408 Prof. A. P. Chattock on the Velocity and 



two straight lines. The one for + discharge cut the axis at 

 1810, and that for — at 1240 volts. The only two points 

 appreciably off the lines were those for the highest value of 

 z. This looks as though z were wrong here, for the points 

 fell below the lines as in the pressure-curves ; and the 

 omission of these points seems therefore reasonable. 



It was possible to increase the field by arranging that the 

 discharging point protruded a shorter distance beyond D. 

 In this way fields of 6300 volts per centimetre for + , and 

 5800 for — discharge were obtained, the values of P for 

 which are shown by crosses in curves III. 



With D removed the field is of course a minimum. This 

 was the case with curves I. a ond c. Under these circum- 

 stances, however, the field was so divergent that the point 

 would not travel far enough in its ebonite box to allow of the 

 whole pressure being measured for values of z greater than 

 about 0'6 centim. Only two pairs of curves were thus available 

 in the set without D, and they are represented by the circles 

 in curves III. Excluding the strong field close to the point, 

 the average value of the field between point and plate in this 

 case was 2530 volts per centimetre for + and 2300 for ~ 

 discharge. 



Lastly, to test the effect of variation in the current, three 

 current strengths were used, respectively 1, 2*5, and 4 times 

 that of the rest of the experiments. The integrated pressures 

 for these, divided respectively by 1, 2*5, and 4, are given by 

 the three points at s = 096 centim. The highest point in 

 each case corresponds with the lowest current, and the lowest 

 with the hiohest. 



All these extra points may be said to fall upon the ruled lines 

 about as well as the majority of the original points. It is true 

 that the coincidence is not very startling, but the discrepancies 

 are not more pronounced for one set of conditions than 

 another, and they are probably due chiefly to the difficulty 

 of obtaining the correct values of the integrated pressures. 

 It will be seen from curves I. that the pressure decreases very 

 gradually at the edges of the pressure-area, where the pressure- 

 intensity is also very small. The areas over which these smnll 

 pressures are to be integrated become greater, however, as 

 they are farther from the centre ; and the smallness of the 

 pressures themselves, while rendering them difficult to 

 measure, does not therefore imply that they are negligible in 

 the integration. In addition to this the pressures are not 

 necessarily symmetrical about the centre line of curves I., 

 yet the integration depends upon the assumption that they 

 are. There is thus plenty of room for the observed irregu- 



