88 Plant Protection Quarterly Vol.12(2) 1997 management strategies, particularly in identifying situations where biological control can be implemented, Biological control of St. John’s wort in New Zealand • a network of trained officers who are able to advise farmers on the integrated Pauline Syrett, Manaaki Whenua – Landcare Research, PO Box 69, Lincoln, management of St. John’s wort and New Zealand. how biological control can be incorpo- rated, and what to expect, and • community participation and a sense of Summary Problem status of St. John’s wort in ownership of the implementation of Three species introduced for bio- New Zealand biological control. logical control of St. John’s wort (Hyperi- The first record of St. John’s wort in New The networks also provide a pathway for cum perforatum) have established in New Zealand was from Great Barrier Island in the exchange of information between the Zealand. They are two leaf-feeding bee- 1869, and the first record on the South Is- Network Co-ordinators and co-operators tles, hyperici and C. land was in 1896 (Jessep 1970). It is now and will enable other biocontrol agents for quadrigemina, and a gall-forming fly, widely distributed through the drier areas St. John’s wort to be distributed should Zeuxidiplosis giardi. The earliest intro- on the eastern side of New Zealand, and they become available in the future. duction, C. hyperici in 1943, is the most more common in the South Island than the To successfully implement biological common and widespread of the three, oc- North Island. From the 1940s it occupied control throughout the entire distribution curring on St. John’s wort plants through- large areas of pastoral land from the of St. John’s wort in Australia, a good map- out their distribution. The second leaf- Awatere Valley in Marlborough to the ping and databasing system is required to feeding (first released in 1963) was Matukituki Valley in Otago. Miller (1970) identify areas which have not yet been tar- not rediscovered until 1984, and occurs in reported it as ‘a weed of considerable im- geted for biological control. Feedback mixed populations with C. hyperici. portance, particularly in the higher re- from network co-operators on the These St. John’s wort occasionally gions of low rainfall.’ He described it as progress of mites is important and needs outbreak to very large populations, caus- occurring in dense stands covering large to be encouraged in the future develop- ing complete defoliation of host plants. areas, and displacing pasture species. Its ment of the network. Experience with this C. hyperici seems to be the dominant spe- toxicity to stock was also of concern. program has indicated that it is difficult to cies in New Zealand, and studies of the Connor (1977) noted that the greatest maintain a flow of information without reproductive diapause of the two species problem of photosensitization (caused by the continued co-ordination and facilita- indicate that this is because C. hyperici’s hypericin in the weed) was with high tion of Network Co-ordinators. reproductive strategy is more successful country sheep on the South Island, al- in areas with colder winters. Z. giardi has though cattle also suffered from the dis- References a limited distribution in the northern part ease. It was particularly troublesome Clark, L.R. and Clark, N. (1952). A study of of the South Island, and where it occurs, when mustering (stock losses occurred the effect of Chrysomela hyperici Forst. plants appear stunted, with fewer flow- when were driven through wa- on St. John’s wort in the Mannus Valley, ers. There are no longer reports of areas ter). In more recent years St. John’s wort N.S.W. Australian Journal of Agricultural in New Zealand where St. John’s wort is a has gradually decreased in prominence. Research 3, 29-59. problem weed, and we conclude that suc- This has been partially attributed to Jupp, P.W. (1992). The biological control of cessful biological control with is changes in land management practices fol- St. John’s wort program. Final report to as least partially responsible for this lowing substantial reduction in stocking the Meat Research Corporation. change. levels on extensively grazed pastoral land. Jupp, P.W. (1996). The establishment of a Fraser (1987) found that few districts in the distribution network for the mite Introduction South Island regarded St. John’s wort as Aculus hyperici for the control of St. New Zealand’s biological control program more than a roadside weed: in only one John’s wort. Final Report to the MRC, for St. John’s wort ( perforatum (Hakataramea) was the plant actively con- pp. 22-7. L., Clusiaceae) was an adjunct of the early trolled on pastoral land. In the last decade Jupp, P.W., Briese, D.T. and Cullen, J.M. part of the Australian program. As each of the significance of the weed has declined (1997). Evidence of resistance in Hyperi- three insect control agents was introduced further: no Regional Council reports it as cum perforatum to a biological control into Australia in the 1940s and 1960s, ship- even a minor problem (L. Hayes personal agent, the eriophyid mite, Aculus ments were sent from Australia to New communication). hyperici. Plant Protection Quarterly 12, Zealand (Wilson 1943). The outcome of 67-70. New Zealand’s program has been some- Introduction and establishment of Jupp, P.W. and Cullen, J.M. (1996). Ex- what different from the Australian experi- biological control agents pected and observed effects of the mite ence, so the purpose of this paper is to re- Two leaf-feeding beetles and a gall-form- Aculus hyperici on St. John’s wort, Hy- view existing knowledge of the fate of St. ing fly were introduced into New Zealand pericum perforatum, in Australia. Pro- John’s wort and its three insect control to control St. John’s wort, and all three ceedings of the IX International Sympo- agents in New Zealand. have established (Table 1). sium on Biological Control of Weeds, pp. 365-70. Table 1. Species of insects introduced into New Zealand for biological control Shepherd R.C.H. (1984/85). The present of St. John’s wort. status of St. John’s wort (Hypericum Species Origin Date introduced perforatum L.) and its biological control agents in Victoria, Australia. Agricul- Chrysolina hyperici Australia ex England 1943 ture, Ecosystems and Environment 12, C. quadrigemina Australia ex France; 1963 141-9. Canada ex USA Wapshere, A.J. (1984). Recent work in Eu- ex Australia ex France 1990 rope on biological control of (Guttiferae) for Australia. Zeuxidiplosis giardi Australia ex California Entomophaga 29, 145-6. ex France 1960–61 Plant Protection Quarterly Vol.12(2) 1997 89

Figure 1. Numbers of stems of St. John’s wort (mean) recorded in insecticide sprayed and unsprayed half-plots, Mt. Gerald Station, Lake Tekapo, 1983–86.

Chrysolina hyperici (Förster) have been recovered from one of these, controlling St. John’s wort, but there were (Coleoptera: Chrysomelidae) near Lake Benmore in South Canterbury. other areas where its impact was less Adult beetles of C. hyperici were imported Mixed populations of the Chrysolina spp. marked (Hancox et al. 1986). An insecti- from Australia in the summer of 1943 and beetles at Clyde, Central Otago, (from the cidal exclusion trial was carried out from after one day in quarantine, were released earlier releases) comprised around 25% C. 1983 to 1986 at Mt. Gerald Station, South directly into the field (Miller 1970). By the quadrigemina in 1991 (Schöps et al. 1996). Canterbury to measure the effect of C. following year there were large popula- hyperici on stem production of St. John’s tions of beetles at the original release site Zeuxidiplosis giardi Kieffer (Diptera: wort (Syrett and Hancox 1985, Fraser in the Awatere Valley, Marlborough, and Cecidomyiidae) 1987). Although total beetle numbers an 80 m2 area cleared of weed. Over the Introductions of this gall midge were were relatively low at this site during the next few years beetles were collected and made in 1960–61, also from Australia. Dif- period of this experiment, in spring and transferred to areas throughout New Zea- ficulties were experienced in rearing the early summer significantly more stems land (Hancox et al. 1986). They established midge, and smaller numbers of flies were were counted in the plots protected from readily on St. John’s wort throughout its available for release than planned. Never- beetles than in those where beetles were range. theless, they soon established in a re- present (Figure 1). A reduction in the stricted area near Nelson, and subsequent growth of St. John’s wort in spring is valu- Chrysolina quadrigemina (Suffrian) releases were made throughout the South able to farmers because in many areas this (Coleoptera: Chrysomelidae) Island (Given and Woods 1964, Hancox et weed is one of the earliest plants to pro- A second species of Chrysolina was intro- al. 1986). The gall midge has established duce new growth, and is heavily grazed duced in 1963, and over 100 000 adult bee- well over a limited region in the northern then. A reduction in the amount of St. tles were released in the period to 1968 at South Island, including at an altitude of 830 John’s wort