Oct. 26, 188: 



NATURE 



625 



bility of the bridge when erected would equal that of the best 

 existing structures of that class. 



The paper referred to contains six points of objection, which 

 are treated in a general way without attempting a scientific 

 criticism. This is to be regretted considering the import- 

 ance of the subject. I take each point in succession. With 

 regard to 



I. I cannot see an objection to the novelty of a system, if, as 

 n this case, the conditions are unprecedented, and if the author 

 of the paper himself is compelled to recommend a system of 

 striking novelty. 



II. What, may be asked, constitutes the enormity of magni- 

 tude of a structural part? Is it the excessive proportion of 

 strain in it arising from its own weight to that arising from other 

 weights and forces ? If so, it will be found that this proportion 

 may here be still very small, although it may not be ignored, as 

 sometimes is done. 



III. The experimental knowledge hitherto derived from struc- 

 tures with rising degrees of magnitude has not upset the theories 

 used in the calculations of strength. It cannot be asserted that 

 the top flange of a common rolled beam, being a strut, we assume 

 twenty times as long as it is wide, would be under a test load in 

 a safer position against buckling than the top flange of the Ohio 

 girder bridge, which is 510 feet long and 20 feet wide, or the 

 bottom flange of the Forth Bridge, n hich is 675 feet long and 

 from 32 to 120 feet wide. 



IV. We constantly rely on the strength of long struts ; they 

 exist in all girders, and many of them are of the same import- 

 ance for the strength of the girders as the links for the strength 

 of a chain. The theory of their strengih, imperfect as it is, is 

 applicable to all with a fair amount of truth, and there is no 

 reason why it should not be applied equally to the struts in the 

 Forth Bridge. 



V. Assuming that the dangers from wind-pressure during the 

 erection do not concern us here, it would be interesting to hear 

 from the author which parts of the e rected bridge would pro- ! 

 bably give way first, and whether I his would take place by 

 crushing, shearing, twisting, or pulling actions. The leverage 

 offered to wind by the long brackets would come into question 

 only when the pressure is different on the two sides of a pier. 

 The difference would produce a twisting action, which would 

 exist in the central pier, but which could be obviated in the two 

 side piers. The resisting leverage of the central pier is 270 feet, 

 or about two-thirds of the acting leverage. Approximately the 

 same proportion obtains with regard to the stability against tilt- 

 ing under uniform wind-pressure, while in the ca-e (of the Tay 

 Bridge the proportion was less than one-third. 



VI. It is highly improbable that Mr. Baker should not have 

 calculated his struts ; in his book on the strength of beams, 

 columns, and arches, he gives a very intelligible deduction of the 

 theory of long struts, which, although elementary and not so 

 elegant as that by the author, seems original. I have found 

 deductions of that kind in most English text books, while in 

 books of foreign origin generally the equation of the line of 

 flexure is taken as the starting point. Its approximate form 

 is — 



_ M __ I _ <t* y 

 El p dx- 



M being the bending moment at any point, E the modulus of 

 elasticity, / the moment of inertia of the section of the strut, and 

 p the radius of curvature. The integration gives the limiting 

 weight Wading endways upon a long strut, as already Navier 

 stated it, 



W=t,EI 

 a- 



where E I = C in the Paper. This formula is not applicable to 

 short struts, since W might exceed the crushing strength of the 

 material. The limiting weight IV 1 for short columns is therefore 

 calculated with IV 1 = fp, where / is the sectional area and p 

 the pressure on the sectional unit. Unfortunately there exists 

 among theorists a difference of opinion as to the proper value of 

 / ; some put for it the crushing strength, and others the limit of 

 elasticity, and now and then there are controversies going on 

 about this matter. Meanwhile it is impossible to mark the limit 

 between short and long struts which theoretically exists. Prac- 

 tically, however, the limit is indistinct, and Rankine, Gordon, 

 and others, taking this into consideration, have put the two 

 formulae together into one empirical formula for W, the limiting 

 weight for struts of any given dimensions. 



