PRE-RELEASE STUDIES ON ZOPHODIA TAPIACOLA (DYAR) ( PYRALIDAE LEPIDCPTERA ),

Transcription

1 PRE-RELEASE STUDIES ON ZOPHODIA TAPIACOLA (DYAR) ( PYRALIDAE LEPIDCPTERA ), A BIOLOGICAL CONTROL AGENT AGAINST JOINTED CACTUS, OPUNTIA AURANTIACA LINDLEY by J. H. HOFFMANN Thesis presented fr the degree f Haster f Science. Department f Zlgy and Entmlgy, Rhdes University, Grahamstwn, Suth Africa. January, 1976

2 Male (left) a nd female (right) z. tapiacl a. (x3) Phtcpy f Miss Drake's painting f 0. aurantiaca publ ished in Lindley's descript in f the plant (1833). (xo,7)

3 ii ACKNOWLEDGEMENTS I wish t extend my sincere thanks t my supervisrs Dr. V. C. Mran and Dr. D.P. Annecke wh initiated and guided me thrugh this prject. Als t Prfessr B.R. Allansn fr allwing me the use f facilities in his department and fr helpful advice. T Mr. H.G. Zimmermann, Plant Prtectin Research Institute, Uitenhage, whse knwledge and experience with jinted cactus was always frthcming. I thank the Cuncil fr Scientific and Industrial Research fr funding this research and the Suth African Department f Agricultural Technical Services withut whm the wrk culd never have taken place. I am very grateful t the fllwing peple: Mr. G. de Kck, Grtfntein Agricultural Cllege, Middleburg, and Dr. J.S. Starke, Thrneycrft farm, Grahamstwn, fr supplying me with spineless cactus plants. Mr. J. Willetts, Thursfrd farm, Grahamstwn, fr allwing me t cllect jinted cactus n his prperty. Mr. H.J. Tanner, Btany Department, Rhdes University, fr supplying varius r namental cacti and ther succulent plants. Mr. D. Frsyth, Mr. B.H. Gunn, Mr. S.B. Malclm, Mr. M. W. Mansell, Mr. A.J. Urban and Dr. A.G. Jact Guillarmd, all f the Zlgy and Entmlgy Department, Rhdes University, and Dr. L. Parlis and Mr. D. McGarvey, Chemistry Department, Rhdes University, fr invaluable help and discussin. Mr. B. Burger, Mr. D. Malan and Mr. W. Kulati, Plant Prtectin Research Institute, fr their generus technical ass istance. Finally I wuld like t thank my wife, Sue, fr her ~atience and help thrughut and fr typing this manuscript.

4 iii CONTENTS PAGE ACKNOWLEDGEMENTS,, RESUME INTRODUCTION MATERIAL AND METHODS 1. THE LIFE HISTORY OF ZOPHODIA TAPIACOLA 2. THE HOST SPECIFICITY OF ZOPHODIA TAPIACOLA 2.1 Starvatin tests 2.2 Negative-vipsitin tests ii iv THE EFFECTS OF ZOPHODIA TAPIACOLA ON OPUNTIA MAXIMA The gum secreting O. maxima cultivars The efficiency f defensive gum in. maxima The surce f gum in O. maxima Cactus cuticle as a barrier t Z. tapiac1a larvae Z. tapiacla attacks n O. maxima flush THE IMPACT OF ZOPHODIA TAPIACOLA ON OPUNTIA AURANTIACA 38 DISCUSSION REFERENCES APPENDIX I APPENDIX II

5 iv ; ; RESUME Jinted Cactus, Opuntia aurantiaca Lindley (see frntispiece), is the mst imprtant weed plant in Suth Africa, infesting apprximately, 1,2 X M 2 and csting apprximately R per annum. Trdn herbicide effectively kills jinted cactus bushes t which it is applied. Hwever, apart frm being expensive and damaging t beneficial vegetatin, spray prgrammes have nt successfully cntrlled the weed because mst small O. aurantiaca plants are impssible t detect in the field. slutin t the prblem. Bilgical cntrl may prvide a Tw insects, the cchineal bug, Dactylpius austrinus De Ltt and the pyralid mth, Cactblastis cactrum Berg., already exercise a degree f cntrl ver the weed. The intrductin int Suth Africa f ther natural enemies such as Zphdia tapiacla (Dyar) frm Argentina, Suth America, may reduce the density f jinted cactus t belw an acceptable ecnmic threshld. Any insect cnsidered fr release shuld nt clnise and destry beneficial plants f which the culivated spineless cacti are the mst vulnerable. Pre-release studies n Z. tapiacla have shwn that it can nly clnise a few species f lw grwing cacti and that it will nt damage the large spineless cacti r ther desirable plants. Further, the mths are relatively fecund and each larva destrys significant amunts f O. aurantiaca during its develpment. Cnsequently, Z. tapiacla is nt nly cnsidered safe fr release but it has the ptential t act as a successful bilgical cntrl a gent f O. aurantiaca in Suth Africa.

6 1 INTRODUCTION Jinted cactus, Opuntia aurantiaca Lindley (1833), is the majr weed prblem in the East Cape Prvince f Suth Africa where the plant invades valuable grazing pasture and, if nt checked, frms dense, thrny, impenetrable thickets. The plants are generally lw grwing incnspicuus shrubs frm which individual jints, r claddes, are readily detached principally by passing animals, wind, rain and flwing water (Pettey, 1947). The height f each plant is determined by the surrunding vegetatin and in semi-arid kar scrub, jinted cactus seldm grws t mre than ne half meter while it may reach tw t three meters when it is supprted and prtected by tall vegetatin. Althugh the yellw O. aurantiaca flwers prduce seed bearing fruits, reprductin is mainly vegetative because every lse jint and f r uit that cmes int cntact with the sil can rt t frm a new bush. The ease with which jints are detached frm the plants, cupled with the ability f each t frm a new bush, facilitates spread and establishment f the weed in uninfested areas. Fr many years, attempts have been made t eradicate, r at least cntrl and check O. aurantiaca. Initially landwners tried t rid their prperty f the weed by mechanical means. This, hwever, prved t be ineffective and time cnsuming. Expensive hrmnal herbicide applicatins (initially 2,4,5-T, nw replaced with Trdn 40) temprarily

7 2 reduce the weed density t levels acceptable t the landwners, althugh in the East Cape Prvince, jinted cactus is still spreading t uninfested areas at a rate f 1 x 8 M2 per annum and eradicatin in already infested veld has never been achieved. Trdn 40 always kills O. aurantiaca plants nt which it is sprayed, hwever, its use never achieves satisfactry cntrl because mst small plants are nt detected during herbicide applicatin. Besides being inefficient, -the herbicide prgramme is very expensive and R has been spent n litres f herbic ide alne between 1957 and This figure des nt include the cst f transprt estimated at R each year and the cst f labur fr which there are n estimates available*. Anther disadvantage f the methd is that many beneficial plants in the prximity f jinted cactus bushes are als killed by the herbicide which may remain effective in the sil fr a number f years. Rising csts cupled with labur shrtages are making the applicatin f herbicidal cntrl s difficult that there is an increasing demand fr cheaper, mre efficient cntrl methds. Bilgical cntrl is an alternative that may prvide a slutin t the jinted cactus prblem (Nes$er & Annecke, ). The practice f intrducing natural enemies t cmbat cactus and ther weeds has met with much success in many parts f the wrld, including Suth Africa (Ddd, 1940; Fullaway, 1954; * Unpublished Data frm, the Eastern cape Regin, Department f Agricultural Technical Services, Queenstwn, Suth Africa.

