· b i formi fol ius . Finally , the uptake o f copper by �· bi formi­ folius was followed through a growing season . 4.2 EXPERIMENTAL METHODS Seeds o f three metallophytes (Haumaniastrum katangense , tl• robertii and Aeol anthus b i formi fo lius ) and corms o f A . b i fo rmi­ fol ius were collected i n Shaba Province , Z aire ( Fungurume or Mine de L ' Etoi l e ) by Pro f . F . Malaisse o f the Universite Nationale du Zai re at Lubumbashi . The seeds were stored under refrigeration until used. . . Before any germination could proceed , the seeds had to be washed in running water for 4-7 days. This is presumed to be necessary because of an adaptation to the cl imatic conditions of their homeland . The seeds germinate at the beginning of the rainy season since most plants- there grow during the rainy season and die-bac� during the dry season . I t i s necessary therefore that seeds do not germinate too earl y . It appears that some inhibitor of germination has been developed and only the continued presence of excess water overcomes this inhibition , prob ably by leaching the inhibitor from the seed . 4 . 2 . 1 Uptake and Accumulation Tri al A potting mixture of 50% peat - 50% perl ite with added nutrients was u sed in this trial . To the basic mixture was added cobalt and/or copper ( as nitrates ) to give the followi ng conditions; 1 00 , 1 , 000 and 1 0 , 000 �g/g of cobal t , o f copper and o f cobalt + copper . B ackground mixture ( i e . n o addi tives ) was used for control purposes . The potting mixture was pl aced in pots each of which contained approx . 200g . Each concen tration of heavy metal s was repl icated five times . One seedl ing was planted per po t . The seedl ings had been germinated on a Copenhagen table after washing and then pl anted i nto b ackground potting mixture unti l the second pai r of l eaves developed. At thi s time the seedlings were transplanted into the experimental pot s . All pots were watered from beneath . All three metallophytes were grown. Leaf samples were collected for analysi s after five weeks and every second week thereafter . All su rvivors were sampled in e ach condi tion and a composite sample was analysed. The soi l s were sampled to determine their cob alt and copper content after two months . A supplementary uptake �rial was made using corms , rather than seeds , of �· b iformi folius . The same condi tions with respect to the potting mix and metal content applied . Five replicates were al so used at each concentration. The corms were planted di rectly into the potting mix with one corm per pot . After three months , the leaves and the soi l s were sampled for analysi s • . . 8 6 All leaf analyses were perfo rmed b y the method used for the f ield survey samples i n Chapter 2 . The soi l analyses were performed by d igesting 0 . 1 g of oven-dried soi l with 20cm3 o f aqua regi a . This digest w a s taken to dryness over a water-bath and then redissolved in 1 0cm3 o f 2M hydrochloric acid ( prepared from r e distilled constant-b o i l i ng hydrochloric aci d ) . Thi s solution was centrifuged to remove the unreacted s il icates and the supernatant decanted . necessary (up to 1 00cm3 ) . The supernatant was then diluted as Analysis was perfo rmed by atomi c absorption spectro photometry a s for Chapter 2 . 4 . 2 . 2 Tolerance Tests A tol erance test was dev ised whi ch used the peat-perli te po tting mix o f the uptake tri al . See d s o f �· k atangense and �· b i fo rmifolius were germinated o n a Copenhagen table and then planted i nto bedding trays. After the development o f a second pai r of leave s , the seedl i ngs were transplanted into v ials containing approx . 2g o f back g round potting mi x . After a fu rther ten days , to allow for recovery after the transpl ant , solutions o f cob al t o r copper were added to g i ve the de sired concentra tions : fo r �· k atangense , background and 1 00-64 , 000 �g/g o f ei ther 8 7 cob al t o r copper w i th four repl icates per concentration ; for A . b i fo rmifolius , b ackg roun d , 1 00- 1 6 , 000 �g/g o f cobalt and 1 00-64 , 000 �g/g o f copper also with fou r repl icates per concentrat ion . The tests were continued . until stable results were achieved ( i e . no change over three weekly readings ) . The tolerance l evel was taken as the h i ghest concentration with at l east half the seedlings surviving . O ther concentrations for which any i ndivi dual specimen survived were also noted . All so ils were analysed for cobalt and copper content after the test. The analyses were perfo rmed as for the soi l s in the uptake t r � al s . Leaves of �· k atangense were also analysed for their cob al t and copper content . Analysis o f these samples was performed a s for field survey samples in Chapter 2 . Leaves of A. b i formi fol ius were not analysed because o f algal p roblems encountered in this tes t . 4 . 2 . 3 Germination Tests After the washi ng period , seeds of two speci e s , H . roberti i and A . b i formi folius , were pl aced on filter papers within petri di shes . The fil ter papers h ad previously b een soaked i n solutions of cobal t , of copper or of cobalt � copper over a range from 0 . 1 - 3%. A control was kept by germinating seeds on a fi lter paper soaked i n deioni zed water . Twenty seeds were u sed per concentration . All fi lter papers were k ept moi st; if necessary de ionized water was added to maintain the moi sture level . The test was continued fo r three weeks . Both the rate and the percentage of germinations were recorded . 4 . 2 . 4 Seasonal Copper Uptake i n Aeol anthus b i fo rmifolius Sampl es from a specimen o f �· b i fo rmifolius growing at Mi ne de L'E toi le were collected by Pro f. F. Malai sse throughout the pl ant ' s three month growing season . These sampl es were anal ysed fo r coppe r by the method employed for fi eld survey sampl es in Chapter 2 . 4 . 3 COBALT AND COPPER UPTAKE The results of the seedling uptake trials are shown in figs . 4 . 1 for �· k atangense , 4 . 2 for �· robertii and 4 . 3 for �· b i formi fol ius . I t shoul d be noted that e ach point on the diagrams is a mean of values over four analyse s . No di fferences in the leaf concentrations , at each soil concentration , were found over the weeks that they were tested , indicating that the uptake of cobalt and copper had reached a limiting value before sampl ing b egan at fi ve week s in the tri al substrates. Only 8 8 �· roberti i survived in all concentrations. H . katangense failed to survive i n any 1 0 , 00 0 pg/g pot i rrespect ive of the metal present; 10 000�------------------------------� • C o p p e r • Coba l t • . 1 000 - CJ') ' CJ') ::L - ....._ 0 � __, c c 0 � 0 L � c � u c 0 u 1 00 • 1 0 • 70 1 1 00 • • 1 10 1 00 1 000 t Concan t ra t i on i n soi l ( J.J g /g ) F i g u re 4. 1 Accumu l a t ion c u r v e for Haumaniastrum kd tan g,ense _ s<2a d l i n g s . 1 0 000 10 000�----------------------------�� - J) ' :n � .... - j 'J) _, --- -- ) " ' - " -- � ) - ) ) 1 000 1 00 1 0 1 • C o p p e r • Coba l t • • • ' 1 5 0 1 20 0 1 0 100 1 000 Con c e n t ra t ion 1 n so i l ( IJ g I g ) £igure 4 . 2 Accumu la t i on c u r ve fo r Haumaniastrum rober fii sczczd l i n gs . 1 0 000 1 0 000 �------------------- • C o p p e r • C o b a l t 1 0 0 0 ::J') ..... ::J') � - - J � 1 00 - --- -- • :::> -- J - • - 1 0 -- � J • -- :::> • .) 70 ' 1 500 1 1 0 1 00 1 000 1 0 000 Con ce2n tra t ion i n so i l ( IJ9 I g ) .£jgu r e2 4 . 3 Accumu la t i on cu rve for A eo/an thus biformifolius s eed l i n g s . A . b i formi fol iu s fai l ed to survi ve i n the 1 0 , 000 �g/g cobalt and coba lt + copper pot s . S i gnifi cant quanti t ies o f the cobalt salts had been l eached from the 1 0 , 000 �g/g pot s . The l e aching o f copper from the 1 0 , 000 pg/g po t s w a s much l e s s . Some coba lt was also lost f rom the 1 , 000 �g/g pots but lower concentrations were not affected . Copper concentrations below 1 0 , 000 pg/g were not affected b y leachin g . H . roberti i survived i n 4 , 000 p g Co/g and 8 , 500 pg Cu/g 9 2 soi l concentration s . The corre spondi ng l eaf concentrations were 7 , 380 pg Co/g and 29 pg Cu/g respective l y . The best growth of the seedl ings o ccurred i n the pots with the hi ghest metal concentrations . There w a s a h eavy l o s s o f seedl ings a t low metal concentration s , parti cularly i n the control g roup . �· k atangense survived i n much lower concentrations with a maximum of 660 �g Co/g and 1 , 320 � g Cu/g . The corresponding leaf concentrations were 260 pg Co/g and 6 �g Cu/g ( the h i gher concentrations o n fi g 4 . 1 come from the tolerance tests ) . The best g rowth occurred a t the 1 00 �g/g metal concentrations . The control group specimens we re only s l i ghtly l e ss in growth . G rowth in 1 , 320 �g Cu/g was better than i n 660 �g Co/g wi th specimens i n the l atter showing s igns o f chloro s i s . Fo r the 620 �g Co/g + 1 , 240 �g Cu/g ( added 1 , 000 ug/g coba l t + copper ) only one s tunted specimen survived. A . b i formi fol ius survi ved i n 8 , 500 �g Cu/g but only 660 pg Co/g . The copper concentration i n t h e l eaves never exceeded 8 5 �g/g compared to the very h i gh values found in wi l d specimens at Mine de L ' Etoi l e (Mal ai sse & Grego i re , 1 978 ) . The h i ghest coba lt concentration in l e af material w a s 2 , 000 pg/g . The b est g rowth was at 8 , 500 pg Cu/g . At this concentration the pl ants were darker i n colour a nd more steady i n g rowth . Flowering was f i r st observed i n a specimen at this l evel . Flowering was slower at lower copper concentrations and no flowering was observed i n specimens on coba lt-enri ched substrate s . 9 3 The accumulation curves ( figs . 4 . 1 , 4 . 2 & 4 . 3 ) do have some characteri sti cs i n common . All show discontinuities . This gives the accumul ation curves the exclusion-breakdown form discussed in the i ntroducti on to thi s chapter . For cobalt , the discontinuiti es are at 1 50 pg/g for H. roberti i and 70 �g/g for �· katangense and A. b i formi folius . The corresponding copper values are 1 , 200 �g/g for �· roberti i , 1 , 1 00 �g/g for �- katangense and 1 , 500 pg/g for A. biformi folius . The most immedi ate observation is the much lower soi l concentrations at which cobalt exclusion breaks down i n compari son to copper exclusion . Thi s i s not surpri sing since pl ants keep a ti ghter control over essenti al el ements ( eg . copper) compared to non-essenti al el ements ( e g . cob alt) . A further piece of evidence for ti ghter control of the essenti al el ement i s the greater gradient of the pre-breakdown curves for cobalt indi cating a less pronounced exclusion mechanism for thi s element. In the wil d , all three species are confined to mineral i zed soils although the pot trials reported here show that all three can exist on substrates with only traces of cobalt or copper. Losses at lower concentrati ons were high howeve r , w ith fungal attacks the predomi nant apparent cause . When one cons iders the tox ic nature of the natural soi l s of these spec ies , particul arly the high copper content which inhib its fungal growth , it is not surprising that the plants have a low natural resi stance to fungal attacks . �· katangense , the least tolerant of the three species , did however have a greater resi stance to the fungi . The results of the uptake of copper in �· b iformi folius were surpri sing . The copper content of leaves in the wild contai n approx . 2 , 000 �g/g plu s , yet the highest content recorded here was only 85 �g/g . It was consi dered that thi s might b e because the specimens here were grown from seed rather than from corms whi ch have an inbuilt store of coppe� whi ch might b e mobilized to give higher leaf concentrations. To test thi s theory corms were grown over a similar concentration range . The results of thi s tri al are shown in fi g . 4 . 4 . The results showed no significant di fference in copper levels to the seedling tri al , with a maximum leaf concentration of 1 51 pg/g at a soil concentration of 9 , 000 pg/g . 10 000----------------------------------� 0') ' 0') ::L - 1- 0 C:J - c c 0 .,_ 0 L... .,_ c � u c 0 ...) 1 000 1 00 1 0 1 • • • • • • • • ' 30 1 30 10 1 00 1 000 Concen t ra t ion in so i l ( IJ 9 /g ) £lgurC2 4 . 4 A ccu mu la t ion c u rve for • C o p p e r • Cobalt 1 0 000 Aeolanthus biformifolius grown f rom corm s . 9 5 This di fference i n b e haviour between the w i ld and cultivated samp l e s must therefore b e d ue to other reasons . W i th the consi derable di fferences in soil s , this i s perhaps understandab l e but a t thi s time the exact reason ( s ) a r e unknown . I t i s no teworthy that the point o f the di scont i nui ties for both cob a l t and , more parti cularl y , copper a r e at much reduced soi l concentrations for the corms when compared to those of the seedl i ng s . It appears that corms h ave a reduced abi l i ty to exclude these metal s . The reason for the di fference i n b ehaviour i s unknown b u t obviously worthy of further study . The g rowth of all the corm specimens was simi l a r al though the specimen at 9 , 000 pg/g was small e r than the others. 4 . 4 TOLERANCE TESTS Only two species were tested fo r tol e rance : H . roberti i was omi tted when the mortal i ty rate among the seedl ings was found t o b e too h i gh t o allow any meani ngful resul ts t o b e obtained . The tole rance l evel s ( g reater than 50% survi val ) for �· k atangense were 300 pg/g for cobalt and 1 , 380 �g/g for copper . A s the cobalt content o f the soi l s used i n the copper tests averaged 1 2 �g/g and the copper content of the soi l s in the cob alt test averaged 60 p g/g , no interference effects were l ikely . At the tolerance l evel s ! the l e aves had concentrations of 259 �g Co/g and 201 pg Cu/g respectivel y . The data fo r soi l and l eaf concen­ trations h ave been i ncluded in fi g . 4 . 1 . One i ndi vi dual specimen survived in a cob a l t concentration above the tol erance level , at 3?0 �g/g w i th a leaf concentration of 2 , 400 �g/g . The copper tol erance l evel was sharply defined w i th death occurring rapi dl y above thi s l evel whi l e the pl ants at 1 , 380 pg/g were growing wel l . The cobal t l evel was much l ess sharply defined. The specimens at the two concentrations immedi atel y above the tolerance l evel ( 3?0 and 800 �g/g ) were very slow in dying ( the survi vo r excepte d ) . The tolerance level s fo r A . b i formi fol i us were 6 1 0 p g/g for coba l t and 8?0 �g/g for copper . I B ackground levels o f cobal t in the copper-test soil s and copper in the cobal t-test soi l s were the same as for �· katangense ( 1 2 �g Co/g and 60 pg Cu/g respectivel y ) . Unlike �· katangense , A . b i formi folius had several indivi dual s survi ving higher level s o f both metal s : 1 , 670 and 9 , 740 �g/g for cobalt and 2 , 460 , 7 , ?80 and 1 6 , 700 pg/g for copper . A further di fference between the two species i s that for �· b i formi fol ius , cobal t appeared to b e more toxic than copper . While the tol erance l evels of both species i s o f the same order of magni tu de , the tolerance of indivi du al specimens of �· b i formi fol ius i s such that this species appears capable of surviving in more toxi c substrates than �· katangense . 4 . 5 GERMINATION TESTS Both H . robertii and �· b iformi fol ius were tested for the ir ab i l i ty to germinate in the presence o f cobal t and copper. The resul ts are recorded in tables 4 . 1 ( H . robertiD and 4 . 2 ( A . b i formi folius ) . In the �· rob erti i test , i t was observed that germination in control ( di st i lled water) did not di ffer from germination in the p resence of 0 . 1% cobal t o r 0 . 1% copper. Germi nation in 0 . 1% cob al t + copper was however lower. Germination can occur i n up to 1% cobalt , al though the percentage here i s low , but no germination occurred for copper concentrations greater than 0 .1% . Thi s di fference cou l d be used to reinfo rce the designation of the species as a " cobal t-flower " rather than a " copper-flower" ( Brook s , 1 977 ) . In the presence of cob al t , germination at a low percentage can occur wi th copper concentrations above 0 . 1%. The overall germi nation rate of �· robertii is low , being not more than 55% under any of the condi tions tested. �· b i formi fol ius is a very ready germinator , with up to 95% germination being observed . I n thi s tes t , the control germination level was most closely matched by the 0 . 1% cobal t + copper germination level , al though the 0 . 1% copper and 0 . 1% cobalt were only marginally l ower. Lower germination level s were recorded in higher concentrations of cobal t ( 0 . 5% , 1%) , copper ( 0 . 5%) and cobal t + copper ( 0 . 5%, 1%) . As for H. roberti i , 9 6 TABLE 4 . 1 Germination of Haumaniastrum roberti i seeds. Time (Weeks ) 0 . 5 1 1 . 5 2 2 . 5 3 Control 0 40 50 50 50 55 55 - --- 0 . 1 2 5 55 55 55 55 55 -- - �--- -�-- Cobalt 0 . 5 1 2 3 0 . 1 - - - - 1 0 20 5 - - 50 20 5 - - 5 5 20 10 - - 55 25 10 - - 55 2 5 1 0 - - 55 M Copper 0 . 5 1 2 3 - - - - - - - - - - - - - - - - - - - - - - - - - --1.....-�- MMetal concentrations i n test condi tions expressed as percent . �esults are expressed in percentage germi nation. TABLE 4 . 2 Germination o f Aeol anthus b i formifol ius seeds. M Time .Control Cob al t Copper (Weeks) 0 0. 1 0 . 5 1 2 3 0 . 1 0 . 5 1 2 3 0 . 5 4 5 35 - - - - 1 5 - - - - 1 75 65 5 - - - 7 5 5 - - - 1 . 5 90 75 10 5 - - 80 5 - - - 2 95 80 1 0 5 - - 80 5 - - - 2 . 5 95 85 1 0 5 - - 90 1 0 - - - 3 95 85 1 5 1 0 - - 90 1 5 - - - Cobalt + Coooer 0 . 1 - 2 5 3 5 3 5 35 35 �-- -- - 0 . 5 - 5 5 5 5 1 0 1 2 3 - - - - - - - - - - - - - - - - - - Cobalt + Copper 0 . 1 0 . 5 1 2 3 4 5 5 - - - 80 5 - - - 9 5 1 0 5 - - 95 1 0 5 - - 95 1 0 5 - - 95 1 5 5 - - MMetal concentrati ons in test cond i ti ons expressed as percent . Results are expressed in percentage germi nation. I I ! '-.() -.....! germination will occur at hi gher cobalt + copper than copper concentrations . The overall tol erance to the presence of copper duri ng germination is however h igher for �· b i fo rmifolius than 9 8 for H . robert i i whi l e the tolerance to cobal t at h igh concentrations may _Q� �?rgi nally l ower. The rate o f germination is rel atively rap id . There were few new germ inations after two weeks of the three week test. There are i ndi cations in the resul ts that the hi gher the metal concentration , t he longer t he germination t ime requ i red . The heal thiest seedl i ngs germinated were under the control condi tions. The seedl i ngs were both l arger and had a greater r oo t development . At the hi gher metal concentrations seedl i ngs were very small and root development was i nh ib i te d . I t must be doubtful whether many of the seedl i ngs that germ inated at h i gher metal concentrations woul d have survi ved to establ i sh themselves if in metal -rich soi l s . It should be noted here that plants in Shab a , and more particul arly the metallophytes studi ed here , begin growth wi th the onset o f the rainy season when the volume of rain i s such that the soi l -water concentrations o f metal s will be l owest . 4.5 COPPER ACCUMULATION BY AEOLANTHUS B IFDRM IFOL IUS �· b i fo rmifol ius was analysed for i t s copper content through / i ts three month growing season at Mine de 1 1 E to i le . The results o f these analyses are shown i n table 4 . 3 . It i s obvious that this species is a hyperaccumul ator of copper (Mal a isse & Gregoi re , 1 978) . I t was ob served that the concentration of copper i n the aeri al parts (basal l eave s , flowers and stems ) decreased during the g rowing season whereas the corm , whi ch developed during the g rowi ng season , i ncreased i ts copper content . The l evel reached by t h e corm a t the end of the g rowi ng season is the hi ghest plant copper concentration currently recorded . TABLE 4 . 3 Copper accumul ation i n Aeol anthus b i formifolius during the growing season . N Date B asal Leaves Flowers and Stems 7/1 2 , 600 3 , 500 2/2 2 , 1 50 2 , 1 50 24/3 " All results expressed as p g/g dry wei ght . 4 . 7 GENERAL DISCUSSION Corms 2 , 600 1 1 , 800 1 3 , 700 The simi l ari ties of the accumulation curves o f the three metall ophy tes s tu di e d , for both cobal t and copper , are notab le . Fu rthermore t h e level of the exclusion l imi t for copper i n these species (�. b i formifol ius corms excepted ) corresponds well with 9 9 the l evel s in other tolerant speci es ; Triodi a pungens R . B r . , Tephrosia s p . nov. , Eri achne mucronata R . Br . (Nicol l s � al . , 1965 ) , Lychnis alpina L . and Silene dioica ( L . ) Cl airv . (Crooks , 1 979 ) . I t remains to b e seen just how wi despread thi s limit ( 1 , 000 - 2 , 000 �g Cu/g soi l ) actuall y is . The study of Becium hombl ei with its rise-to-saturation form does appear to b e di fferent ( Reill y , 1 969 ) . However the uptake was studied at only low copper concentrations ( less than 140 pg/g ) . The possibi l i ty that the " ri se " i s the fill i ng of a requi rement not met until the copper content of the soil is approx . 70 pg/g cannot be di smi ssed. Had this speci es been studied at higher soil copper contents , mi ght i t not have moved upwards from the plateau as some exclusion mechani sm broke down? lVfASSEY UNIVERSITY LISRARY I t i s impossible to compare our tolerance test results with other publ ished data . Mos t tol erance tests have been done in solutions . As soi ls fix a certain quanti ty of the added 1 0 0 metal sal ts , i t i s impossible to make compari sons between tolerance l imi ts obtai ned by t he two methods . The only other soil tolerance test on a copper-tolerant species ( Mimulus gut tatus , Allen & Sheppard , 1 97 1 ) was to look a t indi ces of tolerance of several popul ations at a g iven soil copper content rather than an attempt to find a l imit of tolerance for the species . The germi nation tests on H . robertii and A . b i formi fol ius confi rm t he conclusions of most studies on other tole rant species. With the exception of Horscroft ( 1 96 1 ) for B ecium homblei , no studi es of species tolerant to heavy metals have shown any essentiality for that heavy metal before germination (or growth in general ) will occur (Howard-Will i ams , 1 970 , Allen & Sheppard , 1 97 1 , Shaw , 1 980 ) . The data of Horscroft ( 1 96 1 ) are in conflict with data on the same species given by Howard-Wi l l i ams ( 1 970 ) . From this discussion i t c an be seen that the character of tolerance to copper has many similarities despite the wide vari ety of plant species which have been studi ed . I t appears that there may b e a case for parall el adaptation to the toxi ci ty of copper contaminated soi ls among these di fferent plant species. The l ack of studies on o ther cobal t tolerant species make comparative di scussions for this el ement impossible . CHA P TE R 5 P hyt och e m i c a l S t u d i e s o n Som e M etal l ophy tes fro m S h a ba 5 . 