How do observers become proficient at finding camouflaged targets? Does gaining proficiency in locating a target in one camouflage search situation improve performance in novel camouflage situations? We examined these questions by having participants first perform a difficult camouflage search task for three sessions and, in the last session, search for novel camouflaged targets. Near-perfect transfer of training was observed, with participants finding targets in novel camouflage situations almost as quickly as in highly familiar search situations. Previous research has suggested that an "object search" strategy rather than a "background search" strategy is one reason camouflage search can be slow and inaccurate. Eye movement analyses suggest that a more background-focused strategy was not necessary for improved performance and was not responsible for transfer. Another experiment explicitly instructed participants to search background regions, rather than regions containing salient objects, and provided participants feedback regarding the location of their own fixations. Surprisingly, even in this condition, participants did not devote more oculomotor attention to background regions. Theoretical and practical implications are discussed. Supplemental figures for this article can be downloaded from app.psychonomic-journals.org/content/supplemental.