Introduction

The breed of cattle bred for bullfighting is known as Lidia. It is the only breed exploited for its ethological performance. Since the 18^th Century, individuals have been selected based on behavioral traits observed during what is known as a tienta, a test of aggressiveness¹. Because it has been a genetic improvement process carried out at the herd level, it has produced multiple genetic lines characterized by phenotypic² and ethological³ traits that are both stable and defined⁴. The resulting bulls are raised for four to five years to fight in the bullring for approximately fifteen minutes⁵. The goal of selection is to produce bulls deemed apt for bullfights, which exhibit behavior broadly termed bravura (implying both ferocity and bravery). However, defining ideal behavior for a fighting bull is extremely difficult and highly subjective^6-7.

To produce bulls with the desired bravura, breeders keep exhaustive records of behavior in breeding animals during the tienta, and of bulls during bullfights⁸. This is a complex task that does not necessarily produce high heritability of the desired traits⁹. One challenge in attaining trait heritability is defining what the desired behavioral traits are. Many authors have tried to objectively describe the desired behavioral traits in fighting bulls by defining the ethological patterns, positive or negative, defining the opposing qualities of bravura and meekness^10-14. Some have devised ethological assessment methods based on the frequency of behavioral patterns. For example, there is a fighting bull rating table¹⁵, a bull aptitude test¹⁶, and bull evaluation sheets^17,18. In addition, a computer program and ethological assessment methodology¹⁹ have been implemented^20-22. These studies focus on quantifying or evaluating bull behavior, be it during the tienta or a bullfight. The intent is to select breeding animals or verify bull performance during bullfights.

A bull’s behavior during pre-bullfight handling (transport, unloading and stabling) may reflect the ethological traits it will exhibit during the bullfight. If this hypothesis holds true, observing bull behavior prior to a bullfight could provide valuable information on its potential behavior in the ring. This would be merely predictive since the performance of the bullfighters and the variables in each bullfight also strongly influence on a bull’s behavior. This connection has been suggested previously¹⁶, but without scientific analysis to support the argument. The present study objective was to analyze bull behavior during pre-bullfight handling in an effort to begin defining ethological guidelines that could predict bull behavior in the bullring.

Material and methods

Data were collected for two hundred animals which fought at bullrings in the cities of Valencia and Madrid, Spain. All animals were between three and five years of age. They came from 17 herds, and belonged to 6 different lines of the Lidia breed.

Pre-bullfight handling did not differ, following a standard procedure. This begins with transport to the bullring three days before the bullfight where the bulls are unloaded individually into a small pen (approx. 150 m²), and cooled with water until they have calmed down. They are then moved through a narrow corridor to a scale, where they are weighed individually, and into a larger corral (approx. 300 m²). Here they have an initial veterinary examination. The bulls remain together in the corral for 48 h. On the day of the bullfight, they are moved individually into another (approx. 300 m²) corral where they have a second, and final, veterinary examination.

Bull behavior during the pre-bullfight handling process was documented on a form by three veterinarians who agreed on the evaluation before recording it. Behavior was assessed at four stages in the process: unloading (1); first veterinary examination (2); second veterinary examination (3); and during time in corrals (4).

At stages 1, 2 and 3, mobility, aggressiveness and respiratory rate were assessed using a three-point scale (1= low, 2= intermediate, and 3= high). At stage 4, once all the animals were in the same corral, two types of collective behavior were assessed: nervousness, (1-calm, 2-vigilant, 3-nervous) and fight rate (1-infrequent, 2-moderate, 3-frequent).

Video recordings were taken of each animal’s bullfight and later evaluated by the same person, with experience in ethological assessment. The software and methodology described by Sánchez et al¹⁹, and validated by Alonso et al^23-25, were used. This allows evaluation of an animal’s behavior during each of the bullfight’s three periods (known as tercios). During each tercio, a total of 21 behavioral variables (see below) were graded on a five-point scale and recorded in an independent spreadsheet file, along with the duration of each tercio.

1. Exit speed (rapisal). The speed at which the bull runs through the door into the ring. 0= animal exits walking and stops in corridor. 5= animal exits at a run.

2. Stops at door (parapu). If the bull stops when it steps onto the sand. 0= animal maintains exit speed when stepping onto sand. 5 = animal stops.

3. Ring circuit (recorre). If the bull moves around the ring before being stopped by the first flourishes of a cape. 0= animal remains standing at some point in the bullring. 5= animal makes one or more circuits.

