490 



NATURE 



[September 22, 1892 



formed by rarefaction, and consequent refrigeration, of the 

 metallic gases constituting the stratum in which the cyclone 

 exists. He argues that it is farmed within the mass of cooled 

 hydrogen drawn from the chromosphere into the vortex of the 

 cyclone. Speaking of the cyclones, he says : — ' Dans leur em- 

 bouchure evasee ils entraineront I'hydrogene froid de la chrom )- 

 sphere, produisant partout sur leur trajet vertical un abaissement 

 notable de temperature et une obscurite relative, due a I'opacite 

 de I'hydrogene froid englouti ' {Revue Scientifique, March 24, 

 1883). Considering the intense cold required to reduce hydro- 

 gen to the * critical point,' it is a strong supposition that the 

 motion given to it by fluid friction on entering the vortex of the 

 cyclone, can produce a rotation, rarefaction, and cooling, great 

 enough to produce precipitation in a region so intensely heated." 

 — {Esmys, 1891 Edition, vol. i., pp. 188-9.) 

 Churchfield, Edgbaston. F. Howard Collins, 



Direct Determination of the Gravitative Constant by 

 Means of a Tuning-fork. A Lecture-Experiment. 



The following direct experiment for finding the value of the 

 constant g has proved an instructive one for us2 with students 

 beginning dynamics, and combines extreme simplicity with 

 greater accuracy than might be anticipated. 



A rectangular strip of thick plate-glass with one face lightly 

 smoked is dropped past the end of a sounding tuning-fork of 

 known pitch, and which, by means of a light attached style, 

 traces on the smoked surface a fine rippling line whose undula- 

 tions give a complete record of the relative; motion. From 

 measurements of such a trace the value of ^ can be determined 

 immediately with an error of not more than \ per cent. 



For let /i and /g be the distances fallen through in two equal 



consecutive intervals of time (/). Then J and ^ are the veloci- 



h -A 



ties at the middles of these two intervals, and ^ 



fore the velocity gained in time /, and 



there- 



the accelera- 



With a fork giving 384 complete oscillations per second it 

 was found convenient to take for t the time of 30 oscillations ; 

 /j is then the length of any 30 consecutive waves and 4 that of 

 the next 30. These lengths were measured by means of a milli- 

 metre scale printed on card and held against the trace, tenths of 

 a millimetre being estimated. The value of the difference 

 (4 - 4) was thus determined from several measures made in 

 different parts of the trace, and, after some preliminary trials, it 

 was found that such measures seldom differed by more than \ 

 per cent, from their mean, and that the means of different traces 

 agreed about equally well among themselves. Under the given 

 conditions (4 - Z^) is just under 6 centimetres. The experiment 

 takes only a moment to perform, and the plate can be at once 

 exhibited as a lantern slide. 



In order to obtain good traces a little care must be exercised. 

 The smoking should be very light. A fine bristle from a clothes- 

 or hearth-brush, 2 to 4 cm. long, stuck on with a scrap of wax, 

 may be used as a style, and it should be inclined downwards so 

 as to make an angle of 45° or less with the vertical face of the 

 plate and project well under the plate before this is let fall, so as 

 to be considerably bent while tracing. By furnishing each prong 

 with such a bristle two simultaneous tracings are obtained. 

 Although the method is independent of the actual velocity with 

 which the plate reaches the style, yet it is best to let the plate 

 fall from quite close above the end of the style (within, say, 

 I cm.), so that as many wave-lengths as possible may be marked 

 on the plate. The fork also should be strongly bowed with a 

 violin bow, so as to give sharply accentuated ripples, the posi- 

 tions of whose crests are defined with greater precision than 

 would be those of gentler undulations. The plate itself can be 

 conveniently let go if the upper part of its suspension is a single 

 string with a knot at the top, and to prevent its swinging in the 

 air or turning as it descends, it may be held against a narrow 

 smooth backing of hard wood. Without these precautions the 

 trace is liable to show curvature and other irregularities, and 

 indeed under any circumstances the first one or two undulations 

 traced near the advancing edge of the plate are liable to be irre- 

 gular. The more massive the plate the less is its motion affected 

 by the pressure of the tracing style. 



