Frequency-dependent selection has not to do with gene flow. In evolution you have 4 "forces" mutation migration selection drift. Gene flow is part of migration processes, while frequency dependent selection is part of selection processes.

From what you say, my guess is that what people want to explain on haplochromine cichlids is how polymorphism is maintained.

Gene flow (which translates into introgressive hybridization) is one potential explanation, because classical selection occurs at a place, and reduces the number of phenotypes. But gene flow brings new phenotypes from outside the place, which maintains polymorphism.

Negative frequency-dependent selection is an alternative explanation, because it is a type of selection that maintains polymorphism: with this type of selection, phenotypes are selected because they are rare, so they never become rare enough to disapear, and polymorphism is maintained.

Could you please explain why negative frequency dependent selection would reduce gene flow between polymorphic variants? As I understand it, immigrants may have greater mating success because they are a rare type compared to the residents. This kind of negatively frequency-dependent selection should strengthen gene flow in comparison to Wright’s island model's assumption that immigrants have the same fitness as resident individuals, and that those resident females that mate with immigrant males receive equal fitness to females that mate with members of their own population.

I wonder whether there's confusion going on between negative frequency dependent selection (i.e. where the rarer morph does better due to lower predation, or better access to resources or some such force) and assortative mating (i.e. where each of the two morphs prefer to mate with their own morph). The latter would of course reduce gene flow between the morphs and drive the population towards speciation. However neither negative nor positive frequency dependent selection would do that on its own I don't think. Negative FDS would help prevent either morph from taking over entirely (i.e. going to fixation) so I suppose it would help by keeping both morphs present, but gene flow between the morphs would continue unless mating is non-random.

I agree with the first three responses that there is some confusion in the question. Negative FDS does indeed maintain polymorphisms within populations, but it does not support complete speciation. My point of difference with the previous responses is that negative FDS may in fact promote gene flow among morphs as successful mating is promoted among different morphs (disassortative mating). This may result in hybridization, which also promotes gene flow and can lead to introgression among morphs and species. Hybrid speciation also requires some clarification here. It is commonly associated with extreme phenotypes, habitats and genome/chromosomal variation creating new species with limited gene flow with the parental species. Both these points indicate that there is some confusion in the initial question, and perhaps this needs to be rephrased to assist the discussion

For haplochromine cichlids, I recall these suggestions both coming from the Seehausen lab. The role of introgression among related taxa colonising a lake via different river systems was seen as generating increased genetic diversity for future adaptive radiation. I dont think this was really proposed so much to act in the short-term actual speciation processes, except by just increasing standing genetic variation from which divergence might take place. Introgression could also occur later in a mature radiation bringing new combinations of adaptive traits into a single gene pool, again facilitating subsequent radiation. The only idea I can recall about frequency dependent selection was the rather specific idea that a rare colour morph of the male of a territorial species might recieve less aggression from other territorial males and thus have an advantage which might allow this morph to increase in abundance. If the same signal was used for mate choice by females and females had divergent preferences (some for each kind of male), it was proposed that this might facilitate sympatric speciation (or indeed co-existence of allopatric sister species on secodary contact). Or were you thinking of something else? As Aurelie says, FD selection is more often used to explain co-existence of polymorphisms in a single gene pool.

Yes, negative frequency-dependent selection might promote a rare introgressant morph. In order for this to be associated with a possible mechanism of speciation (i.e. promoting reproductive isolation between the two morphs), some kind of pleiotropic 'magic trait' facilitating both non-random mating and divergent selection (some differential ecological adaptation) would be necessary.

I do not think so,a negative frequency-dependent process will favour intermediate frequencies for alles/genes contributing to the polymoprphism. A negative frequency-dependent polymoprhism is know to favour negative assortative mating, and therefore it should in principle cause/favours the contrary process to speciation (i.e. positive assortative mating).

I apologize for the confussion on the initial question. It needs clarification indeed. I'm considering here the role of sexual selection on Haplochromine cichlids. In most of the cases I'm finding there is assortative mating, so there is a preference for conspecific mates. In intrasexual competition, conspecific males clash with each other for mating, there is no competition between non-conspecific males. In the context of intrasexual competition, by frequency-dependent selection we would find that migrants and hybrids will have advantage. However, considering assortative mating, I see difficult for those hybrids to mate as they are not conspecific to any females. Thus, the difficulty I find here is, how can hybridization promote speciation if, besides they will enjoy advantage by negative FDS, hybrids would not be sexually fitted?

I hope that clarifies the issue.

Basically, I think you are mixing up 3 completely different issues. #1. In classic allopatric/ parapatric speciation, hybridisation causes gene flow between diverging populations and retards speciation. If hybrids have lowered fitness, this will help speciation by reducing gene flow. Basically, the same will hold for sympatric speciation.The Seehausen idea is nothing to do with this. #2.Ecological speciation requires a lot of genetic variation in adaptive traits in the 'starter' population so that it can split into 2 different species. A population that underwent a period of hybridisation in the past might have more genetic variation than normal and so have more raw material for later divergence. There isnt really any proposal that CURRENT hybridisation between populations in the process of diverging accelerates speciation: it always suppresses speciation. The main article here is the Seehausen TREE review, but there are other more recent things. #3. The argument about male-male competition and FDS is completely unrelated to this. The main articles here are Schluter & Seehausen and the various experimental papers by Dijkstra et al. They consider a situation where there is polymorphism in male courtship signals and in female preferences for male signals. This can build up assortative mating which supposedly facilitates sympatric speciation, but it is expected to be helped out when there is some kind of ecological divergence too. But they also consider a case where male-male signalling might be involved too. The proposal is that the more distinct the male signals, the less aggression there will be between them. If each morph is more aggressive to its own kind, it will benefit when rare, hence FDS. Actually, I have a bit of trouble with this, as I would expect that in sympatric speciation, both morphs will be more aggressive to the 'ancestral' signal. I dont think there is actually clear data supporting this for intraspecific morphs, although there are plenty of studies showing differences between full species.As in #1, hybridisation between the different morphs will slow down speciation.