Assessing the risks associated with the release of a ...

Biological Control 28 (2003) 302–312 www.elsevier.com/locate/ybcon

Assessing the risks associated with the release of a flowerbud weevil, Anthonomus santacruzi, against the invasive tree Solanum mauritianum in South Africa Terry Olckers* ARC—Plant Protection Research Institute, Private Bag X6006, Hilton 3245, South Africa Received 26 July 2002; accepted 25 March 2003

Abstract Biological control of Solanum mauritianum Scopoli, a major environmental weed in the high-rainfall regions of South Africa, is dependent on the establishment of agents that can reduce fruiting and limit seed dispersal. The flowerbud weevil, Anthonomus santacruzi Hustache, is a promising fruit-reducing agent, despite ambiguous results obtained during host-specificity evaluations in quarantine. Adult no-choice tests showed that although feeding is confined to Solanum species, normal feeding and survival occurred on the foliage (devoid of floral material) of cultivated eggplant (aubergine), potato, and several native South African Solanum species. During paired-choice tests, involving floral bouquets in 10-liter containers, A. santacruzi oviposited in the flower buds of 12 of the 17 test species, including potato and eggplant, although significantly more larvae were recovered on S. mauritianum than on eight other species. Larvae survived to adults on all 12 species, with survival significantly lower on only four species than on S. mauritianum. However, during multi-choice tests, involving potted plants in a large walk-in cage, A. santacruzi consistently displayed significant feeding and oviposition preferences for S. mauritianum over all of the 14 Solanum species tested. Analyses of the risk of attack on nontarget Solanum plants suggested that, with the possible exception of two native species, none is likely to be extensively utilized as a host in the field. Also, host records and field surveys in South America have suggested that A. santacruzi has a very narrow host range and that the ambiguous laboratory results are further examples of artificially expanded host ranges. These and other considerations suggest that A. santacruzi should be considered for release against S. mauritianum in South Africa, and an application for permission to release the weevil was submitted in 2003. Ó 2003 Elsevier Science (USA). All rights reserved. Keywords: Biological weed control; Bugweed; Curculionidae; Host specificity; Solanaceae; Risk assessment; Woolly nightshade

1. Introduction The South American tree, Solanum mauritianum Scopoli (Solanaceae) (bugweed, woolly nightshade) is a major environmental weed in the eastern, higher rainfall regions of South Africa (Henderson, 2001). The weed is also present in other neighbouring southern African countries, including Mozambique, Swaziland, and Zimbabwe, and also occurs further north in western, central, and eastern tropical Africa and in Madagascar, Mauritius, and the Atlantic islands (Jaeger and Hepper, 1986).

* Fax: +27-33-355-9423. E-mail address: [email protected].

Although S. mauritianum is recognised as a weed in many of these African countries, the worst infestations are in South Africa where the weed has been targeted for biological control since 1984 (Olckers, 1999; Olckers and Zimmermann, 1991). However, expanded host ranges displayed by candidate agents during quarantine tests, aggravated by the high economic importance of the genus Solanum, have resulted in the rejection of most of the agents tested. Only one agent, the leaf-sucking lace bug, Gargaphia decoris Drake (Hemiptera: Tingidae), has so far been released against the weed in South Africa (Olckers, 2000a). The nature of the problems caused by S. mauritianum in South Africa has necessitated the establishment of a complex of biological control agents to reduce the

1049-9644/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S1049-9644(03)00083-5

