Suppose we have a species of finch living in Europe before the last Ice Age and living on, say, various small seeds, pretty much as the goldfinch does now.
The shape and strength of this finch's beak will be determined by a few genes. Maybe one controlling the size and another controlling the muscles which work it. It really doesn't matter for our mind experiment exactly how many or what each does, we can think of them collectively as a 'beak gene' so long as we remember we are using the term 'gene' here as a shorthand for maybe a collection of genes.
As with any other gene there will be slight variations which will be inherited by offspring and which may make the beak better or worse at eating this or that food.
Now, imagine the ice sheets gradually extending from the north as the Ice Age sets in, and pushing the finch's range south, together with some of its food plants, some of which may well become extinct in the process, forcing the finch to adapt to other foods. Also, as its range moves south it may come into contact with new food plants which weren't available in the north.
Our finch will eventually find itself split into two or three isolated populations: one in the Iberian Peninsula; one in Italy and maybe a third in the Balkans, each with its own distinct mixture and availability of different foods.
These food plants will also be adapting, driven by the presence of our finches. Those seeds which have a harder case won't get eaten and so will produce more offspring with those harder seed cases than the soft-cased seeds and the finches with stronger beaks will be able to eat them, so the beaks will tend to get stouter and stronger. So we may have an arms-race developing in, for example Italy which leads to finches with short, strong beaks good at cracking seeds.
Meanwhile in Iberia, another food plant may be be more successful if it can protect its seed at the end of a tube which only the finches with the thinnest beaks can reach, so another arms-race may develop in which the winning finches are those with long, curved beaks.
And maybe in the Balkan Peninsula another arms race has produced finches which didn't need to change much from the original finch.
Now, thinking back to our 'beak gene': all that has happened in our three populations is that variations in this gene have been selected by the different environments in what creationists would call microevolution. Small steps at a time with each generation being filtered by the environment by natural selection so that gradually, and in line with changes in the finches' environments, differences have arisen in the three populations as each had adapted and become specialised for that populations environment. In one population one set of variant will have come to dominate in the gene pool; in another population, a different set will dominate.
Now, have we got three different species, three races of the same species, or three different subspecies? In fact, at that point in the finch's evolutionary history the question is entirely academic and of no biological significance whatsoever because the populations can't interbreed anyway, being physically isolated.
The test will come when the ice retreats and the populations move north again, together with their food plants. If their food plants don't extend their range northwards than the finches might not either, even if they could. But let's assume they do so.
Let's assume also that the only change in their genomes has been in the 'beak gene'. Admittedly, this is unlikely because other environmental factors will have been moulding other genes which will also be microevolving, but, to illustrate a point, let's just stick with the 'beak gene'.
If that has been the only change than the populations would almost certainly be able to interbreed, so at that point in their evolution they would at best be varieties or maybe subspecies, but what sort of beak would their offspring have? The probability is that they would have some sort of intermediate beak. But what use is an intermediate length stoutish beak when you need to reach seeds at the end of a long, thin tube? What use would it be for cracking tough seed husks?
What we would now have is an environment in which the offspring of those finches which DID interbreed were being selected out by starvation whilst anything which acted to prevent interbreeding would be highly favourable and so variations such as different display plumages, mating rituals, territorial songs, etc., which made interbreeding less likely would be favoured. As with 'beak genes', genes allowing interbreeding are now being selected against and variations of those same genes which inhibit it are being selected for, so changing their frequency in the respective gene pools, just as happened with 'beak genes' because of a different set of environmental forces.