SOME NUCLEI OF CLOUDY CONDENSATION. 19 



made its appearance in the flame ; on now testing the products they were found to have 

 greatly increased in impurity. This at once suggested that some piece of solid matter 

 had got into the flame. I now gently tapped the india-rubber tube by which the filtered 

 air was brought to the combustion chamber, when the flame at once responded, showing 

 a few bright sparks in succession, and the test-flask showed the air to be very full of 

 nuclei. The india-rubber tube was therefore removed, and the filter connected direct to 

 the combustion chamber. On now testing the apparatus, I gently tapped the filter 

 itself, with the result that matters were as bad as before — dense clouding made its 

 appearance in the test-flask. The arrangement of the apparatus was now slightly 

 altered. The combustion chamber, filter, and gas-generating apparatus were put on one 

 table, while the test-flask, with its pump, were put on another, to prevent the vibrations 

 produced by working the pump from shaking the other parts of the apparatus. 

 Further, the cotton in the filter was damped by passing damp air through it the reverse 

 way, and when in use a wet cloth was hung over the entrance end of the filter. This 

 was done to prevent dust rising from the dry wool or metal case. 



It may be asked why all these precautions are necessary, when a cotton-wool filter 

 is found to act quite satisfactorily without them when used in the ordinary dust experi- 

 ments. The reason is very simple. If a little bit of any kind of matter happens to be 

 carried with the air out of the filter in making the ordinary experiments, this means 

 ■only one centre of condensation, and that drop may not meet the eye of the observer ; 

 but when the same bit of matter, particularly if it be organic, passes through the 

 hydrogen flame, at once there are created thousands of centres of condensation, and what 

 in the one case may escape detection gives rise in the other to a very manifest effect. 



After many failures, and many alterations and improvements in the apparatus, the 

 result is that, when everything is working right, the test-flask shows the products of 

 combustion of pure hydrogen in dust-free air to be free from nuclei of all kinds. The 

 <iir remains clear on expansion, and only a few drops are seen falling rapidly ; the air is 

 in fact far freer from condensation than the purest air met with in the country. The 

 air was never absolutely free from a single nucleus on expansion, but frequently only a 

 very few drops appeared, and all in view could be easily counted. 



From this experiment it may be concluded, 1st, that pure hydrogen burning in 

 dustless air gives rise to no solid or liquid form of dust ; and 2nd, that any activity 

 the "ions" may have is extremely short-lived, as they are inactive by the time they 

 are cooled. 



Whilst the apparatus, as described, shows that pure hydrogen when burning in 

 dustless air gives rise to no permanent nuclei of condensation, we can, by making a 

 slight alteration in the conditions, show the great importance of dust in the cloudy 

 condensation produced by the products of combustion. If we remove the filter, C, and 

 allow the air of the room to go unfiltered to the hydrogen flame, we get very different 

 results, according to the condition of the air. If the day be calm, and the air of the 

 room be but little disturbed, we get a great increase in the density of the condensation 



