182 PEPPER TREE AS A CAUSE OF HAY FEVEIt. 



Graiiamstowji seems to rank next to Bloemfontein and 

 Kimberley in order of density. Tliere the total number of 

 pepper trees in the whole town is only 470 ! From Maritzbui-g it 



was reported that " there are very few of these trees so 



few in fact that it is felt that tabulated returns such as asked 

 for . . . would not be of much use. There is only one 

 stretch of pepper trees in the public thoroughfares, and the total 

 number of pepper trees in the area referred to is only 100." 

 Repeated requests to the Municipality of Johannesburg for 

 corresponding data have been without avail. From other sources 

 I am given to understand that there are considerable numbers in 

 certain suburbs, such as Doornfontein, but figures are necessary 

 to make a comparison possible. 



The information originally received regarding the number of 

 pepper trees in Maritzburg and Grahamstown (Potts, 1919) 

 proves, on further i.nquiry, to have been somewhat misleading. 

 Comparatively speaking, there are really very few pepper trees 

 in these towns, and the smallness of their numbers is, no doubt, 

 m itself a sufficient reason why the tree does not cause hay fever 

 in them to a serious extent. 



(b) Pollens in the Air of Other Toicns. 



For testing the pollen-content of the air of these other 

 centres, in which it was necessary to send plates through the 

 post, microslides smeared on one side were sent for direct obser- 

 vation, and small sheets of tin-plate of lanteni-slide size (Sc- 

 inches square) smeared on both sides for centrifuging. In this 

 way the risk of breakage involved in sending large plates of glass 

 through the post was avoided. Also, where centrifuging is to be 

 adopted, metal has the advantage over glass, that the plates can 

 be suspended safely from one comer and thus swing freely and 

 offer a sticky surface to the wind, irrespective of the direction 

 from which it blows. The results given by the two methods 

 were in agreement. The word plate is used as a general teiin to 

 cover both microslide and metal sheet. 



Kimberley. 



In all, 28 plates were exposed in various parts of Kimberley 

 during ten days in the first half of December, 1920, when, as 

 reported by four local doctors, a severe epidemic of ha}' fever was 

 raging. The period of exposure varied from 12 to 72 hours, and 

 the weather was described as being hot, dry and windy during 

 most of the period. The wind, it is understood, was of the same 

 dry, westerly type already described for Bloemfontein at this 

 season. Weather records for the month are given in Table XV. 

 The mean maximum temperature and mean difference between 

 the wet and dry bulb from December 1 to December 15 were 

 96'6or. and 16''0°F. respectively. The records, therefore, show 

 great heat and dryness. Examination of the plates showed that 

 pepper tree pollen was present in varying amounts on 22 of the 28 

 plates, the amount being roughly of the same order as had 



