Prof. Jellett on a new Optical Saccharometer. 359 



pieces, and the zero of the vernier made to coincide accurately 

 with the zero of the scale. The milled head is now turned so 

 as to draw back the tube until the tints on the two parts of the 

 circular image seen through 7 become equal. The number on 

 the scale corresponding to the zero of the vernier is then noted. 

 Let this reading be R, and let S be the strength of the known 

 solution. 



Now, let this solution be removed from the tube, which is 

 then to be filled with the solution whose strength is required. 

 The same process having been gone through, let the new read- 

 ing be li'j then the strength required is given by the equation 



S'= £ • S. 



If the experiment be carefully conducted, and if there be no 

 error in the strength of the standard solution, the error in 

 the measurement made as above described ought not to ex- 

 ceed 0-02 gr. per cubic inch for a single experiment. If the 

 mean of a number of experiments be taken, the error will of 

 course be still less*. 



The author has given to this instrument the name saccharo- 

 meter, derived from one important use to which it may be ap- 

 plied. This, however, is but one of its applications ; and there 

 are many others at least as important. It may generally be de- 

 fined to be an instrument by which the ratio of the rotatory power 

 of any transparent fluid to that of a standard fluid may be accu- 

 rately determined. 



It is not desirable to use a very strong solution of the sub- 

 stance to be examined. The reason of this is the imperfect 

 compensation which exists between fluids possessed of opposite 

 rotatory powers. It is generally assumed that the ratio of the 

 rotation produced in the planes of polarization of any two of the 

 simple rays of which a white ray is composed is the same, what- 

 ever be the substance causing the rotation. It follows, indeed, 

 from the law of Biot, that this is not accurately true ; but it has 

 been generally supposed that the error is too small to be per- 

 ceived. If this were true, it would always be possible to assign 

 to the lengths of two columns of oppositely rotating fluids such 

 a ratio that the effect of the one should be accurately compen- 

 sated by the effect of the other. But the author has found that 

 in certain cases the error is very perceptible indeed. This is 

 shown by the impossibility of giving to the tube ff any position 



* [Professor Jellett estimates the error to which even a practised expe- 

 rimenter is liable, in making a similar determination by means of Soleil's 

 saccharometer, at not less than 0'5 grain of sugar per cubic inch of solution.] 