T. Olckers / Biological Control 28 (2003) 302–312

weedÕs high rates of vegetative growth and fruit set (Olckers, 1999). Although chemical and mechanical control methods are effective (Olckers, 1999; Olckers and Zimmermann, 1991), these are mostly negated by the high recruitment of fast-growing seedlings following clearing operations. S. mauritianum populations in South Africa support very low numbers of flower-feeding insects (Olckers and Hulley, 1989, 1991a,b) and thus produce extremely high densities of fruit. These are favored by frugivorous birds which are the main dispersers of the seeds. By contrast, fruiting in South America is considerably lower and appears to be regulated by several florivorous natural enemies. Agents which directly reduce fruiting are thus expected to play a crucial role in the integrated control of S. mauritianum in South Africa. The most important of the florivorous species appear to be flowerbud-feeding weevils of the genus Anthonomus Germar (Coleoptera: Curculionidae) which occur throughout the range of S. mauritianum in South America and contribute to the low fruit set. Two species, Anthonomus santacruzi Hustache and Anthonomus morticinus Clark, have been collected on S. mauritianum in South America, often at the same localities (Olckers, 1999; Olckers et al., 2002). Anthonomus species were imported into quarantine in South Africa on several occasions but, until 1995, were never successfully cultured because of excessive mortality during importations and inadequate culturing procedures. A. santacruzi was cultured and studied for the first time in October 1995, when a small founder colony of 18 adults was collected near the Iguazu Falls in northeastern Argentina. Although this culture was lost in 1997, the importation of a large founder colony in early 1998 permitted the establishment of new cultures in quarantine and the initiation of biology studies and host-specificity tests. In this paper, I discuss the results of these studies and assess the risks associated with the release of A. santacruzi in

303

South Africa. A. morticinus is currently the subject of a separate study in Brazil and will be discussed in a subsequent contribution.

2. Materials and methods 2.1. Field surveys and collections Collections of different species of Anthonomus were made during a 7-day survey of native and cultivated Solanum species in northeastern Argentina and southeastern Paraguay, during February 1998. At each locality where weevils were encountered on a particular Solanum species (Table 1), all other congeneric plants were examined to confirm the weevilsÕ presence or absence. Voucher specimens of plants and weevils are lodged in the collections of the South American Biological Control Laboratory (USDA–ARS, Buenos Aires, Argentina) and the National Collection of Insects, Plant Protection Research Institute (Pretoria, South Africa). 2.2. Laboratory cultures and biology studies Following the surveys, some 200 weevils collected on S. mauritianum were introduced into quarantine in South Africa. Identification of the voucher specimens in late 1998 revealed that both A. morticinus and A. santacruzi had been introduced, causing fears that the cultures were a mixture of both species. However, careful scrutiny of each individual weevil revealed that the cultures comprised solely A. santacruzi and that A. morticinus had either failed to breed or been displaced. Weevil colonies were maintained in 4-liter plastic containers fitted with gauze tops inside a quarantine laboratory at 22 3 °C, 70–80% relative humidity, and a 14-h photophase. Field-collected bouquets of

Table 1 Species of Anthonomus collected during a survey of Solanum species in Argentina and Paraguay during February 1998 Solanum species

Anthonomus species

Total number of sitesa

Number of sites with weevils

Estimated number of weevilsb

Other Solanum species at the same sitesc

S. fastigiatum Willd. (fa) S. mauritianum Scop. (ma)

A. santacruzi A. santacruzi A. morticinus

pa vi me si vi fa me si us vi

—

1 8 6 0 0 2 1 1

50+ 260+ (lar) 160+ (lar)

S. melongena L. (me) S. palinacanthum Dun. (pa) S. sisymbriifolium Lam. (si) Unidentified sp. (us) S. viarum Dun. (vi)

5 14 14 2d 2 8 1 12

a

—

A. sisymbrii A. tenebrosus A. sisymbrii

The localities where each species was sampled are provided in Olckers et al. (2002). Based on relative abundance at each site; (lar) indicates presence of larvae in the buds. c Only sites where the weevils were present were considered. d Cultivations were not treated with pesticides and supported their normal insect faunas. e Weevils also collected on this plant species; all other species did not have weevils present. b

— —

50+ (lar) 3 3 (lar)

vie fa ma me si vi sie

304


S. mauritianum, containing flowers, flowerbuds, and apical leaflets, were placed in glass vials filled with water and presented to the weevils for feeding and oviposition. Fresh bouquets were provided every 4–5 days and the old buds were placed in glass petri dishes to allow the immature stages to complete their development. The petri dishes were supplied with sections of leaves of the succulent plant Carpobrotus sp. (Mesembryanthemaceae), instead of moist filter paper, because of their ability to increase the humidity in the petri dishes and limit desiccation of the buds (S. Neser, Plant Protection Research Institute, South Africa, personal communication). Developing larvae were dissected from buds which had desiccated or rotted and were transferred to fresh buds. Emerging adults were added to the cultures or used in biology studies and host-specificity tests. During biology studies, daily observations on the weevilÕs feeding behaviour, preoviposition period, duration and frequency of oviposition, and duration of development of the immature stages were carried out.

