Host Range Differences Between Populations of Puccinia graminis subsp. graminicola Obtained from Perennial Ryegrass and Tall Fescue

Cool-season grasses are intensively cultivated for seed production in western Oregon, where stem rust is the most damaging disease on several species, including perennial ryegrass (Lolium perenne), annual ryegrass (L. multiflorum), and tall fescue (Festuca arundinacea). Stem rust also occurs on seed crops of orchard grass (Dactylis glomerata) and red fescue (F. rubra) in the region. The causal agent of stem rust in grasses and cereals is Puccinia graminis Pers.:Pers. Within this species, there exist distinct populations that differ in morphology (13) or in the host species they can infect (1). There has not been a uniform approach to nomenclature or taxonomy of these subspecific groups, a situation which also has been noted for the crown rust pathogen (6) and other rusts affecting cereals and grasses (14). Schwartz (14) notes that some taxonomists use a narrow hostbased concept for delimiting rust species and subspecific groups, whereas others use a broader concept based principally on morphology. The existence of subspecific groups of the stem rust pathogen based on host range was first documented by Erickson (5), who called them formae speciales. Each forma specialis, although not restricted to a single host species or genus, has a unique constellation of compatible hosts. The first observation of a forma specialis in P. graminis from Lolium spp. was a 1948 report by Guyot et al. in France (8). In 1951, Waterhouse (17) described an Australian population of P. graminis pathogenic to L. perenne and 30 other grass species (including F. arundinacea, D. glomerata, and Poa annua) but not to species in genus Avena, Triticum, or Secale. He considered its spore morphology sufficiently distinct to designate it a subspecies, P. graminis f. sp. lolii. Shortly thereafter, a brief note by Guyot (7) reported inoculation studies in which his inoculum of P. graminis f. sp. lolii displayed a host range similar to that of the Australian population. Guyot (7) also reported a host range for P. graminis f. sp. festucae that included F. arundinacea, F. ovina, and F. pratensis; this form could not infect L. perenne, oat (Avena sativa), or rye (Secale cereale ) and produced only small, poorly sporulating pustules (termed “low-infection type” in the rust literature; 11) on D. glomerata. Tajimi (15) conducted a study with P. graminis f. sp. lolii in Japan and found that this population had a host range that was generally similar to the French and Australian populations, but that included A. sativa as well as a fully susceptible reaction (“high-infection type”) on S. cereale and Phleum pratense. Urban (16), in a 1967 paper, divided Puccinia graminis into two subspecies, P. graminis subsp. graminis and P. graminis subsp. graminicola, based principally on spore morphology. He placed the lolii populations of Waterhouse (17) and Guyot (7) in P. graminis subsp. graminicola. In a comprehensive treatment of rust fungi on grasses and cereals, Cummins (4) follows Urban in using only morphological characters, thus placing various formae speciales (designated by other authors) into each of the morphological subspecies. Two studies have examined the host range of populations identified as P. graminis subsp. graminicola from grasses. A population from Phleum pratense in Eastern Europe was pathogenic to D. glomerata, Poa pratense, and P. annua (among many others) but did not cause symptoms on L. perenne or F. arundinacea (3). In Oregon, a population of Puccinia graminis subsp. graminicola collected from F. rubra subsp. commutata was found to be essentially nonpathogenic to L. perenne or F. arundinacea (19). The host range of P. graminis subsp. graminicola populations from L. perenne or F. arundinacea in Oregon has not been evaluated. For optimum progress in research on resistance breeding and epidemic management, it is important to know whether the rust pathogen affecting F. arundinacea is the same as that affecting L. perenne, and whether the population or populations can cause disease symptoms on other grass seed crops in the region as well. The objective of this study was to test isolates of P. graminis subsp. graminicola from L. perenne and F. arundinacea for differences in ability to cause disease on a range of grass species.