- Email: [email protected]
Prediction of carcass composition in the rabbit
PII:
SO309-1740(96)00078-2
Meat Science, Vol. 44, Nos 1-2, 7s-83, 1996 Copyright 0 1996 Elsevier Science Ltd Printed in Great Britain. All rights reserved 03W-1740/96 %15.00+0.00
ELSEVIER
Prediction of Carcass Composition in the Rabbit Pilar Hernhdez, Departamento
M. Pla & A. Blasco*
de Ciencia Animal, Universidad Politkcnica de Valencia, P.O. Box 22012, Valencia 4607 1, Spain
(Received 12 January 1996; revised version received 28 May 1996; accepted 7 June 1996)
ABSTRACT Carcass composition of two synthetic rabbit breeds was predicted from retail cuts and external measurements by using regression equations. Breed R has a higher adult weight and reaches slaughter weight 1 week before breed V. Sixty rabbits of each breed were slaughtered when they (approximately) reached the Spanish commercial liveweight of 2kg. The carcasses were measured and retailed according to the norms of the World Rabbit Scientific Association. Rabbit carcass composition is well defined by meat percentage of the commercial carcass and ratio meat/bone. External measurements on the carcass, retail cuts and meat of retail cuts or muscular masses are all bad predictors of carcass meat percentage or ratio meat/bone (R2 < 0.53). The ratio meat/bone of the hind leg can give reasonable predictions for carcass meat percentage and meat/bone ratio (d = 0.60 and 0.69). Dissectible carcass fat weight and dissectible carcass fat percentage can be predicted by the perirenal fat weight (R2 = 0.77 and 0.69). Fat depots had a low predictive power for fat percentage of the dissected meat in the half carcass. Copyright 0 1996 Elsevier Science Ltd
INTRODUCTION
Although rabbit meat has been commercialised hitherto by selling the entire carcass, there is a recent market for retail cuts that points out the interest of carcass evaluation. Prediction equations for carcass composition have been developed in other species, but little work on this topic has been done in rabbits. Some of the studies were carried out in only one breed. Varewyck & Bouquet (1982) studied the relationships between the meat percentage of a part of the carcass and the whole carcass in the breed Blanc de Termonde, and Blasco et al. (1984) and Lambertini et al. (1991) proposed prediction equations for meat percentage and the ratio meat/bone in California and White New Zealand rabbits respectively, based on carcass measurements. These equations may not apply to other breeds. Lukefahr & Ozimba (1991) proposed prediction equations for two purebred and two crossbred types, based on live measurements, but the prediction was not good for the ratios (dressing- ir&ntage and ratio meat/bone). The aim of this paper w& to predict traits related to rabbit carcass quality in two breeds slaughtered at the same commercial carcass weight, but at a different stage of maturity. *To whom correspondence
should be addressed. 75
76
P. Hernhndez et al.
MATERIALS
AND METHODS
Animals
One-hundred and twenty rabbits from two synthetic breeds of different state of maturity were used in the experiment. The breed V was formed by crossing two commercial dam hybrids, and was selected for litter size for 15 generations. The breed R was formed by crossing a commercial terminal sire hybrid with a Californian breed and was selected on growth rate between the 4th and the 10th week of life for 10 generations. Breed R has a higher adult weight and attains the slaughter weight on average 1 week before breed V (Blasco & Gbmez, 1993). Sixty animals of each breed were reared in collective cages of eight animals from weaning to slaughter at the experimental farm of the University of Valencia. Weaning took place at 4 weeks of age for both breeds. Both breeds were fed ad libitum with a standard pelleted diet (16.5% protein, 3.4% fat, 15.5% fibre). Animals were slaughtered at 8 and 9 weeks of age for the breeds R and V, respectively, when they reached the Spanish commercial liveweight of 2 kg. No fasting was practised. Sex was taken at random, since there is no sexual dimorphism at these ages (Lukefahr et al., 1983; Lopez et al., 1992; Blasco & Gomez, 1993). Carcass measurements
The carcasses were refrigerated 24 h at 3°C and were measured and retailed according to the norms of the World Rabbit Scientific Association (WRSA) (Blasco et al., 1993). The European carcasses contain head, liver, lungs, thymus, oesophagus, heart and kidneys, which were removed to obtain the Reference carcass, which only contains meat, fat and bone. The following variables were measured on the carcass: LW: liveweight; CCW: chilled carcass weight 24 h after slaughter; DL: dorsal length (interval between the atlas vertebra and the 7th lumbar vertebra); TL: thigh length (between the 7th lumbar vertebra to the ischion tuberosity); LCL: lumbar circumference; L: length between the neural spines of atlas and last sacral vertebra; LL: length between the neural spines of first and last lumbar vertebra (loin length); HLL: hind leg length; SL: length from the wing-like portion of the ilium to the ischion tuberosity (sacral length); W: width between the 3rd trocanters of both femurs; RCW: reference carcass weight. The reference carcass was cut in the following points: cutpoint 1, section between the 7th and 8th thoracic vertebra; cutpoint 2, section between the last thoracic and the first lumbar vertebra, following the prolongation of the 12th rib when cutting the thoracic wall; cutpoint 3, section between the 6th and 7th lumbar vertebra, cutting the abdominal wall transversally to the vertebral column; cutpoint 4, separation of fore legs including insertion and thoracic muscles. The following variables were measured; FLW: fore leg weight including insertion and thoracic muscles; TW: thoracic cage weight; Pl-2: weight of the part of the carcass between cutpoints 1 and 2; P2-3: weight of the part of the carcass between cutpoints 2 and 3; HPW: hind part weight; HLW: weight of one hind leg. All the retail cuts were dissected, and the following variables measured: MW: meat weight of the reference carcass; BW, bone weight of the reference carcass; LDW, M. longissimus dorsi weight; AWW, abdominal wall weight; MHLW, meat weight of one hind leg; BHLW, bone weight of one hind leg; MFLW, meat weight of one fore leg; BFLW, bone weight of one fore leg; PMaW, Psoas major weight; PMiW, Psoas minor + Quadratus lumborum weight; SFaW, scapular fat weight; PFaW, perirenal fat weight; IFaW, inguinal fat weight; DFaW, dissectible fat weight of the reference carcass (DFaW = SFaW + PFaW + IFaW).
