The effect of crate height on the behavior of female turkeys during commercial pre-slaughter transportation
Abstract
Limited information is available on suitable height of transport crates for turkeys. We compared behaviors and physiological indicators of four groups of 10 female turkeys each confined in either conventional (38.5 cm height) or experimental (77 cm height) crates during six commercial pre-slaughter transportations for 86 km (76 ± 4 min) along two tracts with one-lane streets, crossroads, bends, roundabouts (S1 and S2) and a highway tract (H) between S1 and S2. Only 36% of birds in the higher crates maintained a standing position. In conventional versus experimental crates, the frequency of rising attempts was five/bird/hour versus less than one/bird/hour, while wing flapping was seven/bird/hour versus 20/bird/hour, and balance loss was one versus four/bird/hour. The behaviors of both groups differed significantly according to the route tract, with a lower frequency of stress-related behaviors at H. No scratches, fractures or hematomas were detected in any birds after transportation. Crate height had no significant effect on hemato-biochemical markers. These results suggest that crates enabling a standing position may increase potentially dangerous behaviors. Moreover, busy and curvy routes should be avoided, as they may contribute to increasing the frequency of stress-related behaviors.
Introduction
Live transportation is a major stressor for farm animals and may have deleterious effects on animal health, well-being, performance and product quality (von Borell 2001; Nielsen et al. 2011). For this reason, animal welfare during transportation is one of the biggest arguments against modern livestock production and there is a paucity of empirical data on the transport needs for different animal species (EFSA 2004, 2011). According to the OIE's (world Organization for Animal Health) Terrestrial Animal Health Code, transported animals should be able to assume a natural standing position without coming into contact with the upper deck of the vehicle; these conditions do not normally apply to poultry, but under certain conditions poultry may benefit from having adequate head room (OIE 2016). This is also remarked in the European Union (EU) legislation currently in force (Council Regulation EC 1/2005 of 22 December 2004), requiring that ‘sufficient floor area and height is provided for the animals, appropriate to their size and the intended journey’. However, requirements for poultry do not address any species-specific need, and the minimum floor area is to be provided based on live weight only, while minimum crate height is not provided as such. The aforementioned Regulation also states that space requirements may vary depending on the weight and size of the birds but also on their physical conditions, weather and transport duration.
Two studies (Wichman et al. 2010, 2012) have investigated the effect of crate height on the welfare of male turkeys. Wichman et al. (2010) compared stationary crates of 40, 55 and 90 cm in height placed in experimental facilities and found that the degree of vertical confinement had little effect on physiological parameters, but had a large effect on birds’ behaviors, indicating a better welfare associated with higher crates, as these permitted a standing position. However, Wichman et al. (2012) found that crates of 55 cm height compared to those of 40 cm were associated with significantly more birds with dorsal scratches after transport. Yet, both these studies did not assess animal behavior during transportation, which is likely to be different from lairage as a result of thermal stress, vibrations, vehicle movements and manoeuvers, exposure to flashing lights and other noises (Elrom 2000). Moreover, no study has so far investigated welfare issues in transported female turkeys, which are younger and lighter than males at slaughtering time, thereby being relatively more reactive and prone to injuries. Nonetheless, like their male counterparts, female turkeys are not allowed to maintain a standing position in the conventional crates. The aim of this study was to compare the behavior and physiological response of female turkeys in crates of different heights during commercial preslaughter transportation.
Materials and Methods
Animals and housing
Forty female turkeys (commercial hybrid; 11 ± 0.5 kg live weight) were randomly selected at the time of slaughtering, individually identified by means of a numbered leg ring and assigned to four groups of 10 birds (A, B, C, D). Six commercial transportations were carried out by the same driver in northern Italy during a 3 day period (two travels per day; one at late morning and one at early afternoon) for 86 km (76 ± 4 min) along the same route. The route was characterized by two tracts with one-lane streets, crossroads, bends, roundabouts (S1 and S2; 20 km in 25 min each) and a highway tract (H; 36 km in 25 min) between S1 and S2. No breaks were made between S1, H and S2.
Before each travel, two out of the four groups were manually loaded in either a conventional iron crate (CC) (114 cm × 105 cm; height 38.5 cm) or an experimental iron crate (EC) with the same floor allowance but a height of 77 cm. Each group travelled three times (once a day) by alternating the position in CC and EC. When returning from the travel, animals were manually unloaded and housed under normal farming conditions: ad libitum feed and water in a confined area of around 25 m2 per each group. Wood shavings were used as bedding material.
