There are also some hints that the MTF and FTM cases may not be completely symmetrical: a few years ago there was a study (M. Gahr, 'Male Japanese quails with female brains do not show male sexual behaviors', PNAS 100:13 (2003), p. 7959--7964) where parts of the brain were transplanted from one bird embryo to another. In that study the 'FtM' birds behaved as females but the 'MtF' ones did not behave as males.
This is still very preliminary, though, and it's not at all obvious that the same applies to humans -- especially since the old statistics can be explained rather well by cultural differences and the very different surgical options. Also, the numbers are pretty low, and this too is likely to be because the cultural climate and medical options have changed in the past couple of decades.
As for the prevalence of transsexuality, Olyslager and Conway ('On the Calculation of the Prevalence of Transsexualism', WPATH 20th International Symposium, 2007) suggest something like 1:2000--1:4500 for MtF and 1:5500--1:8000 for FtM -- and this for transsexuality that has resulted in surgery, with estimates as low as 1:100 or more for the various transgendered conditions put together. The Wikipedia entry for 'transsexualism' has some pointers.
So, anyway, it's something between 1:100,000 and 1:100 for FtM, and between 1:35,000 and 1:100 for MtF. Helpful, eh?

Nfr