Biological control of jointed cactus, Opuntia aurantiaca (Cactaceae), in South Africa

Agriculture. Ecos),stems and Environment, 37 ( ! 991 ) 5-27 Elsevier Science Publishers B.V., Amsterdam

Biological control of jointed cactus, Opuntia aurantiaca (Cactaceae), in South Africa V.C. Moran ~ and H.G. Zimmermann b ~University of Cape 7bwn. Rondebosch 7700. Soz.,h Africa bPlant Protection Research Institute. Private Bag X134. Pretoria 0001. South ,4frica (Accepted 25 March 1991 )

ABSTRACT Moran, V.C. and Zimmermann, H.G., 1991. Biological control of jointed cactus, Opuntia aztrantiaca (Cactaceae), in South Africa. Agric. Ecosystems Environ., 37: 5-27. Jointed cactus, Opunzia aurantiaca Lindley, has been a problem weed in South Africa for nearly a century. Its taxonomic status, history of introduction, and chemical and biological control were reviewed in 1979. The present account updates this information and deals with recent South African research on the ecology ofO. aurantiaca and its biological control agents, and on the management of the weed. This review places some emphasis on the primary agent, the cochineal insect DacO,Iopius austrinus De Lotto (Homoptera: Dactylopiidae), with a shorter commentary on the phycitid moth Cactoblastis cactorum (Bergroth) that has also become established on jointed cactus. Research on three other introduced moths, Tucumania tapiacola Dyar (Phycitidae), Mim,,rista pulchellalis Dyar (Pyraustidae) and Nanaia sp. Heinrich (Phycitidae), is abstracted. Of the latter three species, only M. pulchellalis has become established, but its role in biological control is insignificant. It is concluded that: (i) the utility of insects as biological agents against jointed cactus has probably been fully exploited: (ii) the importance and threat of O. aurantiaca as a weed in South Africa may have been overestimated: (iii) future research emphases should be on the role of pathogens as biocontrol agents and on the formulation ofarea-specific integrated management procedures involving D. austrinus and herbicides.

INTRODUCTIO.N

Jointed cactus, Opuntia aurantiaca Lindley, is a low-growing, many-jointed plant (Fig. 1 ), native to South America, that has become a weed in Australia and South Africa. The numerous long barbed spines on easily detached joints (cladodes) are injurious to livestock and high densities of the weed reduce the productivity of pastoral land. Despite sustained control efforts in South Africa over the past century, jointed cactus is still perceived by many as a major weed problem and there is no evidence that its distributional range in South Africa is decreasing. The taxonomic status, history of introduction and control of jointed cactus 0167-8809/91/$03.50 © i 99 ! Elsevier Science Publishers B.V. All rights reserved.

V.C. M O R A N A N D H G. Z I M M E R M A N N

!
J J

i_

/ ,# i 1

/ /

?

'

"

7' "'"

/

i

\ ,.,.

\

Fig. I. Jointed cactus, O, tlllralltitlcti Lindley. (Drawn by R. Weber, National Botanical Institute, Pretoria, )

in South Africa were reviewed by Moran and Annecke (1979), who also described the biology of the insect agents involved up until that time. It is necessary to review this subject again only a decade later because about the same amount of research (number of pages published) has been produced on jointed cactus in South Africa over the past l0 years as was produced during the 90 years preceding 1979, and the research emphasis has changed considerably. From the early 1890s up to 1979, more than 40% of the publications dealt with the control, 'eradication' or destruction of jointed cactus, with un-

BIOLOGICAL CONTROL OF OP/_'NT/A .-I UR.4NIZ.I('4 IN SOUTH AFRICA

7

der 25% reporting on the biology or ecology of the plant and its associated insects, and with the remaining 35% devoted to taxonomy or to miscellaneous related subjects. The research during those earlier years was mostly fuelled by a misguided euphoria that herbicides would provide a quick and easy solution to the jointed cactus problem in South Africa, and much of the research on management was based on a deficient knowledge of the biology and ecology of the organisms involved. In contrast, subsequent to 1979, about 70% of the publications have dealt with the ecology of the weed and its biological control agents, or with their interactions. Recent publications d~aling with control of the weed (less than 10% of the total) have been on the subject of management or on integrated control. The present review thus deals with research developments in the biological control and management of jointed cactus over the past l0 years, with emphasis on the biology and ecology of the most important agent, the cochineal insect Dactylopius austrinus De Lotto, and on three lepidopteran species that have been recently introduced. O R I G I N AND T A X O N O M I C STATUS

Opuntia auramiaca is a variable taxon (Arnold, 1977; Moran and Annecke, 1979) and there is consensus that the plants are hybrids ofplatyopuntia cacti that originated in temperate South America, particularly Eastern Argentina and Southern Uruguay (Lindley, 1833; Moran et al., 1976; Arnold, 1977; Moran and Annecke, 1979; Moran and Zimmermann, 1984). A detailed ordination study of 88 morphological and biological attributes of the weed supported the supposition of a hybrid origin for O. aurantiaca (van de Venter et al., 1984) and showed that, of four taxa studied, Opuntia discolor Britton and Rose and Opuntia sahniana Parm. were the most likely putative parents, with O. aurantiaca having most affinity with O. discolor. However, it has been observed (H.G. Zimmermann, 1974) that the latter species is sterile and, therefore, possibly also a hybrid. Thus, one of the basic issues in biological control of O. attmntiaca, namely the origin of the weed, remains unresolved. T H E SPREAD. DENSITY A N D C O N T R O L OF O. A URANI'IAC,t IN S O U T H AFRICA

Moran and Annecke (1979) presented a detailed history of relevant associations between South African and European botanists, and !~rovided circumstantial evidence to suggest that, as a result of these contacts, O. aurantiaca was shipped to South Africa from established glasshouse cactus collections in the UK, probably in the 1840s. Figure 2 shows the area infested by O. aurantiaca since it was first recog-

8

V.C. MORAN AND H.G. ZIMMERMANN B 800-

0_ .= 6004004

,c

200'

i

IO00

1930

1960

1990

Year

Fig. 2. The estimated area in South Africa infested with O. aurantiaca over the last 100 years. Source: I, MacDonald (1892); 2, Pettey (1948); 3, Schonland (I .°,24); 4, Phillips (1938); 5, Du Toit ( 193 ~ ); 6, Pettey ( 1948 ); 7, Moran and Annecke ( 1979 ), after G.C. Burger (unpublished data); 8, assuming an annual increase of about 80 000 ha, Zimmermann ( 1981 ); 9, assuming an average annual increase of about 2000 ha per annum, from Moran and Annecke (1079).

nized as a weed in South Africa. The apparently rapid increase in area from the 1930s until 1964, when the first intensive surveys were made, reflects a real increase in spread of the weed, but is also partly an artefact of recent increases in searching effort and of more thorough documentation. However, the evidence suggests that an asymptote has now been reached at somewhere between 850 000 and 870 000 ha of infested land, concentrated mainly in the Eastern Cape and Karoo. The spread of the weed in South Africa has been affected by a succession of control practices over the past 100 years that are detailed in Moran and Annecke ( ! 979) and summarized by Zimmermann and Moran (1982). Prior to 1892, and until about 1938, control was limited to mechanical clearing and/or the application of arsenical herbicides. In spite of these practices, the weed spread rapidly and, more importantly, increased in density, so that by 1938 grazing by livestock on many properties was severely inhibited. Following the release, in June 1935, of the cochineal insect, D. austrinus, the density of the weed was spectacularly reduced over large areas and until 1946 biological control was the officially recommended practice. Local resurgences of the weed led to the reintroduction of mechanical control during the decade from 1947. These procedures were only partly effective because ofregrowth, reinfestation and spread from residual populations. Since 1957, landowners have been compelled to participate in a costly chemical control programme (see Moran and Annecke, 1979; Zimmermann et al., 1982; Zimmermann and Moran, 1982, for details). Herbicidal control, together with biological control, has now resulted in an overall reduction in the density of

