ON THE SEASONAL VARIATIONS OF TEMPERATURE. 



481 



The montlis in which the mean monthly temperature of the air and 

 water most nearly approxi- 

 mated were : December, 

 when the difference was 

 only 2° (i.e., of air above 

 water) ; November 2°-5 ; 

 Jnne 2°-8 ; and July 3°. 

 Those in which the differ- 

 ence of temperature varied 

 most were: March 7°"4; 

 September i'>°-6 ; and April 

 5°-5. Mean difference for 

 whole year 4°'3. Thus it 

 will be seen thatthemonths 

 in which the temperatures 

 of the air and water most 

 nearly appi-oximated were 

 those in which the day 

 was at the shortest and the 

 longest. In other words, 

 tliei'e were two maxima 

 and two minima of differ- 

 ence between the tempera- 

 tures of the air and water, 

 the former occurring in 

 the months of March and 

 September, the equinoctial 

 months ; and the latter in 

 December and June, the 

 months of the winter and 

 summer solstice. The 

 former fact is easily ex- 

 plained, but it is rather 

 !i curious circumstance 

 that the same thing shoiild 

 hold good of the month in 

 which tlie sun is longest 

 above the horizon, and 

 most nearly vertical. The 

 explanation, I have no 

 doubt, is that in the latter 

 part of June and the first 

 part of July, when there 

 was a period of drougrlit 

 and warm weather, which 

 lasted more than three 

 weeks, the river fell to its 

 lowest level, and the cur- 

 rent was very sluggish. 

 Hence the water became 

 more heated than in ordi- 

 nary circumstances, and its 

 temperature more nearly approached that of the air. 



1891, J J 



