NICHOLS AND HULL. — PRESSURE DUE TO RADIATION. 579 



T 



For T— 23.75 seconds — = 5.94 and 8 = 0.06. Hence 



4 



A = \/5 4 (1-008). 



If 8 = 0, A = \/2 — , consequently an error of 1 per cent in T causes 



an error of 0.8 per cent in A. 



To make sure that the observed radiation pressures depended only on 

 the intensity of the beam, and were uninfluenced by the wave length of 

 the incident energy, the ballistic observations of pressure, the thermal 

 measurements of intensity, and the determination of the reflection 

 coefficients, were carried out for three entirely different wave-groups of 

 the incident radiation. In the measurements designated " through air," 

 no absorbing medium was introduced in the path of the beam between 

 the lamp and the balance except the glass lenses and plates already men- 

 tioned. In the measurements " through red glass," a plate of ruby glass 

 was put in the path of the beam between L 2 and d 3 (Fig. 3). For the 

 observations "through water cell," a 9 mm. layer of distilled water in a 

 glass cell was placed in the path of the beam at the same point. 



The separate observations entering into a single series of ballistic 

 measurements and their treatment will appear from Table II, which is 

 copied direct from the laboratory notebook and represents an average 

 ballistic series. The designations EVC„ WVD„ EVD g , and WVC g 

 mean that the vane C in the first case was on the east side of the rota- 

 tion axis with its silver face toward the light. The subscript g signifies 

 that the glass face of the vane was toward the light. The second 

 column of the table gives the zero reading of the balance before opening 

 the shutter ; the third, the end of the swing produced by a six-second 

 exposure ; the fourth, the deflection of the balance ; the fifth, the 

 ballistic deflection of the lamp galvanometer G 2 . Columns six and 

 seven give the balance deflection reduced to standard lamp. 



The results of all the ballistic pressure measurements "through air" 

 are collected in Table III. In the fourth and fifth columns two values 

 are given for the constant of the lamp galvanometer G 2 ; since revers- 

 ing the magnet on the balance bell-jar to reverse the suspension within 

 affected the constant of the galvanometer slightly the values for the silver 

 and glass faces forward were never the same. The subscripts show to 

 which series, silver or glass, the constant belongs. The values of the 

 lever-arm I of the balance, in the sixth column, are obtained by mea- 



