On the Mean Density of the Earth. 7 
2° or 4°95 for the mean density of the earth. Or, if we assume 
the density of the mountain still higher, as 2°8 instead of 2°75, 
we then obtain 2 x 2°8 = 5:05, a little more than 5 for the 
earth’s density; which last number 5 I therefore fix upon in 
conclusion, as probably the nearest to the truth; or at least it is 
sufficiently large, as it is grounded on several assumptions that 
are most favourable for the highest result; namely, 2777714 or 
22 for the density of the mountain; also £ a3 rather above the 
calculated ratio of the densities of the earth and mountain; and 
lastly, the assumption of the mountain being quite solid; though 
it is probable that there may be cavities in most mountains, as 
they are generally the production either of volcanoes or of earth- 
quakes. OW 
For all these reasons then, it is highly probable that the earth’s 
mean density is very near five times the density of water; but not 
higher. If any person should still hesitate to adopt this con- 
clusion, his hesitation must arise from doubts either on the data 
obtained by the measurements, or on the accuracy of the com- 
putations made from them. But if any such person attentively 
read over Dr, Maskelyne’s account of the measurements, in the 
Phil. Trans. of 1775, his doubts must be soon removed as to the 
data supplied by the survey of the hill, or by the astronomical 
observations. And as to the accuracy of my own computations, 
made from those data, they are fully and fairly before the public, 
- in the works before mentioned ; and let any person, who doubts, 
look over and repeat the calculations there stated, and try if he 
can find any inaccuracy in them. ‘The only possible ground of 
doubt in the measured data, must be in the observed deviation in 
the plumb line taken by Dr. Maskelyne ; but when we consider 
the accuracy of the observer, and of the instruments, and read 
the account of the use of them, it must be then very difficult to 
doubt of their accuracy. On this point it is commonly acknow- 
ledged, that a good observer, with the best instruments, can observe 
angles to a small fraction of a second. Dr. Maskelyne’s obser- 
vations give 1-6 seconds for the sum of the deviations of the 
plumb line, from a medium of between 300 and 400 observations. 
Now let us suppose it possible to have committed an error of 
four-tenths of a second in this number, and that the true num- 
ber should have been 12 seconds, instead of 11:6, being an error 
of the 29th part of the whole: This then would cause an error 
of the 29th part in the result ; which would reduce the density 5 
to about 4°$; showing that the number 11-6 is not too small, 
but the contrary. Next let us assume 11 seconds only, omitting 
the 6-10ths, being almost the 20th part of the whole, and which 
therefore would give nearly 5*25 for the earth’s density, being 
still far below the number 5°48, as deduced from Mr. Cavendish’s 
experiment, 
