Skinfold measures are conducted extensively on athletes and on gymnasia clients, but the issue of calibration has received little attention. Early attempts at calibration of skinfold calipers concluded that a jaw pressure of ∼ 10 g·mm−2 was most appropriate for measuring skinfold thicknesses (1,10). However, this recommendation used a static (upscale) calibration with the caliper jaws opening that is opposite to the manner in which calipers are used to make measurements. Schmidt and Carter (14) demonstrated that the mean jaw closing (downscale) pressure of 10 new Harpenden calipers was 8.25 g·mm−2 for the range of 10–50 mm of jaw gap. After carefully dismantling, servicing, and reassembling Harpenden calipers, Carlyon et al. (3) suggested that new calipers should also be calibrated before use because of differences in the strength of the springs, the mechanical condition of the caliper pivot, the mechanical resistance of the indicator gauge, and the caliper jaw alignment. Each of these factors can affect the downscale jaw pressure and hence the skinfold readings obtained.
Using a relatively crude method of foam rubber blocks to simulate the density of skinfolds, Gore et al. (7) identified that calipers with upscale jaw pressures that varied by greater than 1 g·mm−2 produced significantly different skinfold readings taken on athletes. However, that study evaluated six separate skinfold calipers and, in addition to differences between the springs fitted to the calipers, differences between the mechanical condition of the pivots and indicator gauges of the various calipers may have confounded the results. Accordingly, this study was designed to isolate the effect of the springs on skinfold measurements by using only one set of serviced calipers fitted with six pairs of springs that produced different downscale jaw pressures as measured by a precision dynamic calibration rig (3). Observations over 5 yr of servicing different Harpenden calipers have also revealed that the spring coefficient (Δ force per unit Δ length) varies among springs, but the reason for the discrepancy is uncertain. Therefore, one aim of this study was to examine the notion that pairs of springs fitted to Harpenden calipers deteriorate as a result of the repeated jaw opening and closing operations that are a function of their use. The second major aim was to quantify the change in skinfold thickness per unit change in downscale jaw pressure on a standard pair of calipers. The third aim was to measure sufficient calipers to provide guidelines about the ideal downscale jaw pressure of Harpenden calipers.
Spring characteristics: repeated use of springs.
Five new Harpenden springs (British Indicators, West Sussex, United Kingdom) were used to evaluate the effect of at least 100,000 full excursions on the spring coefficient. Before use, the coefficient of each spring was measured using a 5-kg load cell (Scale Components, Brisbane, Australia) and a Vernier caliper (Mitutoyo, Japan) as the change in the force exerted by a spring per unit change in its length. The length of each spring was measured at both 29.40 and 37.24 N, since the spring length under such a force is typical of the length of a spring when installed in a Harpenden caliper and the jaws are closed or open to 40-mm jaw gap, respectively (3). A spring coefficient is a numerical description of Hooke’s Law of elasticity, which states that the force exerted by a spring is proportional to its length.
Each spring was then installed on a set of Harpenden calipers (British Indicators) whose jaws were mechanically opened by a DC motor and appropriate linkage. A cam on the motor shaft and the voltage applied to the motor were set so that the jaws were opened to 45-mm jaw gap in 1 s, closed back to zero jaw gap in 1 s, and held closed for 1 s. The number of openings was calculated by dividing this 3-s cycle into the time each spring was installed in the apparatus. When the spring had completed the prescribed number of openings, it was removed from the apparatus and the spring coefficient was remeasured. Five springs were measured before and after ∼100,000 excursions. One of these five springs was remeasured after 172,000 excursions and one after ∼500,000 cycles of opening and closing.
Spring characteristics: springs of different ages.
