This study examines the relationship between citation frequency and the human capital of teams of authors. Analysis of a random
sample of articles published in top natural science journals shows that articles co-authored by teams including frequently
cited scholars and teams whose members have diverse disciplinary backgrounds have greater citation frequency. The institutional
prestige, the percentage of team members at U. S. institutions and the variety of disciplines represented by team member backgrounds
do not influence citation frequency. The study introduces a method for evaluating the extent of multidisciplinarity that accounts
for the relatedness of disciplines or authors.
Building on the findings of recent ethnographic studies of scientific practice, I develop and test theory about the impact
of taken-for-granted-ness on citation practice in scientific communities. Using data gathered from special issues of scientific
journals I find support for the hypothesized differences in the practices of natural and social science communities. Post
hoc analysis uncovers evidence of a third pattern of citation practice associated in part with engineering and technology
research, and evidence that organization studies and strategic management communities tend to employ extreme versions of social
science citation practices. I discuss the implications of the study for our understanding of communities of practice, for
our beliefs about differences between the branches of science, and about science as a productive enterprise.