Discussion of Atmospheric-Electric Observations 



417 



one year, it is hardly possible to look upon the curves as comparable with mean diurnal- 

 variation curves for the year, as obtained at one station, for the present observations 

 correspond to all sorts of different latitudes. 



The curves for X, n+, and R are shown in Figure 26. It naturally turned out that 

 there were many occasions on which observations were started and later in the day it was 

 found necessary to discontinue them as a result of weather conditions. All of the observa- 

 tions have, however, been used in drawing the curves, so that the parts of the curves 

 corresponding to the night depend upon fewer observations than those for the day. Each 

 point on the curves is the representative of observations on a large number of days, and 

 the number of days for the individual points are recorded against them. It is interesting 

 to observe that the parts of the curves determined by points representative of the observa- 



Midnight 2 4 B 8 10 12 U 16 1» 



Hours 



Fig. 26.— Ocean Diurnal-Variation Curves for Ionic Content, Potential-Gradient, and Penetrating Radiation. 



tions on many days are much more definite in form than the parts determined by points 

 corresponding to few observations, a result, of course, to be expected. Happily the 

 doubtful parts of the curves are of sufficiently small extent to render it possible to draw 

 them in with reasonable certainty by noting that, except for the slow progression of the 

 annual variation, the curves must repeat themselves with a period of 24 hours. The curves 

 for X and n+ show definite diurnal variations, but the curve for R does not indicate any such 

 tendency, and this point is emphasized by the fact that the portions of the curve which are 

 fixed by the largest numbers of observations are just those which approximate most closely 

 to the representation of R as independent of the time of day. This result, which is in har- 

 mony with the observations of Sunpson and Wright in the Atlantic and Indian Oceans,* is 

 not necessarily inconsistent with the results of those observers who, on land, have found 

 diurnal variations m R, since, as ah-eady remarked, a large portion of R as measured on 

 land is determined by the radioactive material in the soil and atmosphere. 



Turning now to the potential-gradient curve, it is interesting to compare this curve 

 with the mean diurnal-variation curves for the year for several land stations. This is 

 done in Figure 27. It will be observed that the chief f eature which characterizes the ocean 



^Proc. R. Soc. A, vol. 85, p. 199, 1911. 



