Further analyses were carried out in the R statistical environ- ment (R Core Team, 2014). Total microbial lipids and marker lipids for individual functional groups were analyzed separately using linear mixed effects models with the lmer function in the ‘lme4’ package (Bates et al., 2014). The model was fully factorial, with each combination of cropping system and stover removal/fertiliza- tion represented as a level in a combined factor, while block and plot within block were included as random effects to reflect the split-plot design of the experiment as well as within-plot sampling in 2008. Because of the balanced experimental design, the model estimates equaled the arithmetic mean for each treatment combination