BioControl DOI 10.1007/s10526-014-9578-7 Biology and host specificity of Rhinusa pilosa, a recommended biological control agent of Linaria vulgaris André Gassmann • Rosemarie De Clerck-Floate • Sharlene Sing • Ivo Toševski • Milana Mitrović • Olivier Krstić Received: 20 September 2013 / Accepted: 28 March 2014 Ó International Organization for Biological Control (IOBC) 2014 Abstract Linaria vulgaris Mill. (Plantaginaceae), common or yellow toadflax, is a Eurasian short-lived perennial forb invasive throughout temperate North America. Rhinusa pilosa (Gyllenhal) (Coleoptera, Curculionidae) is a univoltine shoot-galling weevil found exclusively on L. vulgaris in Europe. Under no-choice test conditions, 13 non-native Linaria species exposed to R. pilosa were accepted for oviposition and most were found to be suitable, to varying degrees, for gall and larval development. Adult feeding and survival was minimal on native North American species in the plant tribe Antirrhineae which includes the target plant. In nochoice tests with 63 native North American species and 24 other non-target species outside Linaria, oviposition was limited to four native North American species. Only three larvae developed to the adult stage on Sairocarpus virga (A. Gray) D.A. Sutton, with no negative impact on plant growth. Risks to native flora from the release of R. pilosa are therefore expected to be minimal. The Technical Advisory Group for the Biological Control of Weeds (TAG—BCW) has recommended release of R. pilosa in September 2013. Keywords Common toadflax
Yellow toadflax
Curculionidae
Host range tests
Pre-release studies
Biological control of weeds Introduction Handling Editor: John Scott. Electronic supplementary material The online version of this article (doi:10.1007/s10526-014-9578-7) contains supplementary material, which is available to authorized users. A. Gassmann (&)
I. Toševski CABI, Rue des Grillons 1, 2800 Delémont, Switzerland e-mail: a.gassmann@cabi.org I. Toševski e-mail: tosevski_ivo@yahoo.com R. De Clerck-Floate Agriculture and Agri-Food Canada, Lethbridge Research Centre, 5403 1st Avenue South, Lethbridge, AB T1J 4B1, Canada e-mail: rosemarie.declerck-floate@agr.gc.ca Linaria vulgaris Mill. (Plantaginaceae), common or yellow toadflax, a short-lived perennial forb native to most of Europe and northern Asia (Chater et al. 1972; S. Sing USDA Forest Service – Rocky Mountain Research Station, 1648 South 7th Avenue – MSU Campus, Bozeman, MT 59717-2780, USA e-mail: ssing@fs.fed.us I. Toševski
M. Mitrović
O. Krstić Department of Plant Pests, Institute for Plant Protection and Environment, Banatska, Zemun, Serbia e-mail: milanadesancic@yahoo.co.uk O. Krstić e-mail: oliverk13@yahoo.com 123 A. Gassmann et al. Sutton 1988), rapidly and comprehensively colonized temperate North America after its surmised seventeenth century eastern USA introduction (Mack 2003). This weed now occurs in all mainland states of the USA and most provinces and territories of Canada (USDA, NRCS 2013). Several exotic toadflax-feeding insect species can be found in North America today. Their presence is due both to unintentional and intentional introductions from Europe, along with their hosts L. vulgaris or the congeneric weedy species, Linaria dalmatica (L.) Miller and L. genistifolia (L.) Mill. Three adventitiously introduced seed-feeding beetles, Brachypterolus pulicarius (Linnaeus, 1758) (Coleoptera: Nitidulidae), Rhinusa antirrhini (Paykull, 1800) (formerly Gymnetron antirrhini), and Rhinusa neta (Germar, 1821) (formerly Gymnetron netum) (Coleoptera: Curculionidae), are now ubiquitous on North American L. vulgaris, either self-dispersed or intentionally spread as part of the North American toadflax biological control program (De Clerck-Floate and McClay 2013; De Clerck-Floate and Turner 2013; Sing et al. 2005; Wilson et al. 2005). Five additional insect species have been approved and intentionally released in North America (De Clerck-Floate and Harris 2002; McClay and De Clerck-Floate 2002; Harris 1963). These are: the defoliating moth Calophasia lunula (Hufnagel, 1766) (Lepidoptera: Noctuidae), two congeneric root-boring moths, Eteobalea intermediella (Riedl, 1966) and E. serratella (Treitschke, 1833) (Lepidoptera: Cosmopterigidae), the root-galling weevil Rhinusa linariae (Panzer, 1792) (formerly Gymnetron linariae) (Coleoptera: Curculionidae), and the stem-mining weevil, Mecinus janthinus Germar (Coleoptera: Curculionidae) (De Clerck-Floate and McClay 2013; De ClerckFloate and Turner 2013). Despite the use of the mentioned insect species against invasive toadflaxes beginning in the 1960s (McClay and De Clerck-Floate 2002; Sing et al. 2005; Wilson et al. 2005; De Clerck-Floate and McClay 2013; De Clerck-Floate and Turner 2013), the reported efficacy of the majority of these on L. vulgaris has been minimal, with only M. janthinus recently showing some promise since its first releases in the 1990s in both the USA and Canada (Sing et al. 2005; De Clerck-Floate and McClay 2013). Since 2000, several other European insects have been investigated for potential use against L. vulgaris, and the first of these, 123 the shoot-galling weevil, Rhinusa pilosa (Gyllenhal) (Coleoptera: Curculionidae), was petitioned for release in the USA and Canada in 2012 (De ClerckFloate and McClay 2013). It was chosen as a candidate for testing on the basis of both native range host records suggesting that it is highly host-specific, and on field observations indicating that gall formation by the weevil could severely stunt shoots and prevent flowering of L. vulgaris (Toševski, unpublished data). The current paper presents the results of a combination of field records, laboratory, greenhouse and garden experiments used to