Six species of overseas low starthistle replaces desirable veg- Yellow starthistle, one of California’s more pernicious weeds, infests this field in have been approved for release as etation, in both natural and managed Nevada County. biological control agents of yellow settings. In natural settings, it dimin- starthistle, California’s most per- ishes recreational values, and dense vasive weed. Previously, four stands of this invasive exotic can re- Mayfield 1985). Yellow starthistle’s biocontrol insects were known to duce biodiversity and help carry wild logarithmic range expansion contin- be established in California; we fires. Within California’s agricultural ues. A 1997 survey by California De- now confirm the establishment of community, yellow starthistle most se- partment of Food and Agriculture the peacock , as well as the ac- verely impacts ranchers. While young (CDFA) found this weed in 42% (n = 1,935) of California’s 4,638 townships cidentally introduced false pea- yellow starthistle shoots can be grazed - each 6 by 6 square miles - and in cock fly. Remarkably, the false by cattle, the sharp spines of older plants deter feeding, thereby greatly 22% (1,019 townships) the infestations peacock fly is significantly more reducing the forage value of hundreds are reported as ”high” (Pitcairn et al. widespread and more effective of thousands of acres for most of the 1998a). “High“ abundance was de- against yellow starthistle than the year. Yellow starthistle can be toxic to fined as being, at a minimum, several peacock fly - or any other horses that feed on it, causing a fatal miles of dense roadside infestation. biocontrol to date. How- neurological disorder called ”nigro- Widespread exotic pests, such as ever, since the false peacock fly is pallidal encephalomalacia” (Cordy yellow starthistle, are obvious targets not an approved agent, we will 1978). Starthistle removal also in- for classical biological control, in await completion of our ongoing creases herbicide and labor costs for which natural enemies of a pest are field and laboratory assessments other agricultural enterprises, such as imported from its native territory. Po- of this fly’s safety to crops and vineyards and orchards. tential biocontrol agents for weeds are native plants before recommend- Yellow starthistle is native to the carefully selected and screened in ex- ing use of this promising fly as a eastern Mediterranean region of tensive tests to ensure their safety. biological control agent. Eurasia, and was introduced into Cali- Prior to release, an array of federal fornia more than 150 years ago and state agencies review the host While California’s diverse habitats (Maddox and Mayfield 1985). It is now range tests and other information support hundreds of invasive weed the state’s most widespread weed. The about the proposed agent; formal ap- species, probably none is more wide- infested area increased from an esti- proval for release must be granted by spread nor pernicious than yellow mated 1.2 million acres in 1958 to 7.9 the U.S.Department of Agriculture’s starthistle ( solstitialis L.) Yel- million acres in 1985 (Maddox and and Plant Health Inspection Service (USDA-APHIS), as well as by and at one of the Idaho sites in 1995 they were the result of natural migra- the state where the release will take (Turner et al. 1996). Establishment of tion from the long-established popula- place. this fly was not observed at any of the tions at the Merlin release site (107 This classical biological control ap- six California sites (Turner et al. 1996). miles away), rather than from our re- proach has been used against more At the three sites (in Idaho, Oregon leases earlier that year in Shasta and than 100 weed species worldwide. and Washington) where the peacock Siskiyou counties. California has benefited greatly from fly did establish, bachelor’s button, the successful biological control of in- , was widespread. Discovery of a new fly vasive plants such as Klamath weed Bachelor’s button is another exotic an- The ease with which these (Hypericum perforaturn), tansy ragwort nual, closely related to yellow from Oregon established at all sites, (Seneciojacobaea), and puncture vine starthistle, that is invasive in the Pa- including those that lacked bachelor’s (Tribulus terrestris) (Julien 1992). Over- cific Northwest. It was theorized that button, along with their rapid dis- seas surveys to locate potential the early-blooming bachelor’s button persal from the release sites, was un- biocontrol agents for yellow starthistle flowers were acting as an alternate