Annual Report 2011 Biological control of dyer’s woad, Isatis tinctoria

H.L. Hinz, L. Blair, C. Closca and E. Gerber

March 2012

www.cabi.org

KNOWLEDGE FOR LIFE

CABI Ref: VM01736 Issued March 2012

Biological control of dyer’s woad, Isatis tinctoria

Annual Report 2011

H.L. Hinz, L. Blair, C. Closca and E. Gerber

CABI Rue des Grillons 1, CH-2800 Delémont, Switzerland Tel: ++ 41 32 421 4870 Fax: ++ 41 32 421 4871 Email: Europe-CH@cabi.org Sponsored by: Bureau of Land Management, Idaho Counties of the states of Idaho and Utah USDA-APHIS-CPHST Wyoming Biological Control Steering Committee

This report is the Copyright of CAB International, on behalf of the sponsors of this work where appropriate. It presents unpublished research findings, which should not be used or quoted without written agreement from CAB International. Unless specifically agreed otherwise in writing, all information herein should be treated as confidential.

Table of Contents Summary 1 1.  Introduction 3 2.  Work Programme for Period under Report 4 3.  Ceutorhynchus peyerimhoffi HUSTACHE (Col., Curculionidae) 5 

3.1.  Rearing and fecundity 5 3.2.  Manipulation of dyer’s woad phenology 5 3.3.  Host-specificity tests 5 

3.3.1. No-choice oviposition tests 5 3.3.2. No-choice development tests 8 

3.4.  Discussion and outlook 8 4.  Ceutorhynchus rusticus GYLLENHAL (Col., Curculionidae) 9 

4.1.  Emergence and collections 9 4.2.  Host-specificity tests 9 

4.2.1. No-choice tests established in autumn 2010 9 4.2.2. No-choice tests established in autumn 2011 11 4.2.3. Open-field test established in autumn 2011 12 

4.3.  Discussion and outlook 14 5.  Psylliodes isatidis HEIKERTINGER (Col., Chrysomelidae) 14 

5.1.  Rearing 14 5.2.  Fecundity 15 5.3.  Host-specificity tests 15 

5.3.1. No-choice larval transfer tests 15 5.3.2. No-choice oviposition and development tests 17 

5.4.  Impact experiment 17 5.5.  Discussion and outlook 17 

6.  Combined Open-field Test for Ceutorhynchus rusticus and Psylliodes isatidis 18 7.  Field Impact of Potential Agents and Population Dynamics of Dyer’s Woad 20 

7.1.  Material and methods 20 7.2.  Results 21 7.3.  Discussion and outlook 22 

8.  Work Programme Proposed for 2012 23 9.  Acknowledgements 23 10.  References 24 

1

Summary

1. Work in 2011 concentrated on the three most readily available biological control candidates for dyer’s woad, Isatis tinctoria, i.e. the seed-feeding weevil Ceutorhynchus peyerimhoffi, the root crown-mining C. rusticus, and the stem-mining flea beetle, Psylliodes isatidis.

2. For C. peyerimhoffi we continued with no-choice oviposition and larval development tests. Although we had more weevils available than last year, fewer tests could be established, due to an unusually hot spring and consequently accelerated plant phenology. Tests were established with 13 test plant species, eight native to North America (NA), including one federally listed endangered species. Apart from dyer’s woad, eggs were only found in four European test species. In subsequent development tests, none of the three test species exposed supported larval development, confirming the narrow host range of C. peyerimhoffi. Our rearing worked extremely well in 2011 and nearly 3000 larvae emerged. We are planning on continuing no-choice oviposition and development tests in 2012 and provided sufficient plant material is available to establish a multiple-choice field cage test with species accepted for oviposition under no-choice conditions. 3. In autumn 2010, we established additional no-choice oviposition and development tests for C. rusticus with 21 test plant species, 16 native to NA, including two federally listed endangered species. Larvae (alive or dead) were found in six test species, but adults of C. rusticus only emerged from dyer’s woad in 2011. In autumn 2011 additional no-choice tests were established with 27 test plant species, the majority native to NA, again including two federally listed endangered species. Since once again relatively few C. rusticus emerged from dyer’s woad control plants exposed in an open-field test established in autumn 2010 (see below), we set up an additional test in autumn 2011 with C. rusticus only, in which we dissected exposed plants for eggs. All dyer’s woad control plants were heavily attacked and an average of over 60 eggs and/or first instar larvae were found per plant. Three plants of two test species were attacked as well, but only very few eggs and/or first instar larvae were found. We assume that attack was influenced by the fact that we placed egg laying females onto the exposed test plants and not in between them. A few females might have therefore stayed on the plants they were released upon to lay a small number of eggs. The test nevertheless showed that potential non-target attack by C. rusticus under open-field conditions is a very rare event. We are planning to repeat a similar open-field test in autumn 2012. 4. We continued the rearing and no-choice larval transfer tests for P. isatidis in 2011. Of the 21 test species exposed, adults emerged from 11; however, apart from Stanleya pinnata, successful larval development was much lower than on dyer’s woad. Because S. pinnata is known to hyperaccumulate selenium, which can act as a herbivore defence, half of the S. pinnata plants were treated with 40 micromolar sodium selenate prior to transfer of P. isatidis larvae, while the other half of the plants were left untreated. Plants treated with selenium showed significantly fewer visual attack symptoms and no adults emerged. The fact that even the lowest selenium levels inhibited development of P. isatidis on S. pinnata is encouraging, since these are well within the range found in naturally growing S. pinnata.

2

In addition, an impact experiment was established, transferring three different densities of larvae (0, 30, 60 and 90 per plant) onto potted dyer’s woad plants. Shoot height was reduced by about 22 cm at the highest density, while shoot base diameter, biomass and seed production were not significantly decreased. Although seed production was reduced by 58% at the highest density compared to control plants, results were not statistically significant because of high variability. The accelerated phenology of dyer’s woad due to an unusually hot spring is thought to have hampered the impact experiment with P. isatidis and we are therefore planning to repeat it in 2012. 5. In autumn 2010, we established a joint open-field test for C. rusticus and P. isatidis in southern Germany. All plants that had been exposed and survived overwintering at the centre were individually covered with gauze bags in spring 2011 and regularly checked for adult emergence. Adults of P. isatidis emerged from eight and adults of C. rusticus from six out of nine exposed dyer’s woad plants. Apart from one P. isatidis that emerged from the European Barbarea vulgaris no adults of either species emerged from any of the other test plants exposed, confirming their narrow host range under natural field conditions. 6. We continued the experiment in southern Germany to test the effect of C. rusticus and P. isatidis in combination with interspecific plant competition on dyer’s woad under field conditions. In 2011, there were hardly any dyer’s woad plants left on unweeded plots. Data collection therefore concentrated on weeded plots. As in 2010, plants on unsprayed plots produced shorter and thinner shoots than plants on sprayed plots, while the number of shoots was similar. In contrast to 2010, the reduction in seed output on unsprayed plots was, at 72%, much more severe in 2011. Since C. rusticus was again the dominant herbivore, we assume that most of the reduction in seed output was due to C. rusticus attack. Whether the reduction in seed production will have an effect on the population dynamics of dyer’s woad in the long term would need to be investigated using a modelling approach. In Idaho, USA, a very similar experiment was established in autumn 2009. First results indicate that a higher number of seedlings germinated in spring than in fall, while in southern Germany, a similar number of seedlings germinated in fall and spring. Similar to southern Germany, interspecific competition in Idaho reduced survival, the proportion of reproducing plants and seed output. Instead of herbivory, the experiment in Idaho investigated the effect of the native rust Puccinia thlaspeos. Unfortunately, only a few plants were infected with the rust and therefore no quantitative data on its impact on dyer’s woad could be collected so far.

