431 
o f Edinburgh, Session 1874 - 75 . 
of geometers seems to have been aware. The method followed 
in the calculation of the Cadastre table of logarithms was an 
egregious blunder. The result was in accordance with the 
method. 
After having told us that, for numbers above 10,000, the errors 
in Brigg’s last three figures are not given, M. Lefort goes on to 
say (page 997): — “For all that there do exist numerous diver- 
gencies in the thirteenth and fourteenth places. We must thence 
conclude that the calculators did not suppose the Cadastre tables 
to have enough of precision for correcting Brigg’s last places. In 
that they were perfectly right. The bases of the calculation had 
been chosen so as to make sure of twelve places, and the precautions 
taken were only applicable to the research of these twelve places, 
which are all that are fit for publication.” 
In order, after the method, but not behind it in importance, 
comes the conscientiousness with which the work is performed. 
The careful computer who may have to revise his own work puts 
the first performance aside, even leaves it for a considerable time, 
lest the sight of the figures, or the remembrance of them, should 
lead to the repetition of an error : better still, he arranges the 
operation in another way, so that the same additions and subtrac- 
tions may not recur. But when he has to do with hired assistants 
he must contrive safeguards against carelessness, even against 
simulation. Against the former it is possible, though difficult, to 
protect one’s self, but against the latter there is no protection 
other than in repeating the work, which comes, in effect, to the 
dismissal of the delinquent. The most obvious safeguard is to 
have the work done by two or more computers who have no oppor- 
tunity of intercommunication. 
Now the above described fundamental arrangement of the work 
excluded the possibility of such isolation. The computer^ra very 
ordinary computer, knowing little of mathematics — ^s desired 
to make a chain of calculations involving about 13,000 figures ; 
and the test, to him, of the accuracy of his performance was to be 
found at its conclusion. By the time that the twentieth part of 
this task was done, the calculator must have felt the danger of 
error; perhaps he revised thus far on loose paper, found errors, 
and corrected them. Another division of the task was done in 
3 K 
VOL. VIII. 
