Occasionally, more especially in the case of large bosses, like those in which granite so frequently appears, the eruptive mass may be observed to rise here and there in detached knobs through the surrounding rocks, or to enclose patches of these, in such a manner as to indicate that the large body of eruptive material terminates upward in a very irregular surface, of which only the more prominent parts project through the cake of overlying rocks. In true bosses, unlike sills or laccolites, we do not get to any bottom on which the eruptive material rests. Laccolites, indeed, may be regarded as intermediate between the typical sill and the typical boss. The difference between a laccolite and a boss lies in the fact that the body of the laccolite does not descend into an unknown depth in the crust, but lies upon a platform on which it has accumulated, the magma having ascended by one or more ducts, which generally bear but a small proportion in area to the mass of the laccolite. The boss, on the other hand, is not known to lie on any horizon, nor to proceed from smaller ducts underneath, but plunges as a great pillar or irregular mass, which may frequently be noticed to widen downwards into the crust. There can be no doubt, however, that many masses of eruptive rock, which, according to the definition here given, should be called bosses, would be found to be truly laccolites if their structure below ground could be ascertained. It is obvious that our failure to find any platform on which the body of a boss lies, may arise merely from denudation having been as yet insufficient to lay such a platform bare. It is hardly probable that a boss several miles in diameter should descend as a column of that magnitude to the magma-reservoir from which its material came. More probably it has been supplied through one or more smaller ducts. The large boss now visible at the surface may thus be really a laccolitic expansion on one or more horizons. M. Michel Lévy lays stress on the general widening of granitic bosses as they descend into the crust.[35] While his observations are supported by many illustrations from all parts of the globe, and are probably true of the deeper-seated masses of granite, it is no less true that numerous examples have been met with where a granite boss is sharply marked off from the rocks which it has invaded and on which it may be seen to lie. Apart from the cases where granite seems to form part of a vast internal, once molten mass, into which its encircling gneisses seem to graduate, there are others in which this rock, as now visible, has been injected into the crust as a boss or as a laccolite. Instances will be described in later chapters where such bosses have risen through Cambrian, Silurian, Devonian and Carboniferous formations. It may be said that between such granitic intrusions and volcanic operations no connection can be traced. But reasons will be brought forward in later chapters to regard some of