ELISHA MITCHELL SCIENTIFIC SOCIETY 9 
Successful decomposition should meam much more. It 
should mean the discovery of a method which will decompose 
not one, but many or indeed, all of the elements, and the de 
composition of these must not yield a larger number of sup- 
posedly simple bodies, but a small group of one or two or 
three which are common constituents of all. It is quite idle 
to venture upon any prediction whether such a method will 
ever be discovered. Setting aside, then, the direct experi- 
mental proof of the composite nature of the elements as unat- 
tainable at present, let us next examine the indirect evidence. 
It would seem wisest for the present to introduce under that 
heading the spectroscopic work of Lockyer. The results, 
while highly interesting, are too indefinite as yet to speak of 
as having a direct bearing, Yet a careful study of the spec- 
tra of the elements leads us to a strong suspicion that the 
less plausible assumption is the one that the particles which 
gwive rise to such varied vibrations are simple and unitary in 
nature. Lockyer’s most recent work, following up the line 
of his ‘Working Hypothesis’ of twenty years ago,is very sug- 
gestive and may lead to important results (Chemistry of the 
Hottest Stars, Roy. Soc. Proc., LXI., 148; On the Order of 
Appearance of Chemical Substances at Different Tempera- 
tures, Chem. News, 79, 145). Still too much must be assum- 
ed yet for such work to be very conclusive. He writes of 
‘proto-magnesium and proto-calcium,’ and Pickering discuss- 
es a ‘new hydrogen,’ all with an assurance and confidence 
which proves at least how deeply these changes in the spec- 
tra have impressed some of those who have most carefully 
studied them. 
But a more important method of indirectly testing the 
question is through a comparison of the properties of 
the atoms. Such a comparison has been made as to 
the atomic weights. In other words, the idea of the compos- 
ite nature of the elements followed very close upon the adop- 
tion of a stricter definition of them as simple bodies. Dal- 
ton, Prout, Débereiner, Dumas, Cooke and many others have 
aided in developing the idea, sometimes faultily and harmful- 
ly, at other times helpfully. Some fell into the common er- 
ror of going too far, but all were struck by the fact that 
