Collaborations are pervasive in current science. Collaborations have been studied and encouraged in many disciplines. However, little is known how a team really functions from the detailed division of labor within. In this research, we investigate the patterns of scientific collaboration and division of labor within individual scholarly articles by analyzing their co-contributorship networks. Co-contributorship networks are constructed by performing the one-mode projection of the author-task bipartite networks obtained from 138,787 papers published in PLoS journals. Given a paper, we define three types of contributors: Specialists, Team-players, and Versatiles. Specialists are those who contribute to all their tasks alone; team-players are those who contribute to every task with other collaborators; and versatiles are those who do both. We find that team-players are the majority and they tend to contribute to the five most common tasks as expected, such as "data analysis" and "performing experiments". The specialists and versatiles are more prevalent than expected by a random-graph null model. Versatiles tend to be senior authors associated with funding and supervisions. Specialists are associated with two contrasting roles: the supervising role as team leaders or marginal and specialized contributions.