Conventional crates consisted of a six-crate module; thus, the experimental crate consisted of two conventional crates merged in one single crate by removing the floor in between (see Fig. 1). The lateral sides of the crates had two 10 cm × 21 cm areas with 14 open slots (1 cm), while the front and back sides were made by a 2.5 × 2.5 grid. The floor/roof of the crates consisted of a smooth unperforated surface. The module was fixed in the front part of the truck, where light intensity was approximately 15 lux. Ambient temperature ranged between 9 and 12°C (sunny weather; relative humidity 65%–75%); there was no ventilation system in the vehicle. Windbreak cloths were turned down along the vehicle's sides for assuring animal protection from draughts and direct sunlight. All transportation procedures and the number of animals per crate were established in accordance with Council Regulation (EC) 1/2005 on the protection of animals during transport and related operations (Council of the European Union 2005).
Experimental measurements
A blood sample was taken from the wing vein before (T0) and immediately after (T1) the first transportation using a 22G needle and collected into lithium-heparin coated 3 mL vacutainers. While a subsample of 24 animals was sampled at T0 (12 assigned to CC and 12 to EC), all 40 subjects were sampled at T1 (20 after unloading from CC and 20 from EC).
Corticosterone detection was carried out by means of a commercial enzyme immunoassay kit (Corticosterone Enzyme Immunoassay kit, Arbor Assay, MI, USA). Packed cell volume (PCV) was measured by the microhematocrit method (Campbell 1995) in capillary tubes. Heterophils/lymphocytes ratios (H:L) was calculated by counting at least 200 leukocytes in Wright's-Giemsa stained blood films, by microscopic examination. The level of total protein, aspartate aminotransferase (AST), alanine aminotransferase (ALT), creatine kinase (CK), lactate dehydrogenase (LDH), triglycerides (TG) and uric acid were determined with the automated Cobas C501 clinical chemistry analyzer (Roche Diagnostics, Basel, Switzerland) with dedicated Roche reagents.
Birds in experimental and conventional crates were filmed for behavioral observation by means of two cameras (HERO 3; GoPro Inc., San Mateo, CA, USA) for the whole period between loading and unloading. In accordance with the ethogram proposed by Wichman et al. (2010), the different body postures were recorded as ‘states’ and their total duration was determined over the route tracts by scan sampling at 1 min intervals; other behaviors were recorded as ‘events’ by continuous behavior sampling (see Table 1). Veterinary personnel inspected the body of the animals after each transportation in order to detect any injury (scratches, fractures and hematomas).
| Behavior | Description |
|---|---|
| States | |
| Lying | Bird lying with the body in contact with the floor |
| Sitting | Sternum above the floor with at least one hock on the floor |
| Mounting | Bird lying with the body on the back of another bird |
| Standing | Sternum above the floor in a horizontal position with the front part of the body being in a higher position than the rear part. Legs are straight or just a little bit bent |
| Low standing | Sternum above the floor, legs above the floor but not fully straightened and the frontal part of the body (breast) is positioned lower than the rear part of the body |
| Events | |
| Preening | Beak moves in contact with the plumage |
| Stretching | One leg and often also one wing is lifted out or/and in a rearwards direction from the bird while it is standing |
| Panting | Breathing rapidly in short gasps |
| Stepping | One leg is lifted |
| Turning | Direction of body changes with at least 45° (includes stepping although this is not recorded as such when turning is recorded) |
| Rising attempts | Bird moves from the lying position to either a standing, sitting or low standing position |
| Balance loss | A sudden vibration forces an animal distancing his legs for few seconds to increase the base of support |
| Wing flapping | One or both wings are moved |
All the procedures in this study were approved by the IZSVe Ethics Committee (n. CE. IZSVe 18/2013) and carried out in accordance with Council Directive 86/609/EEC on the protection of animals used for experimental and other scientific purposes (Council of the European Union 1986).
Data analysis
As regards the behavior, each crate under observation during the three route tracts was considered as the statistical unit. At each minute, the total number of animals per crate with a given behaviors (states and events) was recorded. The percent mean and standard error of animal states were calculated for each route tract, and the mean number of events per minute per crate at each route tract was expressed as events/bird/hour. Given that ‘lying’ was not always clearly distinguishable from ‘sitting’ due to the high stocking density, these two states were combined as ‘lying/sitting’.
