rks, constructed presumably
by Kepler's _sub-volvani_, or by other intelligences. There is perhaps
some excuse to be made for the freaks of an exuberant fancy in regard to
objects which, if we ignore for a moment their enormous dimensions,
judged by a terrestrial standard, certainly have, in their apparent
absence of any physical relation to neighbouring objects, all the
appearance of being works of art rather than of nature. The keen-sighted
and very imaginative Gruithuisen believed that in some instances they
represent roads cut through interminable forests, and in others the
dried-up beds of once mighty rivers. His description of the Triesnecker
rill-system reads like a page from a geographical primer. A portion of it
is compared to the river Po, and he traces its course mile by mile up to
the "delta" at its place of disemboguement into the Mare Vaporum. From
the position of some rills with respect to the contour of the surrounding
country, it is evident that if water were now present on the moon, they,
being situated at the lowest level, would form natural channels for its
reception; but the exceptions to this arrangement are so numerous and
obvious, that the idea may be at once dismissed that there is any analogy
between them and our rivers. The eminent selenographer, the late W.K.
Birt, compared many of them to "inverted river-beds" from the fact that,
as often as not, they become broader and deeper as they attain a higher
level. The branches resemble rivers more frequently than the main
channels; for they generally commence as very fine grooves, and, becoming
broader and broader, join them at an acute angle. An attempt again has
been made to compare the lunar clefts with those vast gorges, the
marvellous results of aqueous action, called canyons, which attain their
greatest dimensions in North America; such as the Great Canyon of the
Colorado, which is at least 300 miles in length, and in places 2000 yards
in depth, with perpendicular or even overhanging sides; but the analogy,
at first sight specious, utterly breaks down under closer examination.
Some selenographers consider them to consist of long-extending rows of
confluent craters, too minute to be separately distinguished, and to be
thus due to some kind of volcanic action. This is undoubtedly true in
many instances, for almost every lunar region affords examples of crater-
rows merging by almost imperceptible gradations into cleft-like features,
and crater-ro