168 
MR. H. L. CALLERDAR OR THE PRACTICAL 
Crosses were made, with a writing diamond, on the extremities of the tube, at 
a distance of about 15 cm. from the gas furnace, and were observed with reading 
microscopes. 
The reading microscopes were made by the Cambridge Scientific Instrument Society, 
and are all but perfect. The microscope is fixed in a carriage, which is moveable 
horizontally by means of a micrometer screw reading to ’00002 of an inch. 
They were mounted on solid stone blocks, resting on a thick stone slab which 
supported the gas furnace, so screened with bright tinned plate that the microscopes 
and their supports, and the stone slab, were not appreciably heated. 
After the apparatus has been adjusted the microscopes are placed in position and 
focussed on the crosses, and are not subsequently touched, except with the micrometer 
screws, in taking readings. 
Since the success of the method depends on the distance between the stands of the 
reading microscopes remaining invariable, this was verified, from time to time, by 
means of an auxiliary glass tube, 1 metre long, with diamond crosses, which could be 
adjusted in position beneath the microscopes without disturbing them. No varia¬ 
tions of the distance could be detected, and they were probably less than ’0001 
of an inch. Several independent readings seem to show that the probable error of a 
single reading due to all causes is about of the magnitude of '0002 inch. 
Readings were also taken of the temperature of the air, and of the ah’ between the 
gas furnace and the stone slab; the variations were usually small enough to be 
neglected. The crosses on the glass tube could be very accurately brought into focus 
beneath the microscopes by adjusting the levelling-screws on which the gas furnace 
was supported ; this is a most important provision, as the furnace expands with heat 
and puts the tube out of position. 
With the above apparatus observations were taken on different occasions. The 
method employed was to light the gas furnace and wait until the temperature became 
steady, and then take several readings of resistance and of the microscopes on the 
plan shown in Table I. 
It was soon found that the length of the glass tube was liable to permanent altera¬ 
tion owing to change of structure. For this reason the micrometer-screws were 
usually read only to the nearest toMAo °f an inch, and the other approximations 
were made in the calculation, which would have been inadmissible if the phenomenon 
observed were accurately self-consistent. 
Table I. shows the irregularities produced by these structural changes. Table II., 
however, contains two series of observations strictly comparable with each other, the 
highest temperature observation being taken first to avoid the permanent contraction. 
For the sake of comparison, the third column gives the expansion calculated from the 
formula l/l {) = 1 + ‘0000068 pt -f 27 X 10 -9 p>£ 3 , which very nearly represents the 
observations, and which gives for the cubical expansion 
v/v 0 — 1+ ‘0000 204 pt + 81 X 10 ~ 9 pt?, 
pt being the temperature Centigrade by platinum wire. 