2.3. Test plants With the exception of the no-choice tests (see below), host-specificity testing was largely focused on species in the genus Solanum. Although this genus is in need of taxonomic revision, because of considerable overlap in species descriptions, there are presumed to be some 34 native, 15 exotic, and two cultivated species in South Africa (Arnold and De Wet, 1993). Because it was not possible to obtain all of these species, the most common Solanum species occurring in South Africa were tested, particularly those that occur in the same regions invaded by S. mauritianum. Consequently, 17 species comprising both cultivated, 12 native (35% of the native flora), and three exotic species (20% of the exotic flora) were comprehensively tested. Although it was not feasible to import and test additional Solanum species that occur in other African countries known to harbor S. mauritianum, the majority of the test species, including cultivated, native, and exotic plants also occur in other African countries (Table 2).

Table 2 Physiological host range of A. santacruzi adults as determined by no-choice tests on the foliage (devoid of floral material) of cultivated, exotic, and native solanaceous plants Test plant

Distribution in Africaa

Feeding damage

A. Cultivated species C. annuum L. (green pepper) L. esculentum Mill. (tomato) P. peruviana L. (cape gooseberry) S. melongena L. (eggplant) S. tuberosum L. (potato)

Widespread Widespread Widespread Widespread Widespread

None None None Normal Normal

20.0* 6.7* 20.0* 93.3 76.7

B. Exotic species C. aurantiacum Lindl.d C. betacea (Cav.) Sendtn.d D. stramonium L.d S. mauritianum Scop.d S. pseudocapsicum L.d S. sisymbriifolium Lam.d S. viarum Dun.

Widespread? Widespread Widespread Widespread Widespread Widespread Regional?

None None None Extensive Normal Exploratory Suppressed

16.7* 3.3* 20.0* 82.5 93.3 70.0 30.0*

C. Native species S. aculeastrum Dun.d S. burchellii Dun. S. capense L.d S. chenopodiodes Lam. S. coccineum Jacq.d S. duplo-sinuatum Klotzsch S. giganteum Jacq.d S. incanum L.d S. linnaeanum Hepper and Jaegerd S. cf. linnaeanum Hepper and Jaeger S. nodiflorum Jacq.d S. panduriforme E. Mey.d S. tomentosum L.

Widespread Regional Regional Endemic? Regional Widespread Widespread Widespread Endemic Endemic? Widespread Widespread Regional

Suppressed Normal Suppressed None Normal Normal Normal Normal Suppressed Normal Suppressed Normal Extensive

96.7 63.3 100.0 0* 40.0* 83.3 90.0 83.3 80.0 96.7 86.7 90.0 90.0

a

% Survivalb; c

Where ÔendemicÕ ¼ present in South Africa only, ÔregionalÕ ¼ also present in other southern African countries and ÔwidespreadÕ ¼ also present outside of southern Africa (Jaeger and Hepper, 1986; Wright, 1904). b Percentage of weevils alive after 14 days. c Totals compared by v2 tests (2 2 Tables); test species followed by the asterisk were significantly different to S. mauritianum. d Listed as weeds in South Africa (Wells et al., 1986).