8 3 Hllway, 1964; Wilsn, 1964; Andres & Geden, 1971; Nes.er & Annecke, 1973; Geden et al., 1974; DeBach, 1975). Extensive surveys in Suth America where O. aurantiaca des nt reach p est prprtins (except n the island f Martin Garcia ff Buens Aires) have established that there are several natural enemies f jinted cactus and ther clsely related cacti in that area (Zimmermann, 1972). Tw f these natural enemies, namely the cchineal bug, Dactypius austrinus De Ltt and the pyralid mth, Cactb1astis cactrum Berg., have a lready been released in Suth Africa and are respnsible fr a significant but undetermined degree f cntrl. Huffaker (1974) pints ut that it may be the ttal respnse f a number f natural enemies that cntrls a pest and nt the respnse f anyne. Cnsequently, the pressure already being exerted n O. aurantiaca may be increased if mre natural enemies are intrduced and adequate bilgical r intergrated cntrl f jinted cactus may be achieved. Befre any rganism can be transferred t a new cuntry. fr bil gical weed cntrl, its hst range must be determined t ensure that beneficial plants wi ll nt be clnised and destryed in its new envirnment (Huffaker, 1964). The cacti are usually very gd candidates fr bilgical cntrl because they are seldm ecnmically beneficial plants and almst withu t exceptin, i nsects that feed n cacti d nt utilise ther plant families and vice versa (llann, 1969). Hwever, althugh plants utside the Cactaceae might be safe frm attack by cactphagus insects, in Suth Africa a prblem

9 4 arises because nt all cacti are weeds and there is a beneficial grup knwn as the Burbank spineless cacti, Opuntia maxima Mill., which are grwn as fdder crps in the drier areas f the cuntry (de Kck & Au c amp, 1970). O. maxima, which embraces a wide variety f cultivars r frms, is clsely related t the weed prickly-pear Opuntia megacantha Salm Dyk and they may nt be btanically distinct frm each ther (Hendersn & Andersn, 1966). Hwever, the spineless cacti are generally referred t as. maxima while the thrney weed is knwn as O. megacantha. O. maxima is already damaged by intrduced American cactphagus insects, including C. cactrum, which were released t cntrl ther Opuntia weeds (Pettey, 1939; Pettey & Marais, 1950 ).Therefre, t prevent further damage, nly insects that are unable t c lnise O. maxima may be cnsidered 'safe' fr release and pre-release studies must be cnducted t establish the hst specificity f all ptential cntrl candidates. The first and prbably mst cnclusive indicatins f whether an insect will be a suitable, safe bilgical cntrl agent cme frm surveys cnducted in its indigenus habitat and frm reviews f the literature (Zwlfer & Harris, 1971). Surveys f the pyralid mth, Zphdia tapiacla (Dyar), cnducted in Argentina, Suth America, shw that Z. tapiacla is an ligphagus species restricted t a few lw grwing O. aurantiaca type cacti (Mann, 1969; Zimmermann, 1972). Ddd (1940) recrds that in 193~ Tucumania tapiacla Dyar (= Z. tapiacl~) was intrduced int Australia as a cntrl

10 5 agent f O. aurantiaca in that cuntry. Extensive pre- and pst-release examinatins shwed that the mth never clnised beneficial plants and it nly survived n O. aurantiaca, H. martinii and yung O. inermis plants. Ddd als nted that Z. tapiacla never exercised spectacular cntrl ver O. aurantiaca because the cchineal, Dactylpius sp. near cnfusus (= D. austrinus), had previusly destryed mst f the available O. aurantiaca plants, thereby depriving the mth f suitable hsts. Hwever, he als nted that the mth did seem t be inflicting an unspecified degree f damage t O. aurantiaca in certain areas. Cnsequently, Z. tapiacla was cnsidered t be a hst specific species which, althugh nt spectacular in its cntrl f O. aurantiaca in Australia, was never-the-less amenable t intrductin int new cuntries. As a result, in February, 1973 Z. tapiacla larvae were cllected frm Opuntia disclr and Opuntia retrsa hst plants alng rad Number 81 between Pirane and Ls Lmitas in the Frmsa Prvince, Argentina, by Mr. H. Erb wh was emplyed by the Suth African Department f Agricultural Technical Servi ces fr that purpse. The larvae were used t start a labratry reared clny f Z. tapiacla which was maintained at the Institute f Miquel and Lill, Tucuman, Argentina. In December, 1973, 500 t 00 eggs n frayed clth strips were sent by air t Suth Africa where they were received in Jhannesburg by an fficer f the Department f Agricultural Technical Services frm Pretria. The eggs were Checked fr

11 6 damage befre being frwarded t Mr. H.G. Zimmermann at the Weed Labratry, Department f Agricultural Technical Services, Uitenhage. Here the culture was maintained fr ne generatin and n the 5th February, 1974, eggs f the F1 generatin were mved t Rhdes University, Grahamstwn where the culture has since been maintained. In May, 1974 Mr. H. Erb cllected mre Zphdia sp. larvae frm Opuntia tayapayensis plants arund Cchabamba, Blivia. These were reared thrugh t adults in Tucuman and the eggs btained were sent t Suth Africa in September, This clny was reared i n Uitenhage until the 4th Nvember, 1974 after which it was mved t Grahamstwn. Initially it was suspec~ed that the Frmsa ppulatin and the Blivian ne cmprised tw seperate Zphdia species. Hwever, Mr. P. Whalley, British Museum (Natural Histry), Lndn, determined specimens frm bth clnies as all being Z. tapiacla. Unfrtunately, in June, 1975 the Blivian clny died when C. cactrum larvae, whi ch are mre vigrus feeders than Z. tapiacla, where accidental l y intrduced int the rearing cages with jinted cactus frm t h e field. Individuals frm the Frmsa clny were used t study the life histry, basic bilgy and hst specificity f z. tapiacla with a view t its release as a cntrl agent f O. aurantiaca in Suth Africa.

12 .. '... _clth strip,,' OVIPOSITION - CAGE EGG S ON sliding dr CLOTH STRIPS.....," ~ jinted cactus..... HS LARVAE REARING-CAGE \ NEWSPAPER DESTROYED CLADODES WITH PUPAE & LARVAE light trap / NEWS PUPAE PUPAE ~""O"O PAPER DARK-BOX - DESTROYED \ ~ CLADODES ~ AFTER 35 DAYS ~ BURNT Fi. j. 1. Rearing techniques used t maintain a clny f z. t apiacla in quarantine.

13 7 MATERIAL AND METHODS The clny f z. tapiacla frm the Frmsa Prvince, Argentina was maintained at Rhdes University in a quarantine insectary at a cnstant temperature f 25 C (± 3 C) and a relative humidity f 50% (~lo%) with a 12 hur light/12 hur dark phtperid. During the dark perid a Philips GL 45 B indicatr bulb prvided dim illuminatin in the rm s that the mths were nt subject t unnatural ttal darkness. Adult mths were prvided with a hney-water feed and frayed clth strips in perspex thermplastic 'vipsitincages' (fig. 1). The mths mated in these cages, after which the females depsited mst f their eggs n the clth strips. Eggs were transferred n the strips t 'rearing-cages ' and placed amng O. aurantiaca claddes cllected at Thursfrd farm, Grahamstwn, Suth Africa. The base f each. 'rearingcage~ was lined with five t ten sheets f newspaper and then filled with claddes. Yung z. tapiacla larvae hatching frm the e ggs crawled nt and tunnelled int the jinted cactus in which they develped and pupated. At weekly intervals, all dead claddes and the newspaper were remved and replaced with fresh nes. The dead claddes were placed in a 'dark-bx' with a 20mm diameter 'escape-hle' lcated in the ne tp crner and cvered with a gauze 'light-trapcage'. Adult mths emerging frm the pupae in the dead claddes were attracted thrugh the escape hle t the dim insectary "night light". Cnsequently, they aggregated in

14 0 0 LU It: ~... Ct It: LU Q. :ie LU R. H. 0/0 ' I 0 ~ m.- l> -! < m J: c: 50 3: c -! -< ~ 0 Fig. 2. Temperature, relative humidity and phtperid maintained in the cntrlled e nvirnment rm during all bservatins n Z. tapiacla.