BIOLOGICAL CONTROL OF OPU1VTIA A UR,4NTZ4C-I IN SOUTH AFRICA

9

the weed to levels where, in most places, grazing is not significantly reduced (Zimmermann, 1981; Robertson, 1985a). Dense populations of the weed are presently confined to riverine bush and to higher rainfall regions where cochineal insects are relatively ineffective (Zimmermann et al., 1986), as is the case in Australia (Hosking and Deighton, 1981 ). The biological control of jointed cactus continues to be inextricably linked with current chemical control practices and an overview of the integrated management of O. aurantiaca is provided in the discussion. THE COCHINEAL, DACI'YLOPIUS.4 USTRINUS DE LOTTO

The cochineal insect, D. ausoTnus, is by far the most important biological control agent of O. aurantiaca in South Africa (Pettey, 1948; Zimmermann et al., 1974; Moran and Annecke, 1979; Zimmermann, 1981) and in Australia (Hosking and Deighton, 1979, 1980, 1981; Hosking et al., 1988). The species was described by De Lotto (1974) and is native to Central and Western Argentina (Mann, 1969 ). Dactylopius austrinus do,': not occur on O. aurantiaca in South America and its preferred native host plant may be Opuntia utkilio Spegazzini or Opuntia sulphurea G. Don in Loudon (Mann, 1969; Moran and Annecke, 1979). The biology of D. austrinus is detailed in Gunn (1978), Moran and Annecke (1979) and Moran and Cobby (1979) and, subsequently, Hosking (1984) has documented the effects of temperature on the development of the insect and on its population growth in Australia. The introduction and release ofD. austrinus in South Africa, and its effectiveness and shortcomings as a biological control agent of jointed cactus, are reviewed by Moran and Annecke (1979). The present review emphasizes recent research developments against the backdrop of the remarkably perceptive commentary by Pettey (1948) on the interactions between D. austrinus and its host plant. Cyclical resurgences of jointed cactus

The observations of Pettey ( 1948 ) are clear: the introduction of D. austrinu~ resulted in lower equilibrium densities of jointed cactus in almost all infested areas, but with cyclical resurgences above an acceptable threshold. After 1957, herbicides have been applied to jointed cactus throughout infested areas and it is impossible now, in retrospect, to quantify the reductions in equilibrium densities, of ~he weed, especially as Pettey's ( 1948 ) reports are descriptive. Nor is it possible to apportion these reductions in density to either chemical or biological zontrol. However, a clear indication of the efficacy of D. austrinus on jointed cactus comes from exclusion experiments (Zimmermann, 1981; Zimmermann and Moran, 1982) which are summarized in Table I.

10

V.C. MORANAND H.G. ZIMMERMANN

TABLE i Survival of and cladode production by O. aurantiaca plants over a 5 year period. Of 255 marked plants, 152 were colonized by D. aztstrmus and 103 were kept free of cochineal insects. The data are from Zimmermann and Moran (1982)

Plants surviving (%) Total no. of cladodes produced Mean no. ofcladodes produced per plant Total cladodes produced that rooted (%)

Without cochineal insects

Colonized by cochineal insects

100 17348 168 63

23 3666 24 6

Recently, Zimmermann and Malan ( 1989 ) reported a cyclical interaction, over 12 years, between jointed cactus and the biological control agent D. austrinus (Fig 3 ). An increase in the density of O. aurantiaca in an area facilitates a build-up in cochineal numbers that in turn causes a crash in the host plant population. Consequently, the cochineal is deprived of a host and it in turn experiences a population crash• The host plant population then recovers and the cycle is repeated. Zimmermann and Malan (1989) speculate that the amplitude of the host plant oscillations are mostly dependent on climatic conditions. These host plant resurgences, albeit at different times and at different levels at different places, and the apparent disappearance of the cochineal, are well known to many landowners and cause concern and scepticism about the effi+

60-

I00

+ +

I

.

~"'~ •J " =

"

+

"

~o ~

~+ i

I

60

~+

40 '~ 20-

~:.

2O ~

o

.

.

.

.

1976

.

.

.

~'~'~

1080

101i,t

,

,

";',"*

10118

Year Fig. 3. Average population fluctuations of O. aurantiaca and of the cochineal insect, D. austrim~s, from 1976 to 1988 on twenty 50 m-' p e r m a n e n t plots (after Zimmermann and Malan, 1989 ), On two occasions, indicated by arrows, a supplementary inocutum of D. austrinus was m a n u a l l y introduced.

BI()LOGICAL CONTROL OF 01'~ 'NIL! .-It 'R..IN1Z.I(~.I IN SOUTH AFRICA

II

ciency of the insects and of biological control, making herbicidal control an attractive alternative. Cochineal dispersal Pettey ( ! 948 ) observed that natural recolonization of O. aurantiaca by D. austrinus was unsatisfactory at low host plant densities, necessitating intermittent manual redistribution of the insects. Gunn ( 1977, 1978, 1979) confirmed that the dispersal of cochineal insects, via wind-borne female crawlers from jointed cactus, is limited and showed that natural colonization of O. ato'antiaca is poor at low plant densities (i.e. less than 1 plant m--'). These limitations of cochineal dispersal are exacerbated by chemical control practices against jointed cactus that sustain low densities of the plant (Zirnmermann and Moran, 1982). Moran et ai. ( 1982 ) have documented the morphological, physiological and behavioural adaptations of female crawlers of D. austrinus for dispersal from their host, and they quantified the limited extent of vertical and horizontal displacement of crawlers from O. aurantiaca. In a pasture heavily infested with jointed cactus, vertical displacement of the crawlers seldom exceeded 1 m and less than 20°/0 of the crawlers were blown more than 2.5 m from the source plant. Gunn (1979) developed a method of improving cochineal dispersal. He placed crawler-producing colonies of D. austrinus on host plant cladodes in a wire basket on top of a 5 m high pole. Under these circumstances, crawlers were recorded in high densities up to about 100 m from the source and, over the next 2 years, jointed cactus populations were greatly reduced in these experimental areas (Zimmermann, 1981; Zimmermann and Moran, 1982 ). This elaborate experimental method of inoculation may not be more effective or practical than simple manual redistribution of colonized cladodes, except perhaps in inaccessible situations such as steep slopes. Predation of cochineal insects Geyer ( 1947 ) and Pettey ( 1948 ) reported that two coccinellid beetles, Exochomusflaviventris Mader and Crypotolaemus montrouzieri Mulsant, were important predators of D. austrinus. Nowadays, these predators are rarely associated with D. austrinus in the field (Moran and Annecke, 1979). The only recent research development in this area is the work of Morrison (1984) who demonstrated, in a comparative study of three cochineal species, that the fully developed 'wooly-wax' covering on mature colonies of D. austrinus protects the cochineal insects from predation by E. flaviventris. Individual cochineal colonies, denuded of their protective covering, and crawlers of D. austrinus are susceptible to predation by these beetles. The presence of carminic acid in the bodies of these insects has anti-feedant properties (Eisner et al.,

12

V.C. MORAN AND H.G. ZIMMERMANN

1980; Morrison, 1984) and Moran (1980) has suggested that the function of carminic acid in cochineal insects is to prevent attack by parasitoids because these insects are never parasitized. The effects of predation on the population dynamics of cochineal insects have not been quantified.