IV 1 

 TV 



This formula embellished with some empirical coefficients giv<s 

 good results for struts of ordinary proportions, and as the struts 

 in the Forth Bridge seem to have ordinary proportions, it is 

 quite safe to use it for their calculation. M. am Ende 



3, Westminster Chambers, Victoria Street, S.W., October 24 



Having read with interest Sir G. Airy's article on this sub- 

 ject in the last number of Nature, I am glad to see that it 

 advocates a suspension-bridge in lieu of the proposed structure. 

 It may perhaps interest your readers to give the particulars of 

 the Great International Suspension Bridge over the Niagara 

 River, which supports a carriage-way and a railway-track 

 above. 



The length of span between the towers is 800 feet. There 

 are 4 cables, each composed of 3640 wires No. 9= T55' diaiv., 

 without weld or joint; the cables are 10" in diameter. All tKe 

 wires of each cable were separately brought into position, so that 

 each one bears its full share of the tension. When a cable had 

 been thus 1 uilt up, it was tightly served with soft iron wire 

 to bind the 3640 wires together, and to preserve them from 

 rust. 



Since this bridge was built, great improvements have been 

 made in the manufacture of wire. Whereas the resistance to 

 tensile stress at the moment of fracture of the best qualities r.f 

 iron wire, such as that manufactured at Manchester for this 

 bridge, does not much exceed 27 tons per square inch of section, 

 hardened and tempered steel wire can now be made in large 

 quantities and in long lengths with a minimum resistance at the 

 moment of fracture of 90 tons per square inch. 



Steel plates, rod', or bars cannot be made in quantity with a 

 higher resistance than 34 tons, or less than half that of wire. 

 Hardened and tempered steel wire similar to that used in pianos 

 is thus clearly the most suitable material for suspension bridges, 

 and has been recognised as such in America, where it is to le 

 used in the construction of the New York and Brooklyn suspen- 

 sion bridge, the span of which is the.same as the proposed Forth 

 Bridge. 



Our English railway engineers, however, have not yet recog- 

 nised the great advantages wire possesses over any other form of 

 material such as bars, chains, &c, for resisting tensile stress, 

 and the further advantages that wire can be tested more easily 

 and made of a more uniform quality. 



Some ten years ago I called on Sir T. Bouch, the former 

 Engineer to the Forth Bridge, to point out the advantages of 

 a tempered steel wire suspension bridge over any other form, 

 of structure for the Forth Bridge. The idea was, however, 

 nveer worked cut on paper. William H. Johnson 



Manchester, October 23 



On the Alterations in the Dimensions of the Magnetic 

 Metals by the Act of Magnetisation 



I have read with interest Prof. Barrett's paper in Natufe, 

 vol. xxvi. p. 585. Between his results as to the effect of mag- 

 netisation en the dimen-ions of bars of iron, of steel, and of 

 nickel, and those of Sir William Thomson's experiments ("Elec- 

 trodynamic Qualities of Metals," Part VII., Phil. Trans. R. S., 

 Part I., 1879) on the effects of stress in the magnetisation of bars 

 of the same metals, there exists a remarkable analogy, which, 

 however, seems to break down in the case of cobalt. Accord ing 

 to these experiments (which, I may mention, were carried out 

 under Sir William Thomson's direction by my brother, Mr. 

 Thomas Gray, and myself), the effect of the application of lon- 

 gitudinal pull to a bar of iron, while under the influence of induc- 

 tive force tending to produce longitudinal magnetisation, is, for 

 forces lower than a certain critical value, called from the Italian 

 experimenter who fi-st observed it, the Villari Critical Value, to 

 increase, and of the removal of pull, to diminish, the inductive 

 magnetisation. When the magnetising force exceeded the critical 

 value, these effects changed sign, and tended to a constant value 

 as the magnetising force was increased. 



Again, the effect of transverse pull, produced by means of 

 hydrostatic pressure in an iron tube, is, when applied, to 

 diminish the longitudinal magnetisation, and when removed, to 

 increase it. We see, then, from Joule's result, confirmed by 



