Mycol. Res. 107 (7): 797–802 (July 2003). f The British Mycological Society 797 DOI: 10.1017/S0953756203008025 Printed in the United Kingdom. Puccinia tanaceti : specialist or generalist ? George NEWCOMBE Department of Forest Resources, University of Idaho, Moscow, ID 83844-1133, USA. E-mail : georgen@uidaho.edu Received 17 February 2003; accepted 30 April 2003. Rust is reported for the first time in North America on tansy (Tanacetum vulgare), a widespread and invasive Eurasian plant that was introduced into North America in the 17th century. Morphologically, the rust fungus corresponds exactly to Gäumann’s description of Puccinia tanaceti, and to European specimens of P. tanaceti on T. vulgare. An inoculation study confirmed the narrow host specialization of P. tanaceti that Gäumann described, in that T. vulgare was successfully inoculated whereas four other species of the tribe Anthemidae of the Asteraceae were entirely resistant, i.e. Leucanthemum maximum (Shasta daisy), L. vulgare (ox-eye daisy), Artemisia absinthium (wormwood), and A. tridentata (sagebrush). This confirmation of a specialized P. tanaceti invalidates P. tanaceti s. lat. The latter complex has resulted from synonymies, and compilation of host records of rust fungi resembling P. tanaceti on at least 105 species representing 19 genera from six tribes of the Asteraceae from both New and Old Worlds. INTRODUCTION Do some rust fungi have host ranges broad enough that they could be considered ‘generalists ’ ? This question may seem unusual since rust fungi are generally thought to be narrowly specialized parasites of plants. Nevertheless, apparent ‘generalists ’ have resulted from the practice of compilation of host range from host records in herbaria. One apparent generalist is Puccinia tanaceti. Records compiled by the US National Fungus Collection show that P. tanaceti s. lat. has become a rust fungus that one could expect to find on any species in the tribe Anthemidae of the Asteraceae (Farr et al. 2002). Among the 108 genera of the Anthemidae are Tanacetum, Artemisia, Achillea, Chrysanthemum, and Anthemis itself ; the species total for the tribe is at least 1741 (Bremer & Humphries 1993). Morphology-based determinations of P. tanaceti on many species in the Anthemidae and other tribes of the Asteraceae were largely sanctioned by Cummins (1978) who synonymized P. chrysanthemi and P. absinthii with P. tanaceti, even though others (e.g. Punithalingam 1968) had not. Arthur (1962) had earlier synonymized P. similis with P. absinthii, and Wilson & Henderson (1966) added P. artemisiella to the synonymy and simply united the rusts on Artemisia and Chrysanthemum since ‘there are not clear morphological differences between them. ’ The process of indirect compilation of host range has even spilled beyond the boundaries of Anthemidae because records of P. tanaceti on five other tribes of the Asteraceae (i.e. Astereae, Helenieae, Heliantheae, Lactuceae and Vernonieae) are common in the US National Fungus Collection (BPI). The implications of this process of indirect compilation of host range are that P. tanaceti s. lat. possesses a host range that could potentially total in the thousands of host species. However, there is an opposing view of P. tanaceti. Based on direct, host-range determinations or inoculations, Mayor (1946) and Gäumann (1959) believed P. tanaceti to be narrowly specialized. Mayor and then Gäumann performed experiments in which P. tanaceti s. str. was successfully inoculated onto its type host, the Eurasian Tanacetum vulgare. At the same time, inoculations of T. balsamita, Chrysanthemum cinerariaefolium, C. corymbosum, C. inodorum, C. parthenium and C. segetum resulted in no rust (i.e. these species appeared resistant). T. vulgare, the type host of P. tanaceti s. str., possesses a native distribution in Europe and western Asia (Bremer & Humphries 1993). But through deliberate and accidental introduction, T. vulgare has become one of the most widespread plants in the world. In North America it is found across Canada and in all of the 48 contiguous states except Texas and a few south-eastern states. It was deliberately introduced in gardens of the 798 Puccinia tanaceti Table 1. Summary of the host range of Puccinia tanaceti s. lat. in the world, according to the ‘Fungus-Host Distributions’, and ‘Specimens* ’ databases of the US National Fungus Collection. Host Tribe Genera (no. of species on which P. tanaceti s. lat. has been