Journal of Veterinary Behavior 61 (2023) 19–24 Contents lists available at ScienceDirect Journal of Veterinary Behavior journal homepage: www.journalvetbehavior.com Consistency of Convenience Sampling Order and its Association With Response to Handling and Weaning in Domestic Pigs (Sus scrofa) Heng-Lun Ko∗, Xavier Manteca, Pol Llonch Department of Animal and Food Science, School of Veterinary Science, Universitat Autònoma de Barcelona, Cerdanyola del Vallès, Barcelona, Spain a r t i c l e i n f o a b s t r a c t Article history: Convenience sampling in animal experiments may affect findings due to individual difference in temper- Received 7 April 2021 ament. The aim in this study was to assess the consistency of convenience sampling order (ORDER), and Revised 17 January 2023 to study the association between potential factors (i.e., behaviors related to handling, and coping ability Accepted 17 January 2023 to weaning) and ORDER in pigs (n = 325). ORDER was recorded by catching the closest pig in the pen on Available online 24 January 2023 Day (D) 1, 14, 23, 27, 31, 38, 69 and 79, while weighing pigs. Response to handling was assessed on all Keywords: weighing days except D69. For D1 to 38, vocalization and attempt of escaping were recorded on a Yes/No Coping style basis; for D79, a scoring system from ‘very difficult’ to ‘easy to handle’ was used for 3 handling phases: Handling ‘Moving into the crate,’ ‘In the crate,’ and ‘Leaving the crate’. ORDER within each pen was categorized to Personality 4 quartiles. To study the association between coping ability to weaning and ORDER, salivary cortisol and Salivary stress biomarkers chromogranin A (CgA) were determined on pre- and post-weaning. The response variables for statistical Sampling analysis were ORDER: either 1) the log-transformed percentage of ORDER; or 2) first quartile (1QT, i.e., Temperament first-caught) vs. the rest; or 3) fourth quartile (4QT, i.e., last-caught) vs. the rest. An individual effect on ORDER was found (P < 0.001), suggesting that a pig that had been caught either first, middle, or last was very likely to be caught in a similar order in the following handling sessions. On the other hand, neither responses to handling nor coping ability to weaning were associated with ORDER (P > 0.05). To conclude, although ORDER is not associated with responses to human handling and coping ability to weaning, con- venience sampling in pigs should be cautious as ORDER is highly consistent and long-lasting. © 2023 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Introduction practice to reduce bias (Bespalov, Wicke and Castagné, 2020). Ran- domization should be taken into account, especially during sample A recent report conducted in 2017 indicated that only 11.2% of selection, for several reasons. Animal temperament is among the the analyzed publications clearly described which method of ran- reasons of concern, because the way an animal reacts when being domization was used to allocate animals to different treatment sampled (i.e., handled) would likely affect the study results. groups in the experiments. On the other hand, 64.2% of the ana- Temperament, often known as personality or behavioral syn- lyzed publications stated using randomization (Macleod, The Na- drome synonymously in literature (MacKay and Haskell, 2015), ture Publication Quality Improvement Project (NPQIP) Collabora- refers to the individual behavioral differences within a popula- tive Group, 2017). In animal studies, randomization is a common tion, with the expectation that the differences remain consistent over time and across situations (Réale et al., 2007). It has been observed in a number of species, from insects, reptiles, to fish, birds and mammals (Brehm and Mortelliti, 2018), including pigs ∗ Corresponding author: Heng-Lun Ko, Department of Animal and Food Science, (O’Malley et al., 2019). Temperament might be a source of bias School of Veterinary Science, Universitat Autònoma de Barcelona, Cerdanyola del Vallès, Barcelona 08193, Spain, Tel.: ±34 93 581 1352; Fax: ±34 93 581 14 94. and potentially affect the validity of results especially in behav- E-mail address: henglun.ko@uab.cat (H.-L. Ko). ioral and physiological studies, and their interpretation, due to https://doi.org/10.1016/j.jveb.2023.01.004 1558-7878/© 2023 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) H.