550 Mr. F. B. Young on the Field at 



it fell with increasing rapidity. The value of cr is plotted 

 (curve I. in fig. 5) with the distance of the screen expressed 

 as a multiple of R, the radius of the hemisphere, and the 

 curve shows that for distances of about 5*5 R and upwards 

 the distribution is undisturbed. For the purpose of testing 

 the influence of the length of the needle upon the density 

 curve, a small circular screen was made by stretching paper 

 covered with metal foil over a hoop about a metre in diameter. 

 This was mounted on a wooden clip by means of which it 

 could be clamped upon the needle concentrically with it and 

 in a horizontal plane (P in fig. 3). A wire connected the 

 screen to the earthed casing so that when the screen was 

 raised the needle was virtually made shorter. The screen A 

 was maintained at a distance of about 16 R, whilst the screen 

 P was moved to various heights on the needle. The value 

 of cr for a constant value of 6 (0 = 1*309) was found for each 

 position of the screen, and crve II. in fig. 5 was plotted, the 

 abscissae representing the length of the needle which pro- 

 truded through P in terms of R. The curve shows that the 

 distribution of the charge is practically independent of the 

 length of the needle if that length exceeds 18 R. (The 

 length of the needle used in determining the correction factor 

 a was over 50 R.) 



If the end of a platinum wire is fused in the oxy-coalgas 

 flame it will, under the influence of surface tension, assume a 

 very truly spherical surface. Unfortunately, however, the 

 end is rarely hemispherical since the molten platinum usually 

 forms a blob at the end of the wire. The manufacture of 

 such points is much simpler than the grinding of steel ones, 

 and it was therefore desirable to determine to what extent 

 the value of a was affected by the blob. Three stalks of 

 various diameters were made ; and the large sphere was 

 mounted upon these in succession to represent a blob. 



The curves A, B, C, and D in fig. 6 are the density curves 

 obtained for stalks whose radii are respectively R, *82 R, 

 '68 R, and '51 R. The tension curve D' is the one derived 

 from D, and shows how the charge beneath the hemisphere 

 diminishes the pull of the screen upon the needle. The 

 density of the curve falls off less rapidly as the stalk becomes 

 narrower, and there is consequently an increase in the pull 

 on the upper portion of the sphere which partially but not 

 wholly counteracts the back-pull upon the under surface. In 

 calculating the value of a from the derived curves the negative 

 area was of course subtracted from the positive area. 



