The results for the direct anthropometry-moiré differences in h, are 
tabulated in Table I, whereby Up designates henceforth the mean of the dif- 
ferences (in percentage of the total), 0p their standard deviation, M their 
range; and u, c the mean and the standard deviation of the direct anthropo- : 
metric measurements. 
Table I 
Direct Antropometry - Moiré Height hi Differences 
  
  
1 3 4 5 
u (cm) 9.88 9.81 8.10 6.10 5.93 
og (cm) +0.70 +0.92 +1.20 +0.86 +0. 64 
up (2) -2.2 +0, 7 +1.9 -1.2 -1.1 
0p (%) 12.1 11.6 21.6 12.3 12 2 
Mh (%) -7.1- +3.8 -3.6-43.7 -0.8-+6.0 -5.9-+3.7 -7.1- 12.3 
  
Errors of a systematic nature can be detected in most cases. This is 
attributed to the approximate character of the calculation formula used (i.e. 
Eq.(1) instead of Eq.(2); see Fig. 2). 
It can be shown that (since the vertical leg-lens relation was practi- 
cally the same for all subjects, the legs of which have similar slope proper- 
ties) the moiré calculation tends to systematically overestimate hy» ha» and. 
he, and underestimate hs. The above, caused by central projection, conform 
with the results of Table I and evidently hold true only for the relative 
leg-camera position used here (schematically shown in Fig. 2). These consi- 
derations were not pursued any further, e.g. by using a mean correction 
factor for each hs as the opposite-signed differences compensate partly for 
each other (overall mean difference=-0.4%; cf. Table I). 
Girth Measurements 
  
The leg sections were considered as circular, and circumferences were 
calculated as Ci7TD, . Diameter calculation was based on measurements through 
the marked points in a direction parallel to the image x-axis (hence no 
longer perpendicular to the bone axis). The slope at the leg outlines is 
obviously steep and the fringe density high. This may have introduced small 
errors in the estimation of the diameter to lens distances whichare indispen- 
265