-L. Ko, X. Manteca and P. Llonch Journal of Veterinary Behavior 61 (2023) 19–24 unrepresentative group profile (Biro and Dingemanse, 2009; Carter were dead or lost to follow-up (e.g., transferred to other hous- et al., 2012; Michelangeli et al., 2016). For instance, Montiglio et ings). At 25 days of age, 261 piglets (approximately 5 kg on aver- al. (2012) found that chipmunks (Tamias striatus) that were faster age) were weaned and regrouped (based on similar body size) to a explorers, were trapped more frequently, showed an increased nursery with 8 pens (220 × 200 cm slatted flooring and 100 × 200 sympathetic activity during restraint, and had a relatively stable cm solid heated flooring) in the same room. The initial stocking fecal cortisol level over 5 months, compared to slower explorers. density of the nursery pen was ∼0.20 m2 /animal. During the nurs- As animal temperament is associated with the mechanism of ery period, 53 pigs were dead or lost to follow-up (e.g., transferred how the individuals respond to potential risks or fearful stimuli to other housings). At 71 days of age, 208 pigs (approximately 17 (e.g., human approach or handling) (D’Eath et al., 2009), and how kg on average) were regrouped (based on similar body size) to a they adapt to novelty or an environment (hereafter we refer to fattening unit with 19 pens (314 × 206 cm concrete slatted flooring this mechanism as ‘coping strategy’ or ‘coping style’) (O’Malley and 314 × 60 cm concrete solid flooring) in the same room. The et al., 2019), several studies have suggested an effect of temper- initial stocking density of the fattening pen was ∼0.62 m2 /animal. ament on capture (i.e., trapping) in wild animals during sampling In nursery and fattening pens, pigs had unrestricted access to wa- (Michelangeli et al., 2016). For example, Carter et al. (2012) con- ter from nipple drinkers and were fed ad libitum with commercial firmed that bold wild lizards (Agama planiceps) are likely to be diets corresponding to their production stages: a three-phase feed- trapped sooner than the shy ones. Wilson et al. (2011) also discov- ing program during nursery (2,480, 2,470, and 2,460 Kcal/kg) and ered that angling technique captures more timid bluegill sunfish a two-phase feeding program during fattening (2,488 and 2,477 (Lepomis macrochirus). Recent evidence also discovered that ani- Kcal/kg). mal temperament plays an important role on fitness (e.g., stress response) when the individual interacts with the social and physi- Experimental design cal environment (Michelangeli et al., 2016; O’Malley et al., 2019). In most of the publications, it is considered to be acceptable Sampling order and weighing to report the randomization procedure by giving statements like Piglets were individually identified with numbered ear-tags “X animals were randomly assigned to the Y treatment group” (Importvet, Centelles, Spain) after birth (Day [D] 1). The ORDER of (Bespalov et al., 2020), which may lead to room for doubt about each pig was recorded (ranked from 1 to n animals in each pen) the methodology of randomization. We wondered whether the or- while weighing them individually. Weighing occurred on D1, 14, der of the pigs caught would be consistent over repeated handling 23, 27, 31, 38, 69 and 79. All the handlers were from our univer- sessions, if a handler remain at the same location of the pen, and sity and had the previous experience in pig handling. Only 1 han- randomly catch the closest pig at his/her convenience (i.e., conve- dler could enter the pens and randomly catch/handle the nearest nience sampling), one-by-one. We also asked if convenience sam- pig. Ear-tag number could only be recognized from a close distance pling can affect validity of results in studies due to the individ- but not possible to read it without restraining the pig. Thus, for the ual difference in temperament. Scarce literature is available on the weighing tasks from D1 to D69, piglets were gently lifted from the consistency of convenience sampling order (i.e., handling order, ground by holding 1 of its back legs with 1 hand and supporting ORDER) in group-housed domestic pigs (Sus scrofa), and whether its chest with another hand. Piglets were then placed into a plastic ORDER is associated to other behavioral traits such as tempera- crate (55 × 40 × 40 cm), which sat on a weighing scale (PB-4040- ment and coping strategy. The objectives of the study were to 60, Balanzas Cobos, L’Hospitalet de Llobregat, Spain), one by one. study the consistency of ORDER in pigs, and to study the asso- The bottom of the crate was attached with anti-slip tape to mini- ciation between responses to handling, coping styles and ORDER. mize slipping while pigs were in the crate. After the body weight We hypothesized that ORDER in pigs would remain consistent in was recorded, the upper back of the piglet was marked with a different handling sessions (i.e., first-caught pigs would often be crayon (RAIDEX, Dettingen an der Erms, Germany) to be easily dis- caught first; last-caught pigs would often be caught last). We also tinguished from the remaining piglets, and