complete an in depth study of the biology, ecology and host range of R. pilosa in order to gauge its safety and efficacy before seeking regulatory approval for its field release on L. vulgaris in North America. R. pilosa was recommended for release in September 2013. Target plant systematics and biological information The genus Linaria Mill. was traditionally placed in the Scrophulariaceae (figwort) family (Sutton 1988). Revisions based on molecular phylogenetic analyses indicated that Linaria would be more appropriately included within the expanded Plantaginaceae (plantain) family (Albach et al. 2005; Ghebrehiwet et al. 2000; Olmstead et al. 2001). L. vulgaris belongs to the Antirrhineae tribe, one of eleven currently circumscribed Plantaginaceae tribes (Estes and Small 2008; Ghebrehiwet et al. 2000; Tank et al. 2006). The Antirrhineae tribe consists of six well-supported clades (Cymbalaria, Anarrhinum, Chaenorhinum, Antirrhinum, Galvezia, and Linaria) (Fernandez-Mazuecos et al. 2013; Vargas et al. 2004). In North America, members of the monophyletic Linaria clade are limited to fourteen established non-native species, including L. vulgaris, L. dalmatica and L. genistifolia (L.) (USDA, NRCS 2013). Several of the non-native Linaria spp. that currently occur in North America have found limited use in ethnobotany and/or gardening (e.g., L. purpurea (L.) Mill.; see URL Missouri Botanical Garden 2013; also see Saner et al. 1995, and Vujnovic and Wein 1997 for beneficial economic uses of L. vulgaris and L. dalmatica, respectively). With respect to their ecological value in North America, it is suspected to be minimal, although there are reports of Linaria pollination by native bumble bees (De ClerckFloate and Richards 1997). Three native North Biology and host specificity of Rhinusa pilosa American taxa conventionally placed within Linaria were reclassified in 1988, remaining in the tribe Antirrhineae but moved to a new genus, Nuttallanthus D.A. Sutton (Sutton 1988). Fernández-Mazuecos et al. (2013) suggests however that Nuttallanthus may, pending further investigation, be appropriately circumscribed as a section of Linaria. The tribe Antirrhineae includes 11 native North American genera and 33 native North American species (Vargas et al. 2004; USDA, NRCS 2013). Linaria vulgaris is an erect perennial forb with plants consisting of single to numerous shoots attaining heights of 25–120 cm (Sutton 1988). Dense, persistent patches form primarily through vegetative reproduction from adventitious shoots that arise from tap and lateral roots (Bakshi and Coupland 1960). Sexual reproduction of this obligate out-crossing species is generally less successful (Nadeau and King 1991). Molecular diagnostic techniques have confirmed the occurrence of hybridization between yellow and Dalmatian toadflax from samples field collected at sites in Montana, Idaho and Colorado, USA (Boswell 2013; Turner 2012; Ward et al. 2009). Biological control agent taxonomic information and biology The candidate agent was known as Gymnetron hispidum Brullé, 1832 at the inception of this project. The revised identity of this taxon, presently placed in the genus Rhinusa (Caldara 2001), was taxonomically uncontroversial (Caldara et al. 2008 and references therein) because of its clear differentiation from all other species within the genus, i.e. by the ground color of the body, black with soft dark grey pubescence, densely covered with bristled blackish hairs. However, it was subsequently determined through a study of the type specimens that G. hispidum is a younger synonym of Rhinusa tetra (Fabricius, 1792), while the two closely related species R. pilosa and R. brondelii (Brisout, 1862) were ascribed full species status (Caldara et al. 2008; Caldara et al. 2010). R. pilosa is a univoltine shoot-galling species that overwinters as an adult. The species was described as a gallinducer associated with L. vulgaris, while R. brondelii (also a galler) was primarily associated with L. purpurea in southern Italy, L. genistifolia in Serbia, L. reflexa (L.) Desf. in Tunisia and L. gharbensis Bat. & Pit. in Morocco (Caldara et al. 2010). A detailed description of R. pilosa and morphological differences between R. pilosa and R. brondelii are given by Caldara et al. (2008). Initiated by Lorenza Legarreta and Brent Emerson, University of East Anglia, Norwich, UK (Caldara et al. 2008), a recent population genetic study of weevils from the R. pilosa–R. brondelii complex yielded strong molecular evidence of the highly specific host association of these weevils with their confirmed field host plants (Toševski unpublished data). All known species in the genus Rhinusa utilize Plantaginaceae (Linaria, Antirrhinum, Mesopates, Chaenorhinum and Kickxia) and Scrophulariaceae (Verbascum, Scrophularia) host plants, while Gymnetron from the Palaearctic region are associated only with plants from the genus Veronica (Scrophulariaceae). Species-specific host records from the literature are questionable since they do not adequately distinguish R. pilosa from R. brondelii. According to Hoffmann (1958), R. pilosa induces galls on several Linaria species, but in central and southeastern Europe it mainly occurs on L. vulgaris. The reported alternative hosts are Chaenorrhinum minus (L.) Lange, L. repens (L.) Mill., L. purpurea, L. simplex (Willd.) DC., L. reflexa (Hoffman 1958; Lohse and Tischler 1983) and L. gharbensis in North Africa (Mimeur 1949). A detailed account of the histology of gall induction and gall development by the weevil has been provided by Barnewall (2011) and Barnewall and De Clerck-Floate (2012) using the same insect population. Materials and methods Surveys and agent biology The specimens in the current study were identified