2.4. Adult no-choice tests These were used for the initial screening of A. santacruzi in order to determine whether the weevils can survive on the foliage of plants other than those in the genus Solanum. Test plants were confined to the Solanaceae and included cultivated, exotic, and native plants comprising 24 species from seven genera (Table 2). Because of the limited availability of floral material for testing and because the weevils can survive on foliage alone for several months (see below), the potted plants contained no floral material. The plants were covered by cylindrical plastic sleeves which were fitted into the pots and sealed with gauze tops. Ten newly emerged adults were confined on each test plant for 14 days and feeding and mortality was recorded. Adults were also confined on S. mauritianum as the control. Feeding damage was recorded as either Ônone,Õ Ôexploratory,Õ Ôsuppressed,Õ Ônormal,Õ or Ôextensive.Õ Fresh plants were provided after a week and each test species was tested on three separate occasions. 2.5. Paired-choice tests These tests involved bouquets containing flowers, flower buds, and apical leaflets and were confined to 17 test plant species that included cultivated, exotic, and native species in the genus Solanum (Table 3). Bouquets of S. mauritianum and one test species were placed in glass vials filled with water and presented, with the foliage mostly separate but sometimes touching, to the weevils in 10-liter plastic buckets fitted with gauze tops. The pairs of bouquets contained equivalent amounts of floral material. Twenty reproductively active weevils from the cultures were added to the buckets. After four days, the positions of the weevils and intensity of feeding on the floral material were recorded. As before, feeding intensity was scored in five different categories, where 0 denotes no feeding and 1–4 was used to gauge damage as ÔexploratoryÕ (1), ÔsuppressedÕ (2), ÔnormalÕ (3), and ÔextensiveÕ (4). At the end of each trial, the flowers and flowerbuds of each test plant species were transferred to petri dishes to allow the immature stages to complete their development. After 4–5 days, the floral material was dissected to record oviposition and the developing larvae were transferred to fresh flowerbuds of the species on which oviposition occurred. Fresh flowerbuds were provided when necessary for the duration of larval development. Daily observations were made to determine the number of larvae that survived to adults and the number of days to adult emergence, for comparisons of host suitability. Each paired trial was repeated on 3–5 separate occasions. 2.6. Multi-choice tests On several occasions during the paired-choice tests, the weevils fed and oviposited on nontarget Solanum

305

species (see below). To confirm these results, multi-choice tests were carried out in a large walk-in cage (250 cm 150 cm 200 cm) erected under a skylight inside the quarantine laboratory. Fourteen test species, which included those that were deemed to be most Ôsusceptible,Õ were presented to the weevils (Table 4). During each trial, 2–3 potted plants each of S. mauritianum and four test species were arranged in the cage, with S. mauritianum placed in the center and the test species in each of the corners. The five plant species did not necessarily contain equivalent amounts of floral material and no attempts were made to compensate for unequal numbers of flowers and flowerbuds. Fifteen reproductively active weevils were randomly released in the cage. After five days, the positions of the weevils were recorded and the floral material was transferred to petri dishes and dissected some 4–5 days later to confirm feeding and oviposition. The numbers of flowers and flowerbuds of each species that supported feeding and oviposition were recorded as were the numbers of larvae recovered. Each test plant species was tested on 3–5 separate occasions and the positions of the test species were rotated on each occasion. 2.7. Risk assessment The risks to nontarget Solanum species were quantified (sensu Wan and Harris, 1997) by measuring the weevilÕs performance on each test plant, at different stages in the host selection process, as a proportion of that on S. mauritianum. The relative performance risk of A. santacruzi was determined against 17 nontarget test species which were fed on, or oviposited on, to varying degrees (Table 5). The performance criteria used were: (i) plant preference (R1 ), (ii) food acceptability (R2 ), (iii) oviposition preference (R3 ), and (iv) larval survival (R4 ). Plant preference and oviposition preference were determined by the proportion of weevils present, and the proportion of larvae recovered, on the test plants during the multi-choice tests (Table 4). Food acceptability was determined by the comparative feeding intensities on the test plants during the multi-choice tests, while larval survival was determined by the percentage of larvae that survived to adults following the paired-choice tests (Table 3). In the case of species that were not tested in the multi-choice arena, equivalent data from the paired-choice tests were used. The risk of A. santacruzi utilizing a nontarget plant was assessed in two ways. Firstly, the risk of feeding damage was calculated as the product of the plant preference and food acceptability scores (i.e., R1 R2 ). Secondly, the risk of the weevil establishing viable reproductive populations was assessed by the product of the oviposition preference and larval survival scores (i.e., R3 R4 ). For each criterion, R represents the weevilÕs performance on the test plant relative to that on S. mauritianum (Table 5). To facilitate calculation, zero values were recorded as 0.001 (sensu Wan and Harris, 1997).