15 8 the 'light-trap' and were transferred daily t 'vipsitincages'. Sme larvae pupated between the newspaper sheets frm which they were remved and placed in the 'dark-bxes'. Usually, a number f immature larvae were accidently remved frm the 'rearing-cages' with the dead claddes. In the 'dark-bx' they were attracted t the light and they als accumulated in the 'light-trap' frm which they were returned t the 'rearing-cages'. After fur t five weeks, mths ceased emerging frm the 'dark-bx' which was then fumigated fr six hurs with ethyl acetate befre the dead claddes were remved and burnt. There was a cmplete verlap f generatins in the clny and all stages were cntinuusly available fr investigatin. Observatins n the life histry and cntrl ptential f z. tapiacla were cnducted in a cntrlled envirnment rm with diurnal temperature, relative humidity and light fluctuatins as shwn in fig. 2. Mst f the specificity studies were als cnducted n ptted plants in the envirnment rm althugh field tests were als carried ut n caged plants. O. maxima plants used in the experiments were supplied by Mr. G. de Kck, Grtfntein Agricultural Cllege, Middleburg, Cape Prvince, Suth Africa and als cllected at Thrneycrft farm, Grahamstwn, Suth Africa. Ornamental cacti tested were btained frm Bsch Nurseries, Prt Elizabeth, Suth Africa. Mst ther plants were supplied by the Btany Department, Rhdes University, althugh sme were cllected at varius places in the East Cape Prvince, Suth Africa.

16 EGG LAYING MATING MATE CALLING IMAGINAL ECLOSION INACTI VE 12.00h light I la.ooh dark I lig ht OO.DOh 06.00h l2.0dh Fig. 3. Behaviural activity f Z. tapiacla adults ver a 24 hur perid in the cntrlled envirnment rm. Dark bands represent the times at which each behaviur pattern was nted.

17 9 1. THE LIFE HISTORY OF ZOPHODIA TAPIACOLA Dyar (1925), Heinrich (1939;. 1956), Ddd (1940), Mann (1969) and Gunn (1974) describe the adults and larva f Tucumania tapiac1a Dyar (= z. tapiacla) in sme detail. Hwever, their descriptins f the life histry and bilgy f the mth are patchy. Cnsequently a mre detailed accunt f the life histry is presented here as a preliminary t further studies n the bilgical cntrl ptential f this insect. ADULTS The adults f z. tapiacla (see frntispiece) are dark grey clured with white hind wings and black patches n the legs. The fre wings have brken black transverse lines which cntrast nly slightly with the dark grund clur. There is a rw f black dts alng the terminal edge f the wing. The measured alar expanse f 80 mths varied frm 22 t 33mm, and the females are generally larger than the males (Mean female alar expanse = 28,7mm, minimum = 25mm, maximum = 33mm; mean male alar expanse = 25,Omm, minimum = 22mm, maximum = 29mm). The sexes are readily differentiated because the male's labial palps curl up arund the head while thse f the female are prrect (see frntispiece) Observatins were made n the behaviur f 20 pairs f Z. tapiacla adults in cages in the cntrlled envirnment rm and the perids f activity are shwn in fig. 3. Eclsin f the imag ccurred in the evening at any time

18 between 19.30h and 23.00h. The newly emerged mths tk apprximately 30 minutes t dry their wings, after which they became very active, flying and walking arund the cages. Between 23.30h and 02,00h the mths settled dwn and very little activity ccurred. After 02,00h they became active again and during this perid mating ccurred. The z. tapiacla females initiated curtship by adpting a typical lepidpteran calling psture which expsed the sex phermne membranes in the abdmen (Wigg~swrth, 1972). As a result, phermne was released int the cages and males were attracted t the calling females. While appraching its prspective mate, each male beat its wings and waved its antennae until the tw mths came int cntact. Immediately, the females drpped their abdmens and cpulatin cmmenc ed. During sperm transfer, which lasted fr apprximately 25 minutes, bth mths gripped the substrate with their legs while they jined end t end, heads pinted in ppsite directins. After mating the males left the females wh remained statinary until the fllwing night. After mating the males remained active fr a shrt while, but when the lights came n in the rm they settled dwn and bth sexes were inactive thrughut the day. The male mths were never seen t mate with mre than ne female and they died tw t three days after mating. During the day, the mths adpted a vertically upright, cryptic psture (see frntispiece). Mths that were tuched n the psterir part f their bdy j umped away frm the stimulus and fell t the flr f the cage

19 l50~ , c...i "'!J) C) UJ UJ... 0 II:: UJ III ::IE ::I Z Z UJ "' ::IE NIGHTS Fig. 4. The mean number f eggs laid by each f eight z. tapiacla females n eight cnsecutive nights.

20 11 in a cateleptic state. They remained mtinless fr five t ten minutes befre making a brief flight t regain a perch. At apprximately 18.30h the mths nce again became active and egg laying cmmenced. The females walked abut the cages, depsiting eggs n any rugh prjectin, althugh mst were laid n the hanging frayed clth strips prvided. The females always wandered fr a time between each egg depsitin, which tk 3 t 7 secnds, althugh they ften came back t the same spt and laid eggs clse t and n tp f each ther. Eggs were laid thrughut the night except during the midnight perid f inactivity. OVipsitin usually cntinued fr abut eight nights. The mean number f eggs laid by eight females was 231 (minimum 127; maximum - 284) f which 560/0 were depsited n the first egg laying night, 180/0 n the secnd, 0/0 n the third and the remaining 160/0 n the last five nights (fig. 4). The fecundity and egg laying pattern nted here was very similar t that recrded fr T. tapiacla (= z. tapiacla) by Mann (1969). Many bservatins shwed that, virgin z. tapiacla adults were able t mate successfully n any f the first fur nights after emergence. Hwever, they lse the ability as they age and nne f ten male r ten female: mths were able t cpulate when prvided with yung mates after being kept in islatin fr six nights. All the virgin males died within nine days, while the females lived fr ten t thirteen days. Virgin females always discharged 50 t 0 infertile eggs in the 48 hur s p r eceding death.

21 Fig. 5. Scanning electrn micrgraph f a z. tapiacla egg n an O. aurantiaca thrn. Fig. 6. Cmpsite scanning electrn micrgraph f a first instar z. tapiacla larva. Lateral aspect. ET I~ Fig. 7. z. tapiacla tunnel in a bisected O. aurantiaca cladde. ET- escape tunnel; F- feeding area. Fig. 8.. aurantiaca husk pened t sh w a z. tapiacla ccn. EH- escape hle; E emergence tunnel; P- pupa.

22 12 The O,9mm lng eggs are initially cream clured, ageing thrugh brwn t dark brwn prir t hatching. They have a characteristic sculptured chrin (fig. 5). A system f ridges frms a netwrk f triangles raised abve triangular depressins with aerpyles at the intercese f the ridges. Each egg is pressed against and glued n a thrn s that the cntact surface becmes cncave t accmdate the curve f the thrn. LARVAE After an incubatin perid which always lasted ten days in the labratry, the yung larvae chew their way ut f the egg. Each first instar larva (fig. 6) is apprximately 1, 9mm lng at hatching and it has a white bdy with a black 'head. The larvae are very mbile and they wander abut spinning a silk thread n which they can lwer themselves if disturbed r expsed in an unsheltered area. rhe larvae have a negative phttactic respnse n the hst plant which leads them t the lwer shaded parts f the plant were they seek a suitable entry pint fr initial feeding and penetratin f the cuticle. Table 1 shws that mst z. tapiacla entry sites n a cladde were situated n the shaded side whether the light was directed frm abve r belw. In ttal dark with n shaded area, the larval entry pints were distributed evenly arund the whle cladde.

