To get absolutely below the base of an image was not altogether easy. The first we attempted to dig out was one of the farther ones within the crater (no. 19, fig. 47); it was found that, while the back of the hole into which it had been dropped was excavated in the soft volcanic ash, the front and remaining sides were of hard rock. This rock was cut to the curvature of the figure at a distance of some 3 inches from it, and as the chisel marks were horizontal, from right to left, the workmen must have stood in the cup while preparing it: in clearing out the alluvium between the wall of the cup and the figure, six stone implements were found. The hands, which were about 1 foot below the level of the rim, were perfectly formed. The next statue chosen for excavation was also inside the crater (no. 107, fig. 47); it was most easily attacked from the side, and this time it was possible to get low enough to see that it stood on no foundation, and that the base instead of expanding, as with those which stood on the ahu, contracted in such a manner as to give a peg-shaped appearance; this confirmed the impression made by the previous excavation, that the image was intended to remain in its hole and was not, as some have stated, merely awaiting removal to an ahu (fig. 62). The story was shown not only in the sections of the excavation, but in the degrees of weathering on the figure itself: the lowest part of the image to above the elbow exhibited, by the sharpness of its outlines and frequently of the chisel cuts also, that it had never been exposed, the other portions being worn in relative degrees. Traces of the smoothness of the original surface can still be seen above-ground in the more protected portions of some of the statues, such as in the orbit and under the chin (see frontispiece); but a much clearer impression is of course gained of the finish and detail of the image when the unweathered surface is exposed. The polish is often very beautiful, and pieces of pumice, called “punga,” are found, with which the figures are said to have been rubbed down. The fingers taper, and the excessive length of the thumb-joint and nail are remarkable (fig. 72). The nipples are in some cases so pronounced that the natives often characterised them as feminine, but in no case which we came across did the statues represent other than the nude male figure[23]; the navel is indicated by a raised disc. On the statue with the contracting base, which is one of the best, the surface modelling of the elbow-joint is clearly shown. The orbital cavity in the figures on Raraku is rather differently modelled from those on the ahu; in the statues on the mountain the position of the eyeball is always indicated by a straight line below the brow, the orbit has no lower border (fig. 72). On the terraces the socket is constantly hollowed out as in the figure at the British Museum (fig. 31). EXCAVATED IMAGES. [Illustration: FIG. 62.