306


Table 3 Host selection of A. santacruzi adults as determined by their position, feeding intensity, and oviposition on different species of Solanum during pairedchoice testsa Test plant pairs

Mean ( SE) number of adultsb

Mean ( SE) feeding scoreb;c

% Buds with larvaed;e

Mean ( SE) number of larvaeb

% Larval survivale

Mean ( SE) number of days to adultsb

S. mauritianum S. burchellii

11.3 2.3 a 5.5 1.2 b

2.8 0.6 a 0.8 0.5 b

46.6 a 14.3 b

23.3 8.7 a 2.8 1.7 b

59.7 a 14.3 b

27.3 1.1 a 20.0 a

S. mauritianum S. capense

13.7 2.6 a 3.3 1.5 b

2.7 0.3 a 1.3 0.9 a

48.3 a 30.8 a

27.3 8.9 a 3.3 3.3 b

83.8 a 44.4 b

20.7 0.6 a 25.8 0.5 b

S. mauritianum S. chenopodiodes

13.0 2.0 a 1.0 0 b

3.5 0.5 a 0b

71.4 a 0b

15.0 6.0 a 0b

80.9

21.9 0.4

—

—

S. mauritianum S. coccineum

12.7 0.3 a 4.3 0.9 b

3.0 0 a 2.0 0.6 a

29.3 a 18.6 a

31.7 6.7 a 4.0 0.6 b

69.1 a 66.7 a

20.3 0.5 a 16.3 0.5 b

S. mauritianum S. duplo-sinuatum

15.7 1.5 a 2.3 0.9 b

3.0 0 a 0.7 0.7 b

35.9 a 0b

22.3 7.9 a 0b

40.9

17.9 0.4

—

—

S. mauritianum S. incanum

7.3 2.0 a 10.0 2.1 a

3.3 0.3 a 2.0 0.6 a

31.0 a 51.7 b

18.7 4.5 a 9.7 4.8 a

61.7 a 10.0 b

20.6 0.6 a 19.5 1.5 a

S. mauritianum S. linnaeanum

8.3 1.7 a 6.7 1.2 a

3.0 0 a 2.0 0 a

28.1 a 28.1 a

32.3 6.9 a 9.0 4.5 b

52.4 a 43.8 a

20.8 0.2 a 21.0 0.5 a

S. mauritianum S. cf. linnaeanum

10.7 2.3 a 6.3 1.2 a

3.3 0.3 a 3.0 0 a

28.7 a 23.9 a

28.7 6.2 a 11.7 2.3 b

49.2 a 46.2 a

18.4 0.4 a 20.4 1.2 a

S. mauritianum S. melongena (eggplant)

10.3 0.7 a 4.0 1.0 b

3.0 0 a 0.3 0.3 b

42.4 a 3.7 b

28.0 3.5 a 0.3 0.3 b

54.1 a 100 a

18.9 0.3 a 19.0 a

S. mauritianum S. nodiflorum

11.3 2.0 a 6.8 1.3 a

3.0 0 a 1.3 0.8 b

22.7 a 1.7 b

13.8 3.9 a 1.3 1.3 b

47.6 a 25.0 a

18.7 0.7 a 31.0 b

S. mauritianum S. panduriforme

9.0 1.8 a 8.2 0.9 a

3.0 0 a 3.0 0 a

33.8 a 37.9 a

22.0 10.2 a 11.4 3.8 a

69.6 a 59.6 a

22.5 0.3 a 22.4 0.5 a

S. mauritianum S. pseudocapsicum

6.3 2.7 a 8.3 1.5 a

2.7 0.3 a 2.7 0.3 a

32.0 a 28.3 a

16.3 5.2 a 7.0 3.1 a

60.0 a 83.3 a

21.4 1.4 a 21.5 1.2 a

S. mauritianum S. retroflexum

12.3 2.9 a 3.0 2.0 b

3.3 0.3 a 0b

40.7 a 0b

39.0 5.3 a 0b

45.4

16.9 0.4

—

—

S. mauritianum S. sisymbriifolium

14.7 0.3 a 0.3 0.3 b

3.0 0 a 0b

39.4 a 0b

11.3 2.2 a 0b

62.9

26.7 3.9

—

—

S. mauritianum S. tomentosum

11.0 3.2 a 6.0 3.2 a

3.7 0.3 a 2.0 1.2 a

66.7 a 36.5 b

19.7 8.4 a 8.3 6.9 a

60.0 a 47.4 a

20.5 1.7 a 49.3 8.4 b

S. mauritianum S. tuberosum (potato)

10.5 1.6 a 8.0 1.8 a

3.5 0.3 a 1.8 0.6 b

26.3 a 22.1 a

30.8 5.3 a 9.5 4.6 b

61.9 a 34.5 b

17.9 0.2 a 23.4 1.1 b

S. mauritianum S. viarum

15.0 0.6 a 1.7 0.3 b

4.0 0.6 a 1.3 0.7 b

47.9 a 0b

24.3 14.4 a 0b

58.6

16.9 0.9

—

—

a

Host suitability is indicated by larval survival and development to adults. Means compared by t tests; those followed by the same letter are not significantly different (P > 0:05). c Feeding categories defined in text. d Includes mature and immature flowerbuds. e Totals compared by v2 tests (2 2 Tables); those followed by the same letter are not significantly different (P > 0:05). b

