A major hypothesis to explain the causal initiation of protogynous sex reversal is that females change sex upon reaching a critical size. A study of the coral reef fish Anthias squamipinnis shows that the size hypothesis does not hold. Females from two neighbouring, but spatially discrete and probably genetically homogeneous populations on Aldabra Island changed sex at distinctly different sizes. Previous laboratory and field studies in which sex reversal has followed the removal of a male from social groups have been uncontrolled and thus permit the interpretation that sex reversal is caused by non-specific social disruption or by causes other than male removal. In this study, a male was removed from each of eleven single-male and five multi-male social groups in the laboratory (N= 8 male removals) and in the field (N= 19 male removals). In each group, the result was that one female changed sex. Laboratory controls made it unlikely that sex reversal was induced by non-specific disruption and field observations showed that sex reversals resulted from male removals and were not coincidental, ongoing events. Previous statements that sex change is controlled by the presence or absence of a male, by inhibition of a female's tendency to change sex, or by aggression or dominance are shown, by an analysis of the complexity of issues, to be premature. Gonadal histology on 130 specimens confirmed that this species is a monandric, protogynous hermaphrodite and provided details of gonadal transformation.