largest were produced in the throes of its early violence, and that the smallest are the results of expiring efforts possibly impeded through the deep-seatedness of the ejective source. Another general fact of this chronological order is that the mountain chains are never seen to intrude upon formations of the crater order. We do not anywhere find that a mountain chain runs absolutely into or through a crater; but, on the other hand, we do find that craters have formed on mountain chains. This leads unmistakably to the inference that the craters were not formed _before_ their allied mountain chains; and we might assume therefore that the mountains generally are the older formations, but that there is nothing to prove that the two classes of features, where they intermingle, as in the Apennines and Caucasus, were not erupted cotemporaneously. [Illustration: PLATE XX. OVERLAPPING CRATERS.] Upon the assumption that the latest ejected or extruded matter is that which is brightest, we should place the bright streaks among the more recent features. Be this as it may, it is tolerably certain that the cracks, whose apparently close relation to the radiating streaks we have endeavoured to point out, are relatively of a very late formative period. We are indeed disposed to consider them as the most recent features of all: the evidence in support of this consideration being the fact that they are sometimes found intersecting small craters that, from the way in which they are cut through by the cracks, must have been _in situ_ before the cracking agency came into operation. It is in accordance with our hypothesis of the moon’s transition from a fluid to a solid body to consider that a cracking of the surface would be the latest of all the phenomena produced by contraction in final cooling.