402 Mr. W. J. Humphreys o)t the Absorption. 



any cause other than simple absorption. If, however, as it 

 Avas invariably with my experiments, the plate is equally 

 affected in the two places mentioned, it is clear that the 

 absorptions were equal, or at least that their difference was 

 less than the experimental errors. 



To explain fully what is meant b}' the term " equivalent 

 thickness": — Let C be a compound of the two elements A 

 and B, let d be its density and t the thickness of a slab cut 

 from it, and let a be that fraction which A is of C. Further, 

 let d' be the density of A in the elementary or uncombined 

 form, and t' the thickness of a slab cut from the element A. 

 Then the two slabs have equivalent thicknesses, so far as A is 

 concerned, when adt = d ! t'. Similarly for B, and also for any 

 number of elements of which a compound might be made up. 



The arrangement of the experiment can be easily under- 

 stood by aid of the accompanying sketch — two sectional 

 drawings, at right angles to each other, of the apparatus used. 



V is the vacuum-tube, L a thick lead plate used to protect 

 the dry plate P, which of course was protected from light also 

 by an ordinary plate-holder H. The compound A and the 

 elements E were placed as shown over holes bored through 

 the lead plate. The diameters of these holes were one centi- 

 metre each and their axes two centimetres apart. The thick- 

 ness of the lead plate was one and one half centimetres. 



The current, approximately four and one half amperes, was 

 taken from a storage-battery of about 118 volts, and used in 

 connexion with an excellent coil and rotary break of the 

 "Willyoung pattern. Very high penetrative power was desired 

 from the first, and consequently one of Queen's adjustable 

 tubes was selected for the work. 



To reduce the errors due to inaccurate grinding as far as 



