244 ANNUAL EEPORT SMITHSONIAN INSTITUTION, 1919. 
have been made to crystallize, there is a certain temperature at which 
crystallization proceeds readily, and for small ranges above or below 
that temperature no crystallization occurs. Associated with altera- 
tion of temperature, there is change in the viscosity of the vitreous 
substances, and hence in the freedom of movement of their particles 
among one another. The effect of this can be well seen when a 
crucible full of molten zinc silicate is removed from the furnace. If 
the mass be small so that cooling is quick, it remains a glass and has 
to be reheated before crystallization occurs. With a larger quantity 
and consequently slower cooling, the mass may become wholly 
crystalline, or only partly so if some of the vitreous form reaches a 
temperature at w^hich its viscosity is too great, and freedom of move- 
ment of its particles too small for rearrangement into crystals to 
take place. Some borates are very convenient for illustrating the 
change from the vitreous to the crystalline state. Boric anhydride 
itself has not been made to crystallize, and as might be expected, the 
greater the proportion of it in a borate, the less marked is the 
tendency of a fused mass of that borate to crystallize on cooling. Of 
the three borates, Ca0.2B203, CaO.B203 and 2CaO.B203, the first can 
be obtained vitreous by fairly quick cooling, and pieces of it can, 
with care, be heated again to remove the strain produced, that is to 
say, it can be annealed. The second crystallizes from fusion much 
more readily, cooling has to be quicker to keep it in the vitreous 
state, and only small glassy pieces can be obtained, which crystallize, 
however, on attempting to anneal them. The third borate can only 
be kept vitreous in very small globules. 
Statements, however, of the bulk which can be obtained of any 
readily crystallizable vitreous body require a certain reservation. It 
is well known that crystallizable bodies in the fluid state may be 
cooled considerably below the temperature at which they would 
ordinarily solidify to a crystalline mass, if they are freed from all 
foreign material. Water is a well known example, and among many 
others which could be cited, mention may be made of salol (phenyl 
salicylate). It melts at 43° C, and if the crystallization of a film of 
the molten liquid be watched under the microscope, numerous small 
bubbles of gas can be seen to form during cr3^stallization. The gas 
appears to be modified air. If salol be melted, allowed to solidify, 
and remelted in vacuo, so as to remove all this gas, and the process 
repeated several times, it is found that the molten salol must be 
cooled many degrees (50 or more) below its melting point before 
crystallization takes place. A small crystal of salol will start it 
unless, as can be done on small quantities, cooling has been carried 
far enough to increase the viscosity of the fluid to such an extent 
that the particles have not the freedom of movement necessary for 
