COMPLEX STRESS DISTRIBUTION IN ENGINEERING MATERIALS. 333 
extent by the Air Service, particularly for testing aluminium alloys. It is now pro- 
posed to give a short account of some of the results that have been obtained using 
these shackles. 
Tt will be convenient to consider two classes of material, i.e. ductile and brittle. 
The former is characterised by a yield point at which the stress strain diagram is 
discontinuous, and a stress strain diagram which is practically horizontal at the ultimate. 
For such materials, therefore, the inequalities in the strain during the elastic region will 
only affect the elastic limit and the yield (presuming that the mean strain is measured, 
E would not be affected), Brittle materials, however, have no yield phenomena, 
and there is no discontinuity in the stress strain diagram, which still has a considerable 
ma 
y/ 
La 
(1986.0) Test Specimen 
The gripping device is shown as used with threaded end specimetv, 
A sunilar device fitted with split sockets would be pipe ia 
shouldered specunezris. 
Fie. 13. 
slope at rupture. For such materials one would expect that the ultimate would-be 
affected as well as the elastic limit. 
Ductile Materials. 
The results of three sets of experiments on different steels are given in Table IIT. 
In each case three consecutive specimens were taken in a bar. The two outer ones, 
Nos. 1 and 3, were tested with ordinary screwed grips, the ends of which were resting on 
a spherical seating of fairly large radius, and the seating was lubricated in order to get 
as good a result as possible. The middle one, No. 2, was tested in the special shackles. 
A typical set of stress strain curves for the large specimens is given in fig. 14. It will 
be seen that in every case the values of the elastic limit and yield are greater with the 
special shackles than with the ordinary ones. The disparity is naturally most marked 
in the case of the small specimens, and the use of these shackles or their equivalent is 
almost imperative if accurate determinations are to be obtained on small specimens. 
In every case the same extensometer was used. This was a special one made by the 
author, but attached to the specimen in the usual way, i.e. asin the Ewing or Cambridge, 
go as to measure the strain in the centre line of the specimen. 
