- Email: [email protected]
Use of a Fat Probe to Assess Variation in Abdominal Fat in Broilers1
Use of a Fat Probe to Assess Variation in Abdominal Fat in Broilers 1 K. W. WASHBURN and P. A. STEWART Department of Poultry Science, University of Georgia, Athens, Georgia 30602 (Received for publication November 10, 1986)
1987 Poultry Science 66:1911-1917 INTRODUCTION
Excessive fat deposition in broilers is a problem of concern to the poultry industry. Genetic selection to decrease the proportion of fat mass while maintaining lean mass potentially can solve the problem. Although the study of Becker et al. (1984) indicated that abdominal fat could be reduced by selection, a correlated reduction in body weight was of concern. The use of genetic techniques such as selection index or independent culling may allow the reduction of abdominal fat without reduction in lean body mass (Leclercq et al., 1980; Cahaner and Nitsan, 1986). A measurement of abdominal fat that does not require sacrificing the individual is needed to maximize genetic selection progress. Pym and Thompson (1980) developed a caliper technique for the estimation of abdominal fat in live broilers. They reported a phenotypic correlation of .80 between caliper measure and proportion of abdominal fat. Schwartzberg et al. (1980),
'Supported by State and Hatch funds allocated to the Georgia Agricultural Experiment Stations of the University of Georgia and by a Binational Agricultural Research and Development Fund project.
using their version of the caliper also reported a high correlation (.77) between the fat probe measurement and fat pad as a percentage of live weight. These correlations indicate the calipers or fat probe should be a useful technique as an indirect method of selecting for decreased abdominal fat. Mirosh and Becker (1981) reported correlation coefficients between caliper values and abdominal fat weights of .29 and .45 at 41 and 48 days of age when the probe was inserted to 20 mm. A somewhat higher value (.54) was obtained when the probe was inserted at 35 mm and measurements were made at 48 days. However, Gyles and Maeza (1981) reported that correlations between the probe values and measures of abdominal fat were close to zero. These latter low correlations and personal communication with broiler breeders who attempted to use the calipers as a measure of fat thickness cast some doubt on the use of fat probe calipers as an indirect measure of abdominal fat. Some of the variables that were not controllabe in the studies that used the probe may contribute to the variation in the conclusions of the different studies. The objective of the present study was to determine how well abdominal fat can be estimated from caliper measures obtained from chickens over a wide range of body weights and amounts of abdominal fat. A modification of
1911
Downloaded from http://ps.oxfordjournals.org/ at Univ of Iowa-Law Library on June 2, 2015
ABSTRACT Four trials were conducted to compare the fat probe measure of abdominal fat with that of actual abdominal fat weights and percentage abdominal fat. Various ages, diets, and genetic groups were used to provide experimental chickens with a wide range in body weights and abdominal fat percentages. Mean values obtained by two operators using the probe on the same bird were similar (correlations of .70 to .79). Considering all comparisons, the arithmetic mean of the correlations between fat probe values and percentage abdominal fat was .24; values ranged from .01 to .44. In five comparisons of groups differing significantly in percentage abdominal fat, there were no significant differences in probe values, and correlations between fat and probe values were low. In three other comparisons of groups differing significantly in percentage abdominal fat, differences between probe values were significant and correlations between fat and probe values were moderate. In the comparison groups in which there was an association between probe and fat values, the magnitude of the differences in fat values was about four times as great as that of differences in probe values and the correlations were moderate (.28 to .44). In the comparisons in which higher abdominal fat was associated with higher probe values, higher body weight was associated with higher percentage abdominal fat. In comparisons where higher probe values were not associated with higher abdominal fat, higher body weights were not associated with higher percentage fat. (Key words: variation, fat probe, abdominal fat, broilers, body weight)
1912
WASHBURN AND STEWART
the fat probe caliper that does not require operator pressure to obtain a reading is described and correlations are compared between caliper measurements and actual weight and percentage of abdominal fat for birds with various abdominal fat and body weights. MATERIALS AND METHODS
Downloaded from http://ps.oxfordjournals.org/ at Univ of Iowa-Law Library on June 2, 2015
A modification was made of the fat probe caliper described by Pym and Thompson (1980). There are two primary differences between this caliper and that used by Pym and Thompson (1980). Operator pressure must be exerted to open the calipers. They close automatically after insertion of the cylindrical head into the cloaca and releasing of pressure on the handle. This removes the effect of the operator on the pressure exerted and consequently the reading. Another difference is that the flattened head is mounted so that it can pivot. This should allow better contact with the abdominal area. The cylindrical head and the graduated scale is similar to that of the caliper used by Pym and Thompson (1980). A further modification consisted of a depth set so that the probe would be inserted to a uniform depth in all individuals.The depth of 35 mm was chosen because studies of Mirosh and Becker (1981) indicated that a higher correlation coefficient with abdominal fat was obtained when the probe was inserted to this depth. Four trials were conducted to compare the fat probe measure of abdominal fat with that of actual abdominal fat weights and abdominal fat as a percentage of live weight. Various ages, diets, and genetic groups were used to provide experimental animals with variation in abdominal fat. In these trials, feed was restricted for approximately 8 h before determination of fat pad thickness, body weight, and abdominal fat pad weights. To obtain probe measurements, birds were held in the manner described by Pym (1981) except they were held on the operator's leg rather than on an inclined ramp. However, birds were held at an angle similar to that described by Pym (1981). After probe measurements and body weights were obtained, birds were killed, feathers removed, and abdominal fat (leaf and gizzard) removed and weighed to the nearest .1 g. The percentage fat was calculated as the abdominal fat weight/live body weight. In Trial 1, the variation in repeated probe measures was studied. Two operators independently measured the thickness of the fat pad on
