116 Transactions of the South African Philosoj^hical Society. 



order, the coast being fully twelve days earlier than Kimberley,. 

 although the amplitudes stand in no regular sequence. It is note- 

 worthy, however, that the Philippolis observations determine a break 

 in the uniform increase of a^ with distance from the coast, and also' 

 another in the retardation of phase-time of A^. But since the 

 observations cover a shorter period it is possible that the peculiarity 

 might eliminate itself in time. 



In the case of the maximum temperature the amplitude of the 

 annual term increases with distance from the sea, as is also the case 

 with the term of six-months period. The term of four-months period^ 

 however, increases from Durban to Umtata, falls almost to zero at 

 Queenstown, and then rises again. The term of three-months 

 period behaves irregularly ; but chiefly it would seem from the 

 smallness of the magnitudes it deals with : the influence of a tenth 

 of a degree in the values of p^ and q^ being very considerable in the 

 final evaluation of a^. But both in the term of four-months and in 

 that of three-months period, the greatness of the amplitudes at 

 Umtata, as compared with those of the other places, deserves 

 attention. This not improbably has a solar origin, for the 

 magnitudes dwindle to something like an average in the corre- 

 sponding values of minimum. The relative smallness of a^ at 

 Durban, the fact that a^ is almost as great, coupled with the anti- 

 synchronism of the phase-times of A2 and A3 during January,^ 

 combine to explain why the Durban curve of annual mean 

 maximum temperature is almost a simple sine curve. 



The phase-time of A^ for the curve of maximum temperatures comes 

 earlier with distance from the sea, Kimberley being not quite thirty- 

 five days earlier than Durban. Moreover, it is the distance from 

 the sea in particular, and not the altitude, which is responsible both 

 for the epoch and for the above-mentioned prograssive increase in 

 the value of a^ : for Queenstown, which is less than a third of the 

 distance of Kimberley from the sea, is only 500 feet lower, while 

 Aliwal North and Philippolis, at respective distances of about one- 

 half and two- thirds, are both considerably higher.* The phase-time 

 of A2 is much earlier on the coast than it is at the inland places. 

 Here Umtata forms a sort of connecting link, midway between the 

 coast and the interior. 



The phase-time of A2 for the minimum temperatures is pretty 

 nearly constant everywhere ; but, on the other hand, A^ shows 

 great variation over a small range of differences, falling two days 

 earlier at Umtata and Kimberley than it does at Queenstown, and 



* See the remarks on their altitudes in " Some Pressure and Temperature 

 Results, &c." Also the Reports of the Meteorological Commission. 