77
Prediction of rabbit carcass composition
The WRSA proposes two types of carcass division to facilitate comparisons with other research works (i) Technological (using the retail cuts from cutpoints 1,3 and 4) and (ii) Anatomical (using retail cuts from points 2 and 3). The Technological division has four retail cuts: FLW, TW, Pl-2 + P2-3 and HPW, whereas the Anatomical division has three retail cuts: FLW + TW + Pl-2, P2-3 and HPW. The prediction ability of each variable was evaluated by simple regression on MW, MW/CCW and MW/BW, in a model that also included breed effect and the interaction between breed and regression coefficient. Stepwise regression analyses were performed in order to obtain the best predictive equations with a minimum of variables for these traits. A test of collinearity was employed to detect high collinearity between the variables of the regression equations. When this happened, the stepwise regression was made by only taking into account groups of variables where the coefficient of correlation between them was lower than 0.9. The SAS package (SAS, 1990) was used.
RESULTS
AND DISCUSSION
Prediction of meat content and ratio meat/bone with individual measurements
The mean value of the variables analysed has been given by Pla et al. (1996). In most domestic species variation in carcass fatness is a main factor influencing carcass and meat quality. In rabbits, however, rabbit carcasses have an small dissectible fat content (3.1% and 2.5% of the carcass weight for the breeds V and R, respectively; Pla et al., 1996) and it is not normally used as a quality factor to classify the carcasses. Moreover there is less variation in value of cuts than in pig, beef sheep or poultry carcasses. Therefore the main criteria used to define rabbit carcass quality have been meat percentage in the carcass and the ratio meat/bone (Varewyck & Bouquet, 1982; Blasco et al., 1984; Lukefahr & Ozimba, 1991; Lambertini et al., 1991). As rabbit carcasses are sold with head, liver, lungs, thymus, oesophagus, heart and kidneys in Europe, meat percentage can be referred to the chilled carcass weight or to the Reference carcass weight, which is more similar to the carcass commercialised in the USA. Table 1 shows the correlations between these carcass quality traits. Due to the leanness of rabbit carcasses, meat/bone ratio is highly correlated with the meat percentage of the chilled and reference carcass. Thus, it seems that rabbit carcass quality can be well defined by only using chilled carcass weight (CCW) and meat percentage of the commercial carcass (MW/CCW) or ratio meat/bone (MW/BW). Several external measurements that do not depreciate the carcass were chosen in order to predict these carcass quality traits. None of them showed a good predictive power on meat percentage or ratio meat/bone (Table 2). Some of them were highly related to meat weight, but carcass weight and even liveweight, were good predictors for this trait, and TABLE 1
Correlations Between the Main Carcass Quality Characters. (Breed V Above Diagonal, Breed R Below Diagonal.) luw/ccw
MW/CCW MW/RCW MW/BW
MWIRCW 0.83
0.81 0.79
MWIBW 0.83 0.93
0.97
CCW: Chilled carcass weight 24 h after slaughter; RCW: reference carcass weight; MW: reference carcass meat weight; BW: reference carcass bone weight.
78
P. Hermindez et al.
linear combinations of these measurements do not improve this prediction much, as observed by Bochno et al. (1979) and Janiszewska & Bochno (1979). Neither ratios (LCL/L, W/L) nor linear combinations of external measurements were related to meat percentage or ratio meat/bone. There are few differences in regression coefficients between both breeds, since the interaction breedxcoefficient was non significant in most cases. Blasco et TABLE 2 Coefficient of Determination (R’) and Residual Standard Deviation (RSD) of the Regression of External Measurements on Meat Weight (MW), Meat Percentage of the Carcass (MWjCCW) and Ratio Meat/Bone (MW/BW), Using Two Breeds of Rabbit
LW ccw RCW L LL HLL W DL TL SL LCL LCL/L W/L
R2 MW
RSD
hf.
R*MW,CCW
RSD
ht.
0.78
34.6
0.95 0.97
16.9 12.0
ns ns *
0.55 0.27 0.55 0.60 0.35 0.39 0.46 0.81 0.42 0.22
48.9 62.5 49.3 46.3 58.9 57.2 52.0 31.9 55.7 64.5
ns * ns ns ns ns ns * * **
0.31 0.182 0.192 0.102 0.107 0.222 0.167 0.083 0.175 0.162 0.204 0.179 0.129
0.016 0.015 0.014 0.016 0.016 0.014 0.015 0.017 0.016 0.015 0.014 0.015 0.016
ns ns * ns ns ns ns ns ns ns ns ns ns
R*~w,ew
0.21 0.30 0.31 0.20 0.17 0.34 0.26 0.16 0.26 0.26 0.34 0.29 0.20
RSD
ht.
0.147 0.139 0.138 0.148 0.151 0.134 0.142 0.151 0142 0.143 0.134 0.139 0.148
ns * * ns ns ns ns ns ns ns ns ns ns
Int.: Significance of the interaction breedxregression coefficient. ns: non significant. *: PCO.05. **: PCO.01. LW: Liveweight; CCW: chilled carcass weight 24 h after slaughter; RCW: reference carcass weigh; L: length between the neural spines of atlas and last sacral vertebra; LL: length between the neural spines of first and last lumbar vertebra (loin length); HLL: hind leg length; W: width between the 3rd trocanters of both femurs; DL: dorsal length (interval between the atlas vertebra and the 7th lumbar vertebra); TL: thigh length (between the 7th lumbar vertebra to the ischion tuberosity); SL: length from the wing-like portion of the ilium to the ischion tuberosity (sacral length); LCL: lumbar circumference. TABLE 3 Correlations Between the Weight of the Fore (FPW), Intermediate (IPW) and Hind (HPW) Parts of the Carcass. (Upper Diagonal, Anatomical Division, Down Diagonal Technological Division. Diagonal, Correlations between Anatomical and Technological Division.) Breed V
FPW IPW HPW
Breed R
FPW
IPW
HPW
FPW
IPW
HPW
0.93 0.87 0.90
0.83 0.98 0.93
0.94 0.89 1
0.97 0.94 0.94
0.87 0.98 0.95
0.94 0.91 1
Anatomical division: FPW = FLW + TW + Pl-2, IPW = P2-3. Technological division: FPW = FLW, IPW = Pl-2 + P2-3. FLW: Fore leg weight including insertion and thoracic muscles; TW: thoracic cage weight; Pl-2: weight of the part of the carcass between cutpoints 1 and 2; P2-3: weight of the part of the carcass between cutpoints 2 and 3; HPW: hind part weight.
