An improved method to measure levels of polycyclic aromatic hydrocarbons (PAH) in smokeless tobacco products has been developed and validated. This method utilized base saponification and partitioning steps associated with the analysis of benzo[a]pyrene in whole tobacco (Health Canada Test Method T-307). To extend the number of PAHs determined, eight deuterated PAHs were added as internal standards (IS), as well as a solid phase extraction clean-up through two adsorbents (stacked NH₂ / silica). Analysis was performed by GC/MS in selected-ion-monitoring mode. Recoveries of target PAHs from spiked smokeless tobacco ranged from 84 to 104% with relative standard deviations under 11%. This method was applied to four CORESTA reference products (CR1, CR2, CR3, CR4), a Kentucky reference product (2S3), as well as eight commercially available smokeless products. The target PAHs found in the reference products ranged from 0.82 ng/g for acenaphthene (CR4) to 6754 ng/g for phenanthrene (CR2) on a "dry weight basis". The benzo[a]pyrene results ranged from 1.71 to 100 ng/g and were consistent with values generated by T-307 using HPLC with fluorescence detection. This method differs from that described in a recent publication using direct extraction into cyclohexane and ¹³C labeled PAH as IS. The published data showed reasonable agreement with those generated using this new method for nearly all PAHs with the exception of naphthalene where the reported values were nearly thirty times higher (1980 ng/g and 68.7 ng/g respectively). Attempts were made to reproduce the published method and in our hands, the two methods yielded comparable results for naphthalene. However, for other PAHs, the extraction efficiencies were two to five times lower than those determined using the new method. Possible explanations for the discrepancy in the results for naphthalene were investigated and the results will be presented.