that Matter takes up no Room in the JEther. 853 



In the left-hand member of this inequality we are certainly 

 justified in taking (A, B, C) to be as great as the earth's 

 orbital velocity, for the adverse chance referred to in § 20 

 cannot here cast doubt on our conclusions, inasmuch as a null 

 result was also given by the readings of 1913 Feb., in which 

 the fineness of observation was increased tenfold. The 

 value which may be assigned to A 2 + B 2 is thus roughly 

 (3xl0 6 cm./sec.)' 2 ; 9 the co-latitude is about 38° 15' or 

 1 + cos 2 6 = 1-62; M='0616gram; (7 = 21-6; H (§ 18) is not 

 less than *75. Hence 



F 2 /o<10- 21 certainly. 



23. The second set of: observations used was made with 

 the plate shown in fig. 2 ; and if there were any semi-diurnal 

 term in the azimuth of the mirror, the amplitude was cer- 

 tainly less than 8 seconds. The moment of inertia of the 

 plate was *Q0258 C.G.S., and its period 58 seconds, so that 

 there was no semi-diurnal couple acting on the plate with an 

 amplitude as great as 1*36 x 10 " 9 dyne-cm. Proceeding as 

 in § 22, with the probable assumption that the quadratic 

 velocity -factor A 2 -I- B 2 is at least that due to the earth's orbital 

 motion, we find that 



F~>< 10 " 22 probably. 



24. Let us now provisionally assume for the density of the 

 aether the value 10 12 grams per c.c* From the result of 

 § 22 it would then follow that F 2 <10~ 33 , and that F (whether 

 positive or negative) is numerically less than 3'lxl0" 17 , a 

 limit which is depressed to 10 ~ 17 if we assume the result 

 given as probable in § 22. Thus, if p = 10 12 , the room which 

 one gram of matter f takes up in the aether is certainly less 

 numerically than ± 3*1 x 13 -17 c.c, probably less than 

 ±10 -17 c.c, and possibly nothing whatever. 



25. Though the estimates just given rest on a conjectural 

 value for the density of the aether, it may yet be of interest to 

 express them in a different form. For matter other than 

 hydrogen, the number of electrons in any atom is about half 

 the atomic weight; and if we take 10 ~ 13 cm. as the radius of 

 an electron, the total bulk of electrons in a gram of matter 

 is roughly 1*4 x 10 -15 c.c If p = 10 12 grams per c.c (or any 



* Cf. Lodge, Phil. Mag-, xiii. 1907 ; ' Nature,' lxxv. p. 519 (1907). The 

 estimate is necessarily speculative ; a revision of the argument in the light 

 of more recent experimental data would, I think, tend to raise rather than 

 to lower it. 



t Strictly speaking, we are only entitled to state this of the iridio- 

 platinum alloy used in the experiments. 



Phil. Mag. S. 6. Vol. 27. No. 161. Mai/ 1914. 3 L 



