BiocontrolNews and Information 2003 Vol. 24 No. 4 77N–90N General News Testing Times for Testing were released against St John’s wort The above rare instances of the failure of (Hypericum perforatum) but they had not New Zealand’s past testing procedures The high profile reporting of non-target been tested against native Hypericum spe- occurred because plants that might have impacts of a few classical weed biocontrol cies. Hypericum japonicum and H. grami- been tested were not. There are just two agents over the past few years highlighted neum are uncommon plants and, although recorded cases where plants were tested, but one shortcoming of many, if not most, bio- none of the agents has been recorded on the testing failed to predict the non-target control programmes – related to not inade- them, they could be at risk. In many other attacks that subsequently occurred. The quate testing, but inadequate monitoring. In projects during that period, and in all broom seed beetle Bruchidius villosus and trying to answer the critics of biocontrol, its projects since 1990, it has been standard the gorse pod moth Cydia succedana have defenders could only point to the absence practice to test native plants so the omission unexpectedly attacked seed of other exotic of reported problems for the overwhelming of indigenous species from test plant lists Fabaceae, although pre-release testing sug- majority of the 1150 or more planned should not occur now. However, this does gested that this would not happen. Investi- releases of weed biocontrol agents that not imply that selecting what plants to test gations into both are continuing, but Fowler have been made worldwide. An equally is now always straightforward. noted that a common link is that both use overwhelming absence of post-release sur- seasonally ephemeral resources (young veys means that there is little concrete evi- The first agent to be released in New Zea- pods) whose phenology in comparison to dence about the non-target impact of most land, cinnabar moth, was tested against that of the agent differs slightly between of these releases. eight native species of Senecio before per- Europe and New Zealand, potentially In New Zealand, Landcare Research has mission was given for its release against offering novel no-choice situations to been dusting off old files and investigating ragwort. However, when ragwort is defoli- agents in the field after release. This gives the safety record of weed biocontrol there ated the moth occasionally attacks two spe- us a new concept to consider: Do agents since the first biocontrol agent, cinnabar cies of native Senecio (S. minimus and S. introduced for discrete seasonal resources moth (Tyria jacobaeae), was released biserratus) which were not tested. Why need more careful assessment? Should against ragwort (Senecio jacobaea) in were these two not included in the test plant more rigorous no-choice testing be consid- 1929. This audit will alert the country’s list? They were in a different genus at the ered in these circumstances? Withers biocontrol community to any lurking dan- time. This is the only recorded instance of believes that the answer lies in our ability to gers, and also identify past flaws that will an introduced weed biocontrol agent accurately interpret the results of host spe- help tighten up testing procedures for the attacking a native non-target species in cificity testing, something that retrospective future. New Zealand, but could such a testing analysis of the methods used in the past, omission happen again? It is unlikely. Plant combined with post release field assess- Landcare has been looking at some agents systematics have clearly progressed since ment, will help us with. that are now common in New Zealand, 1929 and if the testing were being con- checking what testing was done before each The message is that if weed biocontrol is to ducted now, these two species would be win the confidence of its often vociferous was released, and whether non-target attack included. However, even with the taxo- has occurred – or might yet. Reporting sceptics, it is likely to require more sophis- nomic relationships clarified, there is ticated yet transparent interpretation of the results at the 11th Symposium on the Bio- another source of uncertainty. logical Control of Weeds in Canberra this host range testing carried out, and poten- tially will become more time-consuming year, Simon Fowler said that, overall, the Alternanthera sessilis, a close relative of reliability of host specificity testing in New and (therefore) more costly as a result. At a alligator weed (A. philoxeroides) was not time when time and budgets are being cut Zealand in the past has been good although included in host specificity tests in that bio- a few gaps have been identified. this fills the biocontrol community with control programme. Both are