79
Prediction of rabbit carcass composition
al. (1984) on a California breed and Lambertini et al. (1991) on a White New Zealand breed did not find either any good relationship between external carcass measurements and meat percentage or ratio meat/bone. Relationships between technological and anatomical carcass division
The WRSA describes two types of carcass division (Blasco et al., 1993). The Technological division corresponds to the commercial cuts, whereas the Anatomical division has been often used in carcass quality experiments. The recommendations of the WRSA describe how to cut the carcass to obtain both, as we did in this experiment. Table 3 shows that the correlations between the weight of the fore, intermediate and hind parts of both divisions are near 1, thus only one division would be enough for carcass studies. As the technological division is more related with the commercial situation, it will be used in this paper. The weights of the retail cuts are also highly correlated. None of them is a good predictor of meat percentage or ratio meat/bone (Table 4), a result which agrees with the results of Blasco et al. (1984) and Lambertini et al. (1991), but they were highly related to meat weight of the carcass, as observed by Blasco et al. (1984) and Niedzwiadek (1984). Prediction equations of meat content and ratio meat/bone using several variables
Stepwise regression analysis for prediction of carcass meat percentage did not select exactly the same set of variables for both breeds, Table 5 shows the best equations for each breed obtained (Eqs 1 and 2). In order to find a single equation for both breeds, the selected set of variables of one breed was used for predictions in the other breed (Table 5, Eqs 3 and 4). This is to examine whether it is possible to find a general prediction equation TABLE 4 Coefficient of Determination (R2) and Residual Standard Deviation (RSD) of the Regression of
Products of Retail Cuts of the Technological Division and Products of the Dissection of the Carcass on Meat Weight (MW), Meat Percentage of the Carcass (MW/CCW) and Ratio Meat/Bone (MW/ SW), Using Two Breeds of Rabbit R2 MW RSD
FPW IPW HPW AWW PMiW PMaW LDW MFLW MHLW MFLW/BFLW MHLW/BHLW
0.90 0.93 0.95 0.76 0.61 0.90 0.93 0.74 0.95 0.22 0.40
23.3 19.2 17.2 36.0 45.7 23.3 19.2 37.0 17.2 64.5 56.4
ht.
** ** ns ns ns ** ** ns ns ns *
RZwwjccw
RSD
Znt.
R2Mw,ew
RSD
ht.
0.46 0.44 0.46 0.35 0.44 0.53 0.51 0.46 0.51 0.41 0.60
0.014 0.015 0.014 0.016 0.015 0.013 0.014 0.014 0.014 0.015 0.012
** * ns
0.33 0.30
0.135 0.138 0.138 0.145 0.133 0.124 0.130 0.129 0.131 0.114 0.092
* * * *
* ns ns * ns ns ns ns
0.31 0.23 0.35 0.44 0.38 0.39 0.37 0.052 0.69
ns ns * ns ns ns ns
Int.: Significance of the interaction breedxregression coefficient. ns: non significant. *: P < (0.05). **: P< (0.01). Fore (FPW), intermediate (IPW) and hind (HPW) part weight of the carcass. AWW, abdominal wall weight; PMiW: Psoas minor + Quadratus lumborum weight; PMaW: Psoas major weight; LDW: weight of the M. Longissimus; MFLW: meat weight of one fore leg; MHLW: meat weight of one hind leg; BFLW: bone weight of one fore leg; BHLW: bone weight of one hind leg.
80
P. Hernhndez et al.
that would hold for both breeds, given that when they reach the commercial slaughter weight they are not at the same stage of maturity. Table 5 (Eqs 16) shows that the differences in accuracy when the equations applicable to of the other breed were used were small, and the R2 values were acceptable in all the cases (ranging from 0.74 to 0.89). Moreover, when the same set of variables was used for both lines, the regression coefficients were not significantly different. This is because some of the variables that were candidates to be selected are highly correlated. Thus, to choose one over another has no appreciable influence in the accuracy of the regression. When fitting the variables of Eqs 1 and 2 to both lines, R2 of 0.88 and 0.86 were found (Table 5, Eqs 5 and 6). However, the high value of the R2 obtained in all of these equations is a result of the poor conditions of this set of data, since the relationships of these variables is very high (see, for example, Table 6, Eqs 2, 3, 5). A test of collinearity showed that the first eigenvalue accounted for more than 99% of the variation of all variables. This quirk of the regression analysis was first observed by Hamilton (1987) and Routledge (1990) for two variables, and it was generalised by Cuadras (1993) for multiple regression. As the stepwise procedure does not prevent this phenomena, the ill-condition state of the data can be unperceived, and this seems to have been the case for some of the equations given by Blasco et al. (1984) and Lambertini ef al. (1991). When the variables are selected under the condition that their relationship should not be higher than 0.9, the resulting equation had a R2 lower than 0.70 (Table 5, Eq. 7), and it includes the meat/bone ratio of the hind leg. This difficulty of predicting meat percentage originates from the high correlation between numerator and denominator of the ratio MW/CCW (0.97 in both lines), which produces in this trait a low variation independent of CCW, a variable highly-correlated to TABLE 5 Prediction
Breed
Equations
of the Carcass Meat Percentage (R2) and Residual Standard
(MWjCCW). Coefficient Deviation (RSD)
Equation
of Determination
RJ
RSD Xl@
IV
2R 3v 4R
SV+R 6V+R
7V+R
0.155x 10e2 PMaW+0.262x lo-* MHLW-0.320x 10e2 HLW+5.18~1@~ HPW-0.467~10-~ CCW+O.514~10-~ RCW + 0.525x 1O-3 HLL + 0.452 0.351 x lO-2 MHLW-0.523x lO-3 CCW +0.688x lO-3 RCW-0.375x10-* HLW+0.387~10-~ HPW-tO.517 0.356x lO-2 MHLW-0.474x 1O-3 CCW +0.557x lO-3 RCW-0.384x lO-2 HLW+0.545x lO-3 HPW+0.537 0.613x 10e3 PMaW + 0.330x 10M2MHLW-0.358x 10m2HLW + 0.394~10-~ HPW-0.513x 10-3CCW+0.648x10-3 RCW+0.330xlO-3 HLL + 0.464 0.363x 1O-2 MHLW-0.499x lO-3 CCW+O.619x lo-) RCW-0.390x 1O-2 HLW+0.493x lO-3 HPW+0.527 0.105x 10e2PMaW + 0.311 x 10e2 MHLW - 0.354x 10e2 HLW + 0.506x 1O-3 HPW - 0.488x lO-3 CCW + 0.566x lO-3 RCW +0.444x lO-3 HLL + 0.453 0.021 MHLW/BHLW + 0.214x lO-2 PMaW - 0.407x 10m4LW +0.622x lO-3 LCL
0.801
7.07
0.880
6.41
0.744
7.89
0.886
6.35
0.864
7.20
0.88 1
6.80
0.693
10.7
PMaW: Psoas major weight; MHLW(g): meat weight of one hind leg; HLW(g): weight of one hind leg; BWHL: bone weight of one hind leg; HPW(g): hind part weight of the carcass; LW(g): liveweight; CCW(g): chilled carcass weight 24 h after slaughter; RCW(g): reference carcass weight; HLL(mm): hind leg length; LCL(mm): lumbar circumference.
