Description of the data
Origin of the animals
In total, 27 sires were used in different frequency over the herd. Nine sires had less than ten offspring and ten sires had between eleven and twenty offspring. Only three sires gave between 21 and 30 calves and five sires gave more than 30 calves (Table 1). Only sixteen calves were born from unknown sire.
In the experiment, 267 females and 249 males for a total of 516 animals were used. The distribution of the animals was rather equal between the four different suckling meadow managements and between the sexes. Indeed, between the suckling meadow managements, the repartition of the animals varied from 113 to 141 and from 53 to 72 for the sexes within each suckling meadow management (Table 2).
The number of animals born over the different calving periods varied from twenty to 76. The distribution of the animals over the different suckling meadow managements within the years was homogeneous except for 2005. That year, no animal was raised in non-irrigated implanted pasture whereas 30 and 27 animals were raised in irrigated implanted pasture and savannah and only seventeen in non-irrigated savannah (Table 3). Before 2000, two calving seasons were applied, and the number of calves from each system was less than 10 calves.
In total, 284 animals were fattened with an intensive feeding regime; 135 females and 149 males. Fattening at pasture concerned 232 animals; 132 females and 100 males. Over the calving periods, the number of animals in each fattening group varied from eight to 42 for the intensive feeding regime and from seven to 36 at pasture (Table 4).
The distribution of the calves according to the calving number of their dam is given in Table 6. The Creole cows showed an exceptional longevity with 126 calves having their dams from parity order seven and higher.
Means and frequencies of the variables
Table 7 represents basic information of the variables of interest in the study such as the number of animals, the observed mean, the standard deviation, the minimum and the maximum. The means of the weaning weight and suckling ADG were calculated on 515 animals only because the weaning weight has not been recorded for one animal in the data set. The means of the weights at fifteen and eighteen months were calculated with the animals not slaughtered yet at that moment.
AGD: average daily gain. Weights are in Kg, age and duration are in days and ADG are in g/day.
Growth traits until weaning were similar between the two fattening environments. The birth weight was equal in groups (27 Kg), the weaning weight and the suckling average daily gain reached 155 Kg and 606 g/day in intensive feeding regime against 159 Kg and 626 g/day at pasture (Table 8). After weaning, all weights reached lower levels at pasture. The growth rate at pasture was about 40% smaller than the growth rate in intensive feeding regime (678 g/day versus 413 g/day).
Analysis of variance: generalized linear models
The average daily gain during fattening, the weights at nine, twelve, fifteen and eighteen months were studied in separate generalized linear model analyses. All models fitted the data with an R2 varying between 86.9% and 93.4%. The Herd/Year/Sex group during fattening and the weaning weight were highly significant for all five variables. The interaction between the fattening environment and the sire was a significant covariable for the weight at eighteen months and showed a tendency for the weight at twelve months (Table 9).
Multivariate analysis: comparison within fattening environments
The analysis included the weights at nine, twelve and fifteen months as dependent variables when the animals were either fattened with an intensive feeding regime or raised at pasture.
The heritability in intensive feeding regime was higher than at pasture for all three weights. In both fattening environments, the trait with the smallest heritability was the weight at nine months (0.23 +/-0.15 in intensive feeding regime and 0.09 +/- 0.11 at pasture) (Table 10). Table 11a shows high genetic correlations between the different weights in intensive feeding regime (between 0.85 +/- 0.09 and 0.98 +/- 0.08). The same observation was given at pasture (genetic correlation between 0.97 +/-0.24 and 1.14 +/- 0.27) although the precision was lower (Table 11b). In both fattening environments, the genetic variance increased with the age of the animals.
Bivariate analysis: comparison between fattening environments
The bivariate analysis included as dependent variables simultaneously two weights measured in the two different fattening environments, in different combinations of the weight at fixed ages of nine, twelve and fifteen months in both environments, and of eighteen month at pasture.
The heritability of the different weights measured in intensive feeding regime was higher than at pasture (Table 12). The heritability of the weights at nine and twelve months at pasture was almost null. It increased to 0.18 +/- 0.15 and 0.28 +/- 0.18 for the weights at fifteen and eighteen months. The genetic variance was notably higher in intensive feeding regime than at pasture and increased with the age of the animals. Indeed the genetic variance for the weight at nine months was equal to 80 in the intensive feeding regime and four at pasture. The genetic variance for the weight at fifteen months was equal to 401 in the intensive feeding regime and 82 at pasture. All genetic correlations showed very poor precision (Table 13).
Random regression: comparison between fattening environments along the fattening period
The heritability for weights during the whole fattening period was higher in intensive feeding regime than at pasture. At pasture, heritability was rather constant varying from 0.18 to 0.23 whereas in intensive feeding regime, the heritability first showed a small increase from 0.40 at 210 days of age to 0.45 at 270 days of age and then decreased to 0.23 at 540 days. The genetic correlation presented an increase from 0.70 to 0.97 between 210 and 300 days of age followed by a decrease until 0.59 at 540 days (Figure 2 and Annex 1).
Genetic correlation for weights within each fattening environment presented high values between closed ages. The genetic correlations decreased comparing weights at more distant ages. This tendency was more visible at pasture where the values dropped from 1 to 0.46 instead of 1 to 0.73 in intensive feeding regime (Figure 3, Annex 2 and 3).
Table of contents :
2 Materials and methods
2.1 Herd management and data collection
2.1.1 Origin of the calves
2.1.2 Suckling period
2.1.3 Fattening period
2.2 Statistical analysis
2.2.1 Generalized linear models
2.2.2 Multivariate analysis: comparison within fattening environments
2.2.3 Bivariate analysis: comparison between fattening environments
2.2.4 Random regression: comparison between fattening environments
3.1 Description of the data
3.1.1 Origin of the animals
3.1.2 Means and frequencies of the variables
3.1.3 Analysis of variance: generalized linear model
3.2 Multivariate analysis
3.2.1 Multivariate analysis: comparison within fattening environments
3.2.2 Bivariate analysis: comparison between fattening environments
3.3 Random regression: comparison between fattening environments
3.4 Breeding value estimations