present in spring therefore sites throughout the South Island (Hancox m in north-west Nelson, but failed in other reduces the chance of stock poisoning et al. 1986). Only one recovery of C. areas. No recent attempts have been through preferential grazing (Fraser quadrigemina was reported during suc- made to spread the midge to other areas, 1987). Field observations showed that ceeding years: four males were collected partly because of difficulties in rearing the early spring prostrate growth was heavily at Ben Ohau in 1974. But in 1984 beetles midge on St. John’s wort plants under damaged by larval feeding, to the extent were identified from widely separated laboratory conditions. Early cultures for that few stems remained. sites in Central Otago and Marlborough field release (by planting out infested (Fraser and Emberson 1987) and since plants) used Hypericum pulchrum L. be- Relationship between C. quadrigemina then from a number of sites in between cause it was more amenable to laboratory and C. hyperici (unpublished data). In an effort to obtain handling (Given and Woods 1964). Because C. quadrigemina occurs only in beetles better adapted to New Zealand’s mixed populations with C. hyperici it is not climate, new material of C. quadrigemina Impact of biological control agents easy to measure its impact on St. John’s from Canada was imported in 1990 and C. hyperici wort. Work described by Fraser and Em- released in 1991. Releases were made at In parts of New Zealand C. hyperici was berson (1987) and Schöps et al. (1996) in- four sites, and C. quadrigemina beetles reported as spectacularly successful in vestigated differences in the reproductive 90 Plant Protection Quarterly Vol.12(2) 1997 diapause of the two Chrysolina species that cause for concern. Today, however, St. Chrysomelidae) in New Zealand. New might explain their relative success at con- John’s wort rarely builds to populations Zealand Entomologist 9, 57-9. trolling St. John’s wort in New Zealand. that become conspicuous. Since C. Given, B.B. (1967). Biological control of quadrigemina has generally been the more weeds and insect pests in New Zealand. Laboratory experiments By subjecting successful of the two Chrysolina species at Eleventh Pacific Science Congress, To- adult C. hyperici that have just entered aes- controlling St. John’s wort worldwide kyo, 1966, Symposium 28: Natural en- tivation to four photoperiod treatments: (Julien 1992), it is tempting to speculate emies in the Pacific area. Mushi 39 short photo phase; long photo phase; that biological control has become more (suppl.), 17-22. short then long photo phase; and long effective in New Zealand as C. quad- Given, B.B. and Woods, G.F. (1964). Gall then short photo phase under both moist rigemina populations have increased. The midge to assist in controlling St. John’s and dry conditions, Fraser and Emberson recent introduction of a more ‘cold wort. New Zealand Journal of Agriculture (1990) showed that day length, rather than adapted’ strain of C. quadrigemina from 198, 61-3. autumn rainfall, is the more likely trigger Canada coincided with the reappearance Hancox, N.G., Syrett, P. and Scott, R.R. for the end of aestival diapause. Similar ex- of C. quadrigemina in significant numbers (1986). Biological control of St. John’s periments by Schöps et al. (1996) with both from the 1960s introductions, possibly as a wort (Hypericum perforatum) in New C. hyperici and C. quadrigemina showed result of adaptation to the New Zealand Zealand; a review. Plant Protection Quar- that under short days C. quadrigemina ter- environment. Shepherd (1985) noted that terly 1, 152-5. minated aestivation about six weeks ear- the proportion of C. quadrigemina relative Jessep, C.T. (1970). St. John’s wort in New lier than C. hyperici. to C. hyperici had increased over a 30 year Zealand. Tussock Grasslands and Moun- period in Victoria, but that in Australia bio- tain Lands Institute Review 20, 76-83. Field observations Schöps et al. (1996) logical control of St. John’s wort is not fully Julien, M.H. (1992). ‘Biological control of sampled a mixed population of Chrysolina effective because beetles do less well in weeds, a world catalogue of agents and spp. at Clyde from early November 1991 shaded situations. In New Zealand there is their target weeds’, 3rd edition, 186 pp. to early April 1992. Beetles were dissected little lightly wooded habitat where St. (CAB International, Wallingford, UK). to determine the stage of their reproduc- John’s wort might do well, but where bee- Miller, D. (1970). Biological control of tive development. Results