http://www.pnas.org/content/early/2013/04/03/1219361110.abstract.html Симпатичная статья.
Там pdf. Картинки выковыривать лень. Если в "циферках":
"A model using only linear and quadratic selection on the shoulder-to-hip ratio
accounted for 79.6% of variation in relative attractiveness scores... The explanatory power of height and penis size when added separately to this model was almost identical. Both traits significantly improved the fit of the model... Each trait, respectively, explained an extra 6.1% and 5.1% of the total variation in relative attractiveness."
По наводке:
http://overscience-mes.livejournal.com/304238.html?view=7599726#t7599726