G90 PROCEEDINGS OP THE AMERICAN ACADEMY. 



ersed first the silver screen, then the aluminium, and vice versa. From 

 the latter measurements it would appear that the effect of transmission 

 through aluminium is actually to "soften" the rays for silver. With no 

 other pair of metals was this anomalous effect observed, and even with 

 this pair it was not always obtained, but depended much on the state of 

 the tube. An experiment performed only a few minutes before the one 

 showing the " softening " effect gave a quite different result, but the 

 change in the condition of the tube between the two experiments was 

 evident to the eye. Direct experiments showed that the secondary radi- 

 ation diffused from the screens cannot account for the observed effect. 



5. Effect of the Order of the Piece? in a Tivo-Piece Screen upon 

 Transmission through the Screen. 



In order to throw some light on the nature of the phenomena which 

 are involved in transmission, the following set of experiments was 

 devised : 



A screen was made by placing face to face two pieces of different 

 metals, and the deflection obtained from the rays transmitted by this 

 screen was compared with that obtained when the order of the two 

 pieces of metal in the screen was reversed. In practice it was found 

 convenient to prepare two screens, one composed of two metals in one 

 order and the other composed of the same metals in the other order, 

 and to compare the deflections accompanying the interposition of the 

 two screens in succession. In using this method it had to be shown 

 by an auxiliary experiment that there was no difference in the thick- 

 ness of the two pieces of either metal, which might produce a spurious 

 effect or mask a real one. This test was usually applied by reversing, 

 in each screen, the order of the pieces of metal composing it. In all 

 of the experiments the deflections were standardized in the manner 

 before described, by comparison with deflections obtained when no 

 screen was interposed. In some of the experiments the three-seconds 

 exposure was used, and in others the rays were allowed to run until 

 the greatest attainable deflection was reached. The quality of the 

 rays used in the different experiments varied over a wide range, and 

 sometimes it was allowed to vary considerably during a single ex- 

 periment, as was the case in the example given in Table X. Copper 

 and silver were the two metals used in this experiment. The cop- 

 per was 0.()()44 cm. thick, and the silver 0.0019 cm. After reversing 

 the order of the plates in each screen, the data summarized in Table XI 

 were obtained. 



From these measurements it appears that the order of the pieces 



