This study assessed whether particulate organic matter (POM) in sand fractions, isolated by wet sieving after treatment with Na hexametaphosphate, can be a sensitive indicator of incipient changes in the content and composition of soil organic matter. In five long-term field experiments including different cropping systems, N fertilizer applications, and organic amendments, we found that C and N in the fine to medium sand fraction (0.063-0.600 mm, "Fraction B") showed considerably larger relative errors according to ANOVA (RMSE was 11-20% of the mean), slightly lower values of the F statistic, and slightly less contrast between treatments than total organic C and N (RMSE 3-9% of the mean). Imprecision in laboratory procedures only explained part of the increase in RMSE for C and N in Fraction B compared with total C and N; within-field spatial variability most likely had a greater influence. Although organic matter in Fraction B had a higher intrinsic sensitivity to soil management, which was partly able to overcome the larger errors, we concluded that an observer would be more likely to detect changes by measuring total organic C and N, when monitoring decadal changes in C and N pools. This makes the investigated POM fractions less suitable as indicators for changes in soil C stocks. However, the C/N ratio of Fraction B showed a distinct signature of the history of organic matter input to the soil, which was absent in the C/N ratio of the total fine earth.