Cryjlallizations obferved on Glafs, 537 
increafed, for the denfity of a friece of cryftallized glafs 
was found, by experiment, to be to that of water as 267 6 
to 1000; whereas the denfity of a piece of uncry ftallized 
glafs, which had been contiguous to the former, and ccn- 
iequently had been compofed of the fame materials, and 
expofed to the fame heat and other circumftances, was to 
the denfity of water as 2662 to 1000. The brittlenefs 
of glafs is diminifhed by cryftallization ; for cryftallized 
glafs is lefs apt to crack by change of heat and cold. 
Cryftallization is always accompanied or preceded by 
an evaporation of the lighter and more fluid parts of the 
glafs; for I found, that, by expofmg a piece of tranfpa- 
rent glafs till it was entirely cryftallized, one fifty-eighth 
part of its weight was loft by evaporation: and I am in- 
duced to believe, from other trials, that glals, which con- 
tains too large a proportion of laline fluxes, is lei's capa- 
ble of cryftallizing than other harder glafles, till it has 
loft its fuperfluous quantity of fucli fluxes by evapora- 
tion. A doubt may therefore arife, whether the change 
of properties, induced by cryftallization, be merely the 
effe&s of altering the texture, that is, the arrangement 
of the minute integrant parts oi glals ; fince this change 
is always accompanied with a lofs of the lighter parts of 
the cryftallizing fubftance. But, although a fupeifluous 
quantity of faline or other fluxes may impede the cryftal- 
lization, yet, that the change of properties, induced by 
cryftallization, is principally or folely the effect of an 
alteration of texture, is evident from this obfervation ; 
that a piece of cryftallized glafs, when expofed to a heat 
confiderably 
