Cooling exhibited by Glass and by Steel. 447 



Discussion. — The salient features of these data are obvious 

 from an inspection of Table III. The density-effect of quench- 

 ing is decidedly negative ; the increase of specific volume is ex- 

 ceptionally large. Moreover, the nine Prince Rupert drops 

 examined exhibit nearly the same initial density, and nearly 

 the same final density. The approximate equality of the val- 

 ues A a for the soft state is easily explained. They indicate 

 that in proportion as we bring the strain to vanish, we reach 

 the normal density of the glass. Not so easy is it to account 

 for the observation, that the density of the Prince Rupert 

 drops shows almost the same degree of equality in the hard 

 (quenched) state. Indeed, if we take the fact into considera- 

 tion that the drops invariably contain bubbles distributed 

 irregularly and without apparent relation as regards size and 

 number, the difficulties of explanation are increased. In other 

 words the remarkable result that the volume-increase due to 

 quenching is quite as much the same for all drops as is at all 

 possible for the case of so complicated an operation, is an ex- 

 ceedingly striking and important result and leads to this infer- 

 ence : Inasmuch as bubbles are present in like total volume in 

 each of the drops, their presence can not be a circumstance of 

 mere accident; they must be regarded as a normal effector, 

 better, a necessary result of the operation of quenching applied 

 to glass ; they must in some very intimate way be connected 

 with the strain which the glass globules have experienced in 

 virtue of sudden cooling. 



Retaining this very plausible surmise in mind, we proceed 

 to a more minute inspection of the effect of annealing the hard 

 Prince Rupert drops. We obtain results very similar to those 

 investigated above for steel. The density effect of annealing is 

 decidedly positive, and is greater as temperature and time of 

 exposure are increased. Again, we readily divide the physical 

 effect of annealing into two parts or phases. The first of these 

 corresponds to the annealing temperatures 0° to 500°, the other 

 to higher temperatures. The range of temperatures corre- 

 sponding to the first phase is therefore larger for glass than for 

 steel ; and if we compare Tables I and II it appears obviously, 

 that for like density effects the annealing temperatures must be 

 chosen higher in the former case (glass) than in the latter. 

 But the change of density encountered in both instances is 

 small ; and yet in spite of the small density-effect during the 

 inferior stage of annealing, by far the greater intensity of strain 

 vanishes. The drops after annealing at 200° are still explosive ; 

 if they be broken after having been annealed in boiling sul- 

 phur (450)° they are found to have lost all traces of the explo- 

 sive properties which the originally quenched drops possessed. 