23 a b c d '.~~.""'_ ;I ~1'.'.' ~ "'"./' r ~. ~. ".. e.' " l!;;. ~ t... ~ Fig. 9. The behaviural sequence fllwed by Z. tapiacla during entry int a hst plant. The larva spins a 'silk-tent ' (a) which it pushes against t pierce the cuticle with its mandibles (b). The larva withdraws frm the tunnel t regur gitate ingested tissue (c) until its head and thrax are accmmdated (d). Excavated tissue is then depsited as frass utside the tunnel (e & f). Finally the entrance is plugged with silk and frass (f). ( x6 ). E- entrance; F- frass; p- plugged entrance hle; R- regurgitated tissue; S- silk. f

24 13 Table 1. The number f Z. tapiacla entry pints n upper and lwer surfaces f O. aurantiaca claddes with illuminatin frm abve and belw r in ttal darkness. NUMBER OF ENTRY POINTS CONDITION LOWER UPPER TOTAL x 2 PROBABILITY SURFACE SURFACE LIGHT ABOVE ,612 P =<0,001 significant LIGHT ,8236 P=>0,001<0,01 BELOW significant TOTAL DARK ,025 P =>0,1<0,9 nt significant Having selected a suitable entry site the yung larva spins a lse 'silk-tent' arund itself (fig. 9a). An uneven area f the plant is always utilised because the 'tent' cannt be cnstructed n a flat surface. Observatins f 250 larval penetratins shwed that 63% entered the plant alngside an arele using the glchids and thrns as a supprt fr the 'tent ' ; 30% spun silk acrss the depressin at the internde f tw adjining claddes and 7% entered the plant at rugh, scar tissue. Figure 9a-f shws the sequence adpted by a Z. tapiacla larva penetrating a jinted cactus cladde at a n arele. The larva spins a number f threads between the thrns and the cuticle, until it is enclsed in a lse silk mesh. At this stage the larva arches its bdy and pushes against the 'tent' (fig. 9b) using it as a brace t frce its mandibles dwn int the cuticle. T facilitate penetratin, the larva chews the cuticle exterir while pushing against the 'tent'. Often, initial attempts t clnise the cladde fail because

25 14 the 'tent' is nt substantial enugh and t little frce is transferred t the mandibles. After several unsuccessful penetratin attempts, the larva reinfrces its 'tent' and tries again t pierce the cuticle. The prcess may be repeated 5 r 6 times, but if n success is achieved after tw r three hurs the larva vacates the 'tent' and begins anew elsewhere. Once the mandibles pierce the cucicle f the hst plant the larva chews deeper int the cladde. Initially the insect -swallws masticated tissue until its fregut is full, it then withdraws frm the tunnel and regurgitates the tissue n the plant surface, (fig. 9c) befre tunnelling cmmences. In this way a hllw is excavated in the cladde which sn accmmdates the head and thrax f the larva (fig. 9d). As the larva tunnels deeper int the plant, fd is passed thrugh the gut and depsited n the utside f the plant as frass (fig. ge & f). After ne t tw hurs the whle larva is able t withdraw int the newly excavated tunnel. Immediately, it turns arund in the tunnel and plugs the entrance hle with a silk net intermeshed with frass (fig. 9f). Feeding, and cnsequent widening f the tunnel, cmmences nce the entrance hle has been sealed. Frass is depsited in the tunnel and is accumulated near the entrance pint. As the larva ages, a verticle escape tunnel leading dwnwards is excavated away frm the feeding site (fig. 7), s that by the third larval instar the escape tunnel usually traverses the cladde frm the feeding site t its lwest extremity.

26 15 Zimmermann (1968) and Mann (1969) have nted in Argentina and Australia that Z. tapiacla larvae cnstruct a silk 'escape' tube frm the invaded cladde "several inches" int the grund. When the clnised cladde is disturbed and during perids f intense heat, the larva retreats int this grund tube. As the larva develps, the feeding and frass depsitin area is extended. The frass is further brken dwn by bacterial actin after which it desiccates and shrinks, frming large hllw cavities in the drying husk. When the whle cladde has been ingested, and the frass has desiccated, all that remains is a hllw husk f cuticle, partially filled with dry frass, cntaining the larva which mayr may nt have cmpleted its develpment. Sme larvae clnise claddes which are nt large enugh t supprt them fr their cmplete develpment. In these cases the larvae can vacate the ld destryed cladde and wander sme distance fr three t five days in search f a new hst, which is then clnised. z. tapiacla larvae mult five times and pass thrugh six instars (see Appendix 1). Mann (1969) described the larvae as being varied in clur frm dark red t purple red, light wine clured, pink, yellw brwn, light range r light yellw. The larval clurs in fact change predictably with age. The white first instar larva turns a pale pink and darkens t a deep red by the third instar. They remain dark red until the final instar when they becme lighter and turn a pale yellw prir t pupatin. The final instar larvae

27 16 attain a maximum length f apprximately 25mm and by the time they are ready t pupate the ccupied cladde is cmpletely hllwed ut and destryed, s that the husk can serve as a pupatin chamber. The larva will nt pupate in a cladde that is nly partially destryed but will vacate it t pupate in the leaf litter r sil. The reasn fr this seems t be that any green tissue that remains in the claddes after larval feeding is cmplet~ is susceptable t bacterial infectin and excessive rt which may be detrimental t the pupa. PUPA Nrmally, the final instar larva chews'an escape hle in the drying cuticle f the cladde. It then cmmences ccn prductin which begins with the prepupa spinning a lse silk net t seal the escape hle. The prepupa then retreats ne t tw centimeters int the cladde where it spins a substantial pupatin chamber which is cnnected t the escape hle by a flimsy silk guide tunnel cnstructed by the prepupa as it backs int the cladde (fig. 8). In the labratry, the prepupa tk frm 36 t 48 hurs t cnstruct their ccns, after which they pupated in the pupatin chamber. Eclsin f the imag tk between 16 and 26 days (Mean 21 days) in the labratry. Each emerging mth is directed t the escape hle in the husk cuticle by the guide tunnel. At the escape hle, it lsens the sealing silk threads with a wetting fluid, befre it crawls ut nt the husk and finds a thrn, r ther prtrusin, n which t spread and dry its Wings. The mean develpmental duratin

28 17 f rm egg hatch t adult eclsin was 81 days fr males and 9 5 days fr females. This difference is nt significant (P =>0,2<0,4).

29 18 2. THE HOST SPECIFICITY OF ZOPHODIA TAPIACOLA Zwlfer & Harris (1971) list a number f methds t determine the hst specificity f an insect. One f these is t cnduct starvatin and negative-vipsitin tests n ecnmic plants. They pint ut, hwever, that the artificial nature f the tests can easily lead t misinterpretatin f the true facts. Cnsequently, ecnmic plants that are eaten r used fr egg depsitin in cnfined cages are nt necessarily suitable hsts because in the field there are many islating mechanisms that restrict insects t particular plffilts (Huffaker, 1959; Harris, 1963). Hwever, the tests are f value in that they shw plant defences which repel insect attacks and als sme f the plants that will nt be utilised, even when the insect is starving. As nt every ecnmic plant can be screened, Harris and Zwlfer (1968) and Wapshere (1974) have utlined principles that shuld be fllwed when chosing plants fr the tests. Starvatin and negative-vipsitin tests were cnducted n Z. tapiacla using tw grups f plants; (il members f the Cactaceae, because they are btanically related t O. aurantiaca, (ii) Suth African indigenus SUCCUlents, because they have never been expsed t Z. tapiacla, and there is always uncertainty abut behaviur f phytphagus rganisms faced with a new plant species (Wapshere, 1974). The tests were nt extended t ther plants r grups f plants fr the fllwing reasns. (a) Ddd (1940) reprts that

30 19 Z. tapiacla never utilised beneficial nn-cactaceus plants during hst specificity tests in Australia. (b) Extensive surveys in Suth America have nly recrded Z. tapiac1a n a few lw grwing cacti (Mann, 1969; Zimmermann, 1972). (c) Almst withut exceptin, cactus feeding insects d nt feed n plants utside the cactaceae and vice versa; this rule is especially true fr the cactus phycitids (Mann, 1969). 2.1 Starvatin tests All starvatin tests were cnducted in the cntrlled envirnment rm. Fr each test, ten newly hatched Z. tapiacla larvae (less than 5 hurs ld) were placed n actively grwing ptted plants pssessing yung flush and mature fliage. Every tw days the plants were scrutinised fr signs f feeding r successful clnisatin. The results f the tests are summarized in Table 2. Apart frm slight 'nibbling' n three f the Euphrbia spp. there was n sustained feeding n these, r any f the ther Suth African succulents, prbably because, like ther lepidpterus larvae (Schnhven, 1973), Z~ tapiacla may have a variety f chemical and physical requirements which are nly met by nrmally acceptable hst plants, in this case certain species f cacti. Attempted feeding and different degrees f develpment ccurred n nearly all the cacti presented during the tests. Hwever, there were a few genera that Z. tapiacla never attempted t clnise (e.g. Epiphyllum spp., Pereskia spp.) and the nly species

