126 



NA TURE 



[June 5, 1890 



(2) Now apply internal pressure. This distends the tube, 

 stretching the roof and floor. The inner wall is compressed with 

 a longitudinal thrust, and the outer wall stretched with a longi- 

 tudinal tension, but the change in the diameter AB, or in the 

 diameter a'b', will be practically unobservable. 



The action on the original gauge and its enclosed fluid of the 

 added part and its fluid, consists now of the forces indicated in 

 the figure, and which amount, as I have in my previous letter 

 pointed out, to a couple (counter-clock-wise in the present 

 figure). 



(3) Now replace the ends at aa' and bb' (this makes no 

 difference in the equilibrium), and holding aa' fixed, remove the 

 added part. 



The gauge will now uncurl, for we are removing the counter- 

 clock-wise couple necessary to maintain equilibrium. Or, to put 

 it in other words, the outer wall ACB, on being released from the 

 tension at B shortens, while the inner wall being released from 

 the pressure at b' becames longer, thus causing the gauge to 

 uncurl. 



As to Gauss's purely geometrical theorem, I fail to see how it 

 is to be of any use in the analysis of the forces, which I take to 

 be the real problem. All that Gauss says to us by his theorem 

 is, " Pure bending in your gauge means uncurling ; if, therefore, 

 you can prove that the forces are such as to produce pure bend- 

 ing> you prove that they produce uncurling." But this is exactly 

 what we cannot prove. Indeed, it is admitted that the bending 

 is not pure. And it is, I think, of no use to urge, with Lord 

 Rayleigh, that the bending is nearly pt(re on account of the 

 comparative inextensibility of the material, for that argument 

 would apply equally to the gauge with both ends fixed, or to a 

 complete annulus which obviously cannot uncurl. In fact, if we 

 could go back to Gauss and ask, "Is it any use showing that the 

 bending is 'nearly pure' ? " he would ask us what we meant by 

 "nearly," and before we could answer that we should have to 

 analyze the whole action. It is for these reasons that I consider the 

 reference to Gauss's theorem not only unfruitful but misleading. 



If we apply the method I have suggested to a tube of elliptical 

 or other than rectangular section, v\ e see that unless longitudinal 

 stresses such as I have dwelt upon would exist in the walls were 

 the annulus completed, the distended gauge will not uncurl on 

 the removal of the added part, and the only reason for consider- 

 ing the curvature of either wall in a plane perpendicular to the 

 circular axis, is that such curvature may, on account of the pro- 

 perties of the material by which it is able to distribute stress in 

 different directions, lead to additions to or subtractions from (and 

 conceivably therefore reversals of) the longitudinal thrust or ten- 

 sion that would exist in a tube of rectangular section. But this 

 is obviously a question of the structure of the material and not of 

 pure geometry. A. M. Worthington. 



R.N.E. College,. Devonport, May 14. 



A Subject-Index and the Royal Society. 



The following brief account of an offer I have made to the 

 Royal Society will, I think, prove of interest to the readers of 

 Nature, and especially to those correspondents who have em- 

 phasized the importance of a subject-index for the progress of 

 science in all its branches. 



Upon the conclusion, last autumn, of five years' work, during 

 which my "Epitome of the Synthetic Philosophy" was com- 

 piled, I commenced to look for some literary work which would 

 be of undoubted practical service to science, and which would if 

 possible aid its further development. The articles and numerous 

 letters then appearing-in your columns urging the importance of 

 a subject classification of the memoirs arranged under the authors' 

 names in the " Royaji Society Catalogue of Scientific Papers," 

 led me to discuss the advisability of my undertaking such a pro- 

 ceeding with several friends upon whose judgment I could rely ; 

 with the result that, one and all agreeing upon the value of such 

 a work, I wrote to Sir G. G. Stokes, P. R.S. (to whom I was 

 directed), offering to compile the manuscript of the greater por- 

 tion of such a work upon the condition that, as soon as the manu- 

 script was completed and approved by the Council of the Royal 

 Society, the Society should guarantee all expenses of print and 

 publication. I was forced to say "the greater portion" of such 

 a work ("70 to 80 per cent."), for examination of a large 

 number of titles had shown me that a certain percentage of them 

 could only be correctly indexed by specialists in their own de- 

 partments, a fact which is emphasized when we call to mind 

 that a title may be in any one of eight European languages. 