were released to their hypothesized that responses to handling and copying styles would home pens with the same lifting technique from the crate. be associated with ORDER. For the weighing task of D79, pigs were unable to be lifted due to their weight but were instead guided to a mobile weighing crate Materials and methods (Meier Brakenberg, Germany) suitable for the body size. The mo- bile weighing crate was placed in front of the entrance of each The study took place in a commercial farm in summer 2017 in home pen. The handler entered the pen with a sorting panel and Lleida (Spain). Pigs (n = 325) were followed from birth to fattening. a paddle stick to facilitate the weighing process. After the body Pigs were managed as per routine practices (feeding, cleaning, and weight was recorded, pigs were spray-marked (RAIDEX, Dettingen caring) by experienced farm staff. an der Erms, Germany) and guided to a corridor until all their pen mates were weighed, and then they returned to their home pen Animals and housings altogether. Piglets were first raised in farrowing pens with their sows and Assessment of the responses to handling during weighing littermates for 25 days, then to nursery pens for 46 days, and then Response to handling was assessed on all weighing days, ex- to fattening pens until they were transported to a slaughterhouse. cept for D69 due to the absence of the observers. For the weighing Raising pigs from birth to slaughter took about 6 months. tasks from D1 to D38, vocalization and attempt of escaping were Twenty-three Danbred sows (10 primiparous and 13 multi- recorded by 2 observers on a Yes (1) / No (0) basis, while a piglet parous) were housed in a farrowing unit of 6 rooms 4 days before was in the plastic crate. An attempt of escaping was considered parturition until weaning. Each sow was confined in a farrowing when the piglet showed a high frequency of locomotion in the crate (190 × 62.5 cm) of a pen (253 × 168 cm) with complete slat- plastic crate and/or lifted and leaned its front legs on the plastic ted flooring. There were 325 piglets at birth (male: 154; female: crate. For the weighing task of D79, a scoring system from D’Eath 171). Within 24 h post-farrowing, new-born piglets were cross fos- et al. (2009) was applied to assess the response to handling dur- tered to standardize the litter size, making the litter size 14.1 ± 0.1 ing weighing by 1 observer (Table 1). The scoring system includes piglets per sow on average. During the suckling period, 64 piglets 3 phases of the handling: ‘Moving into the crate’ (MIC) (from 1 to 20 H.-L. Ko, X. Manteca and P. Llonch Journal of Veterinary Behavior 61 (2023) 19–24 Table 1 Scoring system extracted from D’Eath et al. (2009) to assess the response to handling during weighing. Score Description Moving into the crate (MIC) 1 Pig is very difficult to move and is trying to escape. 2 Pig is difficult to move into the crate. 3 Pig moves into the crate with some assistance from the handler. 4 Pig walks into the crate with little or no encouragement. 5 Pig runs forward into the crate. In the crate (ITC) 1 Pig moves around a lot during weighing, jumping and crashing around. 2 Pig moves around during weighing. 3 Pig stands still during weighing. Leaving the crate (LTC) 1 Pig resists and is very difficult to push out of the weigh crate. 2 Pig moves out of the weigh crate after some pushing. 3 Pig leaves of its own accord once the door is opened. 5), ‘In the crate’ (ITC) (from 1 to 3), and ‘Leaving the crate’ (LTC) Consistency of the sampling order (from 1 to 3), with 1: very difficult to handle, to 3 or 5: easy to To analyze the consistency of ORDER, a Spearman correlation handle. test was performed, in which the ORDERs of an individual pig recorded throughout the study period were correlated altogether. Additionally, a linear model was applied: the response variable was the log-transformed percentage of ORDER (i.e., ORDER/number of Saliva collection around weaning and salivary analysis animals per pen), and the fixed effects were day and individual. Weaning is a common practice in commercial farms which To analyze the degree of consistency of ORDER, several chi- poses multiple social and environmental challenges to piglets, from square goodness of fit tests were applied, by comparing the ob- diet and housing transitions, mixing with unfamiliar conspecifics, served and the expected values of ORDER consistency. ORDER to abrupt separation from the sows (Campbell et al., 2013). It has within each pen was first categorized into 4 quartiles in each han- therefore been considered as an ideal model to study the cop- dling session. The response variables were either the first quar- ing ability (i.e., adaptive stress response) of piglets