ABSTRACT: Both essential and non-essential elements have been associated with female reproductive function in epidemiologic investigations, including among IVF populations. To date, most investigators have used blood or urine to assess biomarkers of exposure, with few employing ovarian follicular fluid (FF). FF may offer a more direct "snapshot" of the oocyte microenvironment than blood or urine, however previous studies report follicle-to-follicle variability in FF constituents that may contribute to exposure misclassification. Our objectives were to investigate sources of trace element variability, to estimate FF biomarker reliability among women undergoing IVF (n = 34), and to determine the minimum number of follicles required to estimate subject-specific mean concentrations. We measured As, Hg, Cd, Pb, Cu, Mn, Se, and Zn in FF samples using inductively coupled plasma tandem mass spectrometry. Inter-subject (between-women) variability contributed most of the variability in FF element concentrations, with ovarian, follicular, and analytical as smaller sources of variability. The proportion of variability attributable to sources between-follicles differed by age, body mass index (BMI), race, and cigarette smoking for Cu, Se, and Zn, by BMI and cigarette smoking for As, by primary infertility diagnosis for Hg, Cu, Se, and Zn, and by ovarian stimulation protocol for Mn and Se. Four to five individual follicles were sufficient to estimate subject-specific mean Cu, Se, and Zn concentrations, while >14 were necessary for As, Hg, Cd, Pb, and Mn. Overall, our results suggest that FF is a suitable source of biomarkers of As and Hg exposure in ovarian follicles. Although limited in size, our study offers the most comprehensive exploration of biological variation in FF trace elements to date and may provide guidance for future studies of ovarian trace element exposures.