A single set of Harpenden calipers was serviced to minimize friction in the pivot joint as well as the in the jaw gap indicator gauge (3). This caliper was used to evaluate the dynamic downscale jaw pressure exerted by six pairs of Harpenden springs using the calibration technique of Carlyon et al. (2). Two were new pairs of springs (N1 and N2), and two were “old” sets of springs (O1 and O2) that were distorted into a “C” shape (Fig. 1) and had been used in a university teaching laboratory for approximately 25 yr. The remaining two pairs of springs (S1 and S2) had been used on calipers for less than 1 yr, but the spring coefficients were well outside the acceptable range of 1.10–1.15 N·mm−1, as suggested by Carlyon et al. (3). At the time of this experiment, springs N1, N2, S1, and S2 were of current manufacture while springs O1 and O2 were ∼ 25 yr old.
The six different pairs of springs (N1, N2, O1, O2, S1, and S2) were used to measure skinfold thicknesses at nine sites (triceps, subscapular, biceps, iliac crest, supraspinale, abdominal, front thigh, medial calf, and mid-axilla) identified using the protocols described by Norton et al. (12). Each of 20 subjects (9 male and 11 female; mean age ± SD of 22.9 ± 8.8 and 22.1 ± 8.4 yr for men and women, respectively) who provided written informed consent was measured using each pair of springs with the order of springs randomized and counterbalanced. At each site, three skinfold measurements were taken at 30-min intervals in an attempt to minimize the affect of repeated measurements. Before any data collection for this study, the investigator undertook formal training in anthropometry (8) and conducted duplicate measurements on 20 subjects to demonstrate precision. The technical error of measurement (TEM) (9) for duplicate measures at each of nine skinfold sites was less than 5%. This precision was improved for actual data collection, since the TEM calculated for the three repeat trials ranged from 2.6 to 1.5% for the nine sites.
Multiple caliper assessment.
Over a period of 3.5 yr, 78 pairs of new and old Harpenden springs were measured to establish the spring coefficients. The 78 spring pairs were then returned to their Harpenden calipers that had the pivot, gauge mechanism and jaw alignment checked and serviced (3). Dynamic downscale jaw pressure was then evaluated at 5, 10, 15, 20, 25, 30, 35, and 40 mm of jaw gap.
Paired Student’s t-tests were used to determine whether there was a change in the spring length at the two standard stresses of at 29.40 and 37.24 N. Three-way repeated measures analysis of variance was used to examine the individual skinfold thicknesses as a function of six different spring pairs, nine skinfold sites, and three trials. Tukey’s post-hoc test was used to determine significant differences between cell means. At each skinfold site least squares linear regression was employed to determine the unit change in skinfold thickness as a function of the dynamic downscale caliper jaw pressure. The “compressibility” at each site was also calculated as the percentage decrease in skinfold thickness per unit increase in jaw pressure relative to the original thickness. All statistical tests were conducted using Statistica for Windows software (Version 5.0, StatSoft; Tulsa, OK). Unless stated otherwise, results are expressed as mean ± SD. Significance was set at P < 0.05 for all analyses.
Spring characteristics: repeated use of springs.
The results of each of the five new springs are shown in Table 1. There was only 0.1-mm change in the length of spring 5 at 29.40 N and no change at 37.24 N after > 500,000 cycles of opening and closing (Table 1). When the five springs were treated as a group, there was no significant difference in the spring length before and after ∼100,000 cycles either at 29.40 N (t = −0.67, P = 0.54) or 37.25 N (t = −1.63, P = 0.18), and consequently there was no change in the spring coefficient.
Spring characteristics: springs of different ages.
The individual spring coefficients were 1.10 and 1.21, 1.14 and 1.19, 2.06 and 1.91, 1.82 and 1.82, 1.74 and 1.57, and 1.57 and 1.57 N·mm−1 for spring pairs N1, N2, O1, O2, S1, and S2, respectively. When used in the serviced Harpenden calipers, the mean dynamic downscale jaw pressure ranged from 8.02 to 9.03 g·mm−2 with each pair of springs recording small differences (up to 0.41 g·mm−2) in pressure for jaw gaps from 5 to 40 mm (Table 2).