81
Prediction of rabbit carcass composition TABLE 6
Correlations
Among Variables (Breed V Above Diagonal, Breed R Below Diagonal) MHL W
MHLW HLW HPW ccw RCW
0.990 0.981 0.949 0.971
HLW
HPW
ccw
RCW
0.980
0.957 0.982
0.932 0.957 0.963
0.952 0.973 0.976 0.990
0.991 0.961 0.981
0.961 0.978
0.986
MHLW: Meat weight of one hind leg; HLW: weight of one hind leg; HPW: hind part weight of the carcass; CCW: chilled carcass weight 24 h after slaughter; RCW: reference carcass weight.
the other variables studied. In other species, the presence of a much higher fatness gives other meaning to the carcass meat percentage. When the ratio meat/bone of the carcass was predicted only the ratio meat/bone of the hind leg gave a reasonable prediction, which was improved a little when adding other measurements not highly correlated with it (R* changed from 0.69 to 0.71). By using the meat of both legs, Varewyck & Bouquet (1982) give R*= 0.66. Given this, the meat/bone ratio of the hind leg can be used to character&e carcass composition and make predictions, as previously suggested by Rouvier (1970). This ratio is the only recommended measurement for carcass composition prediction by the WRSA. Thus, rabbit carcass quality is largely reduced to a two component problem, in which CCW and MW/BW explain the main variation of interest. Individual muscles or other products from dissection are not good predictors of meat percentage or the ratio meat/bone, with the possible exception of the ratio meat/bone of the hind leg, which has a moderately high R* for both traits (Table 4). Prediction of carcass dissectible fat and fat percentage of the meat Total dissectible fat of the carcass is predicted well by using the perirenal fat depot (Table 7). The regression of the perirenal fat depot on carcass dissectible fat percentage shows an R*=0.69. Varewyck & Bouquet (1982) give an R*=0.86when the whole fat tissue of the intermediate part is dissected. The fat contained in the meat (intramuscular and intermuscular fat) is a factor of meat quality, but it is laborious and expensive to evaluate. The predictive power of dissectible fat deposits on fat percentage of the meat dissected in a half-carcass are shown in Table 8. Although the R* values are not good, the relationships between total dissectible fat TABLE 7
Coefficient of Determination (R2) and Residual Standard Deviation (RSD) of the Regression of Fat Deposits on Dissectible Carcass Fat Weight (DFaW) and Dissectible Carcass Fat Percentage (DFaW/CCW)
R2CJW SFaW IFaW PFaW
0.64 0.69 0.77
RSD
Int.
5.52 5.10 4.39
ns ns ns
R*~F~W,CCW
0.54 0.71 0.69
RSD
ht.
0.438 0.343 0.360
ns ns ns
Int.: Significance of the interaction breedxregression coefficient; ns: non significant; *: P < (0.05). **: P < (0.01); SFaW: scapular fat weight; PFaW: perirenal fat weight; IFaW: inguinal fat weight.
82
P. Herncindez et al
TABLE 8 Coefficient of Determination (R2) and Residual Standard Deviation (RSD) of the Regression of the Fat Deposits on Fat Percentage of the Meat (MFaP)
R2LfFd SFaW IFaW PFaW DFaW
0.391 0.550 0.398 0.554
Int.: Significance of the interaction breedxregression
RSD
Int
0.791 0.680 0.786 0.677
* ** ** **
coefficient; ns: non significant. *: P< (0.05). **:
P < (0.01).
SFaW: Scapular fat weight; PFaW: perirenal fat weight; IFaW: inguinal fat weight; DFaW: total dissectible carcass fat weight.
deposits weight and fat percentage of the meat is relatively high (corresponding to a correlation coefficient of 0.74), which could have some implications in selection programs in the future, since fat content is a meat quality criterion. As an increment in the fat contained in the meat will also increase fat carcass content, its economic interest would largely depend on whether and how meat quality is paid in the future.
CONCLUSIONS
This work supports the recommendations of the WRSA on carcass quality characterisation and suggests ways of simplifying them. Two breeds of different degrees of maturity have been used in the experiment, and the conclusions apply for both: carcass composition can be clearly defined by using carcass weight, the ratio meat/bone of the hind leg and the perirenal fat deposit weight. Other multiple regression prediction equations do not show significant improvements.
ACKNOWLEDGEMENTS We are grateful to Luis Valero for his help in the slaughter process and to J. Sahuquillo for the care of the animals. This research is included in the project CAICYT AGF93-0634, financed by the Spanish Ministry of Education and Science.
REFERENCES Blasco, A., Estany, J. & Baselga, M. (1984). Ann. Zootech., 33, 161. Blasco, A., Ouhayoun, J. & Masoero, G. (1993). World Rabbit Sci., 1, 3. Blasco, A. & Gomez, E. (1993). Anim. Prod., 57, 332. Bochno, R., Lewczuk, A. & Janiszewska, M. (1979). Rcz. Nauk Zoot., T6(1), 175. Cuadras, C. M. (1993). Am. Stat., 47, 256. Hamilton, D. (1987). Am. Stat., 41, 129. Janiszewska, M. & Bochno, R. (1979). Zesz. Nauk. ART. Zoot., 19, 119. Lambertini, L., Benassi, M. C. & Zaghini, G. (1991). Rivista di coniglicoltura, 8, 35. Lopez, M. C., Sierra, I. & Lite, M. J. (1992). Options Mediterran., 17, 75. Lukefahr, S. D., Hohenboken, W. D., Cheeke, P. R. & Patton, N. M. (1983). J. Anim. Sci., 57, 899. Lukefahr, S. D. & Ozimba, C. E. (1991). Lives. Prod. Sci., 29, 323.
Prediction of rabbit carcass composition Niedzwiadeck, S. (1984). In: Proc. 3rd World Rabbit Congress, Roma, 2, 585. Pia, M., Hemhdez, P. & Blasco A. 1996. Meat Sci., 44, 85. Rouvier, R. (1970). Ann. GhPt. SIX Anim., 2, 325. Routledge, R. D. (1990). ht. J. Math. Educ. Sci. Technol., 21,403. SAS (1990). SAS/STAT User’s Guide, Version 6, SAS Institute Inc., Cary, NC, USA. Varewyck, H. & Bouquet, Y. (1982). Ann. Zootech., 31, 257.
