This chapter, on the evolution of instinct and behaviour in animals, begins with an important distinction, one which has shown up before: Darwin is not interested here in its origin, rather in how and why it changes over time. This partitioning of a subject into different problems is standard fare in science today; we spend a lot of time in science learning how to judge which aspects of a topic are dependent upon others, which are addressable given the current state of the art (and our ability to access the state of the art!), which are likely to fit into our current research programme (the sorts of things that a grad student can propose to do being very different from those that a post-doc, faculty member, or researcher at an institute or in the private sector), which are likely to be attractive enough to others to be likely to get funding, and so on. For all his wordiness, Darwin is very clear: he will not so much as speculate about the origin of variation, even though that is a critical component of his model of evolution. It is a fantastically interesting question, for which there was no known mechanism in the 19th century; indeed, it was not until well into the 20th century that the first tiny steps were to be taken that would elucidate this problem. Meanwhile, Darwin assumes a model of inheritance which is now known to be completely wrong, but he does not propose to test it: while this was possible in the 19th century (and Gregor Mendel investigated just that at the same time as Darwin was working on natural selection), it involved a great deal of time and effort, and Darwin recognised this (consciously or otherwise) and steered clear of it. Darwin's insight is actually very specific, dealing with a comparatively small aspect of life, and yet he recognised that it was a powerful and flexible concept in spite of the mechanisms for its implementation being (in his time) unknowable or uninvestigated. This is a sign of a remarkable mind, and the best that one can hope for in science: to recognise one's limits, to make the most of what one has available, to minimise the liabilities inherent in one's deficits, and to combine one's own data with that of one's forebears and contemporaries to produce a greater understanding of the world than existed before.
The delineations continue throughout the chapter. The "habit"/"instinct" dichotomy is an exact parallel to the "nature"/"nurture" dualism in humans, but Darwin does not pick sides here; he takes a more balanced approach, judging that each plays a role in every situation, but to varying extents. At the same time, he is careful to separate instinct from reason, and (as most scientists do today) to avoid seeing human capacities in non-human animals. Many of his arguments seem to the modern reader worryingly anecdotal, but much of the study of behaviour in animals is necessarily so. Ethology, like ecology and evolution, had yet to become its own field; Darwin was treading upon virgin territory. He was right to be circumspect. But for all his appeals to folksy common sense, his distinctions are fundamentally sound. He moves the discussion quickly away from the dichotomy between instinct and reason to that between instinct and habit, a much more neutrally approachable topic. Habits are flexible and learned; instincts are stereotyped and inherited. Darwin's latent Lamarckianism does surface here and there, offering the notion that what one organism learns can be passed on to its offspring without being taught, but he does not make so much of this that it drives his theory.
The most important insight in this chapter is Darwin's recognition of instinct as varying and heritable, and thus evolvable -- fundamentally no different from any other character subject to the forces of evolution. He uses this rightly to extend his corollaries on evolution to instinct as well, specifically to the concept that natural selection will act in favour only of the organism possessing the trait being acted upon. Like in earlier chapters, he discusses examples from domesticated animals, arguing that the processes by which humans produced various breeds from a single ancestral stock are the same as those at work in nature. As with much in the Origin, the arguments seem "soft": superficial, suppositional, anecdotal. But to find fault with this is to pursue the wrong problem. Darwin deals with several examples, around which he can occasionally justifiably be accused of spinning "just-so" stories, but his intention is not to show how things happened so much as how they could reasonably be supposed to have happened. His burden of proof, in other words, is very low. "Just-so" stories are completely acceptable, so long as he can demonstrate that his naturalistic examples are more plausible than the creationist alternative.
In fact, the details of his "just-so" stories are largely irrelevant. His supposition of brood-parasitism as initially a habit which became ingrained is mostly wrong, and his examples involving ants demonstrate a complete ignorance of the action of pheremones. He misses an obvious "just-so" story involving ant slaves: the pupae of the future "slave" ants could have been taken in a raid for food, matured and emerged while in storage, and the resulting workers put to use in their "adoptive" colony. But the strength of Darwin's theory is such that these details do not need to be correct. Rather, the principle behind them is the important thing: so long as some aspect of life varies, so long as those variations are inherited, so long as those variations affect the differential survival of individuals expressing them, so long as more individuals are present than can survive and reproduce, the population in question will evolve. How those traits come to vary, be inherited, or affect the survival of their possessors does not matter. What is important is Darwin's awareness of the existence and import of a spectrum of graded states from putatively ancestral ("primitive") through highly derived ("advanced"); from incipient through fully developed, from occasionally useful (facultative) through absolutely necessary (obligate). This argument from homology is one of the most powerful in favour of evolution even today, and this easily makes up for all Darwin's use of anecdotes and supposition. His application of this process to non-material traits such as behaviour is all the more remarkable.