3

1. Introduction

Dyer’s woad, Isatis tinctoria L. (Brassicaceae), is a winter annual, biennial, or short-lived perennial mustard of Eurasian origin that was introduced to North America (NA) by early colonists as a textile dye crop and then accidentally spread as a contaminant of crop seed (Hegi, 1986; McConnell et al., 1999). Today, dyer’s woad is a declared noxious weed in 11 western US states (US Department of Agriculture – Natural Resources Conservation Service [USDA-NRCS] Plants National Database [http://plants.usda.gov]; Invaders Database [http://invader.dbs.umt.edu]). It is particularly troublesome in south-eastern Idaho (Callihan et al., 1984), northern Utah, eastern Washington and Oregon (Hawkes et al., 1985), northern California (Wiggin, 1991) and western Wyoming (Whitson, 1987). McConnell et al. (1999) summarized the distribution, biology, impacts and management of dyer’s woad in NA. Seedlings can emerge in the fall or spring and produce rosettes, which typically flower in April and May of the second year. Seeds are the only means of reproduction and spread, and one plant may produce as many as 10,000. Seed pods of dyer’s woad are thought to have allelopathic effects. Although the stems of dyer’s woad die after seed production, plants can resprout from the thick rootstalk for several years, enabling the plant to behave as a short-lived perennial. New populations of dyer’s woad often establish along roadsides and railways, and from there can quickly spread into crops, rangeland and forests. Unlike many other mustard weeds, dyer’s woad does not depend on disturbance to establish and can readily invade and dominate well-vegetated rangeland sites. Although the plant prefers nutrient-rich, alkaline soils, its 1- to 2-m-deep taproot also allows it to thrive in rocky or sandy soils with limited water-holding capacity. Dyer’s woad can spread extremely quickly, e.g. one infestation in Montana increased from two acres to more than 100 acres in just two years, and the weed is estimated to spread at an annual rate of 14% on rangeland of the Bureau of Land Management (BLM) in the Pacific North-West. It was estimated that dyer’s woad reduced crop and rangeland production in Utah by $2 million in 1981, and the size of its infestation doubled there within ten years. Small infestations of dyer’s woad can be effectively controlled by hand-pulling bolting or flowering plants. On cropland, cultivation can be effective, but must be repeated 2–3 times a year for several years. Mowing is not considered an effective method, because plants can resprout from the crown. Chemical control of dyer’s woad, for example with phenoxy and sulfonylurea herbicides, is effective when applied at the higher label rates during the rosette stage. Metsulfuron in combination with 2,4-D are the herbicides found to be most effective against dyer’s woad in pastures and rangelands. On a large scale, 2,4-D is the most economical herbicide. However, in rangeland or forests, chemical control can be limited by inaccessible terrain, possible undesirable impacts and questionable economic returns. To date, no biological control agents have been introduced to NA for the control of dyer’s woad, although a rust fungus, Puccinia thlaspeos C. Schub., has been found to infect the plant in Idaho and Utah (Kropp et al., 1997), and at high infection levels can prevent seed or fruit production. However, infection rates are quite variable and by itself the rust is not solving the problem. In conclusion, there are no satisfactory methods at this point to sustainably control dyer’s woad and halt its spread. Therefore, an initiative was started in spring 2004 by a consortium of south-eastern Idaho and Utah counties, administered through the

4

Hoary Cress Consortium, to investigate the potential for classical biological control of dyer’s woad. CABI in Switzerland (CABI) was provided with modest start-up funding to conduct preliminary surveys for potential biological control agents against dyer’s woad in Europe. In 2005, Jim Hull (Noxious Weed Superintendent, Franklin County, Idaho) formed a Dyer’s Woad Task Force that was able to obtain funding from the Idaho State Department of Agriculture (ISDA), USDA Forest Service and various counties for a full foreign exploration programme. In 2008, funding by ISDA was unfortunately discontinued. Thanks to Joseph Milan (Biological Control Specialist, BLM/ISDA, Idaho), this funding shortcut was counterbalanced by investment from BLM, Idaho, in 2010 and 2011. In addition, we received funding in 2011 from USDA-APHIS-CPHST (Animal and Plant Health Inspection Service – Center for Plant Health Science and Technology), the Wyoming Biological Control Steering Committee, and counties of the states of Idaho and Utah.

2. Work Programme for Period under Report

Given the unsecure funding situation, we continued to concentrate on the three most advanced and/or easily available potential agents in 2011.

Ceutorhynchus peyerimhoffi (Col., Curculionidae) Continue no-choice oviposition and development tests and try to quantify the

number of seeds damaged by adult feeding; Confirm that test plants that were accepted under no-choice conditions do not

support larval development; Continue the rearing colony.

Ceutorhynchus rusticus (Col., Curculionidae) Monitor quality of plants exposed in no-choice oviposition and development

tests and in the open-field test established in autumn 2010 during winter and record adult emergence in spring 2011;

Depending on results, establish additional no-choice tests and/or an open-field test in autumn 2011.

Psylliodes isatidis (Col., Chrysomelidae) Establish additional no-choice larval transfer tests, especially with test plants

that had supported adult development previously, but where insect specimens were not sent to a taxonomist for species confirmation;

Monitor quality of plants exposed in the open-field test in autumn 2010 during winter and record adult emergence in spring 2011;

Depending on results, establish an additional open-field test in autumn 2011.

Agent impact under field conditions and population dynamics of dyer’s woad Continue monitoring plant survival and seed production; Record C. rusticus and P. isatidis attack in spring 2011.

5

3. Ceutorhynchus peyerimhoffi HUSTACHE (Col., Curculionidae)

This seed-feeding weevil was selected as an additional potential agent for dyer’s woad in 2008, because it has so far only been collected from dyer’s woad in Europe, indicating a narrow host range (E. Colonnelli, pers. comm.). In addition, an agent reducing seed output would be expected to contribute to slowing the invasion of dyer’s woad. Adults emerge in spring, at a time when dyer’s woad is flowering. Females lay their eggs, mostly one per fruit, into the developing seed pods of dyer’s woad. The hatching larva feeds on the developing seed, which it destroys entirely. Mature larvae leave the pods to pupate in the soil, where adults overwinter.

3.1. Rearing and fecundity Between 26 April and 9 May 2010 (a similar time to last year), 128 female and 145 male C. peyerimhoffi emerged from our rearing culture established in summer 2010, i.e. a success rate of 48%. Weevils (10–17 females and 9–16 males) were initially placed for 1–2 weeks onto large, gauze-covered flowering through to reproducing shoots of dyer’s woad so that females could feed on flowers and developing seeds to develop their ovarioles and mature eggs. Between 9 and 17 May, all females were set up individually in little cups with a gauze lid and provided with cut flowering to reproducing inflorescences of dyer’s woad. After 1–2 days, seeds were dissected; 117 females laid eggs, i.e. 91%, similar to last year. Only egg laying females were used in host-specificity tests (see below). To obtain an estimate of the daily fecundity of females, 20 females were set up individually in cups (see above) and all seeds were dissected after 24 h. Each female laid an average of 5.8 ± 0.5 eggs. During the female egg laying period, 26 dyer’s woad plants were established for rearing, from which nearly 3000 larvae emerged (113 per plant!). Larvae were placed individually or ten together into plastic vials filled two-thirds with sifted soil for pupation. Vials are being kept in an underground insectary during winter for adult emergence in 2012.

3.2. Manipulation of dyer’s woad phenology Since the phenology of dyer’s woad plants is usually very homogenous, i.e. they all flower and reproduce around the same time, but reproductive phenology of test species is more spread out, we tried to delay the phenology of some dyer’s woad plants in 2011 so we always have control plants available that are in the right phenological stage for C. peyerimhoffi oviposition. At the end of March, we placed 35 potted, bolting dyer’s woad plants into an incubator at 5 ± 1°C with a 16/8 hour day/night cycle. On 15 April, nine plants were moved out, followed on 26 April by the rest of the plants. On 7 April, we placed an additional 11 plants into a cold room set to around 10°C, which were left there until 27 April. This system worked very well, and ensured that we always had control plants in the right phenological stage available.

