Verified by 
I recently came across another very interesting paper whilst conducting some research for my forthcoming trade book tentatively titled The Consuming Instinct: What Juicy Burgers, Ferraris, Pornography, and Gift Giving Reveal About Human Nature (Prometheus Books, 2011). Fiona Mathews, Paul J. Johnson, and Andrew Neil investigated women's diets prior to falling pregnant to gauge whether it might have an effect on the sex of the offspring. The sample size consisted of 740 British nulliparous women (i.e., yet to have any children) who agreed to provide detailed information about their diets at three stages, namely preconception (covering one year), and early and late gestation. It is important to note that the mothers did not know the sex of their babies except a few who found out this information through amniocentesis or because of abnormal scans. That said this information only occurred later in the pregnancy so for the preconception and early pregnancy data, women were indeed blind to the sex of their child (and hence they did not alter their diets as a response to such information).
The goal was to test the Trivers-Willard hypothesis, which posits that depending on specific parental and/or environmental conditions, natural selection should favor: (1) biased offspring sex ratio; (2) biased differential investment in sons and daughters postnatally. Note that the Trivers-Willard hypothesis posits that given the greater reproductive variance of males, when conditions are good, it makes evolutionary sense to have biased investments toward males. In the Matthews et al. paper, the objective was to test whether the richness of a woman's diet would augment the likelihood of producing sons (i.e., a biased offspring sex ratio).
I shall only report on the preconception data for they are the ones relevant to testing biased offspring sex ratio. The nutritional data was factor-analyzed and two key factors emerged, namely factor 1 captured diets possessing high nutrient levels (protein, fat, Vitamin C, folate, and several minerals including iron, zinc, and potassium) whereas factor 2 corresponded to diets rich in Vitamin A (retinol) and Vitamin B12. Here are the key results:
(1) Only a woman's score on Factor 1 was predictive of the sex of the fetus (p < .001). Specifically, the richer the diet, the more likely was the sex to be male.
(2) Women who scored in the top third of energy intake were 1.5 times more likely to have a son than their counterparts in the lowest third.
(3) When individual food items were analyzed, the only one that was related to fetal sex was cereal! More cereal consumption implies a greater likelihood of male babies.
You might ask yourself what are the mechanisms by which a woman's diet might affect the sex of an offspring (see the following paper for some postulated mechanisms linking parental/environmental conditions and fetal sex). The working hypothesis is that glucose seems to "favor" male conceptuses more so than their female counterparts, so it might be the case that the spermatozoa carrying the Y chromosome (male) get an edge in "rich" dietary environments.
The power of consumer behavior is apparently limitless!
Source for Image:
http://mrbadak.com/wp-content/uploads/2009/06/breakfast.jpg
