What, then, shall we understand by Spence’s phrase that instinct involves the performance of “certain actions”? And how far shall we accept it? We shall take it as implying so much definiteness of behaviour as renders instinctive acts susceptible of scientific investigation, and in this sense shall accept it with some modification of phraseology. We shall freely admit, however, the existence of variations of instinctive behaviour analogous to variations in animal structure. It is the occurrence of such variations that renders the natural selection of instinctive modes of behaviour conceivable. We shall also admit some, nay much, variation in detail. Take, for example, two of the cases which Mr. Marshall cites--nest-building and migration. Both involve, not merely a simple response to a given stimulus, but a complex sequence of actions. In detail there may be much variation even among members of the same species. And yet, can it be questioned that the behaviour as a whole is in each case relatively definite? May we not even say that it is remarkably definite? May we not even go further, and assert that only on the assumption that any given instinctive act is relatively definite, can we regard it as a subject for scientific investigation, and can we hope to distinguish it from other modes of behaviour? The next point for consideration in Spence’s definition, which we have taken as our text, is his characterization of instinctive acts as “tending to the well-being of the individual and the preservation of the species.” Here we have Mr. Marshall with us, for he too lays stress on the fact that instinctive behaviour has reference to a definite biological end. But in saying that the biological end is _the_ objective mark of an instinct,[28] he seems to be in error. Because, in the first place, there are other “objective marks,” and because, in the second place, this objective mark is not restricted to instinctive behaviour. According to Spence, a further characteristic of instinctive acts is that they are independent of instruction or experience; and this serves to differentiate them from other modes of behaviour which are also subservient to a biological end. Intelligent behaviour, not less than that which we term instinctive, has reference to a biological end. Many intelligent acts, for example, have for their object the well-being of the individual; many subserve race preservation; these bear, every whit as much as instinctive acts, the “objective mark” which Mr. Marshall regards as characteristic of instinct. And if we turn to his subjective criterion--the absence of any conception of the biological end which the behaviour subserves--Mr. Marshall’s position is equally untenable. There are thousands of acquired modes of behaviour, dependent on instruction or experience, in which there is, on the subjective side, so far as we can judge, no conception of the biological end to be attained. What can the animal in the early stages