Crab p a r s , consisting of a male grasping another crab (the graspee), were collected by divers during the first breeding season of female snow crab Chionoecetes opllio Different types of graspees were found and were ranked according to their reproductive value to the male. High-value graspees were pubescent females (close to their terminal matunty molt) and nulliparous females ljust molted and close to oviposition). Postn~olt prirmparous females (clean-soft shell and carrying eggs) also mated and were inseminated by males, but they were of less value than pubescent or nulliparous females as there was only a remote chance that the males' stored sperm would be used to fertilize the next egg clutch. Females copulated with up to 6 different males dunng their first breeding season. Another category of graspees including males and juvenile females provided the grasping male with no fecundity benefit. Pubescent females paired with males u p to 13 d before molting. Males grasping the high-value pubescent and nulliparous females were larger, had a harder shell, and were missing fewer limbs than the males grasping low-value primiparous females, other males or juvenile females and than the overall population of adult males on the mating grounds assessed by trawl. Size-assortative mating by male chela size and female carapace width occurred in the predominant pubescent pairs. Moreover, males with larger chelae were associated with pubescent females missing fewer limbs or having relatively narrower abdomens. Both traits may influence female survivorship and lifetime fecundity. The complex assortative mating pattern of snow crabs apparently denves from mate choice and male sexual competition in the context of prolonged attractiveness of pubescent females to males