Climatic effects on cochineal insects It has been known for a long time that cochineal insects are adversely affected by rainfall and seem to flourish under hot dry conditions (De Nobrega, 1848; Pettey, 1948). Recently, Moran et al. (1987) and Moran and Hoffmann (1987) have shown that rain devastates populations of the cochineal Dactylopius opuntiae (Cockerell) and thus reduces its effectiveness as a biological control agent of the prickly pcar, Opuntia ficus-indica (L.) Miller, in wet areas of South Africa. These observations are also applicable to D. au~trinus, in the hotter, drier inland areas (as is the case in Australia, Hosking and Deighton, 1981; Hosking, 1984; Hosking et al., 1988 ), desiccation is the key mortality factor acting independently of density on isolated cladodes and on small O. aurantiaca plants, and this mortality is exacerbated by the rapid growth of D, attstrinus populations. Conversely, at moist coaslal sites the growth of O. aurantiaca is more vigorous, desiccation-induced mortality of the plants is not a significant factor, and proportionately fewer plants are killed by D. austrinus because cochineal population increases are less rapid and because the insects are killed during periods of high rainfall. Consequently, O. aurantiaca infestations increase in density during wet years, which favour the plant but not the cochineal insects, and decrease during dry years. The abundance of O. aurantiaca in South Africa is, therefore, markedly cyclical (Zimmermann et al., 1986). THE PHYCITID MOTH ('.I('70BL..ISTIS C.I('7"ORUM ( BERGROTH )

This well.known phycitid moth, whose native hosts in Argentina are Opuntia delaetiana Weber and another "Monocantha relative" (Dodd, 1940), was the primary agent responsible for the collapse of Opuntia stricta (Haworth) Haworth in Australia (Dodd, 1940; Mann, 1970). It was eventually released in South Africa in 1933 for biological control of O. ficus-indica, where the moth expanded its host range to include O. aurantiaca, and it is now of secondary importance as a biological control agent of the latter weed. The life history of the insect is documented by Dodd (1940), Pettey ( 1948 ) and Mann (1969), and its role as a biological control agent in South Africa is summarized by Annecke and Moran (1978) and by Moran and Annecke (1979) on O. ficus-indica and on O. aurantiaca, respectively. in 1981, H.G. Robertson started a comparative ecological study of C. cactorum in relation to its effectiveness against prickly pear and jointed cactus

BIOLOGICAL CONTROL OF OPUN1Z.1.4 UR.4NTL.I(~-I IN SOUTH AFRICA

13

in South Africa (Robertson, 1985a,b, 1987, 1988, 1989; Robertson and Hoffmann, ! 989). However, these findings are most appropriately reviewed in the context of the biological control of O. ficus-indica (see Zimmermann and Moran, 1991 ). Inter alia, he came to the conclusion that inundative releases of C. cactorum are too costly and would ultimately be ineffective as an augmentation for biological control of jointed cactus. Moran and Annecke (1979) were impressed by reports ofbiotypes or races of C. cactorum that were allegedly far more destructive on O. aurantiaca in South America than in Australia or in South Africa. They record that H.G. Zimmermann had released, in South Africa in 1977 and 1978, a stock of C. cactorum collected from O. aurantiaca in Argentina, but that the fate of this release is unknown. This was followed, in 1978, by an importation of C. cactorum from O. aurantiaca-like cactus plants in South America. The culture was ultimately destroyed when it proved impossible to distinguish, with any degree of reliability, these biotypes from the C. cactorum populations already in the country. These difficulties were experienced in spite of access to an unpublished key, prepared by McFadyen (now published, 1985), that separated C. cactorum biotypes on the basis of larval markings. Robertson (1985a) is sceptical about the utility of alternative biotypes of C. cactorum in South Africa for biological control of jointed cactus because of the overriding detrimental effects of ants, in particular, on C. cactorum populations in South Africa, and because there is no evidence that the supposed biotypes in South America would be more effective than the biotypes already established in South Africa. However, McFadyen (1985) makes the valid point that a closer study of biotypes on different host plants may show them to be separate species, and Moran and Zimmermann (1984) noted that "an increased awareness and improved identification of plant and insect ecotypes or strains might improve the chances for successful biological control of Cactaceae". These comments suggest that it may be unwise to let the matter of possible biotypes of C. cacforum for use against jointed cactus remain inadequately explored and unresolved, as is the case at present. THE PHYCITID MOTH 7"UCUMANL,! 7;,IPIA('OLA DYAR

Dyar (1925) established the genus Tucumania to accommodate two species, Tucumania tapiacola and Tucumania porrecta, that had been recorded from unidentified Opuntia hosts. On the basis of careful morphological comparisons of the adults, eggs and larvae, and as a result of cross-breeding experiments, Hoffmann (1986) concluded that T. porrecta is a junior synonym of T. tapiacola. The species has a wide distribution in Central Bolivia, and in Central and Northern Argentina, and it has also been collected in Eastern Paraguay and Uruguay. Tucumania tapiacola is oligophagous and has been

14

V.C. MORANAND H.G. ZIMMERMANN

recorded from five species of Eriocereus, from Monvilea sp. and from at least 15 species of Opuntia, including O. aurantiaca (Dyar, 1925; Heinrich, 1939, 1956; Dodd, 1940; Mann, 1969; Hoffmann and Moran, 1977; Hoffmann, 1982, 1986). This insect was introduced into Australia in 1935 where it established on 0. aurantiaca and now causes minor damage to its host (Hosking et al., 1988 ). Thirty-eight years later, T. tapiacola was introduced into South Africa from South America (Hoffmann, 1976, 1977, 1982; Hoffmann and Moran, 1977; Moran and Annecke, 1979). These authors, together with Heinrich ( 1939, 1956), Dodd (1940), and Mann (1969), provide accounts on the biology of the species. Two stocks of T. tapiacola were eventually released in South Africa and Table 2 presents the salient details. Two other incipient cultures were introduced from Cochabamba and Parotani, Bolivia, in 1974 and 1978, respectively. The first died out in the fourth insectary generation when their host plants were destroyed by C. cactorum that had adulterated the colonies, and the second was purposely destroyed in 1981 (Hoffmann, 1982). T. tapiacola has not established in South Africa. TABLE 2

The history of introduction and release, in South Africa, of two stocks of 1~ tapiacola from Argentina (after Hoffmann, 1982 ) Origin

lbarretta, Formosa Province

Campana, Buenos Aires Province

Hosts

Opunlia discoh : Oplttltia relrorsa

Opuntia aurantiaca

Introduced as

Eggs m 1973

Larvae in 1981

Releases

Mostly around Grahamstown in the Eastern Cape. Small-scale trial releases I 0 May 1976-May 1977 Large-scale releases from 27 May 1977 until 13 February 1982

Small-scale releases around Grahamstown of pure strain Campana stock and of Cam pana × I baretta crosses during the summers of 1982/ 83 and 1983/84

Total released

22500 eggs 784000 larvae 20250 larvae in cladodes 3076 adults

Some thousands of eggs and larvae, but less than 10000 in total

Establishment

Usually survived tbr two generations (once fer five generations) in the field, but not established

