Appendix 4. The host range of the potential biological control agents Neolema abbreviata and Lema basicostata. 4.1 Summary 4.2 Source of insects 4.3 Selection of plant species to test 4.4 Test methods 4.5 Test results 4.6 Discussion and conclusions 4.1 Summary The ability of young larvae of Neolema abbreviata and Lema basicostata to feed and develop when placed on cut foliage of 18 test plant species (including T. fluminensis controls) was assessed in the laboratory. The acceptability of these plants as a food source for adult beetles was also assessed. There are no native plants in New Zealand that are even remotely related to tradescantia. Apart from some ornamental house plants, there are no closely-related exotic plants in New Zealand that are valued either. Only limited host range testing was required because of these two characteristics. Nīkau palm was tested because it is the most closely related native plant in New Zealand, even though it is not even in the same order as the weed. There was no semblance of attack on this species. Several host plants supported complete development of L. basicostata and N abbreviata larvae. All of these plant species belong within the tribe Tradescantieae. Similarly, when confined on test plants, adult beetles did not feed significantly on any plants outside the Tribe. This suggests that adults would not lay eggs on these plants. These test results indicate that the physiological host range of both N. abbreviata and L. basicostata lies within the Family Commelinaceae, and possibly within the Tribe Tradescantieae. Given the extreme taxonomic distance between this tribe and any valued plant, the risk of significant adverse effects on such plant species in New Zealand (whether native or exotic) is considered to be negligible. 4.2 Source of insects The population of Lema basicostata Monros that was tested combined insects collected from three sites in Brazil: Curitiba, (Paraná), Caxias do Sul (Rio Grande do Sul) and Serra do Rio do Rastro (Santa Catarina). Two populations of Neolema abbreviata (Lacordaire) collected from two sites in Brazil: Curitiba (Paraná) and Lages (Santa Catarina), were tested. The results of these tests were combined (see Section 4.5) 4.3 Selection of plant species to test A list of plants to be included in host range tests was developed. In order for this list to have the best possible scientific basis, a detailed review of the systematics of the genus Tradescantia and the family Commelinaceae was undertaken (Table 4.1). The most important finding of this review was that no native New Zealand plant species fall into the family Commelinaceae, or even the order Commelinales. Similarly, apart from a small range of exotic species sold as houseplants, there are no related plants of significant economic benefit in New Zealand. Because Tradescantia is so isolated from the rest of the flora, host specificity to species, genus, or even subfamily/family is, in theory, not essential to ensure safety in this programme. The phylogenetic relationship between tradescantia and other plants can be found in Table 4.1. The initial list of test plant species was developed for testing in Brazil, and was modified for New Zealand, taking into account the availability of plant species in each country. The plant species tested in containment in New Zealand are listed in Tables 4.2 and 4.3. The native nīkau palm (Rhopalostylis sapida; Order Arecales), was selected for testing because even though it is in a separate order, this is probably the closest New Zealand relative to T. fluminensis (P. Heenan, Landcare Research, pers. comm.). One species belonging to the Lilliales, Zingiberales and Cyperales were tested as 3 assorted representatives of the 18 orders in the class Liliopsida other than Commelinales. Fourteen species of 6 genera belonging to the Family Commelinaceae were tested (including T. fluminensis controls). The common names of test plants can be found in Table 4.4. 4.4 Test methods Adults of both Neolema abbreviata and Lema basicostata are external foliage feeders notching leaves, with Lema basicostata also notching stems. Neolema abbreviata can often be found feeding on the new leaves or resting in unfurled leaves of shoot tips. Young Neolema abbreviata larvae burrow into and mine growing tips or leaves of tradescantia often emerging as larger larvae to feed externally on leaves and pupate on the underside or in furled leaves. Larvae of Lema basicostata enter and mine tradescantia stems only emerging to pupate in the litter or soil surface. Both beetles were observed causing moderate damage in the field, but in laboratory conditions readily caused major die back, often killing potted plants. Informed by this biology, two test procedures were designed that followed standard, internationally accepted protocols. Research