What are these mysterious ray systems? How are we to explain the circumstance that though only the most tremendous forces seem competent to account for bands such as these, many miles broad and many hundreds of miles in length, there are yet none of the usual signs of the action of volcanic forces? If mighty rifts had been formed in the moon's crust by the outbursting action of a hot nucleus, or through the contraction of the crust on an unyielding nucleus (for the effect would be the same in either case), we should scarcely expect to find that after such rifts had formed no signs of any difference of level would appear. If lava flowed out all along the rift, one would imagine that it would form a long dyke which would throw an obvious shadow, except under full solar illumination. If the rift were not filled with lava, the bottom of the rift would certainly remain in shadow long after the surrounding region was illuminated. That lava should exactly fill the rift along its entire length seems incredible. This might happen by a strange chance in the case of a single rift, but not with all the rifts of a radiating system, still less with all the rifts of all the radiating systems. Yet I believe that neither astronomers nor geologists can form any other opinion respecting these mysterious ray-systems than that they were caused by what Humboldt (speaking of the earth) calls the reaction of the interior on the crust. Nasmyth has admirably indicated their appearance, or rather their radiating form, by filling a globular glass shell with water, hermetically closed, and then freezing the water. The expansion of the water bursts the glass shell, and the lines of fracture are found to extend in a series of rays from the part of the shell which first gave way. But this experiment of itself does not explain the mystery of the lunar rays. Accepting the theory that the moon's crust yielded in some such way, we have still to explain how the rifts which were thus formed came to be covered over with matter lying nearly at the same level as the surrounding surface. It appears to me that the only available way of explaining this is somewhat as follows. First, from the way in which the streaks are covered like the surrounding region with craters, we may conclude that the streaks are older than any except the largest craters; from the great extension of many of them, we may safely infer that the lunar crust possessed a large measure of plasticity when they were formed (for otherwise it would have yielded over a smaller area). It was, therefore, probably still hot during the era (which may have lasted millions of years) to which the formation of the rifts belonged: accordingly the lava which flowed out through the rifts remained liquid for a considerable time, and was thus able to spread widely on either side of the rift, forming a broad band of lava-covered surface, instead of a steep and narrow dyke. This seems not only to account for the most striking peculiarity