There is ongoing debate about how and why the menstrual cycle affects women’s attraction to men. According to the dual sexuality hypothesis, women form pair-bond relationships with men who provide care but also obtain genetic benefits by biasing mating effort towards men with high-fitness genes during the fertile phase. By contrast, the commitment hypothesis proposes that attachment bonds with primary partners function to strengthen pair-bond relationships by enhancing in-pair attraction at the fertile phase, rather than extra-pair attraction. We tested these hypotheses by measuring women’s daily sexual and emotional attraction towards men over the whole menstrual cycle. We employed (1) a urinary luteinizing hormone test to determine the day of ovulation, (2) a five-part classification of menstrual cycle that identifies a distinct peri-ovulatory phase, and (3) individualized phase identification for each participant. There was a mid-cycle rise in extra-pair sexual desire. Women gave and received more care from partners during the menstrual than the mid-cycle phases. Partner’s sexual attractiveness and mutual commitment did not moderate these findings. The results do not support either the dual sexuality or commitment hypotheses, and imply that female self-reported sexual desire is not strictly dependent on cyclic hormonal changes. Our results are more consistent with a recently proposed `spandrel’ hypothesis, positing cycle phase effects as a non-functional by-product of raised estradiol. Additionally, we found that, with the date of ovulation estimated by luteinizing hormone tests, 45% of ovulations were misclassified by the backward counting method, which urges caution in interpreting results based on counting methods.