350 PLEASANT WAYS IN SCIENCE. 



But the theory of positive and negative forms of oxygen, 

 though still held by a few physicists, has gradually given 

 way before the advance of new and sounder modes of in- 

 quiry. It has been proved, in the first place, that ozone is 

 denser than ordinary oxygen. The production of ozone is 

 always followed by a contraction of the gas's volume, the 

 contraction being greater or less according to the amount of 

 oxygen which has been ozonized. Regularly as the ob- 

 servers Messrs. Andrews and Tait converted a definite 

 proportion of oxygen into ozone, the corresponding con- 

 traction followed, and as regularly was the original volume 

 of the gas restored when, by the action of heat, the ozone 

 was reconverted into oxygen. 



And now a very singular experiment was made by the 

 observers, with results which proved utterly perplexing to 

 them. Mercury has the power of absorbing ozone ; and 

 the experimenters thought that if, after producing a definite 

 contraction by the formation of ozone, they could absorb 

 the ozone by means of mercury, the quantity of oxygen 

 which remained would serve to show them how much ozone 

 had been formed, and thence, of course, they could deter- 

 mine the density of ozone. 



Suppose, for instance, that we have one hundred cubic 

 inches of oxygen, and that by any process we reduce it to 

 a combination of oxygen and ozone occupying ninety-five 

 cubic inches. Now, if the mercury absorbed the ozone, 

 and we found, say, that there only remained eighty-five 

 cubic inches of oxygen, we could reason in this way : Ten 

 cubic inches were occupied by the ozone before the mercury 

 absorbed it ; but these correspond to fifteen cubic inches 

 of oxygen ; hence, ozone must be denser than oxygen in 

 the proportion of fifteen to ten, or three to two. And what- 

 ever result might have followed, a real absorption of the 

 ozone by the mercury would have satisfactorily solved the 

 problem. 



But the result actually obtained did not admit of inter- 

 pretation in this way. The apparent absorption of the 



