NATURE 



[February 24, 19 16 



and industry, was asked by the authorities to preside 

 over some txjmmittees and to serve on others in order 

 to help the Government, and the country, out of the 

 dilemma. He was too patriotic to refuse, but the 

 strain was too j^reat for one who was far from strong, 

 and he died after a few months of overwork. How 

 can the country be saved from the disastrous conse- 

 quences of the neglect of science? How can the 

 society hope to improve, by means of an enlightened 

 Government, the racial qualities of future generations? 

 The remedy is simple, but there is every reason to 

 believe that it will be effective. All that is necessary is 

 to change the character of the examination for the 

 Civil Service posts and for the Army, giving science a 

 far more important place than it has held hitherto. 

 This change would at once react on our public schools 

 and the old universities, and would bring the mem- 

 bers of future Parliaments under the influence of 

 science. 



POLARISED LIGHT AND ITS APPLICA- 

 TIONS TO ENGINEERINGA 

 ONE of the fundamental questions which arises in 

 the majority of engineering problems is the 

 design of a structure or machine which will carry 

 out some predetermined work in an efficient and 

 economical manner, and whatever the problem may 

 be, it is almost invariably bound up with the arrange- 

 ment of a number of connected parts designed to resist 

 loads which are imposed upon them. 



The machines and structures which the engineer has 

 to construct are almost infinite in their variety, and 

 each one usually presents a new and a difficult 

 problem,, especially as regards the stresses which may 

 be imposed upon its parts, and the way in which these 

 stresses are distributed. 



It is a common experience among engineers to find 

 themselves confronted with a stress problem in their 

 designs which presents almost insoluble difficulties; 

 it often defies mathematical processes, and is beyond 

 the scope of any previous physical investigation. But 

 it must be solved, if only approximately, and the 

 imperative need of an answer renders it advisable to 

 make experirnental investigations before proceeding 

 with an important work of construction. 



It is perhaps somewhat severe, but not untrue, to 

 say that engineers have not always made the fullest 

 use of the discoveries of pure science in their prac- 

 tice; and it is remarkable that a discovery of Sir 

 David Brewster, in 1816, that transparent materials 

 when stressed become doubly refractive, should not 

 have been more frequently pressed into service, for its 

 use was immediately obvious to the discoverer, who 

 pointed out that the stresses in the arched rings of 

 bridges could be rendered visible in a glass model by 

 aid of the doubly refractive effect produced b}' a beam 

 of polarised light. 



Here and there one finds accounts of applications 

 of this property for engineering work, but usually 

 with little success, mainly owing, no doubt, to the diffi- 

 culties experienced in shaping glass models to the 

 required form ; but when these are overcome the 

 value of the information gained is very great, as, for 

 instance, in the recent valuable investigations of the 

 stresses made upon a glass model of a reinforced 

 concrete arch by M. Mesnager, of Paris, who used 

 the results so obtained for the design of an arch of 

 about 310 ft. span, with a most gratifying agreement 

 between the stresses in the actual bridge and its 

 model. The expense and difficulty of constructing 

 glass models are a bar to their general use, and other 

 transparent materials are now available which offer 



' Abridged from a dkroiirse d-livered at the Rojal Institution on Friday, 

 February i8, by Prof. F,. G. Coker. 



many advantages, in that they are strongly doubl\ 

 refracting under stress, are easily fashioned wiiii 

 engineering tools, and are not readily broken 01 

 damaged, while the cost is insignificant. 

 _ Here, for example, is a rough model of an arch 

 ring, made of xylonite, and you observe that wht 1 

 loads are applied, it glows with colour in the polari- 

 scope, and a picture of the state of internal stress is 

 obtained which can be readily interpreted. 

 Measurement by Colour Effect. 



We can estimate simple stresses by the colours 

 observed. 



If, for example, we take a strip of transparent 

 material, and arrange the optical apparatus so that 

 when the strip is unloaded no light is transmitted,, the 

 effect of a moderate tension causes the specimen to 

 appear a greyish-white, and, as the stress increases, 

 the colour changes by insensible gradations to a 

 lemon-yellow, then to a reddish-purple, and, with a 

 very little increase of stress, to a well-defined blue. 

 With a further increase of stress, the scale of colours 

 is approximately repeated for twice the intensity of 

 stress, and the relation of colour to stress can be easily 

 determined. 



For simple tension and compression, therefore, the 

 stress intensity may be inferred by observing the 

 colour bands, bearing in mind that both tension and 

 compression produce similar effects, if changes in the 

 thickness of the material are allowed for. Thus, if 

 we take the case of a transparent beam subjected to a 

 uniform bending moment, a system of colour bands 

 is obtained, distributed as shown in the accompanying 

 experiment, and, by inspection, the distribution across 

 the section can be determined as shown in the 

 diagram. 



This case and others which have been examined 

 afford instances where the results of optical experi- 

 ments can be compared, not only with mechanical 

 measurements of strain, but also with the theory of 

 the distribution of stress in materials ; and the experi- 

 mental determinations for a transparent material show 

 a very good agreement with strain measurements and 

 with the precise theory. We can, therefore, feel very 

 confident that in more complicated cases the stresses 

 in a transparent model are similar to those in a metal. 

 For example, a beam with a notch cut in it may be 

 taken (as shown), and, as might be expected, the 

 effect of the notch is seen to increase the stress in 

 the material very considerably. The distribution Is 

 now much more complicated than it is in a simple 

 beam ; the neutral axis has moved towards the notch, 

 while the colour effects show that the maximum stress 

 is at least twice as great as that in a beam without 

 a notch. 



Laws of Optical Effect. 



In most of the cases arising in engineering work 

 the stress distribution is even more complex, but it is 

 known that any case of stress in the plane of a plate 

 can always be represented by two principal stresses 

 at right angles, and if the magnitude and direction 

 of these are determined for all points the stress dis- 

 tribution is solved. 



In order to obtain an experimental solution of this 

 problem, it is necessary to inquire into the relation of 

 the optical effect to the principal stress intensities at 

 a point, and it is easy to show this by simple experi- 

 ments. If, for example, we take two tension members 

 and subject them to the same uniform stress intensity, 

 the colour effects produced by interference will be pre- 

 cisely the same for each, while if they are superposed 

 the colour effect is that produced on a single 

 member under twice the stress. If, however, 

 two equally stressed tension ♦aembers of the 

 same thickness are crossed, the common area gives a 



NO. 2417, VOL. 96] 



