In the second half of Chapter Five, Darwin discusses the differing degrees of variability found in related organisms. He claims that features exaggerated in one species compared to others of the same genera are more likely to vary within that species, than the same features found in entire genera for which none have the features in question exaggerated. In other words, he says (albeit in a great many more words) that evolution has not yet stopped playing with highly characteristic features: the effects of evolution are still visible after speciation. The features in question will eventually stabilise, but only after enough time has passed to allow the new species to have become new genera. This is an interesting idea, but I am not sure how well supported it is by modern data, and I have yet to look. It would be interesting to have this confirmed.
A related idea is that secondary sexual characteristics -- generally those characteristics visibly distinguishing male from female, but properly those characteristics not directly linked with reproduction (and represented as such by Darwin) -- vary within species more than those characteristics common to both sexes. Again, I am not familiar with the details here, but taking Darwin at his word, one might find some sense in this which Darwin himself seems not to have noted. Here I mean the fact that secondary sexual characteristics include those which individuals (typically but not always females) use in choosing from multiple suitors (typically but not always males) are in fact often species identifiers as well. A bird's song, for instance, carries many messages, which could include that the singer is male, well-fed, and looking for a mate -- but also, and always (at least when signalling specifically to members of the same species), that the singer is of a given species. While aspects of the song might vary from one individual to another (and they invariably will), any significant departures from the cues used to identify species will result in that individual failing to attract the attention of either mates or rivals. What features of a call characterise a species differs from one species to the next; and in the process of speciation, a consensus might form around a particular variant of some feature which previously had no species-descriptive function, and over time this consensus might result in that variant of that feature becoming "fixed" for the population. Ultimately, if the population becomes a distinct species, that feature would become a descriptor of the new species. In other words, the variation of secondary sexual characteristics (of which a bird's mating song is one) is itself one possible engine driving the generation of new species.
The remainder of the chapter features Darwin blindly groping about to connect his theory to his (mis)understanding of genetics. Specifically, he spends a good bit of it discussing reversions, which he rightly supposes to be preserved somehow in the makeup of individuals far removed from their ancestors displaying the features to which they have reverted. There really is not very much to say about this from a modern perspective, or even from a classical-genetics perspective: one may continue to remain completely ignorant of DNA and still be able to follow Mendel's and his successors' explanations. What Darwin does do well here, however, is point out how unrealistic the explanations for his examples must be in the context of what he calls special creation, the notion that each species was created separately. This was, of course, a major purpose to his work -- to establish evolution as a unifying principle in biology, and to show the uselessness of the competing theory of special creation. Much of his work is overkill here, but then the theory of evolution required a great deal of support to replace what had been asserted as truth for literally millennia before.