results, and it provides a saving in time and resources 

 compared to a single challenge per plate. The precision of 

 our results reinforces the need to base clone delineation on 

 repeated challenges of coded isolates. 



The results of the expect-"no" analysis for both condi- 

 tions (tables 3 and 4) can also be interpreted on the basis 

 of ambiguous-separation between closely related geno- 

 types. Both kinds of error indicate uncertain reads, and 

 both kinds of errors are higher within plots than between 

 plots. The interpretation of this pattern, however, is dif- 

 ferent than for the expect-"yes" case. An additional possi- 

 ble source of error in the expect-"no" case is the exist- 

 ence, in the field, of haploid vegetative isolates that could 

 function as "testers." For example, where we use four 

 isolates placed 5 mm apart within the same petri dish they 

 could grow freely with any diploid of the same species and 

 mating could occur, after which the haploid would not 

 react with a different adjacent diploid clone. The existence 

 of such haploids in a population believed to be diploid 

 would generate unstable results, especially within the 

 multiple isolate challenge system used by us. Some of 

 these problems could be circumvented if mating reactions 

 and myceUal growth pattern were also recorded; haploids 

 produce a distinctive growth of aerial mycelium (Korhonen 

 1978; Ullrich and Anderson 1978). 



As stated above, the verification system we present is 

 predicated on the testing of diploid isolates. The existence 

 of a haploid isolate would lead to inconclusive results. If 

 two clones of the same species were being delineated, and 

 one clone contained a haploid isolate, that haploid could 

 grow with both clones. An unexpected "yes" reading 

 would result. The same problem exists in testing the wider 

 extent of clones. According to current theory, a haploid 

 isolate could conceivably give a "yes" read against any 

 diploid isolate of the same species from anywhere in the 

 world. If one were expecting such a mating response, this 

 situation could be monitored. At the very least, any future 

 clonal delineation should include provisions for assessing 

 mating and the presence of haploids in the test population. 

 The existence of haploid isolates could explain the unusual- 

 ly high number of ambiguous readings betw^een plot-to-plot 

 clones vdthin a National Forest (table 5). These challenges 

 were made on an isolate-specific level before the clones 

 were defined to provide data unbiased by reader knowl- 

 edge. This system led to a random selection of isolate pair- 

 ing. If a haploid was selected to represent the clone and it 

 was paired against the same species from a distant plot, a 

 "yes" (when a "no" was expected) would result. Even 

 though most authors believe that vegetative field isolates 

 are diploid (Kile 1983; Korhonen 1978; Ullrich and 

 Anderson 1978; Wargo and Shaw 1985), one must keep in 

 mind that most of these workers have made almost ex- 

 clusive use of cultures derived from sporophores or from 

 vegetative samples obtained in the vicinity of sporophores 

 on sites where Armillaria spp. are in a pathogenic mode. 

 In addition, it is theoretically possible that haploids could 

 have arisen from diploids unstable in culture. This, how- 

 ever, does not appear to be the case, since the diploids of 

 the genus have been shown to possess a high degree of 

 genetic stability (Anderson and Ullrich 1982). Existence of 

 these haploids may not be so surprising in light of the re- 

 cent reports of haploid tissues in Armillaria spp. basidio- 



carps (Peabody and Peabody 1986). Our tentative conclu- 

 sion based on the challenge behavior of some isolates is 

 that our random samples of vegetative structures most 

 probably included some haploids. A thorough understand- 

 ing of the confirmed existence, distribution, behavior, and 

 role of such haploid components would explain much about 

 the genetics and epidemiology of the genus Armillaria. 



Number of Clones per Plot 



A most interesting finding is that an average of 2.3 

 clones per 0.04-ha plot was found. This result supports the 

 current ideas regarding the distribution of clones and 

 species. As seen in tables 7 thru 18 in the appendix, most 

 plots contained one or two clones regardless of the 

 number of isolates obtained. This argues for the coex- 

 istence of two species on most sites, since others have not 

 found two clones of the same species to commonly coexist 

 (Kile 1983; Korhonen 1978; Ullrich and Anderson 1978). 

 Several sites had three clones and could represent the 

 coexistence of three species. The final answer to this 

 generalization will come ^vith the confirmation of species 

 affiliations of the sampled clones. 



Clone Habits 



Associations of clones by hosts and rhizomorph produc- 

 tion (table 6) suggest four Armillaria spp. in our popula- 

 tion of randomly collected isolates. One group of clones 

 comprised 24 percent of aU clones and was restricted to 

 hardwood shrubs. These were generally found as a 

 single rhizomorphic isolate. If the average number of 

 rhizomorphic-derived cultures is a measure of tendency of 

 a clone to produce rhizomorphs, then this group may 

 signify a species that rarely produces rhizomorphs and is 

 restricted to hardwood hosts. A second group was 

 restricted to conifers and produced few rhizomorphs. It 

 averaged 1.4 isolates per clone. This group was about 

 eve'nly split on healthy conifers, conifer detritus, and in 

 pathogenic mode. It included 48 percent of all clones. This 

 pattern fits A. ostoyae as described by Rishbeth (1982) in 

 England and Morrison and others (1985) in British Colum- 

 bia. The third group produced an average of 4.1 isolates 

 per clone. It was found on shrub detritus, conifer detritus, 

 healthy shrubs, and healthy conifers. This group contained 

 18 percent of all clones and behaves like A. bulbosa in 

 England (Rishbeth 1982) where it was found on both hard- 

 woods and conifers. This species is a prolific producer of 

 rhizomorphs, and is not pathogenic. This species was 

 found in British Columbia but w^as restricted to hardwoods 

 as shown by a sporophore population (Morrison and others 

 1985). The fourth group of clones was a prolific producer 

 of rhizomorphs (4.3 isolates per clone) and occurred on 

 shrubs, conifer detritus, healthy conifers, and in a 

 pathogenic mode. Only 7 percent of the clones belonged to 

 this group. This group may represent a species that pro- 

 duces many rhizomorphs, is pathogenic on conifers, and 

 commonly exists in a saprophjrtic mode. 



8 



