Did you know that many teleost fish species are sequentially hermaphroditic - starting life in one sex before switching to the other later on? When individuals start life as a male, but then become female this is called protandry; whereas females who later change to males are called protogynous.
One hypothesis for the evolution of protogyny is that in many species size provides a significant advantage to the mating success of males, but has little impact on mating outcome in females. As such, any mechanism allowing individuals to start life in female form (when they are typically small), but then "mature" into males once they have achieved large size, should be favored by natural selection. This evolutionary scenario is not nearly as implausible as it sounds because teleost fish (unlike most other vertebrates whose gonadal tissues differentiate early in development) develop their sex organs from a single, protogynous tissue type. This hypothesis for the evolution of protogyny has been dubbed the "size-advantage hypothesis."
A recent study by Erem Kazancιoǧlu and Suzanne Alonzo [2010; Evolution Accepted] uses phylogenetic comparative methods to examine the evolution of size-advantage and sequential hermaphroditism in labrid fishes: also known as the wrasses. What they find is that, indeed, the evolution of dioecy (separate sexes) from sequential hermaphroditism is relatively unlikely when the size-advantage of large males is high. However, their evidence for the evolution of protogyny from dioecious species with male size-advantage was somewhat ambiguous.
Although I enjoyed this paper quite a bit, and it seemed perfect fodder for some clever fun in photoshop (actually by E. Lu, see above), I also felt that that the study had some methodologically weak areas. For instance, the authors failed to take advantage of a new phylogenetic logistic regression procedure by Ives & Garland , which seems ideally suited to their data. (In their defense, the method is brand new.) Consequently, however, the authors found themselves of the unfortunate position of using an arbitrary scoring system to estimate size-related reproductive skew: adding 1 point for the presence of "pronounced sexual dichromatism," for example, and subtracting 1 point for "alternative reproductive tactics" (which might decrease the advantage of large male size) . With a phylogenetic multivariable logistic regression, the authors could have tested for an association between the log-odds of protogyny and each of their proxies for size-based reproductive skew (which also included sexual size dimorphism, resource defense, and mate defense), while simultaneously controlling for the phylogenetic non-independence of the species in their sample.
In spite of its limitations, I found this study to be a tremendously interesting read. Due in no small part to its unusual and "sexy" subject matter, I'm sure it is destined to attract the authors considerable attention - among evolutionary biologists and lay people alike.