my expectation, written down before seeing more than seing one (naive) answer, is: greater than 50%. But not necessarily by a lot at first. continuing it on indefinitely, if our population is sufficiently large, will eventually result in a population dominated by females, depending heavily on the details of the distribution of certain abnormalities (aka is there even one two-headed or two-tailed coin, and if not, then what is the exact nature of graph in that region, at which point it becomes ridiculously complicated...)

terribly sloppy notation follows. E(geneticconditions):(chanceofgivingfemalebirth>malebirth). by (havingchildren untill maleborn): (amplifies the prevelance of such conditions in birthing populations.).

But then, i've heard this puzzle before and have had a LONG time to muse about it. 50% is wrong if we're talking about beings with a biology resembling our own. :P

50% is the correct expected distribution, however, if they're using perfect quantum coinflips.

aaanndd...

(please forgive the spelling errors; i just looked at the clock.)