138 LYMAN. — THE SPECTRUM OF HYDROGEN. 
measured wave-length has the value X 1030.8 but beyond this there are some very 
faint lines whose wave-length must be between X 1000 and X 1010. At present these 
lines form the limit of the spectrum. 
The nitrogen-like appearance of the spectrum of air shown in fig. 1, Plate VIII 
deserves attention. The fluted bands are beautifully clear in the original negative 
and their gen 
be seen in the reproduct 
EFFECT OF CAPACITY ON THE SPECTEA. 
The spectra both of air and hydrogen were obtained with no capacity in circuit 
with the discharge tube beyond that afforded by the connections of the apparatus. 
The effect of capacity on the spectrum of both gases in the visible is so striking, how- 
ever, that it seemed worth while to study the phenomenon in this new region of short 
wave-length. Moreover, the recent attempts which have been made to extract from 
the change in spectrum with change in condition some evidence as to the nature of 
the vibrating system of electrons, make such experiments doubly interesting. For in 
on we are dealing with vibrations more than three times as rapid as 
this new reg 
V P 
those studied in the visible spectrum. This difference in rapidity might well be 
expected to differentiate the effect produced by a given change of condition on the 
visible spectrum from the effect produced by the same change on the region between 
X 2000 and X 1030. It is even possible to conceive that this differentiation might 
throw some light on the vibrating system itself. 
The research is unfortunately beset with mechanical difficulties. Reference has 
already been made to the trouble experienced from the spreading of the discha 
into the spectroscope and the resulting fog produced on the photographic plate. This 
difficulty is increased a hundred fold if a disruptive discharge is sent through the tube, 
for in this case the whole interior of the spectroscope seems to become luminous and 
a total fogging of the plate results. With great care as to regulation of pressure some 
spectrographs have been obtained, but they have never been perfectly satisfactory, 
since even if but a single spectrograph is taken on a plate the time of exposure must 
be short. When the investigator turns from the direct connected discharge tube to 
dosed from the receiver by a fluorite plate he is confronted by a new difficulty. 
The fog indeed is prevented, but after 
disci large dep 
of the disrupti 
film on the fluorite window and renders it totally opaque 
T us fita need be hardly visible by transmitted light and yet it will be thick ««, 
absorb all wave-lengths below X 1800. The material of the electrode exercises of cc 
a pronounced mfluence, but even with aluminum, which shows the effect the least 
