316 P. W. Flanagan and F. L. Bunnell 



Linear regression with the same data produces: 



R = 1476+1 17.4 xTmean r' = 0.39, a = 0.01. 



For all microtopographic units, linear regressions of carbon dioxide 

 evolution versus mean daily temperature and maximum daily tempera- 

 ture consistently provided higher coefficients of determination, r\ than 

 regressions involving 2^^ '"and e^'^'" (where T = temperature). The ob- 

 served linear response to temperature may be a result of the summation 

 of a number of exponential responses. 



The contribution of plant roots to observed carbon dioxide produc- 

 tion was analyzed by comparing undisturbed soil cores with cores effec- 

 tively stripped of primary producers. Linear regressions of carbon diox- 

 ide evolution from stripped cores also provide better fits than do expo- 

 nential models for the effect of temperature. The relative response (the 

 predicted response at 10 °C divided by the predicted response at 0°C) is 

 higher for stripped cores than for cores on which the graminoid and moss 

 canopy was left intact. Removing the plant cover increased the relative 

 response from 2.08 to 3.1 in basins of low-centered polygons and from 

 1.59 to 1.89 on rims of low-centered polygons. The intercepts of equa- 

 tions for stripped cores of basin and rim soils were 155 and 634 ml CO2 

 day', about half the level of the intercepts of untreated cores (353 and 

 1121 mlCOjday-'). 



Direct comparisons between treated and untreated cores should be 

 viewed with caution because microbial populations in treated cores are 

 not experiencing the same environment as the controls. Further, there 

 may have been increased root respiration associated with clipping. How- 

 ever, the results do suggest that the near-linear response of carbon diox- 

 ide evolution to temperature changes in soils is in part due to the influ- 

 ence of primary production. The lower relative response of soils on rims, 

 which support higher primary production than do soils in basins, corrob- 

 orates the suggestion, as do the high Qio values observed for microbial 

 respiration in standing dead and litter substrates (Table 9-2). In short, 

 microbial activity below ground appears to respond more strongly to 

 changes in temperature than do processes of primary production below 

 ground. 



Undoubtedly some of the differences in carbon dioxide evolution 

 observed among microtopographic units (Figure 9-9) are associated with 

 differing primary productivity. Rims of low-centered polygons and poly- 

 gon troughs evolve approximately twice as much carbon dioxide as ba- 

 sins of low-centered polygons and support considerably greater primary 

 production. Evolution of carbon dioxide from meadow soils is still great- 

 er, possibly reflecting the higher biomass of bacteria in these soils. We 

 cannot distinguish between the influences of soil moisture and primary 



