the center they are slightly transparent, in the center they 
are, in general, not. The thickness of the cross lines was 
somewhat -less than the diameter of the measuring mark. The 
differences between double readings of negative crosses of 
one plate are given in Fig. 3la, the histograms for the x- 
and y-differences of four measured plates in Fig. 3!b and c. 
The mean values are -0.09 and -0.65 yum and the measuring 
accuracies +1.77 and £2.21 um respectively for n = 2089. The 
same observations as before can be made. The measuring 
accuracy ls again better for the x-values. 
The differenees between double readings of positive 
reseau crosses of one plate are shown in Fig. 35a, the 
histograms for the readings from four plates in Fig. 35b 
and ec. Another decrease in measuring aecuracy is notieed: 
m. £2.18 um and s 12.64 um. Again, m. is smaller than 
mg, this in spite of the fact that the measuring mark was 
first set into the horizontal wing seen to the left of the 
center of the cross and then, without changing the y-setting, 
moved into the center of the cross [69]. Apparently, the 
observer's eye is more sensitive to setting a measuring mark 
on a vertical line than to setting it on a horizontal line. 
This may be the result of the fact that only one pair of 
muscles is necessary for horizontal movement of the eye 
compared to three pairs for vertical movement. The mean 
values of the 2112 double readings of positive reseau crosses