The occurrence of these telling examples of the deceptive appearance of re-set gravels would seem to justify and emphasize the conviction created by a critical examination of the two leading so-called paleolithic sites at Trenton, that Dr. Abbott, notwithstanding his asseverations to the contrary, has been deceived. Very strong support, it seems to me, is given to this conclusion by the recently published opinion of the late Dr. H. Carvill Lewis, a glacialist familiar with the Trenton region, and with the work of Dr. Abbott at the period of his paleolithic castle building. Dr. Lewis is reported to have maintained before an open meeting of the Academy of Science in Philadelphia "that what Dr. Abbott believed to be undisturbed layers (of gravel) were those of an ancient talus."[3] This remark may refer to both the main sites--the one at the railroad station and the other at the river front--or possibly only to the former. I have also heard it stated that that eminent scholar, Dr. Leidy, who must have had ample opportunities of forming correct opinions upon the subject, held pretty much the same views of Dr. Abbott's finds. [3] Brinton, D. G., Science, Oct. 28, p. 249. [Illustration: =Fig. 4.= A freshly formed gravel bluff.] [Illustration: =Fig. 5.= Early stage of talus formation.] [Illustration: =Fig. 6.= An ancient talus.] To make the above criticism entirely clear, a few words of explanation of talus phenomena may be added. As a river cuts its channel deeper and deeper into deposits of gravel a section is gradually exposed, but the gravels break down readily under atmospheric influences and the exposed face does not retain a high angle. The upper part crumbles and descends toward the base, there to rest against the slope or to be carried away by the stream. A supposititious case will be convenient for illustration. A gravel terrace twenty feet in height is encroached upon by the river at high water and undermined, and the face breaks down vertically, leaving an exposure as illustrated in Fig. 4. In a very short time the upper portions become loosened and fall below, giving a steep slope as seen in Fig. 5. The process goes on with gradually decreasing rapidity, and if the river does not again encroach seriously, a practically stable slope is reached, as shown in Fig. 6. Such a talus may be hundreds or even thousands of years old, but there is rarely any means of determining its exact age. If the gravels are homogeneous in character, the talus will simulate their normal condition so completely that the distinction cannot be made out in ordinary gullies or by unsystematic digging. If the gravels contain varied strata the talus will be composite, and will be more readily distinguished from at least portions of the material in place.