Perhaps, however, before entering on the subject-matter of this treatise, my definition of the terms "species" and "variety,"--so far at least as such is practicable,--will be expected of me. I may state, therefore, that I consider the _former_ to involve that ideal _relationship amongst all its members_ which the descent from a common parent can alone convey: whilst the _latter_ should be restricted, unless I am mistaken, to those various aberrations from their peculiar type which are sufficiently constant and isolated in their general character to _appear_, at first sight, to be distinct from it. The _first_ of these enunciations, it will be perceived, takes for granted the acceptance of a dogma which I am fully aware is open to much controversy and doubt,--namely, that of "specific centres of creation." Without, therefore, examining the evidences of that theory which would be out of place in these pages (and which has been so ably done already by the late Professor Edward Forbes), I would merely suggest that the admission of it is almost necessary, in order to convey to our minds any definite notion of the word "species" at all: and that, hence, whilst I would not wish to reject the hypothesis as involving an absurdity (which I believe to be the exact opposite of the truth), I would, in the present state of our knowledge, desire rather to regard it as a _postulate, assumed to illustrate the doctrine of species_, than as a problem capable of satisfactory demonstration. The _second_ of the above definitions may likewise require briefly commenting upon; for I have frequently heard it asserted that everything is to be regarded as a "variety" which has wandered in the smallest degree from its normal state. Now this I contend is essentially an error; for a "variety," to be technically such, must have in it the _prima-facie_ elements of stability,--and to an extent moreover that, without the intermediate links (which, although rarer than the variety itself, _must nevertheless exist_) to connect it with its parent stock, its condition is such that it might be registered as specifically distinct therefrom. Thus, to take an example for illustration, there are many darkly coloured insects which, as every entomologist knows, vary, by slow and regular gradations, into a pallid hue, sometimes into almost white. It also most frequently happens, in such instances, that the _extreme_ aberration is of more common occurrence than the intermediate ones. Here then is a case in point: there is but a _single_ variety involved, namely a pale one,--the gradually progressive shades which imperceptibly affiliate it with its type not being regarded in themselves as "varieties" at all. If this indeed were not so, then would our position be far from pleasant, since we should be compelled to record, as a variety, _every_ separate degree of colour which could possibly be found between the outer limits,--seeing that (increasing, as they did, in an