3. Results 3.1. Field surveys During the survey involving six native South American Solanum species and cultivated eggplant (Solanum melongena), four species of Anthonomus were collected on five Solanum species (Table 1). A. morticinus and

Anthonomus tenebrosus Boheman were collected only on S. mauritianum and an unidentified Solanum species, respectively. A. santacruzi was regularly encountered on S. mauritianum (57% of the sites surveyed) but was also collected in significant numbers on Solanum fastigiatum Willd. at one site where no S. mauritianum was present. However, no weevils were recovered on S. fastigiatum at four other sites where this species was surveyed,


307

Table 4 Host selection of A. santacruzi adults as determined by their position, feeding, and oviposition on different species of Solanum during multi-choice tests in a large walk-in cage Test plant species

Mean ( SE) number of adultsa

% Floral material fed onb;c

% Floral material oviposited onc;d

Mean( SE) number of larvaea

S. S. S. S. S. S. S. S. S. S. S. S. S. S. S.

3.8 0.6 2.3 0.6 0.2 0.2* 0.8 0.2* 1.0 0.6* 1.3 0.9* 3.0 1.1 2.0 0.7 0.7 0.7* 0* 2.0 1.2 0* 3.3 0.6 2.0 0 0*

27.9 7.1 0 1.3 2.1 5.4 4.6 0 0.1 0 2.6 0 12.1 0 0

20.9 2.9 0 0 0 1.6 3.9 0 0.1 0 0 0 11.8 0 0

12.0 3.0 0.8 0.8* 0* 0* 0* 0.3 0.3* 3.0 1.9 0* 0.3 0.3* 0* 0* 0* 5.3 5.3 0* 0*

mauritianum burchellii coccineum duplo-sinuatum incanum linnaeanum cf. linnaeanum melongena (eggplant) nodiflorum panduriforme pseudocapsicum retroflexum tomentosum tuberosum (potato) viarum

a Means compared by one-way ANOVA; test species followed by the asterisk were significantly different to S. mauritianum (P < 0:05; DuncanÕs multiple range test). b Includes all floral material (open flowers; mature and immature flowerbuds). c Totals compared by v2 tests (2 2 Tables); all test species were significantly different (P < 0:05) to S. mauritianum. d Includes mature and immature flowerbuds.

Table 5 Risk analysis on the performance of A. santacruzi adults on nontarget Solanum plants relative to that on S. mauritianuma Test plant species S. S. S. S. S. S. S. S. S. S. S. S. S. S. S. S. S. S.

mauritianumc burchellii capensec chenopodiodes coccineumc duplo-sinuatum incanumc linnaeanumc cf. linnaeanum melongena (eggplant) nodiflorumc panduriformec pseudocapsicumc retroflexumc sisymbriifoliumc tomentosum tuberosum (potato) viarum

Plant preferenceb (R1 )

Food acceptabilityb (R2 )

Feeding risk (R1 R2 )

Oviposition preferenceb (R3 )

Larval survivalb (R4 )

Reproductive risk (R3 R4 )

1.0 0.6052 0.2408 0.0769 0.0526 0.2105 0.2631 0.3421 0.7894 0.5263 0.1842 0.001 0.5263 0.001 0.0204 0.8684 0.5263 0.001

1.0 0.2544 0.4814 0.001 0.001 0.0465 0.0752 0.1935 0.1648 0.001 0.0035 0.001 0.0931 0.001 0.001 0.4336 0.001 0.001

1.0 0.153 0.115 7.6 105 5.2 105 9.7 103 0.019 0.066 0.130 5.2 104 6.4 104 1.0 106 0.048 1.0 106 2.0 105 0.376 5.2 104 1.0 106

