STUDY QUESTION: How are germ cell numbers and initiation of folliculogenesis affected in fetal Turner syndrome (TS) ovaries? SUMMARY ANSWER: Germ cell development was severely affected already in early second trimester pregnancies, including accelerated oogonia loss and impaired initiation of primordial follicle formation in TS ovaries, while the phenotype in TS mosaic ovaries was less severe. WHAT IS KNOWN ALREADY: Females with TS are characterized by premature ovarian insufficiency (POI). This phenotype is proposed to be a consequence of germ cell loss during development, but the timing and mechanisms behind this are not characterized in detail. Only few studies have evaluated germ cell development in fetal TS and TS mosaic ovaries, and with a sparse number of specimens included per study. STUDY DESIGN, SIZE, DURATION: This study included a total of 102 formalin-fixed and paraffin-embedded fetal ovarian tissue specimens. Specimens included were from fetuses with 45,X (N = 42 aged gestational week (GW) 12-20, except one GW 40 sample), 45,X/46,XX (N = 7, aged GW 12-20), and from controls (N = 53, aged GW 12-42) from a biobank (ethics approval # H-2-2014-103). PARTICIPANTS/MATERIALS, SETTING, METHODS: The number of OCT4 positive germ cells/mm2, follicles (primordial and primary)/mm2 and cPARP positive cells/mm2 were quantified in fetal ovarian tissue from TS, TS mosaic and controls following morphological and immunohistochemical analysis. MAIN RESULTS AND THE ROLE OF CHANCE: After adjusting for gestational age, the number of OCT4+ oogonia was significantly higher in control ovaries (N = 53) versus 45,X ovaries (N = 40, P < 0.001), as well as in control ovaries versus 45,X/46,XX mosaic ovaries (N = 7, P < 0.043). Accordingly, the numbers of follicles were significantly higher in control ovaries versus 45,X and 45,X/46,XX ovaries from GW 16-20 with a median range of 154 (N = 11) versus 0 (N = 24) versus 3 (N = 5) (P < 0.001 and P < 0.015, respectively). The number of follicles was also significantly higher in 45,X/46,XX mosaic ovaries from GW 16-20 compared with 45,X ovaries (P < 0.005). Additionally, the numbers of apoptotic cells determined as cPARP+ cells/mm2 were significantly higher in ovaries 45,X (n = 39) versus controls (n = 15, P = 0.001) from GW 12-20 after adjusting for GW. LIMITATIONS, REASONS FOR CAUTION: The analysis of OCT4+ cells/mm2, cPARP+ cells/mm2 and follicles (primordial and primary)/mm2 should be considered semi-quantitative as it was not possible to use quantification by stereology. The heterogeneous distribution of follicles in the ovarian cortex warrants a cautious interpretation of the exact quantitative numbers reported. Moreover, only one 45,X specimen and no 45,X/46,XX specimens aged above GW 20 were available for this study, which unfortunately made it impossible to assess whether the ovarian folliculogenesis was delayed or absent in the TS and TS mosaic specimens. WIDER IMPLICATIONS OF THE FINDINGS: This human study provides insights about the timing of accelerated fetal germ cell loss in TS. Knowledge about the biological mechanism of POI in girls with TS is clinically useful when counseling patients about expected ovarian function and fertility preservation strategies. STUDY FUNDING/COMPETING INTEREST(S): This work was supported by the International Center for Research and Research Training in Endocrine Disruption of Male Reproduction and Child Health (EDMaRC).N/A.