upon which natural selection has full opportunity to produce its effects. Originally it may have simply rested for a time in the direct line of development, on account of finding abundant food. But, as food conditions changed, new enemies attacked it, or old foes adopted new modes of assault, one of two things was necessary for its survival. It must either lose this resting-stage and develop continuously, or it must become adapted to the new conditions. This rendered necessary changes in instinct and in structure. Where the resting-stage, as in the caterpillar, occupied a very large percentage of the total life-duration, and where the process of adaptation had millions of years for its completion, it is not surprising that structural features often very divergent from the typical form were assumed.
There is little or no reason to doubt that all the peculiarities of larval form are due to the two causes here specified: 1. A temporary check to development at some ancestral stage of the animal's unfoldment. 2. An adaptive modification of structure and habit to meet varying conditions in the environment of this stage of development.
Yet in every such case we meet with a difficulty of the same character as that existing in the case of neuter ants. These secondary adaptations are out of the direct line of the animal's development, and it is a question how they can be hereditarily transmitted. The law of phylogenetic development enables us to understand the appearance of certain embryonal peculiarities of structure which do not exist in the mature form. If development is forced to follow its original line, such ancestral features must necessarily appear, though if the development is very rapid only hints of them are perceptible; or they may become utterly obliterated, so far as our powers of observation can decide. Yet such a principle can not apply to secondary structural features, produced in larval adaptation. The latter are in no sense in the direct ancestral line of development, and it is somewhat remarkable that they are so faithfully reproduced, only to be thrown aside again as the animal resumes its temporarily checked development.
It is very evident, from the facts here cited, that the phylogenetic line is subject to disturbing influences. There is no special reason, in the nature of things, why a developing animal should repeat every stage of its ancestral growth. If never disturbed in its development it would naturally do so, since its original evolution from primeval matter lay in that line, and there has been no force since brought to bear upon it to make it deviate. But where any subsequent force causes deviation, that deviation must become persistent. There can be no possible return to the exact ancestral course.!Many such deviations have occurred. Some of them are only apparently such, arising from rapidity of development, and the slurring over of intermediate steps in the line of growth. But many of them are results of subsequent adaptation. Such is the case with many of the peculiarities seen in the unfoldment of the mammalian embryo. It has deviated from
