[Vor. 1 
76 ANNALS OF THE MISSOURI BOTANICAL GARDEN 
dicating that the sanitary quality of the air examined was 
“safe.” 
Experiment 12.—At the time the air sample was being taken, 
a slight drizzling rain was falling, accompanied by considerable 
wind and a temperature of 45°F. Prior to that time it had been 
raining continuously for about 24 hours. A total of 10,800 cc. 
of air was drawn through the sand filter at the rate of 900 cc. 
per minute, the apparatus meanwhile being held about 5 feet 
above the ground level. Тһе sand of the filter was introduced 
into 15 ce. of sterile distilled water, from which platings were 
made. ‘Table vim gives the details of the experiment, together 
with the results obtained. 
TABLE VIII 
No. of bac- 
Plate Quantity Total no. t а eed Coccus No. of sali- 
number plated of colonies melda colonies vary cocci 
1 1 cc. 9 2 7 7 
2 1 ес. 3 3 0 0 
3 5 cc. 0 0 0 0 
4 5 cc. 6 4 2 0 
Attention should be called to the fact that on plate 1, in which 
only 1 сс. of the solution was used, 9 colonies developed—7 
coccus and 2 bacillus—, while on plate 4,in which 5 cc. of the solu- 
tion were used, only 6 colonies appeared,—2 coccus and 4 bacillus. 
Furthermore, the 7 colonies in plate 1 proved to be of salivary 
cocci, whereas none of these organisms were present among the 
cocci of plate 5. These results unquestionably indicatelocal con- 
tamination. It is difficult to say just where the contamination 
took place. Obviously it did not occur during the collection 
of the sample or even during the mixing of the sand solution ; 
for had this been the case all of the plates should have shown 
salivary cocci, and the greater number should have occurred 
on those plates in which larger quantities of the solution were 
plated. In all probability plate 1 was locally contaminated. 
Experiment 13.—While the sample of air was being taken for 
this experiment, the temperature was 63°Е., a light breeze was 
blowing, and the sky was very cloudy although no rain had 