31 Table 2. Results f starvatin tests n z. tapiacla. The clumns indicate; i. the plants. tested; ii. their ecnmic status (if any); 111. the number f replicates fr each test; iv. the extent f clnisatin and develpment n each plant (0 - N penetratin r feeding attempted; X - N develpment; A - attempted feeding; B - larva penetrated int plant tissue; C - larva develped; D - larva pupated; E - adult emerged); v. Remarks. i PLANT SPECIES ii IMPORTANCE iii NO. OF REPLI CATES iv COLONISATION & DEVELOPMENT v REMARKS CACTACEAE Opuntia aurantiaca Lindley Opuntia rsea D.C. Opuntia imbricata (Harv.) D.C. Opuntia mncantha Hawrth Opuntia micrdasys (Lehmann) Pfeiffer Opuntia verschaffeltii Cels Opuntia cylindrica (Lamark) De Candlle Harrisia martinii (Laburet) B. & R. Weed Weed Weed Weed Weed Weed Weed Weed ABC DE ABC DE ABC D E ABC D E ABC D E ABC D E ABC DE ABC D E All mature claddes acceptable fr clnisatin. Each larva destryed a relatively large part f the plant during its develpment. Opuntia megacantha Salm Dyk Opuntia tardspina Griffiths Opuntia maxima Mill. (Burbank Spineless Cactus) Weed Weed Fdder A B A B A B X}-High gum cntent in mature X claddes prevented larvae frm clnising and X damaging these plants.

32 Pereskia aculeat a Mi l l Rhyps al idpsis rsea (Lagerheim) B. & R. Rh ypsal i s cas sutha Gaertner Rhypsali s spp.( 3) Schlumbergera sp. Trichcereus candicans (Gillies) B. & R. Trichcereus spachianus (Lamaire) Riccbn Trichcereus thelegnus (Weber) B. & R. Webercereus sp. Zygcactus truncatus (Hawrth) Schumann We ed Ornament al Ornamental Ornamental Or namental Ornamental Ornamental Ornamental Ornamental Ornamental AB C ABC ABC ABC A X A ABC ABC Larvae nly able t develp up t third ins tar after which t h e thin stems f these species were nt suitable fr develpment. Fluid gum repelled larva. Desiccatin f plant at wund prevented entry f larvae. Whle plant rtted befre larvae develped. Stems t thin t supprt cmplete develpment. AIZOACEAE Carprbrtus edulis (L.) L. Bl. Indigenus Cnphytum spp. (2) Indigenus Delsperma echinatum (Ait. ) Indigenus Schwant Fenestraria spp. (3) Indigenus Glttiphyllum lnqum (Haw. ) N. E. Br. Indigenus Lit hps spp. ( 4) Indigenus

33 Table 2. Cntinued. PLANT SPECIES IMPORTANCE NO. OF REPLI CATES COLONISATION & DEVELOPMENT REMARKS Lema rcereus pruinsus Ott Lbivia sp. Ornamental Ornamental Lphcereus schttii (Engelmann) Ornamental B. & R. Malcarpus sp. Ornamental ABC Yung sht clnised but nt main plant. Mnvillia spegazzinii (Weber) B. & R. Nemammilaria clumbiana (Salm Syk) B. & R. Nemammilaria cllinsii B. & R. Nemammilaria galetii (Scheidweiler) B. & R. Nemammilaria haageana (Pfeiffer) B. & R. Nemammilaria magnimamma (Hawrth) B. & R. Nemammilaria pringlei (Culter) B. & R. Nemammilaria spp. (4) Nepteria spp. (3) Pachvcereus pecten-abriginum (Engelmann) B. & R. Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental 5 5 ABC AB C ABC A B C~ ABC A B X Lcalised rtting prevented cmplete develpment. Plants rtted befre larvae matured. N larval develpment.

34 Ast rphytum myristigma Lauaire Ornamental As trphy tum rnatum ( D. C.) Weber Br zicact us sp. Cephalcereus palmeri Rse Cereus jamac aru De Candlle Cer eus spp. (2)* Cleistcactus spp. (2) Cryphanta spp. (2) Dlicthele sp. Echincactus grusnii Hildmann Epiphyllum sp. Echinpsis rhdtricha Schumann Erdisia sp. Espsta sp. Fercactus flavvirens \Scheidweiler) B. & R. Fercactus latispinus (Hawrth) B. & R. Gymncalcium sp. Hamatcactus setispi nus (Engelmann) Or namenta l Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Ornamental Plant rtted befre A BC l arv a ma t u r ed. A B X A B X~ Larva repelled by f luid A B X. gum exuded at wund in X respnse t attack. A B Dense glchid cncentratin prevented larvae reaching plant cuticle. ABC D E Only c;me individual reached ABC ABC ABC), ABC matur~ty. Plant rtted befre larva matured. Plants rtted befre larva matured. * Number s in brackets rpp re se~t the number f undeter.~ined species tested in sme genera.

35 Table 2. cntinued PLANT SPECIES IMPORTANCE NO. OF REPLI CATES COLONISATION & DEVELOPMENT REMARKS Hesembryanthemum sp. Ophthalmphyllum sp. Indigenus Indigenus APOCYNACEAE Pachypdium bispinsum (L.f.) A.DC. Indigenus pachypdium succulentum A.DC. Indigenus ASCLEPIADACEAE Caralluma keithii R.A. Dyar Stapelia spp. (2) Indigenus Indigenus COHPOSITAE Othnna carnsa (L.f.) Less. Seneci sp. Indigenus Indigenus CRASSULACEAE Bryphyllum tubiflrum Harv. Crassula lactea Sland. Weed Indigenus

36 Crassula rnulticava Lam. e t Versch. Crassula prtulacea Lam. Crassula? spatulata Thunb. Crassula? tetragna L. Crassula spp. (6) Ctyledn rbiculata L. Kalanche? hirta Harv. Indigenus Indigenus Indigenus Indigenus Indigenus Indigenus Indigenus EUPHORBIACEAE Euphrbia gradidens Haw. Euphrbia hrrida Biss. Euphrbia pugnifrmis Biss. Euphrbia stellaspina Bj. Euphrbia spp. (3) Indigenus Indigenus Indigenus Indigenus Indigenus A few web tents spun using thrns f these species, but attempted entries were terminated after the cuticle had been pierced. LILIACEAE Ale? arbrescens Mill. Ale ferx Mill. Agave americana L. Bulbine natalensis Bak. Indigenus Indigenus Extic Indigenus PORTULACACEAE Prt ulacaria afra Jacq. Indigenus ~ dd er 0-- Imprtant grazing p lant (aatten & Bk elrnann, 1966).

37 20 that supprted cmplete develpment were O. aurantiaca, O. rsea, O. imbricata, O. mncantha, O. micrdasys and H. martinii. These six species are the same, r clsely related, plants t thse clnised by z. tapiacla in Argentina and Australia. There were tw larvae that cmpleted their develpment n rnamental cacti, ne n Cryphanta sp., and the ther n Nemammilaria sp., hwever, bth cases were exceptinal. z. tapiacla was nt able t clnise yung flush grwth f any f the cacti, including the acceptable hsts. The reasn fr this failure was that the water retentin mechanism f the flush was destryed when the cuticle was pierced by the attacking larva. As a result, rapid desicca- tin ccurred which caused the tissues arund the wund t shrink and frm a hard callus. The yung larvae either were trapped in the hardened tissue and died, r they were frced ut f their tunnel and subsequently left the plant. The majrity f the cactus species presented t z. tapiacla in the tests supprted sme degree f larval develpment, althugh they were nt successfully clnised fr a number f reasns, sme f which were nted and listed belw. (i) The glchids f the Cleistcactus spp. are very dense and t hey seemed t frm an impenetrable mesh which prevented z. tapiacla frm achieving the plant cuticle. (ii) The narrw stemmed cacti (e. g. Rhypsalis spp.

