PEPPER TREE AS A CAUSE OF HAY FEVER. 175 



susceptible to the hay fever toxins of the pepper tree, as is also 

 indicated by their freedom from hay fever in Bloemfontein dur- 

 ing the epidemic season. The leaf-reaction may be due to the 

 resin present, which other experience proved to be a powerful 

 irritant. 



In the light of these tests it would seem probable that the 

 toxic principles, which induce hay fever in persons susceptible 

 to pepper-tree pollen, are present to a small extent in parts of 

 the flower oth.er than the pollen ; but that neither flowers free 

 from pollen, nor leaves, are of any practical importance in in- 

 tensifying the local epidemics. 



The results are in agreement with the experience of .epidemic 

 sufferers, who find that the commencement of their hay fever 

 synchronises with the coming into flower of the pepper tree on 

 a large scale. If the leaves played an important part in causing 

 hay fever, symptoms would be expected earlier. The results 

 have also an important bearing on prevention, as if th.e epidemic 

 trouble is virtually all due to the pollen of the pepper tree it 

 could be prevented by removing the male tree. 



XI. Pollens in the Air of Bloemfontein during the 

 Eainy Season, 1920. 



The epidemic hay fever season at Bloemfontein extends 

 from about the last week of October until into January (or even 

 later, depending mainly on the rains), and is usually worst dur- 

 ing November and D.ecember. Tables showing the results of 

 exposing pollen plates at Bloemfontein during the epidemic have 

 already been given. For coinparison, plates were exposed at 

 Bloemfontein in March, 1920. The results are given in Table 

 XIII, and the meteorological observations for the same period 

 in Table XIV. 



The University College Grounds are to the west of and be- 

 yond the town. The vegetation in them is typical grass veld, 

 and there were no pepper trees near the place of exposure of 

 the plates. Corresponding particulars regarding the other centres 

 have already been given in connection with the other tables. 



A total of 11 plates were exposed for varying periods between 

 March 19 and April 1. On seven of them grass pollen was found, 

 the amount present being in one instance equivalent to density 

 of 55 pollens per sq. centimetre per 24 hours. No pollen other 

 than grass pollen, not even that of th,e pepper tree, was found 

 on any of the plates. The pollen-content of the air is, therefore, 

 very different from that of November and December. But heavv 

 rains had fallen. The pepper tree had entered on a period of 

 intense vegetative activity, having produced an abundance of 

 new, green, leafy shoots, and had almost ceased to flower. More- 

 over, in the comparatively few flowers produced, th.e pollen re- 

 mained sticky and still lulled the anthers of quite old flowers. 

 The rains had also cavised the grass to grow vigorously, and manv 

 kinds covering extensive areas around the town were in flower. 

 The weather was also verv different. It was cooler, moister and 



