TRANSACTIONS OP SECTION A. 573 



This device cannot be used successfully -with direct currents, since though, 

 theoretically, the voltage distribution would be the same as for alternating 

 potential diflerences, the defective insulation of all condensers (unless constructed 

 with excessive care) allows the charge to leak away, and the actual reading of 

 the voltmeter depends on the ratio of the resistances of the two branches of the 

 circuit. 



Very high insulation of the condensers is not desirable when thej' are used with 

 alternating currents, as an accidental electrostatic charge would be retained and 

 vitiate the readings. 



The two capacities may be combined to form one piece of apparatus. The 

 condenser is arranged with one set of conducting plates all connected together, 

 the other set being split into two groups, one of which forms, with the corre- 

 sponding plates of the first set, one condenser, and the second group, with the 

 remaining plates of the first set, the second condenser. For high voltages the whole 

 may be immersed in oil. 



Experiments have baen made with this device, using various ratios, including 

 one of 20 : 1, in which a P.D. of 10,000 volts was measured with an ordinary 

 oOO-volt electrostatic instrument. 



This method has been previously described by W. Penkert in ' Elecktro- 

 tech, Zeits.' No. 39 (1898), 'Eclairage Electrique,' 1?, p. 332 (1898), and 'Science 

 Abstracts,' 1899, p. 294. 



Sub-section of Astronomy and Meteorology. 

 The following Report and Papers were read : — 



1. Report of the Seismological Committee, — See Reports, p. 77. 



2. Exhibition of Photograjjhs made with the Spectro-Heliograph of the 

 Yerkes Observatory. By A. R. Hinks, M.A. 



3. Radiation through a Foggy Atmosphere. 

 By Arthur Schuster, F.R.S. 



In the theoretical explanation of the appearance of dark lines iu the spectra of 

 the sun and the stars a mass of gas is supposed to act on the incident light by 

 absorption only. When Kirchhofl:' first furnished this explanation it fitted all the 

 facts which were then known, and it was not necessary to go beyond the assump- 

 tion of simple absorption. But difficulties have since arisen. Bright lines are 

 observed to be mixed with the dark lines in some stellar spectra, and even in 

 the sun the H and K line? are bright over a great portion of the disc. 



According to Kirchhoft"'s hypothesis a layer of gas in front of a radiating sur- 

 face can only give bright lines if its temperature is higher than that of the 

 radiating surface; a supposition which in the case of stellar or solar atmospheres 

 is not perhaps impossible, but certainly to he avoided if possible. 



The presence of bright lines admits, however, of easy explanation if we take 

 the scattering of light into consideration, which must, to some extent, take place 

 in a pure gas, and must certainly prevail under the conditions of the condensable 

 vapour in front of stellar photospheres. The scattering of light acts in a different 

 manner from absorption, and should therefore be taken into account. I call a 

 vapour in which scattering plays an appreciable part a ' foggy ' vapour. 



The coefficient of scattering (s) is conveniently defined thus: If streams of 

 radiation of intensity A fall on a plate of infinitely small thickness h, an amount 

 of light is scattered hj the plate which, per unit surface, may be expressed by 

 sAA. Of this ^ sAh is sent backwards and i sA^ forward. Tlie amount of radia- 



