(1a) 
(1b) 
(1c) 
(1d) 
(1e) 
(1f) 
(2a) 
(2b) 
(2e) 
(2d) 
(2e) 
(3a) 
(3b) 
  
  
  
    
  
  
    
    
      
    
     
    
   
   
       
       
    
     
   
    
- 
As previously shown [1] the standard error of the Y-parallaxes is obtained 
ns Vo] (5) 
and the standard error of A 6 
® = : ve (6) 
Of the weight numbers there recur in the following 
2, see)” (me) 
  
  
  
  
Q : (7a) 
4,9, - A? 022 3b^ £^ 
2. 2 
a, COMCETEST )? (2a^-3ae42e?) , are)“ (a°+0°) (a? 4e?-3£?)? d.a bn) 
b 2 2 5 b 
249%, 12 £°g°h 5 1 "g^ n? (a^4e?) of? 7 
A 
Q = (7c) 
Pa 7, "2 
a s ne a(a°1e°) | (ase) (afe?) (pa^-6desse?) _ , ,2 tre) (78) 
w, W, atet(a-e)* | 3(6%e) 5 £2 
Of the correlation numbers there recur only 
Q =0 (8) 
9, w, 
Of special interest is the accuracy in determining w (see chapter 2.2. below). The standard error in 
this element is obtained as 
Sw tA Vw (9) 
3 
The variation in magnitude À for various values of e is demonstrated in fig 2. For e = 0.5d we get 
  
the minimum value Su = 5.774, The same value for s,, is also obtained for e - - 0.26 d. 
2.2 The practical performance of relative orientation 
Relative orientation is carried out in a conventional way by utilizing points 35, 31, 15 and 11. Subse- 
quently adjustment of W is made with overcorrection, utilizing point 91 or 95, 
  
  
i ) - ins = = = = i = = b 
If in formulas (2a) - (2c) we insert Py) = Pig = P33 = Pag 0 and Pay = Pas = Py We o tain 
P d^-3de-2e^ (are) (d^4e^) (2e-d) 
aw 2.1 -3de-2e“  ,  (d«e) 2 (10) 
2 8 3h 31h 
and al, = à ÿ, = 0 and also dby, # O and dbz, #0. 
Note: If in formulas (2), (4) and (7) we put e = - d, the earlier derived formulas are obtained. See e.g. 
[1] , formulas (16) and (17) and table (2).