The interaction between skinfold site and spring pair was significant (F(40,760) = 1.93, P < 0.001), and there was a consistent trend at most of the nine skinfold sites with the lowest thickness for S2 which exerted the highest jaw pressure, and conversely the highest thickness for N1 which exerted the lowest jaw pressure (Fig. 2). Selected post-hoc tests revealed that the mean skinfold thickness measured by N1 (mean downscale jaw pressure 8.02 g·mm−2) was significantly higher than that for S2 (mean downscale jaw pressure 9.04 g·mm−2) at all nine skinfold sites. However, when the difference between the jaw pressure was lower, the skinfold thickness measured by N1 was significantly higher than that measured by O2 (mean downscale jaw pressure 8.44 g·mm−2) for only the abdominal, iliac crest, and mid-thigh sites.
The two-way interaction between the six spring pairs and the three trials was also significant (F(10,190) = 4.37, P < 0.001). Although for springs O1, O2, S1, and S2 there was no difference in skinfolds between trials, in contrast the third trial skinfold was higher than the first trial skinfold for both N1 and N2 by 0.19 and 0.14 mm, respectively. The second trial (14.53 mm) was significantly higher than the first trial for N1 (14.40 mm).
The mean skinfold thickness as a negative function of mean jaw pressure yielded a multiple r2 of 0.89 or greater for each site except abdominal where the corresponding value was 0.79. The linear model for the smallest site (bicep = 6.2 mm) was: skinfold thickness = −0.85 × jaw pressure + 13.4, and for the largest site (mid thigh = 22.0 mm) was: skinfold thickness = −2.2 × jaw pressure + 40.9). That is, greater jaw pressure reduces the absolute skinfold thickness nearly 2.5 times more for a thigh skinfold than for a bicep skinfold. However the “compressibility” per unit increase in jaw pressure (from 8.0 to 9.0 g·mm−2) relative to the original thickness at 8.0 g·mm−2 was 12.7% for biceps and 9.6% for midthigh. Consequently, when the sum of nine skinfolds was calculated there was a significant difference between each of the different pairs of springs (F(5,95) = 157.8, P < 0.001) with largest Σ9 skinfold totals with springs N1 and N2, and the lowest totals with springs S1 and S2, such that the Σ9 for N1 (129.8 ± 9.1 mm) was approximately 10% higher than that obtained for S2 (114.1 ± 8.3 mm) and that for O2 was intermediary (120.4 ± 8.6 mm).
Multiple caliper assessment.
At loads of 29.4 and 37.2 N the length of the 156 springs were in the range of 60.2–69.5 mm, respectively. The corresponding spring coefficients ranged from 1.10 to 1.48 N·mm−1, with a mean of 1.13 ± 0.06 N·mm−1. The mean downscale jaw pressure averaged over all jaw gaps from 5 to 40 mm was 7.82 ± 0.25 g·mm−2 and mean values for each jaw gap are presented in Table 3. The mean jaw pressure at 5 and 40 mm of jaw gap was 7.61 ± 0.21 and 8.03 ± 0.18 g·mm−2.
One major finding of this study is that up to 500,000 cycles of opening and closing new Harpenden springs, equivalent to 10 yr of heavy use, does not alter the spring coefficient. This suggests that the likely cause of spring deterioration is environmental, such as storage in an humid environment or a result of handling the springs which may cause mild corrosion accelerated by perspiration on the fingers and hands. The second outcome is that increasing the mean downscale jaw closing pressure from 8.0 to 9.0 g·mm−2 reduces a skinfold thickness by approximately 10%, independent of the site of the skinfold. Finally, guidelines for the ideal dynamic downscale caliper jaw pressure to allow comparison of skinfold measures conducted with different Harpenden calipers are discussed.
Spring characteristics: repeated use of springs.