4. Distance at start of charge (acudlar). Distance at which the bull begins to charge when first provoked. 0= animal only charges when a bullfighter is very close. 5= animal charges at any distance, no matter distance to the bullfighter.

5. Attacks shields (remata). If bull gores or hits shield behind which bullfighter shelters. 0= under no circumstances does animal attack shield. 5= animal attacks shields every time it gets near them during initial lancing.

6. Head height vis-a-vis horse (humillacab). Estimated height at which bull places horns on horse’s body. 0= animal raises horns towards lancer. 5= animal places horns on lower portion of protective armor or on horse’s abdomen.

7. Push. Extent to which, once it has made contact, the bull uses its dorsal muscles and hind quarters to push the horse. 0= animal does not push at all, remains static or leans slightly into horse. 5= animal uses dorsals and hind legs in attempt to displace horse.

8. Goring. If bull gores horse’s protective armor and to what extent. 0= animal firmly pushes against armor, without changing from point of initial contact. 5= animal insistently gores armor, and even tries to detach the lance.

9. Release (suelto). If animal releases or disengages from horse after feeling pain from the lance, and runs away from horse without bullfighters challenging it. 0= animal remains with horse and does not escape, requires challenge. 5= animal disengages and quickly runs away after feeling pain from lance.

10. Response to pain (crecedol). Extent to which the bull, upon feeling pain, increases its force and aggressivity against the horse. 0= animal decreases aggressivity in response to pain. 5= animal attacks more decisively after lancer’s aggression.

11. Distance to banderillero (largoban). The distance from which the bull charges the banderillero (lancer on foot) when called. 0= animal waits until banderillero is very close. 5= animal charges before being called on all three attempts.

12. Focus on banderillero (fijoban). Extent to which animal focuses on banderillero or is distracted. 0= animal is constantly distracted by the crowd and other bullfighters. 5= animal focuses on banderillero from the first call until the lancing ends.

13. Follows banderillero (sigueban). The tenacity with which the bull follows the banderillero once the darts have been anchored in its back. 0= animal stands still after encounter. 5= animal insistently follows banderillero, normally until he takes cover behind a shield.

14. Gallop. Frequency with which bull gallops. 0= animal never gallops. 5= animal is constantly galloping.

15. Distance of charge at muleta (largomu). Distance from which bull begins charge at muleta (cape supported with stave). 0= animal does not charge until muleta is very close. 5= animal begins charge from large distance when given the option.

16. Head height vis-a-vis muleta (humillamu). Head height of bull when charging muleta. 0= animal keeps head high from beginning to end of charge. 5= animal lowers head when beginning charge and keeps it low until finishing.

17. Determination (codicia). The passes in each “set” of charges at the muleta can continue without pause after each one. This is frequently the case at the beginning of the third tercio, but can decrease as it progresses. 0= animal stops after each set. 5= animal does not pause between passes in all sets.

18. Delay (tardea). Number of calls or challenges required for bull to charge. 0= animal charges the moment cape is shown. 5= animal requires repeated calls, on each pass, before charging.

19. Charge area (embiste). The third tercio occurs in an area chosen by the bullfighter or one preferred by the bull. 0= encounters must happen in area preferred by animal, usually near the shields or the exit to the corrals. 5= animal exhibits no preference.

20. Focus on muleta (fijomul). Extent to which bull remains focused on muleta. 0= animal looks at the bullfighter or the crowd, is constantly distracted. 5= animal remains focused on muleta at all times.

21. Runs away from muleta (huyemul). Extent to which bull runs away from muleta, in search of an exit, after first charges. 0= animal exhibits no intent to avoid muleta. 5= animal constantly runs away, making it impossible to complete passes.

Results for all the variables described above were expressed as a mean with standard deviation. A one-way ANOVA was applied to identify any effect of genetic line or moment of observation (unloading, first examination, second examination) on the ethological variables documented pre-bullfight. If a factor of variation had two levels and was statistically significant (P≤0.05), a Student-Newman-Keuls test (P≤0.05) was run a posteriori to compare groups of homogeneous means. Any possible linear relationships between the ethological variables documented in the corrals and those documented during the bullfight were identified with a Pearson’s linear correlation test. All statistical analyses were run with the IBM^® SPSS^® ver. 19.0 software program²⁶.

Results

Analysis (ANOVA, Student-Newman-Keuls) of the ethological values collected during pre-bullfight handling indicated that animal attitude, represented by mobility, aggressiveness and respiratory rate, exhibited no changes during unloading and the first examination, which take place on the same day and consecutively. Aggressiveness and respiratory rate decreased significantly in the second examination (Figure 1).

