**2. Materials and Methods**

The pilot study was designed to evaluate two different methods of manual broiler catching. LEGS: The broilers were caught by two legs and carried upside-down (Figure 1). Usually, this method allowed two or three broilers in one hand and one in the other hand (range: Two to four birds in one hand, one to two in the other).

**Figure 1.** Catching and carrying broilers by a grip in both legs.

UPRIGHT: The broilers were caught under the abdomen (Figure 2) and carried in an upright position to the transport modules. Mostly two broilers were caught and carried in each catch by the

upright method (range: One to two birds per catch; a catch refers to each time the catchers pick up birds and place them in the drawer).

**Figure 2.** Catching under the abdomen and carrying the broilers upright.

The study was carried out under commercial conditions in one major broiler producing region in Norway (Trøndelag). Two flocks from two different farms were enrolled. The two flocks were of different hybrids; one flock was a conventional fast-growing type (Ross 308), flock size 30,000 broilers, slaughtered at the age of 33 days and 1328 g (mean carcass weight for the flock, estimated live weight 2000 g) and the other flock was a slower growing hybrid (Hubbard JA 787), flock size 16,800 broilers, slaughtered at the age of 44 days and 1539 g (mean carcass weight for the flock, estimated live weight 2300 g). Both flocks were mixed sex and fed ad-libitum. Both broiler houses were from the same manufacturer, had the same size and were equipped with the same ventilation system.

The study sample consisted of 3951 broilers from both houses in total; 2010 caught by LEGS (1031 broilers from the conventional hybrid and 996 broilers from the slower growing hybrid) and 1941 by UPRIGHT (969 broilers from the conventional hybrid and 955 broilers from the slower growing hybrid). This implies that the majority of the broilers in the two flocks were not included in the study. Both catching methods were evaluated in both flocks.

The sampling was carried out in dimmed lighting at the start of the catching process. Only the first eight transport modules in each house were included in the study; the first four modules with LEGS, then the next four modules with UPRIGHT. The birds were caught randomly for each method. The two broiler flocks were caught two subsequent nights by the same catching team, consisting of the same four professional catchers. The team consisted of two women and two men; all had catching as their primary occupation. They had no standard catching method, sometimes by one leg, sometimes by two legs and sometimes upright by abdomen. For this study, they were trained to catch both by LEGS and by UPRIGHT methods. The training consisted of theoretical instructions along with practical demonstrations. For LEGS, the only instruction was to catch by both legs for each bird. For UPRIGHT, the only instruction was to catch under the abdomen and carry the broilers in an upright position. The instruction did not regulate the maximum number of birds allowed per catch. All catchers performed both methods in this study.

Both flocks were caught late night/early morning. The size of the transport modules was identical for both hybrids and methods (Stork®, 2.43 × 1.30 m [length × width], eight open-topped drawers in a module, a planned minimum of 200 cm<sup>2</sup> per kg live bird). The transport modules were marked according to catching the method, and the eight modules were loaded on the same vehicle. The time to fill each transport module with birds was recorded. The modules were loaded in and out of the broiler house with a forklift. The truck driver placed all modules as close as possible to the broilers as per normal industry practice. This process was consistent for both methods and both flocks.

The journey time from farm to abattoir was 30 min for the conventional flock and 10 min for the slower growing flock. Immediately after the transport modules arrived at the abattoir, the broilers were manually lifted out of the modules and investigated in lairage for wing and leg fractures, birds on their back in the drawers and DOA. The birds were assessed by two of the researchers at the same time. Injured birds were immediately stunned with blunt trauma to the head, followed by cervical dislocation. The wing fracture criteria were: open or closed fractures, dislocated wings and detachment of the epiphyseal plates with visible bleeding around the elbow joints [9]. In addition, the number of animals per drawer was recorded. These investigations were performed in lairage and not on the farm, due to better lighting and protection against low temperatures at the abattoir. The assessors of the birds in lairage were aware of the catching method.

Data were entered into an Excel spreadsheet and transferred to STATA 14.2 (College Station, TX, USA). All observations were inspected for deviations and missing data before the analyses. The distribution of the continuous variables (crating time and stocking density) were inspected visually by histograms and by summary statistics. Both variables were approximately normally distributed. Simple linear regression was used to compare means of catching and crating time, the number of birds per cage, DOA and fractures by hybrid and catching method. An interaction term was included in the regression analyses to evaluate whether the univariable relationships between stocking density and hybrid, as well as catching time and hybrid were dependent on the simultaneous effects of hybrid and catching method. Residuals were predicted and visualized by histograms. Since wing fractures were either present or absent in each module, this variable was treated as categorical. Thus, for wing fractures, simple logistic regression was used to study the effect of birds per cage, hybrid, crating time, and catching method. Hybrid was included as a covariate in the study of effects of birds per cage and crating time on wing fractures. Residuals (linear regression) were predicted and displayed on normal quantile plots.
