Comparison of ultrasound and skinfold measurements in assessment of subcutaneous and total fatness.код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 58307-313 (1982) Comparison of Ultrasound and Skinfold Measurements in Assessment of Subcutaneous and Total Fatness GARY A. BORKAN, DAVID E. HULTS, J O H N CARDARELLI. A N D BELTON A. BURROWS Normative Aging Study, VA Outpatient Clinic, Boston, Massachusetts 02108 (G.A.B.,D.E.H.), and Department of Nuclear Medicine, VA Medical Center, Boston, Massachusetts 02130 (J.C., B.A.B.) KEY WORDS Anthropometry Body composition, Obesity, Skinfolds, Ultrasound, ABSTRACT Ultrasound (A-scanmode) and skinfold methods were evaluated in the measurement of subcutaneous fat thickness and prediction of total fat weight (by whole body potassium counting). Based on intraobserver correlations on 39 men at 15 body sites, skinfold caliper measurements were more reproducible than ones obtained by ultrasound. Measurements made with the two techniques at the same site typically produced different mean estimates of fat thickness. However, scores were often highly correlated with each other, indicating similar relative rankings of subjects by each technique. Skinfolds were more highly correlated with total fat weight than were ultrasound measurements, but body weight and anthropometric measures had even higher correlations with total fat weight. Anthropometric measurements were highly correlated with fatness because of their association with body weight, and when this relationship was statistically controlled for, they typically lost their predictive effectiveness. Multiple regression analyses revealed that the best predictors of fat weight were body weight along with skinfold and ultrasound measurements. These results suggest that skinfolds are a more effective means of assessing subcutaneous fat than ultrasound, especially when the large difference in cost of equipment is considered. For the past several decades, the skinfold caliper has been the most frequent method of measuring subcutaneous fat thickness. This device has many distinct advantages, especially in field situations: it is painless, noninvasive, simple to use, portable, and does not require elaborate electronic technology. Moreover, a very substantial literature exists on skinfold measurements and on subcutaneous fat as an index of total body fat (Durnin and Rahaman, 1967; Fletcher, 1962; Keys and Brozek, 1953; Newman, 1956). Disadvantages of skinfolds are that they compress the fat tissue, that they are only useful at certain body sites, and that they may be ineffective in the very obese (Brozek and Kinzey, 1960; Burkinshaw et al., 1973; Durnin et al., 1971). Softtissue radiography represents an alternative and more accurate measure of subcutaneous fat (Garn, 1954; Reynolds, 1944; Stuart et al., 1940). However, this technique is relatively 0002-9483/82/5803-0307$02.500 1982 ALAN R. LISS, INC. cumbersome, and has been less widely adopted. Tissue magnification, radiation exposure, and subject positioning must be carefully monitored in soft tissue radiography (Tanner, 1962). The value of ultrasound in measuring subcutaneous fat thickness has not yet been fully investigated. Ultrasound devices can produce internal images of tissue configuration (B-scan mode) or depth readings of changes of tissue density (A-scanmode). Ultrasound involves no radiation exposure, and is safe and painless to use. This technique has been employed for assessment of body composition of livestock (Stouffer, 1969; Wallace et al., 1977) and validations of ultrasound determinations have been published (Booth et al., 1966; Bullen et al., 1965; Haymes et al., 1976).More recently a Received August 10, 1981; accepted January 21, 1982 308 G.A. BORKAN, D.E. HUUI'S, J. CARDARELLI. AND B.A. BURROWS small portable ultrasound device (Body Composition Meter) has been introduced which is designed specifically for assessment of subcutaneous fat using the A-scan mode. Preliminary reports demonstrate the accuracy of this device through measurements of cadavers (Sanchez and Jacobson, 1978) and surgical patients (Balta et al., 1981). The present study compares subcutaneous fat thickness measurements using a skinfold caliper and Body Composition Meter. Intraobserver reliability was assessed, and site-by-site comparisons were made of fat thicknesses obtained by the two methods. We have also compared ultrasound and skinfolds with