﻿Cooling exhibited by Glass and by Steel. 183 



impaired by tbe removal of less than one-tenth millimeter of 

 shell ; that it vanishes almost wholly with the removal of one- 

 half millimeter of shell. If the radius of the drop be dimin- 

 ished about 0"03 cm the particles of the fractured globule fre- 

 quently cohere, and the original structure may then be inferred 

 from the general direction and distribution of the fissures. The 

 arrangement of the individual fragments' is quite characteristic ; 

 they are found to be flat irregular conoids, with their apices 

 toward the line of symmetry of the drop, their bases in its sur- 

 f ace> — an arrangement something like the e}^e of an insect, or 

 even more like the fruit cone of a spruce tree. In other words 

 the radial structure of the fissured P. E. drop is distinctly 

 marked. This proves that the original (unbroken) drop must 

 have possessed a box-within-box structure; that if the bubbles 

 were symmetrically disposed, particles similarly situated with 

 reference to the line of symmetry would be in like states of 

 strain. But the law according to which matter is distributed 

 from circumference to axis cannot be inferred, since the stated 

 phenomena follow equally well both for surface dilatation and 

 for surface compression. 



All the P. E. drops examined were found to scratch ordi- 

 nary glass with facility ; but on taking the strain out of the 

 drops by annealing them at white heat and slow cooling, their 

 hardness did not materially change. The observed hardness 

 was therefore a property of the glass itself. Indeed by rub- 

 bing together a quenched and an annealed P. E. drop a differ- 

 ence of hardness could not be discerned with certainty. 



Polariscojric Observations. 



Annealing. — Our object in this place is merely to detect vari- 

 ation in the polarization figure, produced either by removal of 

 shells or by annealing. It is therefore sufficient to examine 

 the drops in a given fixed position between polarizing plates.* 

 Disturbances due to diffuse refraction are satisfactorily elimina- 

 ble by submerging the P. E. drop in glycerin, the refractive 

 index of which is nearly that of the glass. The demarcation 

 is then distinct and the colors clear, so that the figures can be 

 drawn. The necessity of grinding special plates is thus fortu- 

 nately obviated ; applied to a quenched P. E. drop the opera- 

 tion would not, of course, be feasible. 



In the case of simple annealing, variations of the polariza- 

 tion figure are due to a diminution of what we may call the 

 birefractive power of the P. E. drop. In the case of removal 

 of shells, variations of the figure result from diminution of 



* We are indebted to Professor Baird for the use of a simple but efficient 

 reflecting polariscope, formerly the property of Professor Henry. 



Am. Jour. Sci. — Third Series, Yol. XXXII, No. 189. — September, 1886. 

 12 