The states were assessed using a Generalized Linear Mixed Model in which the transportation was included as random effect and the repeated observations made on the same crates at every minute during the entire length of the transportation were modelled using a first-order autoregressive covariance structure. Fixed effects included in the model were the crate type, the route tract and their interaction. The Poisson distribution was chosen as the response distribution.
Regarding the events, in order to allow homogeneity in the statistical analyses for the different behaviors, the dependent variable was converted into a binary variable expressing the presence versus absence of the behavior at each minute. As for states, a Generalized Linear Model was then applied with a first-order auto-regressive covariance structure to model the correlations among the repeated measures made on the same crates of each transportation; to allow for model convergence while accounting for the study design, the transportation was also included as a fixed effect. For flapping, balance loss and turning, the other fixed effects in the model were crate type and route tract, while for rising attempts and stepping, due to the limited number of cases the model was run only for CC and EC, respectively.
Regarding the physiological indicators, the animal was considered as a statistical unit. In view of the non-normal data distribution, non-parametric tests were adopted. Specifically, the Wilcoxon matched-pairs signed-ranks test was used to compare T0 and T1, while the Mann-Whitney two-sample statistic was applied to compare the two crate types. All analyses were performed using the software SAS® v. 9.3 (SAS Institute Inc., Cary, NC, USA).
Results
The results of the physiological measures are given in Table 2. All analyses showed a significant variation between T0 and T1, while no difference was evidenced at T1 between CC and EC. However, total protein was higher in the CC group (P < 0.05).
| Predictions (Huff et al. 2008) | T0 | T1 | 77.0 cm | 38.5 cm | |
|---|---|---|---|---|---|
| n = 24 | n = 24 | n = 20 | n = 20 | ||
| Uric acid (μmol/L) | ↑Oxidative stress | 314.0 ± 22.6 | 298.1 ± 20.79 | 321.5 ± 25.5 | 291.5 ± 23.3 |
| ALT (U/L) | ↑Nonspecific damage | 12.9 ± 1.3a | 15.0 ± 1.2b | 13.1 ± 1.4 | 13.4 ± 0.9 |
| AST (U/L) | ↑Muscle damage | 1124.2 ± 93.2 | 1148.2 ± 91.2 | 1040.2 ± 98.7 | 1054.1 ± 60.7 |
| LDH (U/L) | ↑Muscular fatigue | 746.7 ± 95.4a | 784.13 ± 88.7b | 697.5 ± 98.4 | 637.2 ± 64.2 |
| CK (U/L) | ↑Stress | 20 943.1 ± 2009.8a | 23 089.8 ± 2018.6b | 22 949.4 ± 4429.6 | 21 695.4 ± 2212.3 |
| Protein (g/L) | ↑Dehydration | 49.4 ± 1.0 | 50.7 ± 1.3 | 47.7 ± 1.1c | 51.1 ± 1.1d |
| TG (mmol/L) | ↓Feed withdrawal | 17.4 ± 9.5 | 14.91 ± 8.0 | 13.4 ± 1.7 | 15.6 ± 1.7 |
| PCV (%) | ↑Dehydration | 0.36 ± 0.0a | 0.39 ± 0.0b | 0.39 ± 0.0 | 0.38 ± 0.0 |
| Lactate | ↑Hypoxia | 6.0 ± 0.3 | 6.8 ± 0.4 | 6.7 ± 0.4 | 7.6 ± 0.4 |
| Corticosterone | ↑Stress | 11.7 ± 0.7a | 14.8 ± 1.4b | 15.0 ± 1.8 | 16.2 ± 1.6 |
| H/L | ↑Stress | 1.4 ± 0.1a | 2.0 ± 0.2b | 1.9 ± 0.3 | 2.1 ± 0.2 |
- Different exponent letters indicate significant differences for P < 0.05 between time periods (a, b) and crates (c, d). Predictions of physiological responses to transportation were also indicated, on the basis of a previous study which compared T0 versus T1 (Huff et al. 2008). ALT, alanine aminotransferase; AST, aspartate aminotransferase; CK, creatine kinase; H/L, heterophils/lymphocytes; LDH, lactate dehydrogenase; PCV, packed cell volume; TG, triglycerides.
