In a small group (say 60 individuals) stochastic variation in sex at birth can have significan effect. For example a run of 4 or 5 boys in a row can really make the sex ratio in the group unbalanced - the boys will be hard pressed to find marriage partners. And incest prohibitions will eliminate siblings and frequently first cousins as eligible mates. So when expanding into a new territory that is uninhabited, small bands can die out after several generations.
I am trying to get a hold of a simulation some archaeologists
wrote in C++ called ETHNOPOP which simulates the growth (and frequent
extinction) of small hunting and gathering bands.