PEPPER TREE AS A CAUSE OF HAY FEVER. 187 



Kimberley evidence is conclusive : pepper tree pollen obviously 

 dries and finds its way into the air there. But the evidence for 

 the other centres is not very satisfactory — negative evidence never 

 is. The period of trial was short (especially in Johannesburg) ; 

 and the weather, although, perhaps, fairly typical for these 

 centres at this season, was not such as to lead one to expect 

 much pepper tree pollen in the air. The question whether 

 pepper tree pollen becomes powdery in these towns could only 

 be settled by the repeated examination of flowers on the spot, 

 or by the exposure of pollen plates over a longer period. 



It is also very difficult to decide from the meteorological 

 observations whether pepper tree pollen would be likely to 

 become powdery in these other centres. It does so in Kimberley 

 and Bloemfontein, but from the following tables, and from 

 other data already given, it will be seen that, during the epi- 

 demic season, Kimberley and Bloemfontein are much hotter 

 and drier than are the other three towns. As already explained, 

 heat and dryness are the factors concerned in drying the pollen. 



There are no records available regarding the wind, other than 

 those taken at 8.30 a.m., which, for Bloemfontein at least, as 

 already explained, are very misleading, but at this season, Kim- 

 berley, which is only some 100 miles from and due west of 

 BloemJontein, is, I understand, subject to the same hot, dry, 

 westerly wind as is prevalent at Bloemfontein. A measure of the 

 relative drying effect of the climate for the epidemic months at 

 three of the centres is given in the statement which shows 

 the evaporation from free water surfaces. 



Total for the 

 two iTionths. 

 Nov. Dec. Inches. 



Kimberley 13-60 13-06 26-66 



Bloemfontein 9-5 11-8 21-5 



Johannesburg 7-13 6-79 14-92 



The figures for Kimberley and Johannesburg are supplied by the 

 Meteorological Office, which, however, has no data for Bloem- 

 fontein, Grahamstown and Maritzburg. Those for Bloemfontein 

 are calculated from records taken near the Municipal dam at 

 Mazelspoort. 



The greater heat and dryness of Bloemfontein and Kimberley 

 at this season are emphasised by the average figures for 

 December, 1920, given in the last five columns of Table XIX. 

 But it has to be remembered that hay fever occurs in Bloemfon- 

 tein and Kimberley even in November; and the November 

 averages for these towns are not very different from the Decem- 

 ber averages for the other three centres. Averages, however, 

 are perhaps misleading, as pepper tree pollen is not found in the 

 air of Bloemfontein throughout the whole of the hay fever season, 

 but only on hot dry days. Hence it is rather with the records of 

 individual days of such weather at Bloemfontein that the daily 

 figures for other centres should be compared. It is then obvious. 



