a 4 
at 
for computing the Astronomic Refraction. 31 
the first to the second there was an interval of eight days, and of 
eleven from the second to the third. The barometer has 
searcely varied at all, and the thermometer has varied but little, 
yet the refraction fa varied 4 minutes at these zenith distances. 
According to our last tables the refraction changes from 11” to 
12” for each minute of variation in the zenith distance. ‘That 
of Bradley and of all the other astronomers varies from 10” to 
11”. Supposing 11” of variation and reducing all these re- 
fractions to the astronomic horizon, that is to 90° of apparent 
zenith distance, we shall have for the horizontal refraction, 
R52 
30 33 | 
30 33 . of which the mean is 32” 25”. 
Bios8. 3 
184 005° | 
340.42: 1} 
“ From the first to the second day we have a difference of 
3’ 19”, although the barometer aud the thermometer are the 
same. From the second to the third, the refraction has not 
varied, although the thermometer has got up 4 degrees. The 
two last days the refraction has only altered 3”, although the 
thermometer has got up 7°36 degrees. 
“« We cannot therefore be certain of the mean to 2 minutes, 
which agrees nearly with that of Cassini. It is hardly probable . 
that we shall ever be able to compute such anomalies; and 
what would they have been had I observed in winter ? 
« At 79” I could not make the observations of different days 
agree nearer than about 6” or 7” between the extreme values, 
© At 77° I have had variations of 10” and 11”. 
At 79° they were 15”, 
* At 82° they differed as muchas 36”; that is, the table that 
I had constructed representing the observations of several days 
to 1” or 2” nearly, was found once in error —17”, and another 
time +19”. 
© At 84° I have been more fortunate, the error was one half less. 
* At 86° the difference between the extremes was 30”, 
¢ At 88° the errors, which were nothing for several days, in- 
creased then to + 15” and —20”. 
& At 89° from —15” to +30”.” 
‘The tables of Bradley and Mayer give still greater errors, so 
that it appears to me impossible to make any good table for 
these lower degrees, But from the zenith to 82°, we may have 
a number of tables nearly all equally good. 
**In the fifth book of Specola di Palermo by Piazzi, you may 
find a great number of observed refractions, I have recomputed 
all the calculations, and have found them very accurate. Dif- 
ferences at least equal to mine, may be remarked among them. 
