574 



NATURE 



{April iy, 1890 



be bounded by the inner wedge-shaped angles of the longitudinal 

 -vascular bundles constituting the xylem zone, that such an effect 

 ■could be produced. After that change any inorganic substance 

 finding its way into the interior of this cavity had its surface so 

 moulded by the wedges as to produce the superficial ridges and 

 furrows so characteristic of these inorganic casts. 



March 27. — "The Rupture of Steel by Longitudinal Stress." 

 By Chas. A. Carus- Wilson. Communicated by Prof. G.j H. 

 Darwin, F.R.S. 



This paper gives an account of experiments made with a view 

 to determining the nature of the resistance that has to be over- 

 come in order to produce rupture in a steel bar by longitudinal 

 stress. 



The stress required to produce rupture is in every case com- 

 puted by dividing the load on the specimen at the moment of 

 breaking by the contracted area at the fracture measured after 

 rupture ; this stress is called the " true tensile strength " of the 

 material. 



It is well known that any want of uniformity in the distribu- 

 tion of the stress over the ruptured section causes the bar to 

 break at a lower stress than it would if the stress was uniformly 

 distributed. Hence anything that causes want of uniformity is 

 prejudicial ; for instance, a groove turned in a cylindrical steel 

 bar will produce want of uniformity, and will consequently be 

 prejudicial, the stress at rupture being lower according as the 

 angle of the groove is more acute. The most favourable con- 

 dition of test might appear to be that in which a bar of uniform 

 section throughout its length was allowed to draw out freely before 

 breaking, since in this case the stress must be most uniformly 

 distributed. 



Experiment, however, shows that the plain bar is not always 

 the strongest. So long as the want of uniformity of stress is 

 ■considerable, owing to the groove being cut with a very sharp 

 angle, the plain bar is stronger than the grooved bar ; but, if 

 the groove be semicircular instead of angular, the grooved bar 

 is considerably stronger than the plain, in spite of the fact that 

 the stress is more uniformly distributed in the latter. 



It would seem, then, that we can strengthen a bar over any 

 given section by adding material above and below it, the change 

 in section being gradual ; but such an addition of material 

 cannot strengthen the bar if rupture is caused by a certain in- 

 tensity of tensile stress over the ruptured section; the added 

 material cannot increase the resistance of the ruptured section to 

 ■direct tensile stress, but it can increase the resistance to the 

 shearing stress. 



The resistance of a given section of a steel bar doss not, then, 

 depend on its section at right angles to the axis, but on its section 

 at 45° to the axis, for in that direction the shearing stress is a 

 maximum. From this it would seem that the resistance over- 

 come at rupture is the resistance of the steel to shear. 



Experiments were made to see whether the resistance of steel 

 to direct shearing bore to its resistance to direct tension the ratio 

 •required by the above theory ; since the greatest shearing stress 

 is equal to one-half the longitudinal stress, we should expect to 

 find the resistance to direct shearing equal to one-half of the 

 resistance to direct tension. 



A series of experiments were made, with the result that the 

 ultimate resistance to direct shearing was within, on the average, 

 3 per cent, of the half of that to direct tension. 



The appearance of the fracture of steel bars is next discussed. 

 It would appear that when the stress is uniformly disturbed in 

 the neighbourhood of the ruptured section, the fracture is at 45° 

 to the axis, the bar having sheared along that plane which is a 

 plane of least resistance to shear. The tendency to rupture 

 along a plane of shear may be masked by a non-uniform dis- 

 tribution of stress. 



Two plates of photographs are added, showing examples of 

 steel bars broken by shearing under longitudinal stress. 



Physical Society, March 21.— Prof. W. E. Aryton, F.R.S., 

 President, in the chair. — The following communications were 

 read : — The Villa-i critical points in nickel and iron, by Herbert 

 Tomlinson, F.R.S. Villari has shown that the permeability of 

 iron is increased by longitudinal traction provided the magnet- 

 izing force does not exceed a certain limit, but beyond this limit 

 traction produces a decrease of permeability. The value of the 

 force for which traction produces no change in the permea'iility 

 is known as the Villari critical point. As far as the author is 

 aware, no previous observer has found a similar critical point for 

 nickel, but by confining his attention \.o temporary magnetization 



he has detected such a point with comparative ease. He has 

 also examined the variation of the Villari critical points in iron 

 and nickel with change of load, and has investigated the 

 influence of permanent strain on these points. The experiments 

 were niade by the ballistic method, using wires about 400 

 diameters long. In each set of observations the permeability 

 was obtained with various loads, the magnetizing force being 

 kept the same, and with each load the circuit was closed and. 

