Common shrimp, Crangon crangon (L.), were exposed to inorganic arsenic (arsenate), trimethylarsine oxide, or arsenobetaine in sea water (100 l gA s l )1 ) or in food (1 mg As g )1 wet wt) for up to 24 d, followed by 16 d depuration in clean sea water with undosed food, in order to determine the eAciency of uptake and retention of the compounds. Accumulation of arsenic in the tail muscle, gills, midgut gland, exoskeleton, and remaining tissues was found to depend on the chemical form of the arsenic and the route of exposure. No arsenic was accumulated by C. crangon exposed to arse- nate or trimethylarsine oxide in sea water. Shrimps ex- posed to waterborne arsenobetaine initially accumulated a small amount of arsenic in their tail muscle and gills. After 16 d, C. crangon fed arsenate, trimethylarsine ox- ide, or arsenobetaine had accumulated arsenic in their tail muscle to levels 2-, 2-, or 40-times, respectively, that of the control group. A roughly linear rate of ac- cumulation was shown by shrimps fed trimethylarsine oxide or arsenobetaine, but C. crangon fed arsenate ac- cumulated arsenic for 16 d, then lost arsenic such that their concentration on Day 24 was not significantly diAerent from that of the control group. Patterns of arsenic accumulation in the gills of shrimps fed the compounds were similar to those seen in the tail muscle. On a whole animal basis, C. crangon retained1.2% of the arsenate, 1.6% of the trimethylarsine oxide, and 42% of the arsenobetaine consumed over the first 16 d of exposure, with roughly half present in the tail muscle in each case. Data obtained support the view that the direct uptake of arsenobetaine from sea water does not make a significant contribution to the relatively high concentrations of this compound in marine crustaceans, and that food is the primary source. Naturally occurring arsenic compounds in C. crangon and possible trans- formations of the administered arsenic compounds were examined by high performance liquid chromatography using an inductively coupled plasma mass spectrometer as the arsenic-specific detector. Control C. crangon contained arsenobetaine as the major arsenic compound (>95% of total arsenic); tetramethylarsonium ion (0.7%) and an unknown arsenic compound (1.7%) were also present as minor constituents. Shrimp ingesting arsenobetaine accumulated it unchanged. Shrimp in- gesting arsenate did not form methylated arsenic com- pounds; they appeared to contain their accumulated arsenic as unchanged arsenate only, although the pos- sibility that some of the arsenic was reduced to arsenite could not be excluded. C. crangon ingesting trimethyl- arsine oxide biotransformed the compound predomi- nantly to dimethylarsinate.