ABOUT TERRESTRIAL TEMPERATURE. 89 
and water sphere up to latitude 40°. The temperature of the equator is obtained by 
completing the part of the two curves near the vertex. We thus obtain 
Latitude. All Water. All Land. Diff. 
0° a Ses) 115°:3 43°°5 
These numbers, it will be observed, approximate closely to those obtained by a 
very different process for the temperature of the equator in par. 33, especially as 
regards a sphere of water. 
41. The value of the constant C of par. 31, which defines the maximum effect 
of the presence of land on the temperature of any part of the globe, is by this 
latter determination somewhat larger than by our formula of (33). Its value just 
found is 43°°5, while by the approximate formula it is 38°:1—a difference which, in 
a research of this kind, must be regarded as not very material. The greater value 
seems to be most to be depended on. 
42. Some indirect considerations still farther confirm the result that this quan- 
tity C, the maximum effect of land on the temperature of the equator, can neither 
ereatly exceed nor greatly fall short of 40°. Thus, assuming the formula of 
(33), if we project M. Dovn’s temperatures in a curve in terms of the latitude 
(as in Plate II.), and set off from them the effect due to the presence of land, as 
indicated by the term of the formula C . L’ cos 2 A, we leave a curve of terrestrial 
temperature on a water sphere. Now, if we give to C a value much gveater than . 
40°, it will be found that this residual curve becomes sinuous and irregular, 
having three points of contrary flexure instead of one or none at all. This 
complication is little probable under circumstances of remarkable physical sim- 
plicity. Again, if the influence of land were much /ess than that indicated by the 
value of C=40", this would be inconsistent with the existing ‘“'Thermic Anomaly” 
in the higher latitudes (as, for instance, in Siberia), where we can hardly suppose a 
greater cold to reign under existing circumstances than would be the case were 
the earth entirely composed of land. 
43. But one of the most striking confirmations which the argument of (38) affords 
of the accuracy of our first assumptions in § 3, is with respect to the latitude 
at which the “ Thermic Anomaly” vanishes altogether, or the distribution of land 
and water ceases to affect the mean temperature of the parallel. We had already 
(30) deduced from the inspection of the curves on M. Dove’s maps of the 
northern hemisphere, and from them alone, that this critical latitude is between 
40° and 45°. The intersection of the curves of land and water temperature in 
the figure on the last page, deduced almost solely from intertropical tempera- 
tures in both hemispheres, point closely to the parallel of 425°, as that where 
the two curves intersect, or where the third term of our formula (par. 33) ought 
to change its sign.* Hence the predominance of water in the southern over the 
* Jt will be recollected that the change of sign was made to coincide with 45°, merely for the 
purpose of simplifying the formula. (See par. 31.) 