by Roberto Caldara, Milan, Italy. Spring surveys were conducted mainly in Serbia where R. pilosa was found most commonly, as well as in Romania and Hungary between 2003 and 2011. Galled plants were field collected with surrounding soil to keep roots intact, then individually transplanted into plastic pots. Pots were enclosed with a plastic cylinder (10 9 35 cm) fitted with ventilated (gauze) lids to maintain moisture and retain insects emerging from plants. Newly emerged adult weevils were overwintered in mesh field cages (210 9 210 9 190 cm), in plexiglass- 123 A. Gassmann et al. walled outdoor cages (70 9 70 9 70 cm) or in an outdoor insectarium in mobile framed cages (30 9 30 9 45 cm). The latter contained naturally growing L. vulgaris plants and wood blocks configured with grids of drilled 8 mm-diameter holes to serve as overwintering shelters for the weevils. Adult weevils appearing on the walls of overwintering cages early in spring were collected and used in life history and host specificity studies. Adult survival was found to be on average higher in the outdoor insectarium than in either type of field cage, so from 2009 on, all adults were overwintered in the former. Field collected adults were initially used to establish a rearing colony, and from 2008 on, only adult progeny from this rearing colony were used in host specificity tests. Mass-rearing was conducted in field cages (210 9 210 9 190 cm) containing approximately 100 L. vulgaris plants planted directly into the soil. Galled plants were collected in late May and transplanted into pots (as described above) to secure emerging adults. Embryogenesis of R. pilosa and R. eversmanni (Rosenschöld) was studied at 20 °C. native to North America. The so-called critical plants included species in the genus Nuttallanthus (colloquially known as the North American toadflaxes) and other native North American species from the same tribe (Antirrhineae) as the target weed. Host range No-choice oviposition, gall induction and larval development Rhinusa pilosa originating from L. vulgaris sites near Dobanovci (Srem district), approximately 25 km west from Belgrade (N44° 50.3320 , E20° 09.3960 , elevation 72 m), were used in the host specificity tests. The majority of tests were conducted in the greenhouse or garden at the Institute for Plant Protection and Environment (IPPE), Zemun (Belgrade), Serbia. Additional tests were conducted in the Agriculture and Agri-Food Canada quarantine facility at the Lethbridge Research Centre in Lethbridge, AB, Canada. Plant species tested The Linaria test plant list used in the current investigation applied the conventional phylogenetic approach (Wapshere 1974; Briese 2005). The inclusion of closely related plant species identified as economically (=crop or ornamental), culturally or ecologically important (especially threatened, endangered or species of concern) was prioritized in this research. The final list included 87 Plantaginaceae species/subspecies and 35 species from 14 additional plant families. 66 species, across 34 plant genera, were 123 Host specificity tests No-choice adult feeding and survival Feeding was assessed in early spring with reproductive adults that had overwintered in field cages (‘‘posthibernated adults’’) in the garden at IPPE, Zemun (Belgrade), Serbia. The upper part of vegetative 3–6 months old cut stems of 12 different test plant species were each exposed at room temperature in petri dishes (10 cm diameter) to four adult weevils in unreplicated tests for four weeks. Cut stems were changed every day. Feeding (mm2 of plant surface consumed) and the number of living adults by day of study (i.e., daily census on Day 1 through Day 28) were recorded. Individually potted plants with actively young growing stems were used in all oviposition and larval development no-choice tests to ensure that conditions were optimal for gall initiation and formation. Each potted plant was caged with a capped, ventilated plastic cylinder (10 9 35 cm) and the soil surface was covered with coarse sand to reduce buildup of excessive moisture inside the cage and to make the weevils more visible. Plants were exposed for 1–4 days, depending on plant size and weevil oviposition activity, to one newly emerged, post-hibernated and mated female. Only actively egg-laying females were used in this study, which was confirmed by observing each female weevil ovipositing on a L. vulgaris plant prior to being added to caged test plants. All study plants were kept at 21–24 °C with a 16:8 L:D photoperiod and 60 % humidity to allow normal gall development. Each potted plant was covered with a gauze bag in early June, as dictated by weevil phenology, to ensure retention of emerging F1 adults. All plants with galls were examined under a stereo microscope and the number of oviposition marks (those still visible on the surface of galls) was Biology and host specificity of Rhinusa pilosa recorded. All galls or stems with oviposition marks from which no adults had emerged were dissected and gall contents inventoried for dead or alive adults. Each plant with its separate female R. pilosa was considered a replicate within each test, and replicate number per plant species per test was 5–10. A one-way ANOVA followed by a Tukey’s HSD test was used to compare means for the number of adults per replicate and the number of adults per gall. Sequential no-choice oviposition, gall induction and larval development Single female–male R. pilosa pairs retained within a cylinder cage (10 9 35 cm) were placed for 24 h on each of ten replicate potted Sairocarpus virga (A. Gray) D.A. Sutton plants (i.e., one weevil pair per plant). The same single caged pairs of R. pilosa were then transferred and retained for 24 h