the right and left side of males of two broiler strains at 63 days of age. These strains had previously been shown to differ in the amount of abdominal fat. Operator 1 was left handed, whereas Operator 2 was right handed. Both operators had a similar amount of prior experience with the use of the fat probe. In Trial 2, the same procedure was followed except that determinations were made at 7 wk of age. The Athens-Canadian randombred population (AC) and two broiler strains were used and each of these genetic groups was subdivided into dietary groups that received either Diet A or Diet B. Diet A contained 3,210 kcal metabolizable energy (ME) and 23% protein, whereas Diet B contained 3,256 kcal ME and 21.2% protein. In Trial 3, groups that differed in abdominal fat were obtained from an experiment designed to study the effects of dietary NaCl on water consumption and carcass fat. In this trial, which used the two commercial broiler strains used in Trial 1, percentages of abdominal fat were significantly different between sexes and between those groups receiving diets containing 1.6 or 2.4% added NaCl. There were no significant differences between broiler strains. At 49 days of age live body weights, probe values, fat weights, and percentages of fat were obtained from 120 males and 120 females fed the control diet and from 30 males each from the 1.6 and 2.4% NaCl treatments. In Trial 4, the usefulness of the fat probe in detecting differences in abdominal fat between families of a broiler grandparent line was evaluated. This trial was a satellite trial for a primary study involving the genetic basis of efficiency of feed utilization and percentage abdominal fat and used 346 progeny of 35 sire families of a commercial female grandparent line. Chicks were reared in floor pens until 3 wk of age, when they were placed in individual cages until 7 wk of age. They were fed ad libitum the standard University of Georgia (UGA) starter diet containing 3,120 kcal ME and 23% protein from 0 to 3 wk of age and the UGA finisher containing 3,256 kcal ME and 21.2% protein to 7 wk of age. At 7 wk of age, live body weights and fat probe readings were obtained, birds were processed, and the abdominal fat was removed and weighed. Abdominal fat probe values were obtained on the right side by Operator 1 of the previous trials.
FAT PROBE AND ABDOMINAL FAT
sides is presented in Tables 1 and 2. Mean values obtained by the two operators were similar and not statistically different for the different strain and trial comparisons. Correlations between the values obtained by the two operators for the same fat pads were .79 and .70 in Trials 1 and 2. Values for the right and left sides were not significantly (P=£.05) different in Trial 1, whereas in Trial 2 values for the left side were slightly (.35) but significantly (P«.05) larger than the right. This difference between scores was similar for both operators. Correlations between values for the left and right sides were high (.62 to .76) in both trials. The coefficient of variation was somewhat lower for Operator 2 than for Operator 1. In Trial 1, abdominal fat weights and percentages for Strain 1 were significantly greater than for Strain 2 (Table 3). However, differences in probe values between strains were not significant nor was there a trend toward higher probe values for the strain with more abdominal
RESULTS
Variation in probe measures between strains, operators, and right vs. left abdominal fat pad
TABLE 1. Differences in probe measures between strains, diets, operators, and 1right/left (R/L) fat pads Operator 2
Operator 1 Trial
Strain 1
Diet 2
R
L
7.6" 7.2"
7.4 7.2
7.0 7.2
7.2" 7.3"
7.4 a
7.4™ 16.4
7.3 a
7.1 a
7.2™ 14.2
3.5 3.9
4.2 4.3
3.8" 4.iy
4.0 3.9
4.4 4.3
4.2" 4.iy
3.7
4.2
4.0"
4.0
4.3
4.1"
5.2 5.6
5.4 5.7
5.3"
5.3 5.8
5.6 5.8
5.5"
s.ey
x BR2
5.4
5.6
5.5?
5.5
5.7
5.7y
x Trial 2 C.V.
4.5 a
4.9 b
4.7™ 22.7
4.7 a
5.0 b
4.9™ 21.2
R
L
7.6 7.3
7.7 7.0
7.5 a
X
X
(mm) BR1 BR2 x Trial 1 C.V.3 AC
BR2
A B
A B
s.sy
a ' b Indicated significant differences between means for right and left fat pad determined by an operator (P«.05). ^Indicates significant differences between broiler strains or diets within strains determined by an operator (P«.05). ' Indicates significant differences between means of all values obtained by different operators (P<.05). 1
BR1 and BR2 = Broiler strains 1 and 2; AC = Athens-Canadian randombred population.
2
A = Diet containing 3,210 kcal metabolizable energy (ME)/26% protein; B = diet containing, 3,256 kcal ME/21.2% protein. 3
C.V. = Coefficient of variation.
Downloaded from http://ps.oxfordjournals.org/ at Univ of Iowa-Law Library on June 2, 2015
In Trial 4, the value of the fat probe in detecting differences in abdominal fat between families was assessed two ways. In the first, subgroups of the 35 families were grouped in seven high-fat and seven low-fat families based on the sire family percentage fat means and their fat probe values compared. In the second, two types of phenotypic correlations were obtained: individual, using individual weight, probe and abdominal fat values; and sire family, using the mean values from the 35 sire families. Correlations were obtained separately for males and females. Statistical significances of differences were determined by analysis of variance and Duncan's multiple range tests. Analysis was based on within trial individual data for main effects and interactions.
1913
1914
WASHBURN AND STEWART TABLE 2. Phenotypic correlations between repeated probe measures in Trials 1 and 2 Operator 1
Operator 2
Trial
Left/right1
Left/right
o,o 2 2
1 2
.73** .75**
.62** .76**
.79** .70**
1
Left/right refers to correlation between values for the left and right abdominal fat pads.