Not established

BIOLOGICAL C O N T R O L O F OPLWT/,.! A ~ ;RANTZ.I( ~-! IN S O U T H AFRICA

15

Factors preventing establishment of T. tapiacola in South Africa Hoffmann ( 1981, 1982 ) has accounted, in detail, for the mortalities suffered by the various stages of T. tapiacola in the field and has provided a series of partial life tables for the immature stages of 7". tapiacola on, and in, host plants of different sizes and under different conditions in the shade or exposed to full sunlight. The major mortality was egg loss from ant predation and, overall, the cumulative percentage mortality for the immature stages of T. tapiacola in the field always exceeded about 97%. In most situations, it was in excess of 99%. Hoffmann ( 1982 ) records that T. tapiacola is a fecund species ( 309 eggs per female) and that even these high mortalities in the field would, in theory, allow survival of the species and eventual population increases, provided that cumulative mortalities are not coupled with a high adult mortality (in excess of about 70%). In the absence of data on adult mortality, it has, therefore, not been proven that the high mortalities experienced by the immature stages were the direct cause of establishment failure by T. tapiacola in South Africa. The high degree of mortality of the immature stages of T. tapiacola (Ibarretta strain) once they had entered the host plant suggests incompatibility with the new host, O. aurantiaca, which is manifested by an inability to overcome the physical and chemical defenses of the new host, or to derive sufficient nutrients from this host and thus to avoid starvation in large numbers. Also, Hoffmann ( 1982 ) points out that in spite ofthe large numbers of larvae released in the fielra (averaging about 19 000 per release), problems of synchrony in the emergence of adults and possibly their subsequent dispersal from the release sites, may have resulted in the numbers of adults, locally, being below a critical threshold for successful mating. The lbarretta strains of T. tapiacola (Table 2 ) from O. discolor and Opuntia retrorsa Spegazzini may have been 'genetically maladapted' to O. aurantiaca, and Hoffmann ( 1982 ) suggested a strategy of cross-breeding different T. tapiacola strains from South America to improve the chances of establishment. He followed this plan of action by releasing cross-bred Ibarretta/Campana populations in the field, but these releases were no more successful than all the others. Even if there is a slim possibility that, with renewed effort, T. tapiacola may become established in the field in South Africa, judging from the Australian p,c, edent (Hosking et al., 1988 ) and what we now know about the biology of the species, it is extremely unlikely that the establishment of T. tapiacola would have any significant effect on the biological control of jointed cactus. This programme ires been terminated and should not be restarted.

16

v.c. MORAN AND H.G, ZIMMERMANN

T H E PYRAUSTID MOTH ML~IORIST..I PULCItELL,4LIS (DYAR)

There is some unresolved debate among taxonomists about the affinities and generic placement of Mimorista pulchellalis (Mann, 1969; Nieman, 1991a). The biology of this species is treated in some detail by Nieman ( 1983, 1991 a) and Fig. 4 summarizes the history of introduction, release and establishment of the species in South Africa. Following the gloomy p~ognosis for the success of T. tapiacola as a biological control agent against jointed cactus, which was based on an early perception that egg predation by ants would prevent establishment, hopes were high for M. pttlchellalis which lays single saucer-shaped eggs on the host that, in theory at least, should be more resistant to ant predation (Moran, 1980). The ease with which the moth was reared, and its destructiveness to O. aurantiaca 128 larvae ex O. aurantiaca Parana, Argentina (31°47'S

60"29'W), 12 June 1978.

~ ] t o

II 18 adults emerged and used start laboratory culture (33°2YS 126°32'E), on 29 June 1978. lin Grahamstown

J

200 larvae used to start I a st~polementary labora~oryl Iculture at Uitenhage -I

lniti~tl releases in October and November 1979, mass releases 24. March 1980 until

5 December 1982.

1

~'13 774 larvae I I rdeast.'d at 3 sites I I around Grahamstown mainlyI l at Kudu RescrveO3°07'S

I

12~'~'I'~"

I

~Established until January 1987, in low l~umbcrs ill Kudu Reserve only but not

i

onitored more rec~-ntly.

I i ]

l~33°46'~ 's°2"~'r')'

I

249 531 larvae 'released around Uitenhage I'rom 4 February 1980 until 3 February 1981.

I

1

knot established.

I

I Laboratory cultures for

inbreeding experiments in Grahamstown (Wright. 1985).

Fig. 4. Tile history of the introduction, release and establishment of M. puh'hellalis in South Africa. Modified and updated from Wright (1985) and Nieman (1991b).

BIOLO(ilf'ALCONTROLOF OPUN1Z.I ,,IUR.INI'IA(:.I IN S()UTH AFRICA

I7

in the laboratory, raised hopes further for the success of M. pulchellalis in However, these optimistic expectations have not been realized.

q, ,~,~uth Africa.

Laboratory rearing of M. pulchellalis Wright ( ! 985 ) reported the effects of inbreeding and of laboratory rearing on M. puichellalis. She noted that "Considerable problems have been encountered in the laboratory-rearing of other species of cactophagous Lepidoptera for the biological control of jointed cactus. These species were; Amalafrida leithella Dyar, Cactoblastis mundelli Heinrich, Nanaia sp. Heinrich, Sigelgaita sp. Heinrich, and Sigelgaita transilis Heinrich. Initially, the laboratory rearing went well, but in the second filial generation (F2) each species suffered a decrease in numbers and ... entire cultures were lost ...". Table 3 records the details. Wright ( 1985 ) showed that it was mainly an inbreeding depression in egg viability that affected the population fitness of the inbred lines ofM. pulchellalis, such that no viable eggs were produced in the second laboratory generTABLE3

Details of the origin, importation and emergence of five species of cactophagous Lcpidoptera shipped to South Africa for screening as biological control agents against O. aurantiaca. All the cultures were lost (i.e. died out) after one, two or three generations in the laboratory, except Ior one of the Nanaia sp. cultures which suffered a population decline during the second generation. The table is taken from Wright ( 1985 ) Species

Collection locality

Date

No. imported Larvae

Pupae

No. of adults emerged

Generation Decline

Lost

.hnalafrida h,ithella (Ex Opuntia spp. Nr. pesttJi'r )

Pescadero, Colombia Caracas, Venezuela

5.10.82 5.10.82

74 52

0 0

No record No record

F2 F2

('actot~lastis mundelli

Cochamamba, Bolivia Parana, Argent ina

26.6.78 12.6.78

31 12

0 0

31 No record

F2 F2

Nanaia sp. ( Ex O. pascoensis )

Limatambo. Peru Li matambo. Peru Limatambo, Peru Limatambo, Peru Limatambo, Peru

6,7.78 23.9.81 11.7.82 19. ! .84 24.4.84

164 81 145 40 16

0 30 0 0 0

No record 57 64 32 16

Stgt'lgaita sp. (Ex Opuntia sp. )

Chilete, Peru

1.5,84

13

0

Sigelgaita transilis ( Ex Opuntiasp. Nr. pascoensis)

Huancabamba, Peru Huancabamba, Peru Chicamocha, Columbia

14.4.81 1.5.84 1.5.81

56 21 74

0 14 14

( Ex. O.

cochahamhensis )

3

7 No record No record

F2

F2 F3 F2 F2 F2 F2

F2 FI FI

i8

V.C. MORAN AND H.G. ZIMMERMANN

ation. This, also, may have been the factor responsible for the loss of the other lepidopteran cultures in the laboratory. Wright (1985) suggests that the size of the founder population may not have been critical and she notes that the extent to which inbreeding affects the fitness of each of the lepidopteran species is probably highly variable. She recommended that "'In order to improve the success of laboratory-rearing, the chances of inbreeding must be reduced, genetic variability preserved and selection against individuals under artificial conditions minimized", and she suggested a number of practical steps to enhance the chances of successful laboratory rearing for cactophagous Lepidoptera.

