Assessing the Level of Uncertainty of small samples of Multidimensional Biological and Biometric Data

It is shown that small selections of biometric data of the person have essential uncertainty which needs to be considered at multidimensional statistical estimates. Within a hypothesis of normal distribution of the bio-these the table of errors of calculation of a population mean and a mean square deviation as functions of number of examples in test selection is constructed. The table can be used for an assessment of an error of the received results of biometric measurements, and also for a reasonable variation by data in attempts of artificial increase in the amount of small selections by crossing of examples parents and receiving examples descendants. It is claimed that at the small volumes of basic data their uncertainty is a source of mutations, more powerful, than recommended by the standard.