2
O. Oj refers to correlation between mean of values obtained by Operators 1 and 2.
**P<.01.
Correlations between the probe values and the amount of abdominal fat in Trial 2 were moderate for the broiler strain but low for the AC population (Table 4). In Trial 3 there was no significant difference in probe values between sexes, which differed significantly in percentage abdominal fat (Table 5). There were, however, no significant differences in abdominal fat weights in this comparison (Table 5). In the comparison of the groups receiving diets containing different levels of salt, differences between probe values were significant for groups differing in abdominal fat weight and percentage. Correlations of probe values with abdominal fat weight or percentage abdominal fat were moderate (about .4). In Trial 4, differences between the high and low abdominal fat families in fat weight and percentage fat were highly significant (Tables 6 and 7), with large differences between groups.
TABLE 3. Means (± standard error) for fat probe values, abdominal fat weight, percentage abdominal fat, and body weight
Trial
Strain1
1
BR1 BR2
2
AC BR2
Abdominal fat weight
Abdominal fat
Body weight
(g) 68.0 ± 3.3 a 55.9 ± 2.8 b
(%)
7.21 ± .32 a 7.33 ± .19 a
2.37 ± ,10 a 1.93 ± .09 b
(g) 2844 ± 4 0 a 2893 ± 34 a
A B
4.04 ± .08 e 4.11 ± .06 c
2.7 ± 0 . 3 d 4.3 ± 0.4 C
0.39 ± .05 d 0.61 ± .05 c
679 ± 14 c 680± 12 c
A B
5.40 ± .08 b 5.73 ± .11*
14.5 ± 1.2b 25.9 ± 1.7a
0.89 ± .06 b 1.47 ± .09 a
1590± 4 0 b 1747 ± 39 a
Diet 2
Probe (mm)
Column means within trials with different superscripts differ significantly (P<.05). 1
BR1 and 2 = Broiler Strains 1 and 2; AC = Athens-Canadian randombred.
2 Diet A contained 3,210 kcal metabolizable energy (ME)/26% protein; Diet B contained 3,256 kcal ME/ 21.2% protein.
Downloaded from http://ps.oxfordjournals.org/ at Univ of Iowa-Law Library on June 2, 2015
fat. Correlations between probe values and amounts of abdominal fat were low (.24) regardless of operator. In Trial 2, Strain 2 broilers had significantly greater amounts and percentages of abdominal fat than AC birds (Table 3). Fat probe values were also significantly higher in the fatter broiler strain. However, the magnitude of the fat probe differences between strains was much less (broiler strain value was 137% of AC) than that of fat pad weight differences (broiler strain value was 577% of AC). Diet B resulted in significantly increased fat pad weights in both the broiler and randombreds (Table 3). There was a slight but significant increase in probe values in the broilers fed Diet B and a nonsignificant increase in the AC randombreds. The magnitude of the increase in probe values due to dietary differences was much less than that for fat weight or percentage.
1915
FAT PROBE AND ABDOMINAL FAT TABLE 4. Correlation of probe measurements with abdominal fat and body weights Abdominal fat
Abdominal fat Trial 2 Operator
Trial 1
o2
.22 .24 1
BR-2 (g) .52 .35
I
Body weight
Trial 2
Trial 2
AC
Trial 1
BR-2
AC
Trial 1
BR-2
AC
.22 .28
.16 .23
.36 .28
.18 .23
.32 .19
.64 .39
.23 .28
BR-2 = Broiler Strain 2; AC = Athens-Canadian randombred.
DISCUSSION
The results of this study as well as those of Mirosh and Becker (1981) and Gyles and Maeza (1981) indicate that the degree of association of fat probe values with percentage abdominal fat is not generally as high as that previously reported by Pym and Thompson (1980) and Schwartzberg et al. (1980). Correlations between probe values and percentage abdominal
TABLE 5. Means and correlations of fat probe values with abdominal fat weight, percentage abdominal fat, and body weight (Trial 3)
Statistic
Variable
X
Sex Males Females Dietary salt 1.6% NaCl 2.4% NaCl
r2
Sex Males Females Dietary salt 1.6% NaCl 2.4% NaCl
Probe
Abdominal fat weight
Abdominal fat
Body weight
(mm)
(g)
(%)
(g)
7.12 a 7.20 a
34.2 a 33.9 a
1.58 b 1.91 a
2,151 a l,765b
7.41 a 6.78 b
38.7 a 27.9 b
1.98 a 1.45 b
l,956 a l,917b
.41 .40
.44 .42
.34 .26
.26 .42
.37 .43
.28 .06
a ' Means between sexes or dietary salt level followed by different superscripts are significantly different (P<.05).
Downloaded from http://ps.oxfordjournals.org/ at Univ of Iowa-Law Library on June 2, 2015
fat in this study ranged from .01 to .44. The arithmetic mean of all the correlations in the 14 comparisons was .24. Possible reasons for the variation include differences in the construction and operation of the probes, including differences in experience, skill, and methodology of the user. Probes used in this study and by Gyles and Maeza (1981) and Mirosh and Becker (1981) were similar. They differed from that used by Pym and Thompson (1980) primarily in that they required no operator pressure to measure the abdominal fat thickness whereas the one used by Pym and Thompson (1980) required operator pressure to obtain the reading. The report of Pym (1981) emphasizes the effects of the operator skill and methodology on the repeatability of the values obtained. Repeata-
However, differences between these high and low abdominal fat families for probe values were small and not significant. Males had significantly higher body weights, less abdominal fat, and lower percentage fat than did the females. However, there were no significant differences between sexes in fat probe values.
1916
WASHBURN AND STEWART TABLE 6. Mean and coefficient of variation of l^wk body weights, fat weights, and percentage fat with fat probe values in a broiler grandparent population (Trial 4)'
Variable
n
7-wk weight
Males Females x H-families 2 x L-families
188 158
2,037" l,742b 1,900" 1,932"
CV
(g)
Fat weight
CV
% Fat
CV
Probe
(35) (30)
1.23" 1.69b 1.65" 1.09b
(33) (29)
7.34" 7.20" 7.58" 7.06"
(mm)
(g) (11) (9)
CV
25.0" 29.3b 31.3" 21.2b
(15) (15)
ab ' Means between males or H and L families followed by different superscripts are significantly different (P«.05). Coefficients of variation (CV) in parentheses.