This study demonstrated that up to 500,000 cycles of opening and closing new Harpenden springs, equivalent to 10.6 yr of heavy use (10 subjects per day, 5 d·wk−1, with duplicate measures at nine skinfold sites), does not alter the spring coefficient and consequently does not alter the jaw closing pressure of a serviced pair of calipers. At first appearance the data from the 25-yr-old Harpenden caliper springs (O1 and O2), indicate that the spring coefficient increased rather than decreased with age. In these old springs there were a number of coil sections at each end that appeared to have fatigued and had lost their elasticity (see Fig. 1). The fatigued sections reduced the effective spring length and thereby increased the spring coefficient such that when installed on the calipers they produced a greater jaw pressure than new springs N1 and N2.
The 1-yr-old springs (S1 and S2) were not distorted and had spring coefficients that were 1.5 times higher than that of new springs N1 and N2, which indicates that a greater force is required to generate a unit change in spring length. Rather than fatigue with 1 yr of use, the increased spring coefficient suggests that differences in the spring metal manufacture process can alter the spring coefficient. Therefore, to obtain accurate skinfold measures with different Harpenden calipers the spring coefficients of each new spring should be verified to fall in the range of 1.10–1.15 N·mm−1 as suggested by Carlyon et al. (3) and which corresponds to a range of ∼5%. The data from this study of 78 spring pairs also achieved spring coefficients within this range. Although one spring of each pair of N1 and N2 are ∼0.05 N·mm−1 outside the recommended range for the spring coefficient, this will not alter the conclusions about the changes in skinfold thickness relative to jaw pressure as discussed below. Rather, this limitation is a consequence of more stringent criteria for the ideal spring coefficient as more springs were measured after collection of skinfold measures had begun.
A further observation made during the servicing of Harpenden calipers has been that new springs installed on calipers can tarnish after as little as 3 months of use, whereas new springs from the same batch can remain untarnished for over 12 months while in clean and dry storage. In particular, the tarnishing is limited to the exterior surface of new springs, which are handled with normal use, but not the interior spring surface next to the caliper jaws. On the other hand, springs regularly wiped over with a lightly oiled cloth from one pair of calipers used to take ∼ 3000 measurements per year have remained untarnished on both internal and external surfaces for 4 yr. These springs have shown little change in their spring coefficients over the same period; initial versus final spring coefficients were 1.14 and 1.14 N·mm−1, respectively, for one spring and 1.12 and 1.10 N·mm−1, respectively, for the second spring of this spring pair. The pattern and time course of tarnishing a new spring suggests that degradation of the spring metal may result from external contaminates such as perspiration from the fingers and hands. This study examined directly the prolonged use of springs on calipers that were not handled during the test and dispels the notion that repeated use in the range of 0–45 mm of jaw gap causes metal degradation. Based on the results of repeated spring use up to 500,000 openings and observations of spring corrosion from handling, it is therefore recommended that new caliper springs should be regularly wiped with a lightly oiled cloth and that calipers should be stored in a clean and dry environment.
The design of the Harpenden caliper is based on two springs to provide the closing moment that produces the jaw pressure about the skinfold. This study has demonstrated that there is an inverse linear relationship between skinfold thickness and downscale jaw pressure, with an approximate 10% decrease per 1.0 g·mm−2 increase in jaw pressure in the range of 8.0–9.0 g·mm−2. The jaw pressure of a pair of springs is in turn directly proportional to the spring coefficient; provided that the mechanical condition of the caliper pivot and indicator gauge is good. Contrary to the studies of Edwards et al. (6) and Gore et al. (7), which investigated upscale jaw pressure, this study has quantified the effect of a unit increase in downscale jaw pressure on individual skinfold thicknesses. Downscale jaw pressure is most relevant since this is the method by which skinfolds are taken and an ideal calibration procedure should mimic the use.
