Soil’s importance as a major sink in global C cycling has led to a myriad of methods to divide this heterogenous pool into meaningful fractions to characterize soil C dynamics. Turnover rates of soil organic matter (OM)-C have been found to decrease with decreasing particle-size, therefore methods that physically fractionate soil may be used to determine C pools of a range of stabilities. Invariably, these schemes require dispersing aggregated soil by agitation in solution. Sodium hexametaphosphate (HMP) solution is a favored chemical dispersant in many studies. Deionized water (DI) may also be used, sometimes with glass beads. Shaking in HMP or DI causes soil C and N to be solubilized and lost. Whether this is a significant amount of C or N is unknown unless recoveries are calculated, and studies sometimes ignore soluble C loss when determining an organic matter fraction by difference. The amount of C loss probably varies with soil properties such as texture, clay mineralogy, and percent organic carbon. This study compared the effects of using HMP vs. DI on recovery of C and N using a high OM, highly aggregated soil from a tallgrass prairie site to compare to a low OM, loosely aggregated soil from a sandy forest site. It was found that soil properties must be considered when choosing an appropriate dispersing agent to minimize C loss. In addition, HMP can also cause difficulty in completely recovering clay-sized particles. Our results suggest that when choosing an appropriate dispersant, the trade-offs between adequately dispersing soils and the dispersant’s effect on the recovery/loss of soil C should be evaluated with regards to the questions being addressed by each study.