NOTE. ON POINSOT’S MEMOIR ON ROTATION. 283 
and when the South Pole was presented, 
9, = 000067 ; 
giving g = 000086 nearly as the true tempera- 
ture coefficient. 
The remarkable accordance of q,, 7,, with the results of the two 
earlier experiments, makes it very probable that the North Pole was 
presented in the experiments of 1843-44, and the South Pole in 
those of 1846. Should such be the case, the true value of ¢ during 
that period would have been 00009 nearly. But it is shewn on the 
same page, that by the multiplication of the equivalent to a degree of 
temperature by & the scale coefficient, there is obtained 
g = 0001105, 
which agrees much better with g = 000112, the value derived from 
the experiments of 1843-44, than it does with g= 00009; from 
which it would appear that the error that would be committed by 
taking g = 000112, and which is caused by a change in the distribu- 
tion of the magnetism, would be almost completely compensated by 
the superposed effects of temperature on the instrument. 
The discordance above referred to between the results of tempera- 
ture experiments in which the two poles are successively presented, 
may be an exceptional property. Of eight magnets tested at my 
Suggestion, by Mr. Stewart, of the Observatory at Kew, through the 
kind intervention of General Sabine, one only showed any material 
difference in the results derived from presenting both poles ; and for 
other magnets that I have tried, results materially the same have been 
obtained, whichever pole was presented ; nevertheless, the fact that 
it has been occasionally otherwise is a sufficient motive, I think, in 
conducting temperature experiments to present each pole of the de- 
fiector instead of one only. 
NOTE ON POINSOT’S MEMOIR ON ROTATION. 
BY J. B. CHERRIMAN, M.A. 
PROFESSOR OF NATURAL PHILOSOPHY, UNIVERSITY COLLEGE, TORONTO. 
This celebrated memoir of Poinsot’s, which, in connection with 
his invention of couples, has revolutionised our whole system of me- 
chanics, treats the subject partly in an analytical, partly in a geome- 
