B3   >>   BCANZ home   ·   Search database   ·   Browse database

Biocontrol introduction

Target pest: Cirsium arvense (Asterales: Asteraceae), Californian thistle

Agent introduced: Cassida rubiginosa (Coleoptera: Chrysomelidae), green thistle beetle, thistle tortoise beetle



Import source:


Import notes:

Gourlay (2010b) - Cassida rubiginosa, which has a wide native range including Europe, eastern Mediterranean, North Asia and North Africa, was first imported into New Zealand in 2006 by Landcare Research on behalf of the Californian Thistle Action Group.



Release details:

Gourlay (2010b) - released in November 2007 in the South Island as follows: 50 at Waitahuna West near Lawrence, Otago, 50 at Glenomaru near Owaka, South Otago, 50 at Caberfeigh, Catlins, South Otago and 50 in Hokonui Hills near Browns, Central Southland.

Landcare Research (2015b) - earliest releases were in Southland/Otago, but now widely released throughout New Zealand, including 65 beetles near Masterton (Wairarapa) in 2008 (with recently 30,000 collected and redistributed from this site) and 120 beetles in Manawatu-Whanganui in 2008 (with recently collections from this site distributed to 16 further sites in the region) in the North Island.

Landcare Research (2015j) - five releases in 2014-15.

Cripps et al. (2022) - since C. arvense remains one of the most economically damaging pasture weeds in New Zealand, there is high demand for C. rubiginosa, especially for hill-country pasture where spraying herbicide is not practical or economical. Currently, it is redistributed from successful field sites, or reared on host plants in cages and then released at field sites. Trials were undertaken to develop a semi-artificial diet for C. rubiginosa, which would allow the mass-production of the beetle for release. Although no larvae completed development on the diets, some important progress was achieved, providing proof of concept that such a diet can be developed.

Bourdôt et al. (2024) - first released in Southland, with more recent releases in at least seven other regions in New Zealand.


Landcare Research (2015b) - early reports suggest 90% establishment at original release sites, and good establishment at later releases in Wairarapa (where there has been a huge outbreak) and Manawatu-Wanganui.

Bourdôt et al. (2024) - sixteen years after the first releases of C. rubiginosa in New Zealand it is widely established. A total of 627 occurrence records from various sources included 191 separate locations where the beetle has been recorded as having established a persistent population on C. arvense.

Impacts on target:

Landcare Research (2013f) - small beetle population at Lincoln, Canterbury caused striking damage in spring/summer 2012-13, eating all green leaf tissue on some plants by late-December.

Landcare Research (2015b) - heavy defoliation of thistles in some regions, plants demolished in Wairarapa.

Landcare Research (2018j) - since release in 2007 C. rubiginosa has been starting to make good progress towards reducing C. arvense populations in many areas. A trial underway in North Canterbury since 2015 [see Cripps et al. (2019) entry below] indicates that 10 or more larvae per shoot are required to significantly reduce the density and spread of Californian thistle, with shoots incurring greater feeding damage less likely to become reproductive. The trial has also suggested that rainfall is an important factor in determining the level of damage to the thistle, with water-stressed plants more susceptible to the beetle. This might explain why C. rubiginosa has not performed well to date in the south where rainfall is higher.

Cripps et al. (2019) - observational evidence suggests C. rubiginosa is an effective defoliator of C. arvense, occasionally killing individual shoots, in many areas of New Zealand. To provide an experimental assessment of the impact of the beetle, shoot population density and spread was assessed over two years (Nov 2015 - Nov 2017) on sheep-grazed pasture in North Canterbury under typical farm management conditions with an established population of C. arvense. Folivory in the 10 and 20 larvae per shoot treatments caused C. arvense population declines of 29% and 75%, respectively, despite low levels of folivory being achieved in both years (maximum average of 25% folivory). This study shows that average densities of ten or more larvae per shoot can reduce population density and local vegetative spread of C. arvense under typical pasture management.

Cripps et al. (2022) - a controlled field evaluation showed that at least 10 C. rubiginosa larvae per shoot was necessary to reduce the density and local spread of C. arvense under typical sheep-grazed pasture management [see Cripps et al. (2019) entry above]. However, most populations of C. rubiginosa in New Zealand appear to persist at densities less than 10 larvae per shoot, with only occasional outbreak populations exceeding this.

Bourdôt et al. (2024) - the impact of C. rubiginosa on C. arvense in New Zealand has been sporadic. The first observations of high impact were made at two sites in the Wairarapa region of the North Island in 2014. A visit to these sites in late-summer 2017 showed the beetle had completely defoliated all shoots in the thistle populations and had also removed the mesophyll tissue from the stems of the defoliated shoots. However, observations at many other locations where the beetle has established have revealed little or no impact. At 191 geographically separate C. rubiginosa populations its impact on C. arvense has been assessed as ‘none’ at 29 sites (15%),’low’ at 22 sites (12%), ‘high’ at 13 sites (7%) and ‘unknown’ at 127 sites (66%). Models show that the climate throughout most of New Zealand is variably suitable for both species, although everywhere relatively less suitable for the beetle. However, the impact of the assessed populations showed no relationship with the modelled climate suitability, suggesting that the climate suitability bias favouring the thistle does not explain the sporadic impact of the C. rubiginosa. Other possible explanations are absence of suitable overwintering habitat for the beetle, predation, incompatible thistle control operations, and variable vigour of the thistle as affected by variation in soil fertility, grazing and interspecific plant competition.

Paynter (2024) - factors influencing the success of weed biocontrol agents released and established in New Zealand were investigated. Each agent’s impact on the target weed in New Zealand was assessed as ‘heavy’, ‘medium’, ‘variable’, ‘slight’ or ‘none’, where a ‘heavy’, ‘medium’ or ‘variable’ impact have all been observed to reduce populations or percentage cover of their target weed in all or part of their respective target weed ranges in New Zealand. Results showed that: (i) agents that are highly damaging in their native range were almost invariably highly damaging in New Zealand; (ii) invertebrate agents with a closely related ‘native analogue’ species are susceptible to parasitism by the parasitoids that attack their native analogues and failed to have an impact on the target weed, and (iii) agent feeding guild helped predict agent impact - in particular, agents that only attack reproductive parts of the plant (e.g., seed and flower-feeders) are unlikely to reduce weed populations. Damaging impacts of C. rubiginosa, a defoliating beetle, have been reported in its native range, it does not have a New Zealand native ecological analogue and its impact in New Zealand is assessed as ‘variable’.

Impacts on non-targets:

Hill (2006a) - field records in both its native Europe and in North America where it has been established for over 100 years, along with laboratory host range tests, indicate C. rubiginosa hosts are restricted to the tribe Cardueae in the family Asteraceae. New Zealand has no native species in this tribe and almost all imported thistles and knapweeds in this tribe have the potential to become weedy once established in the New Zealand environment. Most if not all species in subtribe Carduinae, and perhaps most in subtribe Centaureinae, appear to be suitable host plants for oviposition and larval development of C. rubiginosa in the field in Europe. In European field cage tests minor adult feeding and a very small number of eggs were recorded on plants in the subtribes Echinopsidinae, Carlininae, and Centaureinae. Larval starvation tests showed that larvae were capable of developing to adult on these plants. Centaurea cyanus supported full development of 80% of larvae tested, and adults could lay eggs on this plant. These results suggest that plants in this subtribe (Centaureinae), including safflower (Carthamus tinctoria), would be capable of supporting populations of green thistle beetles once established in New Zealand.

Landcare Research (2013f) - damage at Lincoln observed on scotch thistle (Cirsium vulgare), though to a much lesser degree than on C. arvense at the same site.

Paynter et al. (2015) - surveys of potential non-target hosts Carduus nutans (nodding thistle), Cirsium vulgare (Scotch thistle) and Cynara scolymus (globe artichoke) show the former two (both exotic weeds) are 'full' hosts (can support breeding populations), and minor 'spillover' feeding on the latter.

Cripps et al. (2015) - in a laboratory trial, survival of C. rubiginosa (first instar larva to adult) was assessed on the beetle’s primary host, C. arvense, and two alternative congeneric hosts C. palustre [marsh thistle] and C. vulgare [Scotch thistle]. Survival of was strongly dependent on plant species with mean survival rates of 47%, 16%, and 8%, respectively, and was primarily explained by leaf trichome [hair] densities and to a lesser extent by specific leaf area [a measure of leaf thickness and density].

