The author has declared that no competing interests exist.
Conceived and designed the experiments: AK. Performed the experiments: AK. Analyzed the data: AK. Contributed reagents/materials/analysis tools: AK. Wrote the manuscript: AK.
String-pulling and patterned-string tasks are often used to analyse perceptual and cognitive abilities in animals. In addition, the paradigm can be used to test the interrelation between visual-spatial and motor performance. Two Australian parrot species, the galah (
The perceptual and cognitive abilities necessary for everyday problem-solving such as foraging vary depending on the ecological niche of a species. For example, estimating distances and spatial relationships between oneself and objects, or between several objects in the environment, requires visual-spatial abilities and is a prerequisite for tracing causal relations among objects. Comparing perceptual and cognitive abilities among species requires a paradigm that allows a broad comparison across species and is easy for a subject to understand and handle [
A patterned-string task in which the subject must choose between two or more strings, only one of which is connected to a reward, requires both perceptual and cognitive abilities as the subject has to determine the difference in the strings and understand which would lead to the reward. The ability to solve patterned string tasks has been tested in numerous mammals [
Several authors have suggested that fine motor skills play an important role in the ability of a species to solve a string-pulling task [
A neural basis for an interrelation between visual-spatial and motor skills may be manifested in brain structures. For example, the cerebellum is not, as traditionally assumed, only responsible for motor coordination and motor control, but is also involved in a wide range of processes [
Patterned-string tasks have been used to assess a variety of capabilities in animals, but a link to motor-skills has not been tested. The single string task is an appropriate method to test the motor-skills of birds in particular, as the set-up requires complex string manipulations and, presumably, fine beak-foot coordination. Thus, enhanced manipulative skills may facilitate the handling of a vertical string. Patterned-string problems are commonly used to examine the visual-spatial aspects of string pulling [
Here, I examine the string-pulling performance of two Australian parrot species which show differences in feeding technique and hence in fine beak-foot coordination, the galah (
In the motor task (T1), individuals of both species performed very well and pulled the rewarded string spontaneously. All but one galah and one cockatiel pulled the string on their first attempt. Although the relative length of the string was the same for each species (twice as long as the body size), the galahs needed significantly longer to pull it (GLMM, factor “species”: Chi2 = 14.189,
The time (a) needed to pull the rewarded string and the relative efficiency (b) of string-pulling shown in the motor task and in the patterned-string tasks. The circles represent the mean values and the whiskers represent the standard errors. The stars indicate the tasks where the differences between the species were significant; * P<0.05, ** P<0.01, *** P<0.001. Relative efficiency was calculated by the formula: (frequency of effective actions – frequency of ineffective actions)/total number of actions.
Although the task appears to lend itself to straightforward solutions, considerable variation in techniques and in the frequency with which they were used were displayed and appeared both across and within species. Generally, the galahs manipulated the string with the foot rather than only stepping on it to fix it to the perch, whereas the cockatiels used the foot only to step onto the looped string (
Three samples for strategies used by birds to obtain the reward (A – upright pulling, occurred only in galahs; B – looping, occurred in galahs and cockatiels, C – sliding, occurred only in cockatiels) .
The galahs scored significantly higher than the cockatiels in patterned string tasks (T2-T7): they had a higher number of successfully solved patterned-string tasks (GLMM with individual as random factor, Chi2 = 5.341
The success criterion means choosing the correct string in the first trial and in at least 8 out of 10 trials in total. The values for the preference task (T5) show the preference for a shorter string when presented with two rewarded strings of different length. The stars indicate the tasks where the differences between the species were significant; * P<0.05.