traditional anthropometry in how well they predict total fat weight (by potassium counting),both at individual sites and in combination. MATERIALS AND METHODS Subjects Subjects were 39 males recruited from the participants in the Normative Aging Study of the Veterans Administration Outpatient Clinic, Boston, Massachusetts. These men, along with almost 2000 other active participants, have been enrolled in a comprehensive longitudinal aging study for an average of 18 years. To gain entry into the Normative Aging Study, men were screened for good health (Bell et al., 1972), although in the succeeding years there has been considerable differentiation in health within the sample. Volunteers from the Normative Aging Study were recruited for a special series of body composition examinations at the VA Medical Center, Boston, and a total of 39 men had all the examinations described below. These men ranged in age from 41 to 76 years, and had a mean age of 57.6 years. TABLE 1. Intraobserver reliability correlations of subcutaneous fat measurement tecniques by sitea Skinfold 1-r2 r Site Arm Anterior (biceps) Lateral Posterior (triceps) Medial Chest Anterior (juxt a-nipple) Mid anterior Lateral (axillary) Posterior (subscapular1 Abdomen Anterior (umbilicus) Lateral (suprailiac) Posterior (dorsal iliac) Leg Anterior Lateral Ultrasound r I-r2 .I1 .81 .83 .I4 .41 .34 .31 .45 .I8 .45 .81 .48. .39 .90 .19 .I1 .41 .85 34 .96 .28 .80 .43 .I3 .36 .29 .08 .41 .11 .14 .24 .64 .48 .62 .59 .I1 .62 .I6 .I9 (n=25) .42 .38 .I2 .I1 .48 50 - - .I8 .74 .39 .45 .91 .93 .87 Posterior Medial .80 .34 .I1 .82 > .32 significant at .05 level. Ultrasound: Ultrasound measurements of subcutaneous fat thickness (including skin thickness) were taken at the same 15 body sites using portable ultrasound equipment (Body Composition Meter, Ithaco, Inc., Ithaca, New York). This lightweight device, designed expressly for assessment of subcutaneous fat thickness, utilizes the A-scan (linear) mode as opposed to the B-scan (two-dimensional) mode. Rather than use the 2.5 MHz transducer supplied with the meter, we used a 3.5 MHz transducer with a 0.5 diameter for greater tissue resolution. Readout from the instrument utilizes a series of light emitting Measurements diodes (LEDs) that allows measurements in Skinfolds: A Lange skinfold caliper (Cam- millimeter intervals to a depth of 100 mm. We bridge Scientific Industries, Cambridge, held the transducer to the skin manually, using Maryland) was used to measure skinfold thick- ultrasound transmission gel, and were able to ness at each of 15 sites located at the levels of maintain good skin contact without depressthe chest at the nipples, umbilicus, and mid- ing the skin surface. Tissue thicknesses less point of the upper leg (halfway between the than 3 mm could not be differentiated due to pubis and middle of the patella). These sites are the blind spot in front of the transducer. We listed in Table 1. Standard anthropometric did not judge this to be a problem, however, techniques were used and all measurements because 3 mm is essentially the lower limit were replicated at the end of the patient exami- of fat and skin thickness encountered. During nation (a 30-minute interval). All measure- the examinations we routinely checked the ments were made by an experienced anthro- calibration of the Body Composition Meter pometrist (G.A.B.),with the subject standing. using a plastic phantom supplied by the Posterior and medial leg skinfolds could not be manufacturer . taken due to skin tautness and lateral leg skinAlthough the recommended mode of utilizfolds could be made on only 25 individuals. ing the Body Compositition Meter is to reduce 309 ULTRASOUND AND SKINFOLD ASSESSMENT OF FATNESS the gain control until only a single diode remains lit (corresponding to the fat-muscle interface), in practice several adjacent or nearby LEDs may remain lit even at low gains (due to spurious echos, or fascial layers). We found that the use of the Body Composition Meter required greater knowledge of underlying anatomy than did skinfold calipers. The anthropometrist was trained on this instrument under the direction of individuals experienced in its use. Whole body potassium counting: Each patient had a single potassium-40 measurement in a whole body counter. The low level counting facility at the VA Medical Center, Boston, consists of two matched 4 x 8 inch sodium iodide crystal detectors mounted