The C. elegans genome encodes an expansive array of potential xenobiotic defense factors that likely help the worm survive environmental toxins. As a consequence, the worm is relatively resistant to xenobiotic perturbation in a laboratory setting. For example, we surveyed over 14,000 small molecules for the induction of obvious phenotypes in the worm. The fraction of compounds with previously characterized bioactivity in other systems that induced phenotype in worms (3.2%, n=3280) is similar to that of 11,000 compounds without documented bioactivity (2.2%). We hypothesized that the low hit rate for the known bioactive compounds is because the worm is generally resistant to small molecule accumulation. To investigate xenobiotic retention in worms, we developed a high-throughput high performance liquid chromatography (HTP-HPLC) assay. We analyzed 1,110 molecules of the Spectrum library (Microsource Inc.) and found that only 177 (16%) are retained. Of these, 15.7% induce phenotype, representing a four-fold enrichment for molecules with bioactivity in worms (p<1.6E-13). These results support our hypothesis that C. elegans is generally resistant to small molecule accumulation. Next, we used our dataset of retained and non-retained molecules to build a model that predicts small molecule accumulation in the tissues of C. elegans. Such a model would allow us, and others, to screen small molecule libraries with greater efficiency, as robust retention is a key prerequisite for robust target inhibition. We compiled over 6,000 physicochemical descriptors of small molecules, and have identified over 100 that distinguish the 177 retained molecules from those not retained. Based on this analysis we built a Bayesian model of retention and used this model to rank 1258 molecules of the LOPAC library (Sigma). The top 50 molecules predicted to be retained are 4.4-fold enriched for molecules that induce phenotype in the worm (p<6.9E-4). We then analyzed the retention of the top 50 and bottom 50 ranked molecules, and found that 48% of the top-ranked molecules are retained in worm tissue. In contrast, only 8% of the bottom-ranked molecules are retained. This demonstrates that we have developed a good model to predict small molecule retention in C. elegans. Our model will hopefully facilitate the design of new nematicides to treat parasitic nematode infections, aid in the design of novel drug leads using C. elegans as a model for disease, and expand the use of small molecules as tools to probe worm biology.