COSMIC RAYS MILLIKAN AXD CAMERON" 221 



very recently brought to light the inhomogeneity of the rays thougli 

 (he latest results by Hoffmann and Steinke lead them to support 

 provisionally our findings and to suggest that in the mixture of 

 cosmic rays some may exist even harder than the hardest brought to 

 light by us. These we found to have an absorption coefficient e({ual 

 to 0.18 per meter of water, which corresponds, if computed by Comp- 

 ton's equations, to a wave length 0.00038 A, or an equivalent generat- 

 ing potential of 32,G00,000 volts. Hoffmann,-" in order to explain 

 his latest sea-level readings, assumes comi)ouents of hardness cor- 

 responding to a wave length, computed in the same way, of 0.00029 

 A, or an equivalent generating potential of 41,000,000 volts. 



We began in the fall of 1926 to build new electroscopes of greater 

 sensibility to the cosmic rays in the hope of determining the intensi- 

 ties of these rays more precisely and studying their spectral distribu- 

 tion more discriminatingly; in particular, we wished to test for 

 the presence of still harder rays than could be brought to light by 

 the sensibilit}^ of our preceding instruments; for there were theo- 

 retical reasons for suspecting that still harder rays might exist. 

 These electroscopes will be described in detail in more technical pa- 

 pers. Suffice it to say here that we can now measure the capacities 

 of electroscopes to a few parts in a thousand (0.791 electrostatic 

 unit is the capacity of the instrument with which the following re- 

 sults have been obtained), and that we are now using from eight to 

 fifteen times the sensibility to cosmic rays that we have heretofore 

 employed; so that at sea level we have in our electroscope say 27 

 cosmic ray ions to play with instead of about 2, and at Pikes Peak 

 about 90 instead of 5 or 6. 



In carrying out experiments with this electroscope in Gem Lake 

 in the summer of 1927, the ionization at the surface of the lake was 

 33.6 per cubic centimeter per second, and it decreased with depth of 

 immersion, regularly and very smoothly, to a zero value of 2.4. But 

 this asymptotic value of the ionization-depth curve was only reached 

 at a depth of IGIf. feet {50 meters) instead of at about 54 feet (16.2 

 meters) as in our preceding 1925 Arrowhead experiments. This 

 does not represent a discrepancy between the two sets of results. It 

 means only that the ionization ordinates of the curve have now been 

 multiplied manyfold by the increased sensibility. In spite of this, 

 the series of ionization-clepth readings taken with tlie new electro- 

 scope falls much more smoothly upon the curve, that is, shows less 

 scattering, than was the case before; so that by improvements ic 

 technique the actual sensibility has been multiplied considerably more 

 than eightfold. It is this increased sensibility and precision of 

 measurement alone which is responsible for the fact that at depths 



» Hoffmann, Ann. der Phys., 82, 417 ; 1927. 



