154 Transactions of the South African Philosophical Society. 



oscillation of the barometer, although it may not be certain that they 

 will actually represent it. 



In this connection the annual harmonic elements of dew-point are 

 added, for purposes of comparison, in the last column of Table 29. 

 It has not been thought necessary to give the monthly values, nor, 

 for this preliminary statement, to go through the great labour of 

 separating clear days from cloudy. So much is therefore reserved 

 for a future communication. 



The dew-point amplitudes have no obvious relationship to those 

 of corresponding denomination in the curves of temperature and 

 pressure, nor, as we shall see presently, in those of the wind. The 

 amplitude of the diurnal wave of moisture is more than double that 

 of the semi-diurnal wave ; the latter is not greatly larger than the 

 eight-hour wave. The relatively great amplitude of the eight-hour 

 wave is noteworthy because of the influence of the clouds over this 

 period. 



The epoch of the diurnal wave of moisture agrees more closely 

 with that of the wind than with anything else : the maximum of the 

 wind wave being less than three minutes earlier (see Table 44). 

 Such an approach to synchronism may be important. There is no 

 such approach at all on the part of the diurnal wave of pressure, 

 which is a singular circumstance taken in conjunction with the fact 

 that the total barometric pressure is necessarily made up of the 

 sum of the individual pressures of dry air and water-vapour. The 

 sympathy between these two constituents of the atmosphere, how- 

 ever, is shown in the semi-diurnal term, the difference between the 

 epochs of barometer and dew-point being little in excess of three 

 minutes of time. Much the same may be said of the eight-hour 

 term, the difference here being about ten minutes. 



It would seem from this comparison that the diurnal wave of 

 barometric pressure does not arise from any access or defect of air 

 to the superincumbent mass, for if it did there would surely be a 

 proportionate access or defect of water-vapour, and a consequent 

 proportionate variation in the dew-point at the same time. This 

 wave is perhaps, then, the outcome of purely thermal actions : an 

 increase of tension to the westward of the place where temperature 

 is rising most rapidly, and a decrease to the eastward of the most 

 rapid fall. There must be also, of course, a concomitant increase of 

 tension of the contained vapour singly, although no increase of its 

 quantity, and therefore no rise in the dew-point. That is to say, if 

 we were to weigh the quantity of air in a given space, as we virtually 

 weigh (roughly) the quantity of vapour per cubic foot by determining 

 the dewpoint, we should not, on this view, find actually any more 



