256 Protection Quarterly Vol.11 Supplement 2 1996 has fallen into three phases. The first was The use of pathogens native to Europe to control conducted at Montpellier, southern France, where the insects and fungi most in southern damaging to populations of the thistles throughout their native range were iden- R.H. Groves and J.J. Burdon, CSIRO Division of Plant Industry, Co-operative tified. None of the insect species appeared Research Centre for Management Systems, GPO Box 1600, Canberra, ACT to limit slender populations (Sheppard et al. 1991); however, the P. 2601, Australia. cardui-pyncocephali was identified as hav- ing the most potential for biological con- Summary trol. Two of a total of 38 isolates of the Foliar pathogens of European thistles We review the foliar pathogens from Eu- pathogen collected in four European coun- that have become significant in rope that are known to attack the intro- tries were selected for further testing, southern Australia are listed and the po- duced thistle groups that are the worst based on their capacity to inflict greater tential for some of them to be classical weeds in southern Australian pastures. damage on accessions of the thistles col- biological control agents is discussed. These include nodding thistle ( lected in southern Australia than the rela- Particular attention is paid to pathogens nutans), the slender thistles (Carduus tively benign strain(s) already present in of slender, variegated and saffron this- pycnocephalus and C. tenuiflorus), Scotch, Australia (Chaboudez et al. 1993a). These tles. In the case of the slender thistles, the Illyrian and stemless thistles ( isolates were each virulent on a different current status of two recently released acanthium, O. illyricum and O. acaulon re- species of slender thistle; they reduced strains of Puccinia cardui-pycnocephali spectively), variegated thistle ( height, dry weight and seed production is presented. Control programs for vari- marianum), spear thistle () significantly under glasshouse conditions egated and saffron thistles are less ad- and saffron thistle (Carthamus lanatus). (Table 2) and also in a garden experiment. vanced. Results of recent studies of P. The specificity of these two aggressive mariana in Europe have thrown doubt on Thistle pathogens in Mediterranean strains (one for C. tenuiflorus from Italy its suitability as a control agent for vari- Europe and the other for C. pycnocephalus from egated thistle. In contrast, a survey of The pathogens which have been recorded France) was then exhaustively tested to pathogens on saffron thistle has shown P. for the various thistles are listed in Table 1. ensure that economically important spe- sommeriana to have some potential for This table draws on existing information cies (such as artichoke) or species native to biological control. Several pathogens in (S. Hasan personal communication), sup- Australia were not vulnerable to attack. Europe may have potential to add to the plemented by some recent effectiveness of biological control of cer- records of pathogens from Table 1. Fungi recorded on thistles in tain thistles in southern Australian pas- Onopordum and saffron thistle Mediterranean Europe with potential for tures, but only if their effects can be inte- (Groves unpublished). It is ob- biological control (in part from S. Hasan, grated with other methods of control, in- vious that there are many unpublished). cluding biological control by insects. pathogens recorded for thistles in Europe, of which the rusts Fungal species Known hosts Introduction form the largest group; the Ophiobolus cirsii Carduus spp., Cirsium spp. The CSIRO Divisions of Entomology and rusts are also one of the most (Sphaeriales) Plant Industry have been researching this- specific of all groups of fungi. It Phyllosticta cirsii Cirsium spp. tle control for over a decade. The program is not surprising, therefore, that (Coelomycetes) has emphasized classical biological con- research on the potential of trol, using insects and fungi, in relation to fungi for biological control has Puccinia acanthii , the and ecology of the major tended to concentrate on the O. acaulon thistle groups as they occur in both Medi- rusts. P. cardui-pycnocephali , terranean Europe and southern Australia. Carduus tenuiflorus The research has been jointly funded by Control of the slender P. carduorum Carduus spp. (incl. nutans) CSIRO and also, for various periods, by thistles by the rust Puccinia P. carthami Carthamus spp. both the International Wool Secretariat cardui-pycnocephali