In 1990 Schmidt and Carter (14) found a mean downscale jaw pressure for 10 new Harpenden calipers of 8.25 g·mm−2 for jaw gaps of 10–50 mm. The corresponding result for current study, which used a precision load cell rather than a spring balance, was 7.82 g·mm−2 for 78 pairs of springs for jaw gaps of 5–40 mm. The current study also indicates that a 10% decrease in a single skinfold and consequently in the total of multiple skinfolds for a 1.0 g·mm−2 increase in downscale jaw pressure. This raises the question as to the appropriate tolerance range for the jaw pressure required for accurate measures when different calipers are used to acquire skinfolds. Edwards et al. (6) suggested a range of 2.0 g·mm−2 for a jaw gap of 2–40 mm, but the results of the present study indicate that this range is too broad.
The caliper jaw downscale pressure range for comparison of results from different calipers depends in part on the precision of the person taking skinfold measures. As demonstrated in this study, a 5% change in jaw pressure would correspond to an approximate 5% change in skinfold thickness. Unpublished data for the intratester TEM of ∼ 520 anthropometrists trained according to the model of Gore et al. (8) indicate that individual skinfold sites have TEM of 5% or less and persons with more refined skills have TEM of 1–2%. In each of the 520 cases, the precision data were collected using a single pair of calipers which therefore removes the confounding effect of different calipers. The 95% confidence interval for a person with a 1.5% TEM is ± 4.2% (13). Therefore, there is a 99% likelihood that an anthropometrist with a TEM of 1.5% would obtain a false decrease in skinfold thickness using spring pair N1 versus S1 and a 95% likelihood of a false decrease with spring pair N2 versus O2. In contrast, the 95% confidence interval for a person with a 4.5% TEM is ± 12.7%, and consequently there is less than 68% likelihood that they would obtain different skinfolds with the different spring pairs of this study. Therefore, to achieve accurate results with different calipers, it is recommended that an anthropometrist with low imprecision should use Harpenden calipers with mean downscale jaw pressure of 7.61 ± 0.21 and 8.03 ± 0.18 g·mm−2 at a jaw gap of 5 and 40 mm, respectively. These recommendations are more relevant than the average value across all jaw gaps (7.82 g·mm−2) since gap specific data are directly comparable with that which is acquired during calibration. This range of approximately ± 0.20 g·mm−2 (or 1 SD) corresponds to approximately ± 2.5% of the mean jaw pressure.
The mean downscale jaw pressure across all jaw gaps was 7.82 ± 0.25 g·mm−2 and is 5.4% lower than that (8.25 g·mm−2) obtained by Schmidt and Carter (14). This suggests that skinfold thickness data from the current 78 Harpenden calipers may not be comparable retrospectively with earlier studies which used Harpenden calipers even if the older calipers were well maintained. However, the recommendations here for a narrow calibration range will enable direct and confident comparison of results from different Harpenden calipers used in any future study. While anthropometrists have previously assumed that data from different Harpenden calipers may be similar, the results of this study suggest that this has not been the case. Careful selection of caliper springs with appropriate spring coefficients and servicing of Harpenden calipers (3) is needed to improve the accuracy of routine skinfold assessment of adiposity.
One limitation of this study is that the significant difference (<0.2 mm) in skinfold thickness among trials for three of the 18 comparisons among the six springs suggests that the 30-min rest interval between trials may have been insufficient. However, compression of subcutaneous adipose tissue from successive trials was not an effect of order because the first rather than the third trial was the lowest skinfold thickness. More importantly, interpretation of skinfold thickness at each site as a function of different springs was not confounded by the two-way interaction between springs and trials. The skinfold thickness for the nine sites by six springs were the same for each of the three trials as indicated by the nonsignificant three-way interaction between springs, trials, and sites (F(80,378) = 1.15,P = 0.17).