1 INTRODUCTION The presence in cel l s of la rge amounts of what are normally trace elements generally resul ts in the ill-health or death of the plant ( Hunter & Vergnano , 1 953 ) . The existence of metallophytes whi ch can accumul ate these elements wi thout such effects raises questions as to the nature o f the excess amount wi thin the plant . The exi stence of free metal ions could b e expected to interfere severely with the physiological processes of the plant b ecause 1 0 2 of the fact that they can readil y compl ex with many organic functional groups ( alcohol i c , amino , carboxyl , disulph i de , imi dazal , phenol ic and sulphydryl groups - Gilbert , 1 95 1 , Mil l s , 1 954 , Timberlake , 1 959 , Enni s , 1 96 2 , Rasheed & See l y , 1 966 , Vallee & Ulmer , 1 972 ) . These functional groups are all important components of compounds within the pl ant ' s metabolic cycl es . Copper i s an e ssenti al element i n pl ants where it i s a n acti vator o f several enzymes (N ason , 1 958) . The requi rement for copper to ful fil! these functions does not however appear to exceed 20 �g/g ( Piper , 1 942 , B eeson et al . , 1 956 , Rasheed & Seel y , 1 966 ) . The abi l i ty of certain cuprophytes to accumul ate this element greatly exceeds this b asic requi rement o f most pl ants . Since excess copper i n "normal " pl ants leads to necrosi s , the healthiness of these .accumul ators requi res an expl anation . As the necrosis i n "normal " pl ants grown on coppe�-ri ch soi l s is due to the interference with , and breakdown of , physiolo gical functions within the pl ant , it is obvious that cuprophytes have developed method s o f avoiding such interference . The method most commonly invoke d to "remove" the copper from circul ation i s complexation (Mi l l s , 1 954 , Gilbert , 1 95 1 , Reill y , 1 969 ) . The evi dence for complexation i s certainly strong i n the studi es of the cuprophyte B ecium homblei (Reill y , 1 969 , 1 972 , Reilly et al . , 1 970) . Reilly ( 1 969 ) showed a strong correl ation between copper and total nitrogen i n leaf sample s which suggested that the copper might b e bound i n a proteinaceous compl ex . Furthermore the fact that almost 50% of the copper was soluble in a series of organic solvents ( dioxan , bu tanol , methanol ) is certainly consi stent with an org anic-copper compl ex . Reilly et al . ( 1 970) found that approx. 17% of the copper was bound with structural components of the 1 0 3 cell wall ( l i gnins , polysacchari des o r proteins ) . By a dialysis experiment , these workers found that 20-25% of the copper existed as ei ther free io_ns o r w_as loose! y complexed while a further 1 0- 1 5% was only l i ghtly complexed . U sing paper chromatography they showed that no free ions wera detectable and that at least two ami no acid or pepti de complexes were present . The number of copper-containing spots found was however dependent on the solvents used . Reil l y ( 1 972 ) · similarly found that the number of spots in chromatographic experiments vari ed wi th the solvents used when he compared copper-tolerant and non-tolerant specimens of Becium homblei . Stu dying the two Indigofera species , !• dyeri and !• seti flora , Ernst ( 1 972) showed b y a sequential solvent extraction process that the bulk of the copper was soluble in ei ther water or dilute hydrochlori c aci d . L i ttle copper was soluble in butanol , i n two reagents capable of exchange processes ( sodium chloride and citri c aci d ) or in the resi due remaining after the extraction series . He consi dered that the water-soluble copper was prob ably located in the cell vacuole but made no attempt to i denti fy the nature of the copper . The hydrochloric aci d-soluble fraction was considered as being exchangeable copper and most prob abl y located on the cell wall . Tbe small amount of copper in the residue was consi dered to be strongly bound to the cell wall . Studies of cobal t i n plants h ave yet to prove that it i s a n essential element for them all . Nason ( 1 958 ) consi ders cobalt to be an activator o f some enzymes . A need for cobalt i n nodul ated leguminous plants i s well known but several groups h ave consi dered that it may a lso be an essenti al el ement for other plants (Ahmed & Evans , 1 960 , Wilson & Nichol as , 1 967 ) . Certainly cobalt complexes have now been recorded in many plant species used b y man (Ballentine & Steven� , 1 95 1 , Bowen � al . , 1 96 2 , D ' Souza & Mi stry , 1 979 ) . The " normal " content of cobalt i n plants i s l ess than 1 pg/g so that the values recorded for many cob altophytes in Shaba easily exceed normal tox i c i t y l evel s . As for copp er , the exi stence o f compl exes has general ly b een cons i dered a s t h e method b y which these spec i e s avo i d the di sruption of the i r physiologi cal functions that free cob al tous ions would crea t e . No specifi c stu d i e s on cob al tophy tes h ave however been c arried o u t . Evidence from non-tol erant spec i e s has suggested that coba l t is compl exed with prote ins and pept ides ( B al l entine & Steven s , 1 95 1 , D ' Souza & Mi stry , 1 979 ) . The possib i l i ty that the coba l t was i n the 1 0 4 form o f vi tamin a 1 2 was di smi ssed on the evi dence avai lable to these researchers and al so to that avai l ab le to B owen et al . (1 962 ) . The wo rk presented i n this chapter i nvesti gates the di stribution of cob alt and copper b e tween the various cell components b y a sequential solvent extraction method. Thi s a l lows some specul ation o n the nature o f these e lements within the l eaves o f f ive metal lophyte s . Stu d i e s i n greater detail on the nature o f cobalt i n Hauman ia strum robert i i are al so repo rted . These s tud ies involved proton probe anal yses fo r e lemental di stri but i on within the l eaf and i sol ation of the water-solubl e cob al t fo r tests t o determine the complexing agent . The nature o f the aci d-solubl e cobalt is al so sub jected to specul at ion . 5 . 2 EXPER IMENTAL METHODS L eaf material of five metallophytes ( Aeol anthu s b i formifol ius , Buchnera metal l o rum Duvi gn. & Van B a ck . , Faroa chalcophi l a , Hauman i a s trum roberti i and S i l ene cobal ti col a ) were coll ected a t various Shaban local i ti e s b y Prof . F . Mal a i s se o f t h e Universi te Nat ionals du Z al re a t Lubumb ashi . The material was freeze-dri ed a nd freighted t o N ew Zeal and . Upon arri val , all t h e material was pl aced in a refrigerator until used . 5 . 2 . 1 S equential Solvent Extraction A sequential solvent extract ion method was used to i nvesti gate the nature o f cob a l t and copper i n the f ive metallophytes . The sequence u sed was developed b y Lee ( 1 977 ) from the method of B owen e t al . ( 1 962 ) . The sequence b egan by macerat ing 2g of freeze-dr i ed leaf mate�ial with 1 Dcm3 o f 9 5% ethanol i n an h omogen i z er fo r five m inutes . The resul tant slurry was centr ifuged and the supernatant decanted . The res i due was washed with two further portions of ethanol wh i ch were similarly centri fuged and decanted . The three ethanol supernatants were fil tered , comb ined and l abell ed fraction A. The resi du e was further extracted w i th two 1 Dcm3 portions of deioni zed water in the homogeni z e r . The supernatants , after centr i fuging , were fi l tered and comb i ne d . The resul tant solution was l abe l led fraction B . The res i due was t hen simil arly extracted with three 5cm3 portions of D . 2M hydrochlor i c acid ( prepared from redi sti l l e d constant-bo i l i ng hydrochlor i c ac id ) . The comb ined supernatants were f il tered and then treated w ith an equal volume o f acetone to preci pitate the proteins and pectates . These precipi tates were labelled frac t i on D . The supernatant after acetone-prec ip i tat ion was labelled fraction C . The resi du e from the hydrochl or i c a c id treatment was then treated with three 5cm3 po rtions o f D . SM perchloric ac i d at 80°C . These d igests were al so centri fuged and the supernatants comb ine d . The supernatant was again treated w i th an equal volume of acetone to preci p i t ate the nucl e i c ac i ds . Th i s prec ip itate was l abelled fraction F and the remaining supernatant fract ion E . The resi due from the perchlor i c a c i d treatmen t wa s final l y d ige sted for t en minutes with boil i ng 2M sodium hydrox i d e . The supernatant col l ected from t h i s was l abe l led frac t i on G whi l e the residu e was l abe l led fraction H . All cobal t , copper and manganese determi nations were made 1 0 5 by atom i c ab sorption spec trophotometry . The instrument used was the V ari an-Techtron model AA5 wi th automatic back g round correct ion a s in Chapter 2 . The l i ne s used were 24D . Bnm and 304 . 4nm fo r cobal t , 324 . Bnm and 2 1 8 . 2nm for copper and 279 . 5nm for manganes e . Where the fractions consi sted o f aqueous solution s they were anal ysed d irectl y . The res i du e , prec i pi tates and organic solutions were treated a s fol l ows : they · were taken to dryness , ashed a t 500°C in a muffl e furnace a nd then redi ssolved in 2M hydrochl o r i c a c id before anal y si s . 5 . 2 . 2 Proton M i croprob e Analys i s T h e proton m i croprobe anal yses were p erformed a t Harvard University and the M I T L i ncoln L aboratory at Camb ri dge , Massachusetts , USA . The proton mi croprob e operates o n a principl e simi l a r to the el ectron mi croprobe ( Horowi tz & Grodz i n s , 1 97 5 , Horow i t z et al . , 1 976 ) . Thu s a s ample i s bombarded b y a m i crobeam of protons whi ch causes exci tation o f the nucl ei within the s ampl e . These nuclei emi t characteri s t i c X-rays as they return to the unexci ted state. The X-rays are then anal ysed to determine the compo si tion o f the sampl e . The proton m icroprobe h a s two important advantages over the e lectron mic roprobe : s en s i t i vi t y i s much improved b e i ng o f the o rder o f 1 - 1 0 �g/g compared to appro x . 1 , 000 pg/g ; and measu rement s do not need to be made i n a vacuum . The i nstrument used here uti l i z ed a 2MeV emergent proton beam from a Van de Graaf accel erator . The samples were mounted and then s canned past the fixed m icrobeam b y a stepping-motor driven XV stag e . I n all cases the characteri s t i c X-rays p roduced b y the el ements w ith in the top appro x . 20 �m of the sampl e were detected b y a Si (L i ) detector and sorted according to pul se h e i ght ( energy ) to determine the compos i t i on . Both the sampl e motion and the data collection were contro l l e d b y mini computer ( A ronson & Horow i t z , 1 980 ) . Onl y �· robert i i was sub jected to proton probe anal ysi s . Two sets of i nformation were gathered : two-dimensional photograph i c scans and one-dimensi onal l i ne scan s . The two­ dimensional scans were done on five-ten e lement s simul taneousl y at l ower resolution ( approx . 1 50 �m ) forming images o f 1 0 , 000 p i xe l s to show the e lemental di stributions . From these mi crograph s , areas of i nterest ( h igh cob alt content ) were i nvesti gated more clo sel y by use of higher resol ution ( appro x . 1 00 �m ) l i ne scans . These scans were sensi t ive to a l l e lements under i nves t i gat ion simultaneous l y . 5 . 2 . 3 Cobal t Complexes in Haumani astrum robert i i ( a ) A c i d-solubl e cobal t . An at tempt was made t o i dent i fy , speculative l y , the nature o f the aci d-soluble cobal t found during 1 0 6 the sol vent extract i on s equence . A further 2g o f l ea f materi a l was t reated a s for the solvent extraction sequence unti l fraction C was ob tained . Thi s was anal ysed for oxalate by a method sim i l ar to that used b y Bornk amm ( 1 965 ) and Mathys ( 1 977 ) . The solution , fraction C , was neu tral i z ed b y the a dd it i on o f a 1 0 7 few drops o f 2M sodium hydrox ide . Thi s prec i pitated the oxalate whi ch was coll ected by fil tering through a Whatman 542 f i l ter paper . The precipitate was washed with warm deioni z ed water b e fo re b eing redi ssolved in a m inimum quant i ty of concentrated sul phuric a c i d . T h e redi ssolved oxal ate solution w a s d i l u t e d until t h e a c i d strength w a s approx . 1 M . Thi s solu t i on w a s then t i trated a t 60°C with 0 . 02M potassium permanganate solution . ( b ) Water- so l ubl e cob a l t . T h e water-solub l e cob al t complex was separat ed with the i ntention of i dentifying the l i gands . The method o f separation is simi l a r to that u sed by Lee et �· ( 1 977b ) in the separa tion o f water-solubl e n i ck el compl exes from n i ckel hyperaccumul ators . Extraction o f the comp l ex requ i red homog eni zation o f 25g o f freez e-dried l eaf materi al , containing 0 . 4 5% cobal t , wi th two 1 00cm3 portions o f deionized water . These extracts were f i l tered and comb i ne d . To remove l i pids , prote ins and other l arge organ i c molecu l e s , t h e fil trate wa s washed w ith a 1 0 : 1 chloro form n-butanol mixture until no fu rther preci pitation occurred at the solvent i nterface . The prec ip i t ated material was anal ysed but contained a negl i gible amount of cob al t . After refi l tering 3 the aqueous solut i on , i t was reduced in volume to approx . 20 cm • Thi s was run in 2 cm3 portion s , through a 30 cm x 2 cm Sephadex G-75 gel f i l tration column . The elu tant was col l e cted in 5 cm3 fractions and the cob al t fract ions comb i ne d . The presence o f coba l t was determined b y atomi c ab sorption spectropho tometry . The recomb i ned cob alt fractions were then recycled throu gh the co lumn . After recomb ining the cob al t fract ions from the second fi l trat ion , the s ampl e was run through a 60cm x 1 cm Sephadex G- 10 column but a s no retardation was found , thi s was unsat i s factory a s a means of separating the cobal t complex from other substances and was d i sconti nu ed . At thi s point the cob alt had not b een greatly concentrate d . Further cleaning was i n i t i a l l y tried with ion-exchange resins but those teste d , Dowex 1-XB in hydrox i d e , carbonate or acetate f o rm , Sephadex CM-25 in ammonium form , and Amberl i te IR-45 i n hydroxi de fo rm , a l l fa i l e d to exchange w i th the complex . A precipi tation cleans ing process was then t r ied . Beg inning wi th a 50 : 50 methanol : water sol u ti on (made b y adding an equal volume of methanol to the sol u t i on conta in ing the cobal t complex ) and i ncreasing the methanol content in 1 0% steps ( ie . 60 : 40 and 70 : 30 ) much preci p i t a t i on occurred . For each pa ir o f prec ip i tat e and sol ution , the cobal t was traced by atomi c ab sorpti on spectrophotometry . Prio r to th i s trac ing anal ysi s , the f il trate was reduced i n vol ume to remove the methanol and the preci pi tate was redi ssol ved in deionized water. The cob alt compl ex b egan precipi tating in the 70 : 30 methanol : water sol u tion . Both the prec ipi tate and the supernatant we re taken to dryness by slow evaporat ion . This precipi tate was anal ysed for i ts cob al t content and u sed for mi croanalysis and el ectropho resi s . Mi croanalyti cal anal yses for carbon , hydrogen and n i t rogen were performgd by the m i croanal ytical chemi stry l abo ratory at O tago Univers ity . O xygen cou l d no t b e determined because of the presence of cob al t i n the compl ex . El ectrophoresis was performed on a water-cooled S avant-type apparatus in whi ch two buffer reservo i rs are separated b y a low fl ash point petrol eum spi r i t . The bu ffer u sed was 500 : 20 : 4 , 500 pyri d ine : aceti c aci d : water at pH6 . 5 . The operat ion was done at 3kV whi ch gave a current of approx . 4DmA . The average t ime 1 0 8 l ength of a run was one hour . T he paper u sed was Whatman No . 1 chromatographi c paper. The electropho retogram was anal ysed for cobal t b y d ivi ding the column o f paper which contained the o r i g i nal " spot '' o f the complex i nto 1 cm x 1 cm squares , ashing the squares at 500°C i n a muffl e furnace , redi ssolving the ash in 2M hydrochl or ic aci d and determining t he presence of cob a l t b y atomic ab sorption spectrophotometry. For preparat ive el ectrophoreti c runs , the origin was a l ine across the centre o f the paper rather than a " spot " on th i s l i ne . After t h e cobalt h a d been l ocated b y u s i n g the s t r i p analys i s , bands o f paper conta ining the cobalt complexes were soak ed in water overnight to l each out the compl exes . The resu l tant solutions were f i l tered to remove the paper and then dri ed to obtain the prec ip i tates which were used i n both HPL C and GLC . H i g h performance l i quid chroma tography ( HPLC ) was performed on a Waters Associates model 660 sol vent prog rammer coupled to a p-bondpak C 1 8 column . The buffer used was 2mM t etra-n-butyl ammonium phosphate at pH 3 . 2 . The flow rate used was 1 . 5 cm 3 /min and the detecto r , a Cecil CE 2 1 2 A spectrometer , was set at 220nm since most sub stances ab sorb around thi s wavel ength . The ana lys is was done b y redi sso l ving 1 0mg of the cobalt prec ip i t ate s , obtained b y prepa rati ve el ectrophoresi s , i n 0 . 5cm 3 o f water and d irectly i njecting 25 � 1 o f thi s solution i nto the system . Gas-l i qu i d chromatography ( GLC ) was performed by a Pye model 1 04 chromatograph equ ipped w i th a 2 . 8m x 4mm column pack ed with 3% SP2340 l iqui d phase on a supel coport , 1 00-220 mesh , PB 34 suppor t . The col umn was operated at 1 80°C w i th dry n itro gen , at a flow rate of 30 cm 3 /mi n , as the carr ier gas . The detector was a fl ame-ioni zat i on detector w i th an a i r-hydrogen fl ame . The fl ow rates for these gases were 350 cm 3 /min and 30 cm 3 /mi n respectivel y . The cobal t fract i ons from the preparative e l ectrophoresi s were derivati zed by di azomethane to produce the methyl esters of the compl exing agents. The d iazomethane was p repared b y the method of Werner ( 1 9 1 9 ) . The cobal t comp l exes ( 1 0mg ) were destroyed by the addi t ion of d i l u te hydrochloric aci d and dri ed on a water-bath under a stream of n i trogen . The precipi tate was then d i ssolved i n 0 . 5cm 3 o f ether and the d iazomethane added . Further addi tions o f di azomethane 1 0 9 were made until effervescence stopped . The ether was then evaporated b y flushing the vi al conta i ni ng the sample wi th n i trogen . The precipi tate which remained ·was redi ssolved i n 0 . 5cm 3 of red i st i l led chl orofo rm and anal ysed i n 2 pl sample s . 5 . 3 D ISTR I BU TION O F HEAVY METALS I N PLANT TISSUE EXTRACTS The d i s t ribution of cob al t and copper in the various fractions o f pl ant t i ssue extracts is shown i n t ab l e 5 . 1 . I t can b e s een that most o f t he heavy metal s are found in fractions B , C and E . The se fractions contain predominantl y polar compounds . Fract ion B conta ins the water-solub l e , low mol ecu lar weight polar compounds eg . o rg an i c ac ids and i no rgan i c sal t s . For a l l the spec i e s anal y s e d , the cobal t percentage extraction exceeded that o f the copper . Thi s is most part i cularly evident in H. rob ert i i and A. b i fo rmi fol ius . The percentage extracted of these metal s is much l ower t han that found for n i ckel in n ick el hyperaccumul ators . Kelly et �· ( 1 97 5 ) found 65-94% of the 1 1 0 n i ckel to b e extractab l e whi l e i n these studi e s the maximum cob alt extraction was 42 . 9% ( H . robert i i ) and the corresponding copper extraction was 22 . 2% ( S . cobal t i col a ) . F raction C contains aci d-solub l e pol ar compounds ( eg . s a lts of o rg ani c a c i d s , pho sphate s , carbonate s ) and i ons released b y exchange processes wi th the cel l wal l s and some prote ins . W i th the exception o f cob al t in A . b i form i fo l i u s and �· roberti i , t he percen tage o f extracted metal s exceeded that of fraction B . Thi s increased percentage extraction i s consi derab ly more noticeab l e for copper than for cobal t . The max imum cob al t extraction was 5 3 . 4% ( S . cobal t i col a ) and the copper extraction was 59 . 3% ( B . metal l o rum ) . Fraction E contains most remaining pol ar compounds and some structural groups eg. some cel lul o se s and l i gni n s . The copper percentage extraction i n thi s fraction always exceeded the percentage ext racted i n fraction B but never that of fraction C . The cobal t percentage extract ion varies much more consi derab l y but only in g. metallorum i s the percentage extracted here the h ighest i n any fracti on for a g i ven speci e s . Maximum extractions were 39 . 6% for cob a l t ( B . metal lorum ) and 26 . 0% for copper ( A . b i form i fol iu s ) . TABLE 5 . 1 The fractionation o f coba lt and copper i n plant ti ssue extracts . Aeol anthu s b i formi fol i u s B u c h e ra m e t a ll orum F a r o a cha l cophi l a Haumani astrum rob ert i i Si l e ne c o b a l t i c o l a Co Cu Co Cu Co Cu Co Cu Mn Co Cu Total Concentrations 2380 3920 1 5 1 0 3520 1 34 700 4690 489 1 98 233 33 (pg/g ) % i n fractions A 0 . 5 0 . 8 0 . 1 0 . 1 1 . 2 1 . 4 0 . 2 1 . 7 0 . 8 0 . 5 2 . 5 8 33 . 8 1 1 . 3 1 0 . 3 9 . 1 2 0 . 4 1 6 . 6 4 2 . 9 1 2 . 6 2 1 . 7 34 . 9 22 . 2 c 3 1 . 7 47 . 8 36 . 3 59 . 3 4 8 . 9 54 . 6 39 . 7 39 . 9 58. 7 5 3 . 4 38 . 8 D 1 . 8 4 . 2 1 . 5 4 . 0 0 . 9 0 . 8 1 . 6 3 . 0 1 . 8 1 . 2 1 . 9 E 22 . 0 26 . 0 39 . fi 20 . 7 2 1 . 3 1 7 . 4 9 . 9 1 8 . 3 1 1 . 6 9 . 0 2 3 . 2 F 0 . 9 1 . 2 0 . 7 0 . 6 0 . 2 0 . 2 0 . 7 2 . 5 0 . 8 0 . 1 0 . 5 G 8 . 5 1 . 8 7 . 0 2 . 0 4 . 4 3 . 6 4 . 5 7 . 1 4 . 0 0 . 7 5 . 4 H 0 . 9 6 . 9 4 . 7 4 . 3 2 . 7 5 . 3 0 . 5 1 4 . 9 0 . 8 0 . 4 5 . 6 A - Neutral small molecules i ncluding amino acids and p igments B - Water-soluble l ow molecular weight pol ar compounds . I I C - Aci d-soluble pol ar compounds and exchangeable i ons D - Proteins and pectates I E - Polar compounds and some structural groups F - Nucleic aci ds G - Remaining proteins and polysaccharides H - Cellulose , l ign i n and immobi l e fracti ons o f cell wal l s . . __.. - - TABLE 5 . 2 Spec ies I I I I I I I V V Stat i s t i cal data for cob al t and copper associations in pl ant fract ions . Intraspecies Assoc i ation Code Spec i e s Variab l e s A . b i formi fol iu s Co vs Cu B . metallorum Co vs Cu - F . chal cooh i l a Co vs Cu H . rob erti i Co vs Cu Co vs Mn s . cobal ti col a Co vs Cu Interspecies A s soc i a t i ons r S i gn i ficance 0 .75 s 0 . 82 s 0 . 99 sxx 0 . 69 s 0 . 85 s� 0 . 9 1 sx ( values of r wi t h s i gnifi cances in parentheses ) I I I . I I I I V I I I I I I I V Co - I I I I I I V 0 . 73 ( S ) 0 . 87 ( Sx ) 0 . 94 ( Sx x ) - 0 . 84 ( Sx ) 0 . 52 ( NS ) - - 0 . 83 (S x ) - - - Cu - I I I I I I V - 0 . 98 ( Sxx ) 0 . 97 ( Sxx ) 0 . 96 ( Sxx ) - - 0 . 98 ( Sxx ) 0 . 95 ( Sxx ) - - 0 . 97 ( Sxx ) - - - Sxx - very h i ghl y signi fican� ( P < 0 . 00 1 ) Sx - highly si gni ficant ( 0 . 00 1 � P < 0 . 0 1 ) S - signifi cant ( 0 . 0 1 � P < 0 . 05 ) NS - not si gnificant ( P # 0 . 0 5 ) V 0 . 89 ( S x ) 0 . 62 ( NS ) 0 . 93 ( Sxx ) 0 . 96 ( Sxx ) V 0 . 94 ( Sxx ) 0 . 9 1 ( Sx ) 0 . 95 (Sxx ) 0 . 92 ( Sx ) 1 1 2 In contrast to the Becium homb l e i studi e s o f Re i l l y ( 1 969 ) , very l i tt l e o f the heavy metal s i n the five metal l ophytes were found in the fracti ons conta in ing amino aci ds ( fracti on A) or prote ins ( fractions D and G ) . General l y ( th e exception b e ing �· b i form i fo l i u s ) even l ess metal was associated w i th the nucl e i c aci d s ( fract ion F ) . The percentage remaini ng i n the resi due ( fracti on H - cel l u l o s e s , l i gn ins and cel l wal l s ) was genera l l y the fourth h i ghest for copper , after the three polar compound fract i ons . Thus i t may be that a s i gn if icant portion o f the copper is bound to the cel l wal l s . The percentage i n the res idue ranged from 4 . 