An epidemiological survey was conducted during a 4-month period of intense malaria transmission in Dielmo, a holoendemic Senegalese village. Two thick blood smears per inhabitant were collected weekly. The sex ratio of Plasmodium falciparum gametocytes (gamete precursors) was studied in 50 gametocyte carriers. All age classes were represented (mean 19·7 years; range: 2 months–75 years); 42 (84%) of them did not receive antimalarial treatment. Overall 668 thick smears were examined until 100 gametocytes had been counted or for 40 min. A total of 11 204 gametocytes were observed with a mean sex ratio of 0·346 (95% CI 0·317–0·374), i.e. 2·89 females per 1 male. Among the 284 thick smears in which at least 10 gametocytes were observed, the mean percentage of male gametocytes was 27·8%, with a range of 0–82%. Great variability was observed between gametocyte carriers and also between thick smears from the same gametocyte carrier. A multivariate analysis was performed which highlighted the fact that only 2 variables had a significant effect on the sex ratio. Anaemia was associated with an increased percentage of males (Prevalence Rate Ratio [PPR] of male gametocytes was multiplied by 1·65 if haematocrit rate <32%) and a wave of gametocytes was associated with an increased percentage of female gametocytes (PRR was multiplied by 0·48 during the peak of gametocytaemia and for the 2 weeks following this peak). The variables without significant effect on sex ratio were: age, sex, clinical status and sickle cell trait status of the gametocyte carrier, density of asexual parasites, quinine treatment, and gametocyte density (when taking account of its waves). These results are discussed in regard of possible differential production, mortality or sequestration of one gametocyte sex and selective advantages for the transmission of parasites.