Microcalorimetry, plate count and PCR-denaturing gradient gel electrophoresis (DGGE) were employed to investigate microbial diversity and activity in soils from the Red Soil Experimental Station of the Chinese Academy of Agricultural Sciences, Hunan Province, China, where a wheat-corn rotation with 12 fertilization treatments was established in 1990. Fertilization greatly increased microbial biomass carbon (C) and nitrogen (N) (C(mic) and N(mic)) as well as the activities of phosphatase, urease, invertase, protease, catalase and dehydrogenase. Manure alone (M) enhanced the number of denitrifying and aerobic bacteria by 54.4% and 20.5%, respectively, whereas fallow (H) increased the number of aerobic cellulose decomposing bacteria by 31.4%. Fallow and soils amended with mineral fertilizers plus pig manure or straw increased both the DGGE band patterns and the Shannon index compared with mineral fertilizers or the control. Mineral treatments with lower bacterial numbers enhanced the values of the peak time (t(max)) more than did organic treatments. The peak height (P(max)) was positively correlated (P<0.01), with soil enzymes, C(mic) and N(mic), and the number of microorganisms, whereas the peak time (t(max)) was negatively connected (P<0.01) with these parameters. The microbial growth rate constant (k) was linked to bacteria (P<0.01), actinomycetes (P<0.05) and catalase (P<0.05). The total heat evolution (Q) had no relationships with any soil microbial properties (except for catalase). We propose that P(max) and t(max) could be used as indices of soil microbial activity, while the values of k and Q are poor indicators.