The maternal immune hypothesis (MIH) argues same sex attraction (SSA) results from maternal immune attack on fetal male-specific brain structures and involves the previous biological influence of elder brothers. One of the surveys supporting this is shown to be based on an unsuitable sample and to contain some strong contrary evidence. The hypothesis relies on at least four speculative ideas and there is evidence against each. (1) Likely immune response prevalence is too low compared with calculated SSA prevalence resulting from the fraternal birth order effect. (2) Testis immune attack would be more likely than brain attack but is not known. (3) Fetal brain structures are practically indistinguishable at birth and subsequent brain anatomical gender differentiation only occurs after birth when no attack is occurring. (4) The hypothesis also predicts unfavourable biology for late birth-order males but in fact the reverse is generally true, and neurological effects are very minor. Studies show aborted fetuses caused by likely maternal immune attack are predominantly girls rather than boys, which also argues against the theory. Studies on identical twins show that common factors such as uterine environment are only a small influence on SSA and post-natal idiosyncratic reactions and non-shared environmental factors are much larger influences.