414 S. F. MacLean, Jr. 



the greatest contribution to total variation in five of the seven cases. The 

 exceptions were mesostigmatid mites and Collembola, in which within- 

 sample variation was greatest. Sampling date made a relatively small 

 contribution to total variation, indicating that spatial variation is more 

 important than temporal variation in determining the abundance of tun- 

 dra soil invertebrates. A similar pattern is seen in the soil microflora 

 (Chapter 8). 



Mean annual density of nematodes, estimated by soil sieving fol- 

 lowed by concentration through sugar solution centrifugation, varied be- 

 tween about 50,000 individuals m"^ in the basins of low-centered 

 polygons and 724,000 m'^ in the mesic meadow. These values are very 

 low, but have been confirmed by the use of three different extraction 

 procedures. In a review of data on nematode density and biomass in a va- 

 riety of terrestrial ecosystems, Sohlenius (1980) found that only in deserts 

 were mean values below one million individuals m ^ with coniferous 

 forest, deciduous forest, and temperate grasslands averaging over three, 

 six, and nine million individuals m"% respectively. Procter (1977) re- 

 ported densities in the range of one to almost five million individuals m"^ 

 in the tundra of Devon Island, N.W.T., Canada. Thus, the low densities 

 found at Barrow are not characteristic of tundra. As at Barrow, the wet 

 meadow at Devon Island supported the lowest density of nematodes. 



Trophic function of nematodes was determined by examination of 

 mouthparts. Although abundance varied by a factor of 15 among sample 

 points, trophic structure of the population was remarkably constant. The 

 free-living nematodes, which are largely bacterial and algal feeders, were 

 most abundant overall. The plant-parasitic nematodes became relatively 

 more abundant on the drier units, and were dominant on the mesic mea- 

 dow. Predatory nematodes made up a small proportion of the popula- 

 tion in all cases. The abundance of all three nematode groups was in- 

 versely correlated with soil moisture; the Spearmann rank correlation be- 

 tween total nematode abundance and moisture was -0.65 {p < 0.05). 



Drier units contain more dicotyledonous plants, many of which are 

 mycorrhizal (Miller and Laursen 1978). This is particularly so for Salix 

 on the mesic meadow. The much greater abundance of plant-parasitic 

 nematodes here may indicate that the root systems of dicotyledonous 

 plants are more susceptible to the attack of nematodes than are roots of 

 the graminoids that dominate wetter areas, or that the nematodes are 

 feeding directly upon mycorrhizal fungi. 



Biomass of nematodes was not directly determined. Based upon 

 studies of nematodes at a variety of other locations (Sohlenius 1980), a 

 mean biomass estimate of 0.1 ^g dry weight per individual was used to 

 approximate population biomass (Table 11-1). More elaborate functions 

 relating density to biomass, for example based upon differences in 

 trophic function, might be used, but hardly seem justified in light of the 