The results of the behavioral observation are given in Table 3. Overall, the frequency of birds lying/sitting was 58% in EC and 95% in CC. Only a minor percentage of birds in higher crates (around 36%) maintained a standing position. The frequency of low standing was 0% in EC and 5% in CC. Mounting behavior was not possible in the conventional crate and was observed on average in 6% of birds in the experimental crate.
| States | 77.0 cm | 38.5 cm | P-value | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| H | S1 | S2 | Total | H | S1 | S2 | Total | Crate type | Tract | |
| Lying/sitting | 68.37 ± 2.42 | 51.96 ± 2.35 | 53.09 ± 2.53 | 57.62 ± 1.45 | 95.32 ± 0.78 | 93.58 ± 0.80 | 94.68 ± 0.79 | 94.51 ± 0.46 | <0.001 | 0.06 |
| Mounting | 3.43 ± 0.54 | 7.82 ± 0.79 | 7.20 ± 0.96 | 6.22 ± 0.46 | 0.00 ± 0.00 | a | a | |||
| Standing | 28.06 ± 2.29 | 39.75 ± 2.23 | 39.35 ± 2.52 | 35.84 ± 1.37 | 0.00 ± 0.00 | a | a | |||
| Low Standing | 0.06 ± 0.06 | 0.06 ± 0.06 | 0.00 ± 0.00 | 0.04 ± 0.03 | 4.47 ± 0.77 | 6.48 ± 0.79 | 5.31 ± 0.79 | 5.44 ± 0.45 | a | a |
| Events | ||||||||||
| Preening | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | ||
| Stretching | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | ||
| Panting | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | ||
| Step/bird/h | 0.82 ± 0.29 | 4.66 ± 0.68 | 2.80 ± 0.59 | 2.88 ± 0.33 | 0.39 ± 0.17 | 0.87 ± 0.27 | 0.16 ± 0.12 | 0.50 ± 0.12 | na | 0.018 |
| Turn/bird/h | 1.80 ± 0.43 | 5.46 ± 0.64 | 3.01 ± 0.48 | 3.57 ± 0.32 | 1.29 ± 0.34 | 2.94 ± 0.53 | 2.26 ± 0.45 | 2.17 ± 0.26 | 0.094 | 0.017 |
| Rising/bird/h | 0.00 ± 0.00 | 0.08 ± 0.06 | 0.31 ± 0.22 | 0.12 ± 0.07 | 2.48 ± 0.59 | 7.05 ± 0.94 | 6.32 ± 0.89 | 5.24 ± 0.48 | na | 0.019 |
| Balance loss/bird/h | 1.89 ± 0.50 | 6.48 ± 0.86 | 4.66 ± 0.78 | 4.46 ± 0.44 | 0.64 ± 0.23 | 2.56 ± 0.52 | 0.61 ± 0.23 | 1.33 ± 0.22 | 0.003 | 0.024 |
| Flap/bird/h | 9.99 ± 2.60 | 30.87 ± 2.74 | 17.96 ± 2.42 | 20.38 ± 1.58 | 3.53 ± 1.05 | 12.44 ± 1.64 | 4.29 ± 0.90 | 7.00 ± 0.77 | 0.021 | < 0.001 |
- a Not analyzed because of the dependence to lying/sitting. na, not analyzed due to the limited number of cases. Significance at P < 0.05.
In CC versus EC, the frequency of rising attempts was five/bird/hour versus less than one/bird/hour, while wing flapping was seven/bird/hour versus 20/bird/hour (P = 0.021) and balance loss was one versus four/bird/hour (P = 0.003). Behaviors of both groups differed significantly according to the route tract (balance loss: P = 0.024; rising attempts: P = 0.019; wing flapping: P < 0.001) with a lower frequency of events at H. No difference was observed between S1 and S2. No scratches, fractures or hematomas were detected in any animal after transportation.
