The explosion of cichlid diversification in the Rift Lakes is referred to as Sympatric speciation. In the case of the African Rift Lakes, you have significant variation of species in a single, isolated lake. Sexual selection is the likely explanation for the biodiversity we find in the Rift Lakes. Under ordinary conditions, species are going to mate with like species. Case in point, look at P. pundamilia and P. nyereria of Lake Victoria. These two species are visibly different in coloration, and the females will select mates of the same species under normal lighting conditions. But, if we place the two species under a monochromatic light that resembles turbid water conditions the females will select mates of either species. Under turbid water conditions, color cannot be differentiated. The barring of the two species is the same and the females cannot differentiate between the two different males.
In addition to different colors and patterns, you have subpopulations that exploit a particular niche not used by the parent population (i.e. mbuna, haps, and peacocks). These species will rarely come across one another in the wild, hence the chance of hybridization in the wild between mbuna and peacocks is small.
The point: mating can and will occur between different species. In the aquarium, we remove the niches of species found in the wild, but the distinct color patterns remain. By providing a sufficient number of females of a particular species, a male not only will cease to harass like he would a single female, but the choice of mating with a like species will significantly reduce the chance of mating with a dissimilar species and producing hybrid offspring. But it still happens.