in P. cirsii Cirsium spp. and the Meat Research Corporation. The southern Australia P. cnici Cirsium spp. program has been centred both in Can- The closely-related taxa P. cruchetiana berra and Montpellier, France, the latter Carduus pycnocephalus and C. P. mariana Silybum marianum being the Mediterranean base of the tenuiflorus both occur as short- P. sommieriana Carthamus lanatus CSIRO European Laboratory. season winter annuals in south- (Uredinales) This paper reviews progress made on ern Australian pastures (Par- Ramularia cardui Carduus spp. the thistle research program as it relates sons 1977, Groves and Kaye R. cirsii Cirsium vulgare to: 1989). Both species have a wide R. onopordi Onopordum acanthium i. the occurrence of certain foliar patho- distribution in Europe and were (Hyphomycetes) gens in Mediterranean Europe, introduced early to Australia. Septoria cardui Carduus nutans ii. the release in southern Australia of two During this introduction stage, Septoria centrophylla Carthamus spp. strains of the pathogen Puccinia cardui- at least one strain of Puccinia Septoria cirsii Cirsium spp. pycnocephali for control of the slender cardui-pycnocephali was also in- Septoria onopordonis Onopordum spp. thistles Carduus pycnocephalus and C. troduced accidentally, but it has Septoria silybi Silybum marianum tenuiflorus, had little effect on the weediness (Coelomycetes) iii.the potential of several other pathogens and spread of slender thistle Ustilago cardui Carduus spp., Cirsium spp., to biologically control populations of populations throughout south- Silybum marianum variegated thistle (Silybum marianum) ern Australia. U. onopordi Onopordum spp. and saffron thistle (Carthamus lanatus) in Research on the biological (Ustilaginales) southern Australia. control of the slender thistles Plant Protection Quarterly Vol.11 Supplement 2 1996 257 The second phase of the program began for which several attempts at Table 2. Effect of Puccinia cardui-pycnocephali in 1992 with the granting of quarantine introduction have been made infection on plant height (cm), dry weight (g) approval to introduce pathogen isolates in Victoria (Bruzzese 1996). and production of viable seeds of the slender into high security quarantine facilities on At one site in southern thistles, Carduus pycnocephalus and C. the CSIRO Black Mountain site, Canberra. Spain, a rust was found on tenuiflorus (from Chaboudez et al. 1993). Host specificity was tested on a total of 46 both variegated thistle and on species, and in September 1993 permission Notobasis syriaca, an annual Control Inoculated was jointly obtained from the Australian thistle similar to and co-occur- Quarantine and Inspection Service and the ring with variegated thistle at C. pycnocephalus (n=20) Australian Nature Conservation Agency the site. This fungus, tenta- Height 87.3 ± 4.5 38.8 ± 2.7 to make field releases (Chaboudez et al. tively identified as Puccinia Dry weight 19.4 ± 1.1 11.3 ± 0.8 1993b). notobasidi, was able to infect Seed production 174 ± 23 32 ± 10 The third phase of the project focused variegated thistle seedlings in C. tenuiflorus (n=21) on field releases of the rust and the con- the field and the glasshouse, Height 52.9 ± 3.7 25.4 ± 4.0 tinuous generation of inoculum in the but always at a level insuffi- Dry weight 16.9 ± 0.8 7.3 ± 0.6 glasshouse to cope with the demand for cient to control seeding. Seed production 84 ± 12 21 ± 8 spores from co-operating agencies and Spores were not carried over Means + standard errors are presented. All pairwise collaborators. Field releases were made from one season to the next. means differ at the 1% level, based on one-way subsequently in all southern Australian In an annual species such as ANOVA. States, albeit at a time of severe drought in variegated thistle, seeding the region. Once the drought broke (in must be significantly reduced and year-to- control biologically because of its close autumn 1995), monitored releases in year survival of the spores must occur if taxonomic affinity to an economically southern New South Wales showed good biological control is to be at all effective. important crop in Australia. Furthermore, pathogen establishment and spread to Subsequently, a further two isolates of surveys over a number of years in France uninoculated thistle stands more than 800 the fungus were obtained—one from N. and Spain of the entomofauna associated m away. At this site disease levels failed to syriaca from Turkey, and one from S. with both safflower and saffron thistle build up