Chemistry and Physics. 767 



longitudinal metal plate. Each compartment was provided with 

 its own insulated electrode. The axis of the "reflected" beam 

 of homogeneous rays coincided with that of the ionization cham- 

 ber so that equal portions of the beam entered the upper and 

 lower compartments. The layers of absorbing material were 

 placed in front of the upper compartment but not the lower. 

 Readings of the electroscope were taken first with the upper elec- 

 trode joined to it and then with the lower one connected with 

 it. In this way the lower compartment furnished the data neces- 

 sary for the standardization of the readings taken with the upper 

 compartment and its associated absorbing screens. "Very con- 

 sistent results were obtained in this way, the points all lying very 

 evenly on a straight line — indeed, the results for the absorption 

 coefficient obtained in independent experiments rarely varied by 

 more than 1 per cent., or very occasionally by 2 per cent." 



The mass-absorption coefficient /x/p is tabulated, for aluminium 

 and copper, for eight wave-lengths extending from 0-431 X 10 _s 

 to 0-627 X 10~ s cm. When fx/p for copper is plotted against fx/p 

 for aluminium, segments of two sensibly parallel straight lines 

 are obtained. The break in the locus occurs at about A = 0-49 A 

 (coef. = 2) which fact is very significant since Barkla found 

 evidence of the emission by aluminium of a J radiation corre- 

 sponding to a mass-absorption coefficient approximately equal 

 to 1-9. Finally, the formula fi/p = a-k n + C gave excellent 

 agreement with the experimental data. For aluminium and 

 copper the values of n were found to be 3 and 5/2 respectively. 

 The latter datum is quite exact and it agrees with Owen's fifth- 

 power law of absorption. — Proc. Roy. Soc, 94 A, 567, 1918. 



H. S. U. 



7. Flame and Furnace Spectra of Iron. — From his exhaustive 

 investigations of the tube-furnace spectrum of iron A. S. King 

 has deduced values for the effective temperatures of the follow- 

 ing flames: air-coal gas (mantle), oxy-hydrogen, oxy-coal gas, 

 oxy-acetylene, and air-coal gas (cone). Since these values differ 

 markedly from the values obtained directly by E. Bauer, who 

 worked with the same flames that had been studied spectroscop- 

 ically by Hemsalech and de Watteville, the whole question has 

 been recently subjected to a very thorough investigation by G. A. 

 Hemsalech. The discrepancies were due, at least in part, to the 

 fact that King used a high-dispersion grating spectrograph while 

 Hemsalech and de Watteville employed an ordinary prism appa- 

 ratus. Consistent results were obtained as soon as Hemsalech 

 set up his own tube-furnace and made all spectroscopic observa- 

 tions of both the flame and the furnace spectra with the same 

 spectrograph. 



The conclusions at which Hemsalech arrived may be sum- 

 marized in the following sentences : 



(a). The spectra of iron given by an electric-tube resistance 



