On the proper use of point-counting and semi-automatic procedures in stereology.

The influence of shape and size of structures as well as the individual effect on accurate results and the speed of evaluation in point-counting and automated planimetry is investigated. The following statements can be made: 1) Planimetry is with the factor 2 to 5 more exact than point-counting. 2) The individual effect on the results is small in planimetry and therefore can be neglected in most cases. However, this is impossible in point-counting. 3) The evaluation-error increases with the decrease of size. It goes over 5% below an area of 40 mm2 in planimetry. The comparable border-values of point-counting are much higher. 4) The shape has little influence on the results in structures with smaller deviations from the form of a circle (stretching-factor below 2.7), but with increasing complexity of the borders, the results show more variability. 5) The border-line examination in planimetry is 5 to 10 times more exact than in point-counting. 6) The time used for equal samplings including calculation of MW and standard deviation is similar for both procedures, but planimetry only needs few measuring steps for small confidence-limits and is able to estimate more parameters in one estimation. 7) The orientation of structures has a planimetry no influence of the results obtained for the single section, in contrast to point-counting of intersections. 8) The psychic condition has influence on both procedures.