Cripps et al. (2106) - in a laboratory trial, survival of C. rubiginosa (first instar larva to adult) was assessed on 16 species of the Cardueae tribe (thistles and knapweeds). The survival of C. rubiginosa decreased with increasing phylogenetic distance from the known primary host, C. arvense, suggesting that its host range, to a large degree, is constrained by evolutionary history. The only trait measured that clearly offered some explanatory value for the survival was specific leaf area. This trait was not phylogenetically dependent [and can vary with environmental conditions], and when combined with phylogenetic distance from C. arvense gave the best model explaining C. rubiginosa survival. The phylogenetic pattern of C. rubiginosa fitness will aid in predicting its ability to control multiple Cardueae weeds in New Zealand.

Paynter et al. (2018) - minor spillover feeding on Cynara scolymus [see Paynter et al. (2015) entry above] was an anticipated risk considered acceptable by regulatory authorities when approval for release was granted.

Hettiarachchi et al. (2018) - adult C. rubiginosa have been observed to feed and oviposit on field populations of nodding thistle, Scotch thistle, winged thistle (Carduus tenuiflorus), marsh thistle (Cirsium palustre), and artichoke. Utilisation of these plants was predicted by pre-release choice and no-choice host testing; however, the potential impact on individual plant, or population performance, is uncertain. A potted plant trial showed C. rubiginosa had minimal impact on the performance of marsh thistle. However, under field conditions, where plants are likely to encounter additional stressors, and with the opportunity to attack smaller rosettes, the beetle may be able to impact this thistle.

Mills et al. (2020) - a trial using potted Carduus pycnocephalus (slender winged thistle) plants showed feeding damage from larvae of C. rubiginosa did not affect the growth or reproductive performance of this plant, regardless of pot size. (The constrained growth conditions of smaller pot sizes was used to mimic stresses plants may encounter in the field.) The maximum feeding damage reached was approximately 50% defoliation in the highest larval density treatment. The results indicate that feeding damage from C. rubiginosa is unlikely to contribute to control of annual thistle weeds such as C. pycnocephalus, even if their growth is constrained.

Hettiarachchi et al. (2023) - the host range of C. rubiginosa was investigated through a series of experiments with adult beetles and 16 potential host plant species in the tribe Cardueae (thistles and knapweeds). Initially, the olfactory recognition (single choice tests) and preference (dual choice tests) of the beetle was tested, and then acceptance (no-choice tests) and preference (dual choice tests) through feeding and oviposition experiments. The olfactory recognition (single odour) and preference (two odours) of the beetle showed a significant phylogenetic relationship, with attraction and preference decreasing with increasing phylogenetic distance from the primary host, Cirsium arvense. In the feeding and oviposition trials, under no-choice conditions there was no phylogenetic pattern to host plant acceptance. However, under choice conditions, phylogenetic distance from C. arvense was a strong predictor of feeding and oviposition preference. These results indicate that host plant utilisation by C. rubiginosa in New Zealand will be mostly restricted to Cirsium and Carduus species (as it is in its native European range), with minimal potential for impact on other Cardueae weeds, despite the fundamental host range (as indicated by the no-choice feeding tests) including a wider range of Cardueae species.

EPA Applications:

EPA (2007b) - 10 Nov 2006: application by the Californian Thistle Action Group (CalTAG) to import for release a weevil, Ceratapion onopordi (Brentidae) and a beetle, Cassida rubiginosa (Chrysomelidae), for the biological control of the weed Californian thistle (Cirsium arvense). EPA Application #NOR06005, approved without controls 17 Apr 2007.


Bourdôt GW, Lamoureaux SL, Cripps MG, Kriticos DJ, Noble A, Kriticos JM (2024). The climatic suitability of New Zealand for Cirsium arvense and its biological control agent Cassida rubiginosa. Biological Control, Volume 188, January 2024, Article Number 105436 https://doi.org/10.1016/j.biocontrol.2023.105436

Cripps M, Mills J, Villamizar L, van Koten C (2022). Initial test of a semiartificial diet for the thistle biocontrol beetle, Cassida rubiginosa. New Zealand Plant Protection 75: 25-30 https://journal.nzpps.org/index.php/nzpp/article/view/11758/11609

Cripps MG, Jackman SD, Roquet C, van Koten C, Rostás M, Bourdôt GW, Susanna A (2016). Evolution of Specialization of Cassida rubiginosa on Cirsium arvense (Compositae, Cardueae). Frontiers in Plant Science, Vol. 7 https://doi.org/10.3389/fpls.2016.01261