collected) Anthemidae Anthemidae Anthemidae Anthemidae Ajania (1) Anthemis (1*) Argyranthemum (3) Artemisia (46+10*) Anthemidae Anthemidae Anthemidae Anthemidae Anthemidae Anthemidae Chrysanthemum (5) Dendranthema (3) Matricaria (1) Gonospermum (3) Sphaeromeria (1*) Tanacetum (2+1*) Astereae Helenieae Heliantheae Heliantheae Haplopappus (1*) Psilostrophe (Riddellia) (1*) Franseria (1*) Helianthus (19*) Heliantheae Heliantheae Heliantheae Lactuceae Vernonieae Heliopsis (1*) Hymenoxys (Actinella) (1*) Verbesina (1*) Lactuca (1*) Vernonia (2*) Location of rust collection Old World New World China Hungary Canary Islands Finland, Poland, UK, South Africa, China, Greece Idaho, Oregon, Canada, Montana, North Dakota, Wyoming, Washington, Arizona, Colorado, Texas, Utah China Finland Canary Islands UK, Germany, Poland, Russia, Czechoslovakia, Austria, Denmark, Finland, France, Hungary, Italy, Norway, Poland, Romania, Sweden, former USSR Czechoslovakia, Germany, Hungary, Romania, Sweden, former USSR, Yugoslavia, Italy colonial settlers in the 17th century, or perhaps earlier, but it is now considered weedy and invasive throughout its range in North America. Records of the US National Fungus Collection show that P. tanaceti is common on T. vulgare, but only in its native range in Europe. In North America rust has not been seen on T. vulgare for the better part of four centuries. In the past two years, uredinia and telia have been observed on leaves and stems of T. vulgare in parts of Idaho. This first report of tansy rust in North America provoked a broader study of opposing views of P. tanaceti as a specialist (i.e. P. tanaceti s. str.) or generalist (i.e. P. tanaceti s. lat.). Wyoming Idaho (on T. vulgare, this report, BPI 842087) Colorado Arizona Arizona Canada, Iowa, Ohio, Kansas, Illinois, Kentucky, Louisiana, Montana, Texas, Washington, Idaho, Vermont, N. & S. Dakota, Indiana, DC, New York, Nebraska, Massachusetts, New Jersey, South Carolina, Virginia, California, Minnesota, Connecticut Minnesota Wyoming DC New Mexico DC, Illinois, Iowa, Nebraska, Indiana Surveys During Sept.–Oct. 2001 and 2002, Tanacetum vulgare was examined for rust in a scattering of locations across the Pacific Northwest : (1) around Priest Lake (northern panhandle of Idaho) ; (2) the Palouse Range (northcentral Idaho) ; (3) along the Clearwater River and its tributaries (north-central Idaho) ; (4) in Puget Sound (near Bellingham, Washington) ; and (5) on Vancouver Island (random sites between Nanaimo and Campbell River, British Columbia). A limited survey was also conducted in Spring 2003. Morphology MATERIALS AND METHODS Host-range records The databases of the US National Fungus Collection (Farr et al. 2002) provided all records compiled in Table 1. Uredinia and telia of the following PUR specimens were examined with light (LM) and scanning electron microscopy (SEM) : (1) Puccinia tanaceti : N2484, N2485, N2486, N58, F16540, F8444, F8445, F8447, F8448, F8449, F8446, F8451, F8450, F17428 ; (2) P. similis : 88903, 48590, 61010, 37681, 37686, 37687, 37689, N2487, 37750, 88915, 88698, 88697, 37799, G. Newcombe 37801, 37802, 37803, 88694, 88916, 37813, 37814, 88918, N2488, 37823, 88695, 63669, 88696, 63668, 88923, 37879, 88920, 88921, 37882; and (3) P. balsamitae : F17845, F8435, F8436, F8437, F8438. Idaho specimens of P. tanaceti, including BPI 842087, were examined in the same manner. 799 of P. tanaceti s. lat. Indirect compilation has thus resulted in a host range of 105 species representing 19 genera from six tribes of the Asteraceae. Surveys Since uredinia and telia were first seen on Tanacetum vulgare in August, the Clearwater River populations of the host were examined for spermogonia and aecia every two weeks through June, July and August in 2001 and 2002, and in May 2003. Tansy plants inoculated in the growth cabinet (see below) were also examined for spermogonia and aecia. During Sept.