 opened until the swings on make and break were equal ; this 

 swing was taken as a measure of the induction under the given 

 load. Several diagrams accompany the paper, in which load 

 and percentage change of permeability are plotted, regard being 

 had to sign. The author finds that for annealed unstrained iron 

 the critical value of the force decreases as the load increases, 

 and that the Villari point is much lower for temporary than for 

 /^/rt/ magnetization. With a load of 47 kilos on a i mm. wire, 

 the value of the force giving the temporary point was 2-8 C.G.S. 

 units. He also found that for a given magnetizing force there 

 are generally two loads which have no effect on the temporary 

 magnetization. With unstrained nickel the critical value of the 

 force is much greater than in iron, being about 114 C.G.S. units 

 for a load of 10 kilos on a wire 0"8 mm. diameter, and 67 for a 

 load of (id kilos. For a force of 21 units no critical point 

 exists. Experiments on a permanently strained iron wire show 

 that for magnetizing forces ranging from 0*03 to 0*3 there is no 

 critical point, and all the resulting curves are identical. There 

 is, however, considerable difference in the observations taken 

 during loading and those taken on unloading. For greater 

 magnetizing forces the curves cease to be identical, and the 

 maximum increase of permeability becomes less and less until 

 for a certain force the curves begin to cut the load line. As 

 the force increases beyond this value the point of cutting ap- 

 proaches the origin, and the curves begin to cut the load 

 line in two points. Further increase of force to 3 C.G.S. 

 units causes the first point to disappear, and the second 

 point recedes from the origin. Finally, with sufficiently high 

 magnetizing forces the second point cannot be reached before the 

 wire breaks, and the curve lies entirely below the load line. 

 With nickel the curves for very minute forces, like those of 

 iron, are exactly the same for different values of the force, but 

 they lie below the load line, i.e. the permeability is diminished 

 by loading ; there is no difference, however, in the loading 

 and unloading curves. Beyond a certain value of the force 

 the identity of the curves ceases, and that part of the curve 

 near the origin bulges towards the load line. For a force 

 a little over 21 C.G.S. units the permeability begins to increase 

 with load, and the curve cuts the line in one point, which point re- 

 cedes from the origin as the force increases. Mr. Shelford Bidwell 

 said that Prof. J. J, Thomson, reasoning from the change of 

 length by magnetization, had predicted a Villari point in cobalt 

 when compressed, and this was verified experimentally. On 

 applying similar reasoning to nickel he, (the speaker) did not 

 expect to find a Villari point, and both S r William Thomson 

 and Prof. Ewing had searched in vain for one. In some experi- 

 ments, not yet completed, he had examined the behaviour of 

 nickel, both loaded and unloaded, when subjected to various 

 magnetizing forces. These show that the metal always con- 

 tracts when magnetized. For no load the contraction at first 

 increased with the magnetizing force, but attains a maximum. 

 With a moderate load the contraction is less for small forces, but 

 for larger forces becomes equal and then exceeds the contrac- 

 tion of the unloaded wire. For greater loads the contraction is 

 less than when unloaded for all values of the force. — On 

 Bertrand's Idiocyclophanous prism by Prof. S. P. Thompson. 

 This hitherto undescribed prism is a total reflection one made 

 of calc-spar, which shows to the naked eye the rings and 

 crosses such as are seen when a slice of spar is examined by 

 convergent light in a polariscope. The spar is cut so that the 

 light after the first reflection passes along the optic axis, and after 

 a second reflection emerges parallel to the incident lijiht. The 

 rings and brushes are present in pairs, but two pairs may be 

 seen by tilting the prism to one side or the other. This was 

 demonstrated before the Society. Prof. Thompson also exhibited 

 a similar prism cut from quartz. Owing to the feeble double- 

 refracting of the substance, no conspicuous rings could be seen, 

 but when examined by the lantern traces of such rings were 

 visible. — On the shape of the movable coils used in electrical 

 measuring instruments, by Mr. T. Mather. The object of this 

 note is to determine the best shape of the horizontal section of 

 swinging coils such as are used in D'Arsonval galvanometers, 



