64 
PROF. T. CARNELLEF, MR. J. S. HALDANE, AND DR. A. M. ANDERSON 
The whole was covered externally with felt. A temperature of 100° was easily main* 
tained inside the cylinder. 
The tubes were at first sterilised by being kept for half an hour at 100° on 
two successive days. Latterly they were simply kept for an hour at 100°. Any 
failure in sterilisation would be easily detected by the appearance of colonies inside 
the jelly, and by their distribution round the tube. But this was never obtained in 
the case of tubes sterilised by the above method (cf. Koch, Gaffry, and Lceffleb, 
‘ Mitth. a. d. K. Gesundheitsamte,’ vol. 1 , p. 332 et seq.), nor did any colonies ever 
appear inside sterilised tubes kept unused for long periods. 
The rate of air-flow employed was, as a rule, about one litre in three minutes. The 
most suitable rate in any given case depends on the number of micro-organisms in the 
air and the rapidity with which they settle. When there are many in the air to be 
examined it is best, as a rule, to make the rate of flow more rapid. The error arising 
from the crowding together of colonies near the entrance is far greater than any likely 
to occur from the passage of micro-organisms into the cotton wool. Not only do the 
colonies coalesce if crowded together near the entrance, but substances appear to 
become diffused through the jelly, which entirely prevent the growth of some of them. 
Again, when the particles to which bacteria are attached are heavy, as occurs, for 
instance, in outside air on a dry windy day in towns, a rapid current is also necessary 
on account of the heavy particles coming down very rapidly on entering the tube. 
On the other hand, a slower current is desirable when there are no heavy particles in 
the air, as in the case of a room which has been standing quiet, or outside ah’ in damp 
weather. It is, of course, quite easy to tell from the distribution of the colonies in 
the tube whether a proper rate of flow has been chosen. 
The quantity of air taken was usually half a litre in bad atmospheres, one or two 
litres in better ventilated rooms, and from one to ten litres in outside ah’, according to 
circumstances. The latter quantity would frequently be insufficient in making 
systematic observations on outside air, as in many cases no colonies develop from ten 
litres. The quantities taken were, however, sufficient as regards the purposes we had 
in view in examining outside air. The tubes after the samples had been collected 
were exposed horizontally on racks in a room the temperature of which was kept 
tolerably constant during the day all winter by means of a warm-air ventilating 
apparatus. They were left until no more colonies made their appearance-—a period 
of, practically speaking, from three to four weeks. In a good many cases the colonies 
had begun to run together before this time had elapsed. In such cases we made an 
allowance for probable increase, based on the percentage increase observed in other 
tubes. 
Probably a better arrangement would have been to have kept them in a large incu¬ 
bator at about 20°, but, having begun by exposing them in the laboratory, we thought 
it best to continue doing so all through our experiments. 