Subject |
Task |
VSA score | ||||||
---|---|---|---|---|---|---|---|---|
T2 | T3 | T4 | T5 | T6 | T7 | |||
Galah | G1m | 9 | 5 | |||||
G2f | 5 | 5 | 4 | |||||
G3f | 7 | 8 | 4 | |||||
G4m | 6 | 6 | 5 | 4 | 2 | |||
G5f | 8 | 6 | 6 | 4 | 2 | |||
G6m | 5 | 5 | ||||||
Cockatiel | C1m | 6 | 6 | 8 | 5 | 2 | ||
C2m | 9 | 8 | 6 | 6 | 2 | |||
C3m | 4 | 6 | 5 | 5 | 4 | 1 | ||
C4m | 5 | 3 | 6 | 4 | 3 | 1 | ||
C5m | 6 | 6 | 8 | 5 | 5 | 1 | ||
C6f | 5 | 3 | 4 | 2 | 2 | |||
C7f | 6 | 4 | 4 | 5 | 4 | 1 | 0 | |
C8f | 6 | 6 | 5 | 8 | 5 | 2 | 0 | |
C9m | 5 | 4 | 8 | 6 | 2 | |||
C10m | 5 | 3 | 3 | 3 |
Numbers represent the number of correctly solved trials (out of 10 in total) per task; the numbers in bold show that the criterion (at least 8 correct choices out of 10 trials) was met; VSA score refers to the number of patterned-string tasks (T2-T7) in which the subject met the criterion; m=male, f=female.
At the individual level, preferences for different solving techniques were found. Several individuals switched techniques between trials, but no consistent pattern was detectable. In both species, a correlation between motor skills in terms of the extent of beak-foot coordination and the overall performance in patterned-string tasks (number of meeting the criterion) was found (Spearman rank correlation;
Spearman rank correlation using the score for beak-foot-coordination (BFC score), measured in terms of the extent of foot usage in the technique preferred when solving the novel motor task (T1), and the score for visual-spatial abilities (VSA score) which reflects the number of correctly solved patterned-string tasks (T2, T3, T4, T5, T6, and T7). Filled circles represent the data for the galahs and the stars represent the cockatiel data; the dashed line represent regression line (
Contrary to expectation, the cockatiels managed the vertical string-pulling problem despite the fact that they do not naturally perform pronounced beak-foot coordination activities. Most individuals of both species pulled the single baited string spontaneously. Hence, I found no support for the hypothesis that the use of feet and beak in the feeding context indicates the presence of the manipulative skills needed for successful string-pulling. However, members of the two species used different sets of techniques to pull the string and the techniques clearly differed in terms of the extent of beak-foot coordination shown. Whereas the galahs manipulated the string with their feet (grabbed it to pull it through the beak in alternate sequences), the cockatiels either just lifted the string by drawing it up through their beak without using their feet at all or they used a foot but only to step on the looped string and to fix it to the perch. Therefore, the performance of the cockatiels showed that using the feet while feeding is not necessary to solve the vertical string-pulling problem, but that it may well determine how the task is solved.
A larger difference between the species was found in the tasks that tested visual-spatial skills (T2, T3, T4, T6, and T7). Measured in terms of how many trials were solved correctly and of how many individuals were successful at a specific task, the galahs outperformed the cockatiels. Three galahs and one cockatiel were able to visually determine a physical connection between objects (T6), whereas only three galahs (and no cockatiel) were able to distinguish between two crossed strings of the same colour (T4). Most of the subjects in both species preferred the shorter rewarded string (T5), thus choosing the most efficient solution. However, only galahs were able to suppress this preference when the shorter string was not connected to a reward (T7), thus showing some kind of understanding that the string must be connected to the reward to work properly. Note that both species pulled the string in patterned string tasks, each using techniques reflecting the extent of their motor skills – in this way both species had the skills to solve the patterned testing problem – but the capacity to recognize the spatial relation between string and reward appeared to differ between them. Of course, birds can fail for other reasons, such as motivation [
Clearly, two species are not sufficient to draw conclusions about an interrelation between motor and specific visual-spatial skills needed to recognize spatial relations between objects in parrots. However, as this is the first attempt to investigate the possibility of such an interrelation, the patterns are encouraging and call for more research. Support can be found in a within-species comparison, as fine beak-foot coordination and performance in patterned-string tasks were positively correlated. The higher the score for motor abilities (defined as a preference for techniques that require finer beak-foot coordination) was, the more successful were the performances achieved by the subject. Furthermore, the published literature on vertical patterned string problems in parrots shows that all species tested so far (keas; [
Contrary to the findings of Magat and Brown [
The patterned string task has been suggested to provide a reasonable simulation of natural foraging situations encountered by frugivore species [
A potential limitation of the present study may be that when testing subjects jointly, different social learning speeds may have influenced the group performance as a whole as well as individual performance. If the birds had used social learning, a sequential pattern of similarity in the techniques applied could be expected. Yet, I found no such pattern: the birds that followed after the first one used different techniques, suggesting that imitation did not play a role. However, types of social learning such as local enhancement, object enhancement or social facilitation – that only guide the attention of the observer to a location or item but still require individual learning by trial and error – likely played a motivational role in galahs. All six birds in the group showed an interest in the tasks, while only 10 of 22 cockatiels participated in the tests even though all subjects were able to observe the successful individuals, and even though monopolization of the set-up was prevented by the presence of several apparatuses. Finally, birds that initially failed a task did not improve their performance in the following trials even though they clearly observed successful companions. Therefore, social learning did not appear to influence the birds’ success rate.
