I ©4 



Report S.A.A. Advancement of vScience. 



tive humidity is lowest. This will be borne out by a referenece tO' 

 Table III., which shows hourly values of "a" for loth February, 

 1903. By far the most interesting daily variation occurs just before 

 sunset. 



The relative humidity has been known to double itself in a 

 litJe over an hour, and in the same time the leak has fallen off from 

 quite a high value down to zero. The negative ions are most sensi- 

 tive. After sunset the air rapidly regains its conductivity, and values 

 of "«" have been got greatly in excess of the usual values. This 

 effect mav take place quite apart from any approach to saturation. 



Contrary to expectations, great disturbances, such as violent 

 hail; rain and thunderstorms, which are fairly frequent in Bloem- 

 fontein, have not, so far, been found to have any great effect on 

 the conductivity of the air. The few potential readings taken in 

 Bloemfontein are of quite the normal kind, positive and steady in 

 fair weather, high " wobblv " readings, usually negative, before and 

 during broken w'eather. 



In dust-storms, the potentials are negative, and so high that 

 the instrument is unable to measure them. If a storm strikes 

 Bloemfontein, the potential is negative and the barometer rises ; if 

 the storm passes on one side, the potential is positive and the baro- 

 meter falls. It seems also that a low potential accompanies good 

 conductivity and a high potential poor conductivity. 



Mr. Logeman has designed and is making an instrument to obtain 

 a continuous record of the conductivity of the air, and with it we 

 hope to push our inquiries further, and so verify or amend several 

 conjectural relations between "a" and the other meteorological 

 phenomena. 



TABLE I.— MEAN MONTHLY VALUES OF 

 CAPE TOWN. 



FOR 