3.3. Host-specificity tests 3.3.1. No-choice oviposition tests METHODS Two females and two males were released onto gauze-covered inflorescences of dyer’s woad control or test plants with developing fruits (see Plate

6

1). After about four days, weevils were retrieved and all fruits (mature and immature) were checked for feeding. For dyer’s woad, a minimum of 25 mature fruits with feeding/potential oviposition holes was dissected for eggs. For test species, all mature fruits were dissected for eggs, since some feeding/potential oviposition holes were hard to see on some test species. As soon as eggs were found, dissections were stopped and the remaining fruits re-covered with gauze to allow larvae to mature. Weevils that had been on dyer’s woad and had laid eggs were placed directly onto test species, while weevils that had been on test plants were placed onto potted, reproducing dyer’s woad plants to ensure that they could feed on their normal host plant in between tests and that they were continuing to lay eggs. Weevils reared from these plants formed part of our rearing colony (also see section 3.1). Between 11 and 31 May, a total of 55 plants were established in this way: 17 dyer’s woad plants and 1–6 replicates of 13 test species (Table 1). RESULTS Nearly all dyer’s woad control plants had eggs and from the majority, larvae emerged (Table 1). On average, a much higher number of eggs was found on dyer’s woad and more larvae emerged per plant than last year. In four test species, all native to Europe, eggs were found but no larvae emerged from any of the plants kept. Most plants showed signs of feeding, two species (Aurinia saxatilis and Lepidium campestre) to a similar degree as dyer’s woad. However, apart from L. campestre, hardly any seeds were damaged due to feeding (Table 1).

Plate 1. Set-up of no-choice oviposition and larval development tests for Ceutorhynchus peyerimhoffi.

7

Table 1. Results of sequential no-choice oviposition and larval development tests for Ceutorhynchus peyerimhoffi in 2011. Plant speciesa #

plants set up

# plants with

feeding

# fruits checked

for feedingb

Mean ± SE fruits with feeding

# fruits diss.c

Mean ± SE seeds

damaged by feeding

# plants with eggs

Mean ± SE eggs

found

# plants kept for larval em.d

# plants with

larval em.d

Mean ± SE larvae

emerged (total #)

Isatis tinctoria 17 17 856 44.2 ± 5.6 493 4.5 ± 2.1 15 7.2 ± 1.4 17 13 15.4 ± 3.9 (261) Aurinia saxatilis 6 6 898 39.8 ± 7.7 347 0 1 0.2 ± 0.2 1 0 0.0 Barbarea orthocerasa 2 0 127 0.0 90 0 0 0.0 0 --- --- Cakile edentulataa 1 1 35 34.3 25 0 0 0.0 0 --- --- Caulanthus ancepsa 4 4 201 11.2 ± 1.7 78 1.25 ± 1.1 0 0.0 0 --- --- Erodium cicutarium 5 0 92 0.0 82 0 0 0.0 0 --- --- Erysimum asperuma 2 0 31 0.0 31 0 0 0.0 0 --- --- Lepidium campestre 3 3 332 50.9 ± 27.6 203 41.0 ± 20.3 1 0.3 ± 0.3 1 0 0.0 Lepidium draba 1 1 57 12.3 3 4 1 1.0 1 0 0.0 Lepidium perfoliatum 2 2 974 14.5 ± 9.2 154 1.5 ± 1.5 2 6.5 ± 5.5 2 0 0.0 Nasturtium gambeliia 5 5 183 7.8 ± 2.5 147 0.4 ± 0.6 0 0.0 0 --- --- Peritoma luteaa 2 0 16 0.0 9 0 0 0.0 0 --- --- Sisymbrium linifoliaa 4 2 90 3.9 ± 3.3 41 0 0 0.0 0 --- --- Thelypodium laciniatuma 1 1 51 9.8 46 0 0 0.0 0 --- ---

a Plant species native to North America, nomenclature after Boufford et al. (2010). b Mature and immature fruits were checked for feeding. c Only mature fruits were dissected (diss.) for eggs. d em. = emergence.

8

3.3.2. No-choice development tests METHODS For two test species in which eggs were found in 2011 (see above) or in 2010 (Boechera holboelii), no-choice oviposition and larval development tests were established using almost the same method as for no-choice oviposition tests. The only difference was that fruits were not dissected for eggs but plants left for larval emergence. Between 13 and 27 May, a total of 21 plants were established in this way: seven dyer’s woad control plants and 4-6 replicates of three test species. RESULTS Mature larvae of C. peyerimhoffi emerged from six of seven dyer’s, while none of the test plants exposed supported larval development of the weevil. After larval emergence from dyer’s woad had ceased, fruits of all test species were dissected. Hardly any seeds were found to be damaged and only one dead first instar larva was found (in L. perfoliatum) (Table 2). Table 2. Results of no-choice feeding and larval development tests for Ceutorhynchus peyerimhoffi in 2011. Plant species #

plants set up

# fruits diss.b

Mean ± SE fruits with feedingb

Mean ± SE # seeds

damaged by feedingb

# plants with larval

emergence

Mean ± SE larvae

emerged (total #)

Isatis tinctoria 7 ---c --- --- 6 16.0 ± 5.1 (112)

Aurinia saxatilis 4 201 0.8 ± 0.7 0.25 ± 0.2 0 0.0 Boechera holboelliia 6 254 1.0 ± 2.4 0.0 0 0.0 Lepidium perfoliatum

4 1759 25.6 ± 37.4 0.5 ± 0.4 0 0.0

a Plant species native to North America, nomenclature after Boufford et al. (2010). b Fruits of test plants were dissected after larval emergence from control plants had ceased. c No fruits of the dyer’s woad plants were dissected (diss.).

3.4. Discussion and outlook Rearing worked extremely well in 2011. We have currently nearly 3000 larvae overwintering for adult emergence in 2012. This should make us completely independent of field collections. In addition, a high proportion of females laid eggs and nearly all exposed dyer’s woad plants were attacked. The only factor that hampered host-specificity tests in 2011 was the extremely warm spring, which accelerated development of plants and shortened the egg laying period of C. peyerimhoffi. Therefore, fewer tests than expected could be set up. Results were again very encouraging, since we found a few eggs in some of the European test species only, and no larval development occurred in any of the test species exposed. Summarizing results of no-choice oviposition tests from 2008 to 2011, we have now exposed 34 test plant species, 20 native to NA. Ten (five native to NA) were accepted for oviposition. However, of the plants that were kept for larval emergence, none supported the development of mature larvae. This renders C. peyerimhoffi the most specific of the three agents currently being studied for the biological control of dyer’s woad. In 2012 we will continue with no-choice oviposition and development tests and we are planning to establish a multiple-choice field cage test with species accepted for oviposition under no-choice conditions.

9

4. Ceutorhynchus rusticus GYLLENHAL (Col., Curculionidae)

Females of this weevil start laying their eggs from September onwards. At CABI, they continue laying throughout winter, until early spring. Most eggs are laid directly into the leaf surface under the epidermis, although some are laid into the leaf stalk. Larvae feed in the leaf stalks and later in the root crown of dyer’s woad. The main shoot of heavily attacked plants often dies, which makes it easy to distinguish attacked from unattacked plants in the field. Mature larvae leave the plants to pupate in the soil, and adults of the F1 generation emerge from the end of May onwards. Newly emerged weevils feed on leaves of dyer’s woad, but then stop feeding, become inactive, and aestivate during July and August. At the beginning of September the weevils start feeding again.

4.1. Emergence and collections Between 18 February and 28 April 2011, 45 C. rusticus adults emerged from the host-specificity tests established in autumn 2010. They were placed in cylinders and provided with cut leaves. Between 13 and 21 September 2011, 238 females and 371 males were collected in southern Germany for host-range tests (see below). The sex ratio was therefore less male biased than in 2010 (about 1:2 in 2010 vs. 1:1.6 in 2011, females:males).

4.2. Host-specificity tests 4.2.1. No-choice tests established in autumn 2010 METHODS All plants exposed in no-choice oviposition and development tests in autumn 2010 (for details of methods see section 4.2.1 in Annual Report 2010: Hinz et al., 2011) were regularly checked during winter and spring and dying plants were dissected for signs of larval mining. Where larvae were found, their head capsule diameters were measured to determine the larval instar. Plants that did not show any signs of attack were disposed of, while all others were covered with gauze bags and kept for adult emergence. RESULTS In the 12 dyer’s woad plants dissected prematurely, larvae were found in 11, and from 9 of 13 plants kept, adults emerged (Table 3). The attack rate on controls was therefore higher than in previous years. Of the 17 test species dissected, larvae were found in six. However, except for Thelypodium laciniatum, a much smaller number of larvae than in dyer’s woad control plants was found. Seven of the 17 larvae found in T. laciniatum were dead and in most other plant species (except for Draba albertina), only first instar larvae were found.

