158 Mr Hicks, On some Irregularities in the Values [May 26, 



vary from 5"2886 to 5 - 87l7, Baily himself says they are " not 

 reliable," and says, — " I made only 23 experiments in this way, for 

 although the principle may be correct, yet the difficulty of seizing 

 the precise moment at which the masses ought to be put in motion 

 occasionally introduced disturbances which affected the accuracy of 

 the results;" and later, "it may become a question whether these 

 experiments are entitled to that confidence which the subject 

 requires." If we had a large number of series from which to take 

 our means it might be well to retain them, but naturally at so low 

 a temperature only a few observations were to be expected. The 

 results of the discussion so far as it has been carried, are given in 

 the following table. The first column gives the temperature, the 

 second the number of series on which the mean given is based, the 

 third the number of daily means, and the fourth the number of 

 separate means employed. The means for the different tempera- 

 tures are given in the fifth column. 



The gradual fall of mean density with rise of temperature is 

 most marked, the only exception being in the case of the lowest 

 temperature (36°) which is slightly smaller than for the tempera- 

 ture of 40°. However, this is based on only four series, but 

 is nevertheless higher than for any temperatures above 40°. 

 It is easily seen from the irregularity at which the means 

 decrease that all disturbances have not been eliminated ; still 

 the numbers suffice to shew that some cause depending on tem- 

 perature has been at work to produce disturbances. Although 

 I undertook the discussion for the very purpose of seeing whether 

 such an effect was present, I must confess I was astonished to find 

 it so marked. It certainly seems too large to be explained by 

 supposing the gravitation between two bodies to be a function 

 of their temperatures, although it is in the direction which we 

 should expect to find, if the gravitation increased with the tempera- 

 ture. For if so, we should suppose from our experiments at a high 



