A comparative study of the calipers GPM (DKSH, Switzerland), Holtain (Holtain Ltd, Great Britain), the caliper designed by V.E. Deryabin, and the sliding caliper GPM (Martin type) was carried out by the four measurers. The study was conducted using a metal caliper checking gauge GPM, a ‘sandwiches’ of soft silicone rubber kSil™ GP250 (Silicon Engineering, UK), as well as by measurements of subscapular, triceps, suprailiac, and calf skinfolds utilized in the Heath-Carter somatotype assessment scheme, in 20 adult volunteers (10 women and 10 men). Results and discussion. When measuring the caliper checking gauge, the V.E. Deryabin’s and the sliding caliper were accurate, the caliper GPM slightly underestimated (by 0.2 mm), and the caliper Holtain overestimated the exact values (by 0.4-0.6 mm). When measuring ‘sandwiches’ of silicone rubber, the GPM and Holtain calipers showed small but statistically significant differences between measurers (up to 1.4 mm), and for the V.E. Deryabin’s caliper, these differences were more pronounced (up to 5 mm). With a sufficiently high reproducibility of skinfold data, the mean values of their total thickness for the GPM and Holtain calipers differed significantly, and for the V.E. Deryabin and GPM calipers, there were no significant differences. The measurement data using the sliding caliper were consistent with those for the Holtain caliper but showed the largest variation of data between measurers: the maximal difference of the total skinfold thickness averages was 5.2 mm and 7.1 mm in the female and male subgroups, respectively, or 10-14% of the total averages. Conclusion. The results obtained support the need for mutual comparison of skinfold calipers and control of technical measurement errors to ensure data comparability. Cross-calibration of skinfold calipers is a necessary procedure that should precede the anthropometric study. When measured by an experienced anthropologist, the technical error of measuring skinfolds with a sliding caliper can be reduced, but this instrument is not recommended for training and subsequent use for skinfold measurements due to the high risk of obtaining inconsistent data.