378 A. W. Ewell — Rotatory Polarization. 



The actual twists were 57, 85, 104, 112, 125, 139 respec- 

 tively. 



Aug. 9. Tube 11. Length of jelly = 3 cm . \' delay between 

 twist and reading. 



Twist.... —60 60 —90 90 —110 110 —130 130 

 Rotation .00 1—1 1—2 4—5 



Sum 2 3 9 



Twist (cont.) —150 150 —170 170 —190 190 



Rotation " 4—13 6—29 16 —57 



Sura " ... 17 35 73 



The actual twists were 57, 86, 104, 123, 142, 160, 178. 

 To illustrate the accuracy of the work, the observations upon 

 tube (43) will be given in full. 



Aug. 11. Tube 43. Length of jelly = 3 crn . J' delay between 

 twist and reading. 







90 



— 90 



120 



— 120 



150 



— 150 



170 



— 170 







84 



— 84 



112 



— 112 



140 



— 140 



158 



— 158 



79 



83 



73 



85 



68 



90 



58 



100 



50 



81 



81 



77 



86 



71 



97 



61 



101 



50 



82 



85 



79 



87 



74 



89 



63 



94 



53 



83 



84 



81 



83 



75 



90 



65 



98 



56 



82 



82 



77 



84 



69 



92 



63 



95 



53 



81 



83 



79 



87 



73 



94 



60 



100 



54 



80 



81 



73 



85 



68 



93 



61 



100 



54 



80 



81 



74 



81 



69 



92 



60 



103 



50 



Means 81 83 77 85 71 92 61 100 52 



Rotation —2 4 —4 10 —11 20 —19 29 



Sum 6 14 31 48 



The natural rotation of all the above tubes was about 2*7 

 per cm. 



Dissymmetry in Rotations.— The dissymmetry in the two 

 directions of rotation for apparently the same numerical twist 

 has been very conspicuous. This is partly due to uncertainty 

 of the true position of the lower clamp for zero twist. A 

 slight error will make the actual twist applied in one direction 

 differ from that applied in the other direction by twice this 

 error, and the difference in the rotations will be still greater. 

 The principal cause is probably dissymmetry in the rubber 

 tube. In most of the cases where the dissymmetry was marked 

 the moment was measured which was required to produce a 

 given twist, and in every case it w T as found that a greater 

 moment was required for a twist in the direction which gave 

 the greater rotation, i. e., for a given twist more force w T as 