Discussion
Previous studies have investigated the effect of crate height on poultry behavior, indicating that hens have a strong preference for a large vertical space allowance (Dawkins 1985) and that spatial restriction may increase the costs of performing certain comfort activities (Nicol 1987). These results seem to be valid also for male turkeys in transport crates at lairage, as Wichman et al. (2010) found a reduced turning, stepping and preening, and an increased number of rising attempts, in birds confined in 40 cm crates compared to birds in 55 and 90 cm crates. Conversely, the utilization of lower crates for birds that are not usually confined (e.g. mallards) was found to be less stressful (as revealed by blood markers) than keeping them in crates that allowed for vertical movements (Bedanova et al. 2014). This seems to be relevant for turkeys, as Wichman et al. (2012) found that transport crates of 55 cm height compared to those of 40 cm were associated with significantly more male turkeys with scratches on the back on arrival at the abattoir after commercial preslaughter transportation.
The present study investigated the behavior of female turkeys associated with crate height during commercial pre-slaughter transportation, as this was not investigated in previous studies. Our results are in agreement with Wichman et al. (2010) regarding the more frequent rising attempts of female turkeys in conventional versus experimental crates. It seems that under transport conditions, such increase is much more pronounced compared to lairage conditions: around five times in a journey of 76 min in comparison with only twice in a 6 h lairage (Wichman et al. 2010).
Interestingly, however, the frequency of standing under transport conditions was similar (i.e. around 30%–40%) to that reported during the 6 h lairage (i.e. 39%–46%) (Wichman et al. 2010). This suggests that animals were not prone to adapt their behavior in response to vibrations, noise and air movement. In the 77 cm height crate we also observed a three times higher occurrence of wing flapping and around three times more losses of balance/bird/hour.
Preening and stretching are often categorized as comfort behaviors; Wichman et al. (2010) found a reduction in preening with decreasing crate height and no stretching was observed in the 40 cm crates. We found that the comfort behaviors observable at lairage such as preening and stretching were completely missing during transportation. Therefore, these welfare indicators cannot be of use in this context to compare the crates, as transportation is a challenging event for birds to cope with. This should be of particular concern in northern Italy, where this study was performed, as it is a densely populated poultry area, with a density of more than 10 000 birds/km2 and a total of 7067 km2, but only a few abattoirs to where animals are transported (Mulatti et al. 2011).
The near absence of comfort behaviors was observed in both route tracts. On the other hand, the one-lane street tract was associated with a three times higher frequency of wing flapping, balance loss and rising attempts in comparison to the highway tract. These results may suggest that crossroads, bends and traffic circles (which are likely to provoke sudden brakes and skids) can have a negative impact on turkeys’ welfare, although the speed is lower than on the highway (i.e. an average of 51 km/h vs. 102 km/h). A possible effect of habituation can be hypothesized. Indeed, although S1 and S2 were identical, animals seemed to be quieter at S2, that is the frequency of the aforementioned stress-related behaviors was lower.
Potentially dangerous behaviors (i.e. wing flapping and mounting) were not frequent enough to cause fractures, hematomas or scratches. Although transportation evidenced a significant physiological stress response, no difference related to crate height was found. This result is in accordance with Wichman et al. (2010) who found no significant variation in lymphocyte ratio, creatine kinase activity or aspartate aminotrasferase activity in male turkeys housed in 40 versus 55 versus 90 cm crates. These authors also suggested that the behavioral indicators are the most sensitive ones to detect differences related to housing conditions. In turkeys, transport-related stress has been found to be associated with gender, genetic line (Huff et al. 2007) and weather conditions (Petracci et al. 2006); thus, these preliminary observations need to be corroborated in different conditions (e.g. cold and hot weather, male gender, etc.).
In conclusion, the present study indicates a clear impact of transportation rather than crate height on poultry physiology. A higher frequency of stress behaviors (i.e. rising attempts) was recorded in response to vertical constraint. However, potentially dangerous behaviors, such as mounting of crate-mates, wing flapping and balance losses, were more frequent in crates enabling a standing position. Hence, freedom of movement is more likely to impair than improve welfare and the actual outcome is to recommend the use of conventional crates. However, further studies will be needed to investigate whether a crate height can be defined which allows a good variety of comfort behaviors while still restricting birds from performing actions that result in injury.
The behavioral repertoire exhibited by the turkeys also changed depending on the tract of the route travelled, with busy and curvy routes increasing the frequency of stress-related behaviors. Therefore, a further recommendation is to avoid as much as possible these types of routes when planning preslaughter transportations.
Acknowledgments
This study was funded by the Italian Ministry of Health (Project RC IZSVe 07/12; B28C13000410001).