NO. 1195, VOL. 46] 



Although as a means of finding the value of g such a method! 

 does not compare for accuracy with the use of a pendulum, yet 

 for the converse process of determining the pitch of a fork from 

 measures of its trace and the known value of ^, it may be of 

 utility ; for, since the length (4 — /j) is proportional to the square 

 of the vibration-number, the percentage error will now be halved 

 or reduced to about i in 400, and I have little doubt that a 

 careful experimenter, by attending to the causes of error, might 

 further improve on this. A. M. WORTHINGTON. 



R.N. E. College, Devonport, September 12. 



A Meteor. 



On Wednesday, September 14, at 7h. 9m. p.m. a large meteor 

 was seen by about twenty people, including myself, who were 

 driving from Penmaenmawr to Conway. It was first observed in 

 the south-east just above the Conway mountain. It was visible for 

 about 30°, fell very slowly in a wavy line inclined at a small 

 angle to the horizon, disappearing behind the mountain. It 

 seemed to be very near the ground as it passed over the 

 mountain. 



The sky was quite bright, so that only Mars was clearly visible 

 in it. The meteor appeared to the eye about the size and bright- 

 ness of Jupiter at the present time, and was of a slightly bluer 

 tint than that planet. There was no perceptible variation in 

 size and brilliance while the meteor was in sight. 



September 19. Grace E. Chisholm. 



Crater-like Depression in Glaciers. 



A PROPOS de la cavite du glacier de Tete Rousse que M. Val- 

 lot et moi avons decouverte et dont vous parlez dans Nature. 

 (September i), M. R. voa Lendenfeld vous ecrit (Nature, 

 September 15) qu'il a trouve des depressions crateriformes sur 

 le glacier de Tasman, dans la Nouvelle Zelande. Permettez-moi 

 de vous signaler que de pareilles depressions existent sur cer- 

 tains glaciers des Alpes et notamment sur le glacier de Corner, 

 ou la carte Suisse en indique 26. Elles sont en general a peu 

 pres circulaires ; leur plus grande dimension horizontale at- 

 teipt parfois 130 metres et leur profondeur 30 metres. L'incli- 

 naison de leurs parois varie en general de 45° a la verticale. 

 Elles re9oivent souvent de I'eau qui s'engouffre au fond dans un 

 moulin ou qui s'ecoule, par une crevasse, dans une depression 

 voisine. Au mois d'aout dernier, I'une d'elles formait un 

 veritable petit lac glaciaire que j'ai sonde avec M. Etienne 

 Ritter au moyen d'un bateau demontable ; la profondeur de 

 I'eau etait presque partout de 5 a 6 metres, sauf dans un trou, 

 vraisemblablement un moulin, ou ma sonde est descendue jusqu'a 

 21 metres. II est probable que, lorsque la pression de I'eau 

 aura elargi le moulin par oil elle s'ecoule, la cavite se videra. 



Les depressions ne me paraissent avoir aucune analogic avec 

 la cavite quej'aivue a Tete Rousse. Leur origine est assez mal 

 connue (voir Heim, " Gletscherkunde," p. 246) ; il est possible, 

 comme le pensait primitivement votre honorable correspon- 

 dant, qu'elles soient d'anciens moalins transformer. 



J'en ai vu une egalement sur la Mer de Glace, entre le Mon- 

 tanvers et le Tacul. 



L'etude de ces depressions, encore tres incomplete, serait tres 

 interessante, et je les signale a I'attention de ceuxqui parcourent 

 les glaciers. 



Veuillez agreer, monsieur, mes civilites empressees. 



Thonon, le 17 Septembre. Andre Delebecque. 



GENERALIZATION OF '' MERCATOR'S" 



PROJECTION PERFORMED BY AID OF 



ELECTRICAL INSTRUMENTS. 



THE following mode of generalizing Mercator's Pro- 

 jection is merely an illustration of a communication 

 to Section A of the British Association at its recent meet- 

 ing in Edinburgh, entitled " Reduction of every Problem 

 of Two Freedoms in Conservative Dynamics to the 

 Drawing of Geodetic Lines on a Surface of given Specific 

 Curvature." An abstract of this paper appeared in 

 Nature for August 18. 



In 1568, Gerhard Kramer, commonly known as " Mer- 

 cator " (the Latin of his surname), gave to the world 



