of Heat over Islands. 245 



tudc X,, be situated ou the same isothermal hue, we must have 

 H=F(0,)-f/(\J; 



whence 



F(c)-F(c,)=/(\.)-/(\). 



It follows from the foregoing conditions that this equation 

 cannot subsist unless we have the inequalities 



X < \j and Ci < c, 

 or 



X > Xj and c, > c. 



Hence that part of the isothermal which has the greatest latitude 

 must be nearer to the coast than any other part, and that which 

 has the smallest latitude must be the most remote from the coast. 

 Isothermals which had been previously near the coast would now 

 no longer be closed curves, at least within the island, and thus 

 several might terminate on the coast as irregular arcs, with their 

 convex sides turned towards the equator. If the dimensions of 

 the island in the direction of the meridian were very great, the 

 isothermals might all terminate on the coast. 



If predominating currents of wind should blow from any point 

 of the compass, it is likewise evident that they will further in- 

 fluence the position of the isothermals in a manner that can be 

 represented by shifting them away from, or by moving them 

 towards, the point of the wind, according as it happens to be a 

 warm or a cold current. 



As the surfaces of islands are usually not flat, but covered 

 with eminences and depressions, and as the temperatm-e of any 

 point depends on its elevation above the sea, as well as on the 

 other elements of its position, the transportation of the isother- 

 mals in an island would necessarily be accompanied by some 

 change in their shapes, and thus, after transposition, they would 

 not in general have the same detailed relations to the coast-line 

 as in their concentric condition. As every piece of land, whether 

 designated as a continent or as an island, is in reality surrounded 

 by water, these views are capable of very general application. 

 But, in the case of continents, as the oceans surrounding them 

 do not possess the same temperature at diff"erent parts of their 

 coasts, the function F(c) should receive difi'erent values for dif- 

 ferent places. In the case of islands of limited extent, we may, 

 however, compare these views with observations without much 

 difficulty. 



It has been long recognized that the warm current, known to 

 mariners as the Gulf-stream, bathes the shores of these islands as 

 well as the greater part of the western coast of Europe. Very 

 recently decisive evidence has been afl"ordcd of its calorific effects, 