E f.lbct ofM. pulchellalis on jointed cactus Damage to, and mortality of, larger O. aurantiaca plants (greater than 300 mm in height) as a consequence of M. pulchellalis attack is insignificant (Nieman, 1991b). It is possible to calculate from Nieman's (1991b) data that of an estimated total of nearly 57 000 small and large plants in his study area, only about 0.02% of the plants would have succumbed as a result of attack from one generation ofM. pulchellalis larvae. His conclusions that "M. puichellalis may fulfill an important role in the current control campaign against jointed cactus" and "In general M. pulchellalis activity restricts the establishment of new jointed cactus plants which may ultimately result in an overall decrease in the density of cactus infestations", thus seem completely unjustified. Besides very low host plant mortality rates, the following factors further diminish any chances that M. pulchellalis will become an effective agent against jointed cactus. (i) Almost invariably, each plant is attacked by a single M. pulchellalis larva which destroys only one or at most a few cladodes, and this explains why only the smallest jointed cactus plants are killed. (ii) The number of individual cladodes damaged or destroyed by established populations of 3/1. pulcheilalis in the field is insignificant, relative to the vast reservoir of cladodes that remain, each with the potential to root and to propagate vegetatively. (iii) The expectation that the flat, saucer-shaped eggs of M. pulcheilalis would reduce the effects ofant predation in the field has not been borne out. High egg predation rates have been recorded, exacerbated by mortality of the larvae and pupae through attacks from indigenous parasitoids ( Hoffmann and Zimmermann, 1990; Nieman, 1991 b). (iv) In spite of large releases ofM. pulchellalis in South Africa (Fig. 4 ), the established population near Grahamstown has apparently dwindled over the five or so years since releases stopped and near Uitenhage, where about 78% of all the releases were made, the species has not established. A realistic expectation may be that M. pulchellalis will die out completely in South Africa. The evidence strongly suggests that Nieman's (1991b) rec-

19

BIOLOGICAL CONTROL OF OP[ "NI'IA .J UK.INTI..I('.I IN SOUTH AFRICA

ommendations that "Ideally an inundative release strategy should be tbllowed over an extended period of time, releases being targeted to areas in which etiolated cactus predominates ...", should not be followed. Overall, the performance of the species in South Africa suggests that there should be no further efforts expended on M. pulchellalis as a biocontrol agent of jointed cactus. T H E P H Y C I T I D M O T H NAN,4L4 SP. H E I N R I C H

Since the renewal of interest in the biological control of jointed cactus in South Africa in the 1970s, but particularly from about 1978 until 1984, there has been an emphasis on exploration for agents from Opuntia host plants that are related, at least in their general morphology, to O. aurantiaca. This, and the fact that all the insect agents from O. aurantiaca had already been tried, explains the importation of several species of Lepidoptera, including Nanaia sp., from various Opuntia hosts (see Table 3 ). Hoffmann (1988) realized, through his experience with the costly programmes on T. tapiacola and on M. pulchellalis, that performance testing of an agent prior to mass-rearing, release and subsequent field evaluation could TABLE 4

Performance of a founder population and two laboratory colonies of Nanaia sp., one reared on the natural host O. pascoensis and the other on O. aurantiaca, as measured by the listed componenls of fitness, over three generations. All the horizontal comparisons of fitness are significantly different, one from the other, except when indicated as non.significant (NS). Derived from Hoffmann ( 1988 ) Founder colony

Opunlia past'oensis

Opunlia al,ramiaca

Survival (%)

NeonataP Larval Pupal First instar to adult

-

Larval plus pupal duration (days) Males Females

Mating (%) Mean no. of eggs per female Egg viability (%) Wing span (mm ) Males Females

92 89 91 74

-

60 59

93 175 93

53 63 83

31 33

30 30

(NS) (NS)

88 60 89 47

66 66 iNS) (NS)

54 62 75 27 28

~From egg hatch, and including penetration into the host plant, until the commencement of feeding.

20

V.C. MORAN AND H.G. ZIMMERMANN

prevent a considerable waste of time, money and effort. He therefore compared the performance of a founder population of Nanaia sp. collected on Opuntia pascoensis Britton and Rose with that of two laboratory populations of the moth, one reared on its native host O. pascoensis and the other on the target weed O. aurantiaca. His data (Table 4 ) unequivocally show that there was a significant loss of fitness in Nanaia sp. populations reared in the laboratory, particularly on O. aurantiaca which is not a natural host for this moth. Furthermore, field trials demonstrated that the heavy mortalities suffered by the eggs and early instars of T. tapiacola through predation and rainfall, particularly, were also suffered by ',he eggs and small larvae of Nanaia sp. (Hoffmann and Zimmermann, 1990). These observations led Hoffmann ( ! 988 ) to recommend that mass-rearing and release ofNanaia sp. should not proceed, and this programme has now ceased. OTHER PHYTOPHAGOUS INSECTS ASSOCIATED WITH O. AUR,4NIZ4C4

Mann (1969) and Zimmermann et al. (1979) provide lists of herbivorous insect species associated with O. aurantiaca in South America. In addition to ~he species treated in the present review, Dactylopius ceylonicus Green, the diaspidid scale Diplacaspis echinocacti (Bouche) and the lonchaeid fly Das.. lops bourguini ( Blanchard ), are listed. The scale, Diplacaspis echinocacti, occurs on a wide range of cacti and was inadvertently introduced into South Africa and elsewhere, but plays no part in biological control (Annecke and Moran, 1978 ). Several attempts were made in South Africa, in the late 1970s and early 1980s, to breed Dasiops bourguini in the laboratory or in large outdoor cages with the provision of a wide range of possible foods for the adult flies and including several elaborate variations in rearing techniques, but all these attempts failed, probably because the flies would not mate properly. Recently, there has been renewed interest in the potential of a weevil, Pachycentrinus conv~:~:icoilisHustache, whose feeding damage was noted on lowgrowing opuntias in Argentina between 1970 and 1973 (Erb and Zimmermann, 1986). The adults and larvae cause only minor damage to the host. The species occurs on several jointed cactus-like hosts, including O. discolor, in the Chaco province of Argentina, but is not known from O. aurantiaca itself because the weevil and the plant are allopatric (Erb and Zimmermann, 1986). FUNGAL PATHOGENS

Ten years ago, research was initiated on pathogenic organisms for the control of O. attrantiaca (Moran and Annecke, 1979). lVd!denha!! et al. (1985, 1987 ) have iselated 144 fungi from six species of cholla cacti and two specie,z of prickly pears that occur in Arizona. Of these, Aureobasidium pullulans (de

BIOLOGICAL C O N T R O L OF OPt WIZ4.4 ['R,.INTI..I('..I I N S O U T H AFRICA

2 1

Bary) from Opuntia acanthocarpa Engelmann and Bigelow, Fusarium pro#feratum (Matshushima) from Opuntia arbuscula Engelmann, and two strains of Microdochium lunatum (Ellis and Everhard) Van Arx from Opuntia ieptocaulis De CandoUe, have proved to be pathogenic to O. aurantiaca. The rationale behind the exploration for pathogens in Arizona, where O. aurantiaca does not occur, is "Because O. aurantiaca has not co-evolved with these pathogens it might lack resistance to them" (Mildenhall et al., 1987), a philosophy which is supported by the arguments of Hokkanen ( 1985 ). Microdochium lunatum has subsequently been collected on O. ficus-indica near Uniondale (33°40 ' S 23°07'E) in South Africa and on O. aurantiaca in Argentina, and all four strains (those from Arizona, South Africa and Argentina) seem to be equally pathogenic on jointed cactus in South Africa (M.J. Morris, personal communication, 1989 ). During September 1982, H.G. Zimmermann and J.H. Hoffinann (personal communication, 1989) noted extensive fungal damage on O. aurantiaca in Colonia, Uruguay. Material was shipped back to South Africa, but came to nothing. In March 1987, S. Neser, M.J. Morris and H. Erb briefly visited Gualeguaychu in Argentina and discovered a pathogenic fungus on O. aurantiaca that exhibited symptoms similar to those produced by Phyllosticta concava (Seaver), but the fungus could not be cultured or reinoculated onto O. aurantiaca in South Africa (M.J. Morris, personal communication, 1989 ). DISCUSSION

The consequences of the recent research emphasis on the biology and ecology of 0. aurantiaca and its associated biocontrol agents have been: (i) the realization that there is virtually no hope of a significant improvement in biological control through further exploration for, and importation of, new and different insect agents; (ii) the important assessment that, after all these years, the status of jointed cactus as a weed may have been overrated and the perception that the plant, in the presence olD. austrinus, still has the potential for rapid spread and invasion of new areas in this country seems to be unfounded; (iii) the conceptt,,ed development of a rational integrated management programme for jointed cactus control in South Afri,~a.

