142 THE STRUOTirRE OF THE NUCLEUS. 



such as arise from the lum-eleotrolyzing liquids, the initial plane may be definitely 

 lowered l»y successive exhaustion and its position in this way fairly well deter- 

 mined. A study was made at some length in the case of benzol, the limit or error 

 of which is tiaced in the following tables. In case of the vapors from the ionizing 

 li(|uids, like water or the alcohols, this method fails, because the api)aratus, after 

 the Hr.st e.xhaustion, is liable to be filled with nuclei. It then becomes necessary to 

 tpiite remove tliem by aspirating filtered air through the vessel before the next 

 e.\i)eriment can be made. So tedious an operation involves much loss of time, but 

 no way of avoiding it was discovered. In these cases, therefore, the position of 

 the plane of demarcation at the beginning of the e.xperiment can only be inferied. 



Indeed, this seems to be an insuperable difficulty incident to the present 

 experiments ; for, even in the former cases, the successive precipitations remove 

 nuclei which are thereafter lost for diffusion. Hence, a modification of method 

 was later introduced to add fresh nuclei instead of filtered air after each exhaustion. 

 Though in the nature of an interference with the experiment, the consequences, 

 either good or bad, remained of small moment. 



8. Apparatus. — The diffusion cylinder is shown in its completed form in 

 figure 1. The glass diffusion tower, AABB, was about 1 meter high and 13-21 

 cm. in diameter, slightly tapering toward the top. Both ends were ground flat, 

 the lower standing on a greased plate.of ground glass, the upper i-eceiving the flat 

 metallic lid, BB, securing the tubes leading to the exhaustion reservoir, C, to the 

 filters, F, etc. A pan, P, of coppei- at the bottom of the i-eceiver, held the liquid 

 from which issued the vapors to be tested. The lid cairies the vacuum gauge, G, 

 graduateil in cm. of mercury and reading to about 40 cm. A vertical centimeter 

 scale (not shown) is marked with India iidc on the outside of the wall of the 

 receiver, and by its aid the height of the fog bank could be read off at once, with 

 the usual precaution for a horizontal line of sight. The illuminating lamp, usually 

 a circular hole in an opaque screen in front of the mantel of a Welsbach burner, 

 was placed at a distance of about 3 meters away in the direction of vision and at 

 the j)robable height of the fog bank, estimated in advance of the observation. 



The main filter F', consisting of a long tube of conqiressed cotton, communicates 

 with the atmosphere through a CaClg desiccator tube, U', so that the air may be 

 both dry and filtered. The tube which conveys this air to the bottom of the 

 receiver, AABB, is provided with two stopcocks, a', b' , one of which, b' , is kept 

 l)ermanently but veiy slightly open, in order that the current of air through the 

 filter may always l)e very slow. On removing the filter (rubber joint) this tube 

 serves additionally as a means of introducing phosphorus or other nuclei into the 

 bottom of the receiver. The air in this case should also be dried by an ordinary 

 CaCl, tube, as shown in figure 2, the phosphorus grid, P, being inserted into the 

 iqiper pait of one shank of the tube. 



The second filter, F, and desiccator, IJ (also to be rei)laced, if desirable, by a 

 nucleatoi"), has a similar construction to the preceding, but the air in this case is 

 discharged into the top of the receiver. 



The exhaustion reservoir C communicates with the receiver AB, through a 



