ABSTRACT: The study of marine bird ecology at sea is complicated by the tendency of many species to follow and otherwise attend vessels. Vessel-attraction likely biases abundance estimates and blurs the correlation between seabird distributions andhabitat features over scales of tens of kilometers. Moreover, ship-following behavior inhibits the statistical analysis of seabird distributions because samples too closely spaced in time and space are not independent. These biases have importantimplications when estimating the size of rare and endangered seabird populations and the prey consumption by abundant species. This paper illustrates how observations of seabird vessel-attendance can be used to mitigate the biases of vessel-attraction andship-following behavior. I quantified the degree of albatross attraction to survey vessels off southern California, and estimated that standard 300 m strip transects overestimated their abundance by at least a factor of 3.57. Additionally, I modeledalbatross ship-following behavior, and determined that 95% of recognizable black-footed (Phoebastria nigripes) and Laysan (P. immutabilis) albatrosses ceased to follow the survey vessel after 60 and 38 min respectively. Using these models, Iestimated that standard survey methods overestimated black-footed albatross abundance by a factor of 1.17 due to their ship-following behavior. When the vessel-attraction and ship-following biases were combined, standard survey techniques overestimatedalbatross squid consumption off southern California by a factor of 4. Determining the degree of vessel-attraction and the temporal scale of ship-following behavior will help design and analyze seabird surveys. These improvements will enhance our abilityto monitor the distribution, abundance, and prey consumption of seabird populations.