towards stress tile (1QT, i.e., first-caught pigs) vs. the rest, or the fourth quartile (Royer et al., 2016). Seventeen out of the 23 litters were selected (4QT, i.e., last-caught pigs) vs. the rest. ORDER in different pro- randomly for saliva collection. From these litters, 6 piglets per duction periods was analyzed separately. To have similar proba- litter were pre-selected for saliva samples: a male and a female bility of being caught in 1QT and in 4QT within a pen, the data piglet of the heaviest, the middle, and the lightest birth weight on D79 were removed because it was the only handling session in within each litter. On the weighing task of D23, before releasing the fattening period. In the suckling period, only the probability of the piglets back to their home pens, those piglets that were se- 0.29 in each pen and each handling session was kept for the anal- lected for saliva sampling were spray-marked in advance. Saliva ysis, meaning only when the pen size was 14 piglets, making it samples were collected on 1D pre-weaning (D24) and 1D post- around 295 piglets. In the nursery period, the probability between weaning (D26). Saliva samples were obtained by first gently lift- 0.24 and 0.26 in each pen and each handling session was kept for ing the marked piglets 1 by 1 and introducing the cotton swabs the analysis, making it around 318 pigs. The response variable of into piglets’ mouths for 1 minute. The cotton swab was available the observed values was created, based on the number of times a in the Salivette tube (Sarstedt AG & Co., Nümbrecht, Germany). pig was in 1QT or 4QT in the total handling sessions (values be- Those piglets were spray-marked again on their upper back on tween 0 and 3 for the suckling, and between 0 and 4 for the nurs- D24, so that the mark remained until D26. Right after the col- ery). The response variable of the expected values was determined lection, each piglet was released, and samples were centrifuged by the goodness of fit test with the theoretical binomial distribu- (Avanti J-20 XP, Beckman Coulter, California, USA) for 10 minutes tion (size = 3 and probability = 0.29 for the suckling; and size = 4 at 30 0 0 rpm and were stored at -20°C until analysis. Salivary cor- and probability = 0.25 for the nursery period). Bonferroni correc- tisol (CORT) (μg/dL) and chromogranin A (CgA) (μg/mL) were de- tion was conducted after the multiple analyses. termined from the saliva samples. An automated chemilumines- cence immunoassay (Immulite 10 0 0 Cortisol, Siemens Medical So- lutions USA, Malvern, USA) was used to detect CORT (Escribano et Association between the responses to handling and the sampling al., 2012) and time-resolved immunofluorometry assays (TR-IFMA) order were used to detect CgA (Escribano et al., 2013). The intra- and For the responses to handling in a crate from D1 to D38, 2 lin- inter-assay coefficients of variations (CV) were lower than 16% and ear mixed effect models (LMM) were applied: the log-transformed 10% for CORT and CgA, respectively. The detection limit was 0.016 percentage of ORDER as the response variable, the individual μg/dL and 4.27 ng/mL for CORT and CgA, respectively. nested in the production period as the random effect, with the following fixed effects in each model: 1) day, sex, relative body weight, and vocalization (0/1); 2) day, sex, relative body weight, Statistical analysis and attempt of escaping (0/1). Relative body weight of an individ- ual was calculated by [actual weight of the individual – mean of Results were analyzed and figures were generated in RStudio the weight on the day]. version 2022.02.3 (R Foundation, Austria). Statistical significance For the responses to handling on D79, the response variable of was accepted when P < 0.05 and a tendency was considered when the following LMMs was the log-transformed percentage of ORDER, 0.05 < P ≤ 0.10. Results were reported with least square means ± and the random effect was the fattening pen, with the following standard errors. fixed effects in each model: 1) sex, relative body weight, and MIC 21 H.