This study investigated the occurrence of the rust fungus Puccinia chondrillina in clonal populations of its host plant, Chondrilla juncea. In 16 populations spread across eastern Turkey, 48 different multilocus isozyme phenotypes were identified in the host. Of these clones, 88% were restricted to single localities, while the remaining 12% were found in 2-11 populations. For 13 of the 16 plant populations the commonest host clone was always infected. Indeed, at ten sites this clone was the only one found to carry disease. In the remaining three populations the rusted plants were all of the second commonest isozyme type. The possibility of such a tight association of rust incidence with host clone frequency simultaneously across a wide geographic area is very low (P≦0.023), supporting the contention that the pathogen P. chondrillina may be imposing negative frequency-dependent selection on these C. juncea populations.

Isozymes were used to discriminate clones of the apomictic plant, Chondrilla juncea L. (Asteraceae), a Eurasian species that has become a major weed of cereal cultivations in Australia. The sample that was screened consisted of the progeny of single plants from each of 123 sites in central Turkey. When used in combination, six polymorphic enzyme systems separated 91 distinct clones. Seventy-four percent of samples detected unique clones, none of these being widespread. Chromosome counts established that these clones are triploid. This pattern contrasted with that found in a survey of 23 populations in southeastern Australia. Apart from the narrow-leaf rust susceptible clone, the survey found two additional multilocus genotypes only, both of which are immune to the introduced strain of the rust.

The resistance structure of a Turkish population of the clonal, apomictic composite Chondrilla juncea and the pathotypic structure of a co-occurring population of its obligate rust pathogen, Puccinia chondrillina, was determined by sequential inoculation of 19 host lines with 15 pathogen isolates each derived from single pustules collected from separate plants among the host population. The resultant matrix of resistant and susceptible reactions provides strong circumstantial evidence for a gene-for-gene interaction. Seven distinct pathotypes were detected in the pathogen population. One of these comprised 53% of the population, a second comprised 13%, while the remaining five pathotypes were each detected only once. The host population was similarly diverse, being composed of eight resistance phenotypes, only two of which were represented by more than one host line. Although C. juncea is apomictic, there was only 58% congruence between host resistance and multi-locus isozyme phenotype categories within this population. Pathotypic phenotypes of 13 other isolates of P. chondrillina collected from ten other Turkish and three more distant populations of C. juncea were markedly different from those found in the population studied in detail. There was no obvious relationship between the degree of geographic separation of pathotypes and their ability to attack particular C. juncea lines in this or three other populations represented by single host lines.

The internal transcribed spacer (ITS) regions of ribosomal DNA were amplified by PCR and used to develop genetic markers for isolates of Puccinia carduorum being evaluated for biological control of Carduus thoermeri (musk thistle). Unique patterns were produced upon restriction of ITS DNA amplified from four separate Puccinia spp. Restriction patterns of ITS DNA of isolates of P. carduorum from Carduus acanthoides and C. thoermeri were distinct from those of P. carduorum from Carduus tenuiflorus and Carduus pycnocephalus. By this technique, isolates of P. carduorum from four different weed hosts can be differentiated from other Puccinia spp. and separated into two host groups.