200 Transactions of the Royal Microscopical Society. 
and in any spectrum under examination, botli being affected in the 
same manner and to the same amount ; and, moreover, the position 
of a band in any part of the spectrum is seen at once, without 
having to measure from it to some more remote fixed point. Another 
great advantage is that the measurements given by the interference 
bands agree far more closely than any others with the law of the 
position of absorption bands, and thus the observer is able at once 
to draw several important conclusions from the actual measurements, 
without previous reduction to another scale. 
3. Wave-length Method. 
Until quite recently, in common with nearly all writers on 
absorption spectra, I have given the position of the bands in 
reference to an arbitrary scale ; but I feel quite convinced that for 
the future we ought all to adopt the plan now employed by many 
in the case of luminous spectra, and express everything in terms of 
wave-lengths. For the purposes of the spectrum microscope, it 
appears to me that the most convenient units are millionths of a 
millimeter. It is just possible in some cases with great care to 
measure to that degree of accuracy, but not to a smaller quantity, 
and therefore it seems undesirable to use four figures when three 
will express all that can be certainly determined. In order to be 
able to make use of this system I have with great care constructed 
a table, giving the wave-lengths of every -y-th division of my quartz 
interference scale, so that, after having measured the position of 
any part of the spectrum by means of this scale, I can by the 
use of the table at once express everything in terms of millionths 
of a millimeter of wave-length. I propose to publish this table and 
give with it the means of correction, in case anyone should have a 
quartz scale not accurately corresponding to my own, so that the 
error would be of no practical importance, and each observer could 
express the measurements made by his own scale in accurate wave- 
lengths. Not only will this, as I trust, lead to the use of an 
uniform system, but measurements thus expressed, having a defi- 
nite relation to the most important physical character of the light 
of the different parts of the spectrum, and not being in any way 
arbitrary, there is a priori a far greater probability that a com- 
parison of the results will lead to the discovery of true general laws. 
Since I have adopted this system I have been led to look on the 
whole subject from an entirely new point of view, and to perceive 
the importance of many phenomena which previously did not appear 
to have any well-marked significance. On the present occasion I 
do not think that I could do better than describe a number of facts 
w'hich seem to point out that the wave-length system is that likely 
to lead to the most perfect development of the spectrum method of 
research. 