10

Table 3. No-choice oviposition and development tests established with Ceutorhynchus rusticus in autumn 2010. Plant speciesa Number of replicates Mean ± SE

larvae found

Mean ± SE adults

emerged set up

validb with feeding (score)c

with eggsc

dissected for larvaed

with larvaee

kept for adult

emergence

with adult emergence

Isatis tinctoria 25 25 25 (1.9) 25 12 11 13 9 12.8 ± 3.0 3.5 ± 1.2 Arabis blepharophyllaa 5 5 4 (0.4) 0 3 0 2 0 0.0 0.0 Barbarea orthocerasa 5 5 5 (1.3) 5 0 --- 5 0 --- 0.0 Barbarea vulgaris 5 5 5 (0.6) 1 2d 0 4 0 0.0 0.0 Boechera hoffmanniia 5 5 5 (1.8) 5 0 --- 5 0 --- 0.0 Boechera ligniferaa 5 5 5 (0.6) 5 0 --- 5 0 --- 0.0 Brassica oleracea var. sabauda 1 1 1 (0.5) 0 1 0 0 --- 0.0 ---

Cardamine cordifoliaa 7 6 4 (0.8) 2 3 1 3 0 0.7 ± 0.7 0.0 Crambe maritima 4 4 3 (0.4) 0 4 0 0 --- 0.0 --- Descurainia incanaa 4 3 0 0 3 0 0 --- 0.0 --- Draba albertinaa 5 5 5 (0.6) 4 3 1 2 0 1.0 ± 1.0 0.0 Erysimum asperuma 1 1 1 (1.5) 1 0 --- 1 0 --- 0.0 Lepidium appelianum 5 5 3 (0.5) 4 2 0 3 0 0.0 0.0 Lepidium draba 6 5 5 (1.3) 4 1 0 4 0 0.0 0.0 Lepidium ostleria 8 8 6 (0.6) 2 6 1 2 0 0.2 ± 0.2 0.0 Lepidium virginicuma 2 2 0 0 2 0 0 --- 0.0 --- Nasturtium gambeliia 5 5 5 (0.6) 1 2d 1 4 0 0.5 ± 0.5 0.0 Parrya nudicaulisa 5 4 4 (0.6) 0 4d 0 2 0 0.0 0.0 Physaria saximontanaa 2 2 0 0 1 0 1 0 0.0 0.0 Smelowskia americanaa 5 5 5 (0.8) 3 3 0 2 0 0.0 0.0 Stanleya confertifloraa 6 4 3 (0.6) 3 4 1 0 --- 0.3 ± 0.3 --- Thelypodium laciniatuma 3 3 3 (1.3) 2 2 2 1 0 8.5 ± 3.5 0.0

a Plant species native to North America, nomenclature after Boufford et al. (2010). b A few plants died soon after infestation and were not regarded as valid replicates. c Based on visual inspection. d Some plants were only partly dissected and then kept for adult emergence. e Based on dissections.

11

4.2.2. No-choice tests established in autumn 2011 METHODS Between 19 September and 21 October 2011, 1–2 females (depending on the size of the plant) and one male were released onto individually potted, gauze-covered test and control plants. All females were tested for their ability to lay eggs prior to use in tests, and only females that laid eggs were used. Each time a series of test plants was established, several dyer’s woad plants were established as controls. After 7–10 days, weevils were retrieved from the plants, and feeding and oviposition recorded visually. Some leaves of some plants were dissected to confirm the presence of eggs. Using this method, we exposed 213 plants: 36 dyer’s woad and 1–10 replicates of 27 test species, including 24 native to NA, two of which are federally listed endangered Brassicaceae, Lepidium barnebyanum and Nasturtium gambelii (= Rorippa gambelii). Since naturally growing Stanleya species accumulate selenium, which has been shown to act as a defence against insect herbivores (Hanson et al., 2003; Freeman et al., 2007), we treated the three Stanleya species used in tests with different levels of sodium selenate (0, 10, 20 and 40 ppm of selenium as Na2SeO4) 4–6 weeks prior to their use in tests (Table 4). The treatment was continued until the beginning of December 2011. All plants were initially kept in an unheated greenhouse. Any dying plants were dissected for signs of larval mining. Where larvae were found, their head capsule diameters were measured to determine the larval instar. Plants that did not show any signs of attack were disposed of, while all others are being kept for adult emergence in 2012. To determine whether adult emergence differs when plants are kept inside or outside during winter, about half of all remaining plants from each species were placed outside at the beginning of December in a garden bed, while the other half were left in the unheated greenhouse. RESULTS As in previous years, a few replicates could not be regarded as valid, because plants died soon after infestation (Table 4). As in previous no-choice tests, feeding and egg laying occurred on nearly all plants offered. Apart from Stanleya viridiflora, where feeding was reduced by increasing levels of selenium, none of the other selenium treatments appeared to have any impact on feeding or egg laying by C. rusticus (Table 4).

12

Table 4. No-choice oviposition and development tests established with Ceutorhynchus rusticus in autumn 2011. Plant speciesa # replicates # valid

replicates with

feeding

Mean feeding score

# valid replicates with eggs set up valid

Isatis tinctoria 36 35 35 1.8 35 Boechera holboelliia 6 6 6 1.6 6 Boechera ligniferaa 4 4 4 1.6 4 Boechera pulchraa 5 5 4 0.4 3 Brassica juncea 4 3 3 1.4 2 Brassica oleracea var. sabauda 4 4 4 0.2 4 Camelina microcarpaa 3 3 2 0.2 0 Cardamine breweria 5 5 5 0.7 5 Cardamine cordifoliaa 4 3 3 2.2 3 Caulanthus pilosusa 5 4 4 2.1 4 Descurainia californicaa 5 5 5 1.1 5 Descurainia nelsoniia 4 3 3 0.4 2 Lepidium barnebyanuma 10 9 9 0.3 7 Lepidium fremontiia 3 3 3 0.6 1 Limnanthes alba ssp. albaa 5 5 3 0.1 1 Nasturtium gambellia 3 3 3 1.0 0 Physaria saximontanaa 4 4 1 0.03 0 Rorippa columbiaea 5 5 5 1.4 4 Rorippa curvisiliquaa 5 5 5 1.0 0 Rorippa sinuataa 5 5 5 1.2 1 Sibara virginicaa 1 1 1 0.5 0 Smelowskia americanaa 2 2 2 1.5 2 Stanleya elataa No selenium added 7 7 7 0.5 7 Stanleya elataa + 10 ppm selenium 6 6 6 1.0 6 Stanleya pinnataa No selenium added 6 6 6 0.8 6 Stanleya pinnataa + 10 ppm selenium 6 6 6 0.8 5 Stanleya pinnataa + 20 ppm selenium 6 6 6 1.2 6 Stanleya pinnataa + 40 ppm selenium 6 6 6 0.8 6 Stanleya viridifloraa No selenium added 8 8 8 1.2 8 Stanleya viridifloraa + 10 ppm selenium 9 9 8 0.9 9 Stanleya viridifloraa + 20 ppm selenium 8 8 8 0.5 8 Stanleya viridifloraa + 40 ppm selenium 7 7 7 0.3 7 Thelypodium laciniatuma 4 4 4 1.5 4 Thelypodium sagittatuma 3 3 3 1.8 3 Thlaspi arvense 9 8 8 0.5 1 a Plant species native to North America, nomenclature after Boufford et al. (2010). 4.2.3. Open-field test established in autumn 2011 Because relatively few C. rusticus emerged from dyer’s woad control plants exposed in an open-field test established in autumn 2010 together with the flea beetle Psylliodes isatidis (see section 6), we set up an additional test in autumn 2011, in which we dissected exposed plants for eggs. The test was established at the same site in southern Germany as in previous years.

