A variety of physical and chemical techniques are used to fractionate soil organic matter, but detailed comparisons of the different approaches and tests of how separation methods influence the properties of isolated organic matter pools are lacking. In this case study based on A horizon samples of 2 California coniferous forests soils, we 1) evaluate the effects of root removal and ultrasonic dispersion on the properties of the <2 g cm3 light fraction and 2) compare the properties of fractions obtained by sequential density separations of ultrasonically treated soil with those obtained by density followed by acid/base hydrolysis (Trumbore et al. 1996). A root-removal effort based on hand-picking visible roots reduced the radiocarbon content and increased the estimated turnover time of the light fraction by roughly 12%. Root-removal protocols that vary between investigators thus can potentially confound variability in carbon cycling for this fraction caused by environmental factors, such as climate. Ultrasonic dispersion did not have a clear effect on the light fraction C and N content or isotopic signature, but led to a decrease in the % C and C/N of the recovered heavy fractions, and losses of 1219% of the total soil C to the sodium metatungstate density solution. Sequentially isolated density fractions clearly differed in mineralogy and organic matter chemistry, but natural-abundance 14C analyses indicated that distinct mineral phases did not correspond to unique C-turnover pools. Density fractions containing kaolinite group minerals alone and in combination with hydroxy-interlayered vermiculite were found to harbor both fast and slow cycling carbon. In contrast, severe chemical treatment isolated a carbon pool with the lowest overall 14C content and longest inferred mean turnover time. Overall, our results show that care must be taken when relying on physical (density) separation to isolate soil fractions with different dynamics, as the details of treatment will influence the results.