We show that a 2D projection is representative of its corresponding 3D
distribution at a confidence level of 90 % if it follows a King profi
le and if we consider the whole spatial distribution. The level is sig
nificantly lower and not decisive in the vicinity of the 2D cluster ce
nter. On the other hand, if we verify the reciprocal statement of the
Mattig's distribution (1958) -i.e. a flux limited sample is represente
d by a 0.6 slope of its count law-, we point out that, due to the usua
l inaccuracy of the slope determination, a slope of 0.6 is not a suffi
ciently strict criterion for completeness and uniformity of a sample a
s often used in the literature.