3% (�. metal l o rum) to 1 4 . 9% (�. roberti i ) . Re i l l y e t a l . ( 1 970 ) found 1 7% o f the copper in �· homb l e i was a ssoci ated w i th the cel l wal l . The percentage of cob alt in the resi due was not general l y s i gn if icant and a greater amount was t o be found with the prote ins and pol y saccharides of fracti on G . 1 1 3 I t i s obvi ous that the cobal t and copper in these p lant s i s general l y comp l exed t o organ ic l i gands t o fo rm polar compounds . The di s tribut i on of the metal s among the various fracti ons wou l d i nd i cate that more than o n e l i gand i s invo l ve d . Th i s contrasts w i th most n i ck e l hyperaccumu l ators which have the nickel bound to only a few simp l e o rganic acids (Lee et al . , 1 977b , L ee et �. , 1 978 , Kersten , 1 979 ) . B ecause v i t amin B 1 2 i s a cobal t complex , the presence o f coba l t i n a pl ant has often ra ised the question a s to whether or not the p lant is produc ing thi s vi tami n . No evi dence has ever b een found to suppo rt vi tamin production and most avai l ab l e evi dence di scounts it (Bowen et a l . , 1 962 ) . A s v i tamin B 1 2 i s readi l y solub l e in 95% ethanol ( fracti on A solven t ) and very l i tt l e cobalt i s solub l e i n thi s so l ven t , i t i s h i gh ly improbab l e that the cobal t forms any such v i tam i n complex i n these spec i e s . Non-accumul ating p l ant speci e s e g . tomato ( B owen et � . , 1 96 2 ) and red k i dney beans ( D ' Souza & M i stry , 1 979 ) have very di fferent percentage d i stributions b etween the various solvent fractions . Fract i on A had general l y o f t he order o f 30-60% of the cob alt whi l e o n ly sl i ghtly less was extracted b y 0 . 2M hydrochlori c aci d (Note : as n e i ther Bowen e t al . , 1 962 , nor D 1 Souza & Mi stry , 1 979 , d i d aqueous extract i ons , the i r hydrochloric aci d fraction i s the equivalent o f fraction B + frac t i on C ) . The d ifference b etween the copper di stributions is no t so marked al though at 9% the fraction A percentage fo r t omato ( B owen et �. , 1 962 ) i s s i gn i fi cantl y h i gher than our metallophyte percentages . Statisti cal anal y s i s of t he data i n tabl e 5 . 1 i s shown i n t ab l e 5 . 2 . For cob a l t and copper w ith in a s ingle spec ies ( intraspecies correlation s ) , there a re signi ficant t o very h ighly s i gn if icant correl ations for all spec i e s . Thi s wou ld appear to i nd icate a parall el uptake for the two elements wi thin a speci e s . For H . roberti i where manganese d i stribution was a l so stu di e d , a h i gh l y s i gnifi cant correl ation was al so ob served w i th cob alt . There wou ld appear to be parallel uptake o f all three e lements in thi s speci es . The i nterspecies correl ations ( correl ations between the same el ement i n d i fferent spec i e s ) are al so general l y s ignifican t . Certainly t h e copper d i stribution b etween the five speci es shows an extraordi nary sim i lar i ty (r > 0 . 9 for all correlations ) . The d i s tribution o f cob alt i s not so uni form however . Despite thi s , most of the correl ations are h i gh l y or very h i gh l y s ign i fican t . B . metal lorum does however a ppear to di ffer from 1 1 4 the other four speci e s in cob a l t di stribut ion . Both non­ s i gn i ficant and the on l y s ignificant co rrel ation i nvolve thi s speci e s and even i t s h i ghly s i gn ifi cant correl ation w ith F . chalcoph i l a i s comparativel y low f o r thi s ser i e s . A s a footnote � i t i s recorded t h a t t h e specimen of g. metallorum used in these experiments was the first of thi s spec i e s t o show hyperaccumul ation of coba l t and copper. 5 . 4 PRO TON M I CROPROBE STUD IES O N HAUMANIASTRUM ROBERTI I The two-dimensi onal scans of a l eaf of H . robert i i revealed several regions in which the cob alt content was h i gher than the general concentration . Cal cium was al so el evated i n these regions but potassium was defi cient . The mi c rographs of these elemental scans are shown i n pl ate 5 . 1 . From these regions , one was chosen on which to carry out a more det a i l ed l ine scan . ····-········-···-· .. -····--················· 0 0 • • • :::�.::::::::::::::::::::::::: :::::::::::::::�:::: :i!:amf:i!J!!!!ii!!!!!!!Hi!!ii!ii:!::::::::::::::: ;; : . ::::........:::. .. ·····-:::::::::=::::::::::::�:: �: :::: : : : . ::: ::.-=::.... ···:r.::::u:::::::::::::::: : : :: : :::: :::-··� ···-=··:me:·::::::::: � : · = : ::::: . . : ::m:==� ... :!!!H::::H:::::::::. � 0 ::::::: . !E!g�gfi!!.g ·= f:::gggggggg�f��:f�HiE�: ·:::�iiiiii!ii::::. :i:::::!:ffi!!ii!!iiiii�� ii � � i� � � i!iiiii!:: . ................ . ••.••.••.•.•............................... . ............................................................. 0 .......................................... .................. 0 ·:uur:gg:=iiij!i££Hrnntiiiiii��i!i�!iiiiiii�i!ii!Eii . ·········"··········:·································· ::::::::::::::::::::.:::::::::::::::::::::::::::::::::: 0 .............................................. ........ 0 . ··:::::::::::::::::=::::::::::::::::::::::::::::::; · ••• • o •············································ · · •ooo o o o•••···-································· ooo o o o o o o••······································· oooo o o o o·-······································ · · ::::::mm:m.=m:==mmm::m:m:::m:: o o o o•••··········· ••••.•...... ........... . . . ............. . .................... .· ·······••··· .................••.. . ............ . .................•. . ........ . .. . ... ... . ........... . ........... . ........ .......... ............ . ................... .••••••••••••• • ••••••••••••••••••• 0 ........... . .................... .···· · · ······-- -······················ ooo ··· ·········--········· ··············· . . . . . . . .................................... 0 . .... . ...................................... oooo o ooooo•••····························· ·•••• • o ••••·······························o ·••••o••••······························•o ooooooo oooooooooooooo······················ � Cc /le,,.,,,..� f"*/.,..J;,. Pla te 5 . 1 Proton microprobe 1 micrographs 1 showing the I distri bution of lcoba l t , ca lc ium I and potassium in 1 a IQaf . The I I l i ght Qr reg ions iare t he reg ions 1of g rQa test ! concen t ra t io n . Ol c � 0 .r:. (/) ·-·- ....._ '- (1) ..Q 0 '- E ::J '- ....._ ll) tJ ·- c:: tJ E ::J tJ :r: "+- 0 (/) "+- c a 0 C:* .....,_ a '-(/) .....,_ (/) c 0 C:* '- u u c a 0 u c 0 -0 u .....,_ (/) c ' C:* a E c C:* ·- - _J . � N . lJ') � '- :::J Ol iLl 0 N 0 c 0 � u � (.) 0 e • D 0 0 0 l{) (/) c 0 L.. ........ '+-0 � 0 (/) 0 '+-0 � - (/) (/) 0 L.. u 0 � u c 0 ........ (/) · - 0 1 5 000 1 0 000 � 5 000 - 0 1 5 000 � 1 0 000 � 5 000 � 0 - 5 0 0 0 � 0 K ea �ea "' K ea n ea lJ\. ea K ea � � Mn � .J 2 56 Ar b i t rar y u n i t s Co eo eo eo -- eo n eo "-A ( a ) Spectrum throu9'1 a sec tion of lower cobalt background. 512 ( b) Spec trum throu g h a sec tion o f higher coba l t bac kgrou n d . ( c ) Spec t rum throu g h a co b a l t anomaly . Fi gu r e 5 . 3 X - ra y sp e c t r a f ro m p ro to n microp ro bQ . Thi s l i ne scan i s shown i n fi g . 5 . 2 fo r cobal t , c a lc ium , manganese and potassium . The d i stribu t i on o f cobal t , c a l c ium and manganese i s remarkably un ifo rm acro s s mo st o f the l eaf with a vari ation of no more than + 1 0% from the mean concentrat ion . Potassium shows a greater var i a t i on b e i ng a s much a s ± 20% from the mean concentrat i on . At the anomal y , the coba l t and manganese concentrat ions ri se sharpl y , in paral l el , whi l e the potas si um concentration drops sharpl y . C a lc ium r i ses i n concentration immedi ately adjacent t o the cob a l t r i se b u t dips bel ow the max imum concentration a t the centre o f the anomal y . The d i ameter o f th is anomaly i s approx . 1mm. A more d i rect compari son of the concentrat ions o f coba l t , manganese , c a l c ium and potassium b etween the cobal t anomal i es and other regions o f the l eaf can b e seen i n the spectra o f fi g . 5 . 3 . I t can b e ob served that manganese shows only i n the coba l t anomal y spectrum. Cob alt a l so shows mo st s trongly i n - t h i s spectrum but does appear i n the other spectra . Potassium and ca lc ium natural l y dom i nate the spectra but i t i s e a s i l y observed that a t the anomal y (wh i ch i s not the same anomaly as i n the l i ne scan ) the cal c i um concentration exceeds the potassium concentrati o n . Th i s i s the reverse of the normal s i tuat ion which i s observed for the back ground spectra . I t appears from these resu l ts that much o f the cobal t i s local i z e d w i th in the pl ant . O n e poss i b i l i t y for such l ocal i z at ion wou l d b e t he precip i tat ion o f cob alt i n some crystal form . The ri se i n cal c i um concentration in the s ame general area and the k nown occurrence o f c a l c ium oxal ate crystal s in p l ants ( Al -Ra i s et a l . , 1 97 1 ) r a ised the que s t i on a s to whether the coba l t cou l d b e eo-prec i pi tating a s a n oxal ate cry stal . Certainly cob a l t oxalate cryst a l s a r e i n solub l e i n aqueous medi a . 5 . 5 COBALT COMPLEXES I N HAUMANIASTRUM ROBERTI I Fol lowing the resu l t s o f the proton probe studi e s , i t was dec i ded to i nvest i g a te the rel a t ive amount s o f a c i d-extractab l e coba l t and oxal a te . Th i s w a s done by determi n ing thei r concentrati o n i n frac t i on C o f the sol vent extraction sequence . 1 1 9 The resu l t s showed that an al i quot o f fraction C contained 1 6 . 6 �moles o f cobal t and 1 8 . 3 �mol e s of oxal ate . It is thu s pos sible that the aci d-extractab l e cob alt cou l d be present in the l eaf as coba l t oxalate crystal s . The removal from the cytoplasm of 30-50% of the cobal t as oxalate crystal s wou l d obviously b e b enefi c i al t o the pl ant . I t wou l d a l so appear that some manganese occurs in these crystal s along with the cob a l t and calcium . The sub s t i tution o f other dival ent ions for cal c ium i n oxal ate crystal s extracted from pl ants has been shown by Al-Ra i s et al. ( 1 97 1 ) . The water-solubl e cabal t complex extracted and separated as in subsection 5 . 2 . 3 was found to contain onl y 1 . 1 4% cob a l t . Th i s wou l d g ive t h e compl ex a mol ecu l a r weight of 5 , 200g per mol e of cob a l t if pure . The m i croanal ys i s of the compl ex gave the resul t s 28 . 05% carbo n , 3 . 95% hydrogen and 1 . 04% n i trogen. Thi s g i ves the complex 1 20 moles o f carbon and 4 mol e s o f ni trogen per mole of cob al t . O n the bas i s o f th i s n i trogen content , i t i s obvious that the compl exing agent i s not protei naceous i n nature . The el ectrophore togram run on the comp lex revealed a rather complex pattern o f cobal t d i s tribution ( fi g . 5 . 4 ) . The dominant peak was an anioni c peak . The c at ion ic fraction appeared to contai n three overl apping peak s . To at tempt further i denti ficati on , several preparat ive-scale el ectrophoretograms were run and the cobal t soaked off a s two fract i ons , anionic and cationi c . These fractions were analysed b y HPLC . The anionic fract ion gave r i se to two peak s which d i d no t coincide with any o f the seven common organic aci ds run as standards ( aconi t i c , c i t ri c , i soc i tri c , mal i c , malonic , qu inic and tartar ic acids ) . T he cond i ti ons u sed i n HPLC spl i t up t he complex , rel easing free aquo-cob al t -( I I ) i ons and the compl exi ng agent ( s ) . Thi s a l l owed d i rect compari sons between the standards and the sampl e s . The cationi c fraction showed a very strong coba l t peak but n o o ther significant peak s which suggested t h a t much o f the coba l t i n thi s fraction might have been aquo-cobgl t-( I I ) i ons . 1 2 0 50 ' ' 40 Ol 30 c · - "0 0 � '- � u c 0 ..c � 0 (/) ..c <( 20 1 0 an 1on1c O R IG I N 0!. ft ft ft .{ .. i - �a o o o o o o o I 1 5 10 1 5 2 0 25 30 Fract ion numb e r £lgure 5 . 4 E le2ct rophore togram o f wa ter - so l u b le c o ba lt comp lex e x t ra c tc2 d from Haumanias trum rober tii. Fu rther samples o f these fract ions were methy l a ted for analysi s by GLC . The resu l t s fo r the anioni c fractinn were simi l ar to those fo r HPLC i e . there were two peak s wh i ch d i d not correspond wi th those o f the standards u sed ( aconi t i c , c i tri c , �-k etog l u tari c , mal i c , maloni c , oxal i c , oxaloaceti c , succinic and tartari c acids ) . The cat ionic fraction showed a singl e 1 2 2 peak whi ch al so d id not correspond wi th any o f the standards . I t must b e noted that a s the "mo l ecular wei gh t " o f the cobalt complex i s o f t he o rder of 5 , 200g per mol e o f cobal t , t he compl ex ing agent ( s ) may have been too l arge to b e sati sfactori l y identi fied by the methods or condi tions used. Furthermore i f the complexing agent i s small and much o f the mass i s in fact some contaminatio n , i t wi l l b e di ffi cu l t to decide whether any sub stance detected i s the comp l exing agent o r the contaminan t . Much further work wi l l b e requ i red b efore a n y i dent if ication o f t h e complex c a n b e made and thi s w i l l requi re the separation of a l arger quantity of the complex from the l ea f . PA R T T WO _A_Iy_ssum robertianum N i ckQ I Accumu la t i on by the Genus A!y_ssum CHAPTER 6 S u rvey of t he G enu s A.Jy_ssum 6 . 1 INTRODUCTION Alyssum L . is one of the l arger g enera in the fam i l y Cru c i ferae (= B rassi caceae ) . S chul z ( 1 936 ) d iv ided th is fami l y i nto n inetaen tr ibes one of whi ch i s named Alysseae after i ts pr inc ipal component genus , Alys sum . The fami l y , tribe and genus h ave been described a s I rano-Turanian i n phytogeographi cal o r i g in (Hedge , 1 976 ) . Alys sum is however a l so wi dely represented in the Mediterranean and Saharo­ S i n d i an reg i ons . Outs ide these reg ions , Alyssum speci e s can b e found i n a b e l t through S i ber ia to t h e Yukon di strict 1 2 5 i n North Ame r i c a . The genus h a s i t s greatest concentration , divers i f i cat ion and pro l i feration i n the eastern Mediterranean and Turk i sh areas ( Dudl ey , 1 964 a , b , 1 965 a , b , Persson , 1 97 1 ) . The Yukon representat ion consi sts of one spec i e s , �· americanum Greens , in a di sjunct popu l a t i on . Th i s spec i es close l y resemb l e s two S i b er i an speci e s , A . obovatum (Meyer) Turcz . and A. b i ovul atum Busch , and may b e conspec i f i c with them ( i n whi ch case A . obovatum would be the spec i f i c epi thet) (Dudl ey , 1 964 a ) . The di stri but i on o f a spec i e s through S ib er i a to North America has been recorded i n other ge nera of the Cruci ferae (Hedg e , 1 976 ) . The genus Alyssum con s i s t s of approx . 1 70 spe c i e s . Dudl ey ( 1 964a ) recogn i z es 1 68 spec i e s i n h i s revi s ion of the genus but ma inta i�s the possib i l i t y o f several more because he was unab l e to examine the type mater i al of some other previousl y-named spe c i e s . Th e genus has b een d ivi ded i nto various sections , sub sections and seri e s . There are s i x sections : ( 1 ) sect ion Meni ocus ( D esv . ) Hook . ; ( 2 ) section Ps i lonema (Meyer ) Hook ; ; ( 3 ) s ect ion A lyssum ; (4 ) section Gamosepalum ( H ausskn .) Dudley wi th the seri e s Connata Dudley and L ib era Dudl ey ; ( 5 ) sect i on Tetraden ia ( Spach . ) Dudley ; ( 6 ) section O dontarrhena (Meyer ) Koch w i th the sub sections Infl ata Dudl e y , Compressa Dudl ey compri s i ng the seri e s Integra Dudley and Crenulata Dudl ey , and Samari fera Dudl ey . The plants themselves are herbaceous in character and may b e annual s , b i ennials o r perenn i al s . Wi thin any g i ven sect ion , however , the l i fe period i s often more c l earl y defined : Meniocus and P s i l onema are composed who l l y o f annual s , A lyssum has all 1 2 6 three , G amosepalum and Tetradeni a perenn i al s only and D dontarrhena mostl y perennial s wi th , rarel y , b i enni a l s . T h e genus Alyssum holds a spe c i al place in the study of n i ckel hyperaccumul ation . The first three species i denti fied as hyperaccumul ators o f t h is e l ement were all from thi s genus �· b ertoloni i ( Mi ngu z z i & Vergnano , 1 948 ) , �· mu rale ( Dok sopul a , 1 96 1 ) and �· serpyl l i fo l i um ssp . lusi tani cum ( Menezes de Sequeira , 1 969 ) . The 1 . 22% ni ckel repo rted by Mingu z z i and Vergnano ( 1 948 ) i n dri ed l eaves of A . b ertol oni i was over an o rder o f magn i tu d e h i gher than that reported for o ther vegetation from the oph i ol i t i c outcrop at Impruneta , near Fi renze (Florence ) , I ta l y . Thi s value greatl y exceeded any o ther n i ckel concentration i n plant materi al recorded at that t ime . In 1 978 , B rooks and R adford ( 1 978a ) surveyed the n i ckel content of European species o f thi s genus and discovered el even new hyperaccumul ators ( see appendix I I ( a ) ) . Furthermore all el even speci e s were from section O dontarrhena . The three previously di scovered t axa are all al so from this section of the genus . Alys sum section Ddontarrhena was or ig inally described as a separate g enus , D dontarrhena Meyer ( Meye r , 1 83 1 ) . Thi s was b ecause i ts memb er species had several important and contrasting d i fferences compared with spec i e s of the genus Al yssum as i t w a s then del imited ( Dudley , 1 964 a ) . Thi s section has a l arge proportion o f speci es which are ecol og ical ly endemi c , and restricted solel y to serpenti n i t i c and other ul trabas ic sub strates . O f the fourteen known hyperaccumul ators o f n ickel witbin this secti on , el even are endemi c to such sub strate s , whi l e three speci e s , � · alpe stre , �· a rgenteum and �· obovatum a re not ( B rooks & R adford , 1 978a ) . Not onl y i s section D dontarrhena the sole section w i th nickel hyperaccumulators but the fourteen taxa with thi s property come from a survey of only twenty-three taxa recorded by B all and Dudl ey ( 1 964 ) in Flora Europaea . Thi s g ives a remark abl y h i g h proportion o f hyper­ accumul ators in t h i s section . A simi l ar proportion exi sts for speci e s of the genus Homal ium i n New Ca l edoni a (B rook s , L e e , e t al . , 1 977 ) . The s i tuation i n whi ch hyperaccumulation o f ni ckel i s restri cted t o a parti cul ar section h a s not b een recorded fo r any other l arge genus but i n Phyllanthus there are some sect i on s which have hyperaccumul ators and others whi ch do not ( Kersten e t �. , 1 979 ) . Speci e s o f other sections o f Alyssum showed no hype raccumulat ion even for spec i es growing on serpent in i t i c substrates ( B rook s & R adford , 1 978a ) . 1 2 7 Thus �· densi ste l l atum ( section Aly ssum ) growing over serpenti n e d i d not accumu l ate more than 4 0 � g N i /g , a concentration which is "normal " for any pl ant growing over such substrates . Desp i te the f act t hat serpent ini t i c sub strates a re al so rel a t i vely rich i n coba l t and chromi um , uptake of these two e l ements i s not notab l e . I ndeed B rook s and R adford ( 1 978a ) used the chromium content of specimens a s an i ndex o f pos s i b l e contamination b y so i l . Specimens w i th greater than 1 0 �g C r/g were a s sumed to b e contami nated and hence rejected. If the speci e s were c hromium accumul ators th i s woul d h ave l ed to an unacceptab l y h i gh rejection rate . I n a cont i nuat ion of t he su rvey o f th is genus , the non-European spec i e s , plus some fu rther European speci e s , were anal ysed for n i cke l , cob a l t a n d chromium . A s a b a si s for determin ing the speci e s t o b e covered b y t h i s survey , i t was decided t o cover on ly those spec i e s l i sted i n Dudley ' s synops i s o f the g�nus (Dudl ey , 1 964a ) . Thi s meant that the survey attempted to cover 1 68 spe c i e s o f Alyssum. 6 . 2 ANALYTICAL METHODS The survey for n i ck el hyperaccumulat i on in Alyssum speci e s was carri ed out on herb ar ium specimen s . Approaches were made to a number of h erb ari a and those l i sted i n appendi x I ( a ) provided spec imens . From the specimens supp l i e d , samples o f 0 . 0 1 - D . D3g dry we ight were w e i ghed and pl aced in 5 cm3 borosi l i cate t e st-tub e s . The s ampl es were then ashed at 500°C i n a muffle furnace . · The ash was then d i ssolved i n 1 cm3 o f 2M hydrochlori c aci d ( prepared from redi s t i l l e d cons tant-b o i l i ng hydroch l o r i c ac i d ) . T h e resul tant soluti on w a s anal ysed b y atomic ab so rption spectropho tometry fo r the el emen ts n i ckel , cobal t and ch romium . The atomic ab s orpti on spectropho tometer u sed was the Vari an-Techtron model AA5 w i t h au tomati c b ack g round c orrector as fo r Chapte r 2 . The l i ne s used for anal ysi s were 232 . Dnm and 35 1 . 5nm fo r n i ck el ( l ow and h i gh concentrations respect ivel y ) , 240 . 8nm f or cob alt and 357 . 9nm for chromium . The chromium content was determi ned as an indi cato r o f pos s ib le contamination al though h a d any spec i e s been rejected too f requent l y fu rthe r i nve s t i g ations o f the chromium contents o f that spe c i e s would have ensu e d . N o spec i es was however rejected wi th any g reat f requency . No spec i es in which i ndivi du al specimens were rej ected had the i r status changed b ecause of that rejection i e . al l spec i e s which showe d h i gh n i ckel contents in contam inated specimens also had at l ea st one specimen w i t h h i g h n i ck el content w i thout contaminat i o n . A s f o r B ro ok s and R adford ( 1 978a ) , a chromi um content o f 1 0 � g/g o r g reater was assumed t o be a resu l t o f contami nat ion and the sampl e was reject e d . The resu l t s o f a l l the anal y se s a r e repo rted on a d r y w e i gh t b a si s . 6 . 3 RESUL TS AND D I SCUSSION The resu l t s of n i ck el anal y ses of several hundred specimens of Alys sum spe c i e s are recorded i n tab l e 6 . 1 . 1 2 8 Thi s tabl e a lso includes data f or the fourteen hype raccumul ators d i scovered b y B rooks and Radford ( 1 978a ) . W i th these resu l t s i nclude d , 1 67 o f t h e 1 68 spe c i e s l i sted b y Dudl ey ( 1 964a ) have now b een su rveyed for thei r n i ckel content . Al though cobal t anal yses were al so performed , the data were never anoma lous ( the majo r i ty had concentrat i on s below 1 �g/g ) and have b een omi tted from the tabl e . The rel at ive l ack o f cob al t i n the serpent ine-g rowi ng specimens i nd icates that Alys sum speci e s preferenti al l y accumulate n i ckel rel at ive to cobal t . T a b l e 6 . 1 N i ckel concentrations i n Alyssum L . s pec i e s Species Section - Meniocus ( D esvaux ) Hooker A . aureum ( Fenzl . ) B o i ss i er A . b l epharocarpum T . R . Dudley & Hub er-Morath �. heterotri chum B o i ssi er �· huet i i B o i ss ier �· l i n i fol ium S tephan ex W i l l denow A. meni ocoides Bo i ssi er A . styl are ( B o i ssier & B al ansa ) B o i ssier mean n i ckel content fo r section Section Psil onema ( C . A . Mey e r ) Hook er �. alyssoides ( L i nnaeu s ) L i nnaeus A . d amascenum Bo i ssier & Gail l a rdot �· dasycarpum S tephan ex W i l l denow �· granatense Bo i ssier & Reu ter �· homalocarpum ( F . Fi scher & C . A . Meyer ) B o i ss ier mean n ick el content for section Section Al yssum B.