83
SO309-1740(96)00078-2
Meat Science, Vol. 44, Nos 1-2, 7s-83, 1996 Copyright 0 1996 Elsevier Science Ltd Printed in Great Britain. All rights reserved 03W-1740/96 %15.00+0.00
ELSEVIER
Prediction of Carcass Composition in the Rabbit Pilar Hernhdez, Departamento
M. Pla & A. Blasco*
de Ciencia Animal, Universidad Politkcnica de Valencia, P.O. Box 22012, Valencia 4607 1, Spain
(Received 12 January 1996; revised version received 28 May 1996; accepted 7 June 1996)
ABSTRACT Carcass composition of two synthetic rabbit breeds was predicted from retail cuts and external measurements by using regression equations. Breed R has a higher adult weight and reaches slaughter weight 1 week before breed V. Sixty rabbits of each breed were slaughtered when they (approximately) reached the Spanish commercial liveweight of 2kg. The carcasses were measured and retailed according to the norms of the World Rabbit Scientific Association. Rabbit carcass composition is well defined by meat percentage of the commercial carcass and ratio meat/bone. External measurements on the carcass, retail cuts and meat of retail cuts or muscular masses are all bad predictors of carcass meat percentage or ratio meat/bone (R2 < 0.53). The ratio meat/bone of the hind leg can give reasonable predictions for carcass meat percentage and meat/bone ratio (d = 0.60 and 0.69). Dissectible carcass fat weight and dissectible carcass fat percentage can be predicted by the perirenal fat weight (R2 = 0.77 and 0.69). Fat depots had a low predictive power for fat percentage of the dissected meat in the half carcass. Copyright 0 1996 Elsevier Science Ltd
INTRODUCTION
Although rabbit meat has been commercialised hitherto by selling the entire carcass, there is a recent market for retail cuts that points out the interest of carcass evaluation. Prediction equations for carcass composition have been developed in other species, but little work on this topic has been done in rabbits. Some of the studies were carried out in only one breed. Varewyck & Bouquet (1982) studied the relationships between the meat percentage of a part of the carcass and the whole carcass in the breed Blanc de Termonde, and Blasco et al. (1984) and Lambertini et al. (1991) proposed prediction equations for meat percentage and the ratio meat/bone in California and White New Zealand rabbits respectively, based on carcass measurements. These equations may not apply to other breeds. Lukefahr & Ozimba (1991) proposed prediction equations for two purebred and two crossbred types, based on live measurements, but the prediction was not good for the ratios (dressing- ir&ntage and ratio meat/bone). The aim of this paper w& to predict traits related to rabbit carcass quality in two breeds slaughtered at the same commercial carcass weight, but at a different stage of maturity. *To whom correspondence
should be addressed. 75
76
P. Hernhndez et al.
MATERIALS
AND METHODS
Animals
One-hundred and twenty rabbits from two synthetic breeds of different state of maturity were used in the experiment. The breed V was formed by crossing two commercial dam hybrids, and was selected for litter size for 15 generations. The breed R was formed by crossing a commercial terminal sire hybrid with a Californian breed and was selected on growth rate between the 4th and the 10th week of life for 10 generations. Breed R has a higher adult weight and attains the slaughter weight on average 1 week before breed V (Blasco & Gbmez, 1993). Sixty animals of each breed were reared in collective cages of eight animals from weaning to slaughter at the experimental farm of the University of Valencia. Weaning took place at 4 weeks of age for both breeds. Both breeds were fed ad libitum with a standard pelleted diet (16.5% protein, 3.4% fat, 15.5% fibre). Animals were slaughtered at 8 and 9 weeks of age for the breeds R and V, respectively, when they reached the Spanish commercial liveweight of 2 kg. No fasting was practised. Sex was taken at random, since there is no sexual dimorphism at these ages (Lukefahr et al., 1983; Lopez et al., 1992; Blasco & Gomez, 1993). Carcass measurements
The carcasses were refrigerated 24 h at 3°C and were measured and retailed according to the norms of the World Rabbit Scientific Association (WRSA) (Blasco et al., 1993). The European carcasses contain head, liver, lungs, thymus, oesophagus, heart and kidneys, which were removed to obtain the Reference carcass, which only contains meat, fat and bone. The following variables were measured on the carcass: LW: liveweight; CCW: chilled carcass weight 24 h after slaughter; DL: dorsal length (interval between the atlas vertebra and the 7th lumbar vertebra); TL: thigh length (between the 7th lumbar vertebra to the ischion tuberosity); LCL: lumbar circumference; L: length between the neural spines of atlas and last sacral vertebra; LL: length between the neural spines of first and last lumbar vertebra (loin length); HLL: hind leg length; SL: length from the wing-like portion of the ilium to the ischion tuberosity (sacral length); W: width between the 3rd trocanters of both femurs; RCW: reference carcass weight. The reference carcass was cut in the following points: cutpoint 1, section between the 7th and 8th thoracic vertebra; cutpoint 2, section between the last thoracic and the first lumbar vertebra, following the prolongation of the 12th rib when cutting the thoracic wall; cutpoint 3, section between the 6th and 7th lumbar vertebra, cutting the abdominal wall transversally to the vertebral column; cutpoint 4, separation of fore legs including insertion and thoracic muscles. The following variables were measured; FLW: fore leg weight including insertion and thoracic muscles; TW: thoracic cage weight; Pl-2: weight of the part of the carcass between cutpoints 1 and 2; P2-3: weight of the part of the carcass between cutpoints 2 and 3; HPW: hind part weight; HLW: weight of one hind leg. All the retail cuts were dissected, and the following variables measured: MW: meat weight of the reference carcass; BW, bone weight of the reference carcass; LDW, M. longissimus dorsi weight; AWW, abdominal wall weight; MHLW, meat weight of one hind leg; BHLW, bone weight of one hind leg; MFLW, meat weight of one fore leg; BFLW, bone weight of one fore leg; PMaW, Psoas major weight; PMiW, Psoas minor + Quadratus lumborum weight; SFaW, scapular fat weight; PFaW, perirenal fat weight; IFaW, inguinal fat weight; DFaW, dissectible fat weight of the reference carcass (DFaW = SFaW + PFaW + IFaW).
77
Prediction of rabbit carcass composition
The WRSA proposes two types of carcass division to facilitate comparisons with other research works (i) Technological (using the retail cuts from cutpoints 1,3 and 4) and (ii) Anatomical (using retail cuts from points 2 and 3). The Technological division has four retail cuts: FLW, TW, Pl-2 + P2-3 and HPW, whereas the Anatomical division has three retail cuts: FLW + TW + Pl-2, P2-3 and HPW. The prediction ability of each variable was evaluated by simple regression on MW, MW/CCW and MW/BW, in a model that also included breed effect and the interaction between breed and regression coefficient. Stepwise regression analyses were performed in order to obtain the best predictive equations with a minimum of variables for these traits. A test of collinearity was employed to detect high collinearity between the variables of the regression equations. When this happened, the stepwise regression was made by only taking into account groups of variables where the coefficient of correlation between them was lower than 0.9. The SAS package (SAS, 1990) was used.
RESULTS
AND DISCUSSION
Prediction of meat content and ratio meat/bone with individual measurements
The mean value of the variables analysed has been given by Pla et al. (1996). In most domestic species variation in carcass fatness is a main factor influencing carcass and meat quality. In rabbits, however, rabbit carcasses have an small dissectible fat content (3.1% and 2.5% of the carcass weight for the breeds V and R, respectively; Pla et al., 1996) and it is not normally used as a quality factor to classify the carcasses. Moreover there is less variation in value of cuts than in pig, beef sheep or poultry carcasses. Therefore the main criteria used to define rabbit carcass quality have been meat percentage in the carcass and the ratio meat/bone (Varewyck & Bouquet, 1982; Blasco et al., 1984; Lukefahr & Ozimba, 1991; Lambertini et al., 1991). As rabbit carcasses are sold with head, liver, lungs, thymus, oesophagus, heart and kidneys in Europe, meat percentage can be referred to the chilled carcass weight or to the Reference carcass weight, which is more similar to the carcass commercialised in the USA. Table 1 shows the correlations between these carcass quality traits. Due to the leanness of rabbit carcasses, meat/bone ratio is highly correlated with the meat percentage of the chilled and reference carcass. Thus, it seems that rabbit carcass quality can be well defined by only using chilled carcass weight (CCW) and meat percentage of the commercial carcass (MW/CCW) or ratio meat/bone (MW/BW). Several external measurements that do not depreciate the carcass were chosen in order to predict these carcass quality traits. None of them showed a good predictive power on meat percentage or ratio meat/bone (Table 2). Some of them were highly related to meat weight, but carcass weight and even liveweight, were good predictors for this trait, and TABLE 1
Correlations Between the Main Carcass Quality Characters. (Breed V Above Diagonal, Breed R Below Diagonal.) luw/ccw
MW/CCW MW/RCW MW/BW
MWIRCW 0.83
0.81 0.79
MWIBW 0.83 0.93
0.97
CCW: Chilled carcass weight 24 h after slaughter; RCW: reference carcass weight; MW: reference carcass meat weight; BW: reference carcass bone weight.
78
P. Hermindez et al.
linear combinations of these measurements do not improve this prediction much, as observed by Bochno et al. (1979) and Janiszewska & Bochno (1979). Neither ratios (LCL/L, W/L) nor linear combinations of external measurements were related to meat percentage or ratio meat/bone. There are few differences in regression coefficients between both breeds, since the interaction breedxcoefficient was non significant in most cases. Blasco et TABLE 2 Coefficient of Determination (R’) and Residual Standard Deviation (RSD) of the Regression of External Measurements on Meat Weight (MW), Meat Percentage of the Carcass (MWjCCW) and Ratio Meat/Bone (MW/BW), Using Two Breeds of Rabbit
LW ccw RCW L LL HLL W DL TL SL LCL LCL/L W/L
R2 MW
RSD
hf.
R*MW,CCW
RSD
ht.
0.78
34.6
0.95 0.97
16.9 12.0
ns ns *
0.55 0.27 0.55 0.60 0.35 0.39 0.46 0.81 0.42 0.22
48.9 62.5 49.3 46.3 58.9 57.2 52.0 31.9 55.7 64.5
ns * ns ns ns ns ns * * **
0.31 0.182 0.192 0.102 0.107 0.222 0.167 0.083 0.175 0.162 0.204 0.179 0.129
0.016 0.015 0.014 0.016 0.016 0.014 0.015 0.017 0.016 0.015 0.014 0.015 0.016
ns ns * ns ns ns ns ns ns ns ns ns ns
R*~w,ew
0.21 0.30 0.31 0.20 0.17 0.34 0.26 0.16 0.26 0.26 0.34 0.29 0.20
RSD
ht.
0.147 0.139 0.138 0.148 0.151 0.134 0.142 0.151 0142 0.143 0.134 0.139 0.148
ns * * ns ns ns ns ns ns ns ns ns ns
Int.: Significance of the interaction breedxregression coefficient. ns: non significant. *: PCO.05. **: PCO.01. LW: Liveweight; CCW: chilled carcass weight 24 h after slaughter; RCW: reference carcass weigh; L: length between the neural spines of atlas and last sacral vertebra; LL: length between the neural spines of first and last lumbar vertebra (loin length); HLL: hind leg length; W: width between the 3rd trocanters of both femurs; DL: dorsal length (interval between the atlas vertebra and the 7th lumbar vertebra); TL: thigh length (between the 7th lumbar vertebra to the ischion tuberosity); SL: length from the wing-like portion of the ilium to the ischion tuberosity (sacral length); LCL: lumbar circumference. TABLE 3 Correlations Between the Weight of the Fore (FPW), Intermediate (IPW) and Hind (HPW) Parts of the Carcass. (Upper Diagonal, Anatomical Division, Down Diagonal Technological Division. Diagonal, Correlations between Anatomical and Technological Division.) Breed V
FPW IPW HPW
Breed R
FPW
IPW
HPW
FPW
IPW
HPW
0.93 0.87 0.90
0.83 0.98 0.93
0.94 0.89 1
0.97 0.94 0.94
0.87 0.98 0.95
0.94 0.91 1
Anatomical division: FPW = FLW + TW + Pl-2, IPW = P2-3. Technological division: FPW = FLW, IPW = Pl-2 + P2-3. FLW: Fore leg weight including insertion and thoracic muscles; TW: thoracic cage weight; Pl-2: weight of the part of the carcass between cutpoints 1 and 2; P2-3: weight of the part of the carcass between cutpoints 2 and 3; HPW: hind part weight.
