﻿876 Dr. E. N. da C. Andrade on the Carriers of 



flame (if a bead of a sodium salt were employed), indicating 

 that the faster carriers are metallic. Owing to the difficulty 

 of avoiding small traces of sodium, the point was sys- 

 tematically investigated with strontium. With a bead of 

 this metal in the flame, after a quarter of an hour in the 

 field, a large metallic current being observed all the time, 

 the strip showed the strontium lines very distinctly on 

 testing in the Bunsen flame with a pocket spectroscope. 

 Control experiments (1) with the bead in the flame, but no 

 field, (2) with the field, but with the strip in a position in 

 which no metallic current was observed (i. e. below the bead), 

 (3) with the isolated plate charged negatively, gave no 

 strontium. To put the matter to a rough quantitative test 

 the galvanometer was calibrated, and the amount of strontium 

 corresponding to a unit deflexion for a unit time calculated 

 on the assumption that the strontium atom had two elemen- 

 tary positive charges, as in electrolysis. Solutions of SrCl 2 

 were made, one drop of which from a given pipette held a 

 known quantity of strontium. Drops of the different solu- 

 tions were brought successively upon a platinum strip similar 

 to that used to receive the carriers, and slowly evaporated ; 

 the brightness and duration of the strontium lines produced 

 on bringing this strip into the Bunsen flame were compared 

 with the lines given by the electrically deposited strontium. 

 The experiments showed that the amount of strontium 

 deposited was, very roughly, that calculated : the experi- 

 mental errors of this method are, of course, large, but it 

 would probably detect a difference if one of the amounts 

 were double the other. 



Further experiments were then made on the transport of 

 sodium. As a control, two platinum strips were arranged 

 close to the earthed electrode, one being earthed through 

 the galvanometer as usual, the other being connected with 

 the insulated electrode, which was then maintained at a 

 positive potential for some minutes. The earthed strip 

 showed the sodium colour strongly in the flame, the other- 

 strip did not. The strips were interchanged, with the same 

 result. Hence there seems no doubt that the transport of 

 sodium is connected with the metallic current. 



Measurements were made of the time in which, for a given 

 galvanometer deflexion, enough sodium was transported to 

 give a just perceptible coloration of the flame. Aselman * 

 has given for the amount of sodium which will give a per- 

 ceptible coloration for one second in the Bunsen flame in a 

 dark room, 1*5 x 10" 8 mg. NaCl, which is equivalent to 

 * E. Aselman, Ann. der Physik [4] xix. p. 960 (1906). 



