Discussion of Atmospheric-Electric Observations 



407 



of the values of f_|_ and !>_ with those obtained from the other passages of the cioiise is evidence 

 in favor of the reliabiUty of the observations. It is further worthy of notice that, of the 

 Atlantic-Ocean values recorded in Table 84, more than half were obtained in the land- 

 locked Caribbean Sea, and the remainder just to the north of the West Indies. The results, 

 therefore, while abnormal as Atlantic-Ocean values, are in harmony with those of the 

 Carnegie's first and thu'd cruises' ua indicating low values for the ionic content and con- 

 ductivity in the regions of transition between sea and land. 



The desirability of basing the potential-gradient reduction factors on several deter- 

 minations made under different conditions has already been emphasized on page 382. It is 

 not often that one can find a location which, from a topographical point of view, is suitable, 

 and on this account the reductions of the potential-gradients to absolute values have been 

 made, thus far, on the basis of only one set of determinations of these factors, made in Colon 

 Harbor, April 2, 1915. Thus the absolute values may be liable to some change as the 

 accumulation of other determinations renders available more reliable values of the reduc- 

 tion factors. There is, however, no reason to believe that any considerable error attaches to 

 the present factors. 



On glancing at Table 84, it appears that, with the exception of the values of the con- 

 ductivity and ionic content for the Atlantic Ocean, there is a much greater uniformity in 

 the values of the various elements at different parts of the globe than is the case with land 

 values. In support of this remark Tables 86-89 and Table 92 are given, showing a 

 collection of land values obtained at different times and in different localities. 



Table 85. — Mean Values of Almospheric-Eleclric Elements Corrected for Diurnal Variations. 



The more complete discussion of the diurnal variation over the ocean is taken up 

 below, but it will be desirable here to anticipate, to some extent, the results for the purpose 

 of reducing the observations to the daily mean values. This reduction has been made in 

 Table 85, in which the quantities have been grouped into two sets — those corresponding 

 to observations on the Pacific Ocean over the course Balboa-Honolulu-Dutch Harbor-Port 

 Lyttelton, and those corresponding to the circumnavigation cruise in the sub-Antarctic 

 Oceans between latitudes 50° and 60° south. For brevity, the former group is designated 

 "Pacific" group, the latter " sub- Antarctic " group. Taking the case of the quantity n+ 

 for the purpose of illustrating the nature of the reduction, the mean value of all the times 

 of determination for the "Pacific" values was 9''6, the diurnal- variation observations 

 themselves being omitted, and the mean value of n_ corresponding to the time 9''6 was 

 845. The daily mean value of n+'^ and the 9^6 value as obtained from the diurnal-varia- 

 tion curve were respectively 810 and 840, so that multiplying 845 by 810/840 we obtain the 

 daily mean value 811 recorded in Table 85. 



This method of procedure depends on the assumption that the diurnal-variation curves 

 are of the same form for the two portions of the cruise referred to, but it appeared on draw- 



'See pages 367 and 375. 



"The mean time throughout the Pacific and sub-Antarctic cruises, of all the determinations of n and X, omitting the 

 diurnal-variation observations themselves, was 9!'7. It is of interest to point out that the values 838 and 123, which corre- 

 spond to the mean 9''7 values of n+ and X for the whole cruise from Balboa onwards are comparatively near to the valuas 

 840 and l.^O as obtained for this time from the diurnal-variation cur\'e8 themselves. 



