Abstract
This thesis proposes that sympatric speciation can arise, in certain circumstances, in diploid sexually-reproducing organisms by mechanisms other than autopolyploidy. This proposition is verified by a series of computer simulations of the natural, biological, genetic processes of reproduction. The research is essentially a series of logical experiments using a computer simulation of natural genetic mechanisms of reproduction. In this methodology there is no direct comparison with biological data but the simulation follows, as closely as possible, natural genetic mechanisms of reproduction. The selected modelling method uses a framework derived from Penna 1995 which represents individual genes as binary digits on digital chromosomes and incorporates randomised simulations of recombination, copying-errors, meiosis and zygote production. The literature search identified some difficulties with defining species but a choice was made to use the `Biological Species Concept', which relies on reproductive isolation as evidence of speciation. Sympatric speciation by recombinational processes is said by some to be rare or unknown in nature, but sympatric speciation by autopolyploidy is said to be common in plants. The simulations reported here exclude the possibility of autopolyploidy and rely only on recombination and mutation processes. The model described by Penna 1995 was shown to be inadequate for representing multiple species. It was therefore modified to allow more genetic variability but to restrict mating to compatible phenotypes and to allow assortative mating. Simulations demonstrated that sympatric speciation can occur in scenarios based on the Dobzhansky-Muller model and that these processes can be facilitated by the intervention of selective predators. Another form of sympatric speciation was also detected where paired, dominant and recessive alleles lead to incompatible phenotypes. No reference has been found in the literature to modelling the Dobzhansky-Muller mechanism, except for Gavrilets 1997 who describes a mathematical analysis rather than a simulation.