the intensity of magnetic forces, &c. 15 
mean and 0.1247, and I thus get .1226 as the mean variation 
of the intensity of the magnets for a change in their tempe- 
rature of 1°, between the temperatures 59.05 and 77-^5 j an 
increase of temperature always causing a decrease of inten- 
sity, and vice versa. 
In the results in the last column of this table there are no 
marked indications of an increase in the values of A . arising 
from an increase of the temperature at which the observa- 
tions were made. Having afterwards, when I carried the 
observations to a greater extent in the scale of temperature, 
clearly ascertained that this was the case, I determined there- 
p 
fore not to take the mean of the values of A . :;rr, as I have 
here pointed out, but as I had made observations at every con- 
venient opportunity , to take out from them, in the first place, all 
the values of A . where the mean between the temperatures 
from which they were derived agreed nearly with the lowest 
temperature of the observations which it was my object to 
reduce ; in the same manner, to take those which agreed 
most nearly with the mean temperature to which these ob- 
servations were to be reduced ; and likewise those agreeing 
with their highest temperature: taking then the mean of 
each of these, from these three means, I derived a value of 
p 
A from which I determined the variation of the angle <p, 
corresponding to any change of temperature. I have men- 
tioned this here, that my reason for giving so many of the 
observations may be apparent. Observations, precisely simi- 
lar to the preceding, were made on the 7th of June : they 
are contained in the following table. 