expected - especially in light of the began in Europe 40 years ago, and, to host until yellow starthistle blossomed complete failure of the earlier releases date, six insect species have been ap- some weeks later (Turner et al. 1996). in California. In 1996, we submitted proved and released in the United Buoyed by these successful estab- specimens of the flies recovered from States for control of this invasive weed lishments, we renewed the coloniza- the field in California to two experts (table 1).All six species attack the tion effort for the peacock fly in Cali- on fly at the CDFA Plant flowers or seeds of yellow starthistle. fornia, with releases at 7 sites in 7 Pest Diagnostics Center. Neither counties in 1995, and 15 more releases thought that these California flies fit The peacock fly in 12 counties in 1996. Sites containing the published description of Ckaeto- The third seed head fly to be ap- both bachelor’s button and yellow rellia australis. The second taxonomist, proved for release was the peacock fly, starthistle were given the highest pri- Eric Fisher, identified them as Ck. Ckaetorellia australis Hering, whose lar- ority. In California, bachelor’s button succinea (Costa), a similar species from vae feed inside the seed head of yel- has naturalized primarily in higher el- Europe and Asia. We then curtailed all low starthistle, destroying most of the evation sites in the northern part of the further releases of Ckaetorellia flies in developing seeds. The mature larvae state. Second priority was given to California. After assembling Chaeto- overwinter in the old heads, with the sites with early-blooming yellow rellia specimens recovered from field adults emerging in the spring. Females starthistle. All flies released (except sites in California, Oregon and Wash- oviposit on maturing buds. Releases of one sample from bachelor’s button) ington, we shipped these, along with peacock flies, reared from yellow were those that emerged from yellow voucher Ckaetorellia specimens from starthistle heads shipped from starthistle heads collected at the Mer- those originally imported and tested at to the USDA Agricultural Research lin, Ore., site. the ARS quarantine in Albany to Ian Service (ARS) quarantine facility in Al- During our surveys at the end of White at the British Museum of Natu- bany, Calif., began in 1988. By 1994, 1995, populations of flies ral History, London, for confirmation. peacock fly had been released at 14 were found at multiple locations in Dr. White is an authority for the sites in California, Idaho, Oregon and Humboldt and Trinity counties in Ckaetorellia and had recently published Washington, but establishment was Northern California. The fly popula- a revision of this genus (White and confirmed only at two sites that year, tions in these counties were so large Marquardt 1989). He confirmed that one each in Oregon and Washington, and widespread that we surmised that the majority of Ckaetorellia specimens from California and Merlin were, in fact, Ck. succinea. White and Marquardt (1989) place the nine known species of Ckaetorellia into two groups - Ck. succinea belonging to one group, and Ck. australis to another. Ck. succinea (and the other two species in its group) has an extra ”spot” on each side of its thorax that is lacking in Ck. australis and the other five species in its group. As no other members of the first group have been recorded in North America, we use this extra “spot” as an easy way to distinguish it from all other Ckaetorellia flies found here. Since Ck. australis is widely known in California as the peacock fly, The quarantine records showed that 207 and 268 sites, respectively, in Cali- there had been 17 shipments of fornia. We revisited not only the origi- Chaetorellia-infested heads from Greece nal Ckaetorellia release sites, but also during 1986-1994. Flies emerging from searched yellow starthistle along prob- nine of these shipments of yellow able flyways, such as major river sys- starthistle heads were eventually re- tems or highways. At each site, we leased in four states. There were 43 systematically swept the yellow voucher specimens from these release starthistle heads with an insect net shipments, and our cooperators in (usually 20 sweeps), and, for at least other states also had retained a similar one site in each county, collected number. We identified the voucher heads that had recently finished flow- specimens, and had representative ering. The latter were held in emer- specimens confirmed by Ckaetorellia gence containers in the laboratory un- expert White. This