Further Spectroscopic Studies on the Luminous Vapour Distilled from Metallic Arcs

L. Rayleigh
1926 Proceedings of the Royal Society A  
I have not succeeded-except as a double summation-in showing what the frequency curve for y'x is whatever be the value of M. That remains a problem for the mathematician who can exhibit in a compact form the result of integrating r+% ir K a,b) = J -b n e~li.a COS 0 b s m 0)2 COgM 1 0 dQ The object of this paper has been to indicate how regression lines far from straight, and correlation surfaces far from normal,* may arise in the case of compound characters even if we are sampling from a normal
more » ... population. Thus we see how readily the old theory of " probable errors " may mislead us. Plates 1-3.
doi:10.1098/rspa.1926.0091 fatcat:xdjeinrgc5grziemq62dld2v6e