fod 
Draper. ] 8 i July 20, 
example, my hydrogen tubes gave a spectrum photograph of fifteen lines 
of which only three belonged to hydrogen. In order to be sure that none 
of these were new hydrogen lines it was necessary to try tubes of various 
makers, to prepare pure hydrogen and employ that, to examine the spec- 
trum of water, and finally to resort to comparison with the Sun. 
The object in view in 1873, at the commencement of this research was 
to secure the means of interpreting the photographs of the spectra of stars 
and other heavenly bodies obtained with my 28 inch reflector. It soon ap- 
peared that the spectra of Nitrogen and other gases in Pliicker’s tubes could 
be photographed and at first some pictures of Hydrogen, Carbonic Acid 
and Nitrogen were made because these gases seemed to be of greatest as- 
tronomical importance on account of their relation to stars, nebulee and 
comets. Before the subject of comparison spectra of the Sun was carefully 
examined there was some confusion in the results, but by using Hydrogen 
the source of these errors was found out. 
But in attempting to make a prolonged research in this direction, it soon 
appeared that it was essential to be able to control the electrical current 
with precision both as to quantity and intensity, and moreover, to have 
currents which when once adjusted would remain constant for hours to- 
gether. These conditions are almost impossible to attain with any form of 
battery, but on the contrary are readily satisfied by dynamo-electric ma- 
chines. Accordingly, I sought fora suitable dynamo-electric machine and 
motor to drive it, and after many delays procured a combination which is 
entirely satisfactory., I must here acknowledge my obligations for the 
successful issue of this search to Professor George F. Barker, who was 
the first person in America to procure a Gramme machine. He was also 
the first to use a Brayton engine to drive a Gramme. 
The dynamo-electric machine selected is one of Gramme’s patent, made 
in Paris, and is a double light machine, that is it has two sets of brushes, 
and is wound with wire of such a size as to give a current of sufficient in- 
tensity for my purposes. It is nominally a 350 candle light machine, but 
the current varies in proportion to the rate of rotation, and I have also 
modified it by changing the interior connections. The machine can pro- 
duce as a maximum a light equal to 500 standard candles, or by slowing 
the rotation of the bobbin the current may be made as feeble as that of the 
weakest battery. In practical use it is sometimes doing the work of more 
than 50 large Grove nitric acid cells, and sometimes the work of a single 
Smee. 
The Gramme machine could not be used to work an induction coil when 
it first reached me, because when the whole current was sent through the 
Foucault interruptor of the Ruhmkorff coil, making 1000 breaks per minute, 
the electro-magnets of the Gramme did not become sufficiently magnetized 
to give an appreciable current. But by dividing the current so that one 
pair of the metallic brushes, which collect from the revolving bobbin, sup- 
plied the electro-magnets, the other pair could be used for exterior work, 
no matter whether interrupted or constant. The current obtained in this 
