XXIX .
Theory of the Parallaxes of Altitude for the Sphere , by Mr. F. Mallet , Professor and Astronomer at Upsal ; Translated from the French by M. Maty , M. D. R. S. Sec. Upsal , October 25 , 1766 .
L ET P be = the moon 's horizontal parallax , or 1 to sin. P , as the moon 's distance to the radius of the terrestrial sphere , on which the spectator is supposed to be placed .
Let A be the distance of the moon from the zenith , and p the parallax of altitude for the same distance .
The astronomers usually compute the value of p in the following manner : let sin. p=sin .
P. Sin. A , and p~ being found by the tables of logarithmic sines , sin. p~sin .
P. sin. A+p~ is found in like manner , p~ being assumed for the true parallax , which is not accurate .
Read Nov. 20 , 1766 .
~ 2 .
In order to show this , I have given another method of computing the parallax of altitude , as exactly as may be , by means of the common tables , in the following manner .
Since sin. p=sin .
P sin. A+p , we have sin. p=sin .
P sin. A cos. p+sin .
P cos. A~ sin. p , or sin. p ( 1 Sin. P cos. A ) = sin. P sin. A cos. p ; hence tang .
p=sin .
P sin. A/ 1-sin .
P cos. A. This formula seems a little difficult to be wrought in numbers , but it is as easy as the above one ; for , supposing sin. B2= sin. P cos. A , the tables will give the angle B , and tang .
p=sin .
P sin. A/ cos. B2 , the computation of which can give no trouble .
Hence it appears , that the calculus for finding the true parallax is not more difficult than that , which gives the said parallax with an error , the value of which is unknown ; for it is evident that the above computation for finding p~ is only an approximation , and that , to make it accurate , it would be necessary to carry it still on by finding sin. p~ = sin. P sin. ( A+p~ ) , and afterwards sin. p~ =P sin. ( A+p~ ) &amp;c. ~ 3 .
I therefore think myself in the right to prefer my method to that hitherto used by astronomers .
To confirm my opinion , I made a trial , by putting P=59~ and A=30~ , and found p-p~ = 0~ , 43 , in which the error of the usual computation amounts to near half a second ; I therefore give the preference to the geometrical calculus .
~ 4 .
Before I quit the formula tang .
p= sin. P sin. A/ 1-sin .
P cos. A , I must observe , that the computation of p may be executed by other methods to the same exactness .
If we take cos. 2 C=sin .
P cos. A , we shall have tang .
p=sin .
P sin. A/ 2 ( sin. C ) 2 , and the computation of this new formula is extremely easy .
~ 5 .
The formula tang .
p=sin .
P sin. A/ 1-sin .
P cos. A , gives besides , sin. p=sin .
P sin. A/ ~1+sin .
P2-2 sin. P cos. A ; make sin. P=2 cos. D , D being a given angle , of which we may have tables ready made , and we shall have sin. p=sin .
P sin. A/ ~1+2 sin. P ( cos. D-cos .
A= sin. P sin. A/ ~1+4 sin. P sin. A+ D/ 2 sin. A D/ 2 ; since cos. D-cos .
A=2 sin. A+ D/ 2 sin. A D/ 2 .
This being found without any logarithmic computation , we shall find tang .
E2 =4 sin. P sin. A+ D/ 2 sin. A D/ 2 , if A~ D , and hence we may easily compute sin. p= sin. P sin. A cos. E ; but if A~D we shall find cos. F2 =4 sin. P Sin. A+ D/ 2 sin. A D/ 2 and hence sin. p=sin .
P sin. A/ sin. F. ~ 6 .
Similar formulae may be found for cos. p , but as the angle p is pretty small , one might easily fall into some error by the usual tables of logarithms .
I shall not say what would be the amount of this error of p , having furnished the manner of avoiding it ; but this remark has not , I think , as yet been made in astronomical calculations ; and I have found it of great consequence in computing eclipses , where the distances to be found are very small arches .
~ 7 .
It may moreover be observed , that if A= D , sin. p= sin. P sin. A ; hence p~ =p in the same case , and p~ ~p , which seems very odd ; but the moon then is below the sensible horizon .
Theory of the apparent Diameters of the Moon .
~ 1 .
First the expression of horizontal diameter of the moon , or of the diameter seen at the horizon , seems to me too vague ; for one ought to understand by it the diameter seen at the center of the terrestrial sphere , rather than the apparent diameter at the horizon , which is not affected by refraction .
Without this , if the one was confounded with the other , an error would arise for the latitude of Paris from 0~,25 to 0~,32 .
~ 2 .
Let us keep the same denominations of P , p , and A , and call D the apparent semi-diameter of the moon at the centre of the sphere , and d= the apparent semi-diameter of the moon at the zenith distance =A .
