, and potentially women with high normal (35)(36)(37)(38)(39)(40)(41)(42)(43)(44) or low normal (<28) CGG repeats, are at risk of premature ovarian aging. The scarcity of population data on CGG repeats <45 CGG, and variation in raceethnicity, makes it difficult to determine true associations. DNA was analyzed for FMR1 CGG repeat lengths from 803 women (386 caucasians, 219 African Americans, 102 Japanese, and 96 Chinese) from the US-based Study of Women's Health Across the Nation (SWAN). Participants had 1 menses in the 3 months before enrollment, 1 pregnancy, no history of infertility or hormonal therapy, and menopause 46 years. Statistical analyses used Fisher exact tests. Among these women with normal reproductive histories, significant FMR1 repeat length differences were found across race-ethnicity for both the longer (P ¼ .0002) and the shorter (P < .0001) alleles. The trinucleotide length variance was greater for non-Asian than Asian women (P < .0001), despite identical median values. Our data indicate that short allele lengths <25 CGG on one or both alleles are more common in non-Asian than Asian women. We confirm the minor allele in the 35 to 39 CGG range among Asians as reported previously. Only 2 (0.3%) premutation carriers were identified. These data demonstrate that FMR1 distributions do vary by race-ethnicity, even within the ''normal'' range. This study indicates the need to control for race-ethnicity in FMR1 ovarian aging research and provides race-ethnic population data for females separated by allele.