106 MEMOIR OF AUGUSTUS DE MORGAN. 



1839. once disrespectfully told that he would fasten on a head, 

 an arm, or a leg, if he lost one, with gutta-percha, he said, 

 ' I should like to see you do as much.' 



When the repetition of the Cavendish experiment was 

 undertaken by Mr. Baily in the year 1837 at his house 

 in Tav.^ stock Place a house thereby rendered memorable 

 Mr. De Morgan gave so many clear descriptions of it 

 and its object^ that Mr. Baily's work in 1838 and 1839 

 requires a longer notice than I have given to those Astro- 

 nomical achievements with which my husband had less 

 to do. A grant of 500L had been made by Government, 

 at the representation of the Astronomical Society, for 

 defraying expenses. The forerunners of this effort to 

 ascertain the mean density of the earth are mentioned by 

 Mr. De Morgan in the Life of Maskelyne, written some 

 time before, and will give some idea of the nature and 

 objects of the undertaking. 



Cavendish ' The labour of deducing an approximation to the 



men"" earth's mean density was undertaken by Dr. Hutton. By 

 getting the best possible estimate of the materials of which 

 Schehallien is composed, and comparing what we must 

 call the weight of the plumb-line towards the mountain 

 with its weight towards the earth, it appeared that the 

 mean density of the latter is about five times that of 

 water. This, considered as a numerical approximation, 

 alone and unsupported, would have been worth little, 

 owing to the doubt which must have existed as to the 

 correctness of the estimation of the mountain's density. 

 It would prove that there was attraction in the mountain, 

 but would give no very great probability as to the value 

 of the earth's density as deduced. But a few years after- 

 wards Cavendish made an experiment with the same 

 object, and by an entirely different method. By producing 

 oscillations in leaden balls by means of other leaden balls, 

 and by a process of reasoning wholly free from astro- 

 .nomical data, he inferred that the mean density of the 



