516 Dr. J. H. Vincent : Electrical Experiments 



the motion of the mercury is of the necklace type, but the 

 interval of approach is in the fan effect much longer than the 

 interval of recession, this being only very slightly longer 

 than the luminous period. The time during which the columns 

 are united is also probably a small fraction of the period, so 

 that nearly the whole time of darkness is occupied by 

 the mercury columns in again coming together to fill the 

 tube. 



The history of the growth of the arc is shown in the 

 picture. It commences with a finite length, grows slowly 

 with increasing velocity until after about a thousandth of a 

 second the velocity of growth is constant, and remains sensibly 

 so until the arc goes out. 



Fig. 10 (PI. VI.) gives a photograph obtained with a lead- 

 glass tube (9 cm., '27 mm. bore). This took a mean current of 

 •25ampere in series with 37 ohms on 105 volts. The time signals 

 are *005 second apart, which scale is adhered to throughout 

 the remainder of the photographs. It will be seen that the 

 luminous interval is not the same for each fan pattern. The 

 dark striae due to the lack of uniformity in the electromotive 

 force are again visible. 



In fig. 11 we have an example of the column breaking in 

 two places. The upper arc is seen to move along the tube at 

 a rate approximately equal to the rate of growth of the length 

 of the main arc in the same direction. This would seem to 

 show that in this case, at least, the mercury in the tube on 

 each side of the main arc" moved as a whole; and the growth 

 of the arc was accompanied by an expulsion of some of the 

 mercury from each end of the tube. One may also regard it 

 as possible in some cases that mercury vapour is present at 

 other places than the portion of the tube occupied by the arc, 

 the growth of the arc being accommodated by the conden- 

 sation of this vapour as the mercury recedes from the centre 

 of the tube. It will be noticed that the healing of the upper 

 column, due to the collapse of the smaller arc, is simultaneous 

 with the sudden increase in luminosity shown by the bright 

 vertical band in the lower fan. The photograph fig. 11 was 

 obtained with a lead-glass tube (9 mm., "24 cm. bore). The 

 volts on the mains were 110 and the current was '2 ampere 

 with 23 ohms in series. 



The phenomena exhibited bv any tube on a given voltage 

 do not change suddenly as we decrease the resistance so as to 

 cross the line AC. The frequency, mean current, and mean 

 volts across the lamp do not change suddenly. The test of 

 having descended below AC is that the arc may go out with 

 the columns still separated. As we decrease the resistance 



