Although the biological mechanism remains unclear, male fetuses appear more sensitive than female fetuses to maternal corticosteroids produced after the twentieth week of gestation. This elevated stress reactivity apparently jeopardizes the viability of males in utero. Consistent with the theory of natural selection, humans may have conserved this male fetal sensitivity to maximize the mother’s total yield of grandchildren.The authors looked at fetal death statistics for pregnancies over 20 weeks from 2006 through 2001 (of which it's estimated that only 20 - 30% are actually reported) and found that the number of deaths of male fetuses was higher than expected in September 2001. However, relative to the total number of pregnancies, the numbers were very small indeed (in the hundreds).
Now, the authors also cite reports in the literature that suggest that major pollution events or economic downturns are correlated with male fetal loss as well, which means that the effect--if it's real--could be confounded. And indeed, in the study population Bruckner et al. used for their analysis, more than half of the mothers had a high school or lower education level, suggesting that economic stressors would be chronic for these women. If the reported effect on sex ratio is real, why isn't the sex ratio among poor women always skewed more in favor of girls, then?
reported vital statistics and see no increase in miscarriages in either 2001 or 2002. If the effect of communal bereavement is as the authors report, this strikes us as unexpected. Indeed, this would reflect an increase in male fetal deaths but with no increase in total deaths, there must have been a corresponding decrease in female deaths. Or, a trivial effect of 'communal bereavement' on fetal deaths.
According to demographic theory, there are typically more males conceived than females, more boys born than girls, and higher male fetal death rate in general. One explanation for this is that the Y-bearing sperm are lighter and hence swim faster than X-bearing sperm (the X is a much larger and hence heavier bit of baggage to carry). But risk of death for male embryos is higher in the best of times. At birth, the sex ratio is about 107:100, but male infants die at a higher rate than females so--the theory and some data go--the sex ratio is about even at puberty when matings start.
But, this brings us to the evolutionary importance of this paper. While claiming such significance, the authors cite just one paper on this, a 1973 Trivers and Willard paper in Science that argued that natural selection favors deviations away from equal parental investment in boys and girls, "rather than deviations in sex ratios per se." But this is based on a model applied primarily to deer. They say it's complicated in humans, and suggest that in humans the "sex ratio at birth correlates with socioeconomic status." They also suggest that differential mortality takes place early in pregnancy, and that it's chronic poor maternal condition that makes the difference.
Now, whether or not the Trivers and Willard model has held up over 40 years we don't know because we don't know that literature, but it clearly doesn't apply to the Bruckner et al. paper. It's not surprising though that an adaptive explanation for the finding--itself so tenuous--is tenuous at best, but it's not unusual that the authors try to claim one anyway. In addition, there is a large literature on an increase in sex ratio--more boys per birth--after traumas like war; the sociobiological explanation has been that this allows a society to make up for valiant heroes lost in combat; whether this idea has stood the test of time we can't say.
The problem here is the usual one of making claims that may be plausible, but go far beyond the data, and the data are inconsistent. But if you simply must have an evolutionary reason, then of course you can always make one up. The problem is that these things are so often empty of seriously rigorous content.