depends for its correctness upon observation. In many cases this correctness depends on the perfect accuracy of instruments, and upon consummate skill in using them. A minute error in the original observation may produce comprehensive error in the conclusion. It is still fresh in the public memory that new and more accurate observation corrected by millions of miles a calculation comparatively so simple as the distance between the earth and the sun. The problem by the solution of which Mr. Hopkins determined that the minimum thickness of the crust is from 800 to 1,000 miles depends for its reliability on certain obscure phenomena connected with precession and nutation. Sir Charles Lyell admits that the problem is a 'delicate' one. Mr. Charles MacLaren remarked, and Miller quotes the remark with approval, that Mr. Hopkins's inference 'is somewhat like an estimate, of the distance of the stars deduced from a difference of one or two seconds in their apparent position, a difference scarcely distinguishable from errors of observation.' Add to this that opinions might be quoted from mathematicians of name as decidedly in favour of the theory that the geological changes which have taken place in the earth's crust are due to central heat, as the deduction of Mr. Hopkins is opposed to it. In the ninth edition of his 'Principles,' _i.e._, in the edition immediately preceding that now current, Sir Charles informs us that 'Baron Fourier, after making a curious series of experiments on the cooling of incandescent bodies, considers it to be proved mathematically, that the actual distribution of heat in the earth's envelope is precisely that which would have taken place if the globe had been formed in a medium of a very high temperature, and had afterwards been constantly cooled.' Sir Charles replied to this in the same edition that, if the earth were a fluid mass, a circulation would exist between centre and circumference, and solidification of the latter could not commence until the whole had been reduced to about the temperature of incipient fusion. We fail to see that this is an answer to Baron Fourier. What necessity is there for supposing that the solidification of the crust commenced before the matter of the globe had been reduced throughout to about the temperature of incipient fusion? The water in a pond must be reduced to about the temperature of incipient freezing before ice can form on the surface, but this does not prevent the formation of a sheet of ice on the top.