was conducted in the Landcare Research containment facility at Lincoln. Details of the test procedures are as follows: Adult no-choice feeding tests Each replicate used one adult beetle from the rearing culture, with a minimum of ten replicates for each beetle species (except for T. cerinthoides and T. sillamontana where only 5 tests could be completed). A freshly excised leaf of the test plant was placed on damp filter paper in a 9-cm Petri dish and an adult beetle was added. Tests were scored after 1 day, and then at intervals of 1–2 days, for a total of 5–8 days. On each occasion, leaf damage was recorded visually into five categories (0, trace = almost zero damage; 1 = minor damage; 2 = <50% of leaf consumed; 3 = >50% of leaf consumed) and the filter paper and leaves replaced with fresh ones. Damage scores in Tables 4.2 and 4.3 are means across replicates for the appropriate time periods (e.g. days 1-3 and 4-7 for N. abbreviata). Any eggs laid were counted and removed. Tests were initiated on a range of dates, with concurrent, matching T. fluminensis controls. The significance of differences between the intensity of feeding on test plants and on T. fluminensis controls was tested by T-test, and the results are presented in Tables 4.2. and 4.3 (NS = not significant; * = P<0.05; ** = P<0.01; ***= P<0.001). Larval feeding/development tests First or second instar larvae of each species were removed from T. fluminensis plants in rearing cages and were placed onto cut test plant material on filter paper in a 9 cm Petri dish, one larvae per dish. Larvae were assessed at 1 or 2-day intervals at first, and then at longer intervals. The filter paper was moistened if dry, and the plant material was renewed if it had visibly deteriorated or been substantially consumed. For each assessment, leaf damage was recorded visually into five categories (0, trace = almost zero damage; 1 = minor damage; 2 = <50% of leaf consumed; 3 = >50% of leaf consumed). Damage scores in Tables 4.2 and 4.3 are means across replicates for the appropriate time periods (e.g. days 1-20 and 21-35 for N. abbreviata). The tests terminated if all larvae were dead or had pupated. Tests were initiated on a range of dates, with concurrent, matching T. fluminensis controls. The significance of differences between the intensity of feeding on test plants and on T. fluminensis controls was tested by T-test, and the results are presented in Tables 4.2. and 4.3 (NS = not significant; * = P<0.05; ** = P<0.01; ***= P<0.001). 4.5 Results Neolema abbreviata The results for the two provenances of N. abbreviata were combined because their levels of attack on different tests plants were very similar (no significant differences between the provenances for plants in the genus Tradescantia – Kruskall-Wallis Test P>0.1; minor differences in other plant species in the Commelinaceae (P<0.05 – indicating a little more attack on the genus Gibasis by adult beetles from provenance 1 versus provenance 2). There were no differences between provenances (and no feeding) in plants outside the Order Commelinales. In 3 of 15 replicates larvae caused insignificant trace feeding damage to leaves of the exotic species Pontederia cordata, which belongs to a family that is closely related to the Commelinaceae. Insignificant trace feeding was observed once by an adult beetle on Anemanthele lessoniana. Otherwise, there was no adult or larval feeding observed, and no eggs were laid, on any species outside the family Commelinaceae. No larvae transferred to these test plants completed development (Table 4.2). With the exception of T. albiflora, adult beetles fed significantly less intensely on test plants within the Commelinaceae than on T. fluminensis controls (Table 4.2), however some feeding was recorded on 11 out of 14 species tested in the family. The plant species not fed on were Dichorisandra thyrsifolia and Cyanotis somaliensis, which are in different subtribes to those attacked (Table 4.1). Thus adult feeding appears to be confined to plants in the same subtribe as the target weed (excluding the insignificant trace nibbling mentioned above). Larvae fed on all test species in the same tribe as the target weed. Successful development to pupation was more variable (9 of 14 plant species) but still occurred across the tribe Tradescantieae. These tests indicate that larvae have a slightly wider host range than adults, with feeding and development to plants within the same tribe as the target weed (excluding the insignificant trace nibbling mentioned above). Lema basicostata Larvae caused insignificant trace feeding damage to Pontederia cordata, but otherwise no adults or larvae fed at all on any plants outside the Family Commelinaceae (Table 4.3). Several eggs were laid, but at only 1-2% of the intensity of oviposition on controls.