-L. Ko, X. Manteca and P. Llonch Journal of Veterinary Behavior 61 (2023) 19–24 score; 2) sex, relative body weight, and ITC score; 3) sex, relative Association between the coping ability to weaning stress and the body weight, and LTC score. sampling order CORT (n = 65, P = 0.36) and CgA (n = 59, P = 0.79) were not as- Association between the coping ability to weaning stress and the sociated with ORDER. Our results suggested that the coping ability sampling order to weaning stress may not be associated with the sampling order Salivary stress biomarkers (CORT and CgA) were used as proxies in pigs. of weaning stress (Yang et al., 2018; Escribano et al., 2019). Two LMMs were applied, the log-transformed percentage of ORDER as Discussion the response variable, the farrowing pen as the random effect, with the following fixed effects in each model: 1) sex and the difference In this study, we explored whether ORDER was consistent in re- of CORT between D24 and D26; 2) sex and the difference of CgA peated handling of group-housed pigs from birth to fattening. We between D24 and D26. also studied possible factors that could be associated with ORDER, combining some behavioral and physiological traits related to re- sponse to handling and weaning stress, which are known to be as- Results sociated with temperament. We found that ORDER was highly consistent in pigs during Consistency of the sampling order eight handling sessions, which lasted almost 3 months. This state- ment is based on the fact that, when comparing the observed val- ORDER was highly consistent throughout the handling sessions ues (i.e., our results, when catching was at our convenience) with (P < 0.001). It reflects the great likelihood that the pigs that were the expected values (i.e., if catching was randomized), we discov- caught first would be caught first again in the following sessions, ered that more pigs than expected were either in the first or the and the same was true for those that were caught in the mid- last quartile more times than it would be in repeated handling ses- dle or the last. The result of the correlation test showed that OR- sions, especially during the nursery period. In the studies of Réale DERs were more positively correlated with each other during the et al., (20 0 0) and Le Cœur et al. (2015), they also found a repeated same production period. During the suckling period, rD1-14 = 0.17 individual trappability in bighorn ewes (Ovis canadensis, multi-year and rD14-23 = 0.18 (both P < 0.01), and during the nursery pe- study) and Siberian chipmunk (Tamias sibiricus, 1-year study), re- riod, rD27-31 = 0.40, rD27-38 = 0.29, rD31-38 = 0.47, rD31-69 = 0.25, and spectively. The phenomenon of this repeatable ORDER in pigs may rD38-69 = 0.40 (all P < 0.001), the correlation coefficients suggest reflect a systematic bias if convenience sampling (e.g., for blood, that ORDER was repeatable while pigs were in the same pen. The saliva, feces, or behavior) is carried out in the captive environment, degree of ORDER consistency in 1QT and 4QT during the suck- which could affect the validity of results and statistical assump- ling and the nursery periods is shown in Figure 1(a) - (d), as tions (Carter et al., 2012). A relatively higher consistency of ORDER they represent the number of handling sessions a pig was caught during the nursery period, compared to the suckling period, could in the same quartile. The expected values (purple bars) repre- be due to the absence of the sow and the farrowing crate in the sent ORDER if that would be randomized, which was significantly environment, which may complicate the catching procedure. different from the observed values (green bars) in both quartiles The practice of weighing consists of several challenges for pigs, ([a] 1QT during suckling: X2 3 = 83.93; [b] 4QT during suckling: which includes novel environment, isolation, confinement and hu- X2 3 = 89.12; [c] 1QT during nursery: X2 4 = 75.43; [d] 4QT during man handling (D’Eath et al., 2009). Previous studies based on 3 nursery: X2 4 = 49.48; all P < 0.001). In Figure 1 (c) and (d), it ap- classical temperament tests of coping style (including human ap- peared that more pigs than expected were repeatedly caught in proach test, novel object test and open door test), described that the same quartile during the nursery period (i.e., the observed val- rapid approach to human or novel object, or rapid exit the home ues were greater than the expected ones) (1QT: when x = 1, 3, and pen was considered as proactive coping style (similar to first- 4; 4QT: when x = 1, 2, 3, and 4). On the other hand, compared to caught pigs), whereas slow approach or reluctant to exit was con- the nursery period, there were less pigs than expected that were sidered as reactive copying style (similar to the last-caught pigs) repeatedly caught in the same quartile during the suckling period (Brown et al., 2009). The association between response to han- (1QT: when x = 1 and 2; 4QT: when x = 0), as shown in Figure 1 (a) dling and coping style (represented by ORDER) was confirmed by and (b). This highly repeatable ORDER pattern in the nursery pe- our results. At the weighing task of D79, we found that the MIC riod than the suckling period is also reflected on the correlation score was negatively correlated with ORDER, which means that test as mentioned above. pigs showing an easier handling (high MIC score) were