13

METHODS On 22 September 2011, three plots were established, consisting of usually three plants each of eight test species and the control, dyer’s woad (Plate 2). For four species, only 1–2 plants were available (Table 5). Between 22 September and 2 October, 16 egg laying pairs of C. rusticus from our institute’s rearing colony were released per plot. On 14 October, all plants were taken back to CABI and were dissected for eggs and weevil larvae. Since we know from previous tests (Hinz et al., 2011) that plants might be attacked by oligophagous weevils occurring naturally in the area, all eggs and larvae were measured to determine whether they were in the range of sizes typical for C. rusticus. In addition, all eggs were placed in Petri dishes, which were regularly checked for larval hatching. Once larvae had hatched, their head capsule diameters were measured. Larvae with head capsule diameters within the range for first instar larvae of C. rusticus were transferred into vials with 95% alcohol and sent to our colleague Ivo Toševski in Serbia for molecular analyses. For any plant where not all eggs eclosed (developed into larvae), but the larvae that did hatch were determined as C. rusticus, all eggs that had been found on the plant were assumed to be C. rusticus.

Plate 2. Set-up of open-field test for Ceutorhynchus rusticus in 2011.

RESULTS All dyer’s woad control plants were heavily attacked and an average of over 60 eggs and/or first instar larvae were found per plant (Table 5). A total of three plants of two test species (Table 5) were attacked as well, but only very few eggs and/or first instar larvae were found in them. Table 5. Results of open-field test established with Ceutorhynchus rusticus in autumn 2011. Plant speciesa # plants

exposed # plants attacked

# C. rusticus eggs/larvae found

(mean ± SE) Isatis tinctoria 9 9 61.6 ± 10.3 Caulanthus crassicaulisa 5 1 1.2 ± 1.2 Descurainia pinnataa 4 0 0.0 Lepidium densifloruma 9 2 0.9 ± 0.5 Lobularia maritima 9 0 0.0 Physaria hemiphysariaa 9 0 0.0 Sisymbrium linifoliuma 9 0 0.0 Stanleya pinnataa 3 0 0.0 Thelypodium sagittatuma 4 0 0.0 a Plant species native to North America, nomenclature after Boufford et al. (2010).

14

4.3. Discussion and outlook No-choice oviposition and larval development tests advanced very well in 2011. Larvae were found in some test species, but none supported development to adult, while development on dyer’s woad control plants was improved compared to previous years. Again, many plants had to be dissected prematurely (before development to mature larva could have been completed). Consequently, we will need to further increase replicates, especially with perennial and biennial native NA species, to achieve conclusive results. Summarizing results from 2005 to 2011, we have now exposed a total of 86 test species, over half native to NA (97 test species when including the plants established in autumn 2011). Of these, six species in four genera, all native to NA, have so far supported development to adult of C. rusticus. In the open-field test established in autumn 2011, two test species (the native NA Caulanthus crassicaulis and Lepidium densiflorum) were attacked to a limited degree. We assume that attack was influenced by the fact that we placed weevils onto the exposed test plants and not in between plants. A few females might have therefore stayed on the plants they were released upon to lay a small number of eggs. The test nevertheless shows that potential non-target attack by C. rusticus under open-field conditions is a very rare event. We believe that it will be most likely to occur at high weevil densities, such as after successful control of the target weed, and when test species occur in close proximity to dyer’s woad. More tests under natural field conditions will be necessary to obtain more conclusive results on the risk of potential non-target attack by C. rusticus.

5. Psylliodes isatidis HEIKERTINGER (Col., Chrysomelidae)

In autumn, females of this flea beetle lay their eggs into the soil, close to the root crown of dyer’s woad. Some larvae hatch in autumn, but most appear to hatch in early spring and feed in the developing shoots. Mature larvae leave the plants to pupate in the soil, and adults of the F1 generation emerge from about mid May onwards. After a short feeding period, beetles become more or less inactive, stop feeding and aestivate until late summer. 5.1. Rearing Of 4590 P. isatidis eggs placed in Petri dishes in autumn 2010, and kept in an incubator at 0.5°C ± 0.5°C, 3138 eclosed in spring 2010, i.e. 68.2% ± 1.3%. This was a higher percentage than in 2010, and comparable to previous years. Newly hatched larvae were used to establish host-specificity tests (see below). The first adults started to emerge from dyer’s woad plants from the beginning of June 2011. A total of 309 P. isatidis adults emerging from dyer’s woad from host-specificity tests and the impact experiment (see below), were placed into cylinders (n = 15) for aestivation and egg laying in autumn 2011. Up to the beginning of September, 83.8% of adults had survived, a slightly higher percentage than in 2010. From the beginning of September until the beginning of December 2011 about 10,000 eggs were collected and transferred to Petri dishes (40 per dish) in preparation for larval transfer tests and a potential additional impact experiment in 2012. They are currently being kept in an incubator at -0.5°C ± 0.5°C.

15

5.2. Fecundity In 2011, we made another effort to sex P. isatidis adults. We found a good technique: immobilizing beetles in a small plastic bag allowed us to look at their abdomens under the binocular microscope. Small differences in the shape of the last abdominal segment make it possible to distinguish males and females reliably. The abdomen of females tends to be more pointed, while that of males is more rounded and has a small additional separation at the last segment (Fig. 1).

Fig. 1. Differences between males and females of Psylliodes isatidis. This feature was used to separate males and females in September after aestivation. To obtain an estimate of female fecundity, individual females (n = 16) were set up for two, five or six days in small cups and provided with horticultural sponge blocks to oviposit in and with a small dyer’s woad leaf for feeding. Results were very homogenous independent of the number of days we left females in cups. Each female laid on average 8.5 ± 0.7 eggs per day.

5.3. Host-specificity tests 5.3.1. No-choice larval transfer tests METHODS Between 30 March and 20 April 2011, 1846 newly emerged larvae were transferred onto 83 plants: 21 dyer’s woad control plants, and 1–4 plants of 21 test species, 14 native to NA, including two federally listed endangered species (Boechera hoffmannii and Streptanthus glandulosus ssp. niger). Because S. pinnata is known to hyperaccumulate selenium, which can act as a herbivore defence, half of the S. pinnata plants were treated for four weeks with 40 micromolar sodium selenate prior to transfer of P. isatidis larvae, while the other half of the plants were left untreated. Depending on plant size, 15–30 larvae were transferred with a paintbrush to the base of a petiole or the shoot base. All plants were kept in an unheated greenhouse. At the beginning of May, all plants were visually checked for signs of larval mining. Plants that did not appear to have been attacked were dissected and all unattacked plants were disposed of. At the end of May, plants were covered with gauze bags to retain emerging adults. Two leaves of each of the four S. pinnata plants treated with selenium were cut, dried, and sent to Prof. Elizabeth Pilon-Smits (Colorado State University, USA) for analyses. Most of the flea beetles emerging from test species and a subsample of beetles emerging from controls (135 in total) were sent to Dr Maurizio Biondi (Università degli Studi dell’Aquila, Italy) to confirm their identity.

16

Table 6. Results of no-choice larval transfer tests conducted with Psylliodes isatidis in 2010. Plant speciesa # larvae

transferred #

plants set up

# plants attackedb

% shoots attacked

(mean ± SE)

% leaf petioles attacked

(mean ± SE)

# plants from which

adults emerged

% larvae developed to

adult (mean ± SE)

Isatis tinctoria 476 21 21 89.3 ± 6.9 16.3 ± 1.6 19 32.4 ± 4.27 Arabis aculeolataa 45 2 1 13.3 ± 13.3 --- 0 0.0 Aubrietia deltoidea 75 3 2 3.2 ± 1.6 --- 1 1.3 ± 3.3 Barbarea orthocerasa 87 4 1 18.8 ± 18.8 --- 0 0.0 Boechera hoffmanniia 90 3 1 0.0 --- 1 1.1 ± 1.1 Boechera holboelliia 60 3 2 66.8 ± 33.6 --- 2 11.0 ± 1.0 Brassica juncea 61 3 3 100.0 --- 1 2.2 ± 2.2 Brassica napus 50 3 2 --- 11.6 ± 6.4 0 0.0 Brassica rapa 60 3 3 --- 28.8 ± 11.1 0 0.0 Cakile edentulataa 60 3 0 0.0 --- --- --- Caramine cordifoliaa 45 2 1 16.7 ± 16.7 --- 1 2.5 ± 2.5 Caulanthus ancepsa 80 3c 2 100.0 --- 1 12.5 ± 12.5 Caulanthus pilosusa 50 3 3 100.0 20.0 1 2.2 ± 2.2 Lepidium appelianum 53 3 0 0.0 --- --- --- Lepidium latipesa 67 3 3 --- 22.3 ± 4.3 2 15.5 ± 8.2 Physaria saximontanaa 16 1 0 0.0 --- --- --- Raphanus sativus 55 3 1 --- 3.0 ± 3.0 1 15.0 Rorippa curvisiliquaa 61 3 3 28.1 ± 11.1 --- 2 4.2 ± 2.3 Rorippa sylvestris 86 3 3 16.8 ± 4.2 --- 0 0.0 Sisymbrium linifoliuma 72 4 1 4.2 ± 4.2 --- 0 0.0 Stanleya pinnataa with selenium 80 4 4 --- 10.2 ± 3.7 0 0.0 Stanleya pinnataa without selenium 80 4 4 --- 52.2 ± 9.9 3 27.5 ± 11.3 Streptanthus glandulosus ssp nigera 37 2 2 100.0 --- 0 0.0

a Plant species native to North America, nomenclature after Boufford et al. (2010). b Based on visual inspection and dissection of plants. c One replicate died prematurely and was not counted as valid.

