THE AID OF THE ACHROMATIC FRINGES. 63 



The low value of E when the smaller pressures are applied is again the 

 result of irregular or insufficient contact of the rod with the abutment of 

 the apparatus. Much greater pressures are apparently needed to secure 

 an adequately fixed seat of so rigid a body as the brass rod. 



33. The same. Glass. Glass rods of about the same size as the sheath 

 were next tried, the dimensions being 



L = 2.3 cm. 2r = o-36 cm. .A=o.io2cm. 2 



Observations were made in triplets for loads between 2.5 and 4.5 kg. The 

 following results are examples: 



io 5 A./V/AP= 70 75 65 69 69 cm. 

 io~ ll jE=i.4 1.3 1.5 1.4 1.4 



A film of pitch was then placed between the end of the glass rod and the 

 cast-iron abutment. The apparatus was heavily loaded for some time to 

 squeeze out all superfluous cement. The triplets measured between 2.5 and 

 4.5 kg. of load showed the same order of value, viz, 



io 5 AJV/AP = 77 70 72 69 65 65 cm. 

 io~ n =i.2 1.4 1.3 1.4 i-5 I -5 



The results throughout in very different adjustments are thus remark- 

 ably consistent, but they are all enormously too low, probably not more 

 than about one-third to one-quarter of the true value of E. 

 These unsatisfactory results were not expected, because 

 the modulus is only about half that of brass. The sec- 

 tional distribution of stress is thus even less uniform in 

 case of glass. 



Experiments were also made in cycles, but the early 

 results, though markedly looped, were not quite trust- 

 worthy. The following values are of later date and better, remembering that 

 the micrometer reads to but io~ 4 cm. : 



= o.5 i.o 2.0 2.5 3.5 4.5 3.5 2.5 2.0 i.o o. 

 io B AAT= o 130 285 345 440 53 475 373 335 *55 20 cm. 



They are given in figure 44 and show both in the curved loci and irregularity 

 of detail, that the true elastics of the glass are masked by the incidental 

 discrepancies. 



The values of E thus obtained for glass, though consistent, are necessarily 

 too small. It appears, therefore, that a brittle solid like glass does not con- 

 duct stress uniformly if the sectional area is relatively large as compared 

 with the length. Thinner rods were next provided, loosely fitting the sleeve 

 already used for brass rods. The new dimensions were : 



2L = 4.84cm. 2r = 0.185 cm. A =0.027 cm. 2 



The rods were placed in the interferometer and the compressions observed, 

 as usual in triplets, between loads of 2 and 4 kg. Most of the individual 