Future prospects for biological control of O. aurantiaca Prospects for enhancing the biological control of jointed cactus using insects do not appear to be good. There are no further ~nsect agents to be found on O. aurantiaca itself and the attempts of South African entomologists to explore all possibilities for suitable agents against jointed cactus, have overst,,~tched the concept of a 'botanically closely related species', resulting in a succession of importe0 Lepidoptera that are simply incompatible with O. au-

22

V.C. MORAN AND H.G. ZIMMERMANN

rantiaca. For example, historical error resulted in a major effort devoted to the release and evaluation of the Ibarretta strain of T. tapiacola (Table 2) from O. discolor and O. retrorsa. In retrospect, it might have been wiser and

more productive to devote equal energies and a higher priority to releases and studies of the Campana strain of T. tapiacola that came from O. aurantiaca itself. Concentration on, and unrealistically high hopes for, insects as agents over the past l0 years, coupled with quarantine problems, may explain the poor progress in the utilization of plant pathogens as biological control agents. The exploration for pathogenic fungi on cacti in Arizona has been a step forward, but their release is a long way ahead and the prime target for exploration for pathogens, which must be m Argentina on O. aurantiaca itself, has only been given scant attention. The clear priority now is to initiate a determined programme to obtain jointed cactus pathogens, and perhaps their insect vectors, from South America for use against O. aurantiaca. The status of O. aurantiaca as a weed in South Africa

The assumption that O. aurantiaca, in the absence of persistent chemical control procedures, still has the potential to expand its range and density, has provided the motivation for the considerable amount of research reported in this review and for the large annual expenditures on herbicides. Robertson (1985a) found that, in an area classified by weed inspectors as 'heavily infested', only about 1.1% of the total ground area was actually covered by jointed cactus; 28% of the ground area was completely devoid of vegetation, mostly as a consequence of overgrazing and erosion. It also appears that present infestation levels do not significantly reduce the carrying capacity of the land and experienced livestock learn to avoid jointed cactus. This study, although restricted to only one site, calls into question the basis and economic validity of the criteria for assessing jointed cactus infestations. Robertson (1985a) argues that the potential for spread and dispersal of O. aurantiaca through the agency of livestock and wild animals has also been greatly exaggerated. It is clear that the status and potential danger of O. aurantiaca as a weed needs reassessment. Abandonment of the expensive and obligatory herbicidal control programme would probably not result in a significant change in the present weed status of O. aurantiaca in South Africa. In most areas, except notably along river courses, cochineal insects, augmented by reinoculations and, as necessary, by area-specific integrated control programmes, would probably be sufficient to ensure that the weed is maintained below a realistic and acceptable threshold.

BIOLOGICAL CONTROL OF OP~ "NIZ.! ..i UR..INTL.t(~4 IN SOUTH AFRICA

23

Integrated management

Details of the research needed for probable success in integrated control strategies for jointed cactus in South Africa can be found in Zimmermann ( 1977, 1979), Moran and Annecke (1979), Zimmermann and Malan (1980), Moran ( 1981 ), Zimmermann ( 1981 ), Zimmermann et al., (1982), Zimmermann and Moran (1982), Kluge et al. (1986), and in Zimmermann and Malan (1989). In essence, however, the distribution of the plant, which is highly aggregated, influences the efficacy of the herbicidal spraying programmes; aggregated plants and large plants are easily located, and are efficiently treated and controlled. Small plants are either overlooked or heavily overdosed. Through a careful analysis of herbicidal spraying procedures and the results achieved, Zimmermann and Malan (1980) suggested strategies which increase efficiency and decrease the costs of herbicidal control by up to 47%. Complicating the recommended integrated management procedures are the facts that: (a) the percentage of O. aurantiaca plants colonized by D. austrinu~ is directly correlated to host plant aggregations; (b) the large plants and aggregations of O. aurantiaca are most efficiently treated and killed during herbicidal applications; (c) it is these categories of plants that provide the main reservoirs for the cochineal insects themselves. Thus, herbicidal control has been directly antagonistic to biological control in most circumstances. This realization has led to recent recommendations: (i) that large plants and aggregations of jointed cactus should be treated with a water-based herbicide, monosodium methanearsonate (MSMA), that has no direct insecticidal effects on the cochineal insects; (ii) that small plants and loose cladodes should be collected, stacked and treated, en masse, with MSMA; (iii) that dense O. aurantiaca infestations should never be sprayed, and time allowed to elapse enabling D. austrinus to reduce the density of these infestations; (iv) lastly, that it is imperative that cochineal be continually present, locally, and reintroduced manually when necessary (Zimmermann and Malan, 1989). Perhaps of most importance in the integrated management programme is the realization that different management procedures are required in each of the many vegetation types and climatic areas infested with jointed cactus. For example, the recommended procedures should be quite different in higher rainfall areas where cochineal insects are ineffective as biological control agents, compared to those recommended in dry areas. The future lies in formulating these tailor-made strategies and in applying them as appropriate in prescribed areas and situations. REFERENCES Annecke, D.P. and Moran, V.C., 1978. Critical reviews of biological pest conlrol in South Africa. 2. The prickly pear, Opunliaficus-indica (L.) Miller. J. Entomol. Soc. South Aft., 41: 161-188.

