Cooling exhibited by Glaus and by Steel. 



449 



Table IV.— Mean volume relations of the Prince Rupert drops and of 

 the hypothetical gas-bubble. 







V 



at 



r 



at 





V A 



V s 



1500° 



500° 



1500° 



500° 





CC. 



cc. 



cc. 



cc. 



c. 



c. 





0-5454 



0-5334 



0-0120 



00052 



0-142 



0-107 



-<2-3 



0-5427 



0-5334 



0-0093 



0-0041 



0-130 



0-099 



-<l-3 

 r 



Mean to = 1-3288- g 

 "Hard:" MeanA /i = 24363 ± 0-0008 

 "Soft:" Mean A s '= 2-4911 ± 0-0016 



Inasmuch as the values v 500 and r 500 are inferior limits, it is 

 clear that if the bubbles were gas or vapor, they would be of 

 the same order of visibility before and after annealing. In 

 other words, if the bubbles were gas, casual inspection would 

 not detect a difference of their size in the hard-quenched and 

 in the "soft" drops. In the experiments, however, the bub- 

 bles all but vanish. Hence they can not be gas. 



To obtain additional assurance on this point we ground flat 

 faces on the annealed drops, ISTos. 1, 5 and 8, and then measured 

 the size of the inclusions. We found these values : 



No. 1 : v s = 0-000037 c. c. 



No. 5 : v s = 0-000062 c. c. 



No. 8 : v s = 0.000040 c. c. 



Mean : v, = 0-000046 c. c. 



From actual measurement therefore we find, 



— >100 



(2) 



a result wholly incompatible with — < 24 as derived above 



(Table IV). It follows conclusively that the bubbles are not 

 accidental gas inclusions ; that only an insignificant part of the 

 variations of volume produced by quenching can be referred to 

 thermal expansion of gas. 



If the gas were aqueous vapor, and if dissociation were com- 

 plete, then the upper limit of (1) would have to be increased 

 in the ratio of 3 : 2. This does not remove the discrepancy 

 between (1) and (2). Moreover, since dissociation of water is 

 only incipient at 1500° and is not even complete at the temper- 

 ature of melting platinum, the volume variation of the bubbles 

 can not be due to the presence of water in the bubbles. 



If on the other hand we endeavor to explain the bubbles as 

 vacuities* left by the contracting glass, we arrive at accordant 

 and plausible results. Let t be the temperature to which the 

 * Very analogous to Torricellian vacua. 