79
Prediction of rabbit carcass composition
al. (1984) on a California breed and Lambertini et al. (1991) on a White New Zealand breed did not find either any good relationship between external carcass measurements and meat percentage or ratio meat/bone. Relationships between technological and anatomical carcass division
The WRSA describes two types of carcass division (Blasco et al., 1993). The Technological division corresponds to the commercial cuts, whereas the Anatomical division has been often used in carcass quality experiments. The recommendations of the WRSA describe how to cut the carcass to obtain both, as we did in this experiment. Table 3 shows that the correlations between the weight of the fore, intermediate and hind parts of both divisions are near 1, thus only one division would be enough for carcass studies. As the technological division is more related with the commercial situation, it will be used in this paper. The weights of the retail cuts are also highly correlated. None of them is a good predictor of meat percentage or ratio meat/bone (Table 4), a result which agrees with the results of Blasco et al. (1984) and Lambertini et al. (1991), but they were highly related to meat weight of the carcass, as observed by Blasco et al. (1984) and Niedzwiadek (1984). Prediction equations of meat content and ratio meat/bone using several variables
Stepwise regression analysis for prediction of carcass meat percentage did not select exactly the same set of variables for both breeds, Table 5 shows the best equations for each breed obtained (Eqs 1 and 2). In order to find a single equation for both breeds, the selected set of variables of one breed was used for predictions in the other breed (Table 5, Eqs 3 and 4). This is to examine whether it is possible to find a general prediction equation TABLE 4 Coefficient of Determination (R2) and Residual Standard Deviation (RSD) of the Regression of
Products of Retail Cuts of the Technological Division and Products of the Dissection of the Carcass on Meat Weight (MW), Meat Percentage of the Carcass (MW/CCW) and Ratio Meat/Bone (MW/ SW), Using Two Breeds of Rabbit R2 MW RSD
FPW IPW HPW AWW PMiW PMaW LDW MFLW MHLW MFLW/BFLW MHLW/BHLW
0.90 0.93 0.95 0.76 0.61 0.90 0.93 0.74 0.95 0.22 0.40
23.3 19.2 17.2 36.0 45.7 23.3 19.2 37.0 17.2 64.5 56.4
ht.
** ** ns ns ns ** ** ns ns ns *
RZwwjccw
RSD
Znt.
R2Mw,ew
RSD
ht.
0.46 0.44 0.46 0.35 0.44 0.53 0.51 0.46 0.51 0.41 0.60
0.014 0.015 0.014 0.016 0.015 0.013 0.014 0.014 0.014 0.015 0.012
** * ns
0.33 0.30
0.135 0.138 0.138 0.145 0.133 0.124 0.130 0.129 0.131 0.114 0.092
* * * *
* ns ns * ns ns ns ns
0.31 0.23 0.35 0.44 0.38 0.39 0.37 0.052 0.69
ns ns * ns ns ns ns
Int.: Significance of the interaction breedxregression coefficient. ns: non significant. *: P < (0.05). **: P< (0.01). Fore (FPW), intermediate (IPW) and hind (HPW) part weight of the carcass. AWW, abdominal wall weight; PMiW: Psoas minor + Quadratus lumborum weight; PMaW: Psoas major weight; LDW: weight of the M. Longissimus; MFLW: meat weight of one fore leg; MHLW: meat weight of one hind leg; BFLW: bone weight of one fore leg; BHLW: bone weight of one hind leg.
80
P. Hernhndez et al.
that would hold for both breeds, given that when they reach the commercial slaughter weight they are not at the same stage of maturity. Table 5 (Eqs 16) shows that the differences in accuracy when the equations applicable to of the other breed were used were small, and the R2 values were acceptable in all the cases (ranging from 0.74 to 0.89). Moreover, when the same set of variables was used for both lines, the regression coefficients were not significantly different. This is because some of the variables that were candidates to be selected are highly correlated. Thus, to choose one over another has no appreciable influence in the accuracy of the regression. When fitting the variables of Eqs 1 and 2 to both lines, R2 of 0.88 and 0.86 were found (Table 5, Eqs 5 and 6). However, the high value of the R2 obtained in all of these equations is a result of the poor conditions of this set of data, since the relationships of these variables is very high (see, for example, Table 6, Eqs 2, 3, 5). A test of collinearity showed that the first eigenvalue accounted for more than 99% of the variation of all variables. This quirk of the regression analysis was first observed by Hamilton (1987) and Routledge (1990) for two variables, and it was generalised by Cuadras (1993) for multiple regression. As the stepwise procedure does not prevent this phenomena, the ill-condition state of the data can be unperceived, and this seems to have been the case for some of the equations given by Blasco et al. (1984) and Lambertini ef al. (1991). When the variables are selected under the condition that their relationship should not be higher than 0.9, the resulting equation had a R2 lower than 0.70 (Table 5, Eq. 7), and it includes the meat/bone ratio of the hind leg. This difficulty of predicting meat percentage originates from the high correlation between numerator and denominator of the ratio MW/CCW (0.97 in both lines), which produces in this trait a low variation independent of CCW, a variable highly-correlated to TABLE 5 Prediction
Breed
Equations
of the Carcass Meat Percentage (R2) and Residual Standard
(MWjCCW). Coefficient Deviation (RSD)
Equation
of Determination
RJ
RSD Xl@
IV
2R 3v 4R
SV+R 6V+R
7V+R
0.155x 10e2 PMaW+0.262x lo-* MHLW-0.320x 10e2 HLW+5.18~1@~ HPW-0.467~10-~ CCW+O.514~10-~ RCW + 0.525x 1O-3 HLL + 0.452 0.351 x lO-2 MHLW-0.523x lO-3 CCW +0.688x lO-3 RCW-0.375x10-* HLW+0.387~10-~ HPW-tO.517 0.356x lO-2 MHLW-0.474x 1O-3 CCW +0.557x lO-3 RCW-0.384x lO-2 HLW+0.545x lO-3 HPW+0.537 0.613x 10e3 PMaW + 0.330x 10M2MHLW-0.358x 10m2HLW + 0.394~10-~ HPW-0.513x 10-3CCW+0.648x10-3 RCW+0.330xlO-3 HLL + 0.464 0.363x 1O-2 MHLW-0.499x lO-3 CCW+O.619x lo-) RCW-0.390x 1O-2 HLW+0.493x lO-3 HPW+0.527 0.105x 10e2PMaW + 0.311 x 10e2 MHLW - 0.354x 10e2 HLW + 0.506x 1O-3 HPW - 0.488x lO-3 CCW + 0.566x lO-3 RCW +0.444x lO-3 HLL + 0.453 0.021 MHLW/BHLW + 0.214x lO-2 PMaW - 0.407x 10m4LW +0.622x lO-3 LCL
0.801
7.07
0.880
6.41
0.744
7.89
0.886
6.35
0.864
7.20
0.88 1
6.80
0.693
10.7
PMaW: Psoas major weight; MHLW(g): meat weight of one hind leg; HLW(g): weight of one hind leg; BWHL: bone weight of one hind leg; HPW(g): hind part weight of the carcass; LW(g): liveweight; CCW(g): chilled carcass weight 24 h after slaughter; RCW(g): reference carcass weight; HLL(mm): hind leg length; LCL(mm): lumbar circumference.