H and L families refers to the seven sire families with the highest and the seven sire families with the lowest percentage fat pads.
bility of multiple caliper readings by Pym (1981), who was very experienced in the use of the calipers, ranged from .73 to .81; repeatability of one operator with relatively less experience ranged from .27 to .73 whereas that of a third operator with a similar amount of experience ranged from .01 to .29. In Trials 1 and 2 of the present study, two operators (one right handed and one left handed) measured both the right and left fat sides. Correlations between values for the left and right fat sides and between operators were high, indicating agreement between the repeated measures. In the present study, the association between probe values and abdominal fat values was greater in some comparisons than in others. The reason for differences in the association between percentage fat and probe values between trials may be the relative magnitude and the direction of the differences between fat weight and body weight in a particular comparison. These data are summarized in Table 8. In Trial 1 there were significant differences between the broiler strains in fat weight and
percentage fat, but not in probe values. In Trial 2 there were significant differences between the AC birds fed Diets A or B in fat weight and percentage fat weight but not in probe values. In Trial 4 there were significant differences between the high and low fat families in fat weight and percentage fat but not in probe values. In all these comparisons in which probe values were not different when fat values were different, body weights of the groups compared were not significantly different. In Trial 2 there were significant differences between broiler Strain 2 birds fed Diets A and B in fat weight, percentage fat, and probe values. Also, there were significant differences between the AC population and the broilers in fat weight, percentage fat, and probe values. In Trial 3, there were significant differences between broilers fed diets with different salt levels in fat weight, percentage fat, and probe values. In all these comparisons higher percentage fat was associated with higher body weight. In Trial 3, higher body weight was not associated with higher probe values in comparisons between
TABLE 7. Phenotypic correlations of 7-u>& body weights, fat weights, and percentage fat with fat probe values in a broiler grandparent population (Trial 4)1 n
7-w k weight
Fat weight
% Fat
Sex
1
S
I
S
1
S
I
S
Males Females
188 158
35 35
.18 .12
.07 .05
.13 .20
.05 .16
.09 .19
.01 .17
1
1 = Correlations based on individual values; S = correlations based on sire family averages.
Downloaded from http://ps.oxfordjournals.org/ at Univ of Iowa-Law Library on June 2, 2015
1 2
FAT PROBE AND ABDOMINAL FAT
1917
TABLE 8. Summary of associations of fat, body weight, and probe values Trial
Comparison
FW1
% FW1
Probe
BW1
Direction of changes
1 2 4 2 2 3 3
Broiler strains AC 2 : Diets A and B 3 High and low fat families Broiler: Diets A and B AC vs. Broiler Salt levels Male vs. female
* * * * * * =
* * * * * * *
= = = * * * =
= =
None None None t BW t t BW t t BW t I BW t
* * * *
% FW % FW % FW % FW
1
FW = Grams of abdominal fat; % FW = abdominal fat as a % of live body weight; BW = live body weight.
2
AC = Athens-Canadian randombred.
•Comparisons are significantly different (P<.05).
males and females. However, in this comparison, higher percentage fat was associated with lower body weight. Relative differences in amount of abdominal fat and body weight may also explain the correlation observed between body weight and probe values in this study and by others. Within a treatment group (and within sex), those birds with heavier body weights would tend to have larger fat pad weights and give a higher probe reading, even if the percentage fat values of the heavier groups as proportions were similar to those of lighter weight groups. There is some evidence that the actual mass of the abdominal fat may also influence the correlation between probe values and abdominal fat percentage, but this is inconclusive. In Trial 1, in which broilers had very large amounts (62 g) of abdominal fat and in Trial 2, where the AC birds has small amounts, correlations were low. In Strain 2 of the second trial and in Trial 3, where abdominal fat weights were moderate (14.5 to 38.7 g), correlations were much higher. However, this trend was not observed in Trial 4.
REFERENCES Becker, W. A., J. V. Spencer, L. W. Mirosh, and J. A. Verstrate, 1984. Genetic variation of abdominal fat, body weight, and carcass weight in a female broiler line. Poultry Sci. 63:607-611. Cahaner, A., and Z. Nitsan, 1986. Evaluation of simultaneous selection for live body weight and against abdominal fat in broilers. Poultry Sci. 64:1257-1263. Gyles, N. R., and A. Maeza, 1981. Abdominal fat in parents and broiler progeny. Poultry Sci. 60:1663. (Abstr.) Leclercq, B., J. C. Blum, and J. P. Boyer, 1980. Selecting broilers for low or high abdominal fat: initial observations. Br. Poult. Sci. 21:107-113. Mirosh, L. W., and W. A. Becker, 1981. Testing a caliper for measuring abdominal leaf fat thickness in live broiler chickens. Poultry Sci. 60:1698-1699. (Abstr.) Pym, R.A.E., 1981. Operator effect upon the prediction of abdominal fat in live broilers using a caliper measurement technique. Pages 156-157 in: Proc. 2nd Conf. Aust. Assoc. Anim. Breed. Genet. Sydney, Australia. Pym, R.A.E., and J. M. Thompson, 1980. A simple caliper technique for the estimation of abdominal fat in live broilers. Br. Poult. Sci. 21:281-286. Schwartzberg, N., R. L. Adams, and W. L. Stadelman, 1980. Correlation of fat thickness probe to actual weight of broiler fat pad. Poultry Sci. 59:1659. (Abstr.)
