June r6, 1910J 



NATURE 



461 



Curve Tracing and Curve Analysis. 



The review of a book on practical curve tracing in 

 Nature of June 9 is tantalising to one who is not in 

 the least interested in drawing the graph of an equation, 

 but who is frequently plotting curves from experiment, 

 and who would like to find formula, not only to fit them, 

 but to explain them. I look through most of the reviews 

 of mathematical books in Nature in the hope of discover- 

 ing one that deals with the practical analysis of curves, 

 and I am continually disappointed. 



Can no mathematician be induced to recognise that for 

 some of us an equation is the end, and not the beginning, 

 of a piece of work? In innumerable cases experimental 

 work ends with a curve, such, for example, as a hysteresis 

 curve, and no attempt is made to find an equation to 

 fit it. 



Half a dozen rules exist, the uses of log. and semi-log. 

 paper can easily be explained, but nobody has gathered 

 them together with explanation of the difference between 

 empirical formulae and rational equations, of interpolation, 

 smoothing, and of the legitimacy of extrapolation. 



London, June 9. - A. P. Trotter. 



A Brush for Collecting Mercury. 



.Since more or less mercury is always spilled around the 

 laboratory, a simple and efficient mercury collector is of 

 great use. I have found a very good one, and, s'nce I 

 have not seen it in use before, I will describe how it is 

 made. 



It is made like a paint-brush, with the difference that 

 140 copper wire is used instead of camel's hair in the 

 brush part. The fine copper wire is then amalgamated 

 with mercury. Use the brush as though painting with it. 

 It will take up large globules and go into cracks and 

 collect the smallest particles, so that none need be lost. 

 Use a cup when collecting, and when the brush is full 

 shake the mercury into the cup. 



George Winchester. 



Washington and Jefferson College, Physical 

 Laboratory, Washington, Pa., May 28. ~ 



LIGHT ALLOYS. 



HTHE problem of producing an alloy which shall 

 ■ combine tjreat strength with a low specific 

 weight has been before metallurgists ever since the 

 commercial manufacture of aluminium became an 

 accomplished fact ; more recently, however, the re- 

 quirements, in the first place, of racing yachts, then 

 of motor cars and of motor cycles, and, finally, the 

 pressing problems of aerial navigation, have added a 

 rapidly increasing importance to the whole question. 

 At the present moment German manufacturers par- 

 ticularly are putting forth claims in regard to achieve- 

 ments in this direction which appear startling at first 

 sight, and it is interesting to examine the whole state 

 of the question. 



The need for a light alloy lies in the fact that pure 

 or nearly pure aluminium is, unfortunately, verv* weak 

 mechanically. Its low specific gravity (2'7i) is more 

 than counterbalanced by the fact that its tensile 

 strength, even in the form of rolled bars, does not 

 exceed 7 tons per square inch. If these figures 

 are compared with those of the best special alloy 

 steels suitable for structural purposes, we find that 

 some of these show tenacities up to 64 tons per 

 square inch, with a density of approximately 7*9. 

 Consequently, a bar of aluminium, to bear the same 

 ultimate load as a bar of such steel having a cross- 

 sectional area of one inch, must have a sectional area 

 of approximately 9 square inches, and would there- 

 fore weigh about three times as much as the steel 

 bar. A light alloy which is to compete successfully 

 with such special steels, therefore, must either be 

 much lighter than pure aluminium or it must combine 

 with the density of aluminium a tensile strength of 

 21 tons per square inch. 



NO. 2120, VOL. 8.-^] 



So far as alloys consisting principally of aluminium 

 are concerned, it does not appear that this tensile 

 strength has ever been attained, except in the case of 

 hard-drawn wires the ductility of which has been re- 

 duced to an excessively low value. It must, however, be 

 borne in mind that the high-tension steels referred to 

 above cannot be employed in excessively thin sections, 

 so that in many special cases, where the scantling of 

 structural parts cannot be reduced to minute dimen- 

 sions, while the strength required is not very great, 

 light alloys may be employed with advantage as com- 

 pared with alloy steel. The same argument applies, 

 however, to a comparison made on similar lines be- 

 tween light alloys and the stronger kinds of wood. 

 These woods are all considerably weaker, per square 

 inch of sectional area, than the light alloys now 

 available, but when their much lower density is taken 

 into account, as well as the advantage of larger 

 scantlings, the result must in many cases be favour- 

 able to the employment of wood. It is for this reason 

 that the frames of most aeroplanes are constructed of 

 wood. When, however, an alloy of density less than 3, 

 and possessing a tensile strength of more than 20 tons 

 per' square inch under conditions allowing of a ductility 

 equivalent to an extension of not less than 15 per 

 cent, on a 2-inch test-piece, becomes available, its 

 employment will become advantageous as compared 

 both with the best alloy steel and the best wood. 



The light alloys available at the present time are 

 somewhat numerous, and, as regards those of them 

 which are patented or otherwise proprietan.- articles, 

 it is difficult to obtain satisfactory data; it is certain, 

 however, that extravagant claims are often advanced 

 for such alloys, and these are not verified when 

 samples are tried in a testing machine. 



The claims of those advertising or selling such 

 alloys must therefore be looked upon with much re- 

 serve. 



Among the earlier alloys of aluminium which found 

 a certain amount of practical application were those 

 with iron and with nickel. One of the racing yachts 

 engaged in one of the later races for the America 

 Cup was built of plates rolled from one of these 

 alloys, but the metal suffered from excessively rapid 

 corrosion, and the presence of iron in aluminium 

 alloys, although it undoubtedly confers considerable 

 strength upon them, is rightly regarded as extremely 

 undesirable. At the present time, the most completely 

 studied of the light alloys are those in which copper- 

 is incorporated with the aluminium, either alone or 

 with the addition of other elements, such as man- 

 ganese. In the form of rolled bars and sheets, these 

 alloys attain a tensile strength of slightly more than 17 

 tons per square inch, with an elongation of 15 per 

 cent, on 2 inches ; these figures apply almost equallv 

 to alloys containing about 4 per cent, of copper alone, 

 or to those containing 3 per cent, of copper and i per 

 cent, of manganese, or 2 per cent, of copper and 

 2 per cent, of manganese, the specific gravities of all 

 these alloys lying close to 2 "8. So far as trustworthv 

 data are available, these figures probably represent the 

 best available alloys of this character. Allovs of 

 aluminium with from 15 to 20 per cent, of zinc mav 

 possibly yield somewhat higher figures, but, owing 

 to the presence of a considerable proportion of zinc, 

 their density is also much higher, so that thev can 

 hardly be classed among the light alloys. 



The light alloys at present employed in practice are 

 principally used in the form of more or less compli- 

 cated castings, such as motor-car engine crank-cases, 

 the corresponding parts of aerial motors, and similar 

 purposes. When thus used the alloys cannot be com- 

 pared with special alloy steels, and still less with 

 wood, and they hold the field quite easily against 

 cast-iron, brass or bronze of anv kind. For these 



