( 680 ) 



In these formulae, A is the thickness of the [)U\te, and 



a{l -\-i)L=:s (31) 



The solution, in so far as it is necessary to our purpose, is 



(;,^l)S'+s — (x— 1)V-* 

 4x 



a^, 





In these cx|n'cssions L and consequently s are now to be sui)posed 

 infinitely small. Replacing e—^ and i&^ l>y 1 — x and l-\-s, one finds 



K-M 



e a. 



The first of ihcse e(piations shows that the amplitudo of the ra^^s 

 reflected l)y the thin plate is infinitely small, so that we may neglect 

 their energy as a quantity of the second order. 



As to the transmitted I'ays, the amount of energy propagated in 

 them will l>e equal to the product of the incident energy by the 

 square of the modulus of the complex expression 



This square is 



'-li' + h' 



C 



whence we deduce for the coefficient of absorption 



C 



On the prohahility icltli irldch one may expect that the centre of 



gravity of a large number of points distributed at random 



on a limited straight line will lie tvithin given limits. 



§ 13. Divide the line into a large number p of equal parts, and 

 call these, beginning at the end .1 of the line, the 1^^ the 2"^^, the 

 3'*^ part, etc. Denote by q the number of points and let q be very 

 much larger than jj. 



We shall imagine the points to be placed on the line one after 

 another, in such a way that, whatever be the position of the points 

 already distributed, a new point may as well fall on one part of 