Skinfolds provides a direct method to evaluate subcutaneous fat, and therefore a sum of skinfolds is suggested as suitable to track longitudinal changes in adiposity of an individual (11). A sum of 6–8 skinfolds is used to monitor subcutaneous fat in high performance athletes, and the sensitivity of this approach is illustrated by the example that 1 month of heavy training caused a fall from 34.7 to 30.6 mm for a cyclist (16). A sum of skinfolds has good face validity, but our data demonstrate a 10% decrease for a sum of skinfolds assessed with a jaw pressure of 8.0 compared with 9.0 g·mm−2. Notwithstanding the problems associated with age, sample, and sex specific equations to estimate body fat (11), it is relevant to consider the likely error in body fat when calipers that exert different jaw pressures are used. If the equation of Durnin and Womersley (5) is used to estimate body density in conjunction with the Siri equation (15), for males a skinfold total (tricep, bicep, subscapular, iliac crest) of 40 and 44 mm yields 18.2 and 19.5% body fat, respectively. Therefore, the 10% difference in skinfold sum that would be achieved with caliper S2 (40 mm) compared with N1 (44 mm) produces a minor difference in estimated percent body fat. The difference is minimized because in the Durnin and Womersley (5) equation the sum of skinfolds is multiplied by a coefficient of minus 0.0004 and added to 1.0988 to estimate body density.
It is falsely reassuring to realize that the 10% errors in a skinfold total can be masked by the mathematics used to estimate body fat. A skinfold total could be over- or underestimated if Harpenden calipers with different jaw pressures are used. A concerned anthropometrist who has a good technique may attempt to counter this problem by using a single pair of Harpenden calipers. Provided that the mechanical condition of the caliper does not deteriorate, they will achieve comparable data over time. However, minimizing measurement error of anthropometry requires not only a precise technique (as indicated by a low TEM for repeat measures) but also an accurate caliper. Servicing a Harpenden caliper (3) and selecting appropriate springs can provide confidence in the accuracy of all measures obtained. Equipment calibration is a fundamental prerequisite for accurate measurement and if several anthropometrists wish to pool their data, for instance, during a large scale survey (4), absolute calibration standards such as those recommended here are required.
A calibration procedure involves measurement of a series of known standards and comparison of the test values against the true values. Calibration is achieved either by adjusting the test instrument or by generating a “look up table” to provide suitable correction factors throughout the measurement range. Calibration of Harpenden skinfold calipers has been problematic since there has been no dynamic downscale “gold standard” available, although we have now identified that a narrow calibration range can be achieved by carefully selecting springs with appropriate spring coefficients and minimizing the frictional losses in the caliper pivot and indicator gauge. The following recommendations are relevant when using different Harpenden calipers or even using one Harpenden caliper since both old and new calipers can have frictional losses or springs that are either too weak or too strong. Our results in the current and previous studies (2,3,7) allow us to make the following tentative recommendations for a skilled anthropometrist to achieve accurate results with any Harpenden caliper:
- • Carefully service the caliper pivot, indicator gauge, and jaw alignment (3) approximately every 12 months, but at least every 24 months;
- • wipe the caliper springs after each day of use with a lightly oiled rag; and
- • select a pair of springs that each have a spring coefficient in the range of 1.10–1.15 N·mm−1.
In a mechanically well-adjusted caliper, these springs should yield a mean downscale jaw pressure of 7.61 ± 0.21 and 8.03 ± 0.18 g·mm−2 at a jaw gap of 5 and 40 mm, respectively.
This study demonstrated that deterioration of new Harpenden springs does not occur after up to 500,000 cycles of opening and closing, which is equivalent to 10 yr of heavy use. This suggests that the likely cause of spring deterioration is environmental, such as storage in an humid environment or a result of handling the springs which causes mild corrosion accelerated by perspiration on the fingers and hands. Routinely oiling the caliper springs may ameliorate spring deterioration. The second outcome of this study is that increasing downscale jaw closing pressure from 8.0 to 9.0 g·mm−2 reduces a skinfold thickness by approximately 10%, independent of the site of the skinfold. The third finding is based on assessment of 78 pairs of Harpenden calipers that had been serviced (3); to obtain accurate skinfolds independent of different Harpenden calipers, it is recommended that the mean downscale caliper jaw pressure should be in the range 7.40–7.82 and 7.85–8.21 g·mm−2 at jaw gaps of 5 and 40 mm, respectively. However, jaw pressure should only be assessed after confirming that the individual spring coefficient is in the range 1.10–1.15 N·mm−1.