–Oct. in both 2001 and 2002, a rust corresponding to Puccinia tanaceti s. str. was found on Tanacetum vulgare around Priest Lake, and along the Clearwater River and its tributaries ; a specimen was deposited in the US National Fungus Collection (BPI 842087). Rust was not observed on T. vulgare in the Palouse Range, nor in the Puget Sound region, nor on Vancouver Island. In May 2003, uredinia of P. tanaceti s. str. were observed on one plant of T. vulgare in Lapwai Canyon. Direct host-range determination Morphology Seeds of Tanacetum vulgare were obtained from plants growing in the riparian zone of the Clearwater River in Idaho. Seeds of Leucanthemum maximum (Shasta daisy), L. vulgare (ox-eye daisy), Artemisia absinthium (wormwood), and A. tridentata (sagebrush) were obtained from Seeds Trust (Hailey, ID). The first four species are of European origin; only A. tridentata is native to North America. Eight plants of Tanacetum vulgare were grown in pots in a greenhouse room as the susceptible control ; 50 plants of each of the other four species were grown interspersed in the same room. Artemisia absinthium was included because it is the type host of Puccinia absinthii that Cummins (1978) synonymized with P. tanaceti. A. tridentata is the most widespread native species of the Anthemidae in the vicinity of the rust-infected T. vulgare of the Clearwater River. Sagebrush might thus be hypothesized to be the source of rust infecting tansy in this region. Some pustules were purely uredinial or purely telial, but many pustules appeared to be mixed (Fig. 1), whether collected in the field in Idaho, or among the greenhouse-inoculated plants of Tanacetum vulgare. Some pustules contained a few mesospores (one-celled teliospores). The ratio of teliospores to urediniospores in the mixed pustules commonly varied from 6 :1 to 30 :1. Relatively pure uredinia were frequently at the center of a circle of telia. Ellipsoid telia were oriented so that their long axes parallelled those of stems and central veins of leaves. Telia were often confluent forming long, dark brown streaks on stems. Urediniospores were broadly ellipsoid, and averaged 26r21 mm. Teliospores varied from 30r21 to 40r 20 mm, but averaged 35r21 mm. Cell wall thickening at the apices of teliospores averaged 7–7.5 mm in width. Ornamentation of urediniospores and teliospores of the Idaho specimens (Fig. 2) appeared to be the same, in the SEM, as that of European specimens (Fig. 1) of Puccinia tanaceti s. str. on T. vulgare. Gentle pits and ridges of teliospores were somewhat variable but spore apices did not appear to be distinguishable from proximal portions of teliospores in ornamentation. In contrast, teliospores of Puccinia similis on Artemisia rigida (Fig. 3) and of P. balsamitae on T. balsamita (Fig. 4) were distinctively ornamented. In the latter two cases, apices were distinct from proximal portions of teliospores, unlike P. tanaceti s. str. Urediniospores of P. tanaceti s. str. (Fig. 1) are very similar to those of native P. similis (Fig. 5) in the SEM, and in the LM (Fig. 6). In the LM, the general appearance of teliospores of P. tanaceti s. lat. are all much like those of P. tanaceti s. str. (Fig. 7), although there may be significant differences in size. For example, teliospores of P. similis (Fig. 3) appear to be larger than those of P. tanaceti s. str. (Fig. 1). Life cycle Inoculations The first inoculation was conducted at the end of August with inoculum from rust-infected Tanacetum vulgare plants from along the Clearwater River. When uredinia and telia developed on inoculated seedlings of T. vulgare, a repeat inoculation was performed with inoculum from the first experiment. Inoculum was applied with a moistened, cotton swab to plants that had been lightly misted beforehand. Inoculated plants were then enclosed in plastic bags overnight so that plant surfaces would stay moist. RESULTS Host-range records The databases of the US National Fungus Collection (Farr et al. 2002) provided records of Puccinia tanaceti s. lat on 77 species among ten genera of the Anthemidae (Table 1). Another 28 species from nine genera from five additional tribes have also been recorded as hosts Life cycle Spermogonia and aecia were neither observed in the field, nor on the inoculated plants of Tanacetum vulgare. 800 Puccinia tanaceti Figs 1–4. Figs 1–2. SEM images of Puccinia tanaceti s. str. in Europe and Idaho. Bars=10 mm. Fig. 1. On Tanacetum vulgare (PUR F17428; Bohemia, Czech Republic, 24 Sept. 1944). Mixed pustule of teliospores and echinulate urediniospores. Teliospore apices are not striate. Fig. 2. On T. vulgare (BPI 842087; Lapwai Canyon, Idaho, 28 Oct. 2001). ‘ Pure’ telium comprising teliospores only, at lower magnification than Fig. 1. Figs 3–4. SEM images of species that are part of the P. tanaceti s. lat. complex. Fig. 3. P. similis on native, North American Artemisia rigida (PUR 88694; Gingko Petrified Forest State Park, Washington, 5 Sept. 1969). Fig. 4. P. balsamitae on Eurasian Tanacetum balsamita (PUR F8436; Berlin, 1889). Their absence was particularly noteworthy in the May 2003 collection that was entirely uredinial. In 2001 and 2002, uredinia, telia, and mixed pustules appeared first in August. In contrast, uredinia of Puccinia similis, appeared on native Artemisia tridentata by June (Fig. 5). Direct host-range determination In both the first and second inoculations, rust appeared on all inoculated Tanacetum vulgare, but not on any of the plants of Leucanthemum maximum (Shasta daisy), L. vulgare (ox-eye daisy), Artemisia absinthium (wormwood), and A. tridentata (sagebrush). The latent period appeared to be relatively long (approx. 