1.0 0.0666 0.1208 0.001 0.001 0.001 0.001 0.025 0.25 0.001 0.025 0.001 0.001 0.001 0.001 0.4416 0.001 0.001

1.0 0.2395 0.5298 0.001 0.9652 0.001 0.1620 0.8358 0.9390 1.8484 0.5252 0.8563 1.3883 0.001 0.001 0.7657 0.5573 0.001

1.0 0.015 0.063 1.0 106 9.6 104 1.0 106 1.6 104 0.020 0.234 1.8 103 0.013 8.5 104 1.3 103 1.0 106 1.0 106 0.338 5.5 104 1.0 106

a

Performance values are proportional to S. mauritianum (see Tables 3 and 4). Zero values were designated 0.001 to facilitate calculations. c Listed as exotic or native weeds in South Africa (Wells et al., 1986). b

suggesting that it may not be a normal host. Anthonomus sisymbrii Hustache was collected on Solanum sisymbriifolium Lam. at two sites and on Solanum viarum Dun. at one site where S. sisymbriifolium and S. viarum grew together. None of the four Anthonomus species were recorded on the two unsprayed eggplant cultivations that were surveyed, despite their host plants growing in very close proximity (i.e., within 20 m) (Table 1). These

preliminary observations suggested that all four species of Anthonomus are likely to have very narrow field host ranges. 3.2. Biology of A. santacruzi Adults of A. santacruzi are black and around 2–3 mm in length. The elytra are striated and the second to sixth

308


interstrial segments contain conspicuous patches of white scales which appear as thin vertical bands. These series of 4–5 white bands on each of the elytra enable this species to be separated from A. morticinus in which a single white band is confined to the fourth interstria (Clark and Burke, 1996). The weevils feed on the stamens of open flowers, but also chew through the petals of unopened buds to feed on the stamens. Feeding, which causes the abortion and abscission of flowers and flowerbuds, is not restricted to floral material. Indeed, the weevils will feed on the apical leaflets and shoot tips of plants that are devoid of flowers for extended periods, causing appreciable damage. However, feeding on nonfloral material was seldom observed when there was sufficient floral material. The adults fly actively and have excellent ability to disperse. The females oviposit in both immature and mature flowerbuds of S. mauritianum, but will occasionally accept open flowers for oviposition. The preoviposition period lasts around four days and the duration of oviposition varied between 43 and 114 days (mean SE ¼ 85.3 15.0). The minute rounded eggs, around 0.3 mm in diameter, are inserted into small depressions which are chewed into the sides of the anthers. The firstinstar larvae feed within the anthers and during their further development consume the entire contents of the flowerbuds. The larvae usually develop singly, although 2–3 larvae are often recovered in a single flowerbud. The larvae have three instars and development takes 10–18 days (mean SE ¼ 14.1 0.2) from oviposition to pupation. Larval feeding prevents the buds from opening and the larvae thus create an enclosed area in which they develop. Pupation occurs in the chamber formed by larval feeding and lasts 4–10 days (mean SE ¼ 6.0 0.1). The emerging adults chew their way out of the infested flowerbuds. The time from oviposition to adult emergence is 15–25 days (mean SE ¼ 19.6 0.2). Four colonies of 10 newly-emerged weevils survived for 99–122 days (mean SE ¼ 110.3 4.8) when sustained on floral bouquets, resulting in an overlap of generations. During this time, the colonies produced between 82 and 471 larvae, with between 16 and 59 larvae produced per female (mean SE ¼ 37.6 8.7). By contrast, colonies survived for 131–182 days (mean SE ¼ 154.3 14.9) when sustained on potted plants devoid of floral material, which was significantly longer (P < 0:05) than colonies kept on floral bouquets. Weevil populations thus seem well capable of persisting on S. mauritianum populations during periods when shortages of floral material may arise. For culturing purposes, colonies of A. santacruzi were best maintained on floral bouquets in containers as opposed to potted plants in cages. Potted plants showed a tendency to abort their flowers and flowerbuds, which may be a response to the artificial laboratory conditions,

