Mr. Ivory on the Astronomical Refractions. 419 



have stated at p. 490 P. T. for 1823. But no mention is 

 made of the reason 1 have there assigned. The sums of the 

 exterior and interior thermometers are respectively 567*6 

 and 669-9; and the difference is 102°-3, being 8° for each 

 star. Now this is a very great difference, and undoubtedly 

 leaves a very considerable uncertainty about the real tempe- 

 rature of observation. When we consider that, at the low 

 altitudes in question, every degree of the thermometer cor- 

 responds to an alteration of 2" or 3" in the star's place, I am 

 persuaded that the opinion I have advanced in my Paper re- 

 lative to Mr. Groombridge's observations will be found to be 

 just. They are not a proper test of the accuracy of a Table 

 of refractions on account of the uncertainty of the tempera- 

 ture. Every shifting of the standard of the Table must ap- 

 proach nearer the truth, or recede further from it, at the rate 

 mentioned. 



I shall now compare my Table with the same observations, 

 on the supposition that the temperature is found more nearly 

 when estimated, not by the exterior thermometer, but 2° 

 nearer the interior one ; or, which is the same thing, I shall 

 lower the temperature of the Table from 50° to 48°. I have 

 also ascertained, by sound principles as I shall show below, 

 that no confidence can be placed in my Table, nor in any 

 other at present in existence, passing 88^° of zenith distance. 

 Nearer the horizon the errors of every Table increase at a 

 prodigious rate. Now, leaving out the last star and estimat- 

 ing the temperature as I have said, the errors of the first 

 12 stars by my Table are + 12"'5 and — 10"-6, or 23"-l. The 

 error of the last star alone is 17"; making the sum of th^ 

 errors of all the 1 3 stars equal to 40". I wonder the inge- 

 nuity of the S. B. L. did not suggest some plausible argu- 

 ment for changing my Table from 50° to 52°, which would 

 have given him a decided advantage. 



The Table in N. A. is one of no authority. No astrono- 

 mer in Europe, except the author, can tell how it was con- 

 structed. Its character rests upon a dictum. We know that 

 there is no correct theory. I will not again venture to say 

 that the formula is empirical for fear of the consequence ; but 

 it is partly empirical, as the author has admitted. The Table 

 must therefore have been constructed either with the as- 

 sistance of other Tables already in use, or by means of ac- 

 tual observations. Now, as no original observations are 

 produced, except one by Mr. Pond, it is very probable that it 

 was constructed in the first of the ways I have mentioned. 



That it may not be thought I am speaking at random, I 



now affirm that the Table in N. A. is the same with that in the 



3 G 2 Finidamcnfa 



