Explaining the older-sister effect on male sexual orientation would need 37 to 58% of eggs miscarried

The two numbers this study puts next to each other are not measured the same way. The model's requirement, 37 to 58 per cent, is a share of fertilised eggs ^[12]. The figure usually quoted for the United States, 10 to 20 per cent, is a share of known pregnancies ^[14]. Set the bounds against each other and the model asks for between about 1.9 and 5.8 times the clinical rate ^[1]. Eggs that fertilise and never implant fall inside the study's denominator and outside the clinic's. Raymond said there is "huge variation, depending on what you consider a miscarriage" ^[15], and some losses happen before a woman knows she is pregnant ^[16].

The model asks only whether miscarriage on its own could generate an older-sister effect of the size eight populations show ^[3][9]. No antibody was measured. "However, it is also possible that the maternal immune hypothesis is not correct, and at this stage, it is not possible to decide," Raymond said ^[19]. He also said the work "shows that miscarriage is potentially an important phenomenon to take into account" ^[18]. The eight datasets did not routinely record whether the participating mothers included trans men and non-binary people ^[13].

The model gives every woman the same probability of miscarriage at every pregnancy ^[7], and Raymond said real risk probably varies between women even once age is accounted for ^[8]. If some women miscarry rarely, a run of older daughters may indicate a mother whose pregnancies mostly survived, which is the reverse of what the proxy needs.

Ray Blanchard at the University of Toronto, who had also proposed that older sisters stand in for miscarried male pregnancies, told New Scientist the study supports his idea ^[20]. "It could have come up with results saying that my hypothesis was basically very unlikely, but it came up with results indicating that, while my hypothesis may not 100 per cent account for the effect of older sisters, we're within the realm of what is possible," he said ^[21].

A 2017 study found that mothers of gay sons with older brothers had higher levels of antibodies against NLGN4Y, a Y-chromosome-linked protein involved in brain development, than mothers of heterosexual sons ^[23]. Elena Jazin at Uppsala University has documented NLGN4Y expression in human embryos of 8 to 11 weeks and says expression probably begins earlier ^[24]. How far a pregnancy has to progress before it can sensitise the mother is unresolved ^[22].

Jan Kabátek at the University of Melbourne has found that older siblings raise a woman's odds of entering a same-sex union, which a response to male-specific proteins does not readily explain ^[25]. Kabátek says the target could be proteins that are not sex specific ^[26]. That version would move the hypothesis off the Y chromosome and away from NLGN4Y. Identical twins share a sexual orientation more often than fraternal twins, which points to a genetic contribution as well ^[27].

Each additional older sister is linked to about a 20 per cent rise in the odds of a man being gay, roughly two-thirds the size of the 30 per cent associated with each older brother ^[2]. Raymond said of the miscarriage explanation: "It's nothing more than a possibility right now" ^[6].