Time to collapse was assessed by two independent observers on the first 112 animals to determine whether there might be any observer bias. It was found that the observations for time to collapse were significantly correlated between the observers, with r = 0.90 (95% CI 0.87–0.94) P < 0.001. Thereafter one operator performed all subsequent recordings of time to collapse. Furthermore, to test for sampling drift the results were analysed for sample day effects, and no significant associations were found (P < 0.564). There were no significant interactions between animal breed, age or gender on the time to collapse or other recorded parameters.
Six hundred and forty four cattle were slaughtered without stunning. The mean time to final collapse for all animals was 18.2 ± standard error (SE) 0.9 s (s) (range 1–257 s). Eighty one percent of all animals collapsed in the initial 20 s after the cut. In total 26 (4%) animals took 60 s or more to achieve final collapse. There was a significant correlation between TR and cervical vertebral position, with r = 0.76 (95% CI 0.73–0.79) P < 0.001. To simplify analysis of neck cut position, the cuts were grouped as either low neck cut (LNC) (TR ≥ 2.5, ≈ C2–C3 +) or high neck cut (HNC) (TR ≤ 2, ≈ C1). In total, data was collected from 561 and 83 cattle that received a LNC and HNC respectively.
Fig. 1 is the frequency distribution for time to final collapse for animals in the LNC and HNC groups. Final collapse in the majority of animals occurred within 20 s (LNC 79%; HNC 89%). There was a significant difference in the mean time to final collapse (Table 1), with animals receiving the HNC (13.5 ± 1.3; range 4–86 s) collapsing sooner than those in the LNC group (18.9 ± 1.1; range 1–257 s) (P < 0.01). The mean number of cuts (P < 0.001) was significantly higher in animals slaughtered with the HNC (2.0 ± 0.1) compared to the LNC (1.7 ± 0.0).