1821.] Phenomena of Heat f GascSj Gravitaiiofif&)C. 413" 



the variation of the temperature ; but they want that extent and 

 variety which are requisite to estabHsh correctly the true mean 

 temperature of the earth in difierent parallels. To settle this 

 point well, a series of observations * are wantnig, in the intersec- 

 tions of about every fifth or tenth degree of latitude and 

 longitude on both sides of the equator, and as near to the poleg 

 as possible. Local circumstances will no doubt always interfere 

 to prevent the perfect coincidence of any general law of temper- 

 ature with observation, which is the case with the one that I 

 have assigned. Indeed were it not so, and were the earth of a 

 uniform constitution, it would be no difficult matter to determine 

 its ellipticity from the temperature alone of different latitudes. 



Philosophers inform us that the southern parts of the earth 

 are colder than the northern. This might possibly be owing to 

 the different structure of the superftcial earth in the two hemi- 

 spheres, with respect to land and water ; but whatever be the 

 cause, an inequality in the mean temperatures of corresponding 

 latitudes cannot fail, if it be considerable, to affect the attraction 

 of the earth at those places, and, therefore, the length of the 

 pendulum, and consequently the figure of the earth, as deter- 

 mined by the pendulum. Whether this be really the cause of 

 the irregularity of figure, which has been said to have been 

 observed in the two hemispheres, I must leave until I have 

 obtained data sufficiently correct to enable me to undertake a 

 complete solution of the problem, respecting the attraction and 

 figure of the earth. Upon the whole, however, this theory of 

 gravitation seems to render it probable, not only that the earth 

 differs in figure much less from a solid of revolution than has 

 lately been imagined, but that its internal structure is also more 

 uniformly homogeneous, than some philosophers have supposed. 



Mathematicians will readily perceive, that a less temperature 

 in the southern than in the northern hemisphere, if considerable^ 

 ought to affect the lunar theory, and accelerate the motion of 

 the moon in n and bl , and retard it in f and Vf ; and that the 

 equation thus produced ought to be the greatest when S falls in 

 the vernal equinox, and the least when in the autumnal ; and 

 that other things being the same, it ought likewise to be greater 

 about the time of the aphelion earth, and less about the time of 

 the perihelion earth. The distance, however, of the moon from 

 the earth, the small obliquity of its orbit to the equator, and the 



* It might be advisable in future experiments to ascertain the temperature of the sea, 

 audits diurnal variation, at different depths; for I do not remember to have seen any 

 notice taken of these things in the experiments that I have hitherto read. One might 

 indeed infer from Dr. Davy's experiments, that if there be a variation at any consider- 

 able depth below the surface, itis probably but trifling. It would, however, at least be 

 satisfactory to decide this question by actual experiment ; and besides, a knowledge of 

 the temperature of the ocean at different depths, and in different places, would possibly 

 enable philosophers to determine the internal temperature of the earth, and its law of 

 variation towards the centre, as well as, perhaps, some other points of great interest ta 

 science. 



