xx, 198 p. A print copy of this thesis is available through the UO Libraries. Search the library catalog for the location and call number. / "Bathymodiolus" childressi is a mixotrophic mussel from Gulf of Mexico cold seeps. There is no genetic differentiation of mussels among the seeps, suggesting wide dispersal of their larvae. This dissertation describes larval biology, ecology, and recruitment dynamics for "B." childressi. Cleavage is spiral at a rate of one per 3-9 hours, with blastula larvae hatching by 40 hours at 7-8 à à °C. At 12-14 à à °C, D-shell veligers developed by day 8 without being fed. Egg size and shell morphology indicate planktotrophy, but feeding was not observed. Embryos developed normally from 7-15 à à °C and 35-45 ppt. Although survival of larvae declined with temperature, some survived at 25 à à °C. Larval survivorship was similar at 35 and 45 ppt. Oxygen consumption increased from blastulae to trochophores and was higher for "B." childressi than for shallow-water mussel trochophores. Estimated energy content of "B." childressi eggs was greater than the energy content of shallow-water mussel eggs. An energetic model predicts that the eggs provide sufficient energy for "B." childressi trochophores to migrate into the euphotic zone. In fact, "B." childressi veligers were found in plankton tows of surface waters. The influence of recruitment on fine-scale distributions of adults at the Brine Pool cold seep was examined through manipulative field experiments. The "Bathymodiolus" childressi population at this site has a distinct bimodal size structure that shifts across an environmental gradient. New recruits of "B." childressi are abundant in the inner zone, where methane and oxygen are high and sulfide is low, leading to the inference that larvae settle preferentially there. Experiments were placed in the inner and outer zones and 2-m away from the bed. The number of larvae collected in traps did not differ among the three zones, nor did settlement density. Juveniles survived and grew in all zones, but more caged than uncaged juveniles survived. Mortality of uncaged juveniles was similar in all zones, suggesting that predation does not cause the bimodal distribution. These results suggest that the bi-modal distribution cannot be attributed to settlement preferences or juvenile mortality, but instead to migration or early post-settlement mortality. This dissertation includes my co-authored materials. / Adviser: Craig M. Young