· a i zo ides Bo i ss ier A . a renarium L o i sel eur-Deslongschamps �· argyrophyllum Scho tt & Kotschy A . armenum Bo i s s i er A . artwinense Busch A . a t l ant icum Desfonta ines A . auranti acum Bo iss ier A . bornmuell eri Haussknecht ex Degen A-. bulbotri chum Haussknecht & B o rnmul l e r N i ck el Content ( �g/g dry we ight ) <. 1 ' 3 8 , 9 6 , 1 2 4(.1 , <. 1 9 , 9 , 1 6 �1 ' 2 3 , 1 4 6 2 , 4 , 4 , 5 , 6 , �1 , 4 2 , 3 , 6 ..(1 , 2 , 2 , 4 , 6 2 , 7 3 2 , 1 0 .( 1 ' .(_ 1 , 2 , 4 � 1 ' 3 c(, 1 , 9 < 1 ' .::::.1 3 , 1 5 4( 1 ' 3 <1 , 10 5 , 5 , 9 1 2 9 A . caespi to sum J . B aumgartner �. calycocarpum Ruprecht �· canescens De Candol l e �· cephalotes Bo i ss i e r A . contemptum Schott & Kotschy A. cunei fol ium Tenore �· densi stel l a tum T . R . Dudley A . d esertorum Stapf A . d i f fusum Tenore A. doerfl eri D egen A. ero sulum Gennari & Pestal o z z a A . fastigi atum Heywood A. fi scheri anum De Cando l l e A . fol i osum Bory & Chaub ard A. ful vescens Sib thorp & Sm i th A . handel i i Hayek A . h i rsutum B i eberstein A . i da eum Boi ss ier & Heldreich A . i ranicum H au ssknecht ex J . B aumgartner A. l anceol atum J . B aumg artner A . l as s i t i cum Hal �csy A. l enense Adams A . l ep i dotum Bo i ss ier A . macrocal yx Co sson & Durand A. macropodon B o i s s i er & Bal ansa A . marginatum S teudel ex Bo i ss ier A . mi c rophyl lum ( C . A . Meye r ) Steudel A . minus ( L i nn aeus ) Rothmal er A . minutum Schl ectendal ex D e Candol l e A . moel l endo r�i anum A scherson ex Beck A . montanum L i nnaeus A . mouradicum Bo i s s i er & B alansa A . muell eri Bo i ssier & Buhse A . nevadense Wilmott ex P . B al l & T . R . Dudley A . ochrol eucum Boi ss ier & Hueter A . ovi rense Kerner 7 ' 10 <.1 ' 1 5 , 6 ..(1 ' 3 ...(1 , 3 .c 1 , < 1 , < 1 , 2 , 2 <1 , .(1 , 4 0 , 4 0 <.1 , 2 , 5 , 1 4 ..(1 , 3 2 .( 1 ' 2 3 �1 , ..(1 , < 1 , 7 .::..1 , .(.1 L..1 , <.1 < 1 ' 5 .c.1 , 2 , 3 , 9 .(1 , 2 , 3 2 ' 4 2 ' 3 3 ' 1 0 1 , 2 , 1 3 .( 1 ' 7 3 ' 5 2 ' 5 �1 ' 22 70 ' 1 52 <1 , <1 , 1 , 2 , 4 <1 , · <. 1 , 5 "- 1 , < 1 , 3 < 1 , .( 1 , 2 , 2 , 5 L. 1 , .(. 1 , 1 2 2 ' 1 1 .(.1 , 2 < 1 , <. 1 , 2 , 2 , 5 a. persi cum B o i ss ier a. praecox B o i s s i er & B al ansa a. propinguum J . B aumgartner a. pseudo-mou radi cum H au ssknecht & Born- mUl l er ex J . B aumgartner a. pulvi nar� Ve l enovsky a. purpureum Lagasca & Rodri guez a. repens J . C . G . Baumgarten a. rost ratum Steven a. scardi cum Wettste in a. scutigerum Durand a. smyrnaeum C . A . Meyer a. sphac i o t i cum B o i s s i e r & H e l dre i ch a. stapfi i V i e rhapper a. str ibrnyi Vel enovsk y A . s tr i ctum W i l l denow �· strigosum B ank s & Sol ander A. s zowi ts i anum F. F i scher & C . A . Meyer A. taygeteum Hel dre ich A . tenu i fo l ium S tephan ex W i l l denow A . tr i chocarpum T . R . Dudl ey & Huber- M o rath A . turk estani cum Regel & Schmalhausen A. umb e l l atum Desvaux �· vou ri nonense Dudl ey & Rechinger �· wi erzb i ck i i H euffel A . wul feni anum Schlechtendal A. xanthocarpum B o i s s i er mean n i ckel content fo r sect ion Section G amo sepalum ( Haussknecht ) T . R . Dudley ( a ) Seri e s Connata T . R . Dud l ey A . l epi do to-stell a tum ( Haussknecht & B o rnmull er ex H aussknecht ) T . R . Dudl ey A. paphlagoni cum ( Haussknech t ) T . R . Dudl ey �· tetrastemon Bo i s s i er 2 , 3 <1 , < 1 67 <..1 , 9 1 , 2 , 5 , 6 , 9 41 , 5 �1 , < 1 , 2 , 2 , 6 < 1 , <. 1 , 4 , 1 0 7 , 26 �1, 9 .( 1 , < 1 , < 1 , 1 <. 1 , �1 , 2 <..1 , 1 0 1 , 1 , 1 5 2 , 3 <.. 1 , 5 2 , 6 2 3 , 4 <.. 1 ' 1 4 <..1 , < 1 , 5 <1 , 1 1 4 <.1 , (1 , < 1 , 1 1 9 , 1 0 , 20 5 3 , 3 2 , 60 3 , 25 , 7 5 A . thymops (Huber-Morath & R e e se ) T . R . Dudley ( b ) Series L ib era T . R . Dudley A . b aumgartneri anum Bornmul l e r A . corningi i T . R . Dudley �· harputi cum T . R . Dudley A . lycaoni cum ( E . E . Schul z ) T . R . Dudley �. n iveum T . R . Dudley a. sulphureum T . R . Dudley & Hub e r-Mo rath mean nickel content f o r section Section Tetraden i a (Spach ) T . R . Dudley A . cochleatum Cosson & Durand a. l apeyrousi anum Jo rdan A . spinosum Li nnaeus mean n i ckel content for section Section Odontarrhena ( C . A . Meyer ) w. Koch I . Subsection Inflata T . R . Dudl ey A . alpe stre L i nnaeus A . ameri canum Greene a. anatol i cum Haussknecht ex E . Nyarady A. b a i c a l i cum E. Nyarady �. b ertolon i i Desvaux a. b i ovulatum B usch I I a. borzaeanum E . Nyarady a. b racteatum B o i ss i e r & Buhse �. cal i acrae E . Ny�rady a. c al l i chroum B o i s s i er & Buhse a. chondrogynum Burtt a. condensatum B o i s s i er ex Haussknecht ssp. condensatum A . condensatum ssp. flexi b i l e (E. N yar�dy ) T . R . Dudley A . condensatum ssp. lyc i um 9 , 1 1 , 1 6 <.1 , 1 , <1 , 3 <.1 , 6 4 , 33 1 , 5 8 , 34 5 2 , 3 L 1 , 7 � 1 ' 9 4 3640 � 6 7 , 1 1 , 2 1 , 36 , 1 47 , 1 84 , 381 <: 1 , 1 0 , 8 170 <. 1 , 4 , 1 5 , 27 1 3400 � 5 .( 1 , < 1 , 2 <.1 , <. 1 , < 1 , 22 2 , 2 , 2 , 1 5 , 32 33 , 1 46 , 1 930 , 1 0900 3980 , 7260 , 1 6300 1 3 , 1 5 , 1 9 , 8 1 , 700 <. 1 , 9 , 37 5 , 2 330 , 4990 <.1 , 1 1 9 , 1 59 A . constel l atum Boi ssier A . corsi cum Duby �· corymbosoides Formanek �· cypri cum E . Nyarady �· davi si anum T . R . Dudley a. di scolor T . R . Dudley & Huber-Mo rath �· eri ophyllum B o i ssier & Haussknecht I a. euboe um Hal acsy a. fal l acinum H aussknecht a. fedtschenkoanum Busch �. fil i forme E. Nyar�dy a. fragi l l imum ( B al dacc i ) R e ch inger f. a. gehamense Federov a. haussknechti i B o i ss ier a. hub er-mo rath i i T . R . Dudl ey A . infl atum E . N ya rady a. l anigerum De Cando l l e / I A . l ib ano ti cum E . Nyarady �· longi stylum ( Sommier & L ev i er ) Grossheim a. markgrafi i O . E . S chulz 8· masmenaeum Boi ssier A . neb rodense T inea A . obova tum ( C . A . Meyer ) Turczaninow A . obtus ifol ium Steven ex De Cando l l e _8 . oxycarpum B o i ssier & B alansa A . pateri E. N yarady 1 3 , 3820 , 5 380 , 1 8 1 00 4600 , 6 1 7 0 , 1 1 400 , 1 1 700 , 2 1 500 2 3900 , 4 14 0 , 7670 1 4400 , 20200 , 2 3600 6500 , 1 1 600 , 1 9600 4450 , 6030 , 87 30 , 1 1 700 8530 , 1 1 500 , 1 1 500 4 550 * 3960 * 6 , 7 1 32 , 1 5 1 , 8 1 4 L..1 , L.. 1 £:.1 , < 1 , < 1 L.. 1 , < 1 , 1 1 5 1 220 , 3540 , 4 99 0 , 1 1 800 ' 1 3500 � 1 ' 20 ' 54 ' 6 3 L.. 1 , 5 , 6 3 , 1 00 < 1 , <. 1 , <: 1 L..1 , < 1 , 6 3 , 9 3 1 3700 � 5480 , 1 5 500 , 1 5600 , 24300 � 1 , < 1 , L.. 1 , 1 1 2 , 6 , 26 , 64 , 1 030 , 4590 60 (1 ' 1 0 ' 4 4 6 0 ' 7 2 9 0 <1 , 30 , 44 , 5 3 , 6 1 1 84 , 484 A. penjwinensi s T . R . Dudl ey <.1 , 1 3 , 1 720 , 7860 A. polycl adum Rechinger f . � . < 1 , 6 , 1 27 �· rob erti anum Bernard ex Gren i er & Godron 1 2 500� A . serpyl l i fol ium Desfontaines ssp. serpyll i- fol i um <1 , � 1 , <1 , 2 , 4 A . serpyl l i fo l ium ssp. lusi tani cum T . R . Dudl ey & P. S i l va 9 ooo" A . serpyll i fol ium ssp. mal a c i tanum R i vas Goday �1 , 1 0 , 4 83 , 1 760 , 8470 A . s i b i r i cum W i l l denow A . s ingarense B o i ss ier & Haussknecht B.· A . A . A . �· smo l ik anum E . N yaradv s�ri acum E . N varadv szarabi acum E . / I Nyarady tavol arae B ri quet tortuosum Waldstein_ & W i l l denow A . t roodi i B o i s s i er A . turgi dum T . R . Dudley K i taibel ex I I . Subsection Compressa T . R . Dudl ey ( a ) S er ies Integra T . R . Dudley A . ak amasi cum Burtt A . a rgenteum A l l i oni A . cassium Bo i ssier A . jancheni i E . Ny �rady A. mu rale W al dstein & K i t aibel A . sub spi no sum T . R . Dudl ey A. tenium Hal �csy ( b ) Series Crenu lata T . R . Dudley A. c i l i ci cum B o i s s i er & B al ansa A . c renul atum Bo iss ier A . gi osnanum E . Nyarady A . hel dre i ch i i H aussknecht A . pterocarpum T . R . Dudley I I I. Sub section Samari fera T . R . Dudl ey A . c ari cum T . R . Dudl ey & Hub er-Morath A . dub e rtreti i Gomb aul t A . fl o ribundum B o i s s i er & B al ansa A. l esbi acum ( Cand�rgy ) Rech i nger f . �· pel tari o i des Bo i ssi er ssp . pel tarioi des I / A . pel tario i des ssp . vi rgati fo rme ( E . Nyarady ) 1 , 1 , 1 8 , 20 , 22 , 487 1 ' 8 , 29 , 1 280 6600� 3380 , 5250 , 1 0200 Not anal ysed 3 1 , 1 , 6 , 1 1 , 1 7 , 27 , 314 5 1 20 , 6560 , 95 1 0 , 9790 36 3660 , 4290 , 9090 1 0800� 5 590 , 6 320 , 7250 , 9 1 50 , 20 000 6 330 , 96 1 0 7080� 7 34 20 w 4260 , 1 3700 7080 , 8360 , 1 0400 4 170 , 4800 , 69 1 0 , 1 2500" 1 1 90 , 486 0 , 6740 3630 , 4780 , 6 1 30 1 6 500 7390 64 1 , 1 370 , 5620 , 7700 7920 , 8200 , 1 4 300 , 2 2400 2 , 4 , 17 Dudley 5300 A. peltarioides ssp. undetermined 7600 A. pi n ifol ium - ( E . N yarady ) T . R . Dudl ey 6670 , 9950 , 1 2600 .8_. B.· B.· samari ferum B o i s s i er & Haussknecht traEe z i forme Bornmu l l e r v irgatum E . N yaradv ex E . N y a rady mean ni ckel content for s ect ion ( excluding hyperaccumulat o r s ) 4220 , 5460 , I 2820 , 5600 , 1 830 , 3080 , 56 �H i ghest value reported b y B rooks & Radfo rd ( 1 978 ) 66 50 6900 , 1 1 900 5480 , 6 2 30 I t i s ob served in table 6 . 1 that hyperaccumul ation o f ni ck el i s confined t o section O dont arrhena . A total o f fo rty-e i ght t a x a of thi s section have shown thi s property . Al l non-O dontarrhena spec i es l i sted have now had the i r ni ck el content determined and of these the h ighest content recorded was 1 52 �g/g in �· m i crophyllum ( section A lyssum ) . Th i s sampl e was the onl y non-O dontarrhena specimen to surpass the 1 00 � g/g concentrati on l evel . �· mi crophyll um i s a speci es found throughout Siberi a . An e a sy compari son between the n i ck el concentrati ons of the O dontarrhena and non-Odontarrhena specimens can b e made b y v i ewing the h i s tog rams of fig. 6 . 1 . 1 3 6 From thi s i t i s readi l y apparent that the bulk of non-Odontarrhena specimens con tained l ess than 1 0 �g/g . The O donta rrhena specimens show a compl etel y d i fferent di stribut i on . I ndeed the di stribution is indicative of two popul at ions of spec i es w i th a separation boundary between 320 �g/g and 1 , 000 � g/g . Th i s boundary co i nc i de s well w i th the value desi gnated b y B rook s , Lee et �· ( 1 977 ) a s the concentration cri terion for n ickel hyperaccumu l at ion . For Alyssum spec i es in sect ion O dontarrhen a , 55% of the spec imens h ave n i ck e l l evel s below 1 , 000 �g/g and 4 5% have co ncent rat i ons above th i s . The h i ghest concentration was r ecorded fo r t he Turk i sh serpen t i ne-endem i c , A . masmenaeum ( 2 . 4 3% n i �k el ) . Further compari sons between O dontarrhena and non-O dontarrhena spec i es can b e made by comparing the mean n i ckel content for each section ( tabl e 6 . 1 ) . Even a fter the val ues o f the hyperaccumul ators have been excluded , the a r i thmet i c mean fo r O dontarrhena at 56 �g/g i s an order of magni tude greater than that for any o f the other five sections : Meniocus , 5 �g/g ; P s i l onema , 3 �g/g ; Al y ssum , 5 �g/g ; G amosepalum , 5 �g/g ; and Tetradeni a , 4 �g/g . O n l y e i ghteen o f the seventy-four taxa l i sted under Odontarrhena fai l e d to reach 1 00 �g/g in any specimen , wherea s , as ment i oned above , only one non-O dontarrhena taxon reached thi s l evel . Tol erance to h i gh n i ckel appears to b e a common al though no t i nvari abl e property o f O dontarrhena spec i e s . ---- - - -------- ---------- 80 (a ) 40 - � l I .1:2 0 I � 1 � 80 ( b) 40 - · I I 0 0.32 1 .0 3.2 10 32 100 320 1000 3200 10000 32000 I I nickel concentra.tions/(J.lg/g} Fig urC2 6 . 1 H i s tog rams o f n i cke2 l concan trat ion in .A_/y_ssum spQc iman s . ( a ) Spec i e s of sect ion Odontarr h<2na . ( b ) Spec i es of sec t i-ons Men i ocus J Psi lonC2maJ A l y ss umJ Gamosep a l um and Tet raden i a . The geograph i cal d i s tribu t ion o f the O dontarrhena specimens i s shown in f i g s . 6 . 2 , 6 . 3 and 6 . 4 . The Turk i sh specimens ( f i g . 6 . 2 ) are shown di str ibuted b y v i l ayets ( provi nces ) rather than actua l local i t i es . A very s t rong correl a t i on between the exi stence of u l t rab as ic rock s and the di str ibut ion of n i ck el hyperaccumul ating specimens i s cl ear � y seen . Specimens i n the western I rano-Turani an phytogeographi cal region ( I ran , I raq , Transcaucasia and eastern Turkey - fi g . 6 . 3 ) and the o u t ly ing specimens (Ukraine , eastern I rano-Tu ran i an reg i on , S i b e r i a and Yukon d i str ict - f i g . 6 . 4 ) show a much l ower hyperaccumu l a t i on p ropo rt i on ( for the d i s t ribut i o n o f the European hype raccumul ators see B ro ok s & Radford , 1 978a ) . From these maps , i t can b e seen that almost al l hyperaccu�u l a t i ng spe c i e s are found in southern Euro pe , the eastern Medi terranean area and Turk ey . O n l y three spec i e s , A . obovatum ( found i n sou theast Russ i a , southern Uk ra ine and S iberi a ) , a. penjwinen s i s a n d a . singarense ( b o th endemi c to Iraq ) , showed hyperaccumu l ation outs ide thi s reg i on . Speci e s wh i ch contain over 1 0 , 000 �g/g ( = 1 % n i ckel ) are a lmost a l l rest r i c ted to east e rn M ed i terranean l ands ( Greece , Aegean I s l ands , Cyprus and the Med i terranean coastal reg i on s 1 3 8 of Tu rkey and northern Syria ) . Some however occur i n central Turkey , northern I ta l y and Co rsi c a . Spec i e s which hyperaccumula te more than 1% n i ckel ( there a r e e i gh teen such speci e s ) are genera l l y very restr icted in the i r d i s tr ibut ion . A . constel l atum i s the excep t i on to th i s rul e but even t h i s spec i es i s found onl y in Turkey and northern i raq . Spec i e s which do not hyperaccumula te n i ckel to t h i s exten t frequentl y have much greater d i s tribu t i on ranges . Some of these speci e s are n o t serpentine endemi c s , eg . �· obovatum , and may have specimens with l ow n i ckel content : 2-4 , 600 �g/g is the range for a. obovatum . There i s l i t tla quest i on that many Ddontarrhena speci e s are very successful i n their adaptation to serpent ine envi ronments . Several t axa are found i n almo s t pure popu l at i ons on serpentine outcrops eg . A . mural e i n the Balkans and A . corsi cum and A . cypricum in southwestern Turk e y . 0 100 � km I "''"' � . . . . - 0 0 0 0 0 0 0 0 0 0 SYRIA . . O D D O DO / / / / IRAQ E_g u r<2 6 . 2 Oi s t r i bu t io n I by v i laye t s J of Tu rk i sh spec i e s of S Q c t io n Odontarr h<2na . Key to F i g u res 6. 2 , 6 . 3 + 6 . 4 • Spe c i mens 0 Specimens • Spe c imen s 0 Spec i mens con ta in in g con ta i n in g con t a in in g cont ain i n g n i ckel concentrat ions > 1 0 . 000 �g /g 1 . 000 - 10 , 000 � g /g 1 00 - 999 1J g I g c 1 00 !Jg / g The b lac k ar e a s i n F i gure 6 . 2 are u l t rabasi c rock s . Iraq d � U.S.S.R. -- -- - --- ------- --- fl_gure 6 . 3 Q.aog raph i ca l d i st r i bu t io n of s pQ c ie s of sQ c t i on Odo n ta rrhQ na 1n t h e wes t c2 rn Iran o - Turan i an rQg i on . U.S.S.R. 0 lOOO km £igurQ 6. 4 GQog rap h i ca l d i s t r i b u t i on o f o u t l y in g spQ c imQn s of s p Q C i Q s o f sect i on Odonta r r he2 na . For t h e d i s tr i bu t i ons in ( a ) s<2 <2 Broo ks a n d R a d fo r d , 1 9 7 8 a ; ( b ) s ee F i gure 6 . 2 ; { c ) see F i g ur e 6 . 3 . W i th i n s e c t i o n O d on t arrhena , i t i s nbservP.d thn t t h P. subsec t i nns Comp ressa and S am a r i fera are each cnmposed 1 4 2 ent i r e l y o f hyperaccumu l a tors w i th one e x cep t i on : � · sub sp i no sum for Compressa and A . pe l t a r i o i d e s f o r Samar i fe ra . N e i t h e r o f these two spec i e s i s found w i th i n t h e c e n t r e o f d i str ibu t i on of the hyperaccumu l ators . �· sub spi no sum i s k nown onl y f rom Jordan whi l e �· pe l t a r i o i de s i s found i n centra l a n d e a s t e rn Turk ey . Subsec t i o n I n fl a t a con t a i ns twenty-seven hyperaccumu l ­ ator s , e i ght s t rong accumul ators ( 1 00 - 999 �g/g ) and twenty non-accumu l a t o rs ( l e ss than 1 �n � g /g ) . The l ectotype spec i e s for sect i o n Ddonta rrhena , � · to rtuosum ( subsec t i �n I n f l a t a ) i s n o t a hyperaccumu l a t o r . T h e e x i s tence o f a non-accumu l ator i n a sub sec t i o n wh ich o therw i s e cons i st s who l l y o f hyperaccumu l ators rai ses t he que s t i on as to whether th i s spec i e s i s the p arent spec i es o f the hyperac cumu l a to rs . W i t�i n sub sect i on Compressa thi s qu e s t i on i s fu rth e r c�mo l i c a t ed b y t h e e x i s tence n f two s e r i e s , I nt e g ra and C renu l a t a . S e r i e s C r enu l a t a cons i s ts whn l l y o f hyperaccumu l a to rs . 5 R r i P. s I n t e q r a c onta i n s t h e non- accumu l a tor � sub spi nosum a n d a l s ,l A . mu ra l e wh i ch i s no t endem i c to serpen t i ne so i l s . T h e g en g r a ph i ca l di st r ibu t i on � f the spe c i e s o f ser i e s I n teg r J i s shown i n f i g . S . 5 . I t i s a lways p oss i b l e that �- sub spi nosum is a re l i c of a former w i d e spread b u t now general l y d e funct spec i es but the no t ab l e spread of �· mural e i n i t s many gu i se s · ra i ses doub t s . �· mural e i s a hi ghl y var i ab l e and a c t i vel y evo l vi ng g ene t i c un i t ( D udl ey , 1 964b ) . Furthermore �· mu ral e has , i n the past , b een d i vi ded into appro x . forty taxa , e i ther mi cro spec ies o r i nfraspec i e s . I t cou l d b e t hat i t i s A . mu r a l e from wh ich the o thers have evo l ved . A . subspi nosum may then have evo l ved from A . mu ral e as an adap t a t i on to the Jordan i an envi ronment rather than A . mu ral e der i ving from A . sub spi nosum . Al t ernat i ve l y some factor may have a l l owed A . mu ral e to supersede a prev i ous parent spec i es o f wh i ch A . sub spi no sum is a rel i c . The hype raccumu l a t i on of n i ck e l b y � . mu ral e i s worthy o f deeper s tudy a s al though the el even specimens ana l y sed b y � rook s and R adford ( 1 978a ) all hyper­ accumu l ated , the specimens anal ysed b y t h i s author d i d n o t . -- - - - - - - ........ I RAN "\ ' mmm \ I I I .· . . . . . . . . . -;. / / / / I I I £i gur e 6 · 5 Th e d i str i bu t ion of s p e c 1 e s : s e ct i on O d o n ta rr h ena, subsect i on Compress a , s e r i e s I n t e g r a . ak = A . akamasicum ar = A . argen teum ea = A . cas$jum j = A .janchenii s = A . subsR_inosum t = A . tenium mmm= A. murale s. s p. murale var. murale mma= A. murale s.sp. murale va r. g/R_inum mmh= A . murale s. sp. murale var . haradjiani mmp =A. murale s. sp. murale var._p_ichleri ms =A . m urate s. s p. stojanoffii T U R K E Y -- - / \ / ci \ - - - - / ...... .... ...... l', .· · · · · · · · · · · P -"--\- -- -.,--.... t ·r · --- g ' fr-:5 SYR IA .. -/ / F i g u r e 6 . 6 The d i s t r i b u t io n of _A/y_ssum speci e s : s e c t i o n O d o n tar r h ena , su bsec t i o n C ompr e s sa, s eri es Cren u l a t a . c 1 = A . c i I ic i cum er = A . crenu la tum g = A . giosnanum h = A . heldreichii p = A . p_terocarR_um 1 4 5 There wa s , howeve r , a consp i cuou s d i f f e r ence i n co l l ec t i on l o cal i t i es for these samp l e s : B rook s and � adford ( 1 978a ) sampl e d spec i e s from the B a l k ans ( p rnbab l y �· mura l e ssp . mu ral e var . p i ch l e r i ) ; th i s autho r samp l e d spe c i e s f rom the USSR and Turkey ( prob ab l y A. mu ral e ssp. mu ral e var. mu ral e or var . a l pi num ) . Assuming t hese i dent i f i cat i ons to b e co rrP.c t , i t i s then no t improb ab l e that , i n t ime , A . mura l e ssp . mura l e var . pi chl eri wi l l deve l o p i n t o a new serpenti ne-endem i c spe c i e s . W i th i n s e r i e s C renu l a t a , � · c i l i c i cum emerges as the most probab l e progeni t or on the b a si s of the al l i ances g i ven b y D u d l e y ( 1 9S 5a ) . The poss i b l e l i ne of evo l u t i on o f the spe c i e s o f this s e r i e s may b e as fo l l ow s : c i l i c i cum � � c renu l a tum gi o snanum p teroc arpum h e l d re i c h i i T h e d i s t r i b u t i o n o f t h e s e s p e c i e s i s shown i n f i g . 6 . 6 . The r e l a t i onsh i p b e tween the two s e r i e s o f sub s ec t i on Compressa i s b eyond the scope o f thi s d i scu s s i on . The quest i on. a s t o whether A . pe l t a r i o i d e s i s the parent spe c i e s of sub se c t i on S amari fera is perhaps l e ss compl i ca t e d . A . p e l tar i o i de s doe s , howeve r , have t h r e e sub speci e s ; A . pel tar i o i d e s ssp . p e l t a ri o i de s i n eastern and northeastern Turk e y , � pel t a r i o i de s ssp . vi rgat i fo rme in central T u rk e y , and a n undeterm ined b u t d i fferent sub spec i e s ( T . R . Dudl e y p e r s . comm . t o R . D . R eeve s ) i n Bu rdu r , sou thwest Tu rk e y . The p o ssi b i l i ty i s that A . pe l tar i o i de s s s p . pe l tari o i des i s the parent t ax on of sub sec t i on S ama r i fe ra . Th i s may b e i nferred from a probab l e past d i s t r i bu t i o n . A . pe l tar i o i de s s sp . p e l tar i o i de s i s found t o day a t h i gh a l t i tudes i n t h e vi c i ni ty o f mel t i n g snow ( 2 , 000 - 3 , 580m , Dudl ey , 1 964b ) . During the i ce-ag e s , i t cou l d e a s i l y have b een more wi despread i n Turk e y . U S S R pv ...... • • • • • • · · - - 0 ...... - - - - ' - --":--...... .........,_ - - - ...... ...... - · · · · · ' - - ....... ........ -"'- ·· · . ' - - ....... - - ·. \ - ........ ' _ _ J' /' 0 ,.... , ,....- - nn : \ """ I - - ,...,,..., . - /' ' - ..... ..... - - , '\ : \ I ,...< v ' -r , J \ :. / y \ ........ / , - - - -...... . . . . . . \ ( I \ - .,...::- - ,,/ 1 -1 - "\ · · · . ....._ I \ - - - - - v/ / ...,. r I \ \ . . · I - -\- - - - - - � ,. "" - - I I .·\ ( \ / - \ .· \ \ TU - - - / - I / : I \ \ R K E Y I - -.. � ...\ / _.,. / I \ '· I \ \ I f (",..: ...,. S 11 / \ · : \ ' I f ' / I I :. I - \ " " \ .... ..... / : ,, -.... ..,, , \ I ' \ ..... '\,, I ) I / I 0 I ...... .... f ,(\VI ' ........ � 7. / I : l \\r,' f ', / - --HL-""" 1 1 \ -.... , c / \ ...... ...... ,.... 1 , 11 /1 /1 0 . · · - . . . . . .\ ':.J ' ,_{"'\......