160 Mr. Schroeter's new Observations 
from A to B, and from D to E, could not, by my calculation, 
amount to more than 0,63 second. 
In general, such accuracy of computation avails nothing, 
because the observations and measurements of such a very faint 
and always decreasing light, cannot be so very exact. This is 
particularly shewn, and strikingly enough, in the two measure- 
ments of 20th May, 1793, given in the paper of my opponent ; 
where the projection of this crepuscular light, taking the appa- 
rent diameter of Venus = 60", was one time 12", 5, and the 
other time only 7", 7. And so much the more unimportant is 
it in the result of my calculation, that I assumed the crepus- 
cular light as having been measured from A. But that in my 
way of measuring, in which the penumbra is abstracted by the 
observation itself, I have been happier and more accurate, is 
testified by the computations to be given presently of my two 
measurements of the years 1790 and 1793, which were made 
under different circumstances, and yet correspond uncom- 
monly well. 
(b) The second objection, that I have measured the projection 
of the twilight too small, is equally unfounded ; for 
(«) The projection found by the author must properly be 
somewhat larger than mine, because he did not, like me, mea- 
sure the magnitude B C, but A C, fig. 20 ; and 
(/ 3 ) It will appear from the following computations, that I 
have found it at least as large as he did, without reckoning 
in the difference from A to B. He did not consider, that three 
years before I had observed under other circumstances, which 
must make the extent of the crepuscule appear less; and in gene- 
ral I do not perceive how he can form such a judgment from his 
