Description and criticism is given of a preliminary design and use of a soil/litter microcosm in which oxygen, temperature, and humidity are kept constant, and oxygen generation and carbon dioxide and heat evolution rates are monitored. Using four microcosms, one acting as a dead control, experiments were performed giving the following results: for 'identically' prepared and incubated microcosms, the coefficient of variation was as small as 3.8 percent for carbon dioxide evolution rate and as large as 9.9 percent for oxygen consumption rates. It was also found that an adjustment period of 7 to 10 days after microcosm preparation was necessary to approach relatively constant production rates. For microcosms adjusted to 10, 30, and 60 percent of field water holding capacity, oxygen and carbon dioxide rates and bacterial densities vary directly whereas the fungi and actinomycetes varied inversely; while for cadmium-amended microcosms, 0.01 ppm and initial stages in the 10 ppm CdCl2 unit, oxygen consumption was stimulated suggesting respiratory enzyme uncoupling while in the later stages the 10 ppm cadmium-amended soils reduced both O2 and CO2 respiration by 40 percent. No organismal density changes due to cadmium were detected indicating the cadmium initially affects respiration, possibly by uncoupling respiratory phosphorylation, and that longer experiments might be necessary to detect population density changes. Copyright (c) 1976 by The Williams & Wilkins Co.