On the other hand, the cause of the increase in length of the tubular corolla and of the butterfly’s ‘tongue,’ lies neither in the flower nor in the butterfly, but it is to be found in those other insects which visit the flower and steal its honey without being of any assistance in cross-fertilization. It may be stated shortly, that non-tubular corollas, with the honey freely exposed—for it must be assumed the ancestral form was of this kind—gradually developed into corollas with the honey deeply concealed. The whole process was presumably first started by the flower, for the gradual withdrawal of the honey to greater depths conferred the advantage of protection from rain (Hermann Müller), while larger quantities of honey could be stored up, and this would also increase the number of insects visiting the flower and render their visits more certain. As soon as this withdrawal occurred, the mouth-parts of insects began to be subjected to a selective process whereby these organs in some of them were lengthened at the same rate as that at which the honey was withdrawn. When once the process had begun, its continuance was ensured, for as soon as flower-frequenting insects were divided into two groups with short and with long mouth-parts respectively, a further increase in the length of the corolla-tube necessarily took place in all those flowers which were especially benefited by the assured visits of a relatively small number of species of insects, viz., those flowers in which cross-fertilization was more certainly performed in this way than by the uncertain visits of a great variety of species. This would imply that a still further increase in length would take place, for it is obvious that the cross-fertilization of any flower would be more certainly performed by an insect when the number of species of plants visited by it became less; and hence the cross-fertilization would be rendered most certain when the insect became completely adapted—in size, form, character of its surface, and the manner in which it obtained the honey—to the peculiarities of the flower. Those insects which obtain honey from a great variety of flowers are sure to waste a great part of the pollen by carrying it to the flowers of many different species, while insects which can only obtain honey from a few species of plants must necessarily visit many flowers of the same species one after the other, and they would therefore more generally distribute the pollen in an effective manner.