144 



OCEAN ATMOSPHERIC -ELECTRIC RESULTS 



been included in a single plot. By this arrangement a 

 total of 1470 simultaneous observations has been utilized. 

 The corresponding frequency curve also is given in the 

 figure. In the case of cruise VII data, for which the 

 curves are given in figure 12B, there is a total of 957 

 simultaneous observations on ion content and conductivity, 

 including both positive and negative elements. As will 

 be seen for both sets of data, the ion content remains 

 essentially constant as the computed mobility increases, 

 until the mean value is reached. The ion content then 

 diminishes as the computed mobility is still further in- 

 creased. The conductivity, on the other hand, gradually 

 increases at the low values of computed mobility and 

 then assumes a more or less constant value at about the 

 mean mobility and remains essentially constant as the 

 computed mobility is further Increased. The curves are 

 identical in character for the early cruises and for the 

 last cruise. Such a relation between the mobility and 

 the ion content of the air is extremely interesting, if 

 real. 



A dimunition in the ion content of the air with an in- 

 crease in mobility, due to the recombinationof oppositely 

 charged ions, is to be anticipated. The ion content, be- 

 cause of this effect, should vary inversely as the square 

 root of the mobility, the variation beginning with the 

 lowest mobility and continuing throughout the entire mo- 

 bility range. The observed variation differs from this 

 in two respects. In the first place no dimunition occurs 

 at the lower mobilities. In the second place the varia- 

 tion which occurs at the higher mobilities is inversely 

 proportional to the first power of the mobility. This 

 lack of agreement, therefore, between the observed and 

 the expected variation in ion content with mobility is 

 sufficient to cast doubt on the reality of the computed 

 mobilities. It seems necessary, therefore, to consider 

 the possibility that the scatter in values of the computed 

 mobilities is due to errors in making the ion content and 

 the conductivity measurements. To explain the results 

 on this basis, it is necessary to assume that, on the 

 average, both elements when measured incorrectly, 

 were measured too small. The computed mobilities 

 which are less than the mean value are due to the use of 

 conductivity values which are less than the true ones. 

 Likewise, computed mobilities greater than the mean 

 value are due to the use of values of the ion content less 

 than the true values of this element. The error of 

 measurement of each element accordingly is systematic 

 In that it is not equally distributed plus and minus. The 

 errors, however, fit extremely well a normal frequency 

 curve and consequently suggest that they are of an acci- 

 dental character. PYom the data a precision index has 

 been computed and also the probable error of the arith- 

 metical mean. A list of the probable errors as deter- 

 mined is given in the table below. The errors do not ap- 

 ply exclusively to either the ion content or the conduc- 

 tivity measurement Init are a compromise of the two. 

 From an examination of the frequency curves, however, 

 it is seen that there is but little difference in the proba- 

 ble error for the two elements in question. 



It is of considerable interest to compare the proba- 

 ble error in the case of the conductivity measurement 

 after the recording of this element began, with the probable 

 error indicated above for data obtained by eye readings. 

 The number of observations secured after the recording 

 began is not large and consequently too much reliance 

 must not be placed on the results obtained. The fre- 

 quency curve, combining the observations on both signs 



Table 1. Values of mean mobility of small ions for 

 cruises IV, VI, and Vn, 1915 to 1929 



Cruise 



Sign 



No. of 

 obsn's. 



Mean 

 mobility 



Probable 

 error of 

 the mean 



of ions is given in figure 13. In this figure also are given 

 plots of the conductivity against mobility and of the ion 

 content against mobility. The data for these plots were 

 obtained by grouping the computed mobilities in the 

 manner explained earlier. The most noteworthy feature 

 of this plot is the fact that there is no suggestion of a 

 change in conductivity with computed mobility. This 

 result, though based on few data, suggests that the sys- 

 tematic error has been largely eliminated from the 

 measurements obtained with the conductivity recorder. 

 The systematic error, in other words, appears to have 

 been introduced through the eye readings. The particu- 

 lar manner in which the errors were introduced is not 

 immediately apparent. 



The elimination of the systematic errors in the case 

 of the conductivity recorder data increases the value of 

 the m.ean mobility, because the lower values of the com- 

 puted mobilities are no longer obtained. Since the sys- 

 tematic errors are still retained in the case of the ion- 

 content measurements, the higher mobilities also are 

 retained. The mean mobility, in consequence, is some- 

 what enhanced as seen in table 2. The question arises, 

 then, as to how much value can be ascribed to a mean 

 where systematic errors have affected the data, although 

 in the case of the Carnegie instruments the errors in the 

 conductivity apparatus about counterbalance those of 

 the ion counter. In view of the occurrence of systematic 

 errors in the measurement of ion content and the conduc- 

 tivity during all cruises of the Carnepie . one should be 

 prepared to accept the differences shown in table 1 in the 

 mean mobilities from cruise to cruise. The small differ- 

 ences between the positive and negative mobilities for 

 each cruise shown in table 1, as well as the magnitudes 

 of the mobilities themselves, are in good agreement with 

 those to be expected from laboratory measurements on 

 moist air. 



Table 2. Increase in mean mobility with change from 

 eye-reading to recording conductivity apparatus 



Apparatus 



Period of use 



No. of 

 obsn's. 



Mobility 



Eye reading 

 Recording 



May 10, 1928- 

 July 28, 1929 



Sep. 5- 

 Nov. 18, 1929 



600 



357 

 46J 

 49b 



k+ 1.30 



k- 1.39 



k+ 1.49 



k- 1.56 



(a) Nov.4-5, (b) Oct. 21-22; 24 abnormal valuesomitted 



I 