___ ...... _ _ , - ./1-r:y/ / _ . . . . - · · · · · · · . . - ·-.;,;:2/: I" � r J ··y( . . · · - . . . . . . . . . . . . . . . . · d i / t/ SY R I A � £lg u r e 6 · 7 The d i s tri bu tion of Aly_ssum spe c i es : s e c t ion Odontarrhen a , subsec t i on- S amari fera. c = A . caricum d = A . duber tre tii f =A floribundum I = A . lesbiacum pp = A .J2eltari01des s.sp. _f2eltarioides p v = A . _f2.eltarioides s . sp. virgatiforme s = A . samari ferum t = A . trap_eziforme p =A._R.I'nifolium v = A . virga tum pu =.A .f2eltarioides s. sp. und et e rm i ne d • Then , a s the c l imate warmed up after the i ce-age s ende d , i t b eg an to retreat i nto the coo l er r eg ions i e . the moun t a i ns of Turk e y . As i t ret rea ted , i t l e f t b eh ind specimens i so l a t e d i n v a r i ous areas . These specimens con t i nu e d to g row and evolve in i so l a t i o n and have g i ven r i s e to the spec i e s seen today . T h a t these specimens survi ved wou l d b e d u e to the i r b e i ng i n envi ronments whi ch were host i l e to t he i nvading spec i e s i e . they were in serpent i n e so i l s . A d i st r i bu t i on of thi s t ype i s p aral l e l e d i n Scandi n avi a today b y Arenar i a norvegi ca Gunn . wh i ch sou th o f 66°N ( the warmer area ) i s found almost excl u s i v e l y on serpen t i ne so i l s but fu rther north ( the cool e r a r e a ) is more w i despread ( Rune , 1 9 53 ) . G i ven t ime , the i so l a t e d south ern popu l a t i ons cou l d evol ve i nto new t a x a . Turk e y , b e i ng re l eased from t h e i ce-ages earl i e r than Scand inav i a , has had the t ime t o deve l o p these n ew t axa . The occurrence o f a sub spec i e s of A. pe l t ar i o i des i n sou thwest Turkey appears to b e an i n d i ca t i on that the p a rent taxon was once mo re w i despread . That these var i n u s subspec i e s e x i s t as W8l l a s the i ndependent spec i e s may be a r esu l t o f th8 � u l t i ­ pl i c i t y o f i ce-ages ( B i rman , 1 958 ) . The i ndependent soec i 8 s may have deve l oped p r i o r to the l a test post-Pl e i stocane 1 4 7 g l ac i a t i on wh i l e the sub spec i e s have a r i sen sub sequent tn th i s g l acfat i on a s A . pel ta r i o i de s o n c e m o r e retreated to cool e r a reas . T he p re sent d i s tr ibu t i on o f �· pe l t ar i o i d e s ssp . pe l t a r i o i des is i t sel f p rob ab l y · o f R ecent o r i g i n . The t axon probab l y mi grated from t h e Taurus Moun t a i n s al ong the Anato l i an d i agonal ( a mount a i nous be l t from Seyhan-N i g de to S i vas ) and thence to the northeast and east o f Turk e y . T h i s m i g ra t i on expl a i n s t he g reater mu l t i p l i c i ty o f spec i e s i n the Tau ru s Moun t a i ns o f sou th e rn Turk ey . T hese r anges are more s t rong l y g l aci a ted than the no rthern Pont i c Mou n t a i n s suggest i ng t h a t the l at t e r have onl y recent l y ri sen to the i r cu rrent h e i g h t ( B i rman , 1 968 ) . Thu s p r i o r to thi s recent up l i f t , �· pel t ar i o i des ssp . pel t ar i o i des may have b e en r e s t r i c te d to the mount a i n s of sou thern Turk ey . Fu rther evi dence for t h i s l ate m i g ra t i on comes from the recogn i t ion by D ud l e y ( 1 965a ) t h a t �· v i rga tum , the most northerl y i ndependent spe c i e s , i s al so the c l o ses t 1 4 8 r e l a t e d spec i e s . I t o l so a ppears t h a t the two sub spec i e s � · pe l t a r i o i de s ssp . p e l t a r i o i de s and s s p . v irgat i forme are c l o se t o separa t i on . The area of o ve r l ap in the i r d i s t r ibu t i ons i s no t a l arge p a r t o f the i r ranges and w i t h t h e al t i t ud ina l p re fe rences a n d t h e t imes o f f lowering d i fferi ng , t h e y appear t o be c l ose to i ndependent deve l o pmen t . Thi s deve l o pment i s l i k e l y t o acce l erate a s the separation i s comp l e t e d and new spec i e s evo l ve . O n l y one specimen i ntermed i a te t o these two subspe c i e s has ever b e en recorde d . T h i s can b e compared w i th i n fraspeci f i c t a x a o f o t h e r A lyssum spec i e s where i n termedi a t e s in areas o f ove r l a p a r e common ( D udl e y , 1 95 5a ) . I t thu s appears p rob ab l e that �· pe l tar io i des s sp . pe l t a r i o i d e s cou l d be t he paren t taxon to the other members o f sub s ec t i on S amari fera . A poss ib l e l i ne of deve l o pment (based on al l i ances g i ven b y D u dl e y , 1 965a ) is a s fol l ows : pe l t a ri o i des unde t e rm ined _./"" sama r i fera 7 ./""' . 7 / " / pel tar i o i de s ee l t a r i o i des ~ pe l tari oi de s vi rga t i fo rme vi rgatum l esb i acum fl ori bundum - trapez i forme - cari cum I p i n i fo l ium dub e r t re t i i B rook s and Radford ( 1 978a ) noted that whi l e �· s e rpyl­ l i fo l i um ssp . serpyl l i fo l i um was not an accumu l ato r , �· s e rpyl l i fol i um s sp . l usi tani cum i nvari abl y wa s . They su ggested that th i s c ou ld be u s e d to support a change of status for the sub species l u s i t an i cum whi ch Pinto da S i l va ( pe r s . comm . to R . R . B rook s , 1 977 ) was a l ready con s i de r i n g . T h i s change o f name i s n ow u n d e r w a y ( T . R . Dudl ey , i n ed . ) and the new name i s to b e A . p i n todas i l vae . S ince these commun i c a t i on s , a second subspec i e s o f �- serpyl l i fo l ium , �· serpyl l i fo l i um ssp . mal ac i t anum , has shown hyperaccumu l a t i on ( tabl e 6 . 1 ) . A s a consequence o f th i s , a n y dec i s i on to spl i t 1 4 9 A . serpyl l i fo l ium i nto a numb er of spec i e s must b e recons i dered . I f � · pintodas i l vae i s a ccepted as a spe c i e s i ndependent o f A . serpyl l i fo l i um then the s tatus o f �· serpyl l i fol ium ssp . mal ac i tanum wi l l b e very o pen to deb ate . However , the f inal cons i dera t i o n s for the spl i t t i ng o f th is spec i e s into several new spec i es mu st b e b ased on more t han the n i cke l content and i t s a ssoc i ated phy s i o l o g i c a l and morpho l o g i cal stresses . I t m a y b e that n i ckel stress a s a cause o f mo rphol o g i cal vari a t i o n w i l l h ave t o b e stu d i e d i n the l ab o ratory b efore the s i tu a t i on b e comes c l e a r . T h a t c e r t a i n sub spec i e s are d i s t i ngu i shab l e morpho l o g i cal l y and al so in n i cke l content can again be seen for �· condensatum where �- co ndensatum ssp . condensatum and A . condensatum s s p . lyc ium do not hyperaccumu l ate b u t �· condensatum ssp . fl ex i b i l e does ( N o t e : the two sub spec i e s conden satum and lyc i um may be o ne and the same , as no reference to the subspec i e s lyc i um c an b e found b u t O dontarrhena lyc i a J ord . & Fourr . i s g i ven by Dudl ey , 1 965a , a s a synonym o f the sub spe c i e s condensatum ) . One specimen of A. condensa tum s s p . condensatum w i th a concentra t i on o f 700 �g/g l ooks m i sp l aced when compared t o the other samp l e s o f t h i s t axon . Thi s spec imen may , o r may not , h ave b e en i nco rrect l y i dent i fi ed . M i s i dent i f i c a t i on s are no t unknown among herbar ium specimens ( cf . B rooks & Radford , 1 978a ) . CHA P TER 7 B i oge ochQm ica l St ud i es on Some Aly_ssum SpQc iQs 7 . 1 INTRODUCTION The uptak e o f n i ck el to hyperaccumul at ion l evel s b y a p art i cu l ar spec i e s can l ead to the u se o f that speci es as a b i ogeochemi cal ind i cator ( Severne & B rook s , 1 972 , Co l e , 1 97 3 , B rook s and W i th e r , 1 977 , L e e et §!. , 1 977a ) . Many other tol erant but non-hyperaccumu l at i ng speci e s have a l so b een used for b i o geochemi c al prospecting ( Lyon e t §!. , 1 968 , Timp e rl ey et §!. , 1 97 0 a , 1 972 a , b , N i e l son e t §!. , 1 97 3 , B rook s , Trow and B o lv i ken , 1 97 9 ) . However , t h e b a s i c study o f t h ese spec i es from a b i ogeochemi cal v i ewpo int i s l im i te d . Such quest ions a s the extent o f poten t i al uptak e o f n i ck e l , the rel ationsh i p b e tween n i ckel in the pl ant and n i ck el i n the so i l , the rate of uptak e of n i ck el b y t h e p l ant and t h e degree of tol erance of these spe c i e s h ave yet to b e answered. Kersten ( 1 97 9 ) reported some i n i t i al i nvesti gati ons on 1 5 1 the p l ant-so i l re l at i onsh i p of the N ew Cal edon ian hype raccumu l ator Psychotr i a dou arre i . I n pot t r i al s wh i ch i nvo l ved adding , i n a t ime sequence , i nc rements o f n i ck el t o a po t t i ng medi um , he showed a l inear r e l at ionsh i p between the l ogari thm i c values o f n i ck e l i n both t h e pl ant a nd t h e s o i l ( expressed on a dry w e i gh t b a s i s ) . Th i s form o f rel at ionsh i p i s cons i dered b y Timperl ey et a l . ( 1 970b ) as common for the uptake o f non-essenti al el emen t s . Lee et §!. ( 1 977a ) have a l so shown st rong rel at ionshi p s between n i c kel in p l ants and extractab l e n i ck e l ( ammonium oxal a te buffer as extractan t ) in so i l s fo r the two N ew Cal edoni an hyperaccumul ators Homal ium k anal i ense and Hybanthus austrocal edoni cu s . Th i s extractab l e n ickel i s prob ab l y comparab l e t o the n i ck e l measu red b y Kersten ( 1 979 ) s i nce the l atter s tudy i nvol ved adding n i ck e l n i t rate whi ch woul d b e read i l y extracted . H . austrocal edonicus a l so had a s i gn i fi cant rel at ionsh i p between the n i ck e l content o f the pl ant and the total n i ck el content of the so i l . A th i rd spe c i es studi ed b y Lee et al . ( 1 977a ) , Homal ium gu i l l a i n i i , showed no s i gn if i cant rel ationsh i p s b etween n i ck el in the pl ant and e i ther 1 5 2 total o r extractab l e n i ck e l in the so i l . Crook s ( 1 97 9 ) and B rook s and Crooks ( 1 980 ) , studying the Fennoscandi an specie s , Lychnis a l p ina and S i l ene di o i ca , showed that �· d i o i ca h a d a l i near rel at ionship b e tween n i ck e l i n the pl ant and i n the so i l s whi l e h· alpina showed an exclus i on-b reak down form of rel ationsh i p ( see Chapter 4 ) . The thresho l d v a l u e fo r the excl usion break down was appro x . 3 , 000 �g/g ( so i l content ) . Whi l e h· a lpina i s n ot known a s a n i ck e l hyperaccumul ator i n t he wi l d (maximum concentration 1 9 2 �g/g - B roo k s , Trow & B o l vi k en , 1 97 9 ) , the pot tri al s of Crooks ( 1 979 ) showed that thi s spec i e s is c ertain ly capab l e of accumu l a t ion � f n i cke l to t h i s l evel . At a so i l concentration o f 4 , 1 00 �g/g , h · alpina had a l eaf concentration of 7 , 300 �g/g . h· alpi n a d i d n o t su rvive i n h igher so i l concentrat i ons o f n i cke l . Psycho tri a douarrei has survived i n 9 , 1 00 �g/g al though a t t h i s concentration , tox i c i ty symptoms became evi dent ( Ke rsten , 1 979 ) . Th e corresponding l e af concentration was 9 , 500 � g/g . B o th �· douarrei and h· alpina were ab l e to su rvi ve in soi l s w i th a very h i gh avai l ab l e ni ck e l content . The N ew Cal edon i an spec i e s stud ied b y L ee et �· ( 1 977a ) had t o t a l so il n i ck e l conten ts o f 1 - 2% but the extractab l e ( and hence readi l y ava i l ab l e ) n i ck e l was g enera l l y of the order of 1 , 000 - 2 , 000 �g/g ( 0 . 1 - 0 . 2% ) . Pro c t o r and Woode l l ( 1 97 5 ) have tabul ated many s erpent ine s o i l n ickel concentrations ( from many source s ) w i th in these ran g e s . It thus appears that hyper­ accumul ators and o ther tol e rant spec i e s h ave adapted to the nickel content of t h e ir envi ronments w i th an excess capaci ty (a margin of safety fo r the varying soil content? ) . Lychnis alpina h as al so b een the subject o f tol erance tests ( Crook s , 1 979 ) . The method u sed in thi s test was a variation o f the roo ting technique deve l oped by W i l k i ns ( 1 9 57 ) and Jowett ( 1 9 58 ) . S e ed l ings with t he ir roo t s exci sed were p l aced in s o lut ions conta in ing 0 . 5g/dm3 ca lcium ni trate and incremental amounts o f n i ck e l . The tol erance l evel was taken as the h ighest concentration i n which measurab l e new root g rowth occurred . Crooks ( 1 97 9 ) found t h a t h• alpina developed good roo t s i n solutions o f 1 � g/cm3 and 2 �g/cm3 but that at 5 �g/cm 3 the devel opment was only sl i g h t . Al though these l evel s of n icke l concen trat ions may s e em l ow , th e y a r e sub stanti al l y h i gher than tho se u s ed i n the devel opment o f i nd i ces of n i cke l tol erance fo r Agrosti s spp . ( J owett , 1 9 58 , G rego ry & B radshaw , 1 965 , �ro c t o r , 1 97 1 ) . Stud i e s of n i cke l to l erance i n o t h e r n i ck el to l erant spec i e s a re very few . E rnst ( 1 972 ) measured the tol e rance of Indigofera s et i flora by compara t i ve proto p l a smatology . H e was ab l e to show that I . set i fl o r a from a n i ck e l i ferous area was more tol erant than the same speci e s from non-ni ck e l i ferous are a s . N i cke l t o l e rance has however been sai d to b e l ess spec i fi c than other metal t o l e rances ( Proc t o r & �oo del l , 1 97 5 ) . The re l at i ve l ack of cobal t i n some hyperaccumu l ators of n i ck el ( e g . Alyssum spp . , Ch apter 6 ) and the rel a t ive enri chment of cobal t in others ( eg . � i norea spp . , B rook s , W i th er & Zeperni ck , 1 977 , Phyl l anthus spp . , Kersten , 1 97 9 ) has a l so to b e exp l a ined . I n many serpent ine so i l s the cobal t content i s approx . 1 0% o f the n i ck e l c ont en t . The po s s i b i l i ty that geographi cal and cl imati c consi dera t i ons a s w e l l as g eo l o g i c a l consi derat ions are i nvolved canno t b e avo i de d . The h i gh cob a l t contents a re found i n n i ckel hyoe raccumu l a t o rs f rom t h e t ro o i c a l a reas of Southeast Asia and N ew Ca l edon i a wh e r e a s tho s e with l ow cob a l t contents come from t h a Medi t e rranean b a s i n wi th i t s ho t , dry summers and mi l d , mo i s t w i n ters . There d o n o t appear t o be any stud i e s pub l i shed as to po s s i b l e cause s fo r these d i fference s . 1 5 3 In thi s chapter , stu d i e s h ave b een made on the b i ogeo chemi cal factors mentioned abo ve . N i cke l uptake i n Alyssum spec i es i s s tu d i ed b o th a s a function o f so i l content and a s a funct ion o f t ime . T he tol erance o f several o f these spec i e s i s al so mea sured b y both a so l u t i on rooting method and a s o i l cul ture metho d . A s a further i nve st i g at i on , the uptak e o f cob a l t was stud i ed under a rel a t i ve absence o f n i ck el . A non-Alyssum spe c i e s B o rnmuel l e ra tymphaea was al so s tud ied f or n i ck el u p take a s a function o f so i l cont en t . T h i s speci e s was di scovered to b e a n icke l hyperaccumul ator b y R eeve s , B rooks and Dudl ey ( 1 981 ) . I t i s c l o se l y rel ated to the Alyssum spe c i e s and the g enus B ornmuel l era i s a memb er o f the tr ibe Al y s sea e of the Cruci ferae . 7 . 2 EXPERIMENTAL METHODS Seeds of e l even A lyssum spec i e s and of Bo rnmuel l era tymphaea were col l e cted b y various persons ( Appendi x 1 ( b ) ) and forwarded to Massey Universi t y . Upon arri val the seeds were p l aced i n cool s to rage unt i l u s ed . Germ ination was done on a Copenhagen tab l e . A l l twel ve speci e s germi nated readi l y . Once g erminated , the seedl ings were t ranspl an ted di rectl y i nto the experimenta l pots unl ess otherwi se s tated . The spe c i e s studi ed were Alyssum montanum L . ( section A l y ssum , a non- accumul ator ) , A . argenteum A l l . , �· corsi cum Dub y , �· euboeum Hal. , A . h e l dre ich i i Hausskn . , �. mura l e W a l ds t . & K i t . , A . s e rpyl l i fo l i um Desf . s sp . serpyl l i fo l i um ( a non-accumul ato r ) , �· serpyl l i fol i um s s p . l u s i t ani cum Dudl ey & S i l va , �· tenium H al . , �· troodi i B o i ss . , �· v irgatum Nyar . ( al l section O dontarrhen a ) and B o rnmue l l era tymphaea ( Hausskn . ) Hausskn . 7 . 2 . 1 N i ck e l Upt ak e Stud i e s A potting mixture o f 1 : 1 p eat/perl i te wi th added nutri ents was used a s the b a s i c medium i n these experiments . T o th i s m ixture was added n i cke l , a s t he ni trat e , to g i ve f ina l ni ckel concentrations i n the range o f 30- 1 0 , 000 p g/g . The mi xture w a s p l aced i n p last ic pot s each o f which contained 200g . A l l twelve speci e s w ere invol ved i n t h e se expe riment s . For each spec i e s , f i ve repl i cates were made fo r e ach concentration . A l l pots were k ept i n a g l a s shouse w i th a temperature range o f 20-2 5°C a nd were watered from b eneath . At the e nd o f s i x week s , a l l p l ants were sampl ed to determine the l ea f and s o i l n i ck e l content s � I n a dd i t i on t o these experimen t s , n i ckel uptake was stu died i n six speci e s g rown on a serpentine soi l from Dun Mounta i n , N e l son , New Zeal and ( containing 900 �g/g ) and i n f ive spe c i e s g rown on a N ew Cal edoni an serpentine soi l ( 4 , 790 �g/g ) . These specimens were l ik ew i se s ampl e d for n i cke l anal ys i s after six week s . 1 5 4 7 . 2 . 2 " Tr igger-Po int " Theo ry T est After t he result s o f t he n ickel uptak e studies became apparent , a test was made to determine at whi ch concent rati on of n i ck el in the soil ( i f any ) the hyperaccumul ation o f thi s e lement was tri ggere d . This test was done using the s ame basic potting mixture a s for the n i ckel uptake studi e s but the range o f n i ckel concentrations was l ower and narrowe r ( 20- 1 50 pg/g ) . All other experimental conditions were the s ame . O n l y two spec i e s , �· ten ium and �· t roodi i were tested. A l l pl ants and s o i l s were s ampled to determine t he ir ni ck el content after six weeks . 7 . 2 . 3 R a te o f N i ck el Uptake The rate o f n i ckel uptake was studied only in the spec ies A . euboeum . Seedlings o f thi s speci e s were g rown fo r f ive weeks i n a b ack ground ( i e . no added n i ckel ) po tting mixture b e fo re being transplanted into pots containing the potting mix ture plus n ickel to g ive a f inal concentration of 7 , 000 �g N i/g . Leaf s amp les were removed dai l y f rom each indivi dual and thei r ni ckel content was determi ned . Th i s experiment was then dupl icated o n a potting mixture containing 6 , 500 �g N i/g . 7 . 2 . 4 N i ck e l Tol e rance Studi e s T h e tol erance of several spec i es t o nickel w a s stu d i ed b y two methods : a solution rooting method modified from that developed b y W i lk ins ( 1 957 ) and Jowet t ( 1 958 ) ; and a soil cul ture method as devel oped i n Chapter 4 . For the solution roo ting metho d , a s eries o f soluti ons containing b e tween 0 � g/cm3 and 1 00 �g/cm 3 o f nickel i n 0 . 5 g/dm3 calcium n i trate were prepare d . Fou r week o l d Alyssum seedl ings were u sed fo r the test . T he seedl ings had had the i r roots exci sed with a s ca lpel and were then suspended in t he t est solu tions by being pl aced i n hol e s through 5mm thi ck pol y styrene rafts . There were five seedlings per species per concent rati on . 1 5 5 - -- ------------ The solutions ( 1 50cm3 ) were pl aced i n 250 cm3 squ at b eakers and aerated continuousl y . L i ght was supp l i ed b y radiation from infrared l amps al so on a continuous b asi s . The temperatu re was maintained a t 2 5°C . All solutions were changed week l y . Between change s , the solu tions were k e p t u p t o volume b y the addition of deion i zed water . After a period o f f ive week s , the l engths of the new roots were measured a s a basis for determining the tolerance o f the speci e s . The s o i l cul ture method i nvo l ves potting seedl ings into vi a l s containing 2g of the b a s i c potting mixture to which was added nickel in the range from 1 00 - 1 0 , 000 �g/g ( fi nal concentration ) . There were five seedl ings per speci es per concentrati on . The seedl ings were watered f rom below in the s ame condi tions a s fo r the nicke l uptak e studi e s . The seedl ings were l e ft to g row until a stab l e su rvi ving di s tribution was obtained . The tol erance l evel for each species was taken as the concentration at whi ch at l e ast hal f o f the seedl i ngs survi ved . I n addition to thi s tol erance l evel f o r the speci e s , the tol e rance of certain indivi dual seedl ings above thi s l evel we re noted . A l l soi l s and seedli ngs were sampl ed for the determi nation o f n ickel a t the conclusion o f the expe riment . 7 . 2 . 5 Cob al t Uptake Studies 1 5 6 Because cob al t i s al so frequently enri ched in serpentine soi l s , the uptake of thi s e l ement b y A lys sum specie s is of intere s t . No specimens o f Alyssum speci e s anal ysed during the survey of th is genu s showed anomalous cobal t concentrations , t hus indicating that those species which hyperaccumul ate n i ckel do so by pre feren tial uptake of th is e lement rather than a general uptake of heavy metal s . However i t was decided to test the uptake o f coba l t i n a s i tuat ion in whi ch preferential uptake o f nickel woul d b e unl i k e l y . T h e b a s i c potting mixtu re ( wi t h low n i ck el content ) w a s u sed f o r t h i s tri al . Cobal t , as the n i t rat e , was added to thi s b a s i c m ixture to g i ve a range of concent rations from 30 - 3 , 200 pg/g . The trial was conducted by pl anting seedl ings in vi a l s co nt a ining 2g o f the experimental sub strates . Five seedl ings were pl anted per speci e s p e r concentration . The seedl ings w ere left t o grow fo r s i x week s before b o th the pl ants and the so i l s were s ampled fo r the determination of the i r cob alt conten t . Six speci e s were used i n thi s t r i al . 7 . 2 . 6 Analyti cal Methods Leaf metal concentrati ons w ere determined b y drying the l eaf s ample s i n an oven a t 80°C until a constant weight was obtained. Sub samples o f 0 . 0 1 - D . D3g were then weighed out for anal y si s . These were ashed at 500°C i n a muffl e furnace and the ash redi ssolved in 1 cm3 o f 2M hydrochlori c aci d ( prepared from redis ti l l e d constant-bo i l ing hydro chloric aci d ) . Thi s sol u t i on was anal ysed b y atomic absorp tion spectro photometry for n ickel o r cobalt as requ i red . The l i ne s u sed fo r anal y s i s w ere 2 32 . Dnm and 3 5 1 . 5nm fo r n i c ke l and 240. 8nm and 304 .4nm for coba l t . The instrument used was the Vari an-Techtron AAS mo del wi th automat i c b a ck g round co rrection a s in Chapter 2 . The soi l s were anal ysed after b e i n g dried i n a n oven a t 1 1 0°C . S amp les o f D . 1 g we re t h en d i gested wi th 2Dcm3 o f aqua reg i a and taken to dryness over a water-bath . The res i du e was then redi s so lved i n 1 Dcm3 of 2M h ydro chloric aci d , p repared as above , and centrifuged to remove the insolub l e materi al ( primaril y undissolved s i l i cates ) . The supernatant was diluted a s requ i re d ( 1 0- 1 00cm3 ) and then analysed for n i ck el o r cobal t b y atomi c ab sorption spectrophotometry . The instrument and the l ines u sed were the same a s fo r the l eaf anal y se s . 7 . 3 N I CKEL UPTAKE I N ALYSSUM SPECIES The resul t s o f the ni ckel uptak e studi e s are shown in f i g . 7 . 1 . The two non-accumulato rs , �· montanum and A . se rpyll i fol ium ssp . s e rpyll i fol ium , were e as i ly di stingu i shed from the o ther spec ies . Thei r rel ationshi p b e tween n i ckel i n t h e plant a n d n i ck el i n the subs trata i s l inear as i s common 1 5 7 1 0 000 - (/) � � � -o � s... -o c c 0 · - ...- a s... ...- c � u c 0 u � � u z 1 20 1 00 e A. argen teum 0 A . corsicum 0 A euboeum · A . heldreichii 0 A_. montanum 0 A. murale • A. serpy_llifolium 6 A_. serpyl/ifolium s. s p. lusitanicum e A. terium 0 A_. troodii 0 A: virgatum 1 00 0 10 000 N i ckel concentrat ion in substra te ( tJg/g ) £jgure 7. 1 N i cke l accumu lat ion in _A_Iy_ssum spec 1 e s as a function of substra te conten t . fo r non-to l erant spec i e s on comparatively metal -rich so i l s ( T imperl ey e t al . , 1 970b ) . The nine hype raccumul ators al l behaved s im i l arly to each o ther . For thi s reason fi g . 7 . 