564 Sir george shuckburgh’s Obfervations 
The mean of thefe two forts of obfervations, made 
with different infer uments, is 2.43, viz. 1000 parts of 
the air at freezing become by cxpanlion from i° of heat 
Pts. Pts. 
equal 1002.43 or 1002.385 with the ftandard tempe- 
rature 39®. 7. Mr. de luces experiments reduced give 
Pts. 
this quantity equal 1002.23^ (fee Tranf.). It may 
be imagined, that I fhould have had a more accurate con- 
clulion by making thefe obfervations in greater dif- 
ferences of temperature than what is fhewn in the fe- 
cond column of the above table ; but it did not appear 
fo to me. On the other hand, I found that it was abfo- 
lutely neceffary that the lame heat fhould be kept up for 
fome hours together, in order that I might be fiire that 
the air within the inftrument, the glafs tube that con- 
tained it, and the air without it, all had acquired the fame 
(h) It has generally been fuppofed, that air expands with each degree of 
the thermometer, commencing from the mean temperature 55°; and, in confe- 
quence of this, aftronomers have computed tables for correcting their meaa 
refraCtions; but, upon reducing the refult of my obfervations to the temperature 
55°, we fhall have the correction, of the refraCtion for i° z= totVW or ttt* 
Now according to Mr. d_e luc this equation is totjVf — tst? which would 
produce a difference of about 4" in the corrected refraCtion, upon an altitude of 
S°j with the temperature 3.5 If my numbers may be fnppofed -to deferve equal 
confidence, ihe error of the tables in common ufe, in the above- circumftances, 
would amount to only half that quantity, and therefore probably will be 
thought fcarce worth correcting. I have mentioned this in order to obviate the 
tcquc Lu£q.&s that have been drawn by fome perfons from Mr.DE luc’s theory. 
uniform 