caught first, and difficult handling (low MIC score) were caught last. However, attempt of escaping and ITC score did not seem to be associated Association between the responses to handling and the sampling with ORDER while weighing. Although immobility or fight/flight order represents the passive or active responses towards fearful situa- tions (Erhard and Mendl, 1999), we did not find the association For the handling sessions in the plastic crate from D1 to 38, sex, between attempt of escaping/ITC score and ORDER. It could be that relative body weight, vocalization (P = 0.72), and attempt of escap- the duration of the isolation in the crate was relatively short (few ing (P = 0.25) were not associated with ORDER. For the handling seconds) to observe different fear responses among individuals in session of D79, the MIC score (P < 0.001) was negatively associ- our study, compared to other studies (e.g., 1 minute in Bolhuis, ated with ORDER, but the ITC (P = 0.95) and LTC scores (P = 0.45) 2004; 10 minutes in Adcock et al., 2015). were not associated with ORDER, indicating that early-caught pigs In the present study, pigs hardly vocalized when being iso- were easier to handle when entering the weight crate (i.e., higher lated in the crate, and we did not find a strong association be- MIC score), and last-caught pigs were more difficult to handle (i.e., tween vocalization and ORDER either. Current knowledge regard- lower MIC score). In addition, sex was associated with ORDER dur- ing the association between vocalization and temperament in pigs ing the handling session on D79 (P < 0.05), in which male pigs is rather inconsistent. In Hessing et al. (1993) and (1994), passive appeared to be easier to handle than the female pigs. pigs (i.e., shy, or reactive coping style) vocalized more. However, 22 H.-L. Ko, X. Manteca and P. Llonch Journal of Veterinary Behavior 61 (2023) 19–24 Figure 1. Bar charts presenting the observed (green) and expected (purple) values of the degree of sampling order consistency in (a) the first quartile (1QT) during suckling, (b) the fourth quartile (4QT) during suckling, (c) the 1QT during nursery, and (d) the 4QT during nursery. The degree of sampling order consistency was calculated as follows: if a pig was caught in the same quartile (either 1QT or 4QT) of the sampling order for three times out of three handling sessions, it got the value of three; if a pig was caught for one time out of three handling sessions, it got the value of one. The distribution of the expected values was determined by using the chi-square goodness of fit tests with the binomial distribution (size = 3, i.e., D1-23 and probability = 0.29 for the suckling period; size = 4, i.e., D27-69 and probability = 0.25 for the nursery period). The observed and the expected values in all figures were significantly different (P < 0.001). in Geverink et al. (2002), high-resisting pigs (i.e., bold, or proactive axis and low reaction of SAM axis (i.e., conservation/withdrawal). coping style) vocalized more than low-resisting ones (i.e., reactive). Nevertheless, Ruis et al. (20 0 0) reported that low-resisting pigs Although the frequency of vocalization could differ due to differ- (i.e., proactive) showed a higher activity of HPA axis than high- ent handling process (backtest restraint in Hessing et al. (1993), resisting ones (i.e., reactive). More studies are therefore necessary and nose sling restraint in Geverink et al. (2002)), more research is to confirm the association between stress axis activation and tem- needed to confirm the association between vocalization and tem- perament traits. perament in pigs (O’Malley et al., 2019). Additionally, as proposed in Koolhaas et al. (2010), animal tem- In terms of coping ability to weaning stress, both ADG around perament is a two-tier model which is divided by 2 independent weaning and salivary stress biomarkers were not associated with dimensions, including the quality of the response towards a chal- ORDER. The association between stress axis activation and temper- lenging circumstance (i.e., coping style) and the quantity of that ament has not been consistent in the literature (Bolhuis, 2004). For response (i.e., stress reactivity). This model suggests the possibil- instance, Koolhaas et al. (1999), Ruis et al. (20 0 0) and Adcock et ity for the proactive individual having a strong tendency to act al. (2015) explained that proactive pigs represent a higher general with low stress reactivity; and the reactive individual initiating lit- activity, and show low reaction of hypothalamic-pituitary-adrenal tle behavior with high stress reactivity, which could be the case (HPA) axis and high reaction of sympathetic-adrenomedullary for last-caught pigs (i.e., reactive) in our study. With the data from (SAM) axis (i.e., fight/flight), whereas reactive pigs, who respond the present study, the response to handling cannot assign the first- more with immobility and avoidance, show high reaction of HPA or last-caught pigs to the two-absolute temperaments (i.e., proac- 23 H.