156 Transactions of the South African Philosophical Society. 



The general features of the curve of vapour-tension, as it is called, 

 in the above two cases, may be readily explained. On the open sea 

 in low latitudes, the lower air is always nearly saturated, and no 

 matter what quantity of moisture any harmonic wave of pressure 

 may be supposed to contain, it cannot add, as vajjour, anything to 

 that already existing, nor, for that matter, add any material percent- 

 age of moisture in any form. The humid conditions seem, indeed, to 

 be determined directly and entirely by the temperature. The elimi- 

 nation of the ocean maximun near land " points unmistakably to an 

 intermixture with the air forming the sea-breeze of descending thin 

 [and drier] air filaments or currents to take the place of the masses 

 of air removed by the currents that ascend from the heated surface 

 of the land."''' 



We should not expect, therefore, to find any traces of barometric 

 affinities in the harmonic components of moisture near the surface of 

 the open sea, although it is not impossible that they may exist at 

 higher altitudes over the sea. 



If we compare the harmonic constants of pressure at Kimberley 

 with those of places having different climates, we shall soon get 

 corroboration of the great influence exercised by the vapour of the 

 atmosphere upon the magnitude of the harmonic constants, more 

 particularly in the second term. Thus Lahore, in a partial degree 

 similarly situated to Kimberley, has amplitudes tt^ and u^ almost 

 identical with those of this station, its -z^i being nearly equal to u^ ; 

 Madras, a coast station, has w^ = '024, n2 — '043 ; Bombay, another 

 coast station, has iCi = "OlS, ic^ — '038 ; Agustia, on the summit of a 

 peak, at an altitude of 6,200 feet, has u^ = -005, 112 = '031 ; Fort 

 Eae, near the Arctic Circle, in the middle of the great Dominion, has 

 Ui = -007, -2^2 = *001. Thus on the whole, other things being equal, 

 the amplitude of the second component increases with the moisture 

 of the air, while the amplitude of the first varies with range of 

 temperature. But because the range of temperature becomes 

 generally greater as the amount of vapour in the air becomes less, 

 it follows that when a resultant curve is constructed from its first two 

 harmonic components, the daylight range from the first maximum 

 to the second minimum will be increased pretty much by a greater 

 range of temperature as it is decreased by a smaller quantity of 

 moisture. And therefore the graphical aspect of this daylight range 

 will be somewhat similar for either the centre of a flat continent or 

 for mid-ocean. On an ocean shore the circumstances are com- 

 plicated by the land- and sea-breeze. It would seem that the 

 tension along the coast at night is lowered partly by the rapidly 

 * Buchan, in the Challenger Beport, p. 12. 



