Two sample preparation methods, acid pretreatment and no pretreatment, for δ^15 N analysis in mollusk shells were compared on sample splits from three common Gulf of Mexico and North Atlantic mollusks (Mercenaria spp., Crassostrea virginica, and Mytilus edulis). In all but one sample, no statistically significant difference (2σ) in δ^15 N values was measured between these two preparation techniques. However, sample splits that were not acid pretreated produced lighter δ^15 N values than their acidified counterparts in 82% of samples studied, and lower N content in small samples correlated with greater differences in method results. In addition, shell biomineralogy directly affected the %N of the samples; calcitic shell material contained greater %N, and produced data with higher analytical precision than aragonitic shell in the analyzed taxa. These data suggest that shell N content controls analytical data precision and that biomineralogy controls shell %N and N content. Within a singe species, N shell content varied as much as 30μg in C. virginica and 24μg in Mercenaria spp., likely as a result of differences in available food supply and N sources to grow-out locations. Because %N can vary greatly among and within species, preliminary analyses are recommended to determine the expected N content in samples and to establish whether omitting acid pretreatment of samples will result in sufficient analytical data precision. N content should also be reported along with analytical error to demonstrate that results are robust.