There can be no doubt that islands are, for the most part, more unproductive (even in proportion) than continents; and that, the smaller the area, the less favourable will it be for the development of insect life. Mr. Darwin has noticed this fact in the Galapagos (which he remarks are only equalled by Tierra del Fuego, in barrenness), on Keeling Island (in the Indian Ocean), where he succeeded in detecting but thirteen species, in St. Helena, and at Ascension; and I have added fresh evidence to the same in the various portions of the Madeiran Group[21]. It is however to geological causes that we must mainly look for the explanation of this phaenomenon; and, therefore, since I propose to examine that branch of our subject in a future chapter of this treatise, we will not discuss it now. It will also be better perhaps to defer for the present the general question of self-_diffusion_, which, at the opening of this section, we proposed to consider, along with that of insect _productiveness_ (as dependent on other local influences, besides climatal ones),--it being scarcely possible to render the problem of dispersion in any degree intelligible without calling in geology to our aid. Having then disposed of this preliminary appendage to our inquiry, by expressing our belief (which I am satisfied that observation will tend more and more to corroborate) _that certain countries and spots are by constitution more favourable than others for the increase_ (apart from the after dissemination) _of the insect tribes_,--and that too through local influences amongst which mere heat and cold are but secondary in importance; let us proceed to consider, how far the _nature of the several districts_ may assist us in accounting for some of those numerous aberrations from the typical state which various insects are accustomed to display, and on which it has too often happened that "species" (so called) have been attempted to be established. I may premise however, that, whilst (as already urged) I would regard climate _per se_ as subsidiary to many other agents, I would not wish to ignore its action altogether even under the present section, since in combination with peculiar circumstances and conditions it may have (and probably has) considerable controlling power: nevertheless I would desire it to be looked upon here as, at any rate, an inferior element, and as working in conjunction with physical influences of greater significance than itself. If therefore under the preceding heads it has been treated (so far at least as the exceptions would permit) as a great geographical principle, possessing a certain modifying quality on a large scale, let us now merely recognize it to the extent in which we are actually compelled to do, when dealing with areas of smaller magnitude,--namely as a _topo_graphical one.