We can hardly speculate with any degree of safety upon the cause of this varying intensity of volcanic discharge. We may ascribe it to variation of _depth_ of the initial disturbing force, or to suddenness of its action; or it may be that different degrees of fluidity of the lava have had modifying effects; or on the other hand different qualities of the crust-material; or yet again differences of period—the quieter extrusions having occurred at a time when the volcanic forces were dying down. There is an alliance between lunar craters and mountains that goes far to show that there has been no radical difference in their origins. For instance, as we have previously pointed out, craters in some cases run in linear groups, as if in those cases they had been formed along a line of disruption or of least resistance of the crust; and the mountain chains have a corresponding linear arrangement. Then we see craters and mountain chains disposed in what seem obviously the same arcs of disturbance. Thus Copernicus (No. 147), Erastothenes (No. 168), and the Apennines appear to belong to one continuous line of eruption; and it requires no great stretch of imagination to suppose that the Caucasus, Eudoxus (No. 208) and Aristotle (No. 209) form a continuation of the same line. Then around the Mare Serenetatis we see mountainous ridges and craters alternating one with the other as though the exuding action there, normally sufficient to produce the ridges, had at some points become forcible enough to produce a crater. Again, upon the very mountain ranges themselves, as for instance among the Apennines, we find small craters occurring. We see, too, that the great craters are in many cases surrounded by radiating systems of ridges which almost assume mountainous proportions, and which are doubtless exuded matter from “starred” cracks, the centres of which are occupied by the craters. The same kind of ridges here and there occur apart from craters (see for instance Plate XVIII., below Aristarchus and Herodotus) and sometimes they occur in the neighbourhood of extensive cracks, to which they also seem allied. We must indeed regard a linear crack as the origin either of a ridge (if the exudation is slight) or of a mountain chain (if the exudation is more copious) or a string of craters (if the extrusion rises to eruptive violence). But the subject of cracks is important enough to be treated in a separate chapter.