August 9, 1900] 



NATURE 



341 



effect is principally confined to a layer of gas near the surface 

 (John S. Townsend, Camb. Phil. Proc, vol. x. Part iv.), but 

 when the pressure is low the secondary rays are not so rapidly 

 absorbed by the gas, and the ionisation (w^) between the plates 

 is nearly uniform. 



The ratios of — were determined for different forces, the air 



being at a pressure of two millimetres. When the strength of 

 the rays was reduced to ^ of its original valu3 it was found that 



the ratios — were unaltered. This shows that a is in- 



dependent of n^ and is some function of X and /. 



The plates were then set at one centimetre apart, and the 

 values of c were determined for different forces. The results, 

 corresponding to a pressure 2 and 8 mm., are given in the 

 second columns of the accompanying tables. The numbers 

 given are the mean between the currents in opposite directions. 

 With this form of apparatus, however, there were only very 

 small differences found in the conductivity when the electro- 

 motive forces were reversed. The plates were then set at two 

 centimetres apart, and the currents found in this case for 

 pressures 2"I4 and '8 mm. are given in the third columns of 

 the tables. 



The force X is given in volts per centimetre. 



X 



20 

 40 

 80 

 120 

 160 

 180 



Table 



X 



!0 



20 



40 



80 

 1 20 



165 ... 

 186 



The tables show that the current increases more rapidly with 

 X when the plates are two centimetres apart than when they are 

 one centimetre apart. This effect cannot be attributed to a 

 surface action which would be independent of ci when X remains 

 constant. 



From the formula — = _ ( E"'^- i ) we can deduce the values 

 c„ cui\ J 



of o from the third columns of the tables, by making rt'=2 and 

 Cn the smallest value of c. From values of o thus obtained, the 



ratios — for the different forces corresponding to plates I centi- 

 metre (af=i) were calculated. The values of c found in this 

 manner are given in the fourth columns, and they show a good 

 agreement with the experimental determinations. 



Other experiments for different pre.ssures have also been 

 made, and they all show an agreement with the present theory. 



For the purpose of deciding whether it is the positive or 

 negative ions which produce oiher ions by their rapid motion 

 through the gas, we may mention the following experimental 

 results. When the lines of force in the gws are not parallel, 

 large differences in current were obtained on reversing the 

 electromotive force. Thus, when the conductivity takes place 

 between two electrodes one inside the other, it was found that 

 for high electromotive forces the current is much greater when 

 the ions go towards the inner electrode. 



Thus, with an apparatus consisting of a small spherical 

 electrode surrounded by a large electrode made of thin aluminium, 

 the currents, when the outside electrode was positive, were 14 

 for a potential difference of 40 volts, and 34 for a potential 

 difference of 300 volts ; when the outside electrode was nega- 

 tive the currents were 14 and 174 for the same voltages. In 

 these experiments the pressure was about 2 mm. The posi- 

 tive and negative ions produced by the rays are generated 

 nearly uniformly throughout the area between the electrodes. 

 When the large electiode is positive only a few of the negative 

 ions pass through the region round the small electrode where 



the force is big, and the current only increases from 14 to 34. 

 When the electromotive force is reversed all the negative ions 

 produced by the rays come into the region where the force is 

 big, and the current is thereby increased from 14 to 174. It 

 is therefore evident that the increase of conductivity must be 

 attributed to the rapid motion of the negative ions. 



I hope in a future paper to give a fuller account of the above 

 experiments, and also to point out some of the applications of 

 this theory to the passage of electricity through gases. I may 

 mention that the high conductivities obtained with ultra-violet 

 light (Stoletow, Journal de Physique (2), 9, pp. 463-473, 

 1890), at pressures of about I millimetre, may be explained by 

 this theory. 



Approximate values of the energy of translation of the nega- 

 tive ion when producing another ion by a collision can also be 

 obtained from the coefficients a. J. S. TOW.VSEND. 



Trinity College, Cambridge. 



A Remarkable Hailstorm. 



I HEREWITH enclose you prints, from untouched negatives, 

 of hailstones which fell at Northampton on Friday, July 20. 



The drawing board measures 19^" by 17", and the average 

 circumference of the hailstones upwards of five inches. These 

 are by no means the largest that fell, according to the state- 

 ments of trustworthy persons, but were typical of what fell in 

 my garden. 



Fig. I. — Group of hailstones which fell at Northampton on July 20. 

 Size of bo.-ird 19J in. by 17 in. 



The majority of the stones were somewhat flattened, Jas shown 

 in the front of the photograph, but many were nearly spherical 

 like those in my hand (Fig. i). 



The stones were extremely dense and well frozen, and buried 

 themselves in the garden soil. Where they fell on hard surfaces, 

 they usually broke into fragments which rebounded to consider- 

 able heights, while glass roofs suffered enormous damage all over 



Fig. 2.— Sections of hailstones (Northampton, July 20). 



the area, some twelve miles by six, covered by the storm. I have 

 a piece of glass 5/i6ths of an inch in thickness many hundred 

 square feet of which were bro'cen at the various factories in the 

 town. 



The sections (Fig. 2) were an afterthought and show the 

 structure exceptionally well in two instances. 



J. G. Roberts. 



Northampton and County School, July 30. 



NO. 1606, VOL. 62] 