1. Behnke, A. R. and J. H. Wilmore. Evaluation and Regulation of Body Build and Composition. Englewood Cliffs, NJ: Prentice Hall, 1984, pp. 1–236.
2. Carlyon, R. G., C. J. Gore, S. M. Woolford, and R. W. Bryant. Calibrating Harpenden skinfold calipers
. In: Anthropometrica, K. I. Norton and T. Olds (Eds.). Sydney, Australia: University of New South Wales Press, 1996, pp. 97–118.
3. Carlyon R., R. Bryant, C. Gore, and R. Walker. Apparatus for precision calibration
of skinfold calipers
. Am. J. Hum. Biol. 10:689–697, 1998.
4. Carter, J. E. L. and T. R. Ackland. Kinanthropometry in Aquatic Sports
. HK Sport Sci. Monograph Series. Vol 5. Champaign IL: Human Kinetics, 1994, pp. 1–184.
5. Durnin, J. V. G. A. and J. Womersley. Body fat assessed from total body density and its estimation from skinfold thickness: measurements on 481 men and women aged 16–72 years. Br. J. Nutr. 32:77–97, 1974.
6. Edwards, D. A. W., W. H. Hammond, J. M. Healy, J. M. Tanner, and R. H. Whitehouse. Design and accuracy of calipers for measuring subcutaneous tissue thickness. Br. J. Nutr. 9:133–143, 1955.
7. Gore, C. J., S. M. Woolford, and R. G. Carlyon. Calibrating skinfold calipers
. J. Sports Sci. 13:355–360, 1995.
8. Gore, C. J., K. I. Norton, T. S. Olds, and N. O. Whittingham. Kinanthropometry Accreditation: An Australian Model
. The International Council of Sport Science and Physical Education 1996 Pre-Olympic Congress, Dallas, Tx, Abstract No. 4018, 1996, p. 129.
9. Knapp, T. Technical error of measurement: a methodological critique. Am. J. Phys. Anthropol. 87:235–236, 1992.
10. Keys, A. Recommendations concerning body measurements for the characterization of nutritional status. Hum. Biol. 28:111–123, 1956.
11. Norton, K. Anthropometric estimation of body fat. In: Anthropometrica, K. I. Norton and T. Olds (Eds.). Sydney, Australia: University of New South Wales Press, 1996, pp. 172–194.
12. Norton, K., N. Whittingham, L. Carter, D. Kerr, C. Gore, and M. Marfell-Jones. Measurement techniques in anthropometry. In: Anthropometrica, K. I. Norton and T. Olds (Eds.). Sydney, Australia: University of New South Wales Press, 1996, pp. 25–75.
13. Pederson, D. G. and C. J. Gore. Anthropometry measurement error. In: Anthropometrica, K. I. Norton and T. Olds (Eds.). Sydney, Australia: University of New South Wales Press, 1996, pp. 77–96.
14. Schmidt, P. K. and J. E. L. Carter. Static and dynamic differences among five types of skinfold calipers
. Hum. Biol. 62:369–388, 1990.
15. Siri, W. E. Body volume measurement by gas dilution. In: Techniques for Measuring Body Composition, J. Brozek and A. Henschel (Eds.). Washington, D.C.: National Academy of Sciences, National Research Council, 1961, pp. 108–117.
16. Woolford, S., P. Bourdon, N. Craig, and T. Stanef. Body composition and its effects on athletic performance. Sports Coach. 16 (4):24–30, 1993.