exotic species dismay. As a direct consequence of the The largest gap in testing probably and indeed there are no native New Zealand greater cost some projects will not be occurred in some projects conducted plants in the family. But what was not fore- undertaken, while the increased scrutiny is between 1943 and 1982. Although early seen was that A. sessilis would subse- likely to raise the bar so high that poten- biocontrol projects included native species quently attain cultural importance as a new tially useful and harmless agents will be in specificity tests, during this period 13 vegetable crop for some sectors of the com- rejected. introductions of natural enemies were made munity. Short of including a fortune-teller which relied heavily on testing in other in the biocontrol project, this would have Source: Hayes, L (ed) (2003) In retrospect. countries, and New Zealand natives were been hard to predict but Toni Withers In: What’s new in biological control of not tested. In one case, thistles, the rationale (Forest Research), who has been working weeds. Annual review. Lincoln, New Zea- was that there are no native thistles in New with Fowler and others on this project, land; Landcare New Zealand Ltd 2003, No. Zealand so there were no closely related believes that the current rigorous process 25, pp. 7-9. native species to test. While there could for drawing up test plant lists means that a Contact: Simon Fowler, have been problems where introductions similar case would not now be overlooked. Landcare Research, Private Bag 92170, were made against other targets with New Of two agents introduced against alligator Auckland, New Zealand. Zealand relatives, no non-target impacts weed, one (the moth Arcola malloi) attacks Email: [email protected] have been recorded on New Zealand native other Alteranthera species, and although Fax: +64 9 849 7093 plants, and only one potential serious damage to A. sessilis has not been observed impact has emerged so far. Three agents it remains a possibility. ❑ Are we on your mailing list? BiocontrolNews and Information is always pleased to receive news of research, conferences, new products or patents, changes in personnel, collaborative agreements or any other information of interest to other readers. If your organization sends out press releases or newsletters, please let us have a copy. In addition, the editors welcome proposals for review topics. 78N BiocontrolNews and Information 2003 Vol. 24 No. 4 Birch Sawflies Seen Off of attack in June-July from the univoltine initiated a successful communication cam- species, P. thomsoni, together with a second paign to dissuade the public from applying An Old World birch leaf-mining sawfly, generation of F. pusilla. Although a third insecticide on the grounds that it would dis- accidentally introduced to Alaska, is the species, the late birch leaf edge miner, Het- rupt biocontrol. Edmonton’s success has subject of a classical biocontrol programme erarthrus nemoratus, is present in Alberta, been noted by authorities in areas still trou- employing a natural enemy from neigh- it is very rare and causes little damage. bled by the sawflies, and NoFC is currently bouring Canada. This programme, capital- working with Northwest Territories and izing on the successful biocontrol of birch In the early 1990s, scientists were surprised Alaska on control programmes. leaf-mining sawflies in Canada, will also to notice a dramatic drop in sawfly damage mark the first instance of a natural enemy to birch trees in Edmonton, Alberta and Alaska remained free of alien birch leaf- being released as part of a major pro- even more surprised to find that a native mining sawfly pests until about the 1990s gramme against an invasive alien species in parasitoid was responsible by reducing when the first damage to birch was noticed. the wild in Alaska. populations of one of the species, P. thom- Subsequently, birches in Anchorage began soni. The parasitoid responsible, Lath- to sustain the greatest damage, which now Birches have a worldwide distribution in rolestes luteolator, appears to be native to extends over more than 14,000 ha in and the North Temperate Zone, and some 12 both Old and New Worlds. It had not previ- around the city. The pest has also spread as native species are found in North America. ously been recorded from P. thomsoni, far north as Fairbanks, east to Glennallen They are an important component of the although it attacks sawflies in the genus and south to Haines and Skagway. It was boreal forests that extend north to the Caliroa in Europe and North America, and not until 2002 that the major culprit was treeline. A number of native and exotic spe- Profenusa alumna on northern red oaks positively identified as P. thomsoni; how- cies are highly valued and widely planted (Quercus rubra) in eastern Canada and the ever, F. pusilla and H.