on ten replicate potted L. vulgaris plants. Caged weevils were sequentially transferred between S. virga and L. vulgaris where they were retained for a 24 h period. Test plants were exposed only once to a specific R. pilosa pair. The test was continued until all the female weevils had died. The number of oviposition marks visible on induced galls was recorded approximately seven weeks after set-up. Single-choice gall induction and larval development L. vulgaris versus L. genistifolia A variety of singlechoice tests were set-up that differed in the ratio of the number of R. pilosa pairs per plants used, and/or the size of the test arena. Five R. pilosa pairs were released into a field cage (210 9 210 9 190 cm) that contained 40 potted plants each of L. vulgaris and L. genistifolia. The plants within the cage were thereafter regularly inspected and any instances of gall development were recorded. All potted, galled plants were removed from the cage approximately 2.5 months after set-up and transferred to a greenhouse to follow gall and larval development. Each pot was covered with a gauze bag so that adult emergence could be recorded separately. In a second experiment using lower host plant densities, six R. pilosa pairs were released into a field cage (210 9 210 9 190 cm) that contained eight potted plants each of L. vulgaris and L. genistifolia. Gall development and adult emergence were recorded. L. vulgaris versus S. virga Five R. pilosa pairs were released into a field cage (210 9 210 9 190 cm) that contained 12 potted plants each of L. vulgaris and S. virga, and left for approximately 2.5 months. Gall development and adult emergence were recorded. A second experiment evaluated host preference at a finer spatial scale. One R. pilosa female-male pair was exposed to two L. vulgaris and two S. virga plants growing together in a large 15 cm diameter pot covered with a gauze sleeve. Weevil pairs were transferred to new pots (same plant species and densities) every 2–3 days until all the females died. The weevils were always placed on the soil surface to allow them to freely orient toward any plant within the cage. Gall development and adult emergence were recorded. L. vulgaris versus Sairocarpus nuttallianus Five R. pilosa pairs were released into a field cage (210 9 210 9 190 cm) that contained eight potted plants each of L. vulgaris and S. nuttallianus, and left for approximately 2.5 months. Gall development and adult emergence were recorded. Multiple-choice gall induction and larval development Six pairs of R. pilosa were released into each of two field cages (210 9 210 9 190 cm), both set up with six potted S. virga, six potted S. nuttallianus and 24 potted L. vulgaris. Plants were left for approximately two months. The design of the experiment was meant to simulate field conditions in which target hosts are typically more abundant than non-target hosts, thus allowing the females to express their preference for ovipositing on unattacked L. vulgaris shoots versus S. nuttallianus. Gall development and adult emergence were recorded. Gall impact on host plant fitness Sixty field-collected, single shoot L. vulgaris plants were transplanted into plastic pots with the root system of each plant cut to 3 cm in length to standardize for pre-treatment growth. Half of the plants were randomly selected to be individually caged and exposed to one mated female for 24 h. All plants were harvested two months later and plant height, number 123 A. Gassmann et al. 59o N 43 N Fig. 1 Confirmed localities for R. pilosa (Gyllenhal, 1838) according to archived material (white dots) and newly recorded populations (black stars). Populations used for screening test (black squares) of shoots and dry biomass were recorded. A one-way ANOVA was used to compare means. Gall impact on the fitness of S. virga To determine the impact of R. pilosa on S. virga plants, sixty same-sized S. virga plants were paired and one of the pair was randomly assigned to receive one mated and ovipositing R. pilosa female for 24 h, while the other was designated as a control. All plants were harvested three months later and plant height, number of shoots and dry biomass were recorded. A one-way ANOVA was used to compare means. Hungary (one population) and in Romania (one population) (Table S1). All galls recorded on L. genistifolia and L. dalmatica were confirmed as being induced by the closely related species, R. brondelii. R. pilosa has been recorded exclusively on L. vulgaris at three sympatric stands of L. genistifolia and L. vulgaris in Serbia. To date, all confirmed localities for R. pilosa are located between latitudes 43°N and 59oN of the Western Palearctic and L. vulgaris is the only confirmed host plant. Biology Life history Results Geographical distribution In total, 20 Linaria species and subspecies from 290 populations were surveyed for R. pilosa galls in 12 European countries (Fig. 1). R. pilosa stem galls were found on L. vulgaris in Serbia (12 populations), 123 Activation of adults occured early in the growing season (i.e., March–May) coinciding with the spring burst of shoot growth from host perennating tap roots. Post-hibernated adults fed 3–5 days on L. vulgaris shoots and foliage before mating began. Oviposition followed approximately ten days later, loosely timed to occur from April to May, depending on local environmental conditions. Gall development was Biology and host specificity of Rhinusa pilosa Table 1 No-choice adult feeding tests with posthibernated adults of R. pilosa Plant species Feeding (mm2) surface area Day when feeding ceased Adult survival when feeding ceased (%) L. vulgaris EU (control)a 367 [27 100 L. vulgaris NAa 235 [27 100 L. dalmatica NAa 240 [27 100 Species from the same