The subject of the average composition of collections of rocks has been dealt with in a paper published ten years ago by Mr. F. W. Clarke on “The Relative Abundance of the Chemical Elements.” The object of that paper was to arrive at an estimate of the mean chemical composition of the earth's crust, but incidentally the author shows a close correspondence between the mean compositions of groups of miscellaneous rocks from different regions. It appears from his figures that a moderate number of analyses—less than a hundred—if taken without any selection, is enough to give value to their mean. It may therefore be of interest to apply the method to our own country.