in a scanning geometry in a 8 x 8 x 8 foot room with 9-inch steel walls. The subject lay supine on a bed between the two detectors for two 20-minute traverses of the counter assembly. Details of counters of this type, as well as scanning geometry and calibration have been published (Tyson et al., 1970). Total millequivalents of potassium-40 were converted to total body potassium based on the known relationship of potassium-40 to total potassium in the body, and based on the previous calibration of the counter with potassium-42. A correction was also made for differential counting efficiency related to body size. Body fat weight was calculated using the formula of Forbes and Lewis (1956): Fat weight = Weight - (Meq K/68.1). Anthropometry: A battery of 44 anthropometric measurements was also taken, including height, weight, a variety of length, breadth, and circumference measures, all using standard anthropometric sites and techniques (Damon et al., 1972).This is the same test battery used from the inception of the Normative Aging Study. The anthropometry was taken only once during the examination. RESULTS Intraobserver correlation coefficients (r)and the proportion of variance attributable to intraobserver error (1 - r2) are presented in Table 1 for each body site and technique. The correlations for skinfolds were higher than for ultrasound for every site except the anterior arm. Four of the seven sites in the trunk achieved correlations of 0.90 or greater for skinfolds. Ultrasound, on the other hand, did not achieve any reliability greater than 0.81 and four sites had correlations less than 0.5. The only sites where reliability of ultrasound compared favorably with skinfolds were on the leg, but this is a part of the body where there is very little fat. Mean fat thicknesses by site and technique are compared in Table 2. For this comparison skinfolds were divided in half, because they TABLE 2. Mean skinfold and ultrasound values b y site for 39 males Site ______ Arm Anterior Lateral Posterior Medial Chest Anterior Midanterior Lateral Posterior Abdomen Anterior Lateral Posterior Leg Anterior Lateral Posterior Medial X Skinfoldsa adj SD SD cv X Ultrasound adj SD SD 2.0 8.4 5.8 2.6 0.7 2.6 1.8 1.0 0.54 2.11 1.49 0.74 0.27 0.25 0.26 0.28 4.1 7.7 6.6 4.9 1.0 2.2 1.8 1.5 0.78 1.0 1.49 0.72 0.19 0.13 0.23 0.15 16.7 16.2 10.9 11.5 4.7 5.2 4.1 4.5 4.23 4.42 3.44 4.31 0.25 0.27 0.31 0.38 32.6 32.3 19.5 9.8 6.7 7.7 6.9 2.4 5.16 6.16 2.97 1.76 14.0 11.8 18.6 3.9 4.5 5.0 3.55 4.19 4.35 0.25 0.35 0.23 27.7 18.5 13.0 8.0 7.0 5.0 7.7 5.3 2.2 2.2 1.67 1.74 0.21 0.34 - - 7.2 4.7 5.9 9.2 1.8 1.5 1.8 2.8 - - askinfold divided by 2 for comparability with ultrasound. bt > 2.02 significant at .05 level. cr > .32 significant at .05 level. adj adj cv tb rc 15.4 4.7 12.5 0.56 0.71 0.81 0.65 0.16 0.19 0.15 0.18 16.0 18.4 9.0 -4.4 0.45 0.70 0.52 0.81 5.12 3.36 3.10 0.18 0.18 0.24 13.7 7.0 - 5.5 0.66 0.53 0.18 1.30 1.07 1.40 2.07 0.18 0.23 0.24 0.23 - 2.4 - 1.6 - 0.76 0.83 - 2.3 - 310 G.A. BORKAN, D.E. HULTS. J. CARDARELLI, AND B.A. BURROWS theoretically represent a double fold of fat, and ultrasound measures a single thickness. Adjusted standard deviations (corrected for intraobserver bias) are included, as are coefficients of variation (based on the adjusted standard deviation). For all sites except the lateral leg, the mean scores between the two techniques were significantly different. Because we had no criterion measure of actual fat thickness, it is not clear which technique yields scores which are closer to the actual fat thickness at each site. Taking the adjusted coefficient of variation as a measure of dispersion of the data, ultrasound generally produced a more limited range of scores than skinfolds. Correlations between ultrasound and skinfolds were strongly positive at many sites, indicating that relative rankings of individuals were similar, even if measurement scales varied. Posterior arm, posterior chest, and lateral leg had correlations greater than 0.80 (i.e., 64% of variance explained). At other sites agreement was much lower, though all but one had significant correlations. There was no apparent systematic relationship between the t-statistics (absolute comparisons) and the correlations (comparison of relative ranking of individuals). To determine whether these techniques are equally effective in predicting overall adiposity, we compared the skinfold and ultrasound sites individually