allows us to put til the flies stopped emerging from C 5: forth the following probable scenario them, about 1 year later. Over the last for the introduction of the false pea- 4 years, we collected more than 700 cock fly. Ckaetorellia australis was the samples from over 450 California sites only fly present among the voucher (fig. 1).We recovered ”true” peacock Above, the false peacock fly, . Below, the false peacock fly specimens from all these shipments, fly at four release sites, and at 20 addi- (left), and the “true” peacock fly, Ch. aus- except for a 1991 shipment of yellow tional, scattered sites in seven corn- fralis (right). Note the extra spot on the up- starthistle heads collected in Oreo- ties, all of which have bachelor’s but- per left side of the thorax of the false pea- kastro, Greece. That shipment con- ton in addition to yellow starthistle. cock fly. tained both the true peacock fly, Ck. On the other hand, the false peacock australis, and the false peacock fly, Ck. fly is well established at 415 of our succinea. The presence of the false pea- sample sites, and is spreading rapidly. cock fly contaminating this shipment We recovered it from 9 counties in was not detected at the Albany quar- 1995, from 21 counties in 1996,34 in antine, and the flies that emerged were 1997, and from 44 counties during sent and released at the Merlin, Ore., 1998. We anticipate that the false pea- site. Both flies became established at cock fly will continue to expand its this site, and a variety of agencies geographic range. We will continue from several states subsequently used to monitor how completely its range flies collected from this Merlin site in will overlap with its host, yellow their attempts to control yellow starthistle. starthistle in their own regions. It is one of the responsibilities of a quaran- Impact on yellow starthistle tine to assure that only the approved How much damage is the false we will refer to Ck. succinea as the false agent is released. However, when peacock fly inflicting on yellow peacock fly. hundreds of live insects must be exam- starthistle? CDFA, ARS, Oregon De- ined, as was the case in 1991, without partment of Agriculture, and univer- Origin of the false peacock fly killing or damaging the insects slated sity scientists in several states are as- Like all the yellow starthistle for release, a cryptic species might be sessing the long-term effects the false biocontrol agents, the shipments of overlooked. The presence of Ck. peacock fly will have on established Ckaetorellia had passed through the succinea at the Oreokastro, Greece, site yellow starthistle infestations. CDFA USDA-ARS quarantine containment had not been observed, and was not recently found that at one site in facility in Albany. There, adults were even suspected. Amador County, 36% of the yellow allowed to emerge from the heads, starthistle heads had been attacked by parasites and other insects were re- Distribution of Chaetorellia f Iies false peacock fly, and the heads with moved, then the Ckaetorellia were col- Once we were aware of the pres- flies had 78% less seed (Pitcairn et al. lected and transferred to the release ence of a second Ckaetorellia fly in Cali- 1998b). During 1998, assisted by col- site. We collated all the shipment fornia, we started extensive surveys to leagues in other states, we conducted records for Chaetorellia flies at the Al- determine the distribution of both spe- similar studies at more than 50 sites bany quarantine, and assembled all cies. During 1996, we surveyed 207 in California, Oregon and Washing- the voucher specimens that had been sites in California, and recovered ton. These samples are still being retained at the quarantine from these Ckaetorellia flies from all 15 of the processed, but we are recording shipments. We also examined voucher 1995-96 release sites, and also 6 other similar seed reductions, although at Specimens retained by our cooperators sites where it had not been released. some sites more than 50% of the heads in Idaho, Oregon and Washington. During 1997 and 1998, we surveyed have been attacked. By contrast, ento- mologist Gary Piper at Washington cartkami, a State University has studied the dam- minor pest of age caused by Ck. australis, the true in peacock fly, for 6 years at a site in the Middle Colfax, Wash., where Ck. succinea still East. They does not occur. He has never re- therefore rec- corded more than 5% of the yellow ommended starthistle heads as being attacked by against import- the peacock fly (G. Piper, personal ing it into the communication) United