Cripps et al. New Zealand Plant Protection 73 (2020) 24–32 https://doi.org/10.30843/nzpp.2020.73.11718 Spill-over attack by the gall fly, Urophora stylata, on congeners of its target weed, Cirsium vulgare Michael Cripps1*, Jovesa Navukula2, Benjamin Kaltenbach1, Chikako van Koten1, Seona Casonato2, Hugh Gourlay3 1AgResearch, Lincoln Science Centre, Private Bag 4749, Lincoln, New Zealand 2Lincoln University, Pest Management and Conservation Department, PO Box 85084, Lincoln, New Zealand 3Manaaki Whenua Landcare Research, PO Box 69040, Lincoln, New Zealand *Corresponding author: [email protected] (Original submission received 19 June 2020; accepted in revised form 19 September 2020) Abstract Urophora stylata, was released in New Zealand in 1998 as a biocontrol agent against the thistle weed, Cirsium vulgare (Scotch thistle). In the summer of 2018, a survey was conducted to assess the The gall fly, sheep and/or beef production, where C. vulgare was present, was surveyed to quantify the attack intensity (gallfield hostsize relativerange of to the seedhead biocontrol size) agent on C. in vulgare New Zealand., and the A attackrandom rate selection on other of thistle18 pasture weeds populations within the under same population. At each location, seedheads were collected from C. vulgare and all other thistle species (Cardueae) present, which included Cirsium arvense (Californian thistle), Cirsium palustre (marsh thistle), Carduus nutans (nodding thistle), and an Arctium species (burdock). In addition to Cirsium vulgare C. arvense (six locations) and C. palustre (one location). The probability of attack on C. arvense was positively correlated with attack intensity on C. vulgare, suggesting that attack on C. arvense, the gall fly was recorded on occurring where seedheads of C. vulgare are in limited supply. is a ‘spill-over effect’ Keywords host range, non-target attack INTRODUCTION host specificity, Post-release monitoring is an important part of classical biocontrol programmes. Monitoring is essential to determine ofspecificity plant species studies on are which carried the biocontrolout to define agent the ‘fundamentalcan complete host range’ of the species under consideration (i.e. the group biocontrol programmes (Morin et al. 2009; Blossey 2016). testing, the fundamental host range of a potential biocontrol Itthe is effectivenessalso important of to biocontrol assess the agents, safety ofand biocontrol the benefits agents of development) (Schaffner 2001). After host specificity to non-target plants including native species and crop plants being restricted to a genus, or tribe of plants. of economic importance. Systematic surveys designed to insect is often phylogenetically defined, such as development assess attack on non-target plants are uncommon globally fundamental host range are seldom attacked (Cristofaro et (Hinz et al. 2019) but New Zealand stands out as an al. Under2013). Biocontrolfield conditions, agents are many unlikely plants to encounter within somethe exception, with several thorough assessments of non-target potential host plants due to different phenologies (temporal attack on valued native and economic plant species (Fowler separation), or different environmental niches (spatial et al. 2000; Paynter et al. 2004; Waipara et al. 2009). Results separation), compared to the primary host (Wapshere of non-target attack assessments in New Zealand have 1989). While temporal and spatial separation of some highlighted some cases of unanticipated attack on native species, and exotic species of minor economic importance (Paynter et al. 2008b; Withers et al. 2008). However, none thepotential expression host plants of preference can limit hierarchies, the realised where field host herbivorous range of of these cases of non-target attack were considered to have insectsbiocontrol preferentially agents, often utilise the field host host plants range that is maximise limited due their to Furthermore, these cases of non-target attack would have beensignificant predictable population with impacts more thorough on the non-target pre-release species. host- fitness (or actively avoid suboptimal hosts that result in range testing, since they occurred on closely related plant reduced fitness) (Sheppard et al. 2005). Where utilisation species (Pemberton 2000; Paynter et al. 2008a; Paynter et of less preferred host plants occurs under field conditions, al. 2015). ait resultis often of depletiona result of of ‘spill-over the primary attack’. host plant(s)Spill-over (Hinz attack et al.is In addition to assessing the safety of introduced 2020).defined In as other transient words, utilisation it is attack of less that preferred is unlikely species, to occur as biocontrol agents, post-release monitoring can be used to when