Downloaded from http://ps.oxfordjournals.org/ at Univ of Iowa-Law Library on June 2, 2015
3 Diet A contained 3,210 kcal metabolizable energy (ME)/26% protein; Diet B contained 3,256 kcal ME/ 21.2% protein.
1987 Poultry Science 66:1911-1917 INTRODUCTION
Excessive fat deposition in broilers is a problem of concern to the poultry industry. Genetic selection to decrease the proportion of fat mass while maintaining lean mass potentially can solve the problem. Although the study of Becker et al. (1984) indicated that abdominal fat could be reduced by selection, a correlated reduction in body weight was of concern. The use of genetic techniques such as selection index or independent culling may allow the reduction of abdominal fat without reduction in lean body mass (Leclercq et al., 1980; Cahaner and Nitsan, 1986). A measurement of abdominal fat that does not require sacrificing the individual is needed to maximize genetic selection progress. Pym and Thompson (1980) developed a caliper technique for the estimation of abdominal fat in live broilers. They reported a phenotypic correlation of .80 between caliper measure and proportion of abdominal fat. Schwartzberg et al. (1980),
'Supported by State and Hatch funds allocated to the Georgia Agricultural Experiment Stations of the University of Georgia and by a Binational Agricultural Research and Development Fund project.
using their version of the caliper also reported a high correlation (.77) between the fat probe measurement and fat pad as a percentage of live weight. These correlations indicate the calipers or fat probe should be a useful technique as an indirect method of selecting for decreased abdominal fat. Mirosh and Becker (1981) reported correlation coefficients between caliper values and abdominal fat weights of .29 and .45 at 41 and 48 days of age when the probe was inserted to 20 mm. A somewhat higher value (.54) was obtained when the probe was inserted at 35 mm and measurements were made at 48 days. However, Gyles and Maeza (1981) reported that correlations between the probe values and measures of abdominal fat were close to zero. These latter low correlations and personal communication with broiler breeders who attempted to use the calipers as a measure of fat thickness cast some doubt on the use of fat probe calipers as an indirect measure of abdominal fat. Some of the variables that were not controllabe in the studies that used the probe may contribute to the variation in the conclusions of the different studies. The objective of the present study was to determine how well abdominal fat can be estimated from caliper measures obtained from chickens over a wide range of body weights and amounts of abdominal fat. A modification of
1911
Downloaded from http://ps.oxfordjournals.org/ at Univ of Iowa-Law Library on June 2, 2015
ABSTRACT Four trials were conducted to compare the fat probe measure of abdominal fat with that of actual abdominal fat weights and percentage abdominal fat. Various ages, diets, and genetic groups were used to provide experimental chickens with a wide range in body weights and abdominal fat percentages. Mean values obtained by two operators using the probe on the same bird were similar (correlations of .70 to .79). Considering all comparisons, the arithmetic mean of the correlations between fat probe values and percentage abdominal fat was .24; values ranged from .01 to .44. In five comparisons of groups differing significantly in percentage abdominal fat, there were no significant differences in probe values, and correlations between fat and probe values were low. In three other comparisons of groups differing significantly in percentage abdominal fat, differences between probe values were significant and correlations between fat and probe values were moderate. In the comparison groups in which there was an association between probe and fat values, the magnitude of the differences in fat values was about four times as great as that of differences in probe values and the correlations were moderate (.28 to .44). In the comparisons in which higher abdominal fat was associated with higher probe values, higher body weight was associated with higher percentage abdominal fat. In comparisons where higher probe values were not associated with higher abdominal fat, higher body weights were not associated with higher percentage fat. (Key words: variation, fat probe, abdominal fat, broilers, body weight)
1912
WASHBURN AND STEWART
the fat probe caliper that does not require operator pressure to obtain a reading is described and correlations are compared between caliper measurements and actual weight and percentage of abdominal fat for birds with various abdominal fat and body weights. MATERIALS AND METHODS
Downloaded from http://ps.oxfordjournals.org/ at Univ of Iowa-Law Library on June 2, 2015
A modification was made of the fat probe caliper described by Pym and Thompson (1980). There are two primary differences between this caliper and that used by Pym and Thompson (1980). Operator pressure must be exerted to open the calipers. They close automatically after insertion of the cylindrical head into the cloaca and releasing of pressure on the handle. This removes the effect of the operator on the pressure exerted and consequently the reading. Another difference is that the flattened head is mounted so that it can pivot. This should allow better contact with the abdominal area. The cylindrical head and the graduated scale is similar to that of the caliper used by Pym and Thompson (1980). A further modification consisted of a depth set so that the probe would be inserted to a uniform depth in all individuals.The depth of 35 mm was chosen because studies of Mirosh and Becker (1981) indicated that a higher correlation coefficient with abdominal fat was obtained when the probe was inserted to this depth. Four trials were conducted to compare the fat probe measure of abdominal fat with that of actual abdominal fat weights and abdominal fat as a percentage of live weight. Various ages, diets, and genetic groups were used to provide experimental animals with variation in abdominal fat. In these trials, feed was restricted for approximately 8 h before determination of fat pad thickness, body weight, and abdominal fat pad weights. To obtain probe measurements, birds were held in the manner described by Pym (1981) except they were held on the operator's leg rather than on an inclined ramp. However, birds were held at an angle similar to that described by Pym (1981). After probe measurements and body weights were obtained, birds were killed, feathers removed, and abdominal fat (leaf and gizzard) removed and weighed to the nearest .1 g. The percentage fat was calculated as the abdominal fat weight/live body weight. In Trial 1, the variation in repeated probe measures was studied. Two operators independently measured the thickness of the fat pad on