24

V+C. MORAN AND H.G. ZIMMERMANN

Arnold, T.H., i 977. The origin and relationships ofOpuntia aurantiaca Lindley. Proceedings of the Second National Weeds Conference of South. Africa, 1977, at Stellenbosch, South Africa, pp. 269-286. De Lotto, G., 1974. On the status and identity of the cochineal insects ( Homoptera: Coccoidea: Dactylopiidae). J. Entomol. Soc. South Afr., 37: 167-193. De Nobrega, G.J., 1848. On the cultivaiion of cochineal. Pharm. J., 1848: 342-348. Dodd, A.P., 1940. The biological campaign against prickly pear. Commonwealth Prickly Pear Board Bulletin, Brisbane, Australia, 177 pp. Du Toit, E., 1935. The jointed-cactus eradication campaign. Farming S. Afr., 10: 344. Dyar, H.G., 1925. Notes on some American Phycitinae (Lepidoptera, Pyralidae). Insecutor Insectiae Menstruus, 13: 220-226. Eisner, T., Nowicki, S., Goetz, M. and Meinwold, J., 1980. Red cochineal dye (carminic acid): Its role in nature. Science, 208: 1039- ! 042. Erb, H.E. and Zimmermann, H.G., 1986. Pachycenoqnus com't:~'icollis Hustache (Coleoptera: Curculionidae), a possible biocontrol agent against jointed cactus. J. Entomol. Soc. South. Air., 49: 392-394. Geyer, J.W.C., 1947. A study of the biology and ecology of l'+)oc'homusflavipes Thunb. (Coccincllidae, Coleoptera ). Part 11. J. Entomol. Soc. South. Afr., 10: 64-109. Gunn, B.H., 1977. Artificial dispersal of the cochineal insect, Daco,Iopius austrinus De Lotto, in infestations of jointed cactus, Opuntia aurantiaca Lindley. Proceedings of the Second National Weeds Conference of South Africa, 1977, at Stellenbosch, South Africa, pp. 287-295. Gunn, B.H., 1978. Sexual dimorphism in the first instar of the cochineal insect Dactylopius austtqnus De Lotto (Homoptera: Dactylopiidae ). J. Entomol. Soc. South. Afr., 41: 333-338. Gunn, B.H., 1979. Dispersal of the cochineal insect Daco'lopius attstrinus De Lotto (Homoptera: Dactylopiidae ). PhD. Thesis, Rhodes University, Grahamstown, 188 pp. ( unpublished ). Heinrich, C., 1939. The cactus-feeding Phycitinae: a contribution towards a revisior of the pyralidoid moths of the family Phycitidae. Proc. U.S. Nat. Mus., 86: 331-413. Heinrich, C., 1956. American moths of the sub-family Phycitinae. U.S. Nat. Mus. Bull., 207: 1158. Hoffmann, J.H., 1976 Pre-release studies on Zophodia tapiacola (Dyar) (Pyralidae: Lepidoptera), a biological control agent against jointed cactus, Opuntia aurantiaca Lindley. MSc. Thesis, Rhodes University, Grahamstown, 57 pp. (unpublished). Hoffmann, J.H., 1977. The pyralid moth, Tm'umania tapiacola Dyar (Lepidoptera), imported into South Af'ica from Argentina tbr jointed cactus control. Proceedings of the Second National Weeds "?onference of South Africa, 1977, at Stellenbosch, South Africa, pp. 297-302. Hoffmann, J.H., 198 I. Release of "l'ucumania tapiacola ( Lepidoptera: Pyralidae ) in South Africa against Opmrtia aurantiaca: the value of detailed monitoring. Proceedings of the Fifth International Symposium on the Biological Control of Weeds, 1980, at Brisbane, Australia, pp. 367-373. Hoffmann, J.H., 1982. Evaluation of Tucumania tapiacola Dyar (Lepidoptera: Phycitidae) for biological control of jointed cactus in South Africa. PhD. Thesis, Rhodes University, Grahamstown, 153 pp. (unpublished). Hoffmann, J.H., 1986. Evidence that the two currently recognized nominal species of Tucumania Dyar (Lepidoptera: Phycitidae) are not distinct. J. Entomol. Soc. South. Aft., 49: 267-274. Hoffmann, J.H., 1988. Pre-release assessment of Nanaia sp. (Lepidoptera: Phycitidae) from Opuntia pascoensis for biological control of Opuntia aurantiaca ( Cactaceae ). Entomophaga, 33: 81-86. Hoffmann, J.H. and Moran, V.C., 1977. Pre-release studies on Tucumania tapiacola Dyar (Lepidoptera: Pyralidae), a potential biocontrol agent against jointed cactus. J. Entomol. Soc. South. Aft,, 40: 205-209.

BIOLOGICAL ('ONTRC, L OF OPI_WTI.4.4URANTIA(.:.I IN SOUTH AFRICA

25

Hoffmann, J.H. and Zimmermann, H.G., 1990. Ovipositional and feeding habits in cactophagous pyralids: predictions for biological control of cactus weeds in southern Africa. In: E.S. Delfosse (Editor), Proceedings of the Seventh International Symposium on the Biological Control of Weeds, 1988, at Rome, Italy, pp. 395-399. Hokkanen, H., 1985. Exploiter-victim relationships of major plant diseases: implications for biological weed control. Agric. Ecosystems Environ., 14: 63-76. Hosking, J.R., 1984. The effect of temperature on the population growth potential of DacO,Iopills austrinus De Lotto (Homoptera: Dactylopiidac), on Opuntia aurantiaca Lindley. J. Austr. Entomol. Soc., 23: 133-139, Hosking, J.R. and Deighton, P.J., 1979. The distribution and control of Opuntia ato'antiaca in New South Wales. Proceedings of the Seventh Asian-Pacific Weed Science Society Conference, 1979, at Sydney, Australia. pp. 195-200. Hosking, J.R. and Deighton, P.J., 1980. Biological control of moisture stressed Opuntia aurantiaca using Dact.vlol~iusaustrinus. In: E.S. Delfosse (Editor), Proceedings of the Fifth International Symposium on the Biological Control of Weeds, 1980, at Brisbane, Australia, pp. 483-487. Hosking, J.R. and Deighton, P.J., 198 I. Tiger pear is a continuing problem. Agric. Gaz. N.S.W., 92: 43-45. Hosking, J.R., McFadyen, R.E. and Murray, N.D., 1988. Distrib.ation and biological control of cactus species in eastern Australia. Plant Prof. Q., 3:115-123. Kluge, R.L., Zimmermann, H.G., Cilliers, C.J. and Harding, G.B., 1986. Integrated control for invasive alien weeds. In: I.A.W. Macdonald. F.J. Kruger and A.A. Ferrar (Editors), The Ecology and Management of Biological Invasions in Southern Africa. Oxford University Press, Cape Town, pp. 295-303. Lindley, J., 1833. Opuntia aurantiaca. Orange-coloured Indian Fig. Bot. Reg. 19 t 1606 ( 1833 ). MacDonald. A.C., 1892. Quoted in Fisher 1892. New Cactus (Prickly Pear). Agric. J. Cape of Good Hope, 5: 93-94. Mann, J., 1969. Cactus-feeding insects and mites. Bull. U.S. Nat. Mus., 256: 1-158. Mann, J., 1970. Cacti naturalized in Australia and their control. Deparlmenl of Lands, Brisbane, 115 pp. McFadyen, R.E., 1985. Larval characteristics ofCactoblastis spp. ( Lepidoptera: Pyralidae ) and the selection of species for biological control of prickly pears ( ()plortia spp. ). Bull. Entomol. Res., 75: 159-168. Mildenhall, J.P., Alcorn, S.M. and Marasas, W.F.O., 1985. An assessment of the pathogenicity to jointed cactus of fungi obtained from chollas in Arizona. Phytophylactica, 17: 54. Mildenhall, J.P., Alcorn, S.M. and Marasas, W.F.O., 1987. Pathogenicity of fungi isolated from Olmntia species in Arizona to Opuntia aurantiaca. Phytophylactica, 19: 485-489. Moran, V.C., 1980. Interactions between phytophagous insects and their Opuntia hosts. Ecol. Entomol., 5: 153-164. Moran, V.C., 1981. The biological control of Opuntia aurantiaca in South Africa: evaluation and emerging control strategies. In: E.S. Delfosse (Editor), Proceedings of the Fifth International Symposium on the Biological Control of Weeds, i 980, at Brisbane, Australia, pp. 383-387. Moran, V.C. and Annecke, D.P., 1979. Critical reviews of biological pest control in South Africa. 3. The jointed cactus, Opuntia attrantiaca Lindley. J. Entomol. Soc. South. Afr., 42: 299-329. Moran, V.C. and Cobby, B.S., 1979. On the life-history and fecundity of the cochineal insect, Dact.rlopius austrimls De Lotto (Homoptera: Dactylopiidae), a biological control agent for the cactus Opuntia aurantiaca. Bull. Entomol. Res., 69: 629-636. Moran, V.C. and Hoffmann, J.H., 1987. The effects of simulated and natural rainfall on cochi-

