234 



NATURE 



[November 6, 1919 



particles in a given volume was found by dividing' 

 the weight of the cloud by the weight of each 

 particle estimated from the rate of fall of the 

 cloud. The charge on each particle was then 

 obtained by dividing the total charge by this 

 number. Various corrections and improvements 

 were later introduced by Millikan, and the 

 charge on an ion in a gas has been found to be 

 47 X iQ-i" electrostatic unit. 



The remarkable discoveries made by Rontgen 

 and Becquerel, which have led to so many ad- 

 vances in the knowledge of molecular physics, 

 were of invaluable assistance in providing means 

 of studying the properties of ions in gases. It 

 was found that X-rays and the rays from radio- 

 active substances made gases conduct, and it was 

 possible to obtain ionisation at a uniform rate in 

 gases under various conditions, with the advan- 

 tage that the mass associated with the ions was 

 not liable to irregular change. 



Special experiments were devised to determine 

 the rate of recombination of positive and negative 

 ions, the velocity of the ions under electric forces 

 and their rate of diffusion ; and various properties 

 of ions in gases were discovered. 



The experiments on diffusion, for instance, led 

 to a method of finding the number of molecules 

 per cubic centimetre of a gas and of comparinsj- 

 the charges on ions in liquids and in gases. If 

 N be the number of molecules per cubic centi- 

 metre of a gas at atmospheric pressure and 15° C, 

 and e the charge on an ion in a gas, a direct 

 determination of the product N x e is given by 

 observing the lateral diffusion of a narrow stream 

 of ions. The value of N x e thus found is 

 1.23x101", and as the charge e was also found 



by other experiments, the value of N is seen to 

 be 2-6 X 10''. 



If E be the charge on a hydrogen ion in a 

 liquid, the total charge 2N x E carried by all the 

 atoms in a cubic centimetre of the gas is equal 

 to the quantity of electricity required to evolve 

 that volume. When expressed in electrostatic 

 units, the latter quantity is 2-46 x lo'". Thus the 

 two charges E and e are the same. 



Another line of investigation was undertaken 

 in order to discover how ions are generated in 

 large numbers, as when small changes of force 

 convert a gas from an insulator to a conductor. 

 It was found that when ions are generated by 

 Rontgen rays or by ultra-violet light a maximum 

 current composed of ions generated by the rays 

 was obtained with small forces, but as the force 

 increased beyond a certain point new ions are 

 generated in the gas by the motion of those pro- 

 duced initially by the rays. At first the new ions 

 are produced by the collisions of negative ions, 

 or electrons, with molecules of the gas, and as 

 the force increases and approaches the value 

 required to produce a discharge, the positive ions 

 also acquire the property of generating others by 

 collision. 



The theory of ionisation by collision was foimd 

 to be in accurate agreement viith the experimental 

 determinations of the forces required to produce 

 spark discharges, brush discharges, and the 

 corona discharge which is accompanied by a glow 

 over the surface of a wire or cylinder. 



Thus the various properties of ions which have 

 been discovered in the last fifty years have already 

 explained many phenomena connected with elec- 

 tric currents. 



SPECTROSCOPIC ASTRONOMY 



By Prof. A. Fowler, F.R.S. 



4 



'T'HE science of celestial chemistry and physics 

 ^ was brought into existence in 1859, when 

 Kirchhoff's famous experiment on the reversal of 

 spectral lines furnished the key to the interpreta- 

 tion of the dark lines of the solar spectrum, and 

 thence to the determination of the composition 

 of the sun and stars. The new science developed 

 with extraordinary rapidity, and within ten years 

 the spectra of all the different classes of celestial 

 bodies had been carefully observed. The gaseous 

 nature of some of the nebulae had been discovered 

 by Huggins, and a spectroscopic classification of 

 stars had been made on such sure foundations by 

 Secchi that it still survives as one of the most 

 convenient modes of describing the main features 

 of stellar spectra. The memorable discovery by 

 Lockyer and Janssen of the method of observing 

 solar prominences without waiting for an eclipse 

 of the sun was also made during this fruitful 

 period, and the possible determination of the 

 radial motions of stars by displacements of the 

 spectral lines had been put to a practical test by 

 Huggins. The demonstration that the immensely 

 NO. 2610, VOL. 104] 



distant celestial bodies were composed, in part at 

 least, of the same kinds of matter as the earth 

 may well take rank among the greatest triumphs 

 of science. 



The half-century which has elapsed since the 

 first issue of this journal has witnessed a progress 

 which must far exceed the highest hopes of the 

 earlier workers. Some of the advances have fol- 

 lowed from the increased apertures of the tele- 

 scopes which collect the light for spectroscopic 

 examination, but many more are to be attributed 

 to the substitution of photographic for visual 

 methods of observation which was made practic- 

 able by the introduction of the gelatine dry plate. 



Great observatories dedicated to astrophysics 

 have Been erected, notably in America, and 

 observational methods have reached a high degree 

 of refinement. In solar investigations, where the 

 great intensity of the light allows of the use of 

 instruments of high resolving power, velocities on 

 the sun's surface can now be measured with a 

 probable error of only a few metres per second ; 

 and even more remarkable is Hale's determina- 



