August 5, 1897] 



NATURE 



321 



It was natural to suppose that the iron vapour pro- 

 ducing the cooler lines was higher up than that respon- 

 sible for the enhanced line at 4924"i. Hence a crucial 

 observation was planned for the eclipse of 1882. If the 

 vapour were higher it should be dimmer, and its lines 

 should, if seen at all, be seen long and faint very near 

 the beginning of totality, while the hotter line, being 

 produced by vapours relatively low and at a higher 

 temperature, should be seen short and bright some time 

 before the beginning of totality. 



Fig. 12 will show how absolutely the prediction was 

 verified by the event. 



J. Norman Lockver. 

 {To be continued.) 



ON LUNAR AND SOLAR PERIODICITIES 

 OF EARTHQUAKES. 

 'T^HE investigation of small periodical changes is 

 -*■ rendered difficult chiefly by the doubt which so 

 often exists, whether the results obtained by the ordinary 

 methods are due to accident or prove some real 

 periodically acting cause. Attention need only be 

 drawn to the many calculations which have been made 

 to trace the sunspot period in terrestrial phenomena, 

 such as rainfall or temperature, to show that widely 

 different conclusions may be drawn from the same 

 evidence according as greater or smaller value is attached 

 to the element of chance. I have been engaged for 

 some time to apply the theory of probability to investi- 

 gations of this nature, with a view if possible to being 

 able in every instance to assign a definite number to the 

 probability that anv periodicity which may be found in 

 the record of some physical phenomenon is of an 

 accidental nature. In a paper recently communicated 

 to the Royal Society, I have applied the results obtained 

 to the periodicities of earthquakes. Mr. Knott {Proc. 

 Roy. Soc, vol. Ix. p. 457) has recently published some 

 investigations, conducted with skill and labour, which in 

 his judgment were favourable to the existence of a true 

 lunar influence on earthquakes. The theory of prob- 

 ability, however, does not support that view. The 

 number of earthquakes treated by Mr. Knott is 7427, 

 and Fourier's series is applied to determine the ampli- 

 tudes of the changes which have periods coincident with 

 the lunar day, the half-day and the third or fourth part 

 of a lunar day. The method employed would always 

 give some results whether a true periodicity existed or 

 not, and I have calculated what the average amplitudes 

 would be if earthquakes were distributed quite at 

 random. These amplitudes depend, of course, on the 

 number of events taken into the calculation, and are 

 found to vary inversely as the square root of that 

 number. The following table will show how the ampli- 

 tudes found by Mr. Knott compare with those calculated 

 by the theory of probability. Ci in the table refers to 

 the lunar day, while Cg, C3 and C4 refer to its sub- 

 multiples. 



Coefficients Ci Cj Q C4 



Expectancy for the coefficients 19*3 157 106 5*2 

 By the theory of probability... io'3 179 109 3*97 



As it may further be shown that cases will frequently 

 occur where the amplitudes found are equal to twice the 

 expectancy, the table may be considered as conclusive 

 that if a lunar effect exists, it must be so small that it 

 is quite hidden by accidental effects. For the present, 

 at any rate, the evidence is against such a lunar in- 

 fluence. A discussion of the periods, coincident with 

 the lunar months, leads to the same conclusion. 



It is otherwise with the annual and daily periods, 

 which have recently been discussed by Mr. Davison. 

 Here the amplitudes found are decidedly too large to 

 NO. 1449, VOL. 56] 



be due to accident ; and we may therefore say, with a 

 degree of probability amounting practically to certainty, 

 that there is a yearly period giving a maximum of 

 earthquakes in December, and a daily period giving a 

 maximum some time between ten o'clock in the morning 

 and noon. Arthur Schuster. 



NOTES. 

 A SMALL committee has been appointed by the Treasury "to 

 consider and report upon the desirability of establishing a 

 National Physical Laboratory for the testing and verification of 

 instruments for physical investigation, for the construction and 

 preservation of standards of measurement, and for the systematic 

 determination of physical constants and numerical data useful for 

 scientific and industrial purposes, and to report whether the work 

 of such an institution, if established, could be associated with 

 any testing or standardising work already performed wholly or 

 partly at the public cost." The following will be the members 

 of the committee :— The Lord Rayleigh, F.R S. (chairman). 

 Sir Courtenay Boyle, K.C.B., Sir Andrew Noble, KC.B., 

 F.R.S., Sir John Wolfe Barry, K.C.B., F.R.S., Prof. W. C. 

 Roberts- Austen, C.B., F.R.S., Mr. Robert Chalmers, of the 

 Treasury, Prof. A. W. Riicker, F.R.S., Mr. Alexander Siemens, 

 Dr. T. E. Thorpe, F.R.S. 



At a meeting of the Royal College of Physicians of London, 

 on Thursday last, the College, on the recommendation of the 

 Council, awarded the Moxon gold medal to Sir Samuel Wilks, 

 F.R.S., the President of the College, for having especially dis- 

 tinguished himself by observation and research in clinical 

 medicine, and the Baly medal to Prof. Schiifer, F.R.S. , for 

 having especially distinguished himself in the science of physi- 

 ology. The Harveian oration will be delivered by Sir William 

 Roberts on October 18, St. Luke's Day. Dr. E. Markham 

 Skerritt, of Bristol, will give the Bradshaw lecture on November 

 4 ; and Dr. F. Pavy will deliver a special lecture, supplementary 

 to the Croonian lectures, delivered in 1894, on November 11. 

 The following were announced as lecturers for next year :— 

 Goulstonian lectures. Dr. John Rose Bradford ; Lumleian 

 lectures. Sir Richard Douglas-Powell; Croonian lectures, Dr. 

 Sidney Martin. 



A Parliamentary paper has just been issued giving an 

 additional Civil Service Estimate, amounting to 10,000/., for 

 Art and Science buildings of Great Britain. The total original 

 net estimate for 1897-98 was 26,000/., and this has been in- 

 creased to 36,000/. The British Museum and the Science and 

 Art Department buildings receive 5000/. each for new works, 

 alterations, and additions. The increased grant to the Science 

 and Art Department is on account of the cost of carrying out 

 certain urgent works and services at South Kensington, designed 

 to give effect to the recommendations made in the first report 

 from the Select Committee on the Museums of the Department 

 It is proposed to remove the more dangerous buildings on the 

 east side of Exhibition Road (including the " boilers " and the 

 electric lighting plant), to displace the occup.ants of the official 

 residences, and reconstruct the entrance to the galleries on the 

 west side of Exhibition Road. 



The Weights and Measures Bill (Metric System), the scope 

 of which has already been described (p. 275), was read a third 

 time and passed in the House of Lords on Friday last. 



Wk notice with regret that the Hon. Ralph Abercromby, who 

 did so much for the advancement of meteorolqjical science, died 

 at Sydney, New South Wales, on June 21, at fifty-four years 

 of age. 



