The differential staining methods for chromosomes have led to the demonstration of more chromosomal polymorphisms. Not rarely, these polymorphisms allow in autosomal trisomies the detection of parental origin of the supernumerary chromosome. In addition, the malsegregation may be ascribed to 1st or 2nd meiotic division in informative families. This approach of analyzing possible causes of trisomies is subject to a considerable bias. Trisomic phenotypes are twice as frequent for 2nd meiotic errors than for 1st meiotic errors. Also, rare chromosome variants seldom occur in matings where malsegregation in 1st meiotic division can be detected. In the present paper this bias is analyzed mathematically on the family as well as on the population level. From this mathematical analysis and from the data in the literature we conclude that Down's syndrome as a whole is caused about 5-10 times more often by a malsegregation in 1st meiotic than by an error in 2nd meiotic division. Mainly from experimental studies in rodents, causes for errors in 1st and 2nd meiotic division are becoming apparent. They are summarized in the context of the results of the present paper.