io 
sidgaalllt in the United States. 19 
* 
neous form, and has been used by Prof. Loomis to show the im- 
perfections of my observations in a “striking light.” It appears, 
on page 89, first as — 42’, second corrected to —46/.5, and again, 
“ By far the greater error here, is —46’.5, which I obtained from 
the first observation.” Here the whole affair i is of no account, 
because the datum on which the calculation was founded, had no 
real existence. ‘The same observation has been corrected by my- 
self, and was reprinted in the January number for 1841, page 150. 
This typographical error should have been detected, as such, from 
its inconsistency with the other items of the same grou 
Prof. Loomis assumes in the next paragraph, that the differen- 
ces of the readings with the face of the compass east, and with 
it west, if they exceed a certain constant quantity by him calcu- 
lated, and called “twice the zero error of the instrument,” are 
“errors of observation,” and upon this assumption, makes two 
tables of errors consisting of 88 items each, at the head of which 
stands the “ typographical error.” Here Prof. Loomis has fallen 
into the prevailing sin of mathematicians, the hasty assumption 
of data, which being granted, they can kill an army “by com- 
putation.” The proper zero error of tw instrument arises from 
a want of exact adjustment of the spirit lev 
graduated circle, and would be a co: 
also another source of error, arising 
ment of the agate planes on which the 
supported ; both of which errors are more or less merged by the 
reversal of the instrument, as they would be plus with the face of — 
the instrument in one position, and minus with it in the opposite 
position. Now this last error is a variable one; being dependent 
upon the total magnetic force, at any place, which force is va- 
riable. Probably Prof. Loomis intended by * zero error” ithe 
conjoint effect of the relation of both the zero of graduation 
and of the agate planes to the spirit level. A nice calculator, 
especially when he is pointing out the errors of other people’s 
labors, should have made a distinction between them, for one is 
a constant quantity, and the other a variable one. But Prof. 
L. has treated the whole asa constant. He has also assumed 
mechanical perfection in the parts of the instrument in contact, 
the pivots and the agate planes, a condition which never exists 
in fact. These vacillations in the various reversals which Prof. 
Loomis has tabulated, were of course known to me, and were 
always compared during the observations. They were apparent 
one = 
¥ 
