298 REPORTS ON THE STATE OF SCIENCE, ETC. 
we find in this case a value of 289 lb. per sq. in., or 2.48 per cent. in excess, 
with a maximum stress of 1,000 lb.+per sq. in. at the notch contour, and a 
concentration stress factor of 3.54, very much less in fact than is afforded 
by the standard notch of the form used in Izod machines and described above. 
This difference in concentration of stress is even more noticeable when the 
notch is of rectangular cross section (fig. 6a) with angles rounded off to a 
radius of 0.2 cm. for a notch 1 cm. deep. In this case the mean stress of 
230 lb. per sq. in. recorded by the testing machine agrees remarkably well 
Fig.6a.STRESS DISTRIBUTION AT THE CONTOUR 
OF A STANDARD RECTANGULAR NOTCH WITH 
ROUNDED CORNERS. _ 
=a 
700600 500 400 
Lbs. per Sg Ins. 
Fig.6b. PRINCIPAL STRESSES AT THE SECTION 
C.D OF FIG.6a. 
ew g POCO EA Hee 7 mY aT? ae a 
pee ee fT UO OOS EY Or VUE RB TTD 
with the mean value of 229 lb. per sq. in. afforded by the curve AB of stress 
distribution across the principal cross section (fig. 68), and we now find the 
maximum stress at points in the contour where a tangent line makes a slight 
inclination to the line of pull. Its value is 640 Ib. per sq. in., and gives the 
low stress concentration factor of 2.78. 
These results seem to show that the form of the notch in an impact test 
not only exercises a very considerable influence on the results, but any one 
type of notch may cause that influence to be exerted in a variable manner 
according to the material tested. 
As a test it is undoubtedly complex and even more difficult to analyse in the 
beam form, and its’ simplification, if that were possible, would be a great 
advantage. 
It might, for example, be worth while to examine the possibilities of an 
impact tension test specimen formed by drilling out a very large hole centrally 
placed in a flat tension member, since this form of discontinuity has been 
shown (4) to cause an approximately linear stress distribution across the 
minimum section with a maximum value at the inner contour, and a minimum at 
the outer contour, which is nearly but not quite zero. In hard materials, 
therefore, where the elastic stress distribution probably remains practically 
unchanged up to fracture, the maximum stress is very approximately twice the 
mean value, while in ductile materials photo-elastic indications appear to show 
that the stress distribution becomes fairly uniform at fracture. 