tribe (Antirrhineae) as the target weed species Anarrhinum bellidifolium EUb Maurandella antirrhiniflora c 0 15 0 0 15 0 17 [27 50 EU European population, NA North American population Nuttallanthus canadensisc 6 [27 25 a Sairocarpus nuttallianusc 35 [27 50 S. virgac 41 23 0 Geographic origin of the L. vulgaris and L. dalmatica populations tested b Native European species c Native North American species Neogaerrhinum strictumc Species from tribes other than that of the target weed species Collinsia heterophyllac 0 15 0 Penstemon procerusc 0 14 0 complete approximately 8–10 days after oviposition under laboratory conditions, corresponding with emergence of the first larval instar. R. pilosa had a total of three larval instars that fed and continued development on host tissues within the developed galls. Pupation was also completed within the gall. Enclosed adults remained within the natal gall for 10–15 days, feeding on remnant host tissues before escaping via holes chewed through the gall’s outer surface. Newly emerged adults fed externally on host shoots for approximately ten days. Thereafter, adult weevils reposed in litter or cracks in the soil during the day. Summer aestivation was interrupted by occasional feeding, mainly in the evening and at night. In late autumn, adults fed shortly before entering dormancy within soil or leaf litter. Galls induced by R. pilosa ovipositing into the stems of L. vulgaris were globular, round or oblong green structures that were positioned between the middle and tip of the host stem. On average each gall produced 2.9 adults during host specificity studies. The highest density of R. pilosa adults to emerge from a single gall was 17. Parasitism Parasitism of R. pilosa in Serbia varied year-to-year. Total parasitism of up to 18.5 and 11.7 % was recorded in 2002 and 2003, respectively. Parasitoids reared from galls included the endoparasitoids Pteromalus sequester Walker (Pteromalidae), Eurytoma curculionum Mayr (Eurytomidae), Tetrastichus spp. (Eulophidae), an unidentified species from the subfamily Entedontinae (Eulophidae), the ectoparasitoids Scambus nigricans (Thomson) and Exeristes roborator (Fabricius) (Ichneumonidae), Bracon (Glabrobracon) pineti Thomson and Ascogaster quadridentata Wesmael (Braconidae). E. roborator was the main parasitoid reared from R. pilosa galls in Serbia, with Eurytoma spp. as the second most abundant. Predation Rhinusa eversmanni is an inquiline, or gall intruder, of R. pilosa. R. eversmanni oviposits in fully developed, ten day-old R. pilosa galls, and its larvae directly compete with resident R. pilosa larvae. At 20 °C, embryogenesis takes significantly longer (F1,25 = 2055, P \ 0.001) for R. pilosa (mean ± SE) (11.8 ± 0.06 days, n = 18) than for R. eversmanni (7.7 ± 0.03 days, n = 9). Faster egg development coupled with a significantly (F1,24 = 26, P \ 0.001) larger size of R. eversmanni first instars (0.26 ± 0.01 mm, n = 12) compared to that of R. pilosa (0.22 ± 0.003 mm, n = 14) reinforce the competitive advantage of R. eversmanni over R. pilosa. The ratio of R. pilosa to R. eversmanni adults emerging from L. vulgaris decreased from 91–86 to 27–16 % from mid April to mid-May in 2004 and 2005. The decrease in R. 123 A. Gassmann et al. pilosa emergence from April to May may be explained by the extended exposure of galls to R. eversmanni. Host specificity tests No-choice adult feeding and survival This unreplicated study suggested a greater total amount of feeding on L. vulgaris plants collected in Europe than on plants originating from established North American populations of either L. vulgaris or L. dalmatica (Table 1). R. pilosa feeding was found on four of the six species tested in tribe Antirrhineae. Prolonged adult survival was recorded on three native North American Antirrhineae. Feeding was not detected and survival was zero on the two native species tested from other Plantaginaceae tribes. No-choice oviposition, gall induction and larval development There was a significant plant effect on the number of adults per replicate (F12,855 = 24.1, P \ 0.001) and the number of adults per gall (F12,1934 = 53.3, P \ 0.001) (Table S2). However, the number of larvae developing to the adult stage for L. vulgaris from European (EU) populations did not significantly differ from L. vulgaris from North American (NA) populations. The mean scores for L. vulgaris from EU and L. vulgaris from NA were significantly different from those of L. genistifolia from EU and L. dalmatica from NA indicating that the later two species were less suitable R. pilosa hosts than L. vulgaris, possibly associated with a defensive reaction by plants (i.e., a hypersensitive response) following oviposition (Barnewall and De Clerck-Floate 2012). This reaction can be described as a rejection by the plant of the insect and/or its gall-inducing stimulus, and was characterized by a distinct longitudinal splitting of the stem region where R. pilosa oviposition had occurred. The number of adults per replicate and the number of adults per gall for L. genistifolia from EU was not significantly different from L. dalmatica from NA. With the exception of Nuttallanus canadensis, S. virga and S. nuttallianus, no gall development was recorded on any native North American species (Table S2). Oviposition marks were recorded on the native North American Epixiphium wislizeni (Engelm. ex A. 123 Gray) Munz but did not result in any gall or larval development. The mean number of galls per replicate was 1.1 on S. virga and S. nuttallianus, but with no larval development reported on the latter species. On S. virga only three of the 297 galls resulted in completion of larval development to adult emergence (Table S2). No dead adults have been recorded within