Having rounded out the basis of his theory, Darwin now proceeds on possible objections to it. He begins this chapter, on "Difficulties on [the] Theory", with a list of four main topics: the nature of transitional forms, and the fact that we do not observe them now; the massive degree of transformation that occured (to give Darwin's example) between the forelimb of the ancestral mammal (much like a modern shrew or opossum) and the wing of the bat, as well as the capacity of natural selection to produce both seemingly inconsequential and highly developed body parts; instinct, and how natural selection might affect behaviour; and what we now call the reproductive barrier between species, which is not apparent when subspecific crosses are made. The latter two points he plans on treating in later chapters.
On the question of transitional forms, Darwin first mentions the trivial case of completely separated populations of the same species adapting to local conditions and thereby forming new species. This he judges to be trivial and worthy of no further discussion, although he does not deny that it can be and has been a potent force for speciation. Rather, he is concerned with the fact that many obviously related species occur with overlapping ranges. He notes that it is possible and even likely that these ranges did not always overlap; changes in sea level may create islands or join them to the mainland, for instance, isolating populations temporarily but long enough to speciate, and then rejoining them with their now-non-conspecific relatives. Again, however, he is not interested in pursuing this line of reasoning; he does not expect us to believe (nor does he believe himself) that something like this has been the case with all closely-related species. Instead he brings up the topics central to ecology: species must relate not just to their physical environments but also to one another; they may eat or be eaten by other species, and they almost certainly will compete with other species for resources. The assemblage of species itself is a stabilising influence, and will tend to limit the amount of acceptable variation within any one component. Also, Darwin posits that for any two populations connected by an intermediate form, the intermediate form's range will be small, and its population size likewise. I am not sure that this is in fact the case, but certainly species specialised for a given area (the extreme populations) will do better in those areas than generalists (the intermediate form); each form will be able to compete (at a disadvantage, but nevertheless) for resources in the adjacent zone, which means that the intermediate form will be dealing with competition from both extreme forms.
Next Darwin comes to a very important insight: evolution does not produce perfection. However miraculously ideal something might look, it only needs to be good enough to work. There is always something better possible, and when it comes along (as it tends to), it will displace the good-enough equivalent of its now-transitional parent. Another point, which I am not sure of myself but which makes sense to me, is that newly adapted features do not lend themselves to an adaptive radiation until they have diverged considerably from their parent type. To put it another way, adaptive radiations do not occur until the transition has completed. This explains the rarity of transitional forms: there will be only one or a few species of a transitional type in existence, compared with countless more already-adapted forms, so the number of individuals available to be preserved is correspondingly smaller. A final point about transitional features is that they may not appear transitional at first. They may lose their function (or, to anticipate another point that Darwin gets to later, change their function) without changing their structure, and thereby be indistinguishable from their predecessors when in fact considerable adaptation is happening elsewhere in the organism.
Now Darwin reaches one of the most famous parts of the Origin: "Organs of Extreme Perfection and Complication". This section focuses primarily on the evolution of the eye, and begins with a line to set the stage that Creationists are fond of quoting to claim that Darwin himself did not believe evolution capable of producing such an organ. Creationists fail to mention the very next sentence, wherein Darwin explains how this only seems to be the case, and sets forth a series of transitions -- with corresponding forms still extant -- through which evolution might arrive at the present state! As he develops his refutation of this objection, he makes a statement in passing that explicitly states what natural selection cannot do: "How a nerve comes to be sensitive to light, hardly concerns us more than how life itself first originated…." This is an important distinction between the source of a feature and its subsequent evolution. The comparison to the origin of life is an apt one: much research is presently being done on it, but the processes involved, although called evolution, are very different from those at play once life has become established. Beck to the eye, Darwin notes that transitional forms between primitive light sensors and image-forming eyes are abundant. He even proposes research into evolving an eye, a simulation which has actually been done! The project in question* concluded that, even with very conservative parameters, such a transition could take place in about half a million years, which is considerably less than the "millions and millions" proposed by Darwin! Along the way, such transitional forms will be considerably generalised compared to their extant equivalents, and Darwin astutely notes that such generalised organs often have multiple functions. Darwin's example is a crustacean that respires using its digestive system; I do not know whether this is physiologically accurate, but the point is a reasonable one, and several other organisms have similarly multifunctional systems. Such principles inevitably bring up a concept articulated in its modern form a decade earlier by Richard Owen: that of homology (or, to use the Victorian equivalent that even he used little, "ideal similarity"). Darwin here argues that homologous structures in different species look similar because they descend from a common ancestor, an important point, although perhaps not obviously so. A more radical point is that many organisms have features which sometimes strongly resemble those of others to which they are not closely related. The evolutionary process leading to this is now called "convergence", and the resemblance itself "homoplasy"; and here Darwin correctly points out that it is always, at some level, fundamentally distinct from homology, and detectable as such. He makes reference to a Latin saying that he claims is current: "natura non facit saltum" ("nature does not make leaps"), to indicate not so much that he is a committed gradualist (although he is) so much as to say that everything has a transitional form, however fleeting.
