152 



SCIENCE. 



[N. S. Vol. XV. No. 369. 



oxygen say two or three times as large as the 

 capacity of the tubing itself. But under cer- 

 tain circumstances this is found to be by no 

 means sufficient, as the following experiment 

 illustrates. 



Ten feet of thin-walled gray tubing having 

 an internal diameter of one fourth of an inch, 

 was used on a burner for half an hour, and 

 was from there transferred immediately to a 

 gasometer of oxygen; the gas was then al- 

 lowed to pass through the tubing and fill over 

 water a cylinder the capacity of which was 

 560 cc. 



As might have been expected the gas so ob- 

 tained in the cylinder exploded violently. The 

 volume of such a piece of tubing is about 95 

 CO., and hence the gas dravm ofi would con- 

 tain something less than one sixth of the 

 mixed hydrocarbons. 



A second cylinder was then dravTn off, and 

 when a taper was thrust into it an explosion 

 was produced which was as violent as the first. 



The third cylinder also exploded, though less 

 violently; the fourth flashed back slowly to 

 the bottom, and the fifth behaved like pure 

 oxygen. 



Thus in this case 2,240 cc. were used to wash 

 out a tube whose volume was less than 100 cc. 

 That is, the contents of the tubing were dis- 

 placed more than twenty times before the gas 

 was removed. 



The experiment obviously points to a solu- 

 bility of the gas in rubber, and this is not 

 surprising in view of the ready absorption by 

 rubber of the low-boiling paraifin hydrocar- 

 bons in the liquid state. 



That a certain amount of gasoline is ab- 

 sorbed in rubber may also be shown by pass- 

 ing a piece of rubber tubing up into a tube 

 filled with the gas and inverted over mercury. 

 It is of course to be remembered that the gas 

 supplied by such machines as that in use here 

 (Springfield Gas Machine) consists of a mix- 

 ture of the vapors of the hydrocarbons with a 

 very considerable proportion of air, so that 

 such absorption experiments as these can only 

 be relative. An evident absorption takes place 

 even with gasoline which does not show any 

 abnormal behavior when conducted through 

 the tubing; but when such behavior was mani- 



fested, the absorption was more than doubled. 



The danger arising from this source lasts 

 for only a short time after the gasoline tank 

 has been filled; and indeed the results re- 

 corded above were obtained only twice, al- 

 though four attempts were made immediately 

 after the filling of the tank; this irregularity 

 is probably due to the varying demands made 

 upon the gasoline machine at diiierent times. 



The rubber tubing employed in the experi- 

 ments was such as is furnished under the 

 catalogue number 8012 by Messrs. Eimer and 

 Amend. The gasoline was that supplied by 

 the Gilbert and Barker Manufacturing Com- 

 pany; hence it is of normal quality; the phe- 

 nomenon in question was observed both with 

 the 86° and 90° products (degrees Baume, 

 equivalent to the specific gravities 0.66 and 

 0.65). 



On the whole these observations point to 

 the conclusion that gasoline of the kind de- 

 scribed contains a small amount of more 

 volatile components, which are given off 

 mainly at first, and being perhaps m.ore 

 soluble in rubber than those which come over 

 later, cause the abnormal behavior above de- 

 scribed. 



It would be interesting to know whether 

 others who use gasoline have had occasion to 

 notice this peculiarity. 



A. P. Saunders. 

 Hamilton College, Clinton, N. Y. 



on the siphon. 



The writer wishes to call attention to an 

 error that has crept into the text-books on gen- 

 eral physics, written for high school and uni- 

 versity classes. Most of the books either state 

 explicitly that a siphon will not work if the 

 shorter of its two legs is longer than the 

 column of liquid that would be supported by 

 the air pressure, or else give explanations of 

 the siphon, from which this follows as a legiti- 

 mate conclusion. As a matter of fact, a 

 siphon can be made to work and draw the 

 liquid to a height considerably greater than 

 that representing atmospheric pressure. 



The writer usually illustrates this fact in 

 his lectures by means of the following simple 

 experiment: Let ABO in the figure be a glass 