-L. Ko, X. Manteca and P. Llonch Journal of Veterinary Behavior 61 (2023) 19–24 tive/reactive; active/passive; bold/shy; low-resisting/high-resisting), Brown, J.A., Dewey, C., Delange, C.F., Mandell, I.B., Purslow, P.P., Robinson, J.A., and thus the association between response to handling and cop- Squires, E.J., Widowski, T.M., 2009. Reliability of temperament tests on finish- ing pigs in group-housing and comparison to social tests. Appl. Anim. Behav. ing style needs further investigation before any conclusion could Sci. 118, 28–35. doi:10.1016/j.applanim.20 09.02.0 05. be drawn. Campbell, J.M., Crenshaw, J.D., Polo, J., 2013. The biological stress of early weaned piglets. J. Anim. Sci. Biotechnol. 4, 1–4. doi:10.1186/2049-1891-4-19. Carter, A.J., Heinsohn, R., Goldizen, A.W., Biro, P.A., 2012. Boldness, trappability Conclusions and sampling bias in wild lizards. Anim. Behav. 83, 1051–1058. doi:10.1016/j. anbehav.2012.01.033. In the present study, we found that pigs handled in repeated D’Eath, R., Roehe, R., Turner, S., Ison, S., Farish, M., Jack, M., Lawrence, A., 2009. Genetics of animal temperament: Aggressive behaviour at mixing is genetically sessions showed a high consistency in ORDER when they were associated with the response to handling in pigs. Animal 3, 1544–1554. doi:10. caught at convenience in a group. This ORDER consistency reflects 1017/S1751731109990528. an individual’s long-lasting effect towards human handling in pigs Erhard, H.W., Mendl, M., 1999. Tonic immobility and emergence time in pigs—more evidence for behavioural strategies. Appl. Anim. Behav. Sci. 61, 227–237. doi:10. (i.e., from suckling to fattening). ORDER was not determined by sex 1016/S0168-1591(98)00196-8. or relative body weight within a group. However, behavior used Escribano, D., Fuentes-Rubio, M., Cerón, J.J., 2012. Validation of an automated chemi- as temperament indicators such as reluctance to enter the weight luminescent immunoassay for salivary cortisol measurements in pigs. J. Vet. Di- crate, occurred more frequent in the last-caught pigs. In conclu- agn. Invest. 24, 918–923. doi:10.1177/1040638712455171. Escribano, D., Ko, H.L., Chong, Q., Llonch, L., Manteca, X., Llonch, P., 2019. Salivary sion, sampling pigs at the handler’s convenience within a group biomarkers to monitor stress due to aggression after weaning in piglets. Res. may lead to a systematic bias. Vet. Sci. 123, 178–183. doi:10.1016/j.rvsc.2019.01.014. Escribano, D., Soler, L., Gutiérrez, A.M., Martínez-Subiela, S., Cerón, J.J., 2013. Mea- surement of chromogranin A in porcine saliva: validation of a time-resolved im- Authorship statement munofluorometric assay and evaluation of its application as a marker of acute stress. Animal 7, 640–647. doi:10.1017/S17517311120 020 05. The idea for the paper was conceived by Heng-Lun Ko and Pol Geverink, N.A., Schouten, W.G.P., Gort, G., Wiegant, V.M., 2002. Individual differ- ences in behavioral and physiological responses to restraint stress in pigs. Phys- Llonch. The experiments were designed by Heng-Lun Ko, Xavier iol. Behav. 77, 451–457. doi:10.1016/S0 031-9384(02)0 0877-6. Manteca, and Pol Llonch. The experiments were performed by Hessing, M.J., Hagelsø, A.M., Schouten, W.G., Wiepkema, P.R., Van Beek, J.A., 1994. Heng-Lun Ko, Pol Llonch, and Qiai Chong. The data were analyzed Individual behavioral and physiological strategies in pigs. Physiol. Behav. 55, 39– 46. doi:10.1016/0 031-9384(94)90 0 07-8. by Heng-Lun Ko, and were interpreted by Heng-Lun Ko, Xavier Hessing, M.J., Hagelsø, A.M., Van Beek, J.A.M., Wiepkema, R.P., Schouten, W.G.P., Manteca, and Pol Llonch. The paper was written by Heng-Lun Ko. Krukow, R., 1993. Individual behavioural characteristics in pigs. Appl. Anim. Be- The paper was reviewed and edited by Heng-Lun Ko, Xavier Man- hav. Sci. 37, 285–295. doi:10.1016/0168-1591(93)90118-9. Koolhaas, J.M., De Boer, S.F., Coppens, C.M., Buwalda, B., 2010. Neuroendocrinol- teca, and Pol Llonch. The final version to be submitted was ap- ogy of coping styles: towards understanding the biology of individual variation. proved by Heng-Lun Ko, Xavier Manteca, and Pol Llonch. Front. Neuroendocrinol. 