206 J. A. POLLOCK. 
to have the same period of electrical vibration if its peri- 
meter is equal to that of the rectangle. If the eccentricity 
of the ellipse is increased, the perimeter has to be decreased 
to keep the period of vibration unaltered, until in the limit 
when the form becomes circular, the ratio of the perimeter 
of the rectangle to the circumference of the circle becomes 
1°11 for a circle 800 cms. in circumference, the circle being 
made of copper wire 0°33 cms. in diameter and the rect- 
angle, 30 cms. wide, of thin brass wire 0°04 cms. thick. If 
the form of the circuit is further altered so that the major 
axis of the ellipse becomes at right angles to the direction 
of propagation of the waves, the perimeter has to be further 
decreased to keep the period unchanged. 
Pocklington’ has calculated theoretically the period of the 
free electrical vibration associated with a closed circular 
ring and has arrived at the result that the wave length is 
rather less than the circumference of the circle. Kiebitz’s 
experiment and the present investigation, give a value 
rather greater than the circumference. The problem of the 
electrical oscillations connected with closed circuits is 
discussed generally by Macdonald,’ but calculations for 
special cases are not given. 
Diameter of the wire forming the circuits.—St. John,’ 
in experiments with waves along wires, for oscillations of 
the same period, obtains a 5*- increase in the value of the 
wave length along parallel copper wires as the diameter 
of the wires changes from 0°04 to 0°12 cms. In St. John’s ; 
investigation however, the problem is complicated by the 3 
presence of extra capacity at the ends of the circuit. No 
difference of period has been found in the present experi- 
ments between a rectangle made of thin brass wire 0°04 
* Polkington—Proc. Camb. Phil. Soc., 1x., p. 324, 1897. 
* Macdonald—“ Electric Waves,” p. 62. 
? St. John—Phil. Mag., 38, 1894. 