38 ~ a Fig.. Gum exuded by spineless cactus as a result f cntinued clnisatin attempts by Z. tapiacla. Fig. 11. A first instar Z. tapiacla encapsulated by spineless cactus gum exuded in respnse t damage by the larva. I~ Fig. 12. Mature spineless cactus frming new claddes after numerus Z. tapiacla larval clnisatin attempts which resulted in large amunts f exuded gum.

39 21 Rhypsa1idpsis spp., Schlumbergera ~. and Zygcactus spp.) were clnised by the small first, secnd and even third instar l arvae. Hwever, the stems were t narrw fr the tunnel s f large, lder individuals which cnsequent ly vacated the plants. (iii) Many cactus species (e.g. Nemammilaria spp., Astrphytum spp. Fercactus spp. ) are very susceptable t bacterial infectin s that when they were clnised by Z. tapiacla, rt usually set in, which brught abu t cmpletb _ cllapse f the plant befre the larva matured. (iv) The ecnmically mst imprtant cactus, O. maxima, a ln -I with the clsely related O. megacantha and O. tardspina, a b well as the Cereus spp., all resisted Z. tapiacla larval clnisatin by exuding a sticky mucilage at the site f a t tack (fig. ). One t three days after the larvae had bee n released nt O. maxima plants, gum appeared and zed t hrugh the damaged cuticle where the larvae had attempted entry. The gum filled the newly excavated tunnels and frced the larvae ut. Smetimes, the larvae were caught up by the sticky substance and being unable t escape they eventually died in the hardened gum (fig. 11). Usual ly, hwev er, the larvae '~ere driven frm their tunnels, whereupn they vacated the plant, r ccassinal ly tried unsuccessfully t penetrate the same plant elsewhere. Althugh larvae were able t excavate small tunnels int the spineless cactus pla nts, the claddes were nt detectably affected by the slight damage a nd subsequent gum prductin. Even claddes that had been a t tacked cntinuusly by numerus larvae (mre than 0) -:k",.lped nrma l ly and gave rise t new shts (fig. 12).

40 Negative-vip sitin tests The ffspring f the mst ligphagus insects will nt survive unless the females place their eggs in situatins s u itable fr the develpment f the Fl generatin. As a result, vipsitinal behaviur in these insects has evlved a cmplex pattern which is dependant fr its release n the receptin f suitable stimuli, usually prvided by acceptable hst plants (Dethier, 1959; Gupta & Thrsteinsn, 1960; Thrsteinsn, 1960; Beck, 1965; Stadler, 1974). The degree f r estrictin t particular plants des nt seem t be abslute a nd bservatins n varius hst specific Lepidptera species by Dethier (1959) have shwn that females d nt always chse the crrect hst plants, but eggs may als be depsited n a variety f bitic and abitic substrat es. Hwever, the eggs that are nt laid n r near the crrect hs"t plants appear t be placed at randm and nt exclu sively in particular alternative situatins. Wiklund (1974) sites examples f species that will nt vipsit n plants which are nutritinally suitable fr larval develpment. Ovipsitinal specificity in many species must therefre be the k e y factr determining the spectrum f plants clnised. Cnsequently, a ptential bilgical cntrl insect which feeds n beneficial plants during starvatin tests might be shwn t be safe by determining whether r nt the adult will vipsit n thse species in the field. Attempts were therefre made t try and determine which plants are chsen by Z. tapiacla f r egg depsitin in the hpe that its hst r a nge may be further substant iated.

41 23 OVipsitinal chice tests were attempted under cntrlled l abratry cnditins where adult females were cnfined in perspex cages with nyln gauze lids. Further tests were cnducted in a 3 x 3 x 2 meter nyln gauze field cage. During tests, Z. tapiacla adults were presented with. aurantiaca and ther Cactaceae, as well as representatives frm the Euphrbiaceae, Liliaceae, Aizaceae and Prtulacaceae. In bth cage situatins, females depsited eggs equally frequently n all the plants, n the gauze sides and lids and.n any small prjectin. The results suggest that either Z. tapiacla lays its eggs haphazardly thrughut its habi tat r alternatively, the cnfine s f the test cages caused the mths t behave atypically frcing them t vipsit n nrmally unacceptable substrates. Zwlfer and Harris (1971) pinted ut that in many attempts t determine the vipsitinal specificity f Lepidptera, the cnfining nature f test cages h as led t cmpletely artificial results. Z. tapiacla, which h as slitary feeding larvae living in small hst plants, may be expected t have a dispersal mechanism built int its vipsi tinal behaviur pattern s that it spreads its eggs ar und, because each plant can nly supprt a limited number f larvae. As a result, there may be sme essential activity by the mth between each egg depsitin and this activity, su ch as a perid f flight, might be necessary t initiate t h e next egg depsitin. The cnfines f the test cages wuld seriusly affect such a behaviural respnse and lead t artificial vipsitin. Whatever the explanatin, n pre f erence fr any particular plant type culd be detected and

42 24 vipsi tin al chice tests were discntinued. Althugh n distinctin was made abut the vipsitinal specificity f Z. tapiacla, when the mths laid n jinted cactus claddes lying n the flr f the cages, they always depsited mst f their eggs n the thrns situated n the lwer surfaces f the claddes (Table 3). This trend was the same bth in dim light and in ttal dark (p =>0,1<0, 9) and did nt appear t be a phttactic respnse. Als, the mths. seldm laid eggs n claddes which were suspended in mid air by threads frm the ceiling f the cages. Less than % f the ttal eggs laid by ten females were depsited n hanging claddes, mst being placed n the cage walls. These bservatins suggest that there may be a behaviural respnse by Z. tapiacla females t place their eggs in sheltered situatins, a feature which may limit them t small, lwgrwing thrny plants in the field. If this cnclusin is crrect and is nt an artifact f the caged cnditins, then vipsitinal behaviur may be a majr factr limiting Z. tapiacla t small O. aurantiaca type plants in the field. Table 3. The number f eggs laid by Z. tapiacla n the thrns f the lwer and upper surfaces f O. aurantiaca claddes lying n the cage flrs in dim light and in ttal dark. LIGHT CONDITION SURFACE DIM DARK UPPER LOWER NUMBER OF FEMALES 12 20

43 25 3. THE EFFECTS OF ZOPHODIA TAPIACOLA ON OPUNTIA MAXI~ffi Defensive gum secretins initiated by insect damage t Opuntia spp. has been dcumented previusly by Pettey (1947) wh nted that C. cactrum was nt able t clnise. tardspina because the first instar larvae were repelled f rm the plants by a mucilagenus sap secreted at the entry w u nd. The same secretins in mature O. megacantha and O. maxima ccassinally frced C. cactrum larvae t vacate t h eir feeding sites. He als bserved that the eggs f the cerambycid brer, Lagchirus fernestus, which were vipsited i n grves excavated by the adults in. megacantha and O. maxima, were ften encapsulated in mucilagenus sap which hardened and suffcated them befre they culd hatch. Gum p rducti n is prbably the mst imprtant feature preventing Z. t apiacla frm clnising O. maxima. Cnsequently, after the mechanism had been detected during starvatin tests, further tests were cnducted t determine the efficiency f i ts defensive value. 3.1 The gum secreting O. maxima cultivars The species,. maxima, embraces a wide range f spineless cac t us varieties r cutivars, f which five r six are grwn extensively fr fdder in the drier areas f the cuntry. One f these cultivars, Chic, shwed during the starvatin test s that i t is able t resist Z. tapiacla clnisatin attempts in t h e labratry by secreting defensive gum. Hwever, beca u se the five ther varieties, Castill, Fusicaulis,

44 26 Mnterey, Mrad and Rbusta are als f ecnmic imprtance and because z. tapiacla may attempt t clnise all six varieties in the field, further tests were cnducted t establish whether the defensive gum_is effective in the field and whether it is fund in all the ecnmically beneficial sp ineless cacti. First instar z. tapiacla were released nt established spineless cactus plants grwing in a cultivated patch. Each p l ant was caged with twenty individuals and after ten days the number f attempted entries was nted (Table 4). Table 4. The number f attempted and successful entries by z. tapiacla first instar larvae n caged plants f different O. maxima cultivars in the field. N. OF LARVAE/ TOTAL N. ATTEMPTED SUCCESSFUL CULTIVAR PLANTS PLANT LARVAE ENTRIES COLONISATIONS castill Chic rusicaulis Mnterey Mrad Rbusta TOTALS A successful entry is ne in which the larva can penetrate the plant cuticle and excavate a tunnel which will accmdate it during its develpment. Out f 1420 first instar z. t apiacla larvae released nt 71 plants, made up f six O. maxima varieties, nne were able t clnise any

