38g Reports on Special Researches 



The fan and air-meter attached to the apparatus are of the same type as those supplied 

 by Messrs. Gunther and Tegetmeyer, with the usual Ebert apparatus; the electroscope is, 

 however, as already stated, of the single-fiber type, and is adjusted to a sensitivity of from 

 5 to 10 divisions per volt. 



In Plate 22, Figure 2, to the left of the figure, are shown the electroscope system, fan, and 

 meter. A box, containing the battery for the plates of the electroscope, is attached to the 

 base of the apparatus, and the whole instrument is suspended from a gimbal system. 

 The open end, which is shielded by a hood, projects through a hole in the roof of the 

 observing-house. 



In making the observations, the fiber is released from earth, and is allowed to move 

 over a fixed range of the scale of the electroscope as in the conductivity observations, the 

 meter readmgs during the transit of the two marks being noted. Apparent leakage may 

 result from two causes, a true leak arising when the fiber departs from zero potential and 

 small alterations in the potential of the battery during the experunent. The former effect, 

 which is in general very small for the small potential increase of the fiber, is at any mstant 

 proportional to that increase, while the latter is mdependent of it. Provided, however, 

 that the "apparent leakage" correction is determined when the fiber is charged to the mid- 

 point of the fixed range, we need not concern ourselves with any distinction between the 

 two types of effects. 



If C is the capacity of the apparatus, and 57 the magnitude of the fixed range, n the 

 ionic density, e the electronic charge, and W the au--flow during the movement over the 

 range 5 V, we have 



JT 



where t is the time of duration of the observations, and -^ is the rate of alteration of 



potential in the leakage experiment. 



If t is the time which would elapse during the entry of CdV units of electricity into the 

 apparatus by apparent leakage, the potential being maintained constant, we have C&V = 

 neW+CSYr/t, and if w is the quantity of air which would flow through the apparatus in 

 the time t, r/t = W/u, so that' 



n = {W ^ — oj ') 



e 



The procedure with regard to taking alternate readmgs for n+ and n_ is exactly 

 analogous to that adopted for the conductivity, and for each set the mean value of W'^ for 

 the set is used in the formula, as also the mean of the values of w~' obtained at the beginning 

 and end of the set. The results are recorded as in the example shown on page 398, the 

 relation between the air-flow and meter-reading being obtained from data suppUed with the 

 meter. It will be noticed that the tables for recording the conductivity and ionic content 

 bear a strong resemblance to each other, a fact of considerable advantage. 



In view of the suspicion which naturally attaches to the mdications of any small meter 

 operating on the fan principle, arrangements have been made during the latter part of 

 1916 to recaUbrate this instrument aboard ship from tune to tune. 



Before making the leakage tests with the conductivity apparatus and the ion-counter, 

 the fans are allowed to draw air through the instruments for 5 minutes, so that the insulat- 

 ing material may attain that degree of dampness which it will have during the experiment. 



'It will be observed, from the mode of deduction of this formula, that, when the apparent leakage acts in the same sense 

 as the ions entering the apparatus, u is to be inserted as a positive quantity. In the usual case, however, where the leakage is 

 a true one, w is negative. 