The next major topic deals with the opposite of highly specialised and adapted features: "Organs of Small Importance". Although this sounds like typical Victorian hyperbole, Darwin claims that this is as big an issue to him as the prior topic. He has dealt with this subject before: such features are possibly vestigial, developing from well-developed functional forms, or they may be correlated with important organs which have driven them to a reduced or less-functional state. Importantly, he notes the possibility of misidentification, in which case functionally important features can appear structurally reduced. He also cautions against what Stephen Jay Gould called "the Panglossian Paradigm", the point that because we see something useful in one context does not mean that it origiinally appeared in that context. In many cases this leads to what is now called preadaptation (or, in an effort to make evolution sound less deterministic, exaptation). One good example that Darwin gives is that of the skull plates of mammals, which in the young are not fused, and thereby impart some flexibility to the head which facilitates the passage of soon-to-be-newborns through their mother's reproductive tract. Such features, although invaluable now, may arise from some unknown state of little or no function. Another, out of which he gets more mileage, is the tail in most mammals, which is in most cases nowhere near as critically important an organ as it is to a fish, in which it generally provides the primary force for propulsion. A similarly unknown solution (to Darwin, anyway) often lies behind "chicken-and-egg" problems, such as (in Darwin's day) the origin of feathers and flight in birds. Darwin brings the point up but has little to say on it. Other features which may seem of little importance in an organism's "struggle for existence" are judged striking by humans, the opinion being that "very many structures have been created for beauty in the eyes of man, or for mere variety." If true, this would be a perfect counterexample to natural selection. Darwin supplies a litany of possible explanations how this is nowhere the case, but ultimately stresses that inheritance is the most important factor. Organisms are the way they are because they inherit their features from their parents, who may or may not have used their bodies in the same way. This leads to a reiteration of a critical principle: features are of utility to the possessing species alone. This is actually a restatement and generalisation of the earlier point, and like it any proven exceptions would be fatal (to use Darwin's term) to the theory, but in reality nonexistant. However, he goes on, one species can take advantage of traits in another, giving the appearance of those traits having evolved for the one taking advantage. In all cases, that trait was originally of more use to its possessor than anything else. This is a foreshadowing of Leigh van Valen's Red Queen hypothesis, and leads to a reiteration of the important point that natural selection only makes things "good enough"; except that here the point is made in the context of other species likewise evolving. If one species takes advantage of a trait of another, that other species (if it is to survive and the advantage taken is sufficiently negative to it) will develop some feature to prevent the first from taking such advantage, and the first will develop some means of getting around that feature, and so on. This is the evolutionary equivalent of an arms race, and can lead to spectacular coadaptations. Meanwhile, to close the point on things having been made for humanity's (or anyone else's) enjoyment, Darwin emphasises the mechanistic aspects of the theory. This is to say that all of this happens without any guidance or predetermined outcome: God is not necessary.
The summary offers a surprisingly concise reiteration of the major points of the chapter, and offers nothing new until the last paragraph. Here Darwin deals with two contemporary terms: "Unity of Type" and "Conditions of Existence". The former he sees to encapsulate homology (and I expect that this was uncontroversial; homology was a new but established idea), and is explained by common descent. The latter Darwin believes to be equivalent to natural selection, although from his brief (and possibly inadequate) description it sounds more properly like what we now call autecology, or an organism's natural history (definitely not what Darwin would have called it, "natural history" being the contemporary term for all of biology!). Whatever one calls it, it is obvious that it influences natural selection.
Next up is an in-depth discussion of another objection to the theory, concerning instinct, or specifically the evolution of animal behaviour.
* Nilsson, D.-E.; Pelger, S. (1994): "A Pessimistic Estimate of the Time Required for an Eye to Evolve". Proceedings: Biological Sciences, 256(1435):53-58.