indicated that C. tles would be ineffective. Perhaps a bal- weeds in New Zealand 1927–48. Infor- quadrigemina females began laying eggs in ance has been achieved between numbers mation Series No. 74, 104 pp. (New Zea- early to mid March, and C. hyperici in mid of the two Chrysolina spp. that results in land Department of Scientific and In- April. Schöps et al. (1996) suggested that better control of St. John’s wort now than dustrial Research, Wellington, New differences in phenology resulting from previously. The impact of Z. giardi is prob- Zealand). differing responses to the onset of short ably minor compared to that of Chrysolina Schöps, K., Syrett, P. and Emberson, R.M. days could explain why one Chrysolina spp. because its distribution is limited. (1996). Summer diapause in Chrysolina species may be more successful than the From being a major agricultural weed hyperici and C. quadrigemina (Coleop- other in a particular region. Eggs of C. in the 1930s and 1940s, St. John’s wort has tera: Chrysomelidae) in relation to bio- quadrigemina hatch in autumn and larvae become a roadside weed of only minor logical control of St. John’s wort, Hy- feed through the winter while C. hyperici significance in the 1990s. Although we can- pericum perforatum (Clusiaceae). Bulletin overwinters in the egg stage. As the egg is not directly attribute this change in status of Entomological Research 86, 591-7. the most cold-hardy stage, C. hyperici is to insect biological control agents, there is Shepherd, R.C.H. (1985). The present sta- well adapted to survive in regions with sufficient evidence to conclude that they tus of St. John’s wort (Hypericum cold winters, while C. quadrigemina has the have been a contributing factor. perforatum L.) and its biological control advantage that, in regions with milder agents in Victoria, Australia. Agricul- winters, it can feed on available host plant Acknowledgments ture, Ecosystems and Environment 12, material earlier than C. hyperici. Thanks to Drs. Eric Scott, Trevor Par- 141-9. tridge, and Megan Ogle-Mannering for Syrett, P. (1989). Hypericum perforatum L., Zeuxidiplosis giardi their helpful comments on an earlier draft St. John’s wort. In ‘A review of biologi- Given and Woods (1964) noted that the of this manuscript. cal control of invertebrate pests and gall midge was most effective on young weeds in New Zealand, 1874 to 1987’, seedling plants, and on autumn and win- References eds. P.J. Cameron, R.L. Hill, J. Bain and ter prostrate growth. Growing points at- Connor, H.E. (1977). The poisonous plants W.P. Thomas, pp. 353-6. Technical tacked by the midge made no recovery, in New Zealand. Bulletin No. 99, 247 pp. Communication No. 10. (CAB Interna- and growth of very young plants was ar- (New Zealand Department of Scientific tional Institute of Biological Control, rested completely by only one or two and Industrial Research, Wellington, Wallingford, Oxford, UK). galls. By 1967 patches of St. John’s wort New Zealand). Syrett, P. and Hancox, N.G. (1985). Effect heavily infested with galls produced no Fraser, B.R. (1987). Aspects of the biology of an introduced beetle on St. John’s flowers, and seedling plants were killed by of the St. John’s wort beetle, Chrysolina wort. Proceedings of 38th New Zealand midge activity (Given 1967). However, at hyperici (Forst.), relating to its role as a Weed and Pest Control Conference, ed. that stage no parasitoids were found biocontrol agent. Thesis, M. Appl. Sci., M.J. Hartley, 13–15 August, Rotorua, (Given 1967). Hancox et al. (1986) noted University of Canterbury, New Zea- New Zealand, pp. 154-7. that Torymoides sp. (Hymenoptera: land, 116 pp. Wilson, F. (1943). The entomological con- Torymidae) had been found in galls. It is Fraser, B.R. and Emberson R.M. (1990). trol of St. John’s wort (Hypericum not known what impact the parasitoid has Photoperiod and reproductive perforatum L.) with particular reference on gall midge populations (Syrett 1989). diapause in the St. John’s wort beetle, to the insect enemies of the weed in Chrysolina hyperici. Proceedings of VII southern France. CSIR Bulletin No. 169, Discussion International Symposium on Biological 87 pp. (CSIR Melbourne, Australia). Although in the years following successful Control of Weeds, 6–11 March 1988, establishment of C. hyperici in New Zea- Rome, Italy, ed. E.S. Delfosse, pp. 323-8. land reports indicated classic and spectacu- Fraser, Bruce and Emberson, Rowan larly successful biological control, there (1987). Rediscovery of Chrysolina remained areas where St. John’s wort was quadrigemina (Suffrian) (Coleoptera: