898 Proceedings of Royal Society of Edinburgh. [sess. 
the end the finer ones are also removed ; but if we test it just 
before the last of them fall, we shall find that the slightest ex- 
pansion is sufficient to make them centres of condensation and that 
higher expansions and supersaturations produce no further cloudy 
condensation — which we would be entitled to expect under the 
higher strain if extremely minute particles were present. We 
may therefore conclude that the finest subdivided solids are still 
acted on by gravitation, and seem to have a definite lower limit 
of subdivision; unless it can be shown that extremely minute 
particles do not form nuclei of condensation, a conclusion for 
which there seems to be no warrant. 
Of the different methods above described for distinguishing 
between solid and gaseous particles, the cloudy condensation test 
seems to be the best; accordingly, the odour of musk was first 
tested by it. The apparatus above described, and shown in the 
figure, was used for the purpose, a little musk being introduced 
into the flask F. It is evident that if the musk gave off solid 
particles, it would now be impossible to get the air in the flask in 
such a condition that no cloudy condensation would take place 
when the air was expanded, because the musk would keep up a 
continued supply of nuclei of condensation. However, on putting 
the musk to the test it was found that the condensation, due to the 
dust-particles, which were admitted along with musk, decreased, 
and finally entirely ceased, just as when no musk was present; 
thus showing that the musk did not give off solid particles. 
It was thought desirable to vary this experiment, as it was 
possible that the vapour-laden air in the test-flask might interfere 
with the result. The musk was therefore removed from the 
test-flask F, and put into a separate bottle, B, by itself. This 
bottle was provided with an air-tight stopper through which 
passed inlet and outlet tubes, and introduced into the apparatus 
between the filter and the test-flask in the manner shown in the 
figure. By this arrangement the musk evaporated in filtered air, of 
the ordinary dryness, before entering the test-flask. Tests made in 
this way also gave no nuclei from the musk, even when the apparatus 
had been left some time for a quantity of musk to evaporate. 
From these experiments we may conclude that musk does not 
give off solid particles, but evaporates as a gas or vapour like 
