June 13, 1901] 



NATURE 



153 



several times ; there are also monthly determinations 

 of the curve scale values. The necessary temperature 

 corrections are based on direct experiments on one 

 occasion when the magnetograph room was artificially 

 heated. Hourly measure ments are made of all the 

 magnetic curves and the results appear in tables, one 

 for each element for each month. In addition to the 

 hourly readings, each table gives the daily mean, 

 maximum, minimum and " range" (maximum less mini- 

 mum) ; it gives also the monthly mean for each hour 

 of the day, and the means for the month of the diurnal 

 maxima, minima and "ranges." The general character 

 of each day, whether quiet or more or less disturbed, is 

 also noted. The monthly means are summarised on 

 p. 59, and there are curves of diurnal variation for each 

 month and for the year. 



The mean monthly diurnal variations are also analysed 

 in Fourier's series. The fact should be noted that the 

 " mean range " of any element for each month being the 

 mean of the differences between the daily maxima and 

 minima, irrespective of the times of their occurrence, is 

 necessarily larger than the range shown m the mean 

 monthly diurnal variation ; it may be largely influenced 

 by the occurrence of magnetic disturbances. For instance, 

 the "mean range" in the horizontal force is given as 

 greater in December than in any other month except 

 .\pril, but the value in December is considerably affected 

 by the occurrence of two or three exceptionally large 

 "ranges." -The mean monthly diurnal variations are less 

 exposed to accidental influences ; their nature is most 

 easily followed in the curves. The amplitude of the 

 regular diurnal variation in both declination and hori- 

 zontal force appears least in November. The amplitudes 

 in January appear surprisingly large compared to those 

 in the last three months of the year. 



The atmospheric electricity installation consists of a 

 Kelvin water-dropper, the discharge tube of which projects 

 "about 2 metres" at a height of 17 metre above the 

 ground. This is connected to the needle of a quadrant 

 electrometer, the quadrants of which are connected the one 

 pair to the positive the other pair to the negative pole of 

 a battery of water cells, the centre of which is to earth. 

 This seems the same arrangement as at Kew. The scale 

 value of the curves is determined weekly. The hourly 

 readings are recorded in tables, one for each month. The 

 daily means, maxima, minima and " ranges " are recorded 

 as in the magnetic tables, also the nature of the daily 

 weather. Hourly means are also given for each month, 

 but in forming these a considerable number of individual 

 results are omitted as being abnormal. Amongst the 

 values omitted are most of the negative potentials, and 

 some entire days are excluded on which negative read- 

 ings were numerous. The measurements of potential 

 being all given in volts, one can follow readily the annual 

 change, which is more than usually pronounced. Thus 

 the mean potential for the year being 47'2 volts, the mean 

 voltage for the three months December to February was 

 93'2, while that for the three months July to September 

 was only 9"o ; the maximum monthly mean was ii2'8 in 

 December, the minimum 6'9 in August. It may be added 

 that the e,\ceptionally low value in August appears in no 

 way due to exceptional occurrences of negative potential. 

 NO. 1650, VOL. 64] 



The mean diurnal variations for the several months 

 are illustrated by curves. As usual there is a marked 

 tendency to a double diurnal variation ; but the principal 

 maximum occurs between 6 and 8 a.m., instead of, as is- 

 customary, in the late evening. Further, the principal 

 minimum is found, the whole year round, in the early- 

 afternoon, usually from 2 to 3 p.m. These results possess 

 a special interest from their apparent irreconcilability 

 with views due to Chauveau which have recently met 

 with considerable recognition. 



Another notable feature is the size of the mean diurnal 

 variation. The largest mean hourly value of the poten- 

 tial is in most months some four or five times the smallest ; 

 for instance, amongst the mean hourly values in Decem- 

 ber the maximum was 1876, the minimum 42'i, while 

 the corresponding values in August were 137 and 2"8. 

 The peculiarities shown by the atmospheric electricity 

 results at Tokio are, to a considerable extent, manifested 

 by observations made during a series of years in the 

 Batavia Observatory, but though Batavia is much nearer 

 to the equator than Tokio, the results at the latter station 

 show the greater departure from those ordinarily recorded 

 in Europe. 



What has been said will suffice to show the interest of 

 the volume, and the evidence it affords of the progressive 

 spirit in Japanese science. The continuance of the ob- 

 servations, and also of the practice of describing them in 

 English (sometimes with Japanese equivalent), is much 

 to be desired. There are, however, one or two points 

 where some friendly criticism may be offered. The 

 variation in the scale value of the horizontal force curves 

 — from io~'^x69to io"'''X57 C.G.S. units for i mm. — is 

 excessive ; and the vertical force scale values show even 

 larger variations (i mm. =io"'^x82 in April, io'"^X2_39 

 in June, and 10"'' x 79 in November). Changes such as 

 these, unless produced designedly at known times, intro- 

 duce uncertainties into at least the annual variation in 

 the amplitudes of the diurnal inequalities. Again, the 

 temperature coefficients are so large for both the 

 horizontal and vertical force magnetographs that ap- 

 preciable uncertainty must be introduced into the diurnal 

 variations unless the changes of temperature are 

 known with extreme accuracy. Under such condi- 

 tions, the employment of six significant figures in 

 the tables of monthly means of the diurnal varia- 

 tions of the force components seems hardly well 

 advised. 



In the case of atmospheric electricity, i mm. of curve 

 ordinate answered, on the average, to 386 volts in Feb- 

 ruary, I '1 5 in August and 2'69 in December. If, as 

 one would rather infer from the preface, the number 

 of battery cells in use varied from 30 to 50, one would 

 not be surprised at a considerable change in the scale 

 value, and it would be only good policy to have the 

 scale contracted in winter when the mean potential is 

 large. The changes recorded in the scale value seem, 

 however, too large to be wholly accounted for in this 

 way, and they do not exhibit so regular a fluctuation 

 as to suggest design. A little further information on 

 these points would enable the critic to pronounce with 

 greater assurance on the value of the results. 



C. Chree. 



