VAMfungal communities

Total counts of VAM fungal spores ranged from 17 to 316 spores/g dry soil, and were lowest in the 1 year field, the oak forest and the hardwood forest (Fig. 2). Within the 12 abandoned fields, total spore count was positively correlated with field age and H2O soluble C, and negatively correlated with non-host abundance (Table 2). Total spore count was not significantly correlated with infectivity (Table 2).

A total of 25 different VAM fungal species in 5 genera were identified from the 15 sites examined (Table 3). We were uncertain of the identity of a thick walled Gigaspora species (Gigaspora sp.) which resembled Gigaspora margarita in wall structure, but ranged in color from yellow to orange. Glomus aggregatum was the most abundant species, comprising between 54% and 95% of the spores from each site (Fig. 2). Relative abundance of some species were unrelated to successional rank, but other species were distinctly more abundant early, or late in the chronosequence (Fig. 3). Of the 25 species observed, seven were significantly negatively correlated to successional rank (early successional) and five were significantly positively correlated to successional rank (late successional) (Table 4). Spore abundance of early successional species were negatively correlated to organic and H2O soluble soil C, soil N and root biomass, while late successional species were positively correlated to these parameters (Table 4).

Between 12 and 22 different VAM fungal species were observed per site (Fig. 4a). Species richness did not increase with successional rank; however, diversity tended to increase (P=0.08), primarily because G. aggregatum became relatively less abundant in the older sites (Table 5). Percent dissimilarity with the rye field increased significantly with successional rank (Fig. 4b). Diversity and dissimilarity with the rye field were also positively correlated to H2O soluble C and total N (Table 5).

VAM Infectivity

Results of the bioassays indicate that densities of infective propagules were relatively low in the fallow field and rye field and generally increased with field age up to 53 years following abandonment. Infectivity was lower in the 60 year old field and in the savanna, and was very low in the upland pin oak and northern hardwood forest sites (Fig. 5). Separate analysis of the 1987 and 1988 bioassays gave the same results. Therefore, data sets were combined [(1987 VAM root length + 1988 VAM root length)/ (1987 root length+ 1988 root length)] for the final correlation analysis. Infectivity was not correlated with either corn shoot weight or corn root weight (r= 0.23 and r=0.22 respectively, N= 14, P>O.O5). Infectivity was positively correlated with field age, soil pH, H2O soluble C and root biomass, and negatively correlated with total P (Fig. 6). Since these five variables were highly autocorrelated (Table 5), we used partial correlation analysis to examine each of their relationships with infectivity while holding constant the other four variables. Using this approach, pH, and H2O soluble C were still positively correlated with infectivity (rp = 0.88, P < 0.001; rp = 0.62, P < 0.05 respectively), but age, root biomass, and total P were no longer significantly correlated with infectivity (rp= -0.36; rp= -0.25; and rp=0.10 respectively).