17

RESULTS All dyer’s woad control plants were attacked and from 19 of 21, adults emerged (Table 6). Successful development of transferred larvae to adult was similar to previous years. Of the 21 test species exposed, many appeared to be attacked based on visual inspection and/or dissections, during which mostly mines were found. Adults of P. isatidis emerged from 11 test species (Table 6). Of these, only Stanleya pinnata not treated with selenium supported a similar level of larval survival to that found in dyer’s woad. Plants of S. pinnata treated with selenium showed significantly lower visual attack symptoms and no adults emerged (Table 6). Selenium content of treated plants was extremely variable and ranged from 337 to 2728 mg/kg dry weight (DW) per individual leaf. None of the Boechera hoffmannii showed visual signs of attack, not even when plants were dissected. Nevertheless one adult of P. isatidis appeared to have emerged. 5.3.2. No-choice oviposition and development tests To determine (i) whether some of the perennial test species that supported development to adult in no-choice larval transfer tests would also be attacked when exposed to egg laying adults, and (ii) whether successful infestation (i.e. egg hatch in spring) is possible when plants are left inside an unheated greenhouse over winter, a few no-choice oviposition and development tests were established on 5 November 2011. One to two egg laying females of P. isatidis were released onto each of 17 gauze-covered plants: six dyer’s woad and 2–3 plants of four test species. After six days, beetles were removed. Plants will be checked for flea beetle attack in spring 2012.

5.4. Impact experiment To determine the impact of different attack levels of P. isatidis on dyer’s woad, an impact experiment was established. METHODS On 21 March 2011, 28 dyer’s woad plants of US origin were selected, the length of their longest leaf measured, and plants randomly assigned to four different larval densities: 0, 30, 60 and 90 per plant. Between 21 and 30 March, larvae were transferred as described in section 5.3.1. On 18 May, shoot height and shoot base diameter of all plants were measured and their state (reproducing or prematurely senesced) recorded. Thereafter, plants were individually covered with gauze bags and emerging adults regularly collected. On 20 June, the above-ground plant parts were harvested, seeds counted and vegetative material dried for 24 h at 80°C and then biomass measured. RESULTS Only shoot height was reduced by about 22 cm at the highest larval density (F3, 23 = 3.50, P = 0.032), while shoot base diameter, biomass and seed production were not significantly decreased by P. isatidis attack. Although seed production was reduced by 58% at the highest density compared to control plants (139 vs. 329 seeds produced on average), results were not statistically significant due to high variability. The successful development of P. isatidis larvae was not significantly affected by the number of larvae transferred, indicating that there was no intraspecific competition.

5.5. Discussion and outlook No-choice larval transfer tests advanced very well in 2011, but unfortunately confirmed the relatively broad physiological host range of larvae of P. isatidis. Because it is not clear whether the one adult that emerged from the federally listed

18

endangered Boechera hoffmannii really developed on this plant species, we are planning to conduct more replicates with this species in 2012. The fact that even the lowest selenium levels of an average of 392 mg/kg DW inhibited development of P. isatidis on Stanleya pinnata is encouraging, since these levels are well within the range found in naturally growing S. pinnata (Feist and Parker, 2001). We are planning to conduct additional larval transfer tests in 2012 with S. pinnata treated with lower levels of selenium in order to determine the lowest level that still inhibits development of P. isatidis. The high variability in selenium content between individual leaves does not appear to be unusual (E. Pilon-Smits, pers. comm.) and might be associated with the variable ability of individual plants to accumulate selenium. It should also be noted that many of the test species supporting development of the flea beetle are annuals that do not occur in autumn, i.e. during the time of female oviposition. It is therefore very unlikely to impossible that they will be attacked under natural conditions. Finally, both open-field tests conducted so far clearly showed that P. isatidis has a very narrow host range under natural conditions (see section 6.1 in Hinz et al., 2011, and section 6 below). We therefore believe that we are justified in continuing work with this species. Due to the exceptionally warm spring in 2011, the dyer’s woad plants developed much faster than usual, which likely hampered the effect of P. isatidis larvae transferred for the impact experiment. The experiment nevertheless gave an indication of how P. isatidis is likely to affect dyer’s woad, i.e. with a reduction in shoot height and seed production. Our observations in previous years have shown that very early P. isatidis attack can completely stunt plants and more or less inhibit seed production. For 2012, we are planning to continue no-choice larval transfer tests, concentrating on plant species from which flea beetle adults had emerged in previous tests but that had not been sent for confirmation of their identity, and on plant species we have not tested yet or that have low numbers of replicates. We are also planning to establish another open-field test with a number of critical test species. In addition, since a large number of eggs was collected in autumn 2011, we are considering repeating the impact experiment transferring different densities of larvae.

6. Combined Open-field Test for Ceutorhynchus rusticus and Psylliodes isatidis

In autumn 2010, an open-field test was established with C. rusticus and P. isatidis. For details of methods and preliminary results see section 6.2. in Hinz et al. (2011). METHODS On 24 September 2010, nine plots with mostly one plant of each test species and the control, dyer’s woad, were established (Table 7). Up until 14 October, about 30 C. rusticus and 14 P. isatidis had been released. On 30 October, plants were taken back to CABI and checked visually and through dissection for attack. All plants were overwintered in an unheated greenhouse and protected against frost. Unfortunately, quite a few plants died prematurely, either during the exposure period or during overwintering (see Table 7). All other plants were individually covered with gauze bags and regularly checked for adult emergence.

19

Table 7. Preliminary results of open-field test for Ceutorhynchus rusticus and Psylliodes isatidis established in autumn 2010. Plant speciesa Plants attacked in

previous tests byb # plants exposed

Number of valid replicates remaining forc

# plants with emergence of

Mean ± SE adults emerged of

C. rusticus P. isatidis C. rusticus P. isatidis C. rusticus P. isatidis C. rusticus P. isatidis Isatis tinctoria 9 9 9 6 8 1.7 ± 0.6 8.2 ± 4.3 Barbarea orthocerasa X 0 9 8 6 0 0 0.0 0.0 Barbarea vulgaris --- X 9 9 9 0 1 0.0 0.1 ± 0.1 Boechera holboeliia X X 9 9 9 0 0 0.0 0.0 Descurainia incanaa --- X 7 5 4 0 0 0.0 0.0 Descurainia nelsoniia X X 9 0 0 --- --- --- --- Lepidium appelianum --- Xd 9 1 1 0 0 0.0 0.0 Lepidium densifloruma X X 9 9 9 0 0 0.0 0.0 Lepidium oblonguma --- X 9 4 1 0 0 0.0 0.0 Lobularia maritima X X 9 4 1 0 0 0.0 0.0 Rorippa sylvestris 0 X 9 9 9 0 0 0.0 0.0 Sisymbrium linifoliuma X --- 9 7 7 0 0 0.0 0.0 Stanleya pinnataa X X 9 7 7 0 0 0.0 0.0 Stanleya viridifloraa X X 7 0 0 --- --- --- --- a Plant species native to North America, nomenclature after Boufford et al. (2010). b Attack = larvae found or adults emerged; X = attacked; 0 = not attacked; --- = plant not previously tested. c Some plants died during the exposure period or during overwintering and were therefore not regarded as valid for full larval development; since C. rusticus completes its larval development during winter or in early spring, any plant surviving until February/March was regarded as valid for development; since P. isatidis completes its development mostly in spring, only plants surviving until the end of April were regarded as valid for development. d Adults that emerged from L. appelianum proved not to be P. isatidis.

