558 



NA TURE 



[Oct. 14, I i 



severally twisted in opposite directions. A straight thick wire 

 passed through the riiiij, the weight of which afforded a ready 

 means of varying the force necessary to balance the torsion 

 of the wire. My first oliject was to prove that the force of 

 the latter was, at any rate to some considerable extent, inde- 

 pendent of the tension. Suppoe that with this arrangement, 

 the wire being horizontal, a balance has been effected when the 

 ring has been turned about the wire as an axis three or four 

 times. What will happen when the wire is further strained? I 

 think it would be a natural expectation (apart from special 

 knowledge) that the weight will rise ; on the other hand, a 

 knowledge of the law of torsion teaches (?) that there is no increase 

 of the force sustaining the weight, which therefore will not rise. 

 But who would suppose that, on the contrary, it would sink? 

 Such, nevertheless, is what takes place. I continued increasing 

 the s'.rain, and the weight continued to sink. I had to go on 

 lessening the weight again and again (by shifting the balancing 

 cross-rod), in order to restore the horizontality of the ring ; until 

 at last there was scarcely any force of torsion left ! To repeat the 

 experiment of course the ring had to receive three or four fresh 

 turns, I did so several ti nes, always continuing, as I thought, 

 to increase the strain. All the time the wire was absorbing the 

 torsion, and did not break. I then thought to try the effect of a 

 high initial torsion. But I did not seem to gel any such by 

 turning the ring more than five or six times. I then thought to 

 see h jw much twisting the vvire would bear. Expecting it 

 every instant to break, I counted up to 100 half turns. By this 

 time the -mre was quite slack! I added another hundred half 

 turns. The wire was now half an inch longer, without any strain 

 having been kept on it except just enough to keep it straight. I 

 went on twisting. At 218 one wire broke. The other then had 

 only sixteen hall-turns of twist in it, out of the 230 or more 

 received. I afterwards went on twisting, mending each time 

 that the wire broke, till the twist (quite visible under the micro- 

 scope) amounted to sixteen turns per inch. The length kept on 

 increasing. After breaking, the wire always untwisted one turn 

 in four inches. 



I feel myself here in presence of laws of which I know 

 nothing ; and my object in writing this short experience is to 

 ascertain whether it is sufficiently in accord with what is known 

 to cause nj surprise to any one but myself. In that case I 

 shall be greatly obliged to any one who will tell me where I 

 can learn all about it. J. Herschel 



CoUingwood, October 4 



I forgot to say that in no case did slackening of the strain 

 reverse the sinking of the weight due to increa-e of strain. 



The Magnetic Storm 



By the mail just arrived from Australia I have received copies 

 of the photographic traces produced by the declination magneto- 

 graph at the Mell)ourne Observatory during the magnetic storm 

 of August 12 to 14, kindly forwarded by Mr. Ellery, the Govern- 

 ment astronomer there. 



A comparison of these curves with those from the Kew 

 instrument for the same period shows that the disturbance 

 conmienced and ended at both places at the same time. 



It is not easy however to trace much similarity in the two sets 

 of curves, as the individual excursions of the magnet east and 

 west of the normal position which form the record of the mag- 

 netic storm, cannot be at all times followed in both curves, but 

 the periods of greater disturbance seem to have been simulta- 

 neous. For example, the commencement of the disturbance 

 was well marked at August lid. 8h. lom. p.m. at Melbourne, 

 which corresponds to iid. loh. Sjm. a.m. G.M.T., whilst here 

 (z'/rfi^Mr. Ellis's letter in Nature, vol. xxii. p. 361) it commenced 

 at loh. 30m. a.m. ; then again the large deviation to the eastward 

 noted in the Rev. S. J. Perry's letter in Nature, which 

 occurred here between I2d. lib. 30m. a.m. and I2h. 30U1. 

 p.m., seems to have had its effect, as a movement of the needle 

 at Melbourne to the westward between I2d. gh. iSm. p.m. and 

 loh. 30m. p.m. The maxi]num delleclion which exceeded the 

 limits of registration of the instrument, I estimate to have taken 

 place at 10 p.m. The corresponding G.M. times for the above 

 are I2d. iih. 38™. a.m., I2h. S3m. p.m., and I2h. 23m. p.m. ; 

 the maximum deflection recorded here seems to have been at 

 I2h. 25ra. p.m. 



