AT KEW OBSEKVATORY, 1898 TO 1912. 143 



derived from 1898 to 1904 and those derived from earlier epochs. As this might arise 

 either from the periods available having been too short to eliminate accidental features 

 or form a real change in the nature of the phenomena, data are given here for the period 

 1905 to 1912 as well as for the whole 15 years. Between 1901'5 and 1909'0, which 

 may be regarded as the epochs to which refer the data in E 15 and those from the years 

 1905 to 1912, there was considerable growth in Western London, and if the difference 

 between EVERETT'S results for 1862-4 and mine for 1898 to 1904 were due to 

 urban extension, then the data for 1905 to 1912 would naturally depart still further 

 from EVERETT'S. Table II. gives the diurnal inequalities for the whole 15 years, and 

 Table III. those for the seven years 1905 to 1912. The highest and lowest hourly 

 values are in heavy type. The hours are really G.M.T., but local time is only 

 1^ minutes after Greenwich. For purposes of comparison data from Table III., p. 306, 

 of Ej must be multiplied by 1'91. 



Besides the diurnal inequalities for the 12 months, Tables II. and III. contain diurnal 

 inequalities for the whole year and for three seasons, winter (November to February), 

 equinox (March, April, September and October), and summer (May to August). The 

 inequalities for the 12 months in Table II. are shown graphically in fig. 1, p. 146. This 

 is immediately comparable with fig. 2 of E 1; provided the line there marked 50 volts be 

 regarded as representing 95. To facilitate recognition of the resemblances and differ- 

 ences between the data from different epochs, fig. 2, p. 147, shows in juxtaposition the 

 diurnal inequalities for the year and the three seasons derived from 1898 to 1904 

 (upper curves), the whole 15 years (central curves), and 1905 to 1912 (lower curves) ; 

 fig. 3, p. 148 contains isopleths for the diurnal variation of the 15 years. For it lam 

 indebted to Mr. E. H. NICHOLS, B.Sc., professional assistant at Kew Observatory. To 

 assist the eye, the isopleth curves are drawn thicker or thinner according as they 

 answer to potentials which are greater or less than the mean for the year, 304 volts 

 per metre. The two broken lines indicate respectively the times of sunrise and sunset 

 throughout the year. 



A careful comparison of Tables II. and III., or of corresponding curves in fig. 1 and 

 in fig. 2 of E,, shows differences between the monthly diurnal inequalities derived 

 from different epochs, but some of these are undoubtedly " accidental." In the seasonal 

 diurnal inequalities in fig. 2, where "accident" is more completely eliminated, the 

 differences that exist except as regards the relative depths of the morning and after- 

 noon minima make small appeal to the eye. It requires an auxiliary such as analysis 

 into Fourier series to show their nature distinctly. It may reasonably be inferred that 

 no large rapid change is in progress in electrical conditions at Kew, and that the 

 inequalities derived from the 15 years give a close approach to normal conditions. 

 Whether the results are sensibly dependent on the particular site selected for the water- 

 dropper it is impossible to say until adequate data are available from a different site. 



A glance at fig. 1 shows a conspicuous difference between midsummer and midwinter. 

 The time elapsing between the forenoon and afternoon maxima and the depth of the 



