In presenting the theory of the foregoing chapters I am aware that there is room for further investigation at many if not all points. But it is in the nature of science, and indeed the great merit of it, that when a substantially correct or valid hypothesis is obtained, instead of closing up the subject and completing our knowledge of it, the very opposite happens and we are confronted with numerous unanswered, though not necessarily unanswerable, questions that beforehand could not even be raised except perhaps in the vaguest terms. In attempting to assess at this stage the present theory of comets a number of considerations have to be borne in mind.

First to be remembered is the inherent difficulty of all cosmogonical problems, both from the mathematical point of view and from that of deriving valid hypotheses. The relevant processes must essentially involve questions of redistribution of the matter of the universe. A system beginning in a stable condition gradually evolves, or through external influence is caused to change, to some other state, and a passage from one stable form to some other form will result. Such a course of development must involve instability and dynamical processes of an irreversible nature, and nearly always will involve motion in the neighbourhood of neutral equilibrium conditions. As is well known, the mathematical difficulties associated with such questions are usually so great that apart from tracing the development up to the point where instability sets in, only considerations of a general nature can be advanced in attempting to decide the future course of the motion. The comet problem proves to be no exception to this. The motion of the cloud near the sun is, as we have seen, of a highly unstable character and subject to large possible changes in its details, but not in its general form, through any slight external disturbances such as must always be assumed to be present. Then again, within the accretion stream gravitational instability is an essential feature of the later development of comets. Thus at both its important stages we have in the process, what is indeed an essential requirement for the phenomena to be explained, a highly flexible mechanism subject to a large possible range of variation in the products that emerge from it, while all the time the mechanism itself preserving its same general character.