sufficiently quickly to prevent marianum from Greece—but the same re- have yielded only one record of a suffi- normal growth or to reduce seeding of in- sult was obtained (Hasan and Groves un- ciently host-specific arthropod that may fected . However, verbal reports published). The same fungus was also con- be able to control the weed species with- suggest that, at some release sites in sistently ineffective on seedlings of S. out affecting in some way the crop species south-western Victoria, the rust has estab- eburneum, the only other species known in (J-P. Aeschlimann personal communica- lished well, has spread by as much as 20 the genus. It is concluded that P. mariana tion). The effectiveness of this insect km and has reduced plant vigour and may be a strain of the wider taxon P. (Botanophilla turcica, a crown fly) is being seeding. Although funding for the project notobasidi. The two fungal taxa are most tested currently (J. Vitou personal com- has formally ceased, new releases con- likely two extremes of the one species, de- munication). Whilst the rust fungi are tinue to be made through Landcare spite having been given different names, sometimes highly specific to taxonomic groups and other collaborators. probably because of the two different variants of a host, no strain of Puccinia We conclude that, except in particularly hosts on which they have been recorded. carthami (Table 1) has yet been identified favourable localities, the main contribu- On the basis of results of our pathogenic- which is specific to the weedy species. Are tion of the recent release of the two strains ity tests, the taxon P. mariana may repre- there other fungi that may fill this appar- of this pathogen will be through a general sent a less aggressive strain of P. notobasidi. ent gap? reduction in the vigour and seed produc- The two Puccinia species are separated on The result of a recent survey commis- tion of the two slender thistle taxa known the basis of spore size. More collections are sioned by CSIRO (Evans, 1995) shows that to occur in southern Australia. needed to determine dimensions of spores Puccinia sommieriana may fulfil some of the collected from the two hosts over a range requirements of a potential biological con- Is Puccinia mariana a potential of site conditions. In conclusion, the trol agent for saffron thistle. Evans sur- biological control agent for chances of finding a strain of this fungus veyed sites in southern Greece where the variegated thistle in southern suitable for biological control of vari- two recognised subspecies of saffron this- Australia? egated thistle in regions such as southern tle and their hybrids overlap at a time Variegated thistle (Silybum marianum) is Australia are slight. when this annual thistle was present widely distributed in Mediterranean Eu- A second fungus, Septoria silybi (Table 1), mainly as a rosette. Evidence of damage to rope and north Africa, where it occurs pri- also occurs commonly on variegated this- young rosette of saffron thistle marily on disturbed and nutrient-enriched tle in Europe as well as in and as a result of infection by P. sommieriana land. The species has been introduced to southern Victoria, where it was appar- was found, despite this taxon not being other regions with a Mediterranean cli- ently accidentally introduced. The fungus listed among the fungi recorded from mate, such as California, Chile and south- commonly occurs on older rosette leaves Greece. Little is known about the biology ern Australia where it is a weed of grazing only. Consequently, it hastens senes- of this pathogen and its biology, espe- lands. cence, but has little effect on plant growth cially its possible effect on reduction of While based at the CSIRO European or seed production (Moscow and Lindow seeding. Laboratory, Montpellier, Groves investi- 1989). Thus, it does not seem a suitable Evans also found Septoria centrophylla to gated populations of variegated thistle in candidate for biological control of vari- be damaging on saffron thistle in Greece, Spain and southern France for the inci- egated thistle. although he considered ‘that the actual dence of the pathogens listed in Table 1, host range may turn out to be wider than especially Puccinia mariana. We considered Is control of Carthamus lanatus by that of the rust Puccinia sommieriana’ that the effects of this rust may comple- fungi possible? (Evans 1995). These two pathogens seem ment those of a host-specific strain of the Saffron thistle (Carthamus lanatus) is closely to be the only possible