PEPPER TREE AS A CAUSE OF UAY FEVER. 155 



III. Determination of Pollens in the Air of Bloemfontein 

 DURING the Epidemic Hay yEVER Season. 



[a) Method and Exper'unents in October and November, 1010. 



To determine the varieties of pollen in the air of the town 

 glass plates smeared with vaseline were suspended in the open 

 air and afterwards examined under the microscope. Consider- 

 able difficulty was encountered in finding a suitable adhesive, 

 glycerine, linseed oil (boiled and unboiled), gelatine, agar, and 

 several others being tried, but the only substance that remained 

 on the plate and retained its stickiness was vaseline. Ordinary 

 photographic quarter-plates (3:^ by 4J inches) were used : these 

 were smeared on one side and hung vertically in a position 

 exposed to the wind at a height usually of about five feet. The 

 length of exposure allowed varied from 14 to 84 hours, depend- 

 ing on the exigencies of College work, and is given in the tables, 

 where the date stated is that on which the period of exposure 

 ended. The plates were collected about 10 a.m. as a rule. 



The comments regarding the weather should be consulted 

 in conjunction with the official meteorological observations, given 

 in separate tables, for the same date. The comments direct 

 attention only to features which are regarded as of importance 

 in connection with the occurrence of pollen in the air, and are 

 not given in the official weather records. The weather con- 

 ditions reported are to be understood as having usually obtained 

 during only a part of the period of exposure. Unless stated to 

 the contrary the wind was westerly, and of the type already 

 described. 



In the microscopic examination of the large plates for pol- 

 len, three areas chosen at random were covered and part of 

 each was searched systematically with the aid of a mechanical 

 stage. Very great difficulty was experienced in this work owing 

 to the large amount of dust which had adhered to the vaseline. 

 After a severe dust storm the particles of dust were so thick 

 as to make it probable that some of the pollen would escape 

 detection, but even when the dust was only one layer in thick- 

 ness its presence greatly complicated the examination of the 

 plates and prolonged the time required. Eventually means were 

 devised for separating the dust from the pollens. This was 

 done by scraping the vaseline from the surface of the plate, 

 dissolving it in petrol and centrifuging, whereby the heavy 

 mineral matter was thrown to the bottom. Some of the clear petrol 

 solution was then transferred to a slide and concentrated by 

 evaporation; and after this had been repeated several times, 

 portions having been taken from various levels in the liquid, the 

 deposit was examined. In this way the dust was got rid of and 

 the pollen concentrated. The method proved very useful where 

 there was a great deal of dust on the plates, but has the disad- 

 vantage of not giving strictly quantitative results, such as can 

 be obtained by direct examination. A comparison, however, of 

 the results of direct examination and of examination after cen- 

 trifuging, made in a number of plates, showed that there was 

 a rough parallelism both as regards varieties of pollen and their 

 relative amounts. 