a gall or shoot of either field or test plants. Sequential no-choice gall induction and larval development Single R. pilosa pairs sequentially transferred between host plant species resulted in 40.3 ± 8.1 oviposition marks and 12.0 ± 2.0 galls on L. vulgaris, compared to 6.0 ± 2.1 oviposition marks and 1.3 ± 0.4 galls on S. virga. A mean number of 31.8 ± 6.1 adults emerged from L. vulgaris. No adults emerged from S. virga. Single-choice oviposition, gall induction and larval development L. vulgaris versus L. genistifolia The release of five R. pilosa pairs onto 40 potted plants of L. vulgaris and 40 potted plants of L. genistifolia generated a total of 42 galls only on L. vulgaris from which 39 adults emerged. When six R. pilosa pairs were released into a field cage set up with eight potted plants of L. vulgaris and eight potted plants of L. genistifolia, a total of 139 oviposition marks and 32 galls were recorded on L. vulgaris, but only 13 oviposition marks and two galls were recorded on L. genistifolia. The two galls observed on L. genistifolia were induced at the end of April during the peak of gall formation on L. vulgaris. 95 adults emerged from the L. vulgaris galls, no adults were produced on L. genistifolia. L. vulgaris versus S. virga Five R. pilosa pairs released into a field cage set up with 12 potted plants of L. vulgaris and 12 potted plants of S. virga produced a total of 39 galls on ten of the 12 L. vulgaris test plants, and four galls on two of the 12 S. virga test plants. The first gall to appear on S. virga was recorded one month later than the earliest galls that appear on L. vulgaris. No larval development occurred in S. virga galls. A total of 83 adult weevils emerged from L. vulgaris galls. Single R. pilosa female-male pairs exposed to two L. vulgaris and two S. virga plants growing Biology and host specificity of Rhinusa pilosa together in a large pot covered with a gauze sleeve then transferred to new pots (same plant species and densities) every 2–3 days produced a total of 68 galls and 224 oviposition marks resulting in 189 adults from L. vulgaris. In contrast, only two galls, without larval development, were recorded on S. virga. versus 71.1 ± 1.9 cm). Attacked plants produced significantly more shoots than controls (n = 30) (12.1 ± 0.6 shoots versus 8.5 ± 0.6 shoots; F1, 58 = 20.8, P \ 0.001). Discussion L. vulgaris versus S. nuttallianus Five R. pilosa pairs released into a field cage set up with eight potted plants of L. vulgaris and eight potted plants of S. nuttallianus produced a total of 33 galls on seven of the eight L. vulgaris plants and one gall each on two of the eight S. nuttallianus plants. No larval development was recorded on S. nuttallianus but 99 adults emerged from the L. vulgaris galls. As on S. virga, gall induction occurred later on S. nuttallianus compared to L. vulgaris. Multiple-choice gall induction and larval development Six pairs of R. pilosa released into field cages (n = 2) set up with six potted plants of S. virga, six potted plants of S. nuttallianus, and 24 potted plants of L. vulgaris produced a total of 274 galls on L. vulgaris from which 622 adult weevils emerged. No galls were recorded on S. virga or S. nuttallianus. Gall impact on host plant fitness Comparison of galled versus control plants showed highly significant reductions in plant height (37.2 ± 3.1 versus 58.4 ± 3.2 cm; F1, 58 = 22.1, P \ 0.001), dry below-ground biomass (0.4 ± 0.06 versus 1.2 ± 0.1 g; F1, 58 = 32.9, P \ 0.001), dry above-ground biomass (galls excluded) (4.6 ± 0.7 versus 12.6 ± 1.4 g; F1, 58 = 26.3, P \ 0.001) and shoots produced (3.5 ± 0.6 versus 13.6 ± 1.3; F1, 58 = 10.7, P = 0.0018). Gall impact on the fitness of S. virga All shoots on S. virga plants exposed to R. pilosa were atypically galled. Dry below-ground biomass was significantly higher (5.8 ± 0.4 versus 4.6 ± 0.4 g; F1, 58 = 5.4, P = 0.02) for attacked plants compared to control plants. Dry above-ground biomass and plant height were similar for attacked and control plants plants (41.9 ± 2.0 versus 39.1 ± 2.8 g; 75.7 ± 2.3 Rhinusa pilosa is a univoltine shoot-galling weevil found exclusively on L. vulgaris in Europe. In a survey of 20 Linaria species and subspecies from 290 populations, including three stands where L. genistifolia and L. vulgaris occur sympatrically, R. pilosa was recorded exclusively on L. vulgaris. Under no-choice conditions, all 13 non-target Linaria species exposed to R. pilosa were accepted for oviposition and most were found to be suitable, to varying degrees, for gall and larval development. In Europe, historical field host records for R. pilosa reported in the literature need to be confirmed, since in most cases no distinction has been made between R. pilosa and R. brondelii. A recent population genetics study of weevils from the R. pilosa–R. brondelii complex, for example, highlighted what appears to be a significant host-associated, genetic divergence of R. brondelii collected from L. genistifolia and L. purpurea (Toševski, unpublished data). Even if only minor morphological differences can be found in the species of the R. pilosa group, it was concluded that all these taxa should be treated as very old and relict species, and may be very specific to the host plant with which they are associated. In two separate single-choice tests, R. pilosa produced significantly fewer oviposition scars and galls and no adults on L. genistifolia compared to L. vulgaris. This confirms that L. genistifolia is not a suitable host. Linaria vulgaris from European and North American populations of L. vulgaris were similarly accepted by R. pilosa for oviposition. Both host populations also were able to support full gall and weevil development, despite adult