30 d). Pustules appeared first on stems, then on the oldest leaves of that stem, and then on progressively younger leaves. DISCUSSION Idaho specimens correspond closely to European specimens of Puccinia tanaceti on Tanacetum vulgare in measurements and teliospore ornamentation (Figs 1 and 2), and to Gäumann’s (1959) description. According to Cummins (1978), teliospores of P. tanaceti are ‘striate at apex’, a character state that was not apparent in Idaho and European specimens from the type host in the SEM (Figs 1 and 2). In the LM, the apices of some teliospores of P. tanaceti s. str. might be construed as striated (Fig. 7), but this was never confirmed in the SEM. Cummins (1978) also stated that teliospores of P. tanaceti are ‘(36–)40–56(–60)r(19–)22–28(–30) mm ’, whereas Gäumann (1959) provided a narrow range with most teliospores around 35r19 mm, just as in G. Newcombe 801 Figs 5–7. Fig. 5. Puccinia similis on native, North American Artemisia tridentata subsp. tridentata (PUR 88696; 30 June 1953). Bar=20 mm. Figs 6–7. LM images of P. tanaceti s. str. on T. vulgare (BPI 842087; Lapwai Canyon, Idaho, 28 Oct. 2001). Bars=10 mm. Fig. 6. Urediniospore in median view. Fig. 7. Teliospores. The arrow points to what might be construed as striation of the apex of one teliospore. the Idaho specimens. For urediniospore size as well, Gäumann’s measurements, based presumably on European specimens, were confirmed in this study rather than Cummins’. Phenologically, P. tanaceti s. str. appears to be largely behaving in Idaho as it always has in its native Europe. Of 71 BPI specimens of P. tanaceti on T. vulgare in Europe, 66 were collected in August or later. The five exceptions were collected in July ; only three noted dates and these were from mid-July onward. All PUR specimens examined in this study were also collected in August or later. The significance of the May 2003 collection in Lapwai Canyon remains to be determined. In contrast, 18 of 40 BPI specimens of P. tanaceti on A. tridentata in its native range, including Idaho, were collected in June or July or even earlier. If P. tanaceti s. lat. were a valid species concept, T. vulgare could not have remained rust-free for so long in North America. Tansy co-occurs in North America with many of the native and introduced species of the Anthemidae that have been reported as hosts of P. tanaceti s. lat. This co-occurrence is of long standing since T. vulgare was introduced into North America in the 17th century. If P. tanaceti s. lat. were truly a ubiquitous generalist of such hosts, then rust would have occurred on T. vulgare in North America long ago, given inevitable exposure to inoculum. 802 Puccinia tanaceti As just one example, Cummins (1978) regarded P. tanaceti as ‘circumglobal on cultivated mums ’. This first report of specialized Puccinia tanaceti s. str. represents the end of four centuries of rust-free ‘enemy release’ for tansy in the New World. According to the enemy release hypothesis, ‘plant species, on introduction to an exotic region, experience a decrease in regulation by herbivores and other natural enemies, resulting in a rapid increase in distribution and abundance ’ (Keane & Crawley 2002). Artemisia absinthium, also introduced and widespread in North America, remains in a state of ‘enemy release’, at least with respect to P. absinthii. P. absinthii should clearly not be considered a synonym of P. tanaceti if these two rust fungi lack a common host. In this regard, it was proven here that P. tanaceti cannot infect A. absinthium, the type host of P. absinthii. In light of the resistance in this study of A. tridentata to P. tanaceti s. str., rust on the native North American sagebrush has also been mistakenly attributed to P. tanaceti s. str. in some studies (e.g. Welch & Nelson 1995). Inasmuch