20

RESULTS Adults of C. rusticus emerged from six of nine exposed dyer’s woad plants, while adults of P. isatidis emerged from eight (Table 7). Since the few dyer’s woad plants from which no adults emerged showed typical signs of P. isatidis or C. rusticus mining, all were regarded as valid controls for the exposed test species. Compared to last year’s test, more C. rusticus adults but fewer P. isatidis adults emerged on average. Apart from dyer’s woad, only one adult of P. isatidis emerged from test plants (the European Barbarea vulgaris). Apart from C. rusticus and P. isatidis, the oligophagous Psylliodes napi (F.) and Ceutorhynchus pallipes Crotch (= C. contractus Marsham) emerged from several test plant species. Results confirm the narrow host range of both species under natural field conditions, even when test species grow in close proximity to the target weed. However, tests will need to be repeated, especially for native NA biennial and perennial plant species that did not survive well (e.g. Stanleya viridiflora).

7. Field Impact of Potential Agents and Population Dynamics of Dyer’s Woad

In autumn 2008, we established an experiment to quantify the field impact of C. rusticus and P. isatidis on dyer’s woad at the site where we are also conducting the open-field tests with C. rusticus and P. isatidis (see section 6 in this report). In addition to assessing the effect of these two potential biological control agents, we wanted to determine the effect of interspecific plant competition and the interaction of these two treatments. 7.1. Material and methods LARGE PLOTS In September 2008, 32 plots (2 × 2 m) containing dyer’s woad were established. In half of the plots, a local grass/forb mixture was sown, while in the other half of the plots, any vegetation apart from dyer’s woad was removed every 5–6 weeks to reduce interspecific competition (plots are hereafter also referred to as ’weeded’ or ’not weeded’). Establishing dyer’s woad plants were counted in the central 50 × 50 cm of each plot. In July 2009, about half of the dyer’s woad rosettes in the central plot were marked with aluminium tags. Plants were regularly measured, and herbivore feeding/attack noted. To reduce attack by herbivores, especially by C. rusticus and P. isatidis that lay their eggs in autumn, half of all plots were sprayed with the systemic insecticide Marshal (24.5% carbosulfan; Maag) at the recommended rate twice in autumn 2009, while C. rusticus and P. isatidis were released on the other half of the plots (plots are hereafter also referred to as ‘sprayed’ or ‘unsprayed’). For details of methods see section 7.1 in Annual Report 2009 (Hinz et al., 2010). 2011, was the last year of this experiment. In 2011, we continued to apply the treatments as described above, i.e. manipulation of interspecific competition and herbivory. Since hardly any dyer’s woad plants remained in the non-weeded plots, we only continued collecting data on the weeded plots. On 11 March, the state of all labelled plants in the central 50 × 50 cm was checked, plants measured and seedlings counted. On the same date, all plots were sprayed with insecticide according to treatments. On 13 April, seedlings were rechecked and usually four plants (range: 2–12) were dug up from the periphery of plots, taken back to CABI and dissected to assess insect attack. On 4 August, seedlings and remaining rosettes were counted one last time and then the experiment was terminated.

21

To estimate the number of seeds produced per plant in the central 50 × 50 cm plot, 3–9 shoots were collected from the periphery of all weeded plots on 26 May and their shoot base diameter taken, since this was determined as the parameter best correlated with seed production in 2010 (see section 7.2. in Hinz et al., 2011). On 17 June, all labelled plants in the central 50 × 50 cm were measured (number of shoots, shoot base diameter, height). The correlation established with the plants collected in the periphery of plots was used to estimate seed production of plants in the central plots separately for sprayed and unsprayed plots. To estimate the number of remaining seeds in the soil, 12 soil samples were taken on 26 May 2011 (i.e. before dispersal of seeds produced in 2011) with a bulb planter (6–7 cm diameter, 10 cm deep) from the periphery of all weeded plots. Soil was stored at 2°C until processing. The soil was washed through a sieve (2-mm mesh size), the remaining material captured in a gauze cloth (0.5-mm mesh size), dried for a few hours at 80°C and then checked under a binocular microscope for seeds of dyer’s woad. SMALL PLOTS In order to determine what proportion of seedlings in later years would be recruited from the original sown seeds, and to investigate the potential influence of germination date on survival and growth of dyer’s woad, 24 small plots (50 × 50 cm) were established in the same way as described for the large plots. From spring 2009 onwards, all dyer’s woad seedlings were regularly marked with different-coloured toothpicks and their fate followed. Seed heads of reproducing plants in these plots were cut before seeds could disperse. During 2011, we continued to mark new seedlings and followed the fate of already established plants.

7.2. Results Hardly any dyer’s woad rosettes present in September 2010 survived until March 2011 in unweeded plots. Subsequent data collection therefore concentrated on weeded plots. Although fewer plants survived the winter on unsprayed compared to sprayed plots (36% vs. 54%), differences were not significant (P > 0.05), and the number of reproducing plants in June 2011 were very similar between sprayed and unsprayed plots (5.9 vs. 5.1). However, plants on unsprayed plots produced 72% fewer seeds than plants on sprayed plots (466 vs. 1645 seeds; F1,12 = 9.59, P = 0.009), and the total number of seeds produced per plot was reduced by 60% (F1,12 = 7.17, P = 0.02). This was probably influenced by the fact that plants on unsprayed plots had reduced shoot base diameters (3.1 vs. 4.8 mm; F1,12 = 25.00, P < 0.001) and height (52 vs. 69 cm; F1,12 = 14.15, P = 0.003). The application of insecticide reduced the proportion of attacked plants (determined by dissection) from 67% to 30% in April 2011 (F1,12 = 7.56, P = 0.018) and the total number of larvae or mines found from 12 to 2.6 (F1,12 = 5.85, P = 0.032). Attack of the root crown was reduced from 45% to 4% and shoot attack from 32% to 15%. All attack of the root crown was caused by C. rusticus, and also some of the attack of shoot bases. Most larvae found were weevils, and only a few were flea beetles. So, similar to last year, most of the attack found was due to C. rusticus. The estimated soil seedbank was slightly higher on sprayed than unsprayed plots, but differences were not significant (549 vs. 691 seeds per m2).

22

7.3. Discussion and outlook In 2011, there were hardly any dyer’s woad plants left on unweeded plots. Other vegetation was extremely dense covering more or less 100% of the soil surface and mainly consisted of perennial grasses. As in 2010, plants on unsprayed plots produced shorter and thinner shoots than plants on sprayed plots, while the number of shoots was similar. In contrast to last year, the reduction in seed output on unsprayed plots was much more severe in 2011, probably influenced by the larger reduction in shoot base diameter (36% reduction in 2011 vs. 20% reduction in 2010). The reduction in seed production on unsprayed plots of 72% was exactly the same percentage of seed reduction we obtained in a common garden experiment where we quantified the impact of C. rusticus on potted dyer’s woad plants (Hinz et al., 2009). Since C. rusticus was the dominant herbivore in our field experiment, we assume that most of the reduction in seed output was due to C. rusticus attack. Whether the reduction in seed production will have an effect on the population dynamics of dyer’s woad in the long term would need to be investigated using a modelling approach. The fact that the soil seedbank did not differ significantly between sprayed and unsprayed plots in 2011 was to be expected, since seed production per plot was not significantly influenced by insecticide application in 2010, and the soil seedbank was estimated before seeds produced in 2011 dispersed. No additional seedlings germinated in the small plots, confirming the relatively short-lived seed bank of dyer’s woad at least under our experimental conditions. In Idaho, a very similar experiment was established in autumn 2009. First results indicate that a higher number of seedlings germinated in spring than in fall, while in southern Germany, a similar number of seedlings germinated in fall and spring. As in southern Germany, interspecific competition reduced survival, the proportion of reproducing plants and seed output. Instead of herbivory, the experiment in Idaho investigated the effect of the native rust Puccinia thlaspeos. Unfortunately, only a few plants were infected with the rust and therefore no quantitative data on its impact on dyer’s woad has been collected so far.

