The Herbivore-Based Trophic System 405 



ure (Bee and Hall 1956, Steere, pers. comm.), but this impact has not been 

 measured nor have long-term patterns been studied. The moss Voitia hy- 

 perborea is associated with musk ox and caribou dung (Steere 1974). 

 Heavy mats of the grasses Arctagrostis, Alopecurus and Calamagrostis 

 cover tops of sand dunes and pingos occupied by ground squirrels, ap- 

 parently in response to disturbance and manuring. Caribou and ground 

 squirrels also may influence the reproductive success of some dicotyle- 

 dons, particularly Pedicularis and Saxifraga, since they seem to be espe- 

 cially fond of their flowering heads. 



A particularly interesting interaction between lemmings, soil and 

 vegetation may be occurring in areas with low-centered polygons. Micro- 

 topography produces large differences in soil moisture within a few 

 meters. The deepest portions of polygon troughs may contain water all 

 summer long, whereas the basins of low-centered polygons contain water 

 only in early summer, and their rims are never submerged. Soil organic 

 matter is greatest in the basins (bulk density < 0.5 g cm"') and least in the 

 troughs (bulk density > 1 g cm''). Exchangeable phosphorus is greatest 

 (240 Mg g'') in the troughs and least (90 ^g g"') in the basins (Barel and 

 Barsdate 1978) Differences in soil conditions produce different vegeta- 

 tional communities. Graminoid shoots are most robust and dense (^^3000 

 m"') in troughs and least robust and dense in basins (1000 to 1500 m"^). 

 Phosphorus concentration in plant tissues and plant production are both 

 highest in troughs. Decomposition and, therefore, nutrient cycling ap- 

 pear to be most rapid in troughs because that is where production of 

 organic matter is greatest and accumulation is least. The highest activity 

 rates of bacteria and the highest standing crops of soil invertebrates also 

 occur in troughs, thus accounting for high rates of decomposition. Lem- 

 mings also concentrate their activities in troughs, apparently because the 

 most palatable and nutritious food is concentrated there. 



All these observations are consistent with one another, and they 

 allow the microtopographic units to be ranked in order of decreasing bio- 

 logical activity: troughs, rims and basins. The one factor that seems most 

 likely to account for the differences in biological activity is soil phos- 

 phorus. Higher levels of phosphorus allow greater production of more 

 nutritious vegetation, which stimulates both decomposition and herbi- 

 vory. Moisture conditions may also influence decomposition, and the 

 troughs maintain greater soil moisture during the warmest part of the 

 summer. But why should phosphorus be concentrated in the troughs? 



One explanation might be that phosphorus is leached from the rims 

 of low-centered polygons to the troughs. But if this were the only factor, 

 phosphorus would accumulate on both sides of the rims, in the basins as 

 well as the troughs. Then the basins and troughs would be expected to 

 show similar levels of available phosphorus, which is not the case. A se- 

 cond possibility is called the nutrient-transport hypothesis: that lemmings 