81
Prediction of rabbit carcass composition TABLE 6
Correlations
Among Variables (Breed V Above Diagonal, Breed R Below Diagonal) MHL W
MHLW HLW HPW ccw RCW
0.990 0.981 0.949 0.971
HLW
HPW
ccw
RCW
0.980
0.957 0.982
0.932 0.957 0.963
0.952 0.973 0.976 0.990
0.991 0.961 0.981
0.961 0.978
0.986
MHLW: Meat weight of one hind leg; HLW: weight of one hind leg; HPW: hind part weight of the carcass; CCW: chilled carcass weight 24 h after slaughter; RCW: reference carcass weight.
the other variables studied. In other species, the presence of a much higher fatness gives other meaning to the carcass meat percentage. When the ratio meat/bone of the carcass was predicted only the ratio meat/bone of the hind leg gave a reasonable prediction, which was improved a little when adding other measurements not highly correlated with it (R* changed from 0.69 to 0.71). By using the meat of both legs, Varewyck & Bouquet (1982) give R*= 0.66. Given this, the meat/bone ratio of the hind leg can be used to character&e carcass composition and make predictions, as previously suggested by Rouvier (1970). This ratio is the only recommended measurement for carcass composition prediction by the WRSA. Thus, rabbit carcass quality is largely reduced to a two component problem, in which CCW and MW/BW explain the main variation of interest. Individual muscles or other products from dissection are not good predictors of meat percentage or the ratio meat/bone, with the possible exception of the ratio meat/bone of the hind leg, which has a moderately high R* for both traits (Table 4). Prediction of carcass dissectible fat and fat percentage of the meat Total dissectible fat of the carcass is predicted well by using the perirenal fat depot (Table 7). The regression of the perirenal fat depot on carcass dissectible fat percentage shows an R*=0.69. Varewyck & Bouquet (1982) give an R*=0.86when the whole fat tissue of the intermediate part is dissected. The fat contained in the meat (intramuscular and intermuscular fat) is a factor of meat quality, but it is laborious and expensive to evaluate. The predictive power of dissectible fat deposits on fat percentage of the meat dissected in a half-carcass are shown in Table 8. Although the R* values are not good, the relationships between total dissectible fat TABLE 7
Coefficient of Determination (R2) and Residual Standard Deviation (RSD) of the Regression of Fat Deposits on Dissectible Carcass Fat Weight (DFaW) and Dissectible Carcass Fat Percentage (DFaW/CCW)
R2CJW SFaW IFaW PFaW
0.64 0.69 0.77
RSD
Int.
5.52 5.10 4.39
ns ns ns
R*~F~W,CCW
0.54 0.71 0.69
RSD
ht.
0.438 0.343 0.360
ns ns ns
Int.: Significance of the interaction breedxregression coefficient; ns: non significant; *: P < (0.05). **: P < (0.01); SFaW: scapular fat weight; PFaW: perirenal fat weight; IFaW: inguinal fat weight.
82
P. Herncindez et al
TABLE 8 Coefficient of Determination (R2) and Residual Standard Deviation (RSD) of the Regression of the Fat Deposits on Fat Percentage of the Meat (MFaP)
R2LfFd SFaW IFaW PFaW DFaW
0.391 0.550 0.398 0.554
Int.: Significance of the interaction breedxregression
RSD
Int
0.791 0.680 0.786 0.677
* ** ** **
coefficient; ns: non significant. *: P< (0.05). **:
P < (0.01).
SFaW: Scapular fat weight; PFaW: perirenal fat weight; IFaW: inguinal fat weight; DFaW: total dissectible carcass fat weight.
deposits weight and fat percentage of the meat is relatively high (corresponding to a correlation coefficient of 0.74), which could have some implications in selection programs in the future, since fat content is a meat quality criterion. As an increment in the fat contained in the meat will also increase fat carcass content, its economic interest would largely depend on whether and how meat quality is paid in the future.
CONCLUSIONS
This work supports the recommendations of the WRSA on carcass quality characterisation and suggests ways of simplifying them. Two breeds of different degrees of maturity have been used in the experiment, and the conclusions apply for both: carcass composition can be clearly defined by using carcass weight, the ratio meat/bone of the hind leg and the perirenal fat deposit weight. Other multiple regression prediction equations do not show significant improvements.
ACKNOWLEDGEMENTS We are grateful to Luis Valero for his help in the slaughter process and to J. Sahuquillo for the care of the animals. This research is included in the project CAICYT AGF93-0634, financed by the Spanish Ministry of Education and Science.
REFERENCES Blasco, A., Estany, J. & Baselga, M. (1984). Ann. Zootech., 33, 161. Blasco, A., Ouhayoun, J. & Masoero, G. (1993). World Rabbit Sci., 1, 3. Blasco, A. & Gomez, E. (1993). Anim. Prod., 57, 332. Bochno, R., Lewczuk, A. & Janiszewska, M. (1979). Rcz. Nauk Zoot., T6(1), 175. Cuadras, C. M. (1993). Am. Stat., 47, 256. Hamilton, D. (1987). Am. Stat., 41, 129. Janiszewska, M. & Bochno, R. (1979). Zesz. Nauk. ART. Zoot., 19, 119. Lambertini, L., Benassi, M. C. & Zaghini, G. (1991). Rivista di coniglicoltura, 8, 35. Lopez, M. C., Sierra, I. & Lite, M. J. (1992). Options Mediterran., 17, 75. Lukefahr, S. D., Hohenboken, W. D., Cheeke, P. R. & Patton, N. M. (1983). J. Anim. Sci., 57, 899. Lukefahr, S. D. & Ozimba, C. E. (1991). Lives. Prod. Sci., 29, 323.
Prediction of rabbit carcass composition Niedzwiadeck, S. (1984). In: Proc. 3rd World Rabbit Congress, Roma, 2, 585. Pia, M., Hemhdez, P. & Blasco A. 1996. Meat Sci., 44, 85. Rouvier, R. (1970). Ann. GhPt. SIX Anim., 2, 325. Routledge, R. D. (1990). ht. J. Math. Educ. Sci. Technol., 21,403. SAS (1990). SAS/STAT User’s Guide, Version 6, SAS Institute Inc., Cary, NC, USA. Varewyck, H. & Bouquet, Y. (1982). Ann. Zootech., 31, 257.
83