Women exposed to the now-banned environmental pollutant polychlorinated biphenyls, or PCBs, are far more likely to give birth to girls. A long-range study of women exposed to PCBs in the 1950s and 60s has shown that the women were 33% less likely to have boy children. This may be a natural reaction that occurs in many species when the environment becomes more hazardous, since having more girls means the species can continue reproducing at a decent clip even if many babies die. [PhysOrg]
"Toxins" in general yielding more female births? No. Epidemiologists would have been hard pressed to miss that. The numbers would have gone bananas during the Industrial Revolution— but of course, they didn't.
What folks have noticed is a broad variety of similarly-structured chemicals sometimes participate in the same reactions hormones do. For example, when DDT starts screwing around with birds' reproductive cycle, you don't get more female births— you get thin eggshells. Biochemistry is complex; dropping a monkey wrench into the endocrine system may well have a sex-related effect, but what effect is a bit tricky to predict.
For an example in humans and/or if you want to scare yourself, note that the USA and India still use endosulfan:
Half-life in the environment of six months to nine years, an anti-androgen, and epidemiologically measurable effects in Kerala— so we're not talking about any of "Maybe", "If there's a spill", or "If someone drinks a gallon of tea each minute for a year".