An experiment was conducted to measure the effects of summer warming on the total population densities of soil-dwelling microarthropods in the high Arctic and to compare these results with those from natural between-year and between-site variations. Small polythene tents were used to elevate summer temperatures over 3 years on polar semi-desert and tundra heath in West Spitsbergen, Svalbard, Norway. Soil cores were taken at regular intervals from tented and untented (control) plots and heat extracted for mites (Acarina: Oribatida) and springtails (Collembola). Species present were similar at both sites, but at the start of the experiment total springtail populations were greater at the polar semi-desert whilst oribatid mite densities were equal at both sites. No significant effect of temperature elevation on oribatid mite populations emerged, even after 3 years. By contrast, springtail numbers were significantly lower on tented versus control plots at the polar semi-desert at the end of year 3, but not so at the tundra heath. Collembola numbers declined at both sites during the warm dry midsummers of 1992/1993 and this was most marked at the better drained polar semi-desert site. Over the equivalent period total oribatid mite populations, while relatively more stable, increased significantly at the polar semi-desert as a result of an increase in the number of juveniles. Results are interpreted in the context of the ecophysiological adaptations of oribatid mites and springtails to soil temperature and moisture. The resulting survival characteristics are considered in relation to the temperature and moisture characteristics of the two sites. The experiment demonstrated that year to year variation in climate, interacting with physical differences between sites, produced an equal or greater effect on microarthropod numbers at any one site than the 8–10% increase in “heat availability” (day degrees above zero) resulting from the summer tent treatment. The limitations of the use of tents to elevate soil temperatures are discussed. Comparisons are made with microarthropod population data from other polar and alpine sites.