26

V.C. MORAN AND H.G. ZIMMERMANN

neal insects (Homoptera: Dactylopiidae): colony distribution and survival on cactus cladodes. Ecol. Entomol., 12:61-68. Moran, V.C. and Zlmmermann, H.G., 1984. The biological control of cactus weeds: achievements and prospects. Biocontrol News Inf. Commonw. Agric. Bur., 5: 279-320. Moran, V.C., Zimmermann, H.G. and Annecke, D.P., 1976. The identity and distribution of Opumia aurantiaca Lindley. Taxon, 25:281-287. Moran, V.C., Gunn, B.H. and Walter, G.H., 1982. Wind dispersal and settling of first-instar crawlers of the cochineal insect DacO,Iopius auso'mus (Homoptera: Coccoidea: Dactylopiidae ). Ecol. Entomol., 7: 409-419. Moran, V.C., Hoffmann, J.H. and Basson, N.C.J., 1987. The effects of simulated rainfall on cochineal insects (Homoptera: Dactyiopiidae): colony composition and survival on cactus cladodes. Ecol. Entomol., 12: 51-60. Morrison, J.F.. 1984. Protection from beetle-predation in cochineal insects (Dactylopiidac: Homoptera). MSc. Thesis, Rhodes University, Grahamstown, 87 pp. (unpublished). N ieman, E., 1983. An evaluation of Mimorista puk'hellalis (Dyar) ( Lepidoplera: Pyraustidae ) as a biocontrol agent against jointed cactus in South Africa. MSc. Thesis, Rhodes University, Grahamstown, 86 pp, (unpublished). Nieman, E., 199 l a. The introduction of Mimorista pulchellalis (Dyar) ( Lepidoptera: Pyraustidae ) into South Africa for the biological control of jointed cactus, Opuntia attrantiaca Lindle~. I. Biology and mass-rearing techniques. Entomophaga, 36: 69-76. Nieman, E., 1991b. The introduction of Mimorista pulcheilalis (Dyar) ( Lepidoptera: Pyraustidae) into South Africa for the biological control of jointed cactus, Opuntia aurantiaca Lindley. 2. Field evaluation. Entomophaga, 36: 77-86. Pettey, F.W., 1948. The biological control of prickly pears in South Africa. Scientific Bulletin of the Department of Agriculture and Forestry, Union of South Africa, 271: 1-163. Phillips, E.P., 1938. Jointed cactus and its eradication. Farming S. Aft., 13:216-217. Robertson, H.G., 1985a. "['he ecology of Cactoblastis cactontm (Berg) (Lepidoptera: Phycitidae ) in relation to its effectiveness as a biological control agent of prickly pear and jointed cactus in South Africa. PhD. Thesis, Rhodes University, Grahamstown. 181 pp., unpublished. Robertson, H,G., 1985b. Egg predation by ants as a partial explanation of the difference in performance of('actoblasti,, cactorum on cactus weeds in South Africa and Australia. In: E.S. Delfosse {Editor), Proceedings of the Sixth International Symposium on the Biological Control of Weeds, 1984, at Vancouver, Canada, pp. 83-88. Robertson, H.G., 1987. Oviposition site selection in ('actoblastis cactorum ( Lepidoptera ): constraints and compromises. Oecologia ( Berlin ), 73: 601-608. Robertson, H.G., 1988. Spatial and temporal patterns of predation by ants on eggs of Cactobhtstts cactorum. Ecol. Entomol., 13:207-214. Robertson, H.G., 1989. Seasonal temperature effects on fecundity of ('actoblastis cactorum (Berg) (Lepidoptera: Pyralidae): differences between South Africa and Australia. J. Entotool. Soc, South. Aft., 52:7 !-80. Robertson, H.G. and Hoffmann, J.H., 1989. Mortality and life-tables of Cactoblastis t'actorum (Berg) (Lepidoptera: Pyralidae ) compared on two host-plant species. Bull. Entomol. Res., 79: 7-17. Schonland, S., 1924. The jointed cactus. J. Dep. Agric. Union S. Afr., 9:2 ! 6-225. Van de Venter, H.A,, Hosten, L., Lubke, R.A. and Palmer, A.R., 1984. Morphology of Opuntia aurantiaca (jointed cactus) biotypes and its close relatives, O. discolor and O. sahniana (Cactaceae). S, Aft. J. Bot., 3: 331-339. Wright, M,D., 1985. The effects of inbreeding and laboratory-rearing on a pyraustid moth, Mimorista pulchellalis Dyar ( LeDidoptera: Pyraustidae ), imported for the biological control of jointed cactus in South Afiita. MSc. Thesis, Rhodes University, Grahamstown, 119 pp. ( unpublished ). Zimmermann, H.G., 1977. A sampling system for field infestations of jointed cactus Opuntia

BIOLOGICAL CONTROL OF OPUNTIA AURANTIAC.I IN SOUTH AFRICA

27

aurantiaca Lindley. Proceedings of the Second National Weeds Conference of South Africa, 1977, at Stellenbosch, South Africa, pp, 203-220. Zimmermann, H.G., 1979. Herbicidal control in relation to distribution of Opuntia aurantiaca Lindley and effects on cochineal populations. "Weed Res., 19: 89-93. Zimmcrmann, H.G., 1981. The ecology and control of Opuntia aurantiaca in South Africa in relation to the cochineal insect Dactylopius austrinus. PhD. Thesis, Rhodes University, Grahamstown, i 54 pp. (unpublished). Zimmermann, H.G. and Malan, D.E., 1980. A modified technique for the herbicidal control of jointed cactus, Opuntia aurantiaca Lindley, in South Africa. Agroplantae, ! 2: 65-67. Zimmermann, H.G. and Malan, D.E., 1989. Population biology of a jointed cactus infestation. Proceedings of the Conference of the Weeds Sciences Society of South Africa, 1989, at Port Edward, South Africa, I p. (unpublished). Zimmermann, H.G. and Moran, V.C., 1982. Ecology and management ofcactus weeds in South Africa. S. Air. J. Sci., 78: 314-320. Zimmermann, H.G. and Moran, V.C., 1991. Biological control of prickly pear, Opuntia ficusimlica (Caclaceae), in South Africa. Agric. Ecosystems Environ., this volume, 37: 29-35. Zimmermann, H.G., Burger, W.A. and Annecke, D.P., 1974. The biological control of jointed cactus in South Africa. Papers presented at the First National Weeds Conference, 1974, at Pretoria, South Africa, pp. 204-21 I. Zimmermann, H.G., Erb, H.E. and McFadyen, R.E, 1979. Annotated list of some cactus-feeding insects of South America. Acta Zool. Lilloana, 33:101-112. Zimmermann, H.G., Malan, DE. and Viljoen, B.D., 1982. Screening of some water based herbicides for jointed cactus (Opuntia aurantiaca Lindley ), control. Proceedings of the Fourth National Weeds Conference of South Africa, 1981, at Pretoria, South Africa, pp. 21-49. Zimmermann, H.G., Moran, V.C. and Hoffmann, J.H., 1986. Insect herbivores as determinants of the present distribution and abundance of invasive cacti in South Africa. In: I.A.W. MacDonald, F.J. Kruger and A.A. Ferrar (Editors), The Ecology and Management of Biological Invasions in Southern Africa. Oxford University Press, Cape Town, pp. 269-274.