The relationship of these tracts to terminal moraines is very clear and specific. They constitute marginal phenomena of the ancient ice sheet. Their distribution completely excludes their reference to floating ice, for they not only undulate over the surface utterly negligent of any horizontal distribution, but they are disposed in loops in crossing the basins of the region, and the convexities of these loops are turned down stream. These basins for the most part open out in southerly or westerly directions which makes it improbable that ice-bearing bodies of water occupied them. But if this were not fatal, certainly the fact that the convexities of the boulder belts are turned down stream and cross the centers of the basins is precisely contrary to the distribution they must have assumed if they were due to floating ice in bodies of water occupying the basins. I hold it, therefore, to be beyond rational question that these tracts were deposited as we find them by the margins of the glacial lobes that invaded the region. If these boulder belts were of the same nature as the average boulders of the till-sheets beneath them, then the simple fact of unusual aggregation might be plausibly referred to the accidents of gathering and deposition. But they are very clearly distinguished from the average boulders of the till by several characteristics. 1. They are superficial. Sometimes they rest completely on the surface, sometimes they are very slightly imbedded, sometimes half buried, sometimes they protrude but a slight portion, and sometimes they are entirely concealed, but lie immediately at the surface. In all cases the aggregation is distinctly superficial. Where they are buried, the burying material is usually of different texture and composition from the subjacent till, and appears to be distinct in origin from it. The superficiality of the tract is very obvious almost everywhere, and is especially so in regions where the subjacent till is of the pebble-clay rather than boulder-clay order, for the comparative absence of boulders below emphasizes the contrast. Throughout most of the region the subjacent till is not of a very bouldery type, so that the distinction is generally a marked one.