1 shows an i nclusive ( h atche d ) area for al l the data rather than indi vi dual curves for each speci es : the consi derab l e overl apping o f t h e curves wou l d b e confus ing . The shape o f the curve i s a ri se-to-satu ration fo rm . Thi s fo rm , characteri z e d by t h e rap i d l inear increase at l ow concentrations t o the p lateau or near-pl ateau a t h i gher concentrati ons , is prob abl y the l east common o f the three forms o f uptak e . The l i near form o f u ptake , a s shown b y the non-accumulato rs , i s b e l i eved to ar i se from an i nabil i ty o f the pl ants to exclude the e l ement so that the concentration in the p lant ri ses i n propo rtion to the i ncreasing concentration i n the so i l . Thi s i ncrease w i l l cont inue unti l t h e p l ant di e s from t h e tox i c effects o f l arge concentrati ons of that e l ement . The second common fo rm is the exclusion-break down form where at low concentrati ons i n the soi l the concentration in the pl ant i s held constant o r nearly- 1 5 9 so by an exclusion mechanism . At some hi gher so i l concentration th i s exclusion mechan i sm b reaks down al lowing rel ative l y free entry o f the e l ement i nto the p l ant . At t h is po int , the concentration in the plant tends to r i s e ve ry rap i d l y ( see Chapter 4 ) wi th death from e lemental po i soning occurring wi th l i tt le fu rther increase i n t he so i l concentration . The ri s e-to­ saturation form i s more di ffi cu l t to int erpret . The n i cke l content i ncreases rap i dl y at l ow soi l concentrations unt i l a p l ateau i s reached where , despite further increases i n so i l concentratio n , l i t t le o r n o i ncrease i n pl ant content occurs . I t appears t hat some exclus i on mechani sm i s operative a t h i g h rather than low s o i l concent rations . Th i s coul d b e i nterpreted as a need fo r high concentrations o f n i ckel fo r these spec i e s . Thus at l o w s o i l n i ckel l evels t h e pl ant has a h i g h concentrat ing ab i l i t y ( the accumul at ion i ndex i s much greater than one , see Section 7 . 4 ) i n an attempt to s a t isfy the plant ' s need. Once th i s need i s ful fi l l e d , the p l ant then excludes the entry o f any ni ckel above that requi re d b y new g rowth . It is now l eft to determ ine why these p lants will not survive i n h i gh e r nickel concentrations in the so il . Those pl ants whi ch show the exclusion-b reakdown form of uptake d ie after the exclusion mechanism i s overl oaded by the h i gh soil concentrat ion , resu l t ing i n l arge quant i t i e s o f the previously excluded element entering and poi soning the plant . This effect shows up as a rapi d increase in the elemental concentrat ion i n the plant at soil concentrations above the exclusion mechanism ' s break down level . The Alys sum speci e s , howeve r , d o no t show a ny sign o f exclusion-breakdown at t h e hi ghest nickel concentrations . I t may b e that the specimens which died a t hi gher soil concentrations o f ni ckel did not d ie from d irect nickel po i soning but from other factors which may o r may no t i nvolve the nickel ( e g . pl asmolysis o f the roots , c a tionic competition e ffec t s ) . Despite thi s l ine o f reasoning , the evi dence fo r a specific need o f n ickel by Alyssum spec ies does no t exi s t . Pl ate 7 . 2 shows the resu l t s o f these uptak e experiments fo r A. serpyll i fol ium ssp . serpyll i folium , �· serpyllifolium ssp . lusi tani cum , �· 0e l dreichi i and two A . murale fo rms . As wel l as the specimens growing in the potting mixture , a further specimen o f e ach spec ies i s shown g rowing in a New Cal e doni an serpentine soil . A . serpyll i fol ium ssp . serpyll i fol ium readily shows i ts low nickel tolerance by the reduction in growth of all specimens g rowing on n ick el-rich sub strates . Even more revealing i s t he very stunted growth o f the specimen in the serpentine soil . A. serpyll i folium s s p . l u s i t anicum showed very even g rowth i n all surviving specimens ( no specimen o f any spec i e s t ested survived in pot s containing 1 0 , 000 �g N i/g dry soil ) . The specimen i n serpentine so il i s only sli ghtly smal l e r than these o ther specimens . The small decrease i n s i z e is most probably due to facto rs o ther than nickel e g . low cal cium , ni t rogen , phosphorus or potassium o r high magnesium content which are common to serpentine soi l s . �· heldreichi i has much reduced growth in the b ack g round mixture . Thi s could b e taken as a l ack o f some e ssential nutrient a nd s i nce t h e o n ly di fference between 1 6 0 Pl a te 7 . 2 Grow t h o f some Alyssum - - -- s pQc i es in n i c kf2 1 uptake tr i a l s . Plan t s grow ing ( from le f t to r i g h t ) in back ground su b s t ra t e 3 0 � g /g ; 70 fj g /g j 1 2 0 � g /g ; 350 � g/g ; 1 ,050 1-Jg lg ; 2 � 680 fjg /g ; and Nczw Ca !Q d on ian sQrpen t inQ so i l 4 � 8 00 fJg /g . Ta x a ( from to p to b ot to m ) are : A_lyssum sergy_lli folium ssp. !usitanicum Afy_ssum ser{2y_llifolium ssp . serpy_llifo/ium A ly_ssum heldreichli' Aly_ssum murale - form B Afy_ssum murale - form A I P l a t e 7. 2 th i s mi x t u re 3nd the o thers i s the n i ckel content , it cou l d b e s a i d t o i nd i c ate a need fo r n i ckel b y thi s speci es . g c o t h e r l ow n i ckel concen t rations i n the sub strate , the p l ant has a s t raggl y growth but thi s reduces to a more compact form at hi gher n i ckel concentrations . The serpenti ne so i l spec imen a lso h as the compact g rowth form . The two A . murale forms show d i fferent behaviours w i th form A h aving increased growth in the b a ck g round mixture whi l e fo rm 8 has reduced 1 6 3 g row th . Abo ve the b ackground l evel , the g rowth o f i n d iv i dual s p ec imens i s much the s ame i rrespective o f the so i l concentrction of n i ck el . The serpent ine soil specimens here are al so onl y s l i ghtly small e r than the other specimens. The other spe c i e s te sted b u t n o t shown i n these photographs had simi l a r ranges o f behaviou r . Tab l e 7 . 1 l i sts the concentrat ions o f ni ckel in the l eaves o f those spec i es g rown in s erpentine so i l s ( e i ther t h e N ew Cal e doni a n , Dun Mounta in o r b o th ) . �I l l speci es , except the non-accumulato rs , hype raccumu l a ted i � t hese te sts . The h i ghest l e af concentration recorded was 1 9 , Q�n �g/g For �- a rgen teum growing in a sub strate wi th 9 , n�� � �/g . All hyperaccumu l a tors reached leaf concen t ra t ions of 8 , 01G - 1 5 , 000 �g/g fo r t h e sub s trates teste d . The l evel o f n i ck e l i n the so i l tol erated b y the hyperaccumulating spe c i e s always exceeded 3 , 000 �g/g but never exceeded 8 , 000 � g/g . These concentration s compare wi th total n i ckel concentrations i n s erpent ine soi l s o f approx . 5 , 000 pg/g o f which only 1 0% i s l ikely to b e ava i l ab l e to the plant (Lee et al . , 1 977a ) . Thus i t appears that these spe c i e s have an excess capaci ty i n thei r adaptation to n i ckel concent rati ons in the sub strate o r that the condi tions used ( i e . the nature o f the sub strate ) have had an ame l i o rating effect on n i ck el tox i c i ty . The two non-accumul ato r s , � · mont anum and �· s e rpyll i fol ium s s p . serpyll i fo l i um survi ved i n sub s trates wi th maximum nickel contents o f 420 �g/g and 350 �g/g respec tive l y . The correspond ing n i ckel concentrations i n the l eaves were 300 �g/g and 60 �g/g . The results a s far a s A . s e rpyl l i fol i um s s p . serpyl l i fol ium are concerned are p a rti cul arly s i gni fi cant . This species i s from section D dontarrhena so that i ts non-accumulat ion o f n i ckel TABLE 7 . 1 N i ckel upt ake b y Alys sum spec i e s grow ing i n serpentine soi l s . Species N i ckel Concentration )C Dun Hountain So i l New Caledonian B.· A . A . A . - A . A . - A . - Soil corsi cum hel drei ch i i mural e A B s e rpyll i fol i um serpyll i fo li um ssp . lusi t ani cum tenium t rood i i 904 4 4 038 9 1 38 1 3 3 1 1 4 7 r 2 56 1 0 n - 1 547 1 2 5 782 ( 0 34 �Al l concentrations expre ssed in �g/g dry we ight . 790 - 390 200 1 40 2 2 3 040 - - 1 6 4 So i l even when on n ickel-rich sub strates indicates that there i s n o genetic factor common i n a l l t a x a o f t h i s section which al l ows fo r hyperaccumul ation . 7 . 4 " TR IGGER-PO INT " THEORY TES T One question inherent t o the ri se-to-saturation form o f uptake i s whether the rise i s precerled by a l ow-l evel pl a teau . If such a pl ateau ex ists then some thing mus t i n i ti ate t h e r i se a nd thi s initi ator ( " trigge r " ) wou l d b e wo rth i nvestigati ng . To test whether such a " trigger-point " exi sts , two species o f Alyssum , �· tenium and �· troodii , were studied at low sub strata concentrations of nickel . The resu l t s of these studi e s are shown in f i g . 7 . 3 . The data have been plotted as the accumul ation indi ces rather than di rectly a s the nickel concentrations i n the pl ant b ecause such a pl o t shows the re su l t s mo re cl earl y . There i s an obvious d i fference in behaviour at very low nickel concentrations when compared to tho se at just sl i ght ly higher concent rati ons . However e ven at the lowest concentra ti ons s tudied , the accumul ation index i s al ready rising i e . the '' trigger-po int " ( i f i t ex ists ) i s at a concentrati on even l ower than 20 �g/g under these condi t ions . There is some c i rcumstanti al evi dence fo r a " trigger-po int " in that the curve is flattening out around and be low 20 �g/g . I t appe ars that a t very low nickel l evel s i n the sub strate even these hyperaccumulators may exclude· nickel from entry i nto the plant . Immediately the " tri gger-point '' ( o r i s it an exclusion-break down point? ) i s reached a very l arge influx o f nickel appears t o occu r . Certainly this type o f behavi our i s common t o spec ies whi ch show an exclusion-break down fo rm of uptak e . I t may then b e that thi s rise-to-saturation fo rm i s indicat i ve o f a second exclusion mechani sm which is tri ggere d by higher internal concentrations of ni ckel and it is the presence of this second mechanism whi ch g i ve s the se species thei r tolerance to excess nickel in the subs t rate . A l ternativel y , i t may b e that at the very low ni ckel concent rations , the nickel i s compl exed wi th the peat component o f t he p o t ti ng mixture and i s not 1 6 5 Q) - 0 '-- (/) ..0 :J (/) z ' - c _g p. z 80�------------------------� • 0 Ni plant = nickel concentration in leaf ( j.J g/g ) N i substrate = n i c kel concentrat ion in s u b s trate ( 1-J9 I g ) • 0 • 0 e A /y_ssum tenium 0 A.Jy_ssum troodii 0 20 40 60 80 100 1 20 1 40 1 60 N i cke l concentration in substrata ( fJ g/g ) _Eigure 7. 3 Plo t of the concentrati o n index a t low su bstra t e concen t rations of n i cke l . avai l ab l e fo r uptak e by the p l an t . Thus no l a rge influx o f nickel into the pl ant will occur until the compl exing capaci ty o f the peat has b een exceeded . An exclusion mechanism wou l d then only come i n to operation once h igh internal concentrations o f n ickel had been reache d . An equi valent mechani sm could e x i s t i n the copper-tol erant speci e s , B e ci um homb l e i , for which the ri se-to-saturation form o f uptake has al so been found ( R e il l y , 1 969 ) . To date 1 6 7 t h i s form o f uptak e h a s not b e en found for non-tol erant specie s . 7 . 5 RATE OF UPTAKE OF N ICKEL The rate of uptak e of n ickel by �· euboeum seedl ings i s shown in fi g . 7 . 4 . Afte r an ini t i al rapid i ncreas e , the concen tration o f n i ckel in the l e af reached a l imiting value . The l ength o f time b efo re the l im i t i ng value was reached vari e d . The vari ation may have been due to e i ther the n i ck e l concentration in t h e substrate , t h e nickel concentration reached in the l eaf o r some other unk nown facto r . The rate o f uptak e in the early l i ne a r r i s e was the s ame for both t r ials ( i e . t he two l i ne s are p aral l el ) . Fo r a n i ck e l concentration i n t h e sub strate o f 6 , 500 �g/g t h e l imi ting value was 850 �g/g whi l e fo r 7, 000 � g/g in the sub strate the corresponding value was 4, 000 �g/g . The value s fo r l e af concentrations , particul arl y the 850 �g/g , a r e lower t han wou l d b e expected o n the b a s i s of t h e n i ckel uptake t e s t s . I t i s possibl e t h a t t h e time di ffe rence in t h e tests , e i gh t day s compared to s i x week s , i s the cause . There may b e a constant ri s e for a time after the ini t i al rapi d rise rather than a complete l evell ing off o f nickel content . I t i s c l e ar , however , t hat the uptake o f n i ckel by t h e se speci e s , and p re sumabl y by the other hyperaccumulating Alyssum speci e s , i s a rel a t ive l y rapi d process . 5000 en 6,1000 :::1 - c: c: 0 ...,._ a s..... ...,._ c: � u c: 0 u 1 0 0 £i.g ure 7. 4 1 2 3 Days after n icke2 l - r i c h The ra te A lx_ssum t 4 5 transplanting so i ls . N i c k e l i n s o i l 6 7000 1-J g /g • • 0 6500 f-J g /g 0 7 8 to of uptakC2 of n i cke l by euboeum. ' 7 . 6 N I CKEL TOLERANCE STUD IES The resul ts o f the solution roo t i ng to l e r ance t e sts are shown i n fi g . 7 . 5 . The four species t e s ted showed s trong tol erances to n i ck e l w ith root growth ceasing ( l ength 3 o f new root l e s s than 0 . 1 cm ) a t b e tween 1 5 � g/cm fo r A . tenium a nd 60 �g/cm3 for �· troodi i . The non-accumul ator �· serpyl l i fo l ium ssp . serpyl l i fo l ium does no t g row roo t s in any o f the solutions containing n i ckel . It doe s howeve r show s t rong roo t g rowth in the b ackground ( no n i ckel ) s o l u t i on . T he r e sul t s o f t he s o i l cul ture tol erance t est a re g i ven i n t ab l e 7 . 2 . A l l the spec ies teste d , except �· serpyl l i fo l ium s sp . se rpyl l i fol ium , tolerated a t l e ast 1 , 000 �g N i /g i n t he soi l . Among t he hyperaccumulators � · corsicum had a surprisingly l ow tole rance l imi t : i ndeed t he l im i t of 1 , 6 50 pg/g was much lower than in the n i cke l 1 6 9 uptake stud i e s ( 3 , 000 �g/g ) . A l l other hyperac cumu l ators to lerated b e tween 3 , 200 �g/g (�. serpyl l i fo l ium s s p . l us i tani cum ) a n d 5 , 000 p g /g (�. troodi i ) a s expected . I ndivi dual s p e ci�ens coul d survive i n n ickel concent rations of up to 7 , 200 �g/g ( A . troodi i ) . One specimen o f �· serpyl l i fo l i um s sp . s e rpyl l i ­ fol ium was able to survive in over 1 , 000 �g/g ( ac tual value 1 , 4 30 � g/g ) . Thus whi l e t h i s t axon does not normal l y g row i n serpenti ne soi l s , i t i s certainly capab l e o f producing i ndividual specimehs whi ch can survive i n t he r ange of avail ab l e n i ckel concentrations found i n them . I t i s thus presumabl y capab l e of g iving r i se to sub spec ies ( b oth �· serpyl l i fol ium s sp . l u s i tanicum and ssp . mal a c i t anum ) more tolerant of these so i l s b y methods o f gene s e lect ion ( Antonovi c s et �. , 1 97 1 ) . A l l specimens , except thos e of �· serp¥l l i fo l ium ssp . s e rpyl l i fo l ium , showed hyperaccumul ation . I t i s not i ce ab l e that i nd ivi dua l specimens of �· t rood i i and A. serpyl l i fo l i um ssp . l u s i t an i cum a ccumulated l e s s n i ckel a t concentrations above t h e tol erance l evel than those at o r below thi s l evel . I t appears that these i ndivi dual s h ave a more effe c t i ve exclusion mechanism a t these h i gher n ickel 1 00 •----------------------� E E - ..c. � CJ) c Cl) � 0 0 L.. c a c::l � 1 0 0 ·1 0 20 0 A . corsicum 6 A. serP.x_llifo/ium s. sp. lusitan icum e A. tenium 0 A. troodii 40 60 80 N ickel concentrat ion in substrate ( 1Jglcm3) £.igure 7· 5 To lerance tests involv ing new ­ root lengths of exciSC2d seed l in gs of AJy_ssum species grown i n vary ing n i ckel solu t ions ( IJ g /cm3 ) 1 7 1 TABLE 7 . 2 .n. • A . - A . - A . - A . - A . - A . - To l e r ance l evels of Alys sum spec i es tested by the so i l cul ture metho d . Species Tol erance Level H Individual Survivo rs H corsi cum 1 6 50 ( 5 793 ) - heldre i c h i i 4 440 ( 1 5 7 30 ) - mural e 3 950 ( 1 0 790 ) 4430 ( 1 3800 ) 5060 serpyl l ifo l i um 675 ( 380 ) 1 4 30 ( 805 ) serpyl l i fol ium ssp . l us i tani cum 3 230 ( 7 7 55 ) 3790 ( 66 3 6 ) ten ium 3 290 ( 5 475 ) 6 1 20 ( 754 4 ) trood i i 4 970 ( 1 9 200 ) 7 1 68 ( 1 8550 ) HAll concentrati ons expressed i n pg/g dry substrata . The corresponding l e af content i s g iven i n parenthe ses as �g/g dry l eaf . ( 1 4 1 50 ) concen t r a t ions i n the sub s t rata than i s gene ral for the spec i e s . When the four spe c i e s common to both methods o f measu r i n g tol erance have the i r resul t s compare d , i t i s appare n t tha t n o common p a t tern emerge s . Thus i n the so l u t i o n roo t i ng method A . t rood i i was mos t tolerant and was fol l owed b y �· corsi cum , � serpyll i fol ium s s p . l u s i tanicum and A . t enium in order . The i r order in the soil cul ture method was A. troodi i (mo s t tolerant ) , �- tenium , �- serpyll i ­ fo l ium ssp . l usi tani cum a n d A . corsi cum ( l east tolerant ) . Th i s d ifference indicates t h a t the tolerance of a speci e s to a p a r ti c u l a r substrata is dependent on more than just the n i cke l conte n t . Many factors have been consi dered a s causes o f s e rpent ine inferti l i t y a nd i n most such so i l s several f a c tors are undoub tedly a c t i ng together (Walker , 1 954 , Proc tor & W o o del l , 1 97 5 ) . Thus whi l e the solution roo t i ng method may have measure d the d irect effect o f n ickel toleranc e , the so i l cul ture me thod may h ave g i ven an indication o f the o v e r a l l effects o f n i ck e l and o ther nutri ents . U ndoub tedl y , i f th i s i s so , the l a t ter m e thod would g i ve a b e tter indicat ion o f the ecolog ica l tol erance o f the spec ies . W i l k i ns ( 1 978) d i s cusses this probl em i n re l a tion to he avy metal tole rance s m ea sured b y means o f roo t g rowth s . More b as i c research into t h e vary i n g t echniques o f measuring t o l erance i s however s t i l l necessary . 7 . 7 COBALT UPTAKE S TUDIES The resu l t s o f the cobal t uptake s tu di e s are shown in f i g . 7 . 6 . The resu l t s show a remark ably simi l a r pattern to the n i ck e l uptake s tudi e s . The major d i fference s are the 1 7 2 l ower soi l concentrat ions o f cob a l t tolerated ( no spec i e s s urvived over 1 , 000 � g Co/g compared w i th 2 , 000-8 , 000 � g N i /g , ) , t he lower l e af concentrations reached ( 1 , 400- 5 , 000 �g Co/g comp are d to 8 , 000-20 , 000 � g N i / g ) and the l ower so i l concentra t ion a t whi c h a rap i d i ncrease i n uptake occurs ( mi d-po i n t fo r cob a l t approx . 1 0 pg/g compa red to approx . 50 pg/g ) . -(/) Q) > 0 � - 'U � L.. 'U c: · - c 0 ........, 0 L.. ........, c: Cl u c 0 u ........, 0 ...a 0 (_) 1 00 1 0 _A. corsicum A . heldreichii 0 A . mura/e 6 A . serpyllifolium s . sp. - - - lusitanicum e A. tenium 0 A . troodii 1 00 1 00 0 Cobalt concentrat ion in substrat<2 ( � g /g ) Figura 7· 6 Cobal t uptaka b y A ly_ssum sp<2c1e s . The l east tol erant spe c i e s studied was A . t rood i i which su rvived only up to 260 � g/g in the so il . The most tolerant species were �· heldrei chi i , �· murale and �· serpyll ifo l ium s s p . l u s i t an i cum which tolerated 7 1 0 �g/g . The h ighest l e af concentrations were found i n � · tenium ( 4 , 83 3 �g/g ) and A . corsicum ( 4 , 1 44 pg/g ) . I t i s notable t h a t a l l the speci e s s t ud i ed for cob a l t uptake hyperaccumulated t h i s e lement de s p i t e the fac t that in the w i l d they show no such incl ination to do so . As i t i s apparent from thi s s tudy t ha t a number o f A lys sum t axa are capabl e o f hyperaccumul at ing cob a l t even though t h i s has not been ob s erved in the wi l d p opul a t i ons , i t i s necessary to exp l a i n t h i s di fference in b ehaviour . The cob a l t content of serpentine soi l s i s general l y only one t enth t ha t of n i ckel . Thus n i ckel coul d b e expected , under c ond i tions of e ither uncontro l l e d or general entry , to be at t e n times the cob a l t concentrat ion wherea s it is found to be a t several thousand times that concent ration . Thus the entry o f these me tal s into the p l ant c anno t be uncontrol l ed i e . they c annot move passively through the roo t epi dermi s into the c y topl asm . Nei ther c an the entry mechanism be general since i t obviou s l y favours accumul a tion of n ick e l over the o ther e l emen t s which include cobal t . The cont ro l mechan i sm fo r the entry o f n i ckel i s however an unknown quant i ty but i t i s c l early ab l e t o operate in the absence o f h igh concentrations of nickel p rovi ded that h i gh concentrations o f an al ternat ive metal are present . A t low concentrations of metal e lement s , 1 7 4 the mechan i sm does no t appear to operate . I t may be that t h i s mechani sm i s concern e d w i th the entry o f metal s i n t he complexed s t a te . Thus one way o f a ccumul ating n ickel r e lat ive to cob a l t wou l d b e t o p roduce an o rganic l i gand which compl exed preferent i a l ly with ni ck el . I n the absence o f n ick el , th i s l i g an d c a n then complex w i t h o ther met a l s to g i ve them entry into the cytopl asm . Some evi dence for this can b e seen in the l evel s to which the respe c t i ve met a l s can b e concentra te d . Thus nickel w i th a mo re s table compl ex i s concentrated to a h i gher degree than cob a l t with a l es s s tab l e compl ex . Thi s a ssumes t h a t the h e avy metal in the free ionic s t a t e i s invol ved a s a cause o f death o f t h e p l ant . L e ss s tabl e complexes w i t h correspondingly greater free met a l ion co ncent rat ions wi l l t here fore b e accumul ated to a l e sser e x tent than the more s t ab l e compl exes b efo re death occurs . Whi l e t h i s hypothesis c an exp l a i n the d ifference i n the 1 7 5 amounts o f the metal s t ak en u p , i t does not account fo r the p l a teaux . E ach p l a teau p re sumes that a t some h i gh concentration free entry of the compl ex into the root is inhib i ted as the soil concent ration con t i nues to r ise . The cause of t h i s i nhibi tion m a y b e t h a t a s t h e free i o n content wi th in t h e p l an t r ises , t h e p l a n t i nvok e s some mechani sm to prevent the l evel b ecomi ng toxi c . This mechani sm may well involve t h e poi.sonlng o f the enzyme ( s ) which p roduce the l i gand by the free ions . Thu s a s the free ion concen t ra tion rise s , l i gand p roduction i s reduced and the amoun t o f metal c ompl exed a nd hence g i ven entry to the p l ant is al so reduced . O ne further p o int must be s t rongl y consi dered : the free metal ion may al so b e an ac tiva t o r , a t l ow concentra tions , for these e n z yme ( s ) s ince the u p t ak e o f the metal to h igh concen trations a ppears to have a " tr igger-poi nt " . Thus with this considera t i on , i t appe ars t h a t low free ion concentrations may a c t ivate the e n z yme ( s ) but that h i gh free i on concentrations poison them. This propo sed mechani sm is summariz e d as below : SO IL M2+ Li gand PLANT L i g an d� Sub strate L Enzyme /ac�i va�ion o r pO l SOnlng ( M-L i g and ) ----� (M-L i g an d ) � M2,.. � � To s to rage From thi s diagrammat ic repre sentation of the uptake mech an i sm , i t i s e a s i l y seen how the stab i l i ty o f the complex (M-L i g and ) contro l s t he amount o f uptake o f the various metal s . A s cobalt i s considered l e s s toxi c to p l ants than n ickel ( A g arwal a e t al . , 1 977, Hunter & Vergnano , 1 953 ) , and as Mathys ( 1 97 5 ) has shown that coba l t i s general l y l e s s toxic 1 7 6 to en zymes than ni ckel , i t woul d appear that the c ob a l t complex i s mu ch l e s s stab l e than t h e ni ckel complex . However cob a l t may be a more effec t ive acti vator o f the en zyme ( s ) i nvol ved a s the rap i d accumul ation o f cob a l t occurs a t a l ow e r concentration than for n ick e l . A l ternati vel y i f the 11 t ri gger-po int " i s due t o the complexing capacity o f the p e a t b e i ng exceeded , then the rapi d accumul ation o f cobal t coul d b e due to the l ess effec tive compl exation o f thi s e lement by the pea t . T h i s less effecti ve c omplexation ( re l a t ive to n i ckel comp lexa t i on ) by the peat wou ld l eave more free cob a l t ions ava i l ab l e fo r compl exation b y the l i gand produced by the p lant . Thus t he cob a l t woul d b e t aken u p at l ower total soil contents th an the n i ck el . Much further work will b e needed b e fore these proposed mechanisms are e ither accepted o r rejectsd . 