You're right that it is difficult. Since there are multiple definitions of species, and many of them are not testable against, lets say, the fossil record, it's hard to pick one as being better than the others.
I think Para's definition is a good one. In this case, the fully speckled one could be a seperate species - BUT...
If only 1 of these super-speckled offspring is born, then they do not form a seperate population. We'd me more likely to classify the single offspring as a freak.
Part of the problem is in the scenario. The changes you suggest are clear and easy to understand, but are also much more rapid than they would happen in the wild.
An example I like to sight is jumping spiders. There are species of jumping spiders which look identicle, which have virtually identicle genetics, but which have slightly different mating dances. As a result a member of group 1 can not mate with a member of group 2.
Now, it's unlikely that the mating dance 1 became mating dance 2 in one or two steps, since the one step would make the performer unmatable.
But a tiny difference here, a tiny difference there. Enough to be different, but still get the point across. 6-7 of these small changes over a few decades and you've got a completely different species.