attack by the weevils, or a combination of the two. In any event, larvae recovered from buds harvested from potted plants generally took longer to develop and were more prone to desiccation than those harvested from bouquets. The biology of A. santacruzi suggests that it is a very promising agent for the biological control of S. mauritianum. Attributes include: (i) the high fecundity, longevity, and mobility of the weevils, (ii) short generation times and overlapping generations to facilitate rapid population increases, and (iii) the destructive nature of adult and larval feeding which prevents fruit set and can thus reduce the weedÕs extensive seed dispersal. 3.3. Host specificity of A. santacruzi Adult no-choice tests. Tests involving colonies of newly-emerged weevils clearly indicated that A. santacruzi was unable to feed on the leaves of any plants outside the genus Solanum and survival after 14 days varied from 3 to 20% (Table 2). Unsuitable species thus included solanaceous crops like green pepper (Capsicum annuum), tomato (Lycopersicon esculentum), and cape gooseberry (Physalis peruviana). However, these tests also indicated that the weevils could sustain themselves on the foliage of several nontarget Solanum species. Although feeding damage on the nontarget Solanum species was mostly lower than the extensive damage recorded on S. mauritianum, there was no significant decrease in survival on 15 of the 18 species (Table 2). This included both cultivated species, potato (Solanum tuberosum) and eggplant (S. melongena), which supported normal feeding and survival. Feeding was considerably reduced on two of the three exotic Solanum species tested, with survival significantly reduced on one species. However, feeding was reduced on only five of the 13 native South African Solanum species tested, with survival significantly reduced on only two species. Although the physiological host range (sensu Cullen, 1990) of A. santacruzi does not extend beyond the genus Solanum, it included several nontarget cultivated and native species. Paired-choice tests. During tests involving S. mauritianum, potato, eggplant, three exotic, and 12 native Solanum species, significantly more weevils were recorded on S. mauritianum than on nine of the 17 test species (Table 3). Adults fed to varying degrees on 14 test species during these trials and on only six of these (eggplant, potato, one exotic, and three native species) was the intensity of feeding significantly lower than on S. mauritianum. In addition, A. santacruzi oviposited in the buds of 12 of the 17 test species, including potato, eggplant, one exotic and nine native species (Table 3). In only four of these 12 instances, were the proportions of floral material that contained larvae significantly higher on S. mauritianum than on the test species.


However, more larvae were always recovered on S. mauritianum than on any of the 12 test species that supported oviposition, and larval recoveries were significantly lower on eight of these, including eggplant and potato. Overall, with some exceptions, these data did not demonstrate conclusively that A. santacruzi displays strong feeding and oviposition preferences for S. mauritianum. Of the 12 nontarget species that were oviposited on during these tests, all supported development to adults, and larval survival was significantly lower on only four of these (including potato) than on S. mauritianum (Table 3). The duration of development to adults was largely unaffected by host plant and only four test species (including potato) caused significantly longer developmental times than recorded on S. mauritianum (Table 3). With few exceptions, these data also did not demonstrate conclusively that the most ÔsusceptibleÕ of the test species were unsuitable for sustaining populations of A. santacruzi. Multi-choice tests. During trials involving S. mauritianum, potato, eggplant, two exotic, and 10 native Solanum species, the weevils were recorded on 11 nontarget species, with significantly more recorded on S. mauritianum than on five of these (Table 4). However, despite regular contact between the weevils and the nontarget plants, six species (including potato and eggplant) did not support any feeding while the remaining eight species supported significantly lower levels of feeding than on S. mauritianum. Similarly, the proportions of floral material that contained larvae were significantly higher on S. mauritianum than on all 14 test species, with no oviposition recorded on nine of these species (Table 4). Consequently, more larvae were recovered on S. mauritianum than on any of the five species that supported oviposition. However, these differences were not significant in the case of two native species despite the recovery of at least 50% fewer larvae on them. Neither potato nor eggplant was selected for feeding or oviposition and some native species (e.g., Solanum coccineum, Solanum incanum, and Solanum panduriforme) that were readily accepted during the paired-choice trials were also avoided in the multi-choice trials. These trials, which were conducted under considerably less restrictive conditions, suggested that only two nontarget species, the native Solanum cf. linnaeanum and Solanum tomentosum, have any likelihood of being selected for feeding and oviposition in the field. 3.4. Risk assessment The feeding risk calculations (Table 5) indicated that 12 of the 17 nontarget species, including potato, eggplant, and seven native species, showed a very low (