the right and left side of males of two broiler strains at 63 days of age. These strains had previously been shown to differ in the amount of abdominal fat. Operator 1 was left handed, whereas Operator 2 was right handed. Both operators had a similar amount of prior experience with the use of the fat probe. In Trial 2, the same procedure was followed except that determinations were made at 7 wk of age. The Athens-Canadian randombred population (AC) and two broiler strains were used and each of these genetic groups was subdivided into dietary groups that received either Diet A or Diet B. Diet A contained 3,210 kcal metabolizable energy (ME) and 23% protein, whereas Diet B contained 3,256 kcal ME and 21.2% protein. In Trial 3, groups that differed in abdominal fat were obtained from an experiment designed to study the effects of dietary NaCl on water consumption and carcass fat. In this trial, which used the two commercial broiler strains used in Trial 1, percentages of abdominal fat were significantly different between sexes and between those groups receiving diets containing 1.6 or 2.4% added NaCl. There were no significant differences between broiler strains. At 49 days of age live body weights, probe values, fat weights, and percentages of fat were obtained from 120 males and 120 females fed the control diet and from 30 males each from the 1.6 and 2.4% NaCl treatments. In Trial 4, the usefulness of the fat probe in detecting differences in abdominal fat between families of a broiler grandparent line was evaluated. This trial was a satellite trial for a primary study involving the genetic basis of efficiency of feed utilization and percentage abdominal fat and used 346 progeny of 35 sire families of a commercial female grandparent line. Chicks were reared in floor pens until 3 wk of age, when they were placed in individual cages until 7 wk of age. They were fed ad libitum the standard University of Georgia (UGA) starter diet containing 3,120 kcal ME and 23% protein from 0 to 3 wk of age and the UGA finisher containing 3,256 kcal ME and 21.2% protein to 7 wk of age. At 7 wk of age, live body weights and fat probe readings were obtained, birds were processed, and the abdominal fat was removed and weighed. Abdominal fat probe values were obtained on the right side by Operator 1 of the previous trials.
FAT PROBE AND ABDOMINAL FAT
sides is presented in Tables 1 and 2. Mean values obtained by the two operators were similar and not statistically different for the different strain and trial comparisons. Correlations between the values obtained by the two operators for the same fat pads were .79 and .70 in Trials 1 and 2. Values for the right and left sides were not significantly (P=£.05) different in Trial 1, whereas in Trial 2 values for the left side were slightly (.35) but significantly (P«.05) larger than the right. This difference between scores was similar for both operators. Correlations between values for the left and right sides were high (.62 to .76) in both trials. The coefficient of variation was somewhat lower for Operator 2 than for Operator 1. In Trial 1, abdominal fat weights and percentages for Strain 1 were significantly greater than for Strain 2 (Table 3). However, differences in probe values between strains were not significant nor was there a trend toward higher probe values for the strain with more abdominal
RESULTS
Variation in probe measures between strains, operators, and right vs. left abdominal fat pad
TABLE 1. Differences in probe measures between strains, diets, operators, and 1right/left (R/L) fat pads Operator 2
Operator 1 Trial
Strain 1
Diet 2
R
L
7.6" 7.2"
7.4 7.2
7.0 7.2
7.2" 7.3"
7.4 a
7.4™ 16.4
7.3 a
7.1 a
7.2™ 14.2
3.5 3.9
4.2 4.3
3.8" 4.iy
4.0 3.9
4.4 4.3
4.2" 4.iy
3.7
4.2
4.0"
4.0
4.3
4.1"
5.2 5.6
5.4 5.7
5.3"
5.3 5.8
5.6 5.8
5.5"
s.ey
x BR2
5.4
5.6
5.5?
5.5
5.7
5.7y
x Trial 2 C.V.
4.5 a
4.9 b
4.7™ 22.7
4.7 a
5.0 b
4.9™ 21.2
R
L
7.6 7.3
7.7 7.0
7.5 a
X
X
(mm) BR1 BR2 x Trial 1 C.V.3 AC
BR2
A B
A B
s.sy
a ' b Indicated significant differences between means for right and left fat pad determined by an operator (P«.05). ^Indicates significant differences between broiler strains or diets within strains determined by an operator (P«.05). ' Indicates significant differences between means of all values obtained by different operators (P<.05). 1
BR1 and BR2 = Broiler strains 1 and 2; AC = Athens-Canadian randombred population.
2
A = Diet containing 3,210 kcal metabolizable energy (ME)/26% protein; B = diet containing, 3,256 kcal ME/21.2% protein. 3
C.V. = Coefficient of variation.
Downloaded from http://ps.oxfordjournals.org/ at Univ of Iowa-Law Library on June 2, 2015
In Trial 4, the value of the fat probe in detecting differences in abdominal fat between families was assessed two ways. In the first, subgroups of the 35 families were grouped in seven high-fat and seven low-fat families based on the sire family percentage fat means and their fat probe values compared. In the second, two types of phenotypic correlations were obtained: individual, using individual weight, probe and abdominal fat values; and sire family, using the mean values from the 35 sire families. Correlations were obtained separately for males and females. Statistical significances of differences were determined by analysis of variance and Duncan's multiple range tests. Analysis was based on within trial individual data for main effects and interactions.
1913
1914
WASHBURN AND STEWART TABLE 2. Phenotypic correlations between repeated probe measures in Trials 1 and 2 Operator 1
Operator 2
Trial
Left/right1
Left/right
o,o 2 2
1 2
.73** .75**
.62** .76**
.79** .70**
1
Left/right refers to correlation between values for the left and right abdominal fat pads.
2
O. Oj refers to correlation between mean of values obtained by Operators 1 and 2.
**P<.01.
Correlations between the probe values and the amount of abdominal fat in Trial 2 were moderate for the broiler strain but low for the AC population (Table 4). In Trial 3 there was no significant difference in probe values between sexes, which differed significantly in percentage abdominal fat (Table 5). There were, however, no significant differences in abdominal fat weights in this comparison (Table 5). In the comparison of the groups receiving diets containing different levels of salt, differences between probe values were significant for groups differing in abdominal fat weight and percentage. Correlations of probe values with abdominal fat weight or percentage abdominal fat were moderate (about .4). In Trial 4, differences between the high and low abdominal fat families in fat weight and percentage fat were highly significant (Tables 6 and 7), with large differences between groups.