candidates from seed-eating , related to the crop congener, safflower Europe for biological control against saf- which had already been introduced to (C. tinctorius). It is the thistle that will be fron thistle in Australia. Given our limited California (Goeden and Ricker 1977) and most difficult (and/or most expensive) to knowledge of pathogen biology and the 258 Plant Protection Quarterly Vol.11 Supplement 2 1996 economic importance of saffron thistle, to biological control. Plant Protection these pathogens warrant further study. Quarterly 11, 245-9. Chaboudez, P., Groves, R.H. and Burdon, Conclusions J.J. (1993a). Puccinia cardui-pycnocephali, Host-specific fungal pathogens occurring a potential agent for the biological con- in Europe may be effective biological trol of slender thistles in Australia. Pro- agents that contribute to the control of at ceedings of the 10th Australian Weeds least some of the major thistles introduced Conference, Volume 1, pp. 84-8. to southern Australia. Slender thistle Chaboudez, P., Burdon, J.J. and Groves, growth in Mediterranean Europe was R.H. (1993b). Application for release of more limited by the incidence of the rust two strains of Puccinia cardui- than it was by the occurrence of more than pycnocephali, a rust fungus for the classi- 20 species of insects (Sheppard et al. 1991). cal biological control of the noxious slen- Though it is still too early to predict the der thistle weeds, Carduus pycnocephalus effectiveness of two strains of that same and C. tenuiflorus. CSIRO Division of fungus on slender thistle populations in Plant Industry, Internal Report, 13 pp. southern Australia, it seems probable that Evans, H.C. (1995). Report on a survey for release of the rust will be insufficient on its fungal pathogens of Carthamus lanatus own to control seeding of these annual (saffron thistle) in Greece. Report of thistles in any but the most favourable of Project CAR010, International Institute environments. Such a conclusion may ap- of Biological Control, Ascot, UK, pp. 1-4. ply even more appropriately to the con- Goeden, R.D. and Ricker, D.W. (1977). Es- trol of other thistle species. tablishment of Rhinocyllus conicus on Bendall (1973) found that a deferral of milk thistle in southern California. Weed autumn grazing of pasture containing Science 25, 288-92. slender thistle was an effective control Groves, R.H. and Kaye, P.E. (1989). Ger- method, a result subsequently shown for mination and phenology of seven intro- some other thistles in pasture in other situ- duced thistle species in southern Aus- ations. A combination of deferred au- tralia. Australian Journal of Botany 37, tumn-early winter grazing with applica- 351-9. tion of an appropriate level of herbicide in Moscow, D. and Lindow, S.E. (1989). Infec- early spring may achieve a greater level of tion of milk thistle (Silybum marianum) control. If such a combined treatment is leaves by Septoria silybi. Phytopathology compatible with the continued effects of 79, 1085-90. rust infection and the establishment of Parsons, W.T. (1977). The ecology and perennial pasture grasses, then even fur- physiology of two species of Carduus as ther control may be achieved and the weeds of pastures in Victoria. Ph.D. the- weediness of at least one group of thistles sis, University of Melbourne. reduced thereby. The challenge to re- Sheppard, A.W., Aeschlimann, J-P., search is to demonstrate the effectiveness Sagliocco, J-L. and Vitou, J. (1991). Natu- of such integrated control methods in the ral enemies and population stability of field not only for slender thistle but also the winter-annual Carduus pycnocephalus for other thistle groups with either annual L. in Mediterranean Europe. Acta or biennial life-cycles. Oecologica 12, 707-26. In conclusion, there is a role for im- ported fungal pathogens in thistle control, but only in an integrated weed manage- ment system. Pathogens operating in iso- lation are likely to be ineffective at control- ling thistle populations in southern Aus- tralia.

Acknowledgments We acknowledge the help of Dr. S. Hasan, CSIRO Division of Entomology, in compil- ing the initial listing used in Table 1. Some of the work reported was supported by the Australian woolgrowers through the IWS.

References Bendall, GM. (1973). The control of slender thistle, Carduus pycnocephalus L. and C. tenuiflorus Curt. (Compositae), in pas- ture by grazing management. Austral- ian Journal of Agricultural Research 24, 831-7. Bruzzese, E. (1996). Ecology of Cirsium vulgare and Silybum marianum in relation