The disturbed period may be considered to have died out at 

 Kew at i4d. 8h. a.m. G.M.T., and at Melbourne at about 



I4d. 7h. a.m., but there is no very distinctive movement which 

 would enable us to fix this limit with accuracy. 



These interesting comparisons are extremely satisfactory, for 

 it is but recently that the Government of Victoria was consider- 

 ing the advisability of discontinuing the system of photographic 

 registration of the magnetometers at Melbourne, and consulted 

 the Kew Committee upon the subject. 



A circular was accordingly issued to the leading physicists of 

 Europe, and their replies being almost unanimously in favour of 

 the continuance of the recording system, the Government erected 

 a new magnetic observatory, and decided upon carrying on the 

 work. 



Mr. Ellery has also forwarded a month's curi'cs for the purpose 

 of assisting in the international comparison of magnetograms 

 now being prosecuted by the Kew Committee. 



The preliminary results of their investigations have been 

 already indicated by Prof. Adams in his recent speech at 

 Swansea (Nature, vol. xxii. p. 416). G. M. Whipple , 



Kew Observatory, October 2 



Coral Reefs and Islands 



I HAVE been greatly interested in Mr. John Murray's paper on 

 coral reefs and islands published in Nature, vol. xxii. p. 351. 

 1 hope you will allow me space to draw scientific attention to the 

 fact that as early as 1857 I published a paper on the Formation 

 of the Peninsula and Keys of Florida {Am. your. vol. xxiii. p. 

 46), in which I maintain that the theory of Darwin, although so 

 beautifully (as I thought) explaining the phenomena of the 

 Pacific reefs, 'whoHy fails to explain those of the Florida coast. 



In 1 85 1 I spent the months of January and February on the 

 Keys of Florida, assisting Pro?. Louis Agassiz in his investiga- 

 tions on the growth of reefs and formation of keys in this region. 

 An abstract of these investigations and their results was published 

 in the Report of the United States Coast Survey for 1S51, p. 

 li,Setseq.^ 



In this report Agassiz shows that the Keys and nearly the 

 whole Peninsula of Florida have been formed by the growth of 

 successive reefs, one beyond the other from north toward the 

 south. In my paper above alluded to, and also in my "Elements 

 of Geology," p. 152, I state further, that the reefs of Florida, if 

 we acc;pt Darwin's theory, are entirely peculiar. For according 

 to Darwin barrier-reefs are formed only l>y subsidence, while on 

 the Florida coast we have well-marked barriers with channels 

 10-40 metres wide where there cannot be any subsidence, for 

 continuous increase of land is inconsistent with subsidence. 

 Again, according to Darwin barriers and atolls always show a 

 loss 0/ land, only a small port. on of which is recovered by coral 

 and wave agency ; while on the Florida coast, on the contrary, 

 there has been a contmuous growth of the Peninsula by coral 

 accretion, until a very large area, viz., about 20,000 square miles, 

 has been added. 



I have attributed the formation of successive reefs from north 

 toward the south to the successive formation of the depth-condi- 

 tion necessary for coral growth ; and this latter, in the absence 

 of any evidence of elevation, to the steady building up by sedi- 

 mentary deposit, and extension southward, of a submarine bank 

 within the deep curve of the Gulf Stream. The formation of 

 barriers instead of fringes on a coast which has certainly not 

 subsided — for continuous land-growth negatives the idea of sub- 

 sidence—I attribute to the shallowness and muddiness of the 

 bottom along this coast. Only at a distance of twenty to forty 

 miles, where the depth of twenty fa horns is reached, and where, 

 therefore, the bottom is no longer chafed by the waves, the con- 

 ditions necessary for coral growth would be found, and here a 

 line of reefs would be formed, limited on one side by the depth 

 aud on the other by the muddiness of the water. 



In brief then, according to my view, the Peninsula and Keys 

 of Florida were formed by the co-operation of several agents : — 

 I. The Gulf Stream building up and extending a submarine bank 

 within its loop. 2. Corals building successive barriers on the 

 bank as the latter was pushed farther and farther southward. 

 3. Waves beating the reefs into lines of islands. 4. Debris from 

 the reefs and keys on t'jc tiie side and the already formed main- 

 land on the other filling up the successive channels and converting 

 them first into swamps and finally into dry land. 



Whether this view is true in all its parts or not, there can be 



' This report has been recently published in full as one of the 

 of the Harvard Museum of Comparative Anatjmy, but I have not yet 