with total fat weight from the whole body counting (Table 3). At all but the anterior arm site, skinfolds had higher correlations with fat weight than did ultrasound. Posterior chest (subscapular) skinfold was the only site that had a correlation with total fat greater than 0.60. The average correlation of skinfold sites with total fat was 0.51, compared to an average of 0.35 for ultrasound. Skinfolds also had higher correlations with body weight than ultrasound for most sites. At the majority of ultrasound and skinfold sites, the correlations with fat weight were greater than those for total weight. A number of skinfold and ultrasound sites remained significant in predicting fat weight, even when their association with body weight was controlled by partial correlation. In particular, significant partial correlations were found for ultrasound and skinfolds of the arms, and skinfolds of the chest. A variety of anthropometric variables were also compared with total fat (Table 4) and several of these were better predictors than any subcutaneous fat measurement. Body weight, with a correlation of 0.72, was more predictive of total fat weight than was any other anthropometric measurement. Abdominal circumference did almost as well, followed by hip breadth and two chest circumferences. In contrast to the subcutaneous fat measures TABLE 3. Correlation coefficients of subcutaneous fat measurements by site with fat weight /from 40K/, body weight and fat wbight partialling for body weight IN = 39)" Site Fat weight SkinF-~ Ultrafold sound Arm Anterior Lateral Posterior Medial Chest Anterior Mid anterior Lateral Posterior Abdomen Anterior Lateral Posterior Leg Anterior Lateral Posterior Medial > .32 significant at .05 level. Body weight Ultrafold sound ~~~ ~ ~ Skin- Fat wt partialling for body wt -~ __ SkinUltrafold sound ~~~ .44 .54 .42 .54 .55 .39 .31 .48 .3 1 .34 .12 .31 .32 .21 - .03 20 .33 .46 .48 .47 .48 .36 .48 .49 .53 .53 .59 .63 .38 .47 .46 .44 .50 .38 .39 .56 .27 .54 29 .35 28 .40 .49 .40 .28 .13 .37 .29 .52 .53 .57 .30 .64 .6 1 .55 .49 .ll .17 .ll .16 .31 - .09 .21 - .02 .17 25 - .05 .21 - 29 .33 .39 27 27 .I8 - .18 .40 23 - .17 - - .28 .09 .02 311 ULTRASOUND AND SKINFOLD ASSESSMENT OF FATNESS TABLE 4. Correlation coefficients of anthropometric measurements with fat weight, body weight and fat weight partialling for body weight (N = 39)a Site Fat weight Weight Height Arm Span Chest depth Hip breadth Maximum chest circumference Resting chest circumference Abdomen circumference Calf circumference Wrist circumference Upper arm circumference .72 .22 .I1 .55 .69 .68 .66 .71 .47 .27 Body weight - - .47 .33 .38 .78 - .19 .79 25 .24 .43 - .08 - .28 - .05 .77 .68 .71 .59 .64 .44 Fat wt partialling for body wt -.la .41 29 ar > 3 2 significant a t .05 level TABLE 5. Prediction o f fat weight lkg) using various sets o f independent variables Predictors by site and method Skinfold Ultrasound Anthropometric circumference Anthropometric length and breadth All predictors multiple r. r2 Prediction equation Fat weight = 2 8 6 chest posterior + ,961 arm anterior ,780 arm medial + ,148 abdomen 4.38 posterior F a t weight = 3.428 arm anterior + ,374 chest mid anterior + 2.759 leg lateral - ,623 leg medial - ,981 leg anterior - 1.981 Fat weight = ,617 abdomen circumference + ,784 calf circumference - 1.687 wrist circumference - 32.705 Fat weight = 1.786 hip breadth + 1.741 chest depth - ,361 arm span + .87 wall height -79.192 Fat weight = ,600 weight 2.008 medial arm ultrasound + ,266 lateral chest skinfold -.420 anterior abdomen skinfold + 1.373 anterior arm ,261 lateral chest ultrasound - ,915 skinfold upper arm circumference - 8.657 + + + r = .71 r2 = .50 r = .76 r2 = .58 r = .75 r' = .56 r = .78 r2 = .60 r = .92 r2 = .85 + in Table 3, the anthropometric measures tended to correlate even more highly with body weight than with fat weight. Consequently, their partial correlations with fat weight, after controlling for body weight, were nonsignificant for most measurements. Several multiple regression analyses were performed with fat weight as the dependent variable and various sets of subcutaneous fat and anthropometric measures as independent variables (Table 5). Results were quite similar (multiple r between 0.71-0.78) between skinfolds, ultrasound, anthropometric circumferences, and lengths. Ultrasound achieved a higher multiple r than skinfolds, despite its poorer showing in the