45 27 f the plants and all were repelled by secreted gum. Gum secretin in respnse t wunding by z. tapiacla is therefre characteristic f the six mst imprtant cultivars f O. maxima, and it may be universal thrughut all the varieties, including the less imprtant nes. The 82 attempted entries recrded abve, represent less than 6% f the ttal number f larvae released. The tugh cuticle f O. maxima, and the inability f z. tapiacla larvae t penetrate it, prbably accunts fr the lw percentage attack. The plant cuticle as a defense mechanism is discussed later (Chapter 3.4) The efficiency f defensive gum in O. maxima Althugh O. maxima and O. megacantha repel slitary feeding z. tapiacla larvae, they are bth successfull y clnised by the immature stages f the pyralid m t h, Cactblastis cactrum Salm Dyk. c. cactrum lays 'sticks' f up t 150 eggs (Ddd, 1940) and the hatching larvae all aggregate at the base f a thrn, usually near the ne n which the eggs were laid. The size f the grup depends n the number f eggs, but there are usually 70 t 90 individuals which pierce the cuticle and enter the plant at the same place. Cnsequently, the grup is able t penetrate the plant tissue mre rapidly than the relatively slw-feeding, slitary i ndividuals f z. tapiacla. The speed at which C. cactrum -tunnels int the plant is quicker than the plant can respnd by secreting gum and therefre the larvae are nt frced ut

46 0 15 3: m > Z *- f- Z z 75 c "' -' 0- m 15 ~ (!) Z ~ l&:: (J ~ 20 (") f- "' 50!;i X 5 G'l w,, ~ II::,,,, -' "',, f-, z 25, w,, (J,, II:: w X "'J( 5 3: til z ",... " > z -t I LARVAL NUMBER ON PLANT Fig. 13. The percentage and number (mean +1 S.E.) f z. tapiacla attacking. maxima when 20, 50 and 0 larvae were released n individual frty centimeter tall plants. Number f replicates is represented next t each pint.

47 28 f their t unnel. The gregarius habit allws C. cactrum t clnise plants which it wuld nt be able t if it was a slitary feeder like z. tapiacla. The impact f large grups f z. tapiacla larvae n spineless cactus was determined by releasing grups f first instar larvae nt 30 t 40cm tall ptted O. maxima plants in the labratry, after which their prgress was recrded. The mean number f larvae attempting t enter each plant was practically the same when 20, 50 and 0 larvae were released n the plants and the percentage attempted clnisatins was inversely related t the number f individuals released n the plants (fig. 13). These trends suggest that there were a limited number f suitable entry pints (e.g. thrn bases and interndes f claddes) at which z. tapiacla culd penetrate the cuticle and nce these were ccupied, the remaining larvae culd nt attempt entry. Hwever, f majr imprtance is the fact that nne f the attempted penetratins was ever successful because the larvae were repelled by the defensive gum secretins characteristic f O. maxima. Als, z. tapiacla did nt aggregate when present in large numbers and never adpted a gregarius habit which may therwise have enabled then t have entered and successfully clnised the spineless cacti The surce f gum in O. maxima Gums and mucilages are lsely defined terms applied t the plysaccharides prduced by plants. Mucilages differ frm

48 MUCILAGE CELLS ruptured by mechanical damage 1 2 NON- MUCILAGENOUS PLANT CELLS I damaged discharge mucilage gummsis (Le. enzymatic break dwn f cellular cmpnents t amrphus / EXUDED GUM I 4 \ plan! injured, resevirs release gum plysaccharide) released frm damaged cells POLYSACCHARIDES IN CELL WALLS Fig. 14. / "" plan! damaged frm gum rest'vi rs nrmal develpment "" / PLANT TISSUES Fur different surces f plysaccharides which cntribute t gum exudates frm flwering plants.

49 29 gums in that they are usually less cmplex, they are maintained within the plant and they are prduced as a result f nrmal metablism. Gums n the ther hand are mre chemically cmplex and are exuded frm the plant as a result f injury. The functin and rigin f gums in plants is nt clearly understd (Smith & Mntgmery, 1959r Aspinall, 1969) hwever, the surce f gum exuded at the wunds f injured plants appears t riginate via ne r mre f fur basic pathways (fig. 14) (Haberlandt, 1928r Fritsch & Salisbury, 1946r Esau, 1964r Fahn, 1967r McLean & Ivimey-Ck, 1967). The fur pathways are as fllwsr ('i) many plants, including the Opuntia spp. and mst ther cacti, have mucilagenus cells scattered thrughut their grund tissue (Bdle & Fritsch, 1908r Metcalfe & Chalk, 1950). The frmatin and functin f these cells is nt clearly understd, (Sctt & Bystrm, 1970) but when they are ruptured they may discharge their cntents and cntribute t the exuded gum. (ii) When the cells f plants are damaged, they underg an enzymatic breakdwn which cnverts all the cntents int an amrphus plysaccharide. This prcess is knwn as gummsis and it may cntribute t the gum. (iii) The cell walls f plant cells are generally very mucilagenus and this mucilage may be released When the cells are brken dwn after damage. Mucilage frm this surce will mix with the prducts f gummsis and may be exuded frm the plant. (iv) Finally, sme plants, including the Opuntia spp.and ther cacti (Stewart, 1919r Archibald, 1936), have gum canals, r resevirs,

50 A B Fig. 15. Cut surface in frntal view (A) and side view (B) f O. maxima shwing gum frced ut f the gum canals by internal pressure. SD MC Fig. 16. A hand sectin thrugh O. maxima tissue at the site f a z. tapiacla tunnel excavated alngside scar damage (x). MC- mucilage cells; SD- scar damage; T- tunnel.

51 30 which are elngated pckets filled with gum and are usually rientated with the lng axis f the stem and assciated with the vascular tissue. Gum-canals are frmed naturally in sme plant species, including the cacti, while in thers they develp as a result f physilgical, chemical r disease damage t the plants tissues. Wunding f the plant results i n the gum, which is under pressure in the canals, being exuded (fig. 15). Any, r all, f the abve pathways culd be respnsible fr the gum prduced by O. maxima when damaged by z. tapiacla. Hwever, there is substantial evidence which suggests that nly the gum-canals (i.e. pathway 4) cntribute significantly t the exudate. Firstly, micrscpic sectins f O. maxima and. aurantiaca stained with th{in (Sctt & Bystrm, 1970) shw that there are apprximately equal cncentratins f mucilagenus cells in bth cacti. If these cells were respnsible fr the prductin f repelling gum in O. maxima, while nt ding s in O. aurantiaca, then there wuld have been a prliferatin f them at the site f damage in O. maxima and nt in O. aurantiaca. Sectins thrugh z. tapiacla excavatins in O. maxima (fig. 16) shw that there is n such prliferatin and that the amunt f gum prduced (figs & 12) culd nt have riginated frm the few ruptured mucilagenus cells in the vicinity f the wund. Gummsis is nt likely t cntribute much t the exudate because all the damaged cells are ingested by the larva and their cntents are nt therefre available fr enzymatic

52 Table 5. The crss sectinal areas f gum canals in O. maxima and O. aurantiaca. Means were derived by sectining five yung flush, five mature and five ld wdy claddes fr bth species, and measuring the fur largest canals in each. CANAL CROSS SECTIONAL AREA mm 2 NO. OF TIMES O. MAXIMA> PLANT STAGE O. MAXIMA O. AURANTIACb O. AURANTIACA FLTJSH (still bearing 0,30 0,03 fleshy leaves) MATURE (fully grwn, 2,43 0,11 22 less than tw years ld) OLD (wdy, hard, 7,07 0,41 17,3 basal claddes. Mre than tw years ld)