TABLE 3. Means (± standard error) for fat probe values, abdominal fat weight, percentage abdominal fat, and body weight
Trial
Strain1
1
BR1 BR2
2
AC BR2
Abdominal fat weight
Abdominal fat
Body weight
(g) 68.0 ± 3.3 a 55.9 ± 2.8 b
(%)
7.21 ± .32 a 7.33 ± .19 a
2.37 ± ,10 a 1.93 ± .09 b
(g) 2844 ± 4 0 a 2893 ± 34 a
A B
4.04 ± .08 e 4.11 ± .06 c
2.7 ± 0 . 3 d 4.3 ± 0.4 C
0.39 ± .05 d 0.61 ± .05 c
679 ± 14 c 680± 12 c
A B
5.40 ± .08 b 5.73 ± .11*
14.5 ± 1.2b 25.9 ± 1.7a
0.89 ± .06 b 1.47 ± .09 a
1590± 4 0 b 1747 ± 39 a
Diet 2
Probe (mm)
Column means within trials with different superscripts differ significantly (P<.05). 1
BR1 and 2 = Broiler Strains 1 and 2; AC = Athens-Canadian randombred.
2 Diet A contained 3,210 kcal metabolizable energy (ME)/26% protein; Diet B contained 3,256 kcal ME/ 21.2% protein.
Downloaded from http://ps.oxfordjournals.org/ at Univ of Iowa-Law Library on June 2, 2015
fat. Correlations between probe values and amounts of abdominal fat were low (.24) regardless of operator. In Trial 2, Strain 2 broilers had significantly greater amounts and percentages of abdominal fat than AC birds (Table 3). Fat probe values were also significantly higher in the fatter broiler strain. However, the magnitude of the fat probe differences between strains was much less (broiler strain value was 137% of AC) than that of fat pad weight differences (broiler strain value was 577% of AC). Diet B resulted in significantly increased fat pad weights in both the broiler and randombreds (Table 3). There was a slight but significant increase in probe values in the broilers fed Diet B and a nonsignificant increase in the AC randombreds. The magnitude of the increase in probe values due to dietary differences was much less than that for fat weight or percentage.
1915
FAT PROBE AND ABDOMINAL FAT TABLE 4. Correlation of probe measurements with abdominal fat and body weights Abdominal fat
Abdominal fat Trial 2 Operator
Trial 1
o2
.22 .24 1
BR-2 (g) .52 .35
I
Body weight
Trial 2
Trial 2
AC
Trial 1
BR-2
AC
Trial 1
BR-2
AC
.22 .28
.16 .23
.36 .28
.18 .23
.32 .19
.64 .39
.23 .28
BR-2 = Broiler Strain 2; AC = Athens-Canadian randombred.
DISCUSSION
The results of this study as well as those of Mirosh and Becker (1981) and Gyles and Maeza (1981) indicate that the degree of association of fat probe values with percentage abdominal fat is not generally as high as that previously reported by Pym and Thompson (1980) and Schwartzberg et al. (1980). Correlations between probe values and percentage abdominal
TABLE 5. Means and correlations of fat probe values with abdominal fat weight, percentage abdominal fat, and body weight (Trial 3)
Statistic
Variable
X
Sex Males Females Dietary salt 1.6% NaCl 2.4% NaCl
r2
Sex Males Females Dietary salt 1.6% NaCl 2.4% NaCl
Probe
Abdominal fat weight
Abdominal fat
Body weight
(mm)
(g)
(%)
(g)
7.12 a 7.20 a
34.2 a 33.9 a
1.58 b 1.91 a
2,151 a l,765b
7.41 a 6.78 b
38.7 a 27.9 b
1.98 a 1.45 b
l,956 a l,917b
.41 .40
.44 .42
.34 .26
.26 .42
.37 .43
.28 .06
a ' Means between sexes or dietary salt level followed by different superscripts are significantly different (P<.05).
Downloaded from http://ps.oxfordjournals.org/ at Univ of Iowa-Law Library on June 2, 2015
fat in this study ranged from .01 to .44. The arithmetic mean of all the correlations in the 14 comparisons was .24. Possible reasons for the variation include differences in the construction and operation of the probes, including differences in experience, skill, and methodology of the user. Probes used in this study and by Gyles and Maeza (1981) and Mirosh and Becker (1981) were similar. They differed from that used by Pym and Thompson (1980) primarily in that they required no operator pressure to measure the abdominal fat thickness whereas the one used by Pym and Thompson (1980) required operator pressure to obtain the reading. The report of Pym (1981) emphasizes the effects of the operator skill and methodology on the repeatability of the values obtained. Repeata-
However, differences between these high and low abdominal fat families for probe values were small and not significant. Males had significantly higher body weights, less abdominal fat, and lower percentage fat than did the females. However, there were no significant differences between sexes in fat probe values.
1916
WASHBURN AND STEWART TABLE 6. Mean and coefficient of variation of l^wk body weights, fat weights, and percentage fat with fat probe values in a broiler grandparent population (Trial 4)'
Variable
n
7-wk weight
Males Females x H-families 2 x L-families
188 158
2,037" l,742b 1,900" 1,932"
CV
(g)
Fat weight
CV
% Fat
CV
Probe
(35) (30)
1.23" 1.69b 1.65" 1.09b
(33) (29)
7.34" 7.20" 7.58" 7.06"
(mm)
(g) (11) (9)
CV
25.0" 29.3b 31.3" 21.2b
(15) (15)
ab ' Means between males or H and L families followed by different superscripts are significantly different (P«.05). Coefficients of variation (CV) in parentheses.