individual correlations in Table 3. The highest predictive effectiveness was achieved when all skinfold, ultrasound, and anthropometric parameters were included as potential independent variables. A multiple r of 0.92 was achieved with body weight entering first, followed by five ultrasound and skinfold variables, and upper arm circumference. No other anthropometric parameters entered the equation significantly, despite their high individual correlations with fat weight. DISCUSSION Results of this study indicate that skinfolds measured by the Lange caliper are a more effective means of measuring subcutaneous fat than ultrasound using the Body Composition Meter. Intraobserver reliability of skinfolds was higher than for ultrasound. Interobserver reliability was not tested in this study, and it is uncertain whether variation in ultrasound technique between technicians would be greater than that for skinfolds. Criterion 312 G.A. BORKAN. D.E. HULTS, J. CARDARELLI. AND B.A. BURROWS methods were not available to assess whether skinfolds and ultrasound are accurate indications of subcutaneous fat thickness at the sites they measure. Therefore we cannot state which provides the best indication of true fat thickness . However, we were able to measure total fat weight, and accuracy of the various site specific composition techniques was assessed. When all variables were evaluated individually, neither ultrasound nor skinfolds predicted fat weight as well as did body weight or several stand ard anthropometric measurements . This has also been reported by Wilmore and Behnke (1970), who found abdomen circumference to be a better predictor of total body density (equivalent to percent fat) than any skinfold measure. In a comparison of ultrasound and skinfolds with total body density (Haas et al., 1981)body weight had the highest correlation with fat mass, and five different anthropometric circumferences had correlations with fat mass higher than any single skinfold or ultrasound measurement. In an investigation of anthropometric indicators of weight loss, Bray et al. (1978) showed higher correlations between anthropometric circumferences and weight loss than with any skinfold site. The data in Tables 3-5 may indicate the origins of the findings described above. Anthropometric measures of circumference were more highly correlated with body weight than with fat weight. Conversely, subcutaneous fat measures were more highly correlated with fat weight than with total weight. Once the effect of body weight was removed using partial correlations, subcutaneous fat measures were more highly correlated with fat weight than was anthropometry. This is substantiated in the last equation in Table 5, which shows that after body weight entered the equation, the only other variables that contributed significantly to explaining fatness were measures of subcutaneous fat (with the exception of upper arm circumference which was the last variable to enter the equation). Thus, while weight proved a good indicator of total fat (explaining 52% of the variance), the addition of five other skinfolds and ultrasound measures brought the explained variance up to 82%. We do not advocate using these equations as predictors of total body fat in other populations. There is ample evidence that such equations are effective only in the sample on which they are based or others of very similar composition (Malina, 1980).Age, ethnic affiliation, sex, economic status, and health must be close- ly matched, as well as the anthropometric techniques used. The present sample, while basically consisting of middle-aged, middleincome white males, is a very small sample size on which to base predictions for the U.S. population. In addition, we are not advocating use of skinfolds and ultrasound in conjuction, for the information they provide is very similar. Results of this study also have bearing on the relationship of subcutaneous to total fat. It has been recognized that subcutaneous fat is not a definitive indicator of total fat. There is a substantial accumulation of fat in the abdomen, and this may vary considerably between individuals and with age (Borkan and Norris, 1977). Anthropometrists of adults are well aware from experience that a large abdomen and a large abdominal skinfold do not always occur together. When this factor is combined with the considerable variability between individuals in fat distribution within the body, it