Abstract
A classification of the total analytical variation into components which can be separated and estimated experimentally, is suggested. Furthermore, when studying time dependent biological variation, it is recommended, if possible, to use an experimental design which prevents a confusion of biological and analytical variation.
Key Words: