224 
cated that an initial cooling rate of 
2.33° C. per day and a maximum rate 
of 60° C. per day would be permissible, 
it was decided to use an initial rate of 
about 1° C. per day and a maximum 
rate of not over 10° C. (preferably 6° 
C.). The reason for such conserva- 
tism was that Adams and Williamson’s 
findings were based on the annealing 
of much smaller pieces of glass, and 
that the validity of extrapolations 
from data obtained on small pieces to 
the case of a much larger piece was 
questionable. 
The cooled disk was removed from 
the mold and the insulating bricks 
stripped off. Before examining it for 
strain, the bottom surface was ground 
with carborundum and water to re- 
move the adherent particles of brick. 
Placed on edge, with the ground 
surface coated with mineral oil, the 
disk was examined for residual strain 
by projecting a divergent beam of 
polarized light against one surface 
and examining it through a Nicol 
prism from the opposite side. ‘The 
interference figure was found to be 
very symmetrical, with its intersection 
in the center of the disk. The maxi- 
mum strain detected produced a 
retardation equivalent to about 6 
millimicrons per centimeter of thick- 
ness, well within the maximum strain 
permissible at that time in optical 
glass used for optical instruments of 
highest precision. 
All that remained was to cut the 8- 
inch hole for the Cassegrainian mount- 
ing (see pl. 4, 1). This was accom- 
plished in about 70 working hours and 
without any unforeseen difficulties. 
So ended an operation that took over 
15 months: 6 months and better for 
making and air-drying the pot; about 
15 days for burning the pot, melting, 
and pouring; 7}; months for annealing 
and cooling; and for strain inspection 
and cutting the hole, spare time as it 
became available. And this does not 
include the time necessary for making 
and drying the mold for the pot. 
The task presented to the Corning 
Glass Works by the California Institute 
ANNUAL REPORT SMITHSONIAN INSTITUTION, 1948 
of Technology in 1931 was to make a 
mirror blank of low-expansion boro- 
silicate glass and of sufficient size to 
produce a telescope disk 201 inches in 
diameter of materially less weight than 
the 40 tons which would result from 
the customary thickness of one-sixth 
the diameter. As an aid to proving 
the feasibility of whatever program 
might be adopted for manufacture, the 
Observatory Council of the California 
Institute of Technology ordered two 
small disks of 30 and 60 inches in 
diameter, respectively. The design of 
these was to be similar to that of the 
large blank to demonstrate the prac- 
ticability of reducing the weight with- 
out sacrificing rigidity and precision 
of the figure in the finished mirror. 
Also, a disk 120 inches in diameter, of 
similar design, was desired for use as 
an optical flat with which the 200-inch 
disk could be tested by the opticians 
during the figuring process. Thus, 
along with the task of making the large 
disk, there were furnished stepping- 
stones by which fulfillment could be 
reached. 
For this purpose, tank melting was 
the logical choice. It was felt that the 
composition from one pot to another 
would be subject to more variation 
than would be found in glass from a 
single filling of a large day tank. The 
glass employed required a higher melt- 
ing temperature (1580° C.) than was 
obtainable with available pot furnaces. 
Also, ladling equipment was less costly 
than installing means for pouring the 
glass from pots. Ladling permits one 
to rake the surface of the bath when 
necessary to free it of objectionable 
scum or floating stones before filling 
the ladle. It also affords an opportu- 
nity to inspect the quality of the glass 
prior to pouring it into the mold. 
The mold was complicated because 
of the ribbed structure required in the 
lower portion of the finished disk. 
The cores were made of high-tempera- 
ture insulating brick of standard size 
laid with high-temperature cement, 
and they were anchored to the iron 
bedplate of the mold by means of a 