31, 307–321. doi:10.1016/j.yfrne.2010.04.001. Koolhaas, J.M., Korte, S.M., De Boer, S.F., Van Der Vegt, B.J., Van Reenen, C.G., Hop- Acknowledgements ster, H., De Jong, I.C., Ruis, M.A.W., Blokhuis, H.J., 1999. Coping styles in ani- mals: current status in behavior and stress-physiology. Neurosci. Biobehav. Rev. 23, 925–935. doi:10.1016/S0149-7634(99)0 0 026-3. We are grateful to Ramaderies Miqueló S.L. for offering the op- Le Cœur, C., Thibault, M., Pisanu, B., Thibault, S., Chapuis, J.L., Baudry, E., 2015. Tem- portunity to carry out the experiment in their farm and taking care porally fluctuating selection on a personality trait in a wild rodent population. Behav. Ecol. 26, 1285–1291. doi:10.1093/beheco/arv074. of the animals. We thank Prof. Irene Camerlink for giving advice on MacKay, J.R.D., Haskell, M.J., 2015. Consistent individual behavioral variation: the experimental design, Dr. Damián Escribano for conducting salivary difference between temperament, personality and behavioral syndromes. Ani- analysis, and Qiai Chong for assisting with data collection. We also mals 5 (3), 455–478. doi:10.3390/ani5030366. Macleod M.R., The Nature Publication Quality Improvement Project (NPQIP) Collab- gratefully acknowledge the statistical consultation provided by Ap- orative Group, 2017. Findings of a retrospective, controlled cohort study of the plied Statistics Service (SEA) of Universitat Autònoma de Barcelona. impact of a change in Nature journals’ editorial policy for life sciences research on the completeness of reporting study design and execution. bioRxiv:187245. doi:10.1101/187245 Ethical Considerations Statement Michelangeli, M., Wong, B.B., Chapple, D.G., 2016. It’s a trap: sampling bias due to animal personality is not always inevitable. Behav. Ecol. 27, 62–67. doi:10.1093/ All the experimental procedures were approved by the ethi- beheco/arv123. Montiglio, P.O., Garant, D., Pelletier, F., Réale, D., 2012. Personality differences are cal committee of Universitat Autònoma de Barcelona (FUE-2016- related to long-term stress reactivity in a population of wild eastern chipmunks, 00441221). Tamias striatus. Anim. Behav. 84, 1071–1079. doi:10.1016/j.anbehav.2012.08.010. O’Malley, C.I., Turner, S.P., D’Eath, R.B., Steibel, J.P., Bates, R.O., Ernst, C.W., Sieg- Conflict of Interest ford, J.M., 2019. Animal personality in the management and welfare of pigs. Appl. Anim. Behav. Sci. 218. doi:10.1016/j.applanim.2019.06.002, Article 104821. Réale, D., Gallant, B.Y., Leblanc, M., Festa-Bianchet, M., 20 0 0. Consistency of tem- The authors declare no conflict of interest. perament in bighorn ewes and correlates with behaviour and life history. Anim. Behav. 60, 589–597. doi:10.10 06/anbe.20 0 0.1530. Réale, D., Reader, S.M., Sol, D., McDougall, P.T., Dingemanse, N.J., 2007. Integrat- References ing animal temperament within ecology and evolution. Biol. Rev. 82, 291–318. doi:10.1111/j.1469-185X.20 07.0 0 010.x. Adcock, S.J., Martin, G.M., Walsh, C.J., 2015. The stress response and exploratory be- Royer, E., Barbé, F., Guillou, D., Rousselière, Y., Chevaux, E., 2016. Development of an haviour in Yucatan minipigs (Sus scrofa): Relations to sex and social rank. Phys- oxidative stress model in weaned pigs highlighting plasma biomarkers’ speci- iol. Behav. 152, 194–202. doi:10.1016/j.physbeh.2015.09.033. ficity to stress inducers. J. Anim. Sci. 94, 48–53. doi:10.2527/jas2015-9857. Bespalov, A., Wicke, K., Castagné, V., 2020. Blinding and Randomization. In: Be- Ruis, M.A., te Brake, J.H., van de Burgwal, J.A., de Jong, I.C., Blokhuis, H.J., spalov, A., Michel, M.C., Steckler, T. (Eds.), Good Research Practice in Non- Koolhaas, J.M., 20 0 0. Personalities in female domesticated pigs: behavioural Clinical Pharmacology and Biomedicine, Handbook of Experimental Pharmacol- and physiological indications. Appl. Anim. Behav. Sci. 66, 31–47. doi:10.1016/ ogy. Springer Nature, Cham, pp. 81–100, 257. doi:10.1007/164_2019_279. S0168-1591(99)0 0 070-2. Biro, P.A., Dingemanse, N.J., 2009. Sampling bias resulting from animal personality. Wilson, A.D., Binder, T.R., McGrath, K.P., Cooke, S.J., Godin, J.G.J., 2011. Capture Trends Ecol. Evol. 24, 66–67. doi:10.1016/j.tree.20 08.11.0 01. technique and fish personality: angling targets timid bluegill sunfish, Lepomis Bolhuis, J.E., 2004. Personality in pigs – Individual characteristics and coping with macrochirus. Can. J. Fish. Aquat. Sci. 68, 749–757. doi:10.1139/f2011-019. environmental challenges. Ph.D. thesis. Wageningen University, The Netherlands Yang, C.H., Ko, H.L., Salazar, L.C., Llonch, L., Manteca, X., Camerlink, I., Llonch, P., Retrieved from. 2018. Pre-weaning environmental enrichment increases piglets’ object play be- Brehm, A.M., Mortelliti, A., 2018. Mind the trap: large-scale field experiment shows haviour on a large scale commercial pig farm. Appl. Anim. Behav. Sci. 202, 7–12. that trappability is not a proxy for personality. Anim. Behav. 142, 110–113. doi:10.1016/j.applanim.2018.02.004. doi:10.1016/j.anbehav.2018.06.009. 24