feeding being lower on the North American L. vulgaris tested. Barnewall (2011) found that R. pilosa was able to gall and develop successfully on all Canadian populations of L. vulgaris tested, even though the plant growth variables measured were different among the populations at the set-up of the experiment, and there were population-dependent differences in plant growth responses post-galling. The test plant list has taken into consideration, both ecologically and economically, the importance of the 123 A. Gassmann et al. now more speciose Plantaginaceae family, and also is more phylogenetically relevant for delineating the host range of candidate Linaria spp. biological control agents (Gaskin et al. 2011). Adult feeding and survival was minimal on native North American species in the weed tribe Antirrhineae. Moreover, oviposition was limited to only four out of 63 native North American species in no-choice tests. Of the native species attacked by R. pilosa, only three larvae developed to the adult stage on only one species, S. virga. Very little oviposition and gall development without larval development was recorded on N. canadensis, whose phylogenetic relationship with Linaria has been recently disputed (Fernández-Mazuecos et al. 2013). Oviposition on native North American species was further reduced in choice experiments. Relative to impact on R. pilosa’s preferred host, L. vulgaris, gall development had apparently no negative impact on the growth of the one native plant species supporting gall and larval development, S. virga. In contrast, stem-galling had a highly significant impact on the vegetative growth of L. vulgaris. In a similar experiment, Barnewall (2011) found that galled L. vulgaris plants produced fewer flowering stems in comparison to controls. Although there was no significant difference between the galled and control plants in the total amount of dried, above-ground biomass, approximately 40 % of the total produced by galled plants constituted gall tissues. Thus, if gall biomass is subtracted from the total above ground biomass produced by the galled plants, it could be deduced that there was a significant loss in net plant productivity due to galling. Under no-choice conditions, four plant species native to North America were found to be suitable for oviposition, and three supported gall development, i.e. N. canadensis, S. nuttallianus and S. virga. When galls form on these sub-optimal hosts, they are abnormal in shape and growth. Externally, they appear as elongated, slight swellings of the stem, which sometime cause stem contortions. Internally, the central gall cavity when present is often narrow and partially occluded by parenchymous tissue. Occasional abnormal stem tissue proliferation was observed to crush weevil eggs deposited within the pith of sub-optimal hosts (Barnewall 2011). This resulted in high egg mortality, and with very few exceptions, the plant’s response did not allow normal larval development to the adult stage. 123 Host-range studies demonstrated that L. vulgaris is the most suitable host plant for R. pilosa. Choice tests with L. vulgaris and non-target species have confirmed that R. pilosa clearly prefers to oviposit on L. vulgaris, which was also the most suitable host for successful gall and larval development. R. pilosa is highly specific to L. vulgaris and the potential impact on non-target species is negligible under these conditions. R. pilosa has been associated in the field exclusively with L. vulgaris, and demonstrates the potential to exert significant impact on its host plant if introduced. Populations are expected to increase rapidly where they will be free from regulation by the gall intruder R. eversmanni and its other native range natural enemies. The Technical Advisory Group for the Biological Control of Weeds (TAG—–BCW) has recommended release of R. pilosa in September 2013. Acknowledgments We are most grateful for the laboratory assistance of J. Jović in Serbia, and E. Barnewall, E. Pavlik and C. Durand in Canada. R. Caldara (Milano, Italy) is acknowledged for his important contribution to the classical taxonomy of Rhinusa, and B. Emerson, G. Hernández-Vera and L. Legarreta (University of East Anglia, Norwich) whose timely molecular analyses resolved key issues related to the phylogeny of Rhinusa pilosa and relatives. Funding for this research was enabled by the Toadflax Biological Control Consortium, through long-persistent support and interest from Agriculture and Agri-Food Canada, British Columbia Ministry of Forests, Lands and Natural Resource Operations, British Columbia Ministry of Agriculture, California Department of Food and Agriculture, Montana Noxious Weed Trust Fund (through Montana State University), USDA Forest Service Rocky Mountain Research Station, USDA APHIS PPQ Center for Plant Health Science and Technology, and Wyoming Biological Control Steering Committee. References Albach DC, Meudt HM, Oxelman B (2005) Piecing together the ‘‘new’’ Plantaginaceae. Am J Bot 92:297–315 Bakshi TS, Coupland RT (1960) Vegetative propagation in Linaria vulgaris. Can J Bot 38:243–249 Barnewall EC (2011) Plant-insect interactions between yellow toadflax, Linaria vulgaris, and a potential biocontrol agent, the gall-forming weevil, Rhinusa pilosa. M.Sc. Thesis, University of Lethbridge, Lethbridge, AB, Canada Barnewall EC, De Clerck-Floate RA (2012) A preliminary histological investigation of gall induction in an unconventional galling system. Arthropod Plant Interact 6:449–459 Boswell A (2013) Development of PCR-RFP and DNA barcoding chloroplast markers for yellow toadflax and Dalmatian toadflax. M.S. Thesis, Colorado State University, Fort Collins, CO, USA Biology and host specificity of Rhinusa