Decembee 29, 1911] 



SCIENCE 



901 



with much rejoicing, only to find the next 

 trial a failure. In fact, more time has 

 been lost through such premature exhibi- 

 tions of docility than in all the frank dec- 

 larations of stubborn opposition ! 



One comes to regard the machine as hav- 

 ing a personality — I had almost said a 

 feminine personality — requiring humor- 

 ing, coaxing, cajoling — even threatening! 

 But finally one realizes that the personality 

 is that of an alert and skilful player in an 

 intricate but fascinating game — who will 

 take immediate advantage of the mistakes 

 of his opponent, who "springs" the most 

 disconcerting surprises, who never leaves 

 any result to chance — but who nevertheless 

 plays fair — in strict accordance with the 

 rules of the game. These rules he knows 

 and makes no allowance if you do not. 

 When you learn them and play accord- 

 ingly, the game progresses as it should. 



As an illustration of the measure of suc- 

 cess attained in this work, I would call at- 

 tention to a recent comparison by Messrs. 

 Gale and Lemon of the performance of a 

 grating of 6|-ineh ruled surface with that 

 of the echelon, the Perot and Fabry inter- 

 ferometer and the Lummer plate. The test 

 object is the green radiation from incan- 

 descent mercury vapor. The spectrum of 

 this radiation had been supposed a simple 

 line, until the interferometer showed it to 

 be made up of five or more components. 

 The whole group occupies a space about 

 one fifteenth of that which separates the 

 sodium lines. 



The grating clearly separates six com- 

 ponents while the more recently devised 

 instruments give from six to nine. Two of 

 these components are at a distance apart 

 of only one hundred and fiftieth of the dis- 

 tance between the sodium lines, and these 

 are so widely separated by the grating that 

 it would be possible to distinguish doublets 

 of one half to one third this value ; so that 

 the actual resolving power is from 300,000 



to 400,000 — of the same order, therefore, 

 as that of the echelon. 



It may well be asked why it is necessary 

 to go any further. The same question was 

 put some twenty years ago when Rowland 

 first astonished the scientific world with re- 

 solving powers of 100,000 — and it was his 

 belief that the width of the spectral lines 

 themselves was so great that no further 

 "resolution" was possible. But it has 

 been abundantly shown that this estimate 

 proved in error, and we now know that 

 there are problems whose solution depends 

 on the use of resolving powers of at least 

 a million, and others are in sight which 

 will require ten million for their accurate 

 solution, and it is safe to say that the 

 supply will meet the demand. 



To return to our comparison of the teles- 

 cope and the spectroscope; while the prog- 

 ress of investigation of the stellar universe 

 will be ever furthered by increased size 

 and resolving power of the telescope, this 

 is very seriously hampered by the turbu- 

 lence of the many miles of atmosphere 

 through which the observations must be 

 made. But there is no corresponding limit 

 to the effective power of spectroscopes and 

 the solution of the corresponding prob- 

 lems of the sub-atomic structures and mo- 

 tions of this ultra-microscopic universe 

 may be confidently awaited in the near 

 future. 



The message we receive from the depth 

 of the stellar firmament or from the elec- 

 tric arcs of our laboratories, come they in 

 a millionth of a second or in hundreds of 

 light years, are faithful records of events 

 of profound significance to the race. They 

 come to us in cipher — in a language we 

 are only beginning to understand. 



Our present duty is to make it possible 

 to receive and to record such messages. 

 When the time comes for a Kepler and a 

 Newton to translate them we may expect 