becomes clear why no single site, or even cornbination of subcutaneous fat sites, gives a truly accurate indication of total fatness. ACNOWLEDGMENTS The authors wish to thank the staff of the Normative Aging Study for their helpful comments in preparing this manuscript. We thank Beth Dolan and Joan Sodergren for typing, and Susy Chan for keypunching the data. We appreciate the assistance of Jere Haas, Ed Frongillo, and James Stouffer for their assistance using the Body Composition Meter. This research was supported by the Medical Research Service of the Veterans Administration. LITERATURE CITED Balta. PJ, Ward, MWM, and Tomkins. AM (19811 U l t r a sound for measurement of subcutaneous fat (Letter to the Editor). The Lancet, Feb.28, 1981. 1:504. Bell, B, Rose, CL, and Damon, A (1972) the Normative Aging Study: an interdisciplinary and longitudinal study of health and aging. Aging and Hum. Dev. 3:5-17. Booth, RAD, Goddard, BA. and Paton, A 11966)Measure ment of fat thickness in man: a comparison of ultrasound, calipers and electrical conductivity. Br. J. Nutr. 203 19-727. Borkan, GA. and Norris, AH (1977) Fat redistribution and the changing body dimensions of the adult male. Hum. Biol. 49:495-514. Bray, GA. Greenway, FL, Molitch. ME, Dahms. WT, Atkinson, RL. and Hamilton, K (1978) Use of anthropometric measures t o assess weight loss. Am. J. Clin. Nutr. 31.769-773. Brozek. J. and Kinsey, W (1960) Age changes in skinfold compressibility. J. Gerontol. 1545-51. Bullen, BA. Quaade, F, Oleson. E, and Lund. SA (1965) UItrasonic reflections used for measuring subcutaneous fat in humans. Hum. Biol. 37:377-384. ULTRASOUND AND SKINFOLD ASSESSMENT OF FATNESS Burkinshaw, L, Jones, PAM, and Krupwicz, DW (1973)Observer error in skinfold thickness measurements. Hum. Biol. 45273. Damon. A, Seltzer, CC. Stoudt. HW. and Bell, B (1972) Age and physique in healthy white veterans at Boston. J. Gerontol. 27202-208. Durnin. JVGA, Armstrong, WH. and Wormorsley, J (1971) An experimental study of the variability of measurements of three observers on twenty-three young women and twenty seven young men. Proc. Nutr. SOC.30:9A. Durnin, JVGA and Rahaman. MM 11967)Theassessment of the amount of fat in the human body from measurements of skinfold thickness. Br. J. Nutri. 21:681-689. Fletcher, R F (1962)The measurement of total body fat with skinfold caliper. Clin. Sci. 2233-346. Forbes, GA. and Lewis, AM (1956) Total sodium, potassium, and chloride in adult men. J. Clin. Invest. 35:596-600. Garn. SM (1954)Fat patterning and fat intercorrelations in the adult male. Hum. Biol. 2656-59. Haas. JD, Frongillo, E J r , Andrup, K. Mainigi. K, Roe, DA. and Buttel, L (1981)Sensitivity of the Ithaco Body Composition Meter. Report submitted to National Cancer Institute (Contract No. N01-CP-85652). Haymes. EM, Lundgren, HM, Loomis, JL. and Buskirk. ER (1976) Validity of the ultrasonic technique as a method of measuring subcutaneous adipose tissue. Ann. Hum. Biol. 3245-25 1. Keys, A, and Brozek, J (1953) Body fat in adult man. Physiol. Rev. 33,245-325. 313 Malina, RM (1980) The measurement of body composition. In: FE Johston, A F Roche. and C Susanne (eds):Human Physical Growth and Maturation: Methodologies and Factors. New York: Plenum Press, pp. 35-59. Newman, RW (1956) Skinfold changes with increasing o b e sity in young American Males. Hum. Biol. 2753-63. Reynolds, EL (1944) Differential tissue growth in the leg during childhood. Child Dev. 15:181. Sanchez, CL, and Jacobson, HN (1978) Anthroponietry measurements, a new type (Letter to the Editor). Am. J. Clin. Nutr. 31.1116-1117. Stouffer. J R 11969) Techniques for the estimation of the composition of meat animals. J. Anim. Sci. 29.207-219. Stuart, HC, Hill, P, and Shaw, C (1940)The growth of bone, muscle and overlying tissues a s revealed by studies of roentgenograms of the leg area. Mon. SOC. Res. Child Dev. 5:3. Tanner, J M (1962)Growth a t Adolescence (2nd ed.).Oxford Blackwell Scientific Publications. Tyson, I, Genna, S , Jones, RL. Bikerman, V. and Burrows, BA (1970) Body potassium measurements with a total body counter. J. Nucl. Med. 111255-259. Wallace, MA, Stouffer, JR. and Westervelt, RG (1977)Relationships of ultrasonic and carcass measurements with retail yield in beef cattle. Livest. Prod. Sci. 4,153-164. Wilmore, J H , and Behnke, AR (1970)An anthropometric estimation of body density and lean body weight in young women. Am. J. Clin. Nutr. 23267-274.