Here is something a wrote for my website a while back:
One of the biggest questions about evolution today is, "How did sex evolve?" The first place we should start is the definition of sex: exchanging genetic material with other members of a species. Bacteria have a way of exchanging genetic material; although it is more akin to the copy and paste functions of a computer than it is to sex. The next step would be for a population to evolve which went through cycles of giving and recieving genetic material. This population would be akin to the species of frogs that can spontaneously change sex from male to female. The final step would be for individuals to be born of only a single sex (Males that stay as males and females that stay as females). Richard Dawkins describes roughly the same scenario in "The Ancestor's Tale".
Guess what? Now this picture of the evolution of sex is being partly confirmed. A newspiece on Livescience says:
We all came from hermaphrodites, organisms with both male and female reproductive organs. And though the origin traces back more than 100 million years, biologists have scratched their heads over how and why the separate male and female sexes evolved.
Now, research on wild strawberry plants is providing evidence for such a transition and the emergence of sex, at least in plants. And the results, which are detailed in the December issue of the journal Heredity, likely apply to animals like us, the researchers say.
The study showed that two genes located at different spots on a chromosome can cast strawberry offspring as a single sex, a hermaphrodite or a neuter (neither male nor female, and essentially sterile). The researchers suspect the two genes could be responsible for one of the earliest stages of the transition from asexual to sexual beings.
"All of the animals and plants that are bi-sexual, or have two sexes, are theorized to have evolved according to a particular set of steps," said researcher Kim Lewers, a plant geneticist at the USDA's Genetic Improvement of Fruits and Vegetables Lab in Maryland. "Until now, no example had been found of the very earliest steps. Therefore, those steps were undemonstrated to be true."
She added, "Finding this example of the very earliest stage allowed us to say the theory is probably right."