Studies on the ecology of soil fungi have traditionally been concerned with the distribution of taxa in the soil, and the influence of the environment on community structure and composition. Relatively little emphasis has been placed on the role of the fungal community and the effects of community composition on decomposition processes, even though it is recognized that the decomposition of natural compounds is a major function of the fungal community. One approach to defining the role of soil fungi in the decomposition process is to take an inventory of the fungi in the soil, evaluate their enzymatic capacity under laboratory conditions and then use these data to make inferences about the role of the community in decomposition of natural substrates, as has been done by Flanagan and Scarborough (1974) and Gochenaur (1984). Using this approach, and comparing soils, Flanagan (198 1) concluded that “it appears that taxonomic differences in fungal communities between ecosystems do not significantly influence the decay rate of plant remains between sites. The biomass of fungi is related to quantity and quality of organic matter, and the decay rate of a given organic matter is controlled primarily by climates”. Although such studies provide useful information, they suffer from the fact that soil fungi are being tested under artificial conditions for their potential for decomposing specific substrates. To interpret the role of fungal community structure on actual decomposition rates, it would be better to examine fungi actually colonizing particular substrates, from the soil, and to monitor simultaneously the amount of decomposition that is taking place. An opportunity to do this arose as part of an ongoing study on the effects of amendments on the decomposition of cotton (cellulose) strips, buried in a Scottish moorland soil. Cotton strips of uniform weave, 1Ocm wide by 30 cm, were prepared (Latter and Howson, 1977) and inserted into an acid (pH 4.345) Scottish moorland soil near Banchory (Grid Ref. No. 637903). The general area of the study has been described in detail by Miles (1973). The top segments of the strips were placed in the old litter/F horizon, the side segments in the variable O/H/Ah horizon, and the bottom segments in the E/Es horizons (nomenclature according to Hodgson, 1974) using the method of French and Howson (1982). On 25 August 1981, five replicate strips were placed randomly in each of three plots that had been amended as follows: (i) N aDDlied as NH.NO, at a rate of 40 am-‘: (ii) P applied as KH;FCJ, at 60 gm-*, (iii) Ca applied& Ca;cb; at 100 g m-*. A further 5 strips were distributed between three plots treated with a 1: 1 mixture of glucose and potato starch: two in a plot treated at 4 g m-*, one in a plot treated at 16gm-* and the remaining two in a plot treated at 40 g m-*. Data from these three plots were combined as a single “carbohydrate” treatment. All amendments had been applied on three occasions during 1979; on 19 September, 15 October and 30 October. Five strips were also inserted in unamended control plots. The strips were removed on 8 December 198 1 and the side segments were cut, frayed and tested for tensile strength (Latter and Howson, 1977). Unburied control strips were simultaneously tested for tensile strength and the percent tensile strength loss for test strips was calculated as follows: