103 
1. The collimator, prism stage and telescope of the 
spectroscope were mounted on separate stands; only one 
prism was used whose angle is 59° 33', and the magnifying 
power of the telescope is 26‘9. 
Only about a quarter of the length of the spectrum could 
be seen in the spectroscope at once, and there being 
no scale, the position of the lines sought after had 
to be measured from known gas lines, by a micro- 
meter screw and pointer attached to the eyepiece. 
2. An induction coil was used with a small Leyden 
I jar in the circuit to obtain the spectrum. Under 
ordinary conditions the length of spark in air, with 
jar in circuit, was l*5c.m, while without the jar a 
spark = 5’2c.m. in length could be obtained. 
3. Two eudiometers were used of the shape shown 
in the margin, being TOc.m. long, l*9c.m. wide, gradu- 
ated up to 160 c.c.s, and having aluminium electrodes 
•25c.m. apart sealed into the side at a distance of 30 
c.m. from the bottom ; the tap at the top serving to 
let in, from a burette, the required amount of gas. 
To get rid of all the air, the eudiometer was filled 
with mercury, by exhausting the air above it with a 
water pump. 
In most cases when the spectrum was observed, 
the gases were under a pressure reduced by one inch 
of mercury, i.e., the mercury stood about one inch 
higher in the tube than in the trough. 
Unless specified to the contrary the gases were 
thoroughly dried by standing for a few hours over 
phosphoric anhydride. 
The phosphoric anhydride was rammed tightly into 
a small tube which can be slipped into the eudiometer 
through the mercury ; a hole being bored through the mass 
of PaOs to allow the mercury to run off the top of it, and 
thus expose its surface to the gas. 