We shall have sin. A : sin. A+ p : : tang .
D : tang .
d , or if one will , sin. A : sin. A+ p : : D : d : the error not exceeding an 100th part of a second .
~ 3 .
We had above sin. p=sin .
P sin. A+ p. Hence sin. P sin. A : sin. p : : ( tang .
D:tang .
d ) : : D:d , or because sin. p=cos .
p sin. P sin. A/ 1-sin .
P cos. A , 1 sin. P cos. A : cos. p : : D , d , and d= D. cos. p/ 1-sin .
P cos. A. ~ 4 .
Mr. Euler , in the Memoirs of the Academy of Berlin , 1747 , page .
175 , makes this same value = V/ 1-p2 sin. h , and according to him , V=D .
M= sin. P sin. ~=cos .
A+ p ; from whence it appears , that the true value of the apparent diameter of the moon , is not more difficult to be computed than the approximated one of Mr. Euler , the exact and geometrical formula being tang .
d= tang .
D cos. p/ 1-sin .
P cos. A and that of Mr. Euler d= D/ I-sin .
P cos. A+ p ; for in both , the values of D , A and p must be employed .
~ 5 .
It likewise appears to me , that since cos. p/ 1-sin .
P cos. A =sin .
p/ sin. P sin. A and therefore tang .
d= tang .
D sin. p/ sin. P sin. A , astronomers ought no less to employ this last formula , than any other more troublesome , in practical computation .
The simplest is tang .
d= tang .
D sin. A+ p/ sin. A , upon the supposition of an exact table of the parallaxes of altitudes ready made ; and I believe it will be as easy to compute with tangents as with arches , by means of logarithms ; and therefore this simplification in putting arches instead of tangents is unnecessary .
~ 6 .
To try the consequences of this theory , I made A=30~ , D=15~ , and taking the vertical of Upsal to the terrestrial axis for the radius of the sphere , I found P=55~ , 10~,3 , supposing that the axis of the earth , is to the diameter of the equator as 199 to 200 , and by the formulae tang .
d= tang .
D sin. A+ p/ sin. A= tang .
D cos. p/ 1-sin .
P cos. A = tang .
D sin. p/ sin. P sin. A , I found d=15~ , 12~.664 , but by the formula d= D cos. p/ 1-sin .
P co~ .
A , I had d=15~ , 12~ , 675 .
and lastly by that of Euler d= D/ 1-sin .
P cos. A+p we have d=15~ , 12~,635 ; from whence it appears that the error is very small , but that with the same trouble one may avoid any error whatsoever .
~ 7 .
The present case did not give an error of ,001 in substituting 1 or the radius instead of cos. p. Hence I conclude that d= D/ 1-sin .
P cos. A will be a more exact formula than that of Euler d= D/ 1 sin. P cos. A+p ~ 8 .
By taking d= D/ 1-~ .
P cos. A , we have d-D =D sin. P cos. A/ 1-sin .
P cos. A=D sin. p cos. A/ sin. A co~ .
p=D tang .
p/ tang .
A , which affords an elegant theorem , to find the increase of the apparent diameter of the moon .
~ 9 .
I have found others by the following methods .
Since sin. A : sin. A+ p : : tang .
D : tang .
d , and sin. A : sin. A+p sin. A : : tang .
D : tang .
d tang .
D : : sin. D cos. d : sin. d D ; but cos. D=cos .
d without any sensible error , and sin. D cos. D= 1/ 2 sin. 2 D , and sin. A+p sin. A=2 sin. 1/ 2 p cos. A+ 1/ 2 p , we shall have sin. dD= sin. 2D sin. 1/ 2 p cos. A+ 1/ 2 p/ sin. A. In the same manner , as I before found sin. p~ = sin. P sin. A and sin. P sin. A : sin. p : : tang .
D : tang .
d , hence sin. p~ : sin. p sin. p~ : : sin. p~ : 2 sin. p-p~/ 2 cos. p+p~/ 2 : : 1/ 2 sin. 2 D : sin. dD= sin. 2D sin. p p~/ 2 cos. p+ p~/ 2/ sin. P sin. A. ~ 10 .
Lastly let L= the distance of the moon from the center of the sphere , l its radius , that of the sphere being = 1 , we have 1 : L : : sin. P:1 and L : l : : 1 : tang .
D or 1 : l : : sin. P : tang .
D=l sin. P ; hence l = tang .
D/ sin. ~ being once found , since sin. A : sin. A+ p : : tang .
D : tang .
d , and sin. A+ p : sin. p : : 1 : sin. P , we ~ have sin. A : sin. p : : tang .
D : sin. P tang .
d : : l : tang .
d=l sin. p/ sin. A. I found the logarithm of l= 9.4343965 at Upsal , by putting 10 for that of the radius of the sphere determined as before .