7 . 8 BDRNMUELLERA TYMPHAEA STUD IES Bo rnmuellera tymphaea ( H au s sk n . ) Haussk n . is a membe r o f another g enus of the tribe A lysseae and i s thus c losely r e l a ted to the Alys sum genu s . I ndeed g. tymphaea i ncludes A lys sum tymphaeum ( Hau sskn. ) Held . & Haussk n . ex Formanek amongst i t s synonyms . Thi s speci e s is found in northern G reece on the L i ngo s R ange and surrounding are as o f the P indus Mountains and further east on M t . Vourino s . g. tymphaea i s s e rpentin e-endemic and both areas in w h i ch i t i s found are noted fo r thei r serpentine-flora . S eve ral Alys sum spec i e s i ncluding A . mural e a n d A . heldreichii are al so found i n t h i s a re a . The resul ts o f the ni ckel uptak e study on t h i s spe c i e s i s shown i n fi g . 7 . 7 which a l so g i ve s a compari son of thi s uptake to those o f the Alyssum species previ ously studi e d . I t i s e a s i l y recognized t h a t t h e uptake o f n i ckel i s i denti cal i n form and s imilar i n amount to the se o th er n i ck e l hyperaccum­ u l a to rs . Thi s f inding ra ises the question as to the e xtent to whi ch t h i s form of uptak e i s w ide spread . I s it confined to tribus Alysseae? I t w i l l be most i ntere s t ing to s tudy other ni ckel hyperaccumul a tors o f the Cruc i fe rae . In particular Peltaria emarg inata o f the tr ibe Lunarieae i s most wo rthy of study s i nce Hayek ( 1 9 1 1 ) and J anchen ( 1 94 2 ) have p roposed the amalg amation o f the tribes Lunarieae and Aly sseae . O ther species whi ch would also be worthy o f s tudy are T h laspi s s p . o f t h e t r i b e L e p i dieae a nd Streptanthus polygalo i des o f the Streptantheae . This l a tter spe c i e s being N ew Worl d i n di stribut ion , whereas t h e o thers a r e al l O l d W orl d , may show whether the geographical l ocation h as an e ffect on the form o f uptak e . 1 7 7 -(/) CJ > 0 CJ - ""0 CJ '­ ""0 c: c: . a � 0 '- -+- c: � u c 0 u 1 0 lm Bornmuellera j Y!!Jp_haea 1�------�----------�----------� 20 1 00 1 000 1 0 000 N i cke l concentration in substrate ( J.,Jg/g ) _figure2 7· 7 Nickel accumulation in Bornmuel/era _ fYJI1{2haea, as a function of substrate conte2nt in comparison with the Aly_ssum species in fi gure 7 · 1 . CHA PTER 8 Phyt och<2mica l Stu d i e s on Some Afy_ssum Sp<2c ie s I 8 . 1 IN TRDDU C T ID� The nickel w i t h i n p l an t s o f t he genus Alyssum is mos t concentrated i n the l e aves . Thi s fa c t w a s fi rst noted b y Mi ngu z z i and V ergnano ( 1 94 8 ) i n t h e o r i g i nal paper on hyperaccumu l a tion o f ni ckel b y �. b e r to l on i i . They further noted that seE!ds , fl mJJe rs a nd f ru i t s t ue re a l so ve ry hi gh but t h a t the cnncen t r � t i ons i n the roots were low when compared to the r e s t of the p l a n t ( a l though sti l l ve ry h i g h when compGred to other p l an t speci e s ) . Tri s has been confirmed by a sub sequent s tudy ( IJ e rgnano Gamo i et nl . , 1 977 ) . Th i s l at ter study a l so showed t h � t very young l eaves a t t h e b e g i nn i n g of a new growing season h a d l e s s ni ckel t h3n the o l der l e aves of the p rev iou s season bu t t h i s d i f f8 rence d i sappea red w i th i n 2 - 3 weeks o f g row t h . S t 8�S a f A . n o r t � � n n 1 � ,�pe a red t o have n i ckel l eve l s s i m i l a r to tho se o f t he f l nwers and fru i t s . V ergnano �amo i ( 1 9G7 ) b e � � n s�cc i f i � rhy tnchsmi c a l i nve s t i g a l;i ons o f ni ck e l h y p e ::- <:J c c:umu l a t ors b •J 'ih < JttJ i ng t h e di s trib u t i o n o f n i c k e l w i t h i n � he s tem t i s s� c � J f � . b e r to l o ni i . U s i ng dime thy l g l yo x ime as a s t a i n , the n i c� e l was found to be preferential l y l oca ted i n t h e e p i de rmi s and scl erenchymat i c ti s sues between the vascular bundl e s . S tudies on the d i s tribution of n i�k el wi thin Hyb anthus fl o ribundus t i ssu g s h ave shown that h ere t o o nickel i s a ssoc iated w ith ep idermal cel l s (Farago e t �. , 1 97 5 ) . 1 8 0 S tud ies o f the d i str ibut ion o f n i ckel w ithin the c e l l have b een done b y S haw ( 1 980 ) . S he showed by d ifferent i al centr ifugat i on .tha t Alys sum serpyl l i fo l ium ssp . s e rpyl l i fol ium ( a non-accumul ator ) and �· serpyl l i fol ium ssp . m a l a c i t anum ( an a c cumul ator ) have di fferent n ickel d i s t ributi ons . The a ccumul ator had the majority ( 72% ) o f the nickel in the supernatant indicating a ready s o lub i l i ty fo � the ni ckel c omplex . Thi s is e a s i l y expl ai n e d i f the nickel found is w i th i n t he vacuol a r s y s tem . The non-accumul ator however h ad o n l y 34% of the n ickel i n the superna tan t . A g reater perc entage o f n i ckel ( 36% ) was found w ith in the r e s idue w ith a fur ther 27% found wi thin the cell wall frac t i on . The compl exation of he avy m e t a l s by the c el l wall h a s b een fr equentl y postulated ( Pe te rson , 1 96 9 , Reilly et al . , 1 970 , Turner & Marshal ! , 1 97 1 , 1 972 ) and may well play a s igni fi cant role in the complexation o f metal s i n anomalous concentrations but i t appears that more speci f i c compl exations a l so occur at the h i gher anomalous concen t rat ions . T h is may well be neces sary s i nce the cell wall has obvious rol es in the movement o f various physiologi cal compounds and ex cessive compl exation wi thin the c e l l wall coul d undoub tedly i nterfere with these p rocesse s . O rgani c acids appear t o b e the most common complexing agent found wi thin hyperaccumula to r plants . I n �· bertolon ii , o rgani c a c ids were f i r s t postul ated to b e the comp l e x i ng agents for ni ckel b y Pelosi et 9�· ( 1 974 ) but i t was not fo r a fu r ther two years ( Pe l o s i e t al . , 1 976 ) th a t mal i c and malonic a c i d s were a c tual ly i denti f i e d . � t h i rd o rg ani c 2 c i d , found in the greate s t quanti t y , was also p re sent � u t cou l d not be i denti fi ed . Pancaro e t al . , ( 1 977 ) confi rmed the invol vement of mal i c and mal onic a c i d s i n the n i ckel complexation w i th in A . b ertolon i i a nd extended t h e rese arch to �· se rpyll i fol ium s s p . l us i tan i cum concl uding that mal i c acid was a l so i nvolved i n n i ckel complexation here . Lee e t al . , ( 1 978) sub sequentl y confi rmed thi s i nvol vement . Shaw ( 1 980 ) investi gated the n i ck el compl exes of �· s e rpyll i fo l ium sub speci e s and confi rmed the i nvolvement of mal i c a c i d in �· serpyll i fol ium ssp . l u s i tani cum . In �· s e rpyl l i fol ium ssp . mal a ci tanum , malon i c 1 8 1 a c i d was found to b e the domi nant complexing agen t . Small amounts of c i tr ic acid were a l so found to b e involved. This is the f i r s t report o f ci tr ic aci d invol vement in n i ck el complexation w i th in A lys sum taxa al though this acid i s a common compl exing agent in New Caledoni an hype raccumulators (Lee et al . , 1 9?7b , 1 978 , K erste n , 1 979 ) . Tha t o rganic a c ids should b e t h e complexing agents may b e partly expl a ined b y t he ob serva t ion tha t ni ckel can s t imul ate o rg ani c acid p ro duct ion ( Dek ock & Mor r i son , 1 9 58 ) . To r i i and L a t ies ( 1 966 ) have a lso no ted tha t o rgan ic ac ids a re synthes i zed to b a l ance excess c a t ions absorbed b y p l ants . N i ck el content in hyperaccumul ating Alyssum pl ants has been l i nked w i th calcium an d magnes ium contents ( Vergnano Gambi et �. , 1 977 , Shaw , 1 980 ) al though such s tat i s t i c al analy s e s as h ave been done have vet to show any rel a t i onships invo l ving these three el ements . Shaw ( 1 980) showed t h a t n i ckel had no s t rong co rrel ations wi th other metal elemen t s i n Alys sum speci e s from sect i on O dontarrhena . The l ack of rel at i onships between ni ckel and other metal elements i s also known from New C al edonian hyperaccumul ators ( L ee et al . , 1 977a , Lee , 1 977 ) and i t may b e that the factors wh ich control nickel accumula t i on a re e i ther o rgani c rather than mineral o r external to the plant such that mineral analyses cannot reveal the rel a t i onshi p . I n t h i s chapter work i s p resented on the d i stribution o f n ickel wi thin specimens o f three Alys sum speci e s . Seve ral inve s t i gations a re also reported for mineral and organic componen t s i n an attempt t o d i scover any rel ationsh i p s wi th n i ckel conten t s . Lastl y , nickel comp lexes were i so l a ted from e ight Alyssum speci e s and from B ornmuell era tymphaea w i th the intention of i dent ify ing the compl exing agents for n i ck el fo r each o f these speci e s . 8 . 2 EXPERIMENTAL METHODS 1 8 2 All p l ant specimens used i n these experiments were g rown from seed col l ected by various peopl e , as l i sted in Appendix I (b ) . The seeds were germinated on a Copenhagen table and then p lanted i n to the appropriate experimental p ots . The pot t i ng mixture o f 50-50 peat-perl ite w i t h added nutri ents and n i ckel ( as the n i trate ) was used in all c a se s . The pots were kept i n a 0 glasshouse a t 20-25 C and watered from beneath. 8 . 2 . 1 N i ck e l D i stribution W i th in the P l ant O rgans Specimens o f three speci es ( A . hel dreichi i , A . mu rale - - and A . s erpyll i fol ium ssp . l us i tani cum) were grown in the p o tt ing mixture made up to 1 , 000 � g N i /g . A fter fi ve months , these specimens were h a rve sted , washed and pressed . When dry , they were divi ded i nto ten parts ( s e e f i g . 8 . 1 ) : the api cal bud ; the uppe r , mid and l ower s tem ; the upper stem l e ave s , wh ich are bo rne directl y on the upper s tem ; the mi d-stem l a teral s , in w h i ch the l a teral growths are devel oping stems ; the lower stem l a teral s whi ch were d i vi ded into stems and l e aves s ince the i r g rowth is mo re advanced here ; the upper roo t s , wh ich were t he l arge r , coarser root s ; and t he l ower roots whi ch were smal l e r and f i ner al though longer t h an the upper root s . The pods and seeds from the samples sent to u s were also anal ysed . Samples o f e ach o rg an were w e i ghed and then a shed at 500°C i n a mu ffle furnac e . The a sh was di ssolved i n 1 cm3 of 2M h ydrochl ori c acid and anal ysed b y atomi c absorption spe c t ropho tome t r y . The spectrophotometer used was the Vari nn­ Techtron model AA5 w i t h au tomati c b ackground correction 3CS attachment as u sed throughout th i s wo rk . The n i ck el l i ne used was at 3 5 1 . 5 nm b ecnuse o f the h i gher concentrations of ni ckel found in these specimens . 8 . 2 . 2 M ineral and · D rgani c Component Analyses Fo r the se analyse s , the spec imens g rown i n the ni ckel uptake trials ( Chapter 7) for three spec ies , the non-accumul ator A . serpyll i fo l i um ssp . s e rpyl l i fol ium and two hyperaccumul ators A . serpyll i fo l ium ssp . lusi tani cum and A . mural e , were harve sted and thei r leaves free z e -dri e d . Thus for each spec i es , a s e r i es o f nickel concen trations w a s ava i l ab l e for compari son purposes. These freeze-dri ed l eaves w e�e used for all the component anal yses . 1 8 3 ( a ) Mineral Compone n t s . App ro x . O . OSg o f the freeze-dri ed material was 3Shed and d i ssolved in 2M hydrochl o r i c a c i d ( 2 cm3 ) . This s o lut ion w a s u s e d di rectl y fo r i ro n , c op p e r , z i nc a n d manganes e . F o r calc ium and magn e s i um the solution was d i l uted to one-tenth w i th 0 . 8% strontium n i t rate in 2M hydrochl o ri c aci d . Fo r so dium and potassium , a further d i l ut ion to one-fourth was made u s i ng 2M hydrochl or i c aci d . A l l el ements except so dium and p o t assi um were ana l ysed b y a tomi c abso r p t i on spec tropho tometry us i ng the Vari an-Techtron AAS w i th au toma t i c b ac k g round co rrec t i o n . The l ines used for the anal y s e s wer� iron 248 . 3nm , cop p e r 324 . 7nm , zinc 2 1 3 . 9nm , manganese 279 . Snm , cal c ium 4 2 2 . 7 nm and magnes i um 285 . 2nm . S o dium and pot assium were anal ysed b y a tomic emi s s i on spectro scopy u s i ng a Gal l enk amp mo del EEL 1 00 fl ame pho tome ter . All resul ts are g i ven on a dry weight b as i s . ( b ) O rg an i c Component Anal y ses . n nl y two o rgani c components uJe re an gl y sed . These we re t h e o rg ani c ac ids and the glucosinol R t es . The o r g a n i c ac i ds we re e x trac ted b y shak i ng t h e d r i e d l e a f m a t e r i a l wi th 80% e thanol ( 0 . 2g i n 5 cm3 ) fo r 2 hours . T h i s w as then cen t r i fug ed an� the 3upe rnatant decanted . The residue was washed wi th successi ve por t ions of water and 50% ethano l . The comb ined supernatants were f il tered and reduced i n vol ume b y ro tary evaporation at 3 5°C . The solution was clean�ed b y passage through a 1 0mm x SOmm cat ion exchange column , using Amberl i te I R- 1 20 ( H+ ) resi n , which dripped d irectly onto a 1 0mm x 1 00mm anion exchange column , u s ing Dowex 1 - X8 ( format e ) . Thi s l at t er resin exchanged wi t h the o rganic aci ds which were then e luted with 2 5cm3 of 20% form i c aci d , 2 5 cm3 o f 50% formi c a c id a n d final l y washed with deionized water . The el utant was then dried b y rotary e vaporat i on a t 3 5°C . The d r i ed product was redi ssolved in 1 cm3 o f de ioni zed water and u sed f o r HPL C . The HPLC instrument u s e d w a s a W aters Associates model 660 solvent p rog rammer coupl ed to a p-bondpak C - 1 8 column. The buffer used was 2mM t e tra-n-butyl ammonium phosphate a t pH 2 . 8 . 1 8 5 The fl ow rate w a s 1 . 5 cm3/mi n . The � e t e c t o r was a Cec i l CE 2 1 2 A spectrometer set at 220 nm . The gluco s inolates w e re anal ysed b y the me thod o f Schul t z and Gmel i n ( 1 954 ) . T h e process began wi th the soxhlet extract i on o f 1g o f dried l e a f mate r i al with 50 cm3 o f methanol . The extraction w a s continued until the solvent w a s colourl e s s . The methanol w a s d i st i l led o f f a n d t h e res i due redi s so l ved i n deioni z e d water . Thi s sol u t i o n was f i l tered and made u p to 1 00 cm3wi th further deion i z e d 0ater . Fo r the specimens which d id not have 1g o f dri ed leaf mater i al , t he process above was carr i e d out u s i ng D . 1 g and having all o ther amounts at one-tenth of those abo ve . A 1 Dcm3 a l i quot was then taken for aci d-alumi na chromatography ( the column conta i ned 5g o f ac id- Al 2D3 ) . The al i qu o t was run through the column wh i ch was then washed w i th 20 cm3 o f deion i zed water to remove the non-glucos i nol ates . The g l uco s i nol ates were then e l u ted w i t h 0 . 1 M p otass ium hydro x i de unti l the yel l ow col our was no longer evi den t . The elutant was made 3 up o r tak e n down to 1 0 cm • An a l iquot o f 3 5 cm was then tak en and p laced i n n tes t- tube surrounded b y 3 mel t i ng i ce and 1 0 cm o f anth rone reagen t ( D . 2g o f anthrone i n 1 00 cm3 of concentra t ed su l p�ur i c aci d ) was added careful l y t o g i ve two l ayers . These were mi xed b y gentl y b lowi n g a i r through t h e solu t i o n . O n c e mixed t h e tubes we re pl aced i n a vi gorously b oi l i ng water-b a t h fo r 1 0 min + 1 5 s . They were then cooled in a cold water b a th and the blue Bolour determined photometr i cal l y at 620 nm on a Spectronic 20 spec trome ter . Thi s metho d measures the glucose o f the glucosi nolates s o that the standards used c an b e made u s i ng glucose solutions . To g et an e st imate o f t he amount o f g luco sinolate s , thi s resul t i s mul t i pl i ed b y 2 . 4 a s an average conversion facto r , o n a wei ght b a s i s ( Schul tz & Gmel i n , 1 954 ) . 8 . 2 . 3 N i ck el Complexati o n The n ickel compl exe s were extracted from the dri ed l e a f mater i al b y t h e method o f L e e et �· ( 1 977b ) adapted to l ower amounts of material . Thus D . 5g of the material was shaken for two hours w i th 5 cm3 o f d e i on i z e d water to extract the compl exes . Thi s p rocess was then repe� t e d one mo re t ime . Ths cnmb i nc d superna t an t s were fi l t e red �nd c l eansed o f l i p i d s , p ro t e i n s and vari ous o t h e r compounds by e x tract ion wi th a 1 n : 1 chlorofo rm : n-butanol so l u t ion . The c leans ing was co n s i d er e d compl ete when no further precipi tation o c curred at t h e i n ter­ face . The aqueous solution was then f il tered and reduced in volume . The so l u t i on wa s run t h rough a 50cm x 1 . 5 cm Sephadex G - 1 0 gel fi l t ration column . The elutant was coll ected i n 4 cm3 fractions i n whi ch the n i cke l was located b y use o f atomic ab sorption sp e c t rophotometry . T h e ni ckel-contai n i ng fractions were recomb i ned and recycled excepting that fractions of the minor second peak ( p re sumed to b e aquonickel ( I I ) but too smal l to i denti fy ) were omi tted . The comb ined fracti ons from the second run were dri ed . Thi s product was then deriva t i z e d fo r i dent i fication by gas c hromatog raphy and mass spectrome try . I t was dec i d e d to u s e the �ethyl derivati ves 1 8 6 s i nce they are e as i e r to handle and s t o re than the trime thyl s i l yl deri vatives . The comp l e x w a s i n i t i a l l y de s t royed b y the addi t i on o f a few drops of 2M hydroch l o r i c a c i d . Thi s 2 c i d was then remo ved I J n d e r 3 s t r aam o f a i r � nd the res i du a d i s sol ved i n red i s t i l l e d d i e t h y l P. th � r . D i � z ome thnne in e t h e re a l solu t i o n , made b y t he me thod o f J erne r ( 1 9 1 9 ) , was added dropwi se until the y e l l ow co l ou r pers i sted . The e t h e r was removed b y fl ushing the vial w i th ni tro g e n . The residue was di ssolved i n redisti l l ed chlorofo rm and used for the anal yses . The GLC was pe rformed o n a Pye model 1 04 gas chromatog raph using a 2 . 8m x 4mm column containing 3% SP 2340 l i qui d phase on a supe lcopo r t 1 00-220 mesh PB 34 suppo rt . Thi s column was operated a t 1 80°C wi th n i trogen ( 30 cm3/mi n ) as the carrier g a s . The de tecto r w a s a n a i r-hydrogen ( 350 cm3/mi n , 3 0 cm3/mi n ) fl ame i oni z a tion detecto r . L o w resolution mass spectra were recorded o n e i ther a n AE I MS 3 0 dual-beam spectrome ter o r a VG M i cromass 1 2F s i n g l e­ beam spec trometer . B oth spectrometers w ere coupled to GLC instrument s : the MS 30 to a Pye chromatograph equ i pped w i th an DV 1 7/2 1 0 mixed l i qu i d phase c o lumn ; the Mi cromass 1 2F to a Vari an-Techtron 1 700 ch romato g r8ph w i th an SP 2 340 l i qui d phase co l umn . Both ch romatog raphs were t emp erature prog rammed u p to 2 50°C at 4-5°C/m i n . A h i gh resol u t i o n measu rement was made on an AE I MS 9 spe c t rome ter . A chemi ca l ioni z ation mass s pectra l s can was made o n the M i cromass 1 2 F spectrometer using i sobutane as the reactan t gas . 8 . 3 NICKEL D I STR I BU TION W I THIN THE PLANT ORGANS A di agram o f the various o rg ans analysed i n i s g i ven as fi g . 8 . 1 . The resu l t s o f the analyses i n t ables 8 . 1 - 8 . 3 . I t i s seen that the g reatest o f n i ckel occurs i n the l e af mater i a l and the least i s i n the roo t s . I t i s fur ther no t i ceable that the th is s e c t i on are g i ven accumul a ti o n accumu l a t i o n stem areas LS and MS have l ower accumul ations than the areas US and LLS . Thi s i s o f great interest s ince the l at t er two areas are g reen stems whil e the fo rmer are b rown and woody . I t thus appears 1 8 7 that n ickel may b e p re fe renti a l l y ac cumu l a ted i n photosyn the t i c t i ssues rather than non-pho tosynth e t i c t i ssues. Thi s o b s e rva t i o n i s compatible w i th t h e resul ts o f M i ngu z z i and Ve rgnano ( 1 94 8 ) and Vergnano G amb i � �· ( 1 977 ) wh i ch showed p refe ren t i al accumulation o f n i ckel b y the green t i ssues o f A . b erto l o ni i . 8 . 4 MINERAL AND ORGAN I C COMPONEN T ANALYSES 8 . 4 . 1 Mineral Analyses T h e resul t s o f t h e mineral a n a l y se s are shown i n t a b le 8 . 4 . I t c an e a s i l y b e s e en t h a t no o ther elements show a ny changes o f concentration simi l a r to that o f ni ckel . Several poi nt s are howeve r worth noting . The c a l cium concentrations o f the hyperaccumul ato r s , A . s erpyll ifol ium s sp . l us i tani cum and �· mural e, are h i gher than t he concentrations i n t he non-accumul ato r , �· serpyl l i fol ium s s p . serpyll i fol ium. The rel at ive magnesium concentrat ions are t he reverse in the i r d i s tribut ion . T hus i t would appear t h a t t h e hyperaccumul ators h ave developed more effi cient mechani sms Bud ( B ) U ppe r - s t em ( US ) Upper- stem leaves ( U L ) M i d - s t em laterals ( M L ) Mid - s t em ( M S ) Lower s t e m ( L S ) U pper root ( U R ) Lo wer root ( L R ) L o wer laterals - s t em s ( LLS ) - leaves ( LLL ) Figu re 8· 1 for n icka l . P l ant organs an'a l ysa d 1 8 9 TABLE 8 . 1 N i cke l d i stribu t i on w i t h i n Alyssum hel dre i chi i . O rgan Cone o' W t %Amt D A I n/ Wt %Amt 0/ DAI ;0 /0 /0 LR 4330 1 1 . 9 4 . 5 0 . 38 } R 2 1 . 3 1 2 . 2 0 . 57 UR 9 1 50 9 . 4 7 . 6 0 . 8 1 LS 7 1 90 1 0 . 0 S . 4 O . S4 ) MS 9660 5 . 4 4 . S 0 . 85 s 28. 9 3 1 . 2 1 . 08 us 1 6740 3 . 4 5 . 0 1 . 47 LLS 1 7060 1 0 . 0 1 5 . 1 1 . 5 1 LLL 1 2 1 50 27 . 1 29 . 1 1 . 07 } ML 1 1 890 1 2 . 9 1 3 . 5 1 . 07 UL 1 4070 7 . 6 9 . 5 1 . 2 5 L 49 . 8 5 6 . 7 1 . 1 4 8 23400 2 . 2 4 . 5 2 . 09 S eeds 1 880 - - - Abb revi a ti o ns used i n t ab l es 8 . 1 , 8 . 2 and 8 . 3 . Plant O rgans . LR = lower roots Cone = concentration of UR = upper roots ni ckel i n f-lg/g , dry wei g h t . R = total roots % Wt = percentage o f total LS = lower s tem wei ght o f parti cul a r MS = mi dstem o rgan . us s tem %Amt = percentage o f total = upper amount o f n i ckel LLS = lower l a teral stems present in a s = total s tems parti cul ar o rgan . LLL lower l ateral l eaves DAI = di stributi ve accumu-= l ation index ML = l eaves on mi dst em % Amt = UL = l eaves on upper s tem % W t 8 = apical bud L = total l e aves ( see al so fi g . 8 . 1 ) TABLE 8 . 2 1 9 0 N i ckel d i stribution w i thin Alys sum serpyl l i fol i um s sp . l u s i t an i cum . O rgan Cone . % W t %Amt DAI % Wt %Amt DAI LR 1 650 22 . 