53 31 breakdwn and mucilage frmatin. Further, gummsis is a universal phenmenn thrughut the plant kingdm, prbably ccuring equally in bth O. aurantiaca and O. maxima. It is therefre unreasnable t assume that the prcess may be respnsible fr prducing repelling gum in ne and nt the ther. The nly remaining surce f exudate in O. maxima is frm the gum -canals. Gum canals ccur in O. maxima and. aurantiaca but the number per c1adde in bth species varies cnsiderably. In large O. maxima pads cunts f between 20 and 40 canals per c1adde have been made, while O. aurantiaca has between 5 and 15 canals in each f its smaller claddes. Gum canals are differentiated early in the develpment f the flush and they are readily discernab1e in sectins f yung claddes. The canal size seems t be the critical factr as t whether gum will be exuded r nt. Measurements f canals in O. maxima and O. aurantiaca (Table 5) shw that the canals get brader as the plants age, hwever, in. aurantiaca the gum canals never reach the same dimensins as thse in O. maxima. Mature O. maxima has large canals which always exude enugh gum t repel z. tapiacla attacks. O. maxima flush and mature O. aurantiaca, n the ther hand, nly ccassinally repel z. tapiacla, prbably because they have small canals which d nt stre enugh gum t frm an exudate. (Only twice during hundreds f bservatins ver tw years has z. tapiacla been seen t be repelled by gum in O. aurantiaca).

54 32 The penetrating larva des nt have t actually rupture ne f the canals in. maxima t initiate exudatin. As sn as the cuticle f an. maxima plant is pierced, a lw pressure area is created in the tissues. Immediately, gum starts flwing frm the high pressure canals thrugh the i ntercellular spaces t the wund. Cnsequently, the larvae are always repelled even when they excavate their tunnel sme distance away frm the clsest gum canal. The delay f ne t three days between the larva piercing the cuticle and the appearance f gum might be explained by the delay fr gum t frce its way thrugh the intercellular spaces. 3.4 Cactus cuticle as a barrier t Z. tapiacla larvae The tughness f plant cuticles has been implicated in preventing r reducing insect feeding (Tantn, 1962; Agarwal, 1969; Feeny, 1970; Suthwd, 1973). The cactus feeding pyralid brers may als be affected by cuticular tughness because Pettey (1947) nted that C. cactrum clnised yung O. megacantha claddes mre successfully than they did lder nes. Als, Mann (1969) recrded that while rearing Z. tapiacla prir t its release in Australia, the first instar larvae had t be prvided with yung jints f O. aurantiaca as they were unable t penetrate the lder nes. These bservatins were extended during this investigatin. The number f larvae penetrating jinted cactus was recrded by releasing sixty first instar Z. tapiacla nt fifty O. aurantiaca claddes f differing cuticular thickness.

55 33 (Abercrmbie et al, 1966 defined the cuticle as the superficial, nn-cellular layer which frms a cntinuus cvering ver the aerial parts f the plant). All the claddes were in cntact with each ther s that the larvae culd chse thse claddes prefered fr clnisatin. Ten days after release, the number f larvae in each cladde was recrded and the cuticular thickness near the penetratin site was measured. Thickness was measured by stripping lcm2 sectins f cuticle frm the plants. Any "green tissue attached t the sample was scraped ff lightly with a sharp blade befre measurements were made using a Mercer micrmeter accurate t 2pm. The mean f three t five measurements n each blck f cuticle was taken. The relative tughness f the same r similar blcks was measured using a penetrmeter which recrded the mg pressure needed t pierce the cuticle with a flat-tipped 'minuten' pin f 0,254mm diameter (Mran & Buchan, 1975). Mst f the larvae (77%) penetrated claddes with a cuticle that was less than O,lmm thick, while n larvae penetrated claddes with a cuticular thickness f 0,15mm (Table 6). Cuticular thickness seems therefre t limit yung z. tapiacla larvae t particular claddes. Cnsequently, the thickness and tughness (= penetrability) f. maxima and O. aurantiaca cuticle was measured and cmpared with respect t their vunerability t z. tapiacla larvae.

56 0,3 y: O,0047x + 0,0795 r:: 0,6951 0,2 - E E ~... (J) (J) Z :.:: U ::r:: ~ 0:: '"... ::l U 0,1 0,3 0,2 j: 0, 1 ::l U Z 0{ w :IE 0,3 (i) Y:: 0,0097X + 0,0548 r.. 0,6865 X......,. O. MAXIMA (i i) Y:: 0,1199x + 0,0002 r (i i) O. AURANTIACA 0,1. maxima 0,1 ~~= au-,-a-n~ti-a-ca CLADODE LENGTH (em) 30 Fig. 17. vari atin in cuticular thickness with cladde age ( = l ength; see text) f r O. maxima and O. aurantiaca. Each pint represents the mean f three t five measurements.

57 34 Table 6. Z. tapiacla larval numbers clnising O. aurantiaca claddes f different cuticular thickness. The numbers f claddes in each grup are in parentheses. Penetratins at aereles, intendes and damaged areas were nted. COLONISATIONS CUTICLE THICKNESS AT AREOLE AT INTERNODE AT DAMAGE TOTAL rm 0,09 ( 8) , ( 5) ,11 () ,12 (6) ,13 ( 12) ,14 (6) 1 1 0,15 ( 3) TOTAL ( 50) In plants generally, the cuticle thickens with age (Martin & Juniper, 1970) and this trend was quantified specifically fr O. aurantiaca and O. maxima. The mean cuticular thickness f different length claddes was measured and the results are pltted in fig. 17. Cladde length was used as a measure f age because the lengths f Opuntia spp. claddes grwn under the same cnditins are prprtinal t their age until they are fully grwn. The regressins shw that the thickness f the cuticle f beth species increase significantly with age (P =<:0,05). Hwever, the cuticle in mature O. maxima (i.e. with claddes greater than 25cm lng) achieves a greater thickness than that in mature O. aurantiaca (i.e. with claddes greater than cm). Althugh O. auranti aca

58 200 - Y = 1403,9x + -72,7 r = 0,8741,,,,,, Q. MAXIMA 0 ~ g) E ~....J 0 I- ::I t.)... t.) a:: w ii: 0 I- 0 w 0 '" u. :z VA It; ::I III III w 0:: n y ~ 792.7x ,7 0 2 ()Ol-- r:: O. AURANTIACA O. ma xima,,,,,,,, 0- l ~I~~ ~----~--~. 0,05 0,1 0,15 0,2 CUTIC!.E THICKNESS (mm) Fig, 18" Relative cuticular hardness pltted against cuticular thickness fr O. maxima and. aurantiaca.

59 35 claddes grw up t mre than 30cm under certain cnditins, their cuticle des nt thicken significantly (P =:>0,80) abve that f cm claddes. The penetrability, r tughness, f bth O. aurantiaca and O. maxima cuticle was als measured and fig. 18 shws that the penetrability f cuticle is prprtinal t its thickness. AS a result, O. maxima develps nt nly a thicker cuticle than O. aurantiaca, but as it thickens it becmes t ugher s that in mature claddes it serves as an effective barrier t z. tapiacla. Cnsequently, the percentage larvae that enter O. aurantiaca and O. maxima when placed n flush grwth is much higher than when they are placed n mature grwth, especially n O. maxima (Table 7). Table 7. The percentage penetratin f cuticle by first instar Z. tapiacla larvae n flush.and mature. aurantiaca and O. maxima. NUMBER OF NUMBER PLANT STAGE LARVAE ON PENETRATING PLANTS CUTICLE PERCENTAGE PENETRATION O. aurantiaca FLUSH ,8 O. maxima FLUSH ,8 O. aurantiaca MATURE ,0 O. maxima MATURE ,3

60 A B 5cm C D Fig. 19. O. maxima flush claddes damaged during a ttempted clnisatins by Z. tapiacla first instar larvae. (A) Healthy flush cladde' bearing small fleshy leaves (L). (B) Flush with tw attempted clnisatins resulting in lcal ised desiccatin but little damage. (C) Destryed grwing pint and limited gum prductin. (D) Whle flush cl adde desiccating 24h pri r t drpping frm the plant.