The facts as to the general distribution of radioactive substances at the surface, and in rocks which have come from considerable depths in the crust, lead us to regard as certain the widespread existence of heat-producing radioactive elements in the exterior crust of the Earth. We find, indeed, in this fact an explanation--at least in part--of the outflow of heat continually taking place at the surface as revealed by the rising temperature inwards. And we conclude that there must be a thickness of crust amounting to some miles, containing the radioactive elements. Some of the most recent measurements of the quantities of radium and thorium in the rocks of igneous origin--_e.g._ granites, syenites, diorites, basalts, etc., show that the 125 radioactive heat continually given out by such rocks amounts to about one millionth part of 0.6 calories per second per cubic metre of average igneous rock. As we have to account for the escape of about 0.0014 calorie[1] per square metre of the Earth's surface per second (assuming the rise of temperature downwards, _i.e._ the "gradient" of temperature, to be one degree centigrade in 35 metres) the downward extension of such rocks might, _prima facie_, be as much as 19 kilometres. About this calculation we have to observe that we assume the average radioactivity of the materials with which we have dealt at the surface to extend uniformly all the way down, _i.e._ that our experiments reveal the average radioactivity of a radioactive crust. There is much to be said for this assumption. The rocks which enter into the measurements come from all depths of the crust. It is highly probable that the less silicious, _i.e._ the more basic, rocks, mainly come from considerable depths; the more acid or silica-rich rocks, from higher levels in the crust. The radioactivity determined as the mean of the values for these two classes of rock closely agrees with that found for intermediate rocks, or rocks containing an intermediate amount of silica. Clarke contends that this last class of material probably represents the average composition of the Earth's crust so far as it has been explored by us. [1] The calorie referred to is the quantity of heat required to heat one gram of water, _i.e._ one cubic centimetre of water--through one degree centigrade. 126 It is therefore highly probable that the value found for the mean radioactivity of acid and basic rocks, or that found for intermediate rocks, truly represents the radioactive state of the crust to a considerable depth. But it is easy to show that we cannot with confidence speak of the thickness of this crust as determinable by equating the heat outflow at the surface with the heat production of this average rock.