2 5 . 7 0 . 26 } R 34 . 9 1 4 . 5 0 . 4 2 U R 4 4 30 1 2 . 7 8 . 8 0 . 69 LS 4850 5 . 8 4 . 4 0 . 75 } MS 6050 6 . 1 5 . 7 0 . 9 3 s 22 . 0 2 3 . 7 1 . 08 us 1 0830 2 . 1 3 . 6 1 . 7 1 LLS 8 1 1 0 7 . 9 1 0 . 0 1 . 27 LLL 9 1 80 2 2 . 5 32 . 1 1 . 4 3 } ML 89 1 0 1 6 . 7 2 3 . 1 1 . 38 L 43 . 1 6 1 . 8 1 . 4 3 UL 1 1 460 2 . 9 5 . 2 1 . 79 8 881 0 1 . 1 1 . 4 1 . 27 PODS 1 8 1 0 - - - SEEDS 2730 - - - 1 9 1 TABLE 8 . 3 N i ck el di stribut i on wi thin Alys sum mural s O rgan Cone % Wt %Amt DAI % Wt %Am t DA I LR 3760 1 1 . 5 5 . 1 0 . 44 } R 1 8 . 1 9 . 3 0 . 5 1 UR 5380 6 . 6 4 . 2 0 . 64 LS 6960 9 . 3 7 . 6 0 . 82 I MS 7730 9 . 4 8 . 6 0 . 9 1 � s 2 5 . 7 26 . 4 1 . 0 3 u s 1 3 1 00 3 . 0 4 . 7 1 . 57 LLS 1 1 390 4 . 1 5 . 5 1 . 34 LLL 9420 22 . 5 2 5 . 1 1 . 1 2 f'-1L 9900 5 . 7 6 . 7 1 . 18 ( L 5 '- ? .:;4 . 3 1 . 1 4 } u . - UL 1 0 1 00 2 1 . 7 2 5 . 9 1 . 1 '] G 8920 6 . 4 6 . 7 1 . 05 ) Pods & 3330 - - - Seeds TABLE 8 . 4 Concentrations o f mineral elements i n three Alys sum tax a . N ix Cax Mgx N ax Kx Mnx x � - serpyllifol ium ssp . serpyll i fol ium ( 1 ) 0 . 0003 2 . 58 0 . 7£? 1 . 0 3 0 . 46 72 ( 2 ) 0 . 0007 2 . 2 1 0 . 59 0 . 95 0 . 20 65 ( 3 ) 0 . 0022 3 . 1 7 0 . 7 3 1 . 04 0 . 30 1 1 1 ( 4 ) 0 . 0065 3 . 66 0 . 76 1 . 44 0 . 44 1 1 2 �· serpylli!ol i um ssp . l usitani cum ( 1 ) 0 . 0004 ( 2 ) 0 . 07 1 9 ( 3 ) 0 . 232 ( 4 ) 0 . 54 3 ( 5 ) 0 . 694 ( 6 ) 1 . 1 4 2 A . murale ( 1 ) 0 . 0003 ( 2 ) 0 . 0 903 ( 3 ) 0 . 28 1 ( 4 ) 0 . 604 ( 5 ) 0 . 620 3 . 52 0 . 4 1 1 . 39 0 . 65 2 1 0 3 . 97 0 . 40 1 . 1 0 0 . 29 1 69 4 . 34 0 . 36 1 . 05 0 . 36 2 2 1 3 . 86 0 . 49 1 . 27 0 . 32 1 70 3 . 20 0 . 38 1 . 24 0 . 6 1 227 4 . 02 0 . 53 1 . 1 1 0 . 54 1 60 4 . 09 0 . 23 1 . 60 0 . 5 1 224 4 . 05 0 . 30 1 . 26 0 . 35 1 68 3 . 54 0 . 38 1 . 2 5 0 . 39 1 73 2 . 98 0 . 30 1 . 1 3 0 . 38 265 4 . 77 0 . 24 1 . 22 0 . 4 5 363 xConcentrati ons in % dry wei ght . xxConcentrations in �g/g dry wei ght . Znxx 40 1 7 5 40 79 322 89 86 29 72 35 75 78 93 53 33 Cuxx 1 3 1 0 1 0 1 3 1 3 1 1 7 6 4 8 1 6 1 1 14 7 7 Fexx 57 49 48 48 73 55 46 49 34 37 63 50 5 1 5 1 3 9 ......0 N fo r c a l c ium ab so rp tion and magnes ium e xc lus ion than the non-accumul a t o r . Th i s wou l d b e benefi c i a l t o pl ants g rowi ng on s erpen ti ne so i l s ( w� i ch these h ype raccumul ato rs do ) as these soi l s t end to be poor in cal cium and r i ch in magnesium. The improved mechani sms fo r c a l c i um ab sorpt ion and magnesium exclusion a l l ow the pl ants to ma inta in a more " no rmal '' r a t i o for t hese two el ements t h an woul d o therwi se b e possib l e . T h i s b ehaviour h a s b een recogn i z e d previou s l y by K ruckeberg ( 1 9 54 ) . I t may be that sod ium and po tassium a re a l so mo re ab so rb e d by hype raccumu l at ors than non-accumul ators but the d i fference is small and it i s d i ffi cul t to be certain that i t i s o f s i gn ifi cance . There i s no d i ffi cul ty i n recogn i z i ng a g reater ab sorption o f manganese by the hyperaccumu l a t o r s . A s manganese i s frequent l y a component o f e nzymes i nvol ved i n o rg ani c a c i d cycl es , i t s i n c rease shou l d al l ow a greater turnover of the cycl es wh i ch i n tu rn woul d allow a greater p ro duction o f o rgan ic a c i d s to counterbal ance the i nc re a se in cation conten t ( D e k ock 2. �-1 o rri s o n , 1 9 58 , To rii & L a t i e s , 1 966 ) o r to compl ex w i th the excess m e t al c a t i ons . Fu rthermo re t h e i nc rease i n manganese conc�n t ra t i o n �uy he l p to c ffse t any compet i t ion b y n i cke l in r e s p e c t of en z yme ac t i vat ion and thus p revent mo re serious i n te r ference to b i o chemical cycl e s . None of i ron , copper o r z inc tend t o show di fferences b e tween t he i r concentra t i ons i n the hyperaccumul ators or the non-accumul a to r . Both i ron and coppe r howe ver appear to decrease i n concent rat ion as the amount of nick e l increases a l t hough t he rate o f decrease i s g radual r� ther than sharp . The z i nc d i s tribution i s l e ss c l e ar ; the r:o ncentration appears t o v a ry markedly wi thout a ny p a t t ern b e ing obvi ou s . 8 .4 . 2 O rganic Component Analyses The resu l t s o f the o rgani c a c i d anal yse s are shown i n t ab l e 8 . 5 . Within t he l i s t o f �dentified acids ( i den t i fi c a t i on b e ing b y compari son wi th s tandards ) , there are several interesting resul ts . It shou l d be noted here that the r e sul ts 1 9 3 TABLE 8 . 5 O rganic acids and glucosinolates i n three Alys sum tax a . U ( 1 ) U ( 2 ) Qui nic U ( 3 ) Mal i c U (4 ) I so- U ( 5 ) Ci tri c U ( 6 ) Malonic f"1alon i c U ( 7 ) U ( B ) U ( 9 ) U ( 1 0 ) U ( 1 1 ) Gx citr ic ( 1 ) ( 2 ) �· s_erpvll i fol iLJ_m ssp . serpyl l i fol i um ( 1 ) 9 9 BD - 1 28 2 1 3 1 1 0 1 07 - 230 - 1 0 3 4 3 7 . 09 ( 2 ) 206 1 52 82 - 1 83 G2 95 - 2 1 2 - 1 29 - 9 . 9 1 (3 ) 137 89 6 1 - 87 75 93 27 1 - - 523 - 9 . 00 ( 4 ) 4 9 66 49 1 06 44G 1 32 95 2 30 - 2 1 6 - - 9 . 69 �· �erpyl_l_i fol ium ssp. l us i tani cum ( 1 ) - 1 89 1 27 58 59 - BB - 1 08 897 1 . 70 ( 2 ) 7 1 93 72 92 48 - 36 72 33 420 2 . 74 (3) 71 56 59 79 40 - 22 4G 76 95 2 . 1 4 ( 4 ) 80 82 65 6 1 11 8 - 43 Ei 5 ss 53 2 . 29 ( 5 ) 63 1 1 3 86 6 3 4 0 G3 44 C,9 39 53 2 . 02 ( 6 ) 1 00 76 73 70 - 55 - 86 79 4 0 2 . 1 7 A . murale ( 1 ) 63 8 1 65 6 6 96 7 3 4 6 - 37 49 25 200 2 . 28 ( 2 ) 7 5 96 74 67 84 73 5 5 - 4 5 44 32 2 1 6 4 . 2G ( 3 ) 49 60 69 60 54 5 5 30 - 26 39 1 2 24 3 2 . 5G ( 4 ) 64 85 77 79 72 90 54 47 74 63 42 346 3 . 2n ( 5 ) 47 65 56 58 69 62 38 67 5 1 46 22 1 54 1 . 74 Resul t s expre ssed a s arb i trary uni ts per gram , dry weight . Alys sum s ample s numbered as for tab l e 8 . 4 xG = % Glucosinolates , dry weight basi s . ......0 � are g i ven i n a rb i trary un i t s pe r g ram o f d r i ed ma t e ri al and h ave not b een transl a t e d i n to �g / g . The non- accumul a to r , a. serpyl l i fo l i um s s p . serpyl l i fol i um , h a s no t i ce ab ly h i g her c i tric and i soci t r i c a c i d l evel s t h an the t wo hyp e r accumul a t or s s tudi ed , a. se rpyl l i fol i u� s s p . l u s i tani cum and �· �ral e . The malonic a c i d conte n t o f �· se rpyl l i fol i um ssp . s erpyl l i fo l ium 1 9 5 i s h i g h but a s malo n i c aci d shows two p e ak s i n the other two t axa ( the s tandard so l u t i o n a l so v a r i e d from one to two peak s ) d i rect comp ari sons � r e d i ff i cul t to mak e . Neither qui n i c nor mal i c ac ids appe ared to show any s i g n i f i c a n t di fferences b e tween the di f ferent t ax a . A l arge numbe r o f un i den t i fi e d o rganic acids w e r e also presen t . Many o f these a c i d s ( U ( 1 ) , U ( 3 ) , U ( 4 ) , U ( S ) , U (6 ) , U (7 ) , U (8 ) and U ( 1 1 ) ) were confi ned to one taxon . The remain i ng three ac i ds , U ( 2 ) , U ( g ) and U ( 1 Q) , we re found in �11 three t a x a . A c i d U ( 2 ) was g e nera l l y found i n h i gher concen t ra t i on s i n the non- accumu l a to r whi l e U ( 9 ) was recorded o n l y once in t h i s t�xon b u t w�s p resen t in al l s ampl e s o f both hyper2c cunu l n t o rs . � c i d U ( 1 0 ) is mos t common in A . mu ral e but is a l so p r e s e n t i n t h e o t he r t ax n . I t doe s not appear to di f fer s ign i fi c: c�n t l '1 on � non- 2 c cumu.l :-3 to r-h ype raccumul a to r b a s i s . T h e r e sul t s o f the g l u co s i no l a t e analyses a r e shown in tab l e 8 . 5 . In a l l three tax a there i s an increase in the glucosino l a te content in the presence of h i gh l evel s of n i ckel i n the so i l . Thus a. serpyl l i fo l ium ssp . serpyl l i fol i um shows a major i ncrease i n gluco sinolate content even when there is l i t t le d i fference in the concen tration of n i ck el w i t h i n the l eaf t i ssues . The soil concentration o f n ickel which effects thi s change appears to b e between 30 and 1 00 �g/g of dri ed soi l . Even more noticeable than t h i s i ncrease i s the d ifference i n concentrations b e tween the non-accumul ato r , a. serpyl l i fol ium ssp . serpyll i fol ium , a nd t he two hyperaccumul ato r s , �· serpyll i ­ fol ium s s p . l us itanicum and A . mural e . This d i fference may result from the use the hyperaccumul ators coul d make o f the ni ckel absorb ed . Glucosinol ates are common secondary metab o l i te s i n the p l ant o rder Capparal e s ( o f which the Cruciferae are a member famil y ) and are bel ieved to act a s fungi c i de s , b ac t e r i ci des (M i tschne r , 1 97 5 ) and i n sec t i c i des ( E r i ck son & Feeny , 1 974 ) fo r the p l a nt s ' protec t i on . Heavy metals c a n al so s e rve these pu rpo s e s so t h a t i n the pre sence of h igh n i cke l concentrations l ess glucos i no l ate wou l d b e requ ired . Thi s expl anation i s however i n confl i ct wi th the p revious observation t ha t gl uco s inol ate production is stimulated by the p resence o f n ickel i n the soi l . N o expl anation of t h i s confl i c t i s currentl y apparent al though o n e possi b i l i t y i s tha t t h e vari a t ions i n concentra t ion wi thin a species a r e norma l rather t h a n ni ckel i nduce d . Thi s looks unl ikel y i n that o n l y � · mural e shows any w i d e vari ation in concentrations . Fur ther studi e s w i l l be requ ired b efore a de finite conc l u s i on can b e made . 8 . 5 N I CKEL COMPLEXATIDN The GLC traces of the components of t he n i ck el comp l exes ob t a i ned a re s hown i n fi g . 8 . 2 . Two peak s a re common t o a l l t h e ru n s : the se p e a k s a r e E a n d X . I n g enerAl , p e ak X i s gre a ter than peak E . g varying numb er o f l esse r peaks are obtai ned for a l l speci e s . T he peaks D and c i t r ic c c i d a r e th e m o s t common of t h e s e l e sser peak s . rhese l e sser peaks h ave a tendency to group themsel ves i n seri e s among the spec i e s . The resu l tan t groups o f s p e c i e s are then found to correspond i n the i r c omponent s pec ies t o t h e sub sect ions and seri es o f s e c t i on D donta rrhen a . Thus A . vi rgatum ( subsec tion Samari fera ) differs from the o thers b y having the series of peaks H , J , K , L and V . �· a rgenteum , A . tenium and A . mu ral e ( subsect i on Compressa , s e r i e s I n te gr a ) have p eaks M , B , D and c i tr i c aci d ( al t hough 8 and D are weak i n �· tenium ) when comp ared to A . h e ldre i chi i ( subsection Compres s a , s er ies Crenu l a t a ) which h a s peaks A , B , C , D and ci t�i c aci d . a. corsi cum , �· euboeum and A . troodi i ( sub section I nfl at a ) h ave the p e ak s D , F 8nd c itric aci d i n common . O ther earl i e r e lut ing ac ids were a lso common i n these l atter three spe c i e s but a s no c l e ar mass spectra were obtained for them , t hey h ave remained unt i tl ed . 1 9 6 • F i g ure 8 · 2 Gas - l i q u i d ch romatographs of t he me t h y latcz d de r iva t i ves of t he2 n i c ke l - complex i ng l i g and s from _A_/y_ssum spec 1 C2 S . A. A. corsicum B . A . euboeum C . A . froodii D . A . argen teum E. A . murale F. A . fenium G . A . heldreich1i' H . A . virga fum I . Bornmuel/era _ty_mp_haea o- 0 L{')- l.U u '- ........, u o_ � � (.) · - V) � (.) c · - E �I . <( w � ...._ L(') _ ....... lL w X o - l.{') - l.{') - ...- O ­ N l.{') _ N c E w �I u x.:__ _ a - L!) - c E w � I- a _ N L!) _ N u u · - ' L: '- ......, ........ . _ ,_ u u 0 E 0 ...c + 0 LD- w 2: I- a_ N l.(')-N w . w X u u 0 E 0 ..c + o - l.{)- o_ c E w :L t- a_ N l.{)_ N o- Ln- u .S E w 2: t- . � - c: LD- ......- (J.) ...._ �I L.L <( 0 w u u · - · - '- '- ...... ...... · - · -u u 0 E 0 ...c + X :::::: ..c: l> ·- (]) '- \:) - (]) ..c: � I � u ' o - l.(') - a - .....- c::: E w � 1- l.(') -......- o _ N l.(') ..,.. N :r: X ' o - l!'> - c E w L 1- � - U1 . N - lL w 0 ()) 0 ..c:: � �I ..._ I en I t-t o - U"> - c E w 2: 1- U"> - U"> N- The cl o se l y rel a ted ao rnmue l l e ra tymphae a hcd onl y pe ak F i n add i t i on to t ! 1 e fl 8 a k s E :mrl X . Hm.Jever the p eak F i n t h i s s pe c i e s h2d a w i de l y d i fferent retRn t i on t ime ( rel a t i ve to peGk E ) when compa re d w i t h the peak F o f the Alyssum spe c i e s ( the compo s i t i on o f th e p e ak s w a s d i fferen t i ated on t h e mass spectra ob t a i ned fo r them ) . I t may b e that the p eaks t i t l ed F are ac tual l y o f i somer s . R e p r esenta t i ve s amp l e s o f the mass spectra o f the t i t l ed p eak s are shown i n f i g . 8 . 3 . The re was a consp i cuous l ack of correl a t i o n b e twe e n the spec t ra p resented here and those r e corded in the l i t e r a tu re such tha t few o f the e s te rs , and he nce th P. r a r e n t ac i d s , cou l d be pos i t i vel y i denti f i e d . Mal i c , mal oni c a n d c i t r i c a c i d s h ave a l l b e e n previ ous l y i mpl i cA te d i n ni ck el Gompl exat ion by Alys sum speci e s ( P e los i e t a l . , 1 976 , Panca ro e t a l . , 1 977 , Shaw , 1 980 ) and were -- -- -- obvious a c i ds to l o u k fo r . C i tr i c a c i d t rimethyl e s t e r was r e ad i l y a p p a ren t in t h e GLC traces and i t was r ead i l y confi rmed b y th2 na ss spe c � r a o f �hose pe ak s . I n A . hel drei ch i i and 2 0 1 A . � r� � n t eur ( r g r re sen t i ng the two s er i e s o f subsec t i on Compres s a ) t he c i t r i c � c i d L r imethy l e s t e r p e ak had a n homo l o gous cont �m i nan t , homoc i t ri c a c i d t r ime t h y l R ste r . Th i s p a rent aci d has b een i dent i f i e d �s the comp l e x i ng agen t in another n i ck e l hyperaccumul ato r , P e a rsoni a me t al l i fe r a ( Stack l ey , 1 980 ) . �. corsi cum ( represent i ng subsec t i on I n f l a t a ) d i d not show the p resence o f thi s c ontam inant . S pectrum 8 bears some re semb l an c e t o spect ra previ o u s l y recorded fo r t h e mal i c a c i d d imethy le s t e r . However t h e s i z eable p eaks a t m/e = 1 1 7 and 1 1 3 requ i re some explana t i on . T he p eak a t m/e = 1 1 3 i s n ormal l y f ound i n mal i c aci d dime thyl e s te r spe c t r a a t 1 0 - 2 0% o f the b ase p eak compared t o the 57% found h e re . Th i s enhancemen t i s h owever mo s t p robab l y an i nstrumental a r t i fac t . The p e ak a t m/e = 1 1 7 i s not known f rom the mal i c aci d d imethy l e st e r b u t i s k nown from the spec t rum of the m ixed monoethylmonome thyl e s t e r of mal i c a c i d (CH300CCHOHCH2COOCH2CH 3 , Webb e t al . , 1 9�7 ) . Fi g ure 8· 3 Mass s pec tra o f the methylated de r ivat ive s o f the n i cke l - complex i ng l i gands . ( b y t he pe a k desi g nat ions of fig . 8· 2 ) ( i ) c i t r i c ac i d ( i i ) c i t r ic ac i d + homoci t r ic ac i d ( i i i ) B ( i v ) M ( v ) A (v i ) H ( v ii ) F ( 1 ) (vi i i ) F ( 2 ) ( i x ) J ( x ) C ( x i ) 0 ( x ii ) E ( xv ) K ( x i i i ) L (xv i ) Y ( x i v ) G (xv i i ) X - r I I I I I I I • .· c , I I I I I I I .. X 10 200 180 1 60 140 1 20 I ' I I I I I I I I I I I 100 80 60 - 11 I I I 40 ( i ) ci tr ic acid ; ( i i ) ci t ric aci d + homoci t ric ac i d ( i i i ) B � :r: <( -> --- · - > > - - 0 "" -� N - - LL LL - · - " > > - - -, - X - t 0 .....:t 0 CO 0 N � (_) - X - � - � - � � - - - 1- - 0 - X - 1------ ---· ·-- - - - . f UJ - X - -� - - -� - � - � - � - � � � 0 � I o i CO \ a ! 0 ; � · 0 N � . I -X - - - .� - - - 1- f.- .... � f.- 1- f- � > X - � __............. - ----- f- - 1- - � � - !'- 1- - > X - 0 (.£) 0 CO 0 0 � 0 N � 0 CO � ( xv i ) Y 1 00 5 0 ( xvi i ) X O • note chang e 2 00 18 0 1 6 0 1 4 0 1 20 1 00 80 · 6 0 40 in intensi t y ------------------------�------------�----------�--------------�·1 � 11 I I I 4 20 400 3 8 0 • Ill I ' I 3 6 0 3 40 .I T T 320 3 00 j I I I 2 8 0 2 6 0 ' i I I ' �·· - .. . ' 2 4 0 2 2 0 1 1 0 200 Thi s e ster wou l d have fo rme d b y meth y l ; J t i on o f Lhe a c i d rnonoethylmal ate whi ch i s known to o c cu r i n uine s Uuebb e t 5!.!· , 1 967 , D r�we rt e t � . , 1 974 ) . Spec t rum B m�y the re fo re b e a mi x ture o f the mal i c ac i d d imethy l e s te r and the monoe thylmal a t g monome thyl ester o f w h i c h t h e forme r i s domi nant . 2 1 5 Spectrum M shows the c l o ses t s p ec trum to those p revi ous l y recorded for mal on i c a c i d d imethyl e s t e r . This spec t rum c onta ins a number o f peak s whi ch are not a sso c i ated w i th t he s pectrum expec ted of malon i c a c i d . I t may w e l l be t hat t h i s spec trum i s o f a mixture o f w h i ch mal o n i c a c i d d ime thyl ester i s a componen t . I f so t h e most ch3rac t e r i s t i c p eaks o f the o t h e r component a r e t he p e ak s at m/e = 85 a n d 83 ( see al so the d iscussion of s p e c t rum F ) . Spectrum A i s domi n a t e d by a peak a t m/e = 1 1 0 wi th the h i ghest p e ak at m/e = 1 4 0 . A p e ak at m/e = 59 i s characte r­ i s t i c o f methy l e s t e rs o f carboxyl i c ac i ds ( and several o the r g roups , Beynon __::!; � . , 1 968 ) . O th e r tvpP.s o f co11pounds w h i ch g ive r ise to t h i s p e ak are t e rt iary a l i phat i c al cohol s , s ome e th e rs , a l i p h a t i c monob asic c arbox y l i c ac i ds �nd e s t e r s o f no rmal chain d i bas i c carboxyl i c ac i ds . The nl coho l s and e thers are u n l i k e l y candi dates a s they do not fo rm stab l e compl exes wi th n i ckel i n aqueous envi ronments . Any monobas i c c a rboxyl i c a c i d wou l d have been e s t e r i f i e d by the d i azomethane hence it cannot b e the source of thi s p eak . The d ibasi c c a rboxy l i c a c i d methyl e sters have been subjected to m a s s spectrome try ( Tress! e t �. , 1 97 5 ) and b ear no resembl ance t o thi s spec trum. I t thus appears that the spect rum is that of a methyl e s t e r of an aci d of unknown compo s i t i o n . I f the p e ak at m/e = 1 40 is the parent i on p eak then the compound must have a h i gh degree of unsatu rat i on ( at least ' two C=C u n i ts or p o s s ib l y a fur anose or p y ranos e structu re ) . Pos s i b l e formul ae cou l d then b e c8H 1 2o2 , c7H8o 3 o r c6H4o4 , a l t hough t he l ast i s l ess p rob able . S pec trum H has i ts b ase p e ak a t m/e = 1 0 1 . The h i ghest recorded p eak was at m/e = 1 3 1 but as the e s t e r s u nder i nve s ti ga t i on conta in o n ly c arbon , hydrogen and oxygen 2 1 6 thi s c anno t b e the p a ren t i o n peak . As mos t ( b u t nn t Q l l ) methy l e s t e r s fregmen t w i th an i n i t i al CH30 " or " COOCH 3 l oss , the paren t compound may have a molecul ar wei ght of 1 6 2 or 1 80 . When CH3o• i s i n i t i al l y l o s t , i t i s general l y fol lowed b y CO l o s s . As CO h a s a mass o f 2 8 and t h e m a s s l o s t from the m/e = 1 3 1 peak i s 30 , g�v�ng the m/e = 1 0 1 peak , i t appears unl i k e l y that t h e m/e = 1 3 1 peak i s d u e to ( M - CH30 ) + • Thus i t i s mo r e p rob ab l e t h a t the e ster has a mol e cul ar weight o f 1 80 . Aga in the a c i rf o f such a n e s t e r woul d b e h ighly unsaturated or h e t e rocycl i c . Spect rum F ( o f which two versi ons are shown ) may b e a mi x tur e . The s p ec t rum l ab e l l e d F ( 1 ) was found i n �· tymphaea and on the ri s i ng s i d e o f the peak in �· corsicum . The spe c t rum F ( 2 ) was found on the decl i n i ng s i de of the p e ak in A. corsi cum . The no ti ceab l e di fferences b e tween the spectra are the r i se of the pHak at m/e = 88 i n F ( 2 ) wi th the decl i n e o f th e peaks a t m/e = 58 and 40 and t h e appearance o f a numb e r o f sma l J p e a k s obove m/e = 88 . I f t h e peak a t m/e = 88 i s from one component o f a mi xturP. and the peaks at m/e = 53 and 40 are frnm ano ther , the d i fferences are easi l y sx p l a i n8d . Fur the rmo re the component w i th peak s at m/e = 58 and 40 h a s peaks a t m/e = 85 and 83 and may wel l b e t h e other component wi th the mal on i c ac id d i methyl e st e r i n spectrum M . The p eak at m/e = 88 i s characte r i s t i c o f 2 -methylcarboxyl i c a c i d methyl este r s . I f' m/e = 1 30 ( the h ighest peak ) i s the parent mol ecu l a r i on then 2-methylpentano i c aci d monome thy- l e s ter i s a possib i l i ty . T h i s a c i d coul d expl a in the m/e = 1 0 1 peak by CH3CH2" los s , and the m/e = 98 peak b y CH 30H l o s s a s well as the rearrangemen t i o n peak at m/e = 88. Furthe rmore this portion of the spectrum corresponds reasonab l y well wi th p revious ly tabul ated spec tra . The small peak at m/e = 1 1 6 may b e the parent ion for the o ther component o f th i s mi x ture . Thi s woul d co rrespond to a formu l a o f c6H 1 2o2 o r c5H8o 3 • R epresentat ives o f the fi rst formul a incl ude the methyl e sters o f pentanoi c aci d , the me thylb utano i c acids and d ime thyl­ p ropanoi c ac i d . The esters o f the methyl-sub sti tuted a c ids all have l a rgR peak s 2 t �/e = 74 wh i ch i s not found i n s pectrum F . The pentanoic ac i d e s t P. r h a s a strong peak a t m/e = 5 7 whi ch i s no t found in spe c t rum F . The second formula i s t hat o f the o xobutano i c a c i d methyl e sters . T h e 2-oxobut ano i c a c i d e s ter h a s n o peak s i n t h e m/e = 83-85 range which spectrum F has . The 3-oxobu tanoi c a c i d e s ter h a s peaks at m/e = 85 and 84 compared with spectrum F which has peaks at m/e = 85 and 8 3 . Also the 3-oxob u t anoi c a c i d e s t er h a s a b ase peak to i ts spectrum at m/e = 4 3 whereas spectrum F has a base p e ak a t m/e = 4 0 . T h e spectrum o f the 3-oxobutanoic aci d e st e r i s undoub tedl y t h e closest o f these s p ec tra t o that w h i c h i s requ i re d but t h e evi dence i s no t conclus ive . Spec trum J has much i n common with spectrum F ( 2 ) . The �ajar di f fe rence i s the appearance o f peaks a t m/e = 1 4 5 , 1 44 and 1 1 5 . P eak s a t m/e = 8 8 and 5 8 again domina t e the spectrum . A we ak metastab l e p e ak , m� = 9 1 . 2 , i nd icates the t rans i t ion from m/e � 1 4 5 to m/e = 1 1 5 . A high res o l u ti on mass spec tr a l �a l y s i s o f t h e m/e = 1 4 5 peak gave a formula o f c7H 1 3o 3 • As suming a g a i n that the ester has i n i t i al l y l os t 59 m a ss · J n i t s ( the l o s s f rom t h e h i ghest peak i s 3 0 mass u n i t s compa red to the ?.8 ( CO ) wh ich woul d be expected i f o n ly t h e CH 3o • h a d b e e n l o s t ) wou l d g i ve a mol ecular w e i gh t o f 2 04 and a fo rmul a of C9H 1 6o 5 • The l o s s o f 5 9 mass uni t s ( assumed ) fol l owed b y the l oss o f 30 mass u n i t s (measured ) appears frequently i n these s p ec tra . Thi s 30 mass uni t i s e i ther c2H6 or CH2o . The l a t te r poss ibi l i ty i s mos t i nt ere s t ing . 2 1 7 L o s s o f CH2o after an i ni ti al l o ss o f · coocH3 wou l d a l low the ester to have the s t ru c ture CH300CCHOHR whi ch would g i ve a p arent aci d s tructure o f HOOCCHOHR . This woul d allow chelation through both the carboxylate and hydroxyl oxygens . The ac ids are not however l i k e ly to b e a l iphat ic 2-hydroxycarboxy l i c aci d s a s t h e e s ters o f these a c i di g i ve a characteri s ti c p e ak a t m/e = 90 which i s no t s e en i n t h i s spec trum. Th i s wou l d tend to indicate that a further funct i onal g roup is c lose to the 2-hydroxycarboxyl g roup ( prob ab l y � - o r �- to the hydroxyl group ) . One such g roup cou l d be a s econd carboxylate g roup . Th i s woul d g i 11 e t; he I + < S � ( N i T ) // ,, >(N i T)JI ,jf ( Ni T ) + A i ( SIN i T ) i � ( N i t s l > ! S I N i l + T� � < '/ // ,, · r