January i6, 1902] 



NA TURE 



247 



Thus at Karang Sago, where was situated the Putch eclipse 

 party. Dr. \V. van Bemmelen, assistant director of the Batavia 

 Magnetic Observatory, observed the changes in the magnetic 

 declination and in horizontal intensity, and he reports the oc- 

 currence of "an extremely interesting magnetic effect." He 

 has courteously sent me an extract of his observations, made 

 during several days before and on the day of the eclipse, and 

 there certainly appears evidence of a magnetic effect in both 

 elements different from that observed on the days prior to the 

 eclipse. 



At Sawah Loento, the site of the Massachusetts Institute of 

 Technology party of Boston, the variations in magnetic declina- 

 tion were observed by Mr; G. L. Hosmer on May 17 and 18. 

 Comparing the two days' results for the interval of the eclipse, 

 there is indisputable proof that something different occurred on 

 the day of the eclipse than on the day before. Namely, at 

 this station, situated so close to the magnetic equator, the range 

 of the diurnal variation of the magnetic declination is about one 

 minute of arc. The magnetic effect during the time of the 

 eclipse was of about the same amount, so that a steady decrease 

 of east declination resulted during the time of day when, normally, 

 there is a steady increase. 



There was but one magnetic observatory directly within the 

 belt, viz. the one at Mauritius, and this was situated not far 

 from the place of beginning of the eclipse. No special mag- 

 netic observations were made at this place, however, but regular 

 photographic curves giving the variations in the magnetic 

 elements were obtained. The declination and the vertical intensity 

 curves apparently do not show any disturbance that could 

 easily be picked out and referred to the eclipse. Regarding 

 the horizontal intensity curve — the more sensitive one — Mr. 

 Claxton states " that the original curve shows slight tremors 

 between 7.15 and 7.50, and occasionally between 8.5 and 

 9.0 a.m." I have plotted this intensity curve on a larger 

 scale, and find that the curve shows no very marked disturbance 

 that might be readily referred to the eclipse, with the exception 

 of one producing an easily perceptible bulge in the curve 

 amounting to about 3-4 units in the fifth decimal C.G.S. units 

 and lasting about 30 minutes. Anyway the effect, if there be 

 one, is very minute, and will not be so readily separated from 

 the usual diurnal variation as in the case of the two previous 

 stations. Whether this is due to the fact that owing to the 

 vicinity of Mauritius to the beginning of the eclipse the minute 

 eclipse magnetic storm did not have time to develop itself or 

 was just in the embryonic state cannot be said. 



The magnetic effect observed at Karang Sago and at Sawah 

 Loento does not appear to have extended very far outside the 

 belt of totality, it being scarcely appreciable at the Batavia mag- 

 netic observatory. 



My grateful and appreciative acknowledgments are due to 

 all who have participated in this interesting investigation — one, 

 to my mind, of fundamental importance to the theory of the 

 diurnal variation of the earth's magnetism as elaborated by 

 Schuster and von Bezold. L. A. B.\uer. 



U.S. Coast and Geodetic Survey, Washington, D.C. 

 December 30, 1901. 



The Roots of the Equation ;( = tan u. 



I.N many treatises on optics it is stated that the roots of the 

 equation, K = tan u, were calculated by Schwerd. Verdet 

 ( " Oeuvres," t. v. p. 265), says : — " These roots have been cal- 

 culated by Schwerd, who arrived at the following values : «,,%= 

 I '4303, «o/»= 2"4590, ..." up to «j/ir. Preston (" Light," 

 p. 255, second edition) says : — "The values of /(corresponding 

 to the maximum values of the illumination have been given by 

 Schwerd as follows : — " The values given are precisely the 

 same as Verdet's. Rayleigh (" Encyc. Brit.," vol. xxiv. p. 430, 

 art. , Wave Theory) gives a method for calculating the roots of 

 the equation, and remarks that they were obtained in another 

 manner by Schwerd. (There is a misprint in Rayleigh's value 

 for u^). Other references might be added. 



Will someone kindly indicate where Schwerd gives the results 

 referred to ? 



In his "Beugungserscheinungen" he shows that the roots of the 

 equation are approximately the values of (27z-l- l)ir/2, obtained 

 by giving integral values to n ; and he remarks (in § 63, p. 

 28) that for values « = i, and « = 2, the true values of » differ 



by I2^'5 and 7''5 from 3ir/2 and 5ir/2, respectively. In table 

 i. , at the end of the book, he gives values of the expression for 

 the intensity, sin'-«/;«-, for values of u increasing by 15° ; and 

 at the foot of the table he states that the first and second 

 maximaareat257°'5 (/.e;. 270°- I2°'5)and442°'5(«'.^.450°- 7°-5). 

 Further details I have not found. 



It may be remarked that the roots of the equation under con- 

 sideration were given long before Schwerd's time. In Euler's 

 " Introductio in Analysin Infinitorum " (Lausanne 174S) the 

 question is fully discussed. See t. ii., cap. xx., prob. ix. 



C. A. Ch.\nt. 



University of Toronto, December 27, 1901. 



A Model Geyser. 



If the following working model, which was the outcome of a 

 good many trials, is in any way original it may prove of interest, 

 as it works automatically and with excellent effect as a geyser of 

 regular period, suitable for a lecture table. The figure needs 

 but little description. A small aspirator with a bent glass tube 

 exit acts as an intermittent syphon. The water is discharged 

 into a half-inch iron pipe, the long horizontal limb of which 

 measures some 13 centimetres. The glass syphon tube slips 

 through a rubber ring at the top of the pipe (gauge fitting), or a 

 cork would doubtless answer the purpose. The lower closed end 

 of the tube is heated by the equivalent of about four ordinary 

 Bunsen burners, and should be placed as shown, as if placed 



under the exit, steam is generated too fast and the water may 

 be blown back into the aspirator. Water drips into the aspirator 

 at such a rate that the syphon discharges about 300 c. c. of water 

 once in every ten minutes. A jet of steam some six feet high 

 and water some two feet high results, with many appropriate 

 gurglings. The diameter of the exit is about 6 millimetres. 

 Of course the apparatus is concealed ; a large circular tin can- 

 ister to which the iron pipe is screwed forms a good foundation 

 and serves to keep the water off the burners. Furnace clinker, 

 which is not wholly unsuitable, forms a readily obtainable 

 material for completing the external features of the geyser. 

 Felsted, January. A. E. Munby. 



Birds Capturing Butterflies in Flight. 



With reference to Mr. Latter's letter in Nature of 

 November 16, 1899 (vol. Ixi. p. 55), which has been brought to 

 my notice, I would say that the supposition that birds do not 

 attack butterflies in flight is not strictly correct. 



The common King Crow {Dicriiriis ater, I believe) invariably 

 captures butterflies on the wing ; I have seen these birds scores 

 of times do this. Their usual prey seems to be a small deep 

 yellow butterfly with black on the tip of the wings, but I have 

 occasionally seen other butterflies so captured by them. 



India, December iS, 1901. A. E. McK.^v. 



NO. 1 68 I, VOL. 65] 



