Hartley — On the Constitution of Electric Sparks. 367 



foil is circular. Under the action of a rapid succession of sparks 

 this form is generally maintained, but occasionally the perforation 

 is irregular, but never angular, the mica being fused. The fused 

 mica forms a lip round the hole, which sometimes resembles a 

 little crater when examined by the microscope. Sparks passed 

 through paper form a circular perforation as far as the fibres will 

 admit of this. The fibres are torn up at the surface, and below 

 these the perforation is seen. The edges of tin-foil through which 

 the spark has passed are fused and thickened ; in the centre is a 

 dark spot but no visible perforation. 



The Size of the Sparks. — On measuring the area covered by the 

 inner portion of metal, supposed to be gold deposited on tin-foil 

 by single sparks, it was found to be generally 0*016 of an inch in 

 diameter, and the exterior diameter of the surrounding annular 

 raised portion of metal was 0*04 inch. They have been seen as 

 large as 0*02 and 0*048, though the majority appeared to measure 

 no more than 0*002 and 0*006 inch. 



The Rapidity of Passage. — In estimating the rapidity with which 

 a succession of sparks pass when spectra are being photographed, 

 two methods have been employed on different occasions. In 1878 

 I used a vibrating spring as a commutator, and, from the number 

 of times the circuit was broken, the number of sparks necessary to 

 yield satisfactory results was found to be not less than 600 per 

 minute. Latterly the number has been estimated by drawing 

 pieces of paper and tin-foil between the electrodes, and counting 

 the sparks in spaces which have passed in one second of time. 

 These were found to average 23 ; they were as many as 27, and 

 occasionally as few as 16. Paper, however, somewhat interferes 

 with the freedom in passage of the sparks, and we may safely say 

 that fairly long sparks are generally as numerous as 1500 per 

 minute. Short sparks, calculated from the musical note of the 

 commutator, must be very considerably greater in number than 

 this ; but it is not possible to make sure of a spark discharge 

 occurring at each making of contact ; and indeed it is almost a 

 certainty, though this has not been proved, that several contacts 

 occur before the passage of the spark. 



On the Chemical Constitution of Electric Sparks. — We are aware 

 from the spectrum of condensed sparks that they are composed of 

 a mixture of air and the vapour of the electrodes. It is possible, 



