On level, undulating and moderately broken ground, the tops of the trees, though they may be one hundred and fifty feet from the ground, are level as the top of a hedge. Even an isolated hill if it be rounded in shape presents hardly better facilities, the trees at the base and on the sides, in their effort to reach the sunlight grow taller than those on the summit, and there is no one tree that commands all the others. If however an isolated hill of several hundred feet in height be found, its steep sides culminating in a sharp peak, one day's work by three or four good axmen, in cutting neighboring trees, will prepare the way for a study of the general relief and topography of the adjacent country. If after these preliminaries have been completed the engineer imagines that he has only to climb the tree and sketch what he sees, to obtain reliable knowledge of the country, he is doomed to serious surprises in the future. If he makes the ascent during the middle of the day, he will, after he has cooled off and rested from his exhausting efforts, see spread out before him a shimmering landscape in which the uniform green carpet and the vertical sun combined, have obliterated all outlines except the more prominent irregularities of the terrene, and have blended different mountain ranges, one of which may be several miles beyond the other, into one, of which only the sky profile is distinct. Naturally under these conditions estimates of distance may be half or double the truth. There are two ways of extracting reliable information from these tree-top reconnaissances. If it be in the rainy season the observer must be prepared to make a day of it, and when he ascends the tree in the morning he takes with him a long light line with which to pull up his coffee and lunch. Then aided by the successive showers which sweep across the landscape, leaving fragments of mists in the ravines, and hanging grey screens between the different ranges and mountains, bringing out the relief first of this and then of that section, an accurate sketch may gradually be made. The time of passage of a shower from one peak to another, or to the observer, may also be utilized as a by no means to be despised check upon distance estimates. If it be the dry season, the observer may take his choice between remaining on his perch in the tree from before sunrise to after sunset, or making two ascents, one early in the morning and the other late in the afternoon. In this case the slowly dispersing clouds of morning, and the gradually gathering mists at sunset, together with the reversed lights and shadows at dawn and sunset, bring out very clearly the relief of the terrene, the overlapping of distant ranges, and the course of the larger streams.