the primary resource is plentiful. Urophora stylata (F.), a biocontrol etassess al. 2012;if the pre-releaseSchaffner ethost al. specificity 2020). Prior testing to matchesintroducing the agentThe introduced purpose of tothe New present Zealand study for was control to assess of the the thistle field arealised biocontrol ‘field hostagent, range’ laboratory-based in the introduced pre-release region (Fowler host weed,host range Cirsium of the vulgare gall fly, (Savi) Tenore (Scotch thistle). The ©2020 New Zealand Plant Protection Society (Inc.) www.nzpps.org Refer to http://www.nzpps.org/terms_of_use.html New Zealand Plant Protection 73 (2020) 25 reported to attack the seedheads of C. vulgare, its primary threatened, or nationally vulnerable species (de Lange fly is native to the Palearctic region, where it is commonly 2020).food plants As expected, (known no to attack Māori was as detected pūhā), onand native classified Sonchus as the bracts of developing seedheads, or sometimes inserts species (Table 1). Of three thistle weeds (Cirsium arvense (L.) host plant (Redfern 1968). The fly oviposits eggs between Scop., Carduus nutans L. and Silybum marianum (L.) Gaertn.) tested in addition to Cirsium vulgare, attack was detected developingeggs in the tipsseeds. of florets.Multiple The larvae larvae often hatch develop as second in a instars single only on Cirsium arvense (Table 1). In Australia, there are two seedhead,and move creatingdown the a hardened floret tubes, multi-chambered and then feed gall onwithin the Cardueae species recorded as native, Rhaponticum australe the seedhead. By mid-summer the next generation of adults (Gaudich.) Soskov and Saussurea lyrata (Bunge) Franch., emerge, completing a partial second generation per year although their native status is doubtful (Hidalgo et al. 2006; (Redfern 1968). Susanna & Garcia-Jacas 2007). No attack was reported on Surveys in central Europe have recorded U. stylata from these Cardueae species, or an additional 10 species from Cirsium and Carduus species (Zwölfer 1965; Korneyev & nine different Asteraceae tribes (Harris et al, unpublished)2. White 2000). However, different biotypes of U. stylata have In New Zealand, U. stylata was released in 1998 via a been reported from its native range, including a biotype shipment from Australia, followed by a supplemental release from the eastern Mediterranean region considered to be sourced from the USA in 1999 (Winston et al. 2014). The oligophagous, developing on a wide range of Cardueae host original sources of U. stylata released in Australia and North plants (Knio et al. 2002). native range in France, Germany, and Switzerland. Thus, it is likelyAmerica that were the fundamentalcollected from host populations range of U. of stylata the fly releasedfrom its speciesPre-release (including host 10 specificity non-Cardueae testing species). carried Oviposition out in the by in New Zealand corresponds with host records from surveys native range, evaluated oviposition of the fly on 44 testArctium plant Carduus tomentosum Miller, Carduus acanthoides L., and Onopordum and Cirsium species. Urophora stylata is the only species acanthiumthe fly was L.),reported but development on three Cardueae was completed test plants only ( on the deliberatelyin central Europe, released and for that biocontrol the fly is ofrestricted Cirsium tovulgare in primary host plant, C. vulgare (Zwölfer, unpublished)1. The New Zealand, although the weevils, Rhinocyllus conicus (F.) and Trichosirocalus horridus (Panzer), have been reported to (Delémont) exclusively using the Swiss Jura biotype of the attack the weed (Zwölfer & Harris 1984; Groenteman et al. host specificity testing was carried out at CABI SwitzerlandC. vulgare 2008). Rhinocyllus conicus was released along with Urophora (Zwölfer, unpublished)1. solstitialis (L.) to control thistle weeds in New Zealand. Both fly, which was considered to be highly specific to are seedhead-feeding biocontrol agents but diverge in their out in New Zealand and Australia, primarily to ensure safety host range. Urophora solstitialis has not been recorded on to Additionalnative plant pre-release species. In host New specificity Zealand, testing the closest was carried native any host plants other than Carduus nutans in New Zealand, relatives to Cardueae belong to the Cichorieae tribe and although closely related Carduus species are within the its include Sonchus species. While Sonchus species are very host range (Woodburn 1993; Korneyev & White 2000). distantly related to Cirsium vulgare (Funk et al. 2009), and Rhinocyllus conicus is known to have an oligophagous host unlikely to be attacked, it was considered prudent to test range attacking many plants in the subtribe Carduinae (true these species since some