arrangement of pulleys would have to be extremely complex, so as to enable the elements to be determined which were dependent upon so many considerations. It is, however, quite plain that if we are ever to succeed in subjecting meteorological phenomena to numerical precision it must be in some such direction as I have indicated. To put the matter a little more plainly; we have reason to believe that a system of pulleys could be so arranged, and the relative movements of them could be so actuated, that a cord passing over those pulleys and adjusting a pencil could be made to show the height of the barometer for every day in the year at a given place. A similar machine might also be conceived which should show the temperature at any stated locality for every hour in the year. I do not for a moment assert that the information at present at our disposal would enable us to construct such machines. All I am now contending for is that mathematical theory seems to declare the possible realisation of such contrivances. The fact that an engine has already been constructed for the comparatively simple case of the tides leads us to hope that the time may arrive when meteorological engines shall have been designed by which meteorological prediction shall become as determined as the prediction of high water.
This discussion will at all events enable us to make some reply to the question which has been often asked, as to what was the cause of the great heat-wave. I do not indeed think that the question admits of any off-hand answer of the kind that is frequently expected by those who ask it—the only kind of answer that seems possible is of a somewhat indirect character. We may here again revert to our illustration of the tides. It sometimes hap-