Further studies using standardized paradigms to test visual-spatial and motor skills across a wide range of parrot species are necessary to show whether the patterns found at the individual level are consistent across a wider range of species and to support the possibility that specific abilities such as visual-spatial skills may interrelate with motor skills not only in mammals but also in birds.
Six galahs and ten cockatiels were tested. All parrots were hatched in a zoo and were raised by their parents. No artificial toys were available, but green branches were provided regularly for playing and nibbling.
The galahs were kept in a walk-through outdoor aviary (12 x 7 x 5 m) with an adjacent indoor aviary (6 x 1.6 x 2.5 m) at Tierpark Gettorf, Germany. The group contained five adults and one subadult (3 males, 3 females), which were not related to each other. All individuals participated in the study. Birds were fed every day between 9 a.m. and 11 a.m. with parrot pellets and fruits. The indoor aviary was lit by several windows and provided with several perches and a nestbox. The outdoor aviary contained several trees and a trunk. Water was available
The cockatiels were kept together with budgerigars (
The galah and cockatiel aviaries were close to each other, so that the birds were not acoustically isolated, but a barrier prevented any visual cues from one group to the other during the experiment in the outdoor aviary. All subjects of each species could be individually recognized at all times during the experimental sessions.
No subject had contact with string-like objects or had been trained in any object-pulling task prior to the present experiments. The animal care during the study was performed by the regular zoo keepers. The daily feeding conditions were adapted to the testing situation. The experiments reported were integrated into the daily routine as part of the regular animal welfare activities. After the study, all tested birds remained in their respective flocks.
The birds were given their regular variety of seeds on test days, but they were deprived of their preferred fruits and vegetables on those days. Water was available
The subjects were tested jointly in their respective groups to simulate conditions in which subjects deal with a novel problem (e.g. new food sources) in the natural environment, where usually a set of individuals is faced with a new situation at the same time.
Motor task (T1): This task tested the parrots’ ability to pull up a reward suspended from a horizontal perch by a single string and examined the techniques used to obtain this reward.
Parallel strings (T2): To test if string-pulling behaviour is food-directed, two strings, one with the reward attached as before and one without, were simultaneously presented to the birds. Pulling up the string with the reward more frequently than expected by chance would indicate that the subject could recognize the string as a means to obtain the reward even if string-pulling behaviour in T1 had been self-rewarding.
Crossed strings – a (T3): To assess whether the parrots’ choice was based on the spatial or the functional relationship between string and reward, I crossed the strings. If their choice was based on the functional connection between food and string, they would pull the baited string. If the choice was based on the spatial relationship only, they would pull the string directly above the bait, as in the earlier trials. In T3, two differently coloured strings (green/white or green/yellow or white/red) were used to allow the birds to trace the strings from one end to the other more easily, assuming that both species have similar colour sensitivity [
Crossed strings – b (T4): This was in principle the same test as in T3, but with two crossed strings of the same colour (white/white or red/red or green/green), and thus expected to be more difficult for the birds to discern.
