An edge correction is often required in determining the dielectric constant of an insulating material by the capacitance method. Theoretical formulas by which the edge correction may be computed have been considered for a number of electrode arrangements. Some of these apply to circular electrodes and some to rectangular electrodes. The error in the dielectric constant which may arise from the use of these formulas is discussed for each of the different arrangements.The accuracies of the methods were experimentally determined by making measurements on the same specimens with the various arrangements of electrodes. The guard-ring arrangement was assumed to give the correct value of the dielectric constant, and the values obtained with the other arrangements were compared with it to determine their errors. Only two methods gave values which were in error by less than 1 percent. Both methods were for arrangements of electrodes in which one electrode was smaller than the other, one being for circular electrodes, the other for rectangular electrodes.Empirical equations were set up which gave values of the dielectric constant which were in error by less than 1 percent for all electrode arrangements, both for circular and rectangular electrodes. These equations are much simpler than the theoretical equations.

The absolute value of a current has been measured by the current balance, using improved coils. The previous work had shown that improvements in the coils were required in order to make them fulfill more nellrly the conditions assumed in developing the equations for computing the absolute value of a current from the force measured by the current balance. These new coils consisted of one moving coil made in the form of a short solenoid and one moving coil and one pair of fixed coils wound from anodized aluminum ribbon in the form of a flat spiral.The results obtained with various combinations of these new coils and one pair of the old fixed coils are consistent among themselves, and their weighted average can be expressed as 1 NBS international ampere=O.999 86 absolute ampere.This result differs appreciably from that obtained in the preceding work, but it is felt that this new value is more reliable because of the improved coils which were used and the greater consistency of the results. CONTENTS Page

The dielectric constant, power factor, and conductivity of purified rubber and of a series of its compounds with sulphur were determined at temperatures from -75°to 235°C. The dielectric constant and power factor were measured at five frequencies from 60 to 300,000~and the apparent conductivity at about 0.002 second and at one minute after the application of potential. The results of the measurements are expressed in both tabular and graphic form. At 25°C . and 1,000^the dielectric constant of purified rubber containing no sulphur was 2.37.With increasing sulphur content the dielectric constant increased to a maximum of about 3.75 at 11.5 percent sulphur, then decreased to a minimum of 2.70 at 22 percent sulphur and again increased to 2.82 at 32 percent sulphur. Under similar circumstances, the power factor increased from 1.6X10 -3 for the rubber alone to a maximum of 93.8X10 -3 for the compound containing 13.5 percent sulphur; it then decreased to about 4.0 X 10~3 at 20 percent sulphur, and again slowly increased to 5.1 X 10 -3 at 32 percent sulphur. The 1-minute conductivity was 2.3X10 -17 mho/cm for the rubber alone. With 12 percent sulphur, it was only 0.5X10 -17 , while at 18 percent sulphur it passed through a sharp maximum of 38X10 -17 , and then decreased to values between 1 and 1.5X10 -17 for compositions between 22 and 32 percent sulphur. Changes of the temperature or the frequency at which the measurements were made shifted the maxima and minima in these curves and modified their heights. For example, at -25°C , the maximum dielectric constant at 1,000-^was 2.8, and was obtained with a compound containing 4 percent sulphur, while at 145°C. the maximum was 4.5 and was obtained for a compound containing about 28 percent sulphur. Comparison of the results of this investigation on purified rubber with previous work done with crude rubber indicates that purification alters the values obtained for the electrical properties, but it does not modify the general manner in which these properties vary with changes in composition, temperature, or frequency.The results may have practical bearing on the selection of rubber compounds for specific uses and in pointing out the manner in which the properties of rubber are related to temperature or frequency. CONTENTS Page 179 3. Change in dimensions with temperature 179 IV.3. ELECTRODES The electrodes were circular disks cut from aluminum sheet and were approximately 0.025 cm in thickness. They were used in pairs 24 and 26 cm in diameter, respectively, as indicated in figure 1. 205-33

The current balance originally used by Rosa, Dorsey, and Miller in 1911 has been somewhat modified and used to determine the absolute value of the ampere. The value in absolute amperes of the current in the coils of the balance was determined from the measured constants of the coils and the electromagnetic force between them. The most important constant which could be directly measured was the ratio of the radii of the coils, which was measured by an electrical method. The values obtained on four sets of coils indicated that there was no error in the measurement of the ratio of the radii greater than three parts in a million.

In the decade preceding 1944 there were published the results of eight determinations of the absolute ohm and seven of the absolute ampere. These determinations were made in the national standardizing laboratories of England, France, Germany, Japan , and the United States. However, only preliminary results of some of the researches have been published, and one value of the absolute ampere was obtained in an experiment that ~as originally developed for another purpose.Hence to obtain the most probable values of the absolute ohm and ampere, it is necessary to ignore some of the published results and to consider some of those remaining as more reliable than others. A critical analysis has been prepared of each of the determinations, and this analysis has been used as a basis for weighting the results.The most probable value of the absolute ohm is given by the relation 1 mean international ohm=1.000 490 absolute ohms.The mean deviation from the mean of the results used in obtaining this value is only 14 parts per million. Further confirmation of the probability of this result is shown by the agreement with the results of two other recent compilers of the absolute-ohm determinations.The most probable value of the absolute ampere is given by the relation 1 mean international ampere=0.999 853 absolute ampere.The mean deviation from the mean of the three results used in obtaining this value is 3 parts per million. Other recent compilers. however, have given larger values, one, 27 parts per million larger, and the other, 119 parts per million larger.