585 
of Edinburgh , Session 1874 - 75 . 
Guyetant had not noticed the numerous last place errors in the 
Arithmetica Logarithmica. In regard to this I now find in M. 
Lefort’s paper inserted in the Annales de l’Observatoire the follow- 
ing statement : — The comparison made by MM. Letellier et 
Guyetant was really only to 12 figures. It might have been ex- 
tended to 14 figures for the first 10,000 numbers whose logarithms 
had been computed to 19 places in the Bureau du Cadastre,” 
from which it seems that the object of the comparison was not to 
correct Brigg’s tables but to verify, in so far, the Cadastre tables 
themselves. 
The only other point to which I would refer is as to my mistake 
concerning a third copy . The explanation is simple. In common 
with many others, I had understood that the two copies of the great 
tables were deposited in separate libraries. Having read only the 
papers in the Comptes Rendus, which contain no notice whatever 
of the loss and recovery of one of these copies, nor of the important 
service rendered by M. Lefort in that recovery, I naturally regarded 
the presentation to the Academy as that of a third copy. The de- 
tail of these matters, interesting to all classes of computers, is con- 
tained in the Annales de l’Observatoire, a sectional work consulted 
by only a limited class. From this paper we learn that one of the 
two copies, so like as to be hardly distinguishable, had been long 
amissing, its whereabouts unknown, until M. Lefort, by untiring 
perseverance, traced it to the possession of the Heirs of Prony, to 
whom it had been allowed by way of minutes, “ Cet exemplaire 
avait ete laisse a Prony & titre de minute.” That is to say, the 
Director had taken away one half of the result of this enormous 
labour, lessening greatly the value of the remaining half by 
depriving it of the means of verification ; and that the so-called 
presentation was only the restitution of what should never have 
been taken away. 
I crave leave to add one word in regard to the nineteen-place 
table. On comparing the logarithms of primes from 1163 to 
10007 as given by Legendre in his “ Exercises de Calcul Integral,” 
Tome III., with my own to twenty-eight places, it is found that, 
for primes above 1900, hardly a logarithm is true to the nineteenth 
place; so much so, that to make a list of the errors would be to 