Preference task (T5): To test if the subjects show a preference for the shorter string (with a reward which could be obtained more easily), two rewarded strings of different lengths were presented.
Broken strings – equal (T6): To test the ability to visually determine whether or not objects are physically connected, two strings of equal length were presented to the subjects. While one string was connected to a reward, the other one had a gap between string and reward. Both rewards were placed on a small platform on a wire (attached to the lateral walls or poles). The distance between the string and the unconnected reward was 5 cm.
Broken strings – different (T7): To test if the birds realize that the string must be connected to the reward in order to work properly, I presented two strings of different length as in T5, but the shorter string was disconnected from the reward. To succeed the birds would abandon any preference for the short string, and chose the longer, rewarded string instead.
The position of the rewarded string in choice tasks was determined randomly across the sessions by tossing a coin.
For each species, I calculated the proportion of birds that met the criterion of choosing correctly in the first trial and in at least 8 out of 10 trials in total. For the analysis of quantitative differences between species, I performed a generalised linear mixed model (GLMM) analysis using lmers (package ‘lme4’, [
Each individual received a score for its relative efficiency in solving the task by comparing frequencies of effective reactions, namely “pulls”, and ineffective reactions, namely “drops”. The score was calculated using the formula: (frequency of effective actions – frequency of ineffective actions) / total number of actions (see also [
To quantify the extent of beak-foot-coordination (BFC), a score was calculated quantifying the extent of foot usage in the technique preferred when solving the novel motor task (T1). A score of 0.5 was assigned to pulling the string with the beak and using the foot just to fix the string on the perch: this was considered moderate coordination. Conversely, pulling the string first with the bill and then using the foot to pull the rest of the string while holding it in the bill and repeating the foot movements (i.e. to the bill to hold the string in the foot and away from the bill with the string in the foot to gain more string, repeating this action up to seven times) was considered highly coordinated and scored as 1.0 (being in general the same movement as the touching of the nose used as a part of the LOS test measuring fine motor skills in children [
Technique | Definition | BFC score |
---|---|---|
Sliding | pulling up the string through the bill without fixing or holding it with the foot | 0 |
Flip | reaching down and flipping the string to the other side of the perch | 0 |
Looping | reaching down, pulling up string with the beak, placing the foot on the string, letting go of the string with the beak, remaining in place, reaching down again | 0.5 |
Side walking | reaching down, pulling up the string with the beak, walking to the side of the perch, placing the foot on the string, and reaching down again | 0.5 |
Turn | turning the whole body 180° while holding the string and stepping on the additional string with the feet | 0.5 |
Upright pull | pulling up the string till the body is in a completely upright position, holding with the beak, and gaining more string by grabbing it with the foot | 1 |
BFC score is the beak-foot coordination coefficient.
To quantify the performance in patterned-string tasks, a score for visual-spatial abilities (VSA) was calculated using the number of patterned-string tasks (T2, T3, T4, T5, T6, and T7) in which the subject met the success criterion – i.e. reaching the reward in at least 8 out of 10 trials.
Finally, Spearman rank correlations were computed between motor and visual-spatial skills using both scores, to assess the interaction between motor and visual-spatial skills within both species.
All data collection was carried out in accordance with the guidelines of the University of Hamburg and with permission of the Tierpark Gettorf, Germany. The present study was strictly non-invasive and based on behavioural observations; all reported experiments were classified as non-animal experiments and required no approval from the relevant body in accordance with the German Animal Welfare Act (Federal Law Gazette I, p. 1094, Section V, Article 7).
(DOCX)
I thank Jutta Schneider, Wiebke Schuett, and the members of the Division of Animal Behaviour at the University of Hamburg for fruitful discussion, two anonymous referees for valuable comments on an earlier draft of this manuscript, and Zoe Demery for improving the English. I also thank Gabriele Ismer from the Tierpark Gettorf, Germany, for her constant support and enthusiasm and for permitting me to conduct research on the animals under her care.