States, es- Since yellow starthistle is an annual pecially since, at that produces seeds copiously, creat- that time, there were ing a persistent seed bank, it will take many other potential many years to determine the impact agents to consider. Ck. false peacock flies are having on yel- cartkami is not known to low starthistle infestations in Califor- be present in the United nia. However, our preliminary obser- States. Prudence, how- vations of the damage this fly is ever, dictates that we evalu- causing to individual heads and plants ate the host range of false pea- of yellow starthistle indicate that Ck. cock fly, especially in respect to succinea should contribute to the even- potential damage it may cause to tual control of yellow starthistle at safflower or native Cirsium thistles. sites where it establishes. Once this fly We and our colleagues in other states arrives at a site, its populations can are currently evaluating the host range quickly increase, and a large percent- of false peacock fly. We expect to com- age of yellow starthistle heads are at- plete our evaluations during 1999. Pre- tacked by the larvae. A single , liminary analyses of the first 2 years of while feeding inside the maturing seed field observations and laboratory tests head, will destroy most, but usually indicate, however, that damage to not all, of the developing seeds. In ad- commercial safflower varieties in Cali- dition, it has multiple generations per fornia by this accidentally introduced year, and its attack is sustained false peacock fly is improbable, but if throughout yellow starthistle’s it occurs, damage will be, at the most, lengthy flowering period in California. minuscule. Frequently, at many sites in California, Pitcairn MJ, O’Connell RA, Gendron JM. but less often in northern states, yel- 1. Balciunas is Research Entomologist and 1998a. Yellow starthistle: Survey of statewide low starthistle continues to flower un- Yellow Startkistle Biocontrol Project distribution. In: Woods DM (ed.). Bio/ogica/ til late summer or early autumn. This & Control Program Annual Summary, 1997. Sac- Leader at the Exotic lnuasiue Weed Re- ramento: California Department of Food and is very important, as none of the other search Unit, USDA-ARS Western Re- Agriculture, Division of Plant Industry. p 64-5. approved and established yellow gional Research Center, Albany; and a Pitcairn MJ, Joley DB, Woods DM, Popescu starthistle biocontrol agents commonly member of the UC Berkeley Center for V. 1998b. Impact of Chaetorellia succinea on yellow starthistle, Centaurea solsfitialis,in Cali- attack these late-blooming yellow Biological Control; B. Villegas is Associ- fornia. In: Woods DM (ed.). Biological Control starthistle flowers. ate Environmental Research Scientist, Program Annual Summary, 1997.Sacramento: California Department of Food and Ag- California Department of Food and Agriculture, Damage to other plants? riculture, Biological Control Program, Division of Plant Industry. p 56-7. Sobhian R, Zwolfer H. 1985. Phytophagous The false peacock fly is not an ”ap- Sacramento, insect species associated with flower heads of proved” biological control agent. It yellow starthistle ( L.). has not been fully tested to determine References Zeitschrift fuer angewandte Entomologie if it will attack plants other than yel- 99:301-21. Cordy DR. 1978. Centaurea species and Turner CE, Piper GL, Coombs EM. 1996. low starthistle. Brief host-range tests equine nigropallidal encephalomalacia. In: (Diptera: ) for were conducted in Europe, when this Keeler RF, Van Kampen KR, James LF the biological control of yellow starthistle, Cen- fly was still known as a ”semi-species” (eds.). Effects of Poisonous Plants on Live- taurea solsfitialis (Compositae), in the western stock. New York, Academic Press. p 327-36. USA: establishment and seed destruction. Bull of Ckaetorellia cartkami. The investiga- Julien MH (ed.). 1992. Biological Control Entomol Res 86: 177-82. tors (Sobhian and Zwolfer 1985) found of Weeds: A World Catalogue of Agents and White IM, Marquardt K. 1989. A revision of that it would not attack safflower, Their Target Weeds, 3rd ed. Wallingford, UK: the genus Chaetorellia Hendel (Diptera: Carthamus tinctorius, a close relative of CAB International. 186 p. Tephritidae) including a new species associ- Maddox DM, Mayfield A. 1985. Yellow ated with spotted knapweed, Centaurea yellow starthistle. However, they were starthistle infestations are on the increase. maculosa Lam. (Asteraceae). Bull Entomol Res able to hybridize Ck. succinea with Ck. Cal Agr 39(11-12):lO-2. 79:453-87.