soMi: COM i:Mi'OK.iKy .ii>i .ixcis i\ I'livsns ni .Ji7 



In ilif |>lu>i(ii;raphs whirh I rcpnuliice," the imprints of tlu-si' two 

 Ih-.uiis st.md side l)y side. In tlu' tirst of them, l-"ij;. 7. the spectriini 

 of the primary rays is s|x-iially depicted on the upper lialf of the 

 plate: one sees the n, d, and ■> hnes of the A'-series of molyhdenuin. 

 three hnes (the first a tloiihlet ) of wliirh the \va\eleiii;th> ,iro respec- 



% 



Fig. 7 — At)ove, the A.'-siK>ctrum of niolylKienum (a-doublct, /S-liiic. v-line from left 



to right); below, the s[)cctruni of this same radiation after scattering at 90° from 



aluminum (each line doubled). (P. .■\. Ross) 



tively .710— .7l4.\. .()33A, .(ilSA. Below, the spertnini of the 

 sec«>ndary rays scattered at the angle 9 is sprearl out: to each of the 

 primary rays there correspmnds a scattered ray of the same wave- 

 length, and beside it another ray of which the wavelength exceeds 

 that of its companion by the required amount. 



thing. The Compton effeit has lieen demonstratetl only where there are electrons 

 associated with atoms. It may lie that the rebound occurs only from an electron 

 which is connectefl to an atom by some [K-culiar liaison, weak so far as the energy 

 required to break it is concerne<l, but able to control the rcs|>onse of the electron 

 to an impact. Something of this son may have to be assumed to explain why the 

 effect is apjjarcntly not greater for conductive substances than for insulating ones 

 and is certainly feebler for massive atoms with numerous loosely-liound electrons 

 than for light atoms with few. 



" I am indebtcti to I'rofessor Ross for thcsf photographs. 