23

8. Work Programme Proposed for 2012

Ceutorhynchus peyerimhoffi (Col., Curculionidae) Continue no-choice oviposition and larval development tests; Conduct a multiple-choice field cage test with species attacked under no-choice

conditions.

Ceutorhynchus rusticus (Col., Curculionidae) Monitor quality of plants exposed in no-choice tests in autumn 2011 during winter

and record adult emergence in spring 2012; Try to complete no-choice oviposition and larval development tests (results will

only be available in 2013); Conduct an additional open-field test with species attacked under no-choice

conditions.

Psylliodes isatidis (Col., Chrysomelidae) Conduct additional no-choice larval transfer tests with plant species not exposed

yet or where flea beetles emerged that were not verified to be P. isatidis; Conduct an additional open-field test with species attacked under no-choice

conditions; Repeat impact experiment transferring different numbers of larvae onto potted

dyer’s woad plants to quantify the effect of P. isatidis.

9. Acknowledgements

We thank Danielle Fife and Célien Montavon for additional technical assistance, and Florence Willemin, Christian Leschenne and Jean Bartlé (all CABI) for plant propagation and maintenance. We are especially grateful to Dr Elizabeth Pilon-Smits and her group (Colorado State University, Fort Collins) for analyses of selenium concentrations of Stanleya pinnata plants used in host-range tests with Psylliodes isatidis. We would like to thank Dr Maurizio Biondi (Università degli Studi dell’Aquila, Italy) for identification of Alticinae. We are grateful to Dr Dale Woods (Californian Department of Food and Agriculture, USA), Dr John Gaskin (USDA-ARS-NPARL [Agricultural Research Service – Northern Plains Agricultural Research Laboratory], Sidney, Montana, USA), Mary Goshorn (Denver Botanic Gardens, Colorado, USA), and Alison E. Stanton (BMP Eccosciences, South Lake Tahoe, California) who provided us with additional seeds of indigenous NA Brassicaceae. Dr John Gaskin also kindly started investigating the genetic variability of Isatis. In 2011, financial support for this project was provided by the Bureau of Land Management, Idaho, USDA-APHIS-CPHST, the Wyoming Biological Control Steering Committee, and counties of the states of Idaho and Utah.

24

10. References

Boufford, D.E., Freeman, C.C., Gandhi, K., Hill, M.J., Kiger, R.W., Poole, J.M., Schmidt, H.H., Shultz, L.M., Strother, J.L. and Zarucchi, J.L. (2010) Flora of North America: North of Mexico. Vol. 7: Magnoliophyta: Salicaceae to Brassicaceae. Oxford University Press, New York.

Callihan, R.H., Dewey, S.E., Patton, J.E. and Thill, D.C. (1984) Distribution, biology and habitat of dyers woad (Isatis tinctoria) in Idaho. Journal of the Idaho Academy of Science 20(1/2), 18–32.

Feist, L.J. and Parker D.R. (2001) Ecotypic variation in selenium accumulation among populations of Stanleya pinnata. New Phytologist 149, 61–69.

Freeman, J.L., Lindblom, S.D., Quinn, C.F., Fakra, S., Marcus, M.A. and Pilon-Smits, E.A.H. (2007) Selenium accumulation protects plants from herbivory by Orthoptera via toxicity and deterrence. New Phyologist 175, 490–500.

Hanson, B, Garifullina, G.F., Lindblom, S.D., Wangeline, A., Ackley, A., Kramer, K., Norton, A., Lawrence, C. and Pilon-Smits, E.A.H. (2003) Selenium accumulation protects Brassica juncea from invertebrate herbivory and fungal infection. New Phytologist 162, 655–662.

Hawkes, R.B., Whitson, T.D. and Dennis, L.R.J. (1985) A Guide to Selected Weeds of Oregon. Oregon Department of Agriculture, Salem, p. 34.

Hegi, G. (1986) Illustrierte Flora von Mitteleuropa. Spermatophyta, Band IV, Teil 1. Angiospermae, Dicotyledones 2. Paul Parey, Berlin/Hamburg, Germany, pp. 126–131.

Hinz, H.L., Cortat, G., Edelmann, L., Gerber, E. Gundelwein, F. and Schaller, S. (2009). Biological control of dyer’s woad, Isatis tinctoria. Annual Report 2008. Unpublished Report, CABI Europe – Switzerland, Delémont, Switzerland.

Hinz, H.L., Gerber, E. and Svejcar, L. (2010) Biological control of dyer’s woad, Isatis tinctoria. Annual Report 2009. Unpublished Report, CABI Europe – Switzerland, Delémont, Switzerland.

Hinz, H.L., Gerber, E., Joye, S. and Ye, Z. (2011) Biological control of dyer’s woad, Isatis tinctoria. Annual Report 2010. Unpublished Report, CABI Europe – Switzerland, Delémont, Switzerland.

Kropp, B.R., Hansen, D.R., Wolf, P.G., Flint, K.M. and Thomson, S.V. (1997) A study on the phylogeny of the dyer's woad rust fungus and other species of Puccinia from crucifers. Phytopathology 87, 565–571.

McConnell, E.G., Evans, J.O. and Dewey, S.A. (1999) Dyer’s woad. In: Sheley, L. and Petroff, K. (eds) Biology and Management of Noxious Rangeland Weeds. OSU Press, Corvalis, Oregon, pp. 231–237.

Whitson, T.D. (ed) (1987) Weeds and Poisonous Plants of Wyoming and Utah. Cooperative Extension Service, University of Wyoming, Laramie, pp. 56–57.

Wiggin, R.A. (1991). On the woad again. Agricultural Research June 1991, p. 9.

25

Distribution list John L. Baker Bruce Helbig Shawna Bautista Jim Hull (5) Dan Bean Cami King Larry Beneker Clair Kofoed Maurizio Biondi Boris Korotyaev Gary Brown Brian Marschman Tim Butler Christian Maus Larry T. Cain Joseph Milan Craig McClure Anna Owsiak Massimo Cristofaro Adrianne Peterson Lauri Coates Mike Pitcairn Tim Collier Carol Randall Enzo Colonnelli Brett Richardson Eric Coombs Warren Ririe Andy Currah Mark Schwarzländer (2) Alecu Diaconu Bruce Shambaugh Margarita Dolgovskaya Josh Shorb Nina Eckberg John Simons Liz Galli-Noble Sharlene Sing John Gaskin Jeanne Standley Arnie Grammon Janet Valle Levent Gültekin Matt Voile David Hallinan David Weaver Rich Hansen USDA ARS EBCL Rüstem Hayat CABI library (2)

www.cabi.org

KNOWLEDGE FOR LIFE

contact CABI europe CABI Head Office Nosworthy Way, Wallingford, Oxfordshire, OX10 8DE, UK T: +44 (0)1491 832111 CABI Bakeham Lane, Egham, Surrey, TW20 9TY, UK T: +44 (0)1491 829080 CABI Rue des Grillons 1, CH-2800 Delémont, Switzerland T: +41 (0)32 4214870

asia CABI C/o Internal Post Box 56, Chinese Academy of Agricultural Sciences, 12 Zhongguancun Nandajie, Beijing 100081, CHINA T: +86 (0)10 82105692 CABI 2nd Floor, CG Block, NASC Complex, DP Shastri Marg, Opp. Todapur Village, PUSA, New Delhi – 110012, INDIA T: +91 (0)11 25841906 CABI PO Box 210, 43400 UPM Serdang, Selangor, MALAYSIA T: +60 (0)3 89432921 CABI Opposite 1-A, Data Gunj Baksh Road, Satellite Town, Rawalpindi-PAKISTAN T: +92 (0)51 9290132

africa CABI ICRAF Complex, United Nations Avenue, Gigiri, PO Box 633-00621, Nairobi, Kenya T: +254 (0)20 7224450/62

americas CABI UNESP- Fazenda Experimental Lageado, Rua: José Barbosa de Barros, 1780 Botucatu – SP, CEP: 18610-307, BRAZIL T: (14) 3882 - 6300 / 3811 - 7127 CABI Gordon Street, Curepe, TRINIDAD AND TOBAGO T: +1 868 6457628 CABI 875 Massachusetts Avenue, 7th Floor, Cambridge, MA 02139, USA T: +1 617 3954051