H and L families refers to the seven sire families with the highest and the seven sire families with the lowest percentage fat pads.
bility of multiple caliper readings by Pym (1981), who was very experienced in the use of the calipers, ranged from .73 to .81; repeatability of one operator with relatively less experience ranged from .27 to .73 whereas that of a third operator with a similar amount of experience ranged from .01 to .29. In Trials 1 and 2 of the present study, two operators (one right handed and one left handed) measured both the right and left fat sides. Correlations between values for the left and right fat sides and between operators were high, indicating agreement between the repeated measures. In the present study, the association between probe values and abdominal fat values was greater in some comparisons than in others. The reason for differences in the association between percentage fat and probe values between trials may be the relative magnitude and the direction of the differences between fat weight and body weight in a particular comparison. These data are summarized in Table 8. In Trial 1 there were significant differences between the broiler strains in fat weight and
percentage fat, but not in probe values. In Trial 2 there were significant differences between the AC birds fed Diets A or B in fat weight and percentage fat weight but not in probe values. In Trial 4 there were significant differences between the high and low fat families in fat weight and percentage fat but not in probe values. In all these comparisons in which probe values were not different when fat values were different, body weights of the groups compared were not significantly different. In Trial 2 there were significant differences between broiler Strain 2 birds fed Diets A and B in fat weight, percentage fat, and probe values. Also, there were significant differences between the AC population and the broilers in fat weight, percentage fat, and probe values. In Trial 3, there were significant differences between broilers fed diets with different salt levels in fat weight, percentage fat, and probe values. In all these comparisons higher percentage fat was associated with higher body weight. In Trial 3, higher body weight was not associated with higher probe values in comparisons between
TABLE 7. Phenotypic correlations of 7-u>& body weights, fat weights, and percentage fat with fat probe values in a broiler grandparent population (Trial 4)1 n
7-w k weight
Fat weight
% Fat
Sex
1
S
I
S
1
S
I
S
Males Females
188 158
35 35
.18 .12
.07 .05
.13 .20
.05 .16
.09 .19
.01 .17
1
1 = Correlations based on individual values; S = correlations based on sire family averages.
Downloaded from http://ps.oxfordjournals.org/ at Univ of Iowa-Law Library on June 2, 2015
1 2
FAT PROBE AND ABDOMINAL FAT
1917
TABLE 8. Summary of associations of fat, body weight, and probe values Trial
Comparison
FW1
% FW1
Probe
BW1
Direction of changes
1 2 4 2 2 3 3
Broiler strains AC 2 : Diets A and B 3 High and low fat families Broiler: Diets A and B AC vs. Broiler Salt levels Male vs. female
* * * * * * =
* * * * * * *
= = = * * * =
= =
None None None t BW t t BW t t BW t I BW t
* * * *
% FW % FW % FW % FW
1
FW = Grams of abdominal fat; % FW = abdominal fat as a % of live body weight; BW = live body weight.
2
AC = Athens-Canadian randombred.
•Comparisons are significantly different (P<.05).
males and females. However, in this comparison, higher percentage fat was associated with lower body weight. Relative differences in amount of abdominal fat and body weight may also explain the correlation observed between body weight and probe values in this study and by others. Within a treatment group (and within sex), those birds with heavier body weights would tend to have larger fat pad weights and give a higher probe reading, even if the percentage fat values of the heavier groups as proportions were similar to those of lighter weight groups. There is some evidence that the actual mass of the abdominal fat may also influence the correlation between probe values and abdominal fat percentage, but this is inconclusive. In Trial 1, in which broilers had very large amounts (62 g) of abdominal fat and in Trial 2, where the AC birds has small amounts, correlations were low. In Strain 2 of the second trial and in Trial 3, where abdominal fat weights were moderate (14.5 to 38.7 g), correlations were much higher. However, this trend was not observed in Trial 4.
REFERENCES Becker, W. A., J. V. Spencer, L. W. Mirosh, and J. A. Verstrate, 1984. Genetic variation of abdominal fat, body weight, and carcass weight in a female broiler line. Poultry Sci. 63:607-611. Cahaner, A., and Z. Nitsan, 1986. Evaluation of simultaneous selection for live body weight and against abdominal fat in broilers. Poultry Sci. 64:1257-1263. Gyles, N. R., and A. Maeza, 1981. Abdominal fat in parents and broiler progeny. Poultry Sci. 60:1663. (Abstr.) Leclercq, B., J. C. Blum, and J. P. Boyer, 1980. Selecting broilers for low or high abdominal fat: initial observations. Br. Poult. Sci. 21:107-113. Mirosh, L. W., and W. A. Becker, 1981. Testing a caliper for measuring abdominal leaf fat thickness in live broiler chickens. Poultry Sci. 60:1698-1699. (Abstr.) Pym, R.A.E., 1981. Operator effect upon the prediction of abdominal fat in live broilers using a caliper measurement technique. Pages 156-157 in: Proc. 2nd Conf. Aust. Assoc. Anim. Breed. Genet. Sydney, Australia. Pym, R.A.E., and J. M. Thompson, 1980. A simple caliper technique for the estimation of abdominal fat in live broilers. Br. Poult. Sci. 21:281-286. Schwartzberg, N., R. L. Adams, and W. L. Stadelman, 1980. Correlation of fat thickness probe to actual weight of broiler fat pad. Poultry Sci. 59:1659. (Abstr.)
Downloaded from http://ps.oxfordjournals.org/ at Univ of Iowa-Law Library on June 2, 2015
3 Diet A contained 3,210 kcal metabolizable energy (ME)/26% protein; Diet B contained 3,256 kcal ME/ 21.2% protein.