pilosa Briese DT (2005) Translating host-specificity test results into the real world: the need to harmonize the yin and yang of current testing procedures. Biol Control 35:208–214 Caldara R (2001) Phylogenetic analysis and higher classification of the tribe Mecinini (Coleoptera: Curculionidae, Curculioninae). Koleopterol Rundsch 71:171–203 Caldara R, Desančić M, Gassmann A, Legarreta L, Emerson BC, Toševski I (2008) On the identity of Rhinusa hispida (Brullé) and its current synonyms (Coleoptera: Curculionidae). Zootaxa 1805:61–68 Caldara RD, Sassi D, Toševski I (2010) Phylogeny of the weevil genus Rhinusa Stephens based on adult morphological characters and host plant information (Coleoptera: Curculionidae). Zootaxa 2627:39–56 Chater AD, Valdés B, Webb DA (1972) Linaria Miller. In: Tutin TG, Heywood VH, Burgess NA, Walters SM, Webb DA (eds) Flora Europaea, vol 3. Cambridge University Press, Cambridge, UK, pp 226–236 De Clerck-Floate RA, Harris P (2002) Linaria dalmatica (L.) Miller, Dalmatian toadflax (Scrophulariaceae). In: Mason PG, Huber JT (eds) Biological control programmes in Canada, 1981–2000. CABI Publishing, Wallingford, UK, pp 368–374 De Clerck-Floate RA, McClay AS (2013) Linaria vulgaris Mill., yellow toadflax (Plantaginaceae). In: Mason PG, Gillespie DR (eds) Biological control programmes in Canada 2001–2012. CABI Publishing, Wallingford, UK, pp 354–362 De Clerck-Floate RA, Richards KW (1997) Pollination ecology and biocontrol: developing release strategies for seed feeding insects on Dalmatian toadflax. Acta Hortic 437(379):384 De Clerck-Floate RA, Turner SC (2013) Linaria dalmatica (L.), Mill., Dalmatian toadflax (Plantaginaceae). In: Mason PG, Gillespie DR (eds) Biological control programmes in Canada 2001–2012. CABI Publishing, Wallingford, UK, pp 342–353 Estes D, Small RL (2008) Phylogenetic relationships of the monotypic genus Amphianthus (Plantaginaceae tribe Gratioleae) inferred from chloroplast DNA sequences. Syst Bot 33:176–182 Fernández-Mazuecos M, Blanco-Pastor JL, Vargas P (2013) A phylogeny of toadflaxes (Linaria Mill.) based on nuclear ITS sequences: systematic and evolutionary consequences. Int J Plant Sci 174:234–249 Gaskin JF, Bon M-C, Cock MJW, Cristofaro C, Biase A, De Clerck-Floate RA, Ellison CA, Hinz HL, Hufbauer RA, Julien MH, Sforza R (2011) Applying molecular-based approaches to classical biological control of weeds. Biol Control 58:1–21 Ghebrehiwet M, Bremer B, Thulin M (2000) Phylogeny of the tribe Antirrhineae (Scrophulariaceae) based on morphological and ndhF sequence data. Plant Syst Evol 220:223–239 Harris P (1963) Host specificity of Calophasia lunula (Hufn.) (Lepidoptera: Noctuidae). Can Ent 95:101–105 Hoffmann A (1958) Faune de France, vol. 62: Coleoptères, Curculionidae. In: Lechevalier P (ed) Paris, France, pp 1264–1311 Lohse GA, Tischler T (1983) Die Kafer Mitteleuropas. Bd. 11: Curculionidae. In: Freude H, Harde KW, Lohse GA (eds) Goecke & Evers, Krefeld, Germany, p 260–271 Mack RN (2003) Plant naturalizations and invasions in the eastern United States: 1634–1860. Ann Mo Bot Gard 90:77–90 McClay AS, De Clerck-Floate RA (2002) Linaria vulgaris Miller, common toadflax (Scrophulariaceae). In: Mason PG, Huber JT (eds) Biological control programmes in Canada, 1981–2000. CABI Publishing, Wallingford, UK, pp 375–382 Mimeur J-M (1949) Contribution à l’ètude des zoocécidies du Maroc. Mémoire hors série de la Société des Sciences naturelles du Maroc Paul Lechevalier, Paris, France Missouri Botanical Garden (http://www.missouribotanical garden.org/PlantFinder/PlantFinderDetails.aspx?kemper code=a242. Accessed 20 Dec 2013 Nadeau L, King JR (1991) Seed dispersal and seedling establishment of Linaria vulgaris Mill. Can J Plant Sci 71:771–782 Olmstead RG, dePamphilis CW, Wolfe AD, Young ND, Elisons WJ, Reeves PA (2001) Disintegration of the Scrophulariaceae. Am J Bot 88:348–361 Saner MA, Clements DR, Hall MR, Doohan DJ, Crompton CW (1995) The biology of Canadian weeds. 105. Linaria vulgaris Mill. Can J Plant Sci 75:525–537 Sing SE, Peterson RKD, Weaver DK, Hansen RW, Markin GP (2005) A retrospective analysis of known and potential risks associated with exotic toadflax-feeding insects. Biol Control 35:276–287 Sutton DA (1988) A revision of the tribe Antirrhineae. British Museum (Natural History), London, UK Tank DC, Beardsley PM, Kelchner SA, Olmstead RG (2006) L.A.S. Johnson Review No. 7: review of the systematics of Scrophulariaceae s.l. and their current disposition. Aust Syst Bot 19:289–307 Turner MFS (2012) Viability and invasive potential of hybrids between yellow toadflax (Linaria vulgaris) and Dalmatian toadflax (Linaria dalmatica). Ph.D. Dissertation, Colorado State University. Fort Collins, CO, USA USDA, NRCS (2013) The PLANTS Database. National Plant Data Team, Greensboro, NC, USA. (http://plants.usda.gov. Accessed 10 Aug 2013) Vargas P, Rosselló JA, Oyama R, Güemes J (2004) Molecular evidence for naturalness of genera in the tribe Antirrhineae (Scrophulariaceae) and three independent evolutionary lineages form the New World and the Old. Plant Syst Evol 249:151–172 Vujnovic K, Wein RW (1997) The biology of Canadian weeds. 106. Linaria dalmatica (L.) Mill. Can J Plant Sci 77:483–491 Wapshere AJ (1974) A strategy for evaluating the safety of organisms for biological weed control. Ann Appl Biol 77:201–211 Ward SM, Fleischmann CE, Turner MF, Sing SE (2009) Hybridization between invasive populations of Dalmatian toadflax (Linaria dalmatica) and yellow toadflax (Linaria vulgaris). Invasive Plant Sci Manage 2:369–378 Wilson LM, Sing SE, Piper GL, Hansen RW, De Clerck-Floate RA, MacKinnon DK, Randall CB (2005) Biology and biological control of dalmatian and yellow toadflax. USDA Forest Service, FHTET-05-13, Morgantown, USA 123