ZenMonkey has neatly picked up on a key issue with Faith's speciation views that I think has gone unnoticed so far and that looks promising for clarifying the problems to Faith. Here he describes why it's unlikely for an allele to never be expressed:
ZenMonkey writes:
I have always assumed this to be true; no allele can be so recessive that it never manifests at all. Sooner or later you have to have two parents both contribute the recessive allele. If they didn't, if an allele were both relatively recessive and also rare, I deem it highly likely that drift would remove it eventually if it were neutral. On the other hand, if it did affect reproductive success, then natural selection would keep it from being rare if the related trait was beneficial or eliminate it if the trait was disadventagous. Do you want to contest this, or can we both accept that there are no hidden alleles, only relatively rare and relatively common ones? It seems obvious to me that diploid organisms, by definition, carry only two alleles for each gene and no more, and donate one and only one to any individual offspring it produces. There is no other place for an allele to hide. And yet you say this (emphasis mine):
Faith writes:
I start with the argument itself, the idea that you have a built-in complement of alleles, age unspecified, that are available in all species for making a huge array of interesting variations,
most of which never get expressed in this world, and do it simply by isolating portions of the gene pool, which is what ultimately brings about "speciation" and the inability to vary further along a particular genetic path.
Kudos!
--Percy