PEPPER TREE AS A CAUSE OF HAY FEVER. 151 



taking the nearest whole figures. The method is, of course, 

 purely arbitrary and is open to criticism. It is really only a 

 first attempt to eliminate the relatively valueless part of the 

 rain. Small showers within a few days after a growing rain 

 no doubt do very appreciable good, but it would be difficult 

 to agree on the limits of amount and period ; on the other hand, 

 it is very doubtful whether a really heavy thunderstorm, of say 

 four inches within perhaps as many hours, does as much good 

 as a fall of two inches of average steadiness. ]\Iuch, too, de- 

 pends on the nature of the soil and on the amount of vegetable 

 covering present. As the daily rainfalls are given, the estimate 

 can be readjusted at discretion. 



Evidence is adduced that in Bloemfontein these light 

 showers play a role in the hay fever epidemics. 



The evaporation is very high, as is indicated by the 

 following figures giving the annual evaporation from free water 

 surfaces : 



Bloemfontein (^Mazelspoort) 82"5 inches. 



Johannesburg (Cox, 1922) 74'67 inches. 



The figure for Johannesburg is from ^Meteorological Office 

 data. That for Bloemfontein is calculated from records taken 

 by the City Engineer's Department near the Municipal Dam 

 isome 16 miles east of the town, and is the average of the three 

 years 1919-1921. 



The high evaporation figure is to be attributed to the dry 

 rarefied air, high temperature and, especially no doubt, to the 

 prevailing strong, dry, westerly wind. 



The result is that, although Bloemfontein has a moderate 

 rainfall, the outstanding feature of the climate is its dryness. 

 This is indicated in the low relative humidity of the air, shown 

 graphically in the chart on p. 190. Additional evidence is to 

 be found in the general jubilation occasioned by a good rain, 

 and the columns of detail to which it gives rise in the daily 

 press; as also in the local practice of describing as "Promising" 

 weather which is cloudy and overcast. 



It may not be genei-ally realised that the average rainfall 

 of Bloemfontein (21''73 inches) is almost identical with that of 

 some parts of the county of Norfolk, England: e.g., Cromer, 

 latitude 52° 56' N. has an average of 21"77 inches. Yet how 

 different is the climate ! But the latitude of Cromer is higher, 

 there is no doubt more cloud, and it is on the sea and, virtually, 

 at sea-level (196 feet). The temperature will, therefore, be less, 

 and the air (and winds) more moist. Hence the evaporation 

 will be less. The two centres nearest Cromer, for which evap- 

 oration figures are available, are Worstead and Ormesby ; and 

 at these stations the average evaporation for 1919 and 1920 was 

 respectively 16'3 and 20"8 inches. 