as direct determinations of host range have been made within the complex of fungi which Gäumann (1959) called ‘ Formenkreis der Puccinia tanaceti ’, narrow specialization appears to be the norm. For example, Gäumann (1959) reported P. chrysanthemi as a rust fungus that could infect some species of Chrysanthemum but not others. P. pyrethri, P. gaeumanni, Uredo neocomensis, and P. matricariae were also regarded by Gäumann as specialists, on the basis of evidence from inoculation studies. Even P. balsamitae, which Gäumann did not regard as ‘strongly specialized ’, was not able to infect C. cineariaefolium, C. corymbosum, and T. vulgare in inoculation studies. Punithalingam (1968) also commented on inoculations that had been conducted with P. chrysanthemi in which more species of Chrysanthemum were resistant than susceptible. In any event, no one has yet shown that P. tanaceti s. str. can infect any plant species other than T. vulgare. By the same token, no one has yet shown that a species other than P. tanaceti s. str. from the P. tanaceti s. lat. complex is able to infect or cause rust on T. vulgare. If narrow specialization proves to be the norm for all autoecious rust fungi resembling P. tanaceti s. str. that are found on the Anthemidae and five other tribes of the Asteraceae, then P. tanaceti s. lat. must be a truly large complex of cryptic or sibling species. Anthemidae (1741 species), Astereae (2800), Helenieae (830), Heliantheae (2500), Lactuceae (1550), and Vernonieae (1300) are clearly speciose tribes of the Asteraceae, the most speciose of plant families (Bremer 1994). The six tribes currently recorded as hosts of P. tanaceti s. lat. are even spread across two subfamilies. Lactuceae and Vernonieae belong to the subfamily Cichorioideae whereas the other four tribes to host P. tanaceti s. lat. belong to subfamily Asteroideae (Bremer 1994). Thus, this host distribution is suggestive of an even bigger P. tanaceti s. lat. complex than current records would indicate. Only time will tell whether the P. tanaceti s. lat. complex is bigger than that of P. hieracii s. lat. (Hennen & Cummins 1969), that is centered on host genera of the tribe Lactuceae. Estimates of the magnitude of fungal diversity include consideration of cryptic species (Hawksworth 2001). Among rust fungi, there may be especially large complexes that have received little attention to date. ACKNOWLEDGEMENTS Claudia Nischwitz translated relevant pages of Gäumann’s Die Rostpilze Mitteleuropas from German to English, and made numerous collections of P. tanaceti s. str. in Lapwai Canyon. Valerie LynchHolm helped with SEM studies. Markus Scholler provided a loan of specimens from the Arthur Herbarium. The Idaho Research Foundation paid some of the travel costs associated with surveys. REFERENCES Arthur, J. C. (1962) Manual of the Rusts in the United States and Canada. [Cummins, G. B., illustrator.] Hafner Publishing, New York. [Reprinted from the 1934 edition with additional material by Cummins.] Bremer, K. (1994) Asteraceae: cladistics and classification. Timber Press, Portland. Bremer, K. & Humphries, C. J. (1993) Generic monograph of the Asteraceae-Anthemidae. Bulletin of the British Museum (Natural History), Botany 23: 71–177. Cummins, G. B. (1978) Rust Fungi on Legumes and Composites in North America. University of Arizona Press, Tucson. Farr, D. F., Rossman, A. Y., Palm, M. E. & McCray, E. B. (2002) Fungal Databases, Systematic Botany & Mycology Laboratory, ARS, USDA. http://nt.ars-grin.gov/fungaldatabases. Gäumann, E. (1959) Die Rostpilze Mitteleuropas. Büchler, Bern. Hawksworth, D. L. (2001) The magnitude of fungal diversity: the 1.5 million species estimate revisited. Mycological Research 105: 1422–1432. Hennen, J. E. & Cummins, G. B. (1969) The autoecious species of Puccinia and Uromyces on North American Senecioneae. Mycologia 61: 340–356. Keane, R. M. & Crawley, M. J. (2002) Exotic plant invasions and the enemy release hypothesis. Trends in Ecology and Evolution 17: 164–170. Mayor, E. (1946) Analyse des travaux parus en Suisse de 1939 à 1945. Uredineana 2: 75–108. Punithalingam, E. (1968) Puccinia chrysanthemi. CMI Descriptions of Pathogenic Fungi and Bacteria 175: 1–2. Welch, B. L. & Nelson, D. L. (1995) Black stem rust reduces big sagebrush seed production. Journal of Range Management 48: 398–401. Wilson, M. & Henderson, D. M. (1966) British Rust Fungi. Cambridge University Press, London. Corresponding Editor: D. L. Hawksworth