After several interviews with Sir G. G. Stokes, and a some- 

 what protracted correspondence, I agreed to arrange a sample 

 index of some 20CO entries upcn a plan suggested by him, and 

 warmly approved by that eminent bibliographer Dr. Garnett, 

 of the British Museum, the plan being to take the leading word 

 or words in the title of each paper, with a reference to the 

 volume, page, author's name, and number of the paper, in the 

 Royal Society's Catalogue, for subsequent arrangement in alphgi- 

 betical order, by which means the subject-key would occupy but 

 a quarter of the bulk of the Catalogue as now published. It 

 would extend — that is, approximately— to three quartos of the 

 size of the present volumes, in similar type, &c. 



In the early part of May this plan was discussed by the 

 Catalogue Committee of the Royal Society, when the following 

 resolution was passed: — "That the offer of Mr. Collins be 

 declined, and that the President be requested to convey to Mr. 

 Collins the best thanks of the Committee for the trouble which 

 he has taken." 



The foregoing account will be sufficient to show that, con- 

 trary to an opinion expressed more than once in these pages, 

 something more is needed than an offer to compile the subject- 

 index. Were the manuscript now completed, and approved by 

 the Royal Society, there would still be wanting a sum sufficient 

 to bring it before the public. 



In conclusion, I should like to express my warm thanks to 

 Sir George Stokes for the kind and courteous way in which he 

 has assisted me in my endeavour to develop what I am still con- 

 vinced would be of immense service to science in all parts of the 

 civilized world ; and not only to science, but to many industries 

 besides. For would not the chemical manufacturer and the 

 dyer profit by a complete list of all the papers that had been 

 written on the coal-tar colours ; the agricuhurist, by knowing 

 the researches which had been undertaken to ascertain the 

 nutritive powers of the bones and phosphates, and the fattening 

 properties of the various cakes and foods ; the engineer, the 

 analyses of iron and steel with their accompanying properties ; 

 the physician, the physiological action of the various drugs ; 

 and the electrician, all the papers, for instance, which had been 

 written upon that little understood subject, induced cuirents? 

 Finally, how many millions might have been saved in the con- 

 struction of harbours all over the British Empire had all the 

 scattered information upon the flow of water in rivers and tidal 

 estuaries been so gathered together as to make reference possible, 

 not to say easy? F. Howard Collins. 



Churchfield, Edgbaston. 



Stream Lightning. 



If a candle-flame is put between the poles of an electrical 

 machine, while it is giving rough angular sparks, the discharge 

 at once changes into a smooth single line of very easy curvatuie : 

 it suggests the difference between sinuous and stream lightning : 

 it is not merely that the spark is as if shortened by the conduct- 

 ing flame ; the whole nature of the discharge is changed. If the 

 flame is held two inches beneath the poles, the spark will go 

 down to it. \V. B. Crofp. 



Winchester College, May 30. 



NO. 1075, VOL. 42] 



Atmospheric Circulation. 



On March 9 and 10, 1887, the barometer rose to 30-92 inches 

 over Iceland — a very exceptional height for that locality at that 

 time of the year. The United States daily maps of the iiorthern 

 hemisphere show that a storm to the southward of this great 

 anticyclone was carried westward a distance of over six hundred 

 miles within twenty-four hours, in a manner similar to that in 

 which West Indian hurricanes follow the course of the trade 

 winds in August and September, although this storm was located 

 in latitude 40° N., or in the usual situation of the anti-trades. 

 Other instances of a similar character have been noted, but this 

 one was unusually well defined, and throws much light upon the 

 laws governing the atmospheric circulation. 



Lyons, N.Y., May 20. M. A. Veeder. 



Testing for Colour-Blindness. 



In answer to Prof. Lodge's query (May 29, p. 100), why those 



concerned in testing for colour-vision do not avail themselves of 



instrum.ents like Lord Rayleigh's, having tested some thousands 



in this city, I may state that experience has shown that they are 



