The Temperature of Water, etc. 259 



natural and admissible ; but if the density of water is 

 greater at the temperature of melting ice than at any 

 other more elevated temperature, as some philosophers 

 assert, it is evident that the vertical descending currents 

 of warm water which I have described cannot exist, and 

 my explanation must be rejected. 



This inquiry interested me all the more, because the 

 fact in question had served as the foundation of the 

 theory which I gave in my seventh Essay on the pe- 

 riodical winds of the polar regions, and as the basis of 

 my conjectures on the existence of currents of cold water 

 in the depths of the sea coming from the polar regions 

 to the equator, and on the cause of the great difference 

 which is found in the temperature of different countries 

 situated in the same latitude and at the same height 

 above the level of the sea. 



After meditating on the means which I should employ 

 to establish this important fact beyond doubt, I thought 

 of the experiment which I am about to describe, and 

 which is all the more interesting, since it not only 

 demonstrates the existence in a mass of water which is 

 warmed or cooled, of the currents assumed by my 

 theory, but proves at the same time that the temper- 

 ature at which the density of water is at a maximum is 

 actually some degrees above that of melting ice. 



Having provided a cylindrical vessel (A, Plate VI.), 

 open above, made of thin sheet brass, 5^ inches in 

 diameter and 4 inches deep, supported on three strong 

 legs \\ inches high, I placed in it a thin brass cup 

 (B) 2 inches in diameter at its bottom (which is a 

 little convex downwards), 2^ inches wide at its brim, 

 and I T \ inches deep; this cup stands on three spread- 

 ing legs made of strong brass wire, and of such form 



