exit from the craterlet it is three miles in breadth. It is about 120 miles long. Its apparent dimensions in a direction at right angles to the limb near which it is situated are considerably foreshortened. When near the terminator, it is seen that this riverbed originally took its origin in the crater of Herodotus itself, but that the exit was closed by a subsequent eruption of black material, nearly choking up the original valley. This cannot be seen as well when the Sun rises higher, and the valley actually at present terminates in the pear-shaped craterlet above mentioned. Although this formation is many times larger than any other member of its class, it is not so characteristic in form as some of the smaller ones, on account of its lacking the zigzag and winding appearance, which seems to have been nearly obliterated by its great width.
Returnng now to the smaller riverbeds, we find that they have a marked tendency to occur in groups. Thus, five are fotmd lying close together in the Harbinger Mountains, 13A [1.8, 5.0], and three more just to the east of them. One of the former group is forked, the two large southern branches uniting to form the northern one, which soon dwindles away and disappears. In all, thirty-five riverbeds, or possible riverbeds, have been detected upon the lunar surface. Of these, the next largest, after that upon Mount Hadley, occurs in Petavius, 2 A [2.8, 4.4]. Most of them are only a few miles in length, and a few hundred feet in width at the widest part; and unless they are pretty deep, they are very difficult objects to detect—undoubtedly, taken as a whole, the most difficult class of objects to be found upon the Moon.
