494 Intelligence and Miscellaneous Articles, 



Nevertheless our numbers a little exceed those of M. Stefan. These 

 slight divergences may, we think, be attributed to two causes : — 



(1) Our experiments must have been made at a notably higher 

 temperature ; for we worked on some of the very hot days of this 

 summer (outside temperature 20° to 25° C), and, besides, the sun's 

 rays concentrated by a lens upon the plate of quartz necessarily 

 raised its temperature. JSTow t it has been shown by Von Laug 

 (Sitzungsberichte der Wiener Akademie, vol. lxxi. April 1875) that 

 the rotatory power of quartz slightly increases with its tempera- 

 ture ; and, to express the increment, he gives the formula 



0,=0 O (1 + -00001490, 

 where </> and <pt are the angles of rotation at zero and at t°. By 

 admitting a difference of 20° between the temperature of M. Stefan's 

 experiments and that of ours the difference between the results 

 would be reduced about 3 thousandths. 



(2) The plate of quartz w T e use shows a slight imperfection in 

 its cutting, with respect to parallelism of the two faces and the di- 

 rection of the axis ; and the rays did not pass through it quite 

 normally : they were rendered convergent by a lens of long focus. 

 This little cause of error also tends to augment the numbers 

 obtained. 



For the expression of the angle of rotation as a function of the 

 wave-length A, the following formula has been proposed — 



♦ — A+ = 



in which A and B are two constants. This formula, which accords 

 in a manner but little short of satisfactory with the observations 

 made between B and H, becomes inaccurate between more extended 

 limits. When the values of A and B are calculated from those of 

 (p for lines a and M, it is found that, for rays of intermediate re- 

 frangibility, the values given by the formula are constantly higher 

 than the numbers observed ; the divergence exceeds 1 degree for 

 the line Gr. On the contrary, for the lines A and N the calculated 

 values are lower than those resulting from observation. 



Starting from the idea that the rotation should be zero for an 

 infinite wave-length, Boltzmann has proposed the formula 

 BCD 



He has shown that, reduced to its first two terms, it conducts to 

 values well accordant with the observations of M. Stefan between 

 B and H. "We have found that it likewise agrees very satisfactorily 

 with our results. Calculating B and C from the numbers we have 

 obtained for the fines a and M, w 7 e get 



_ 740533 0-151227 

 * 10 G \ 2 10 12 X 4 ' 



from which are deduced the numbers of the eighth column of our 

 Table. 