Figure 1 Ethological values (means) during unloading and two veterinary examinations 

Analysis of bull behavior by genetic origin (line) identified differences between them during pre-bullfight handling (Table 1), but not during the bullfight. In the corrals, the Santa Coloma and Albaserrada lines exhibited the highest values for mobility, aggressiveness and respiratory rate, especially at unloading and first examination. Values for the nervousness variable were highest in Santa Coloma in the corrals, and highest in Albaserrada and Santa Coloma animals during the bullfight (P<0.05).

The Pearson’s linear correlation analysis identified multiple significant correlations (Tables 2, 3, 4, 5 and 6). Results have been simplified by grouping the results from the unloading and first examination stages since their means did not differ (Figure 1).

In the pre-bullfight handling results, the most valuable information in terms of predicting animal behavior during the bullfight is that for unloading and first examination. At this time, the animal is in a state of stress after transport, which brings out its intrinsic characteristics of mobility and aggressiveness. In the opening of the bullfight, mobility during unloading and first examination positively correlated to the parameter “Exit speed”, as well as to “Distance to banderillero” in the second tercio (banderillas) and “Charge area” during the third tercio (muleta). In contrast, the ethological parameter values for the second examination were not strongly predictive of animal behavior during the bullfight. However, the mobility parameter in this stage positively correlated with the “Gallops” parameter in the second tercio, a reflection of good physical condition in the corrals and the bullring. Also, the aggressiveness parameter in the second examination positively correlated with the “Delay” parameter in the third tercio.

Discussion

Research done in production animal management facilities to assess animal stress in response to transport to the slaughterhouse indicates that animals adapt to their new surroundings within a few hours^27-29. This agrees with the present results in which bull aggressiveness and respiratory rate values were significantly lower at the second examination after they had adapted to the corrals for two days, which coincides with previous reports³⁰.

In the pre-bullfight stages, bulls of the Santa Coloma and Albaserrada lines exhibited the highest mobility, aggressiveness and respiratory rate values during unloading and first examination. This coincides with genetic studies of the Lidia breed which state that this difference in behavior has resulted from selection focused on more temperamental or fierce behavior^17,31, which has raised consanguinity rates in these lines³². The literature addressing the characteristics of the Lidia lines also indicates that the Santa Coloma and Albaserrada lines have a different capacity to adapt to the stress of transport and corralling^12,33. Bulls from these lines also exhibit distinct behavior during the bullfight²⁷.

Aggressive behavior during unloading and the first veterinary examination is common in domestic animals, and even more so in those raised in extensive systems. In the Lidia breed, the lack of space in the corrals generates greater aggressiveness between individuals³⁴. The present results showed a positive correlation between aggressiveness in the pre-bullfight stages and the ethological parameter “Attacks shield” during the bullfight. However, it exhibited a negative correlation with the parameters “Distance to banderillero”, “Follows banderillero” and “Gallops” in the second tercio (banderillas), and with “Distance of charge at muleta” in the third tercio (muleta). A bull’s aggressiveness in the corrals was apparently linked to its entrance into the ring and goring of the shields, which, in principle, would result from breed line and bravura. But the multiple negative correlations suggest that the most aggressive individuals during pre-bullfight handling did not perform well in the bullfight; indeed, they tended to be meeker as shown in their lack of mobility and zeal during both the second and third tercios.

As suggested in a previous study, bulls exhibiting greater aggressiveness during pre-bullfight handling may be experiencing greater stress, preventing them from resting adequately and resulting in their being less aggressive during the bullfight³⁵. Perhaps the animals included in the present study that were more aggressive in the corrals were more stressed. They expended energy during pre-bullfight handling and were therefore more tired during the bullfight itself. This is supported by the negative correlation between nervousness in the corrals and lower “Determination” values during the third tercio. Animals exhibiting nervousness and aggressiveness in the corrals may be badly adapted to handling, undermining their physical condition and causing them to perform below expectations during the bullfight. Considering this, a bull’s pre-bullfight behavior may affect its physical performance, pushing any correlation between ethologies into the background.

Respiratory frequency showed no correlation with the analyzed ethological parameters, although its relationship to animal physical condition during the bullfight, indicated by time in motion in the ring^35-36, is worth further study.

Conclusions and implications

No previous data, be it published or personal testimony from breeders, is available on bull behavior in the field or during pre-bullfight handling^37-38. The present results suggest some correlation between certain ethological patterns during pre-bullfight handling and behavior during the bullfight. Broader studies including analyses of bull behavior in the field and prior to the bullfight could continue to improve predictability of behavior in the ring.