necessary physiological condition must arise at an early stage in the cycle of events which follow one another in ordered sequence and make towards the goal of reproduction, and that the behaviour to which it leads must be one of the earliest visible manifestations of the seasonal development of the sexual instinct. When does this seasonal development occur? For how long does the instinct lie dormant? In some species there is evidence of this first step in the process of reproduction early in February; there is reason to believe that in others the latter part of January is the period of revival; and the possibility must not be overlooked of still earlier awakenings, marked with little definiteness, though nevertheless of sufficient strength to call into functional activity the primary impulse in the sexual cycle. Here, then, we meet with a difficulty so far as direct observation is concerned, for the duration of the period of dormancy and the precise date of revival vary in different species; and, if accurate information is to be obtained, the study of the series of events which culminate in the attainment of reproduction ought certainly to begin the moment behaviour is influenced by the internal changes, whatever they may be, which are responsible for the awakening of the sexual instinct.
In considering how this difficulty might be met, the importance of migratory species as a channel of information was gradually borne in upon me; for it seemed that the definiteness
