48 
PHYSICS: C. BARUS 
It seemed therefore, worth while to further decrease the section. This 
was done, the dimensions being, length L = 1.8 cm., diameter 0.175 cm., 
A = 0.0199 cm.2 The results gave A e / A X 10^= 26.0 and therefore 
£ = 5 X 10^^ / A e. The rates per turn lie between A g = 5.6 (returning) 
and A e = 6.1 (outgoing), so that E X 10~^^ is between 8.9 and 8.2, respec- 
tively. This is so near the normal value for brass that* a further decrease of 
section of the rod figure 3, was undertaken. The final dimensions were L = 
1.8, diameter = 0.138, A = 0.015 cm.^ The results are given in figure 7, 
care having been taken not to overstrain the thin rod. Here A e / A N X 10' 
= 25.8, E X 10"^^ = 66.4 / A e and A e per turn lies between 6.7 (outgoing) 
and 8.1 (returning). Hence E X 10~" = 9.9 and 8.2, respectively, so that 
the normal modulus of brass has actually been reached. In fig. 7 the 
cycles have been spaced as shown by the arrows, to show the separate obser- 
vations. 
5. Glass. — The glass rod first tested was L = 2.33 cm. long, 0.37 cm. in 
diameter, so that A = 0.107 cm.^ 
With a robust interferometer the outgoing and return data were nearly 
coincident; but the graphs were not as a rule straight. The mean rate per 
turn was found to be A e = 8.6. The fringes were of moderate size 
(Ae / AN X 10» = 27.5), so that lO-^i E = 12.75 / Ae = 1.5, a very low 
result. Larger fringes were now installed, giving A e / A N X W = 34.8. 
The results after regrinding the contact face are shown in figure 8 and are 
again nearly coincident, but lie on curved loci. In the first series the larger 
rate is A e = 8.4 per turn; in the second series A e = 10.0 per turn. Hence 
10-11 £ = 15 1 / A e = 1.9 and 1.6, respectively, 
larger than the preceding; but this is still only about one- third of the normal 
modulus of glass. 
The endeavor was now made to proceed as in the case of brass above, with 
a shouldered rod and thinner sections. With this, in view, the glass rod was 
fixed in a small hollow cup, figure 4, with fusible metal. The cup being 
threaded was thereupon screwed into the cross-piece A, figure 1. With a 
glass rod L = 1.9 cm., A = 0.28 cm.,^ moduli as high as E X lO-^^ = 3 to 4 
were obtained. On taking the rod out however, it became clear that there 
had been gradual yielding of the fusible metal clutch. Hence I returned 
finally to the sheath method (fig. 2) using a thin glass rod, L = 2.54, cm. 
long, 0.185 cm. in diameter, A = 0.026 cm.^ The results are given in figure 9. 
The graphs are nearly coincident but curved. The mean rates for the higher 
loads are per turn of the screw, A e = 10.4 (incoming) and A ^ = 8.6 (out- 
going). A e / A N X W 28.8, being the fringe factor, E X lO-^^ = 
57.86 / A e = 5.5 and 6.8, respectively. Hence here also, as in the case of the 
brass rod above, the normal value of the modulus has been reached; i.e., 
one may expect the data for E to be correct on their absolute values, if the 
ratio of length of rod to diameter is of the order of 10 to 1. 