Cripps MG, Jackman SD, Rostás M, van Koten C, Bourdôt GW (2015). Leaf traits of congeneric host plants explain differences in performance of a specialist herbivore. Ecological Entomology 40(3): 237-246 https://doi.org/10.1111/een.12180

Cripps MG, Jackman SD, van Koten C. (2019). Folivory impact of the biocontrol beetle, Cassida rubiginosa, on population growth of Cirsium arvense. BioControl 64: 91–101 https://doi.org/10.1007/s10526-018-09915-z

EPA (2007b). Application to EPA (NOR06005) to import and release two insects, Ceratapion onopordi (Brentidae) and Cassida rubiginosa (Chrysomelidae), for biological control of the weed Californian thistle (Cirsium arvense). Environmental Protection Authority website https://www.epa.govt.nz/database-search/hsno-application-register/view/NOR06005

Gourlay H (2010b). Green thistle beetle: Cassida rubiginosa. The Biological Control of Weeds Book - Te Whakapau Taru: A New Zealand Guide (Landcare Research) https://www.landcareresearch.co.nz/discover-our-research/biodiversity-biosecurity/weed-biocontrol/projects-agents/biocontrol-agents/green-thistle-beetle/

Hettiarachchi D, Cripps M, Jackman S, van Koten C, Sullivan J, Rostás M. (2018). Impact of the biocontrol beetle, Cassida rubiginosa, on the secondary weed target, marsh thistle (Cirsium palustre). New Zealand Plant Protection 71: 66-71

Hettiarachchi DK, Rostás M, Sullivan JJ, Jackman S, van Koten C, Cripps MG (2023). Plant phylogeny determines host selection and acceptance of the oligophagous leaf beetle Cassida rubiginosa. Pest Management Science, published online 15 July 2023

Hill R (2006a). Application to EPA (NOR06005) to import and release two insects, Ceratapion onopordi (Brentidae) and Cassida rubiginosa (Chrysomelidae), for biological control of the weed Californian thistle (Cirsium arvense). Environmental Protection Authority website https://www.epa.govt.nz/assets/FileAPI/hsno-ar/NOR06005/9287e22698/NOR06005.pdf

Landcare Research (2013f). Beetles decimate Californian thistles at Lincoln. What's new in biological control of weeds? 64: 4 http://www.landcareresearch.co.nz/publications/newsletters/biological-control-of-weeds/issue-64

Landcare Research (2015b). First major green thistle outbreak. What's new in biological control of weeds? 72: 4-5 http://www.landcareresearch.co.nz/publications/newsletters/biological-control-of-weeds/issue-72

Landcare Research (2015j). Biocontrol agents released in 2014/15. Weed Biocontrol: What's New? 73: 2 http://www.landcareresearch.co.nz/publications/newsletters/biological-control-of-weeds/issue-73

Landcare Research (2018j). How effective is the green thistle beetle? Weed Biocontrol: What's New? 86, November 2018. https://www.landcareresearch.co.nz/publications/newsletters/biological-control-of-weeds/issue-86/how-effective-is-the-green-thistle-beetle

Mills J, Jackman S, van Koten C, Cripps M. (2020). The leaf-feeding beetle, Cassida rubiginosa, has no impact on Carduus pycnocephalus (slender winged thistle) regardless of physical constraints on plant growth. New Zealand Plant Protection 73: 49–56 https://doi.org/10.30843/nzpp.2020.73.11722

Paynter Q (2024). Prioritizing candidate agents for the biological control of weeds. Biological Control, Volume 188, January 2024, Article Number 105396 https://doi.org/10.1016/j.biocontrol.2023.105396

Paynter Q, Fowler SV, Groenteman R. (2018). Making weed biological control predictable, safer and more effective: perspectives from New Zealand. BioControl 63: 427-436 (first published online 8 Aug 2017) https://doi.org/10.1007/s10526-017-9837-5 https://link.springer.com/article/10.1007/s10526-017-9837-5

Paynter QE, Fowler SV, Gourlay AH, Peterson PG, Smith LA and Winks CJ (2015). Relative performance on test and target plants in laboratory tests predicts the risk of non-target attack in the field for arthropod weed biocontrol agents. Biological Control 80: 133-142 https://doi.org/10.1016/j.biocontrol.2014.10.007