Least squares means and standard errors of the differences for each trait in all lines at all ages are in Additional file 2.
Figure 1 shows changes in body mass for each line over 7 weeks. There was a significant difference in body mass between lines in both experiments (P < 0.001). The broiler chicken and all three Pekin duck lines grew at a faster rate than both the layer chicken and the mallard (P < 0.001).
Differences in leg bone length mirrored those of body mass; the tibiotarsus (Fig. 2a) was significantly shorter in the layer chicken and mallard compared to the broiler chicken and Pekin lines (P < 0.001). There was an age effect (P < 0.001) i.e. duck lines in each experiment showed a decline in tibiotarsal growth from 5 weeks of age whereas chicken tibiotarsi continued to grow throughout the experiment. When analysed allometrically, chicken tibiotarsi grew with positive allometry and the tibiotarsi of all four duck lines grew with negative allometry (Table 1). A sex effect (P < 0.001 in Experiment 1 and P = 0.002 in Experiment 2) was also observed; males of all lines had longer tibiotarsi than females. In Experiment 1 an age by sex interaction was observed (P < 0.001) i.e. females (broiler chicken and Pekin hybrid, but not layer chicken) had longer tibiotarsi than males at 3 weeks but not at 5 and 7 weeks of age (P < 0.001).
In the cranio-caudal plane, the tibiotarsus of the broiler chicken was significantly more curved (cranially) than the tibiotarsus of the layer chicken, which was in turn more cranially curved than that of the Pekin hybrid (P < 0.001). In Experiment 2, the mallard tibiotarsus displayed significantly more caudal curvature than that of the male Pekin line (P = 0.013) but did not differ from that of the female line (Fig. 3b). Male birds in Experiment 1 exhibited greater cranio-caudal curvature of their tibiotarsi than females. Both species differed in the direction of tibiotarsal curvature in this plane, i.e. all four duck lines curved caudally whereas both chicken lines curved cranially. In the medio-lateral plane (Fig. 3d), the tibiotarsi of both the broiler chicken and Pekin duck displayed greater lateral curvature than their lighter conspecifics; however, this difference was statistically significant (P < 0.001) only between the mallard and Pekin breeding lines. In this plane, the duck tibiotarsi were more laterally curved than those of both chicken lines (P < 0.001).
Tibiotarsal torsion occurred to a similar extent in both chicken lines. The Pekin hybrid differed significantly from the chicken lines (P < 0.001). There was a line by age interaction, i.e. at 3 weeks of age, the Pekin and chicken lines displayed a similar range of tibiotarsal torsion but by 7 weeks of age the distal part of the tibiotarsus of the Pekin hybrid had rotated internally in relation to the proximal tibiotarsus (P = 0.005) (Fig. 4b). Internal rotation occurs when the cranial aspect of the distal part of the tibiotarsus turns to face medially. No difference in tibiotarsal torsion was observed between the mallard and Pekin breeding lines in Experiment 2; however, the distal tibiotarsi of the male Pekin line rotated internally to a greater extent than that of the female Pekin line (P = 0.024). There was an age interaction (P < 0.001) with tibiotarsi in all three duck lines of Experiment 2 rotating internally as they aged. A line by age interaction also occurred in Experiment 1, i.e. the distal tibiotarsi of the Pekin commercial hybrid rotated internally as the bird aged (P = 0.005) whereas the tibiotarsi of the chicken lines did not. The R2 values from regressions of the log of bone torsion on the log of body mass were very low for both chicken lines (Table 1), suggesting no relationship. In the duck lines, tibiotarsal torsion deviated slightly from isometric growth.
Tibiotarsal bone quality
Tibiotarsal stiffness differed significantly within both chickens and ducks (P < 0.001). Lines selected for rapid growth had stiffer tibiotarsi than lines with a slow growth (Fig. 5a). There was an age effect in Experiment 2, i.e. the tibiotarsi of the fast-growing Pekin lines had the same stiffness as the mallard at 3 weeks of age but, by 7 weeks, they were significantly stiffer than those of the mallard (P < 0.001). Tibiotarsal stiffness scaled isometrically in all lines except for the layer chicken, which displayed very positive allometry (Table 1). Tibiotarsal strength (maximum load to rupture) for all lines at all ages was significantly greater in fast-growing lines compared to their slow-growing conspecifics (P < 0.001). Tibiotarsal strength scaled with positive allometry for all lines except for the male and female Pekin breeding lines, which scaled with isometry (Table 1).
Bending stress (Fig. 5b) did not differ significantly between the Pekin hybrid and both chicken lines at all ages. However, there was a line by age interaction (P < 0.001), i.e. Pekin hybrid tibiotarsi tolerated greater bending stresses as the animals grew older whereas the bending stresses tolerated by the tibiotarsi of both chicken lines decreased. The mallard tibiotarsi resisted significantly more bending stress than those of the heavier male Pekin line and female Pekin line (P < 0.001). There was also an age effect, i.e. the tibiotarsi of all three duck lines tolerated more bending stress as they aged (P < 0.001). Data for the male Pekin line at 5 weeks of age was not analysed since these bones moved during loading, causing error.
Data on tibiotarsal ash content for the layer chicken line at 5 weeks of age is not available due to measurement error. The ash content of the bone before drying (Fig. 5c) was significantly greater in the broiler compared with both the layer chicken and Pekin hybrid (P < 0.001). All lines had increased bone mineralisation as they aged (P < 0.001), although the broiler chicken reached its 7-week level of ash content earlier than the Pekin commercial hybrid (P < 0.001). In Experiment 2, all duck lines had increased bone mineralisation as they grew older (P < 0.001); however, for the mallard, bone mineralisation increased until 5 weeks of age and then decreased from 7 weeks of age (P < 0.001). The molar Ca:P ratio across all lines over all ages ranged from 1.4 to 1.8 (see Additional file 2).
Porosity differed significantly between lines in Experiment 1 (P = 0.003), i.e. the tibiotarsi of broiler chicken were more porous at the mid-diaphysis than those of both the layer chicken and Pekin hybrid (Fig. 5d). In Experiment 2, all three duck lines differed significantly in tibiotarsal porosity (P < 0.001) with the male line having the highest mid-diaphyseal porosity and the mallard having the lowest. An age interaction was also observed in the duck lines with the tibiotarsi becoming less porous as the birds aged (P < 0.001).
The length of the femur was significantly shorter in the layer chicken and the mallard than in the broiler chicken and Pekin male and female lines (P < 0.001). Age effects, sex effects and line-by-age interaction effects were observed in both experiments (P < 0.001). Femoral length scaled with negative allometry for all duck lines (Table 1). The femoral length of the broiler chicken increased isometrically with body mass and that of the layer chicken showed slightly positive allometric growth.
There was no difference in cranio-caudal curvature of the femur between the broiler and layer chicken in Experiment 1 (Fig. 3a). However, the femora of the chicken lines were more cranially curved than the Pekin hybrid in this plane (P < 0.001). In Experiment 2, the femora of the mallard and the female Pekin line displayed more cranial curvature than those of the male Pekin line (P = 0.001). In the medio-lateral plane, no significant differences in femoral curvature of the broiler chicken, layer chicken or Pekin duck hybrid were observed. However, there was an age effect (P < 0.001), i.e. lateral curvature of the femora of all three lines decreased as the birds aged. In Experiment 2, the femora of the male Pekin line were significantly more laterally curved than those of the female line and the mallard (P < 0.001). An age effect and a line-by-age interaction effect were observed in these lines (P < 0.001); the femora of the male line became less curved in this plane as the birds aged whereas the female line and mallard maintained the same curvature. By 7 weeks of age, the femora of all lines were curved to a similar degree in the medio-lateral plane (Fig. 3c).
Femoral torsion differed significantly between chicken lines in Experiment 1 (Fig. 4a); the distal femur of the broiler was rotated more externally to the proximal end when compared to that of the layer at all ages (P < 0.001). There was also a line-by-age interaction (P < 0.001); at 3 weeks of age, the distal femur of the Pekin commercial hybrid was rotated internally compared to that of both chicken lines. However, as the Pekin individuals aged, the distal femur rotated externally, reaching a similar degree of femoral torsion as for the broiler chicken by 7 weeks of age. A similar age interaction occurred in Experiment 2, with the distal femur of all duck lines rotating externally in relation to the proximal femur (P < 0.001). The mallard femur underwent less rotation as it aged (P = 0.002), reaching a degree or femoral torsion similar to that of the layer chicken by 7 weeks of age.
Femoral cortical area grew isometrically in the broiler chicken and with positive allometry in the layer chicken. The R2 value for this trait in duck lines was low, suggesting a weak relationship with body mass.