Flexible tool set transport in Goffin’s cockatoos

The use of tool sets constitutes one of the most elaborate examples of animal technology, and reports of it in nature are limited to chimpanzees and Goffin’s cockatoos. Although tool set use in Goffin’s was only recently discovered, we know that chimpanzees flexibly transport tool sets, depending on their need. Flexible tool set transport can be considered full evidence for identification of a genuine tool set, as the selection of the second tool is not just a response to the outcomes of the use of the first tool but implies recognizing the need for both tools before using any of them (thus, categorizing both tools together as a tool set). In three controlled experiments, we tested captive Goffin’s in tasks inspired by the termite fishing of Goualougo Triangle’s chimpanzees. Thereby, we show that some Goffin’s can innovate the use and flexibly use and transport a new tool set for immediate future use; therefore, their sequential tool use is more than the sum of its parts.

We tested if birds would transport the tools as a set: therein, we were also interested in whether the probability of transporting both tools would increase relative to the presence of the box with a membrane. Transporting the tools individually when a tool set is needed (depending on the box type) requires a higher energy investment. The highest investment would be expected in the third phase (requiring a vertical flight).Likewise, a wrong decision is also costly when transporting both tools and only one is needed. Therefore, if their recognition of the problem includes the use of a tool set, we expect them to eventually transport both tools at the same time to the apparatus, and we expected them to transport the tools together more often when both tools are required to solve the task than when one tool is sufficient. Throughout this series of experiments, we revealed the ability of Goffin’s to innovate the use of a tool set, as well as to use and transport it in a flexible way, thus suggesting the ability to categorize both tools as a tool set.

The third and final experiment was similar to the second (either the two-tool or the single-tool apparatus presented in semi-randomized order), but reaching the box with the tools required additional movement. After two initial phases, one requiring climbing and the other requiring a horizontal flight, they faced a vertical flight phase to reach the box ( Figure 2 ). The reason for the incremental effort increase from the walking (phase 1) to the horizontal and finally the vertical flight (phases 2 and 3)was to be able to identify (to some extent) the minimal investment required for the birds to switch from transporting the tools individually to transporting two tools at the same time. We decided against counterbalancing the phases to avoid birds starting with the highest investment phase already having fallen into a habit of transporting both tools when later tested in the low investment phase.

In the second experiment, cockatoos faced two alternative apparatuses in a randomized sequence, one requiring the use of both tools ( Figure 1 A) and the other requiring only the use of the long, flexible probe ( Figure 1 B). Our goal was to test their flexibility in selecting the correct tool(s) from the tool set for each task.

In the first experiment (fishing cashews), the cockatoos were exposed to a previously unknown tool set problem requiring the use of two different tools, one short and rigid (to tear open a membrane) and one long and flexible (to reach a reward behind that membrane—notably, it is too flexible to perforate the membrane directly; Figures 1 A and 1D ). Birds had to innovate and constantly master the use of the required tool set to proceed to further steps. We studied their learning progress in selecting the correct tool (the short one) for the first insertion, as well as their movements before inserting the tool for the first time.

Inspired by the kinds of tool sets that Congo Basin chimpanzees use and transport for termite fishing, we designed a series of experiments to test the flexible use and transport of a tool set in Goffin’s under controlled circumstances.

In the aforementioned tool set, Goffin’s were holding the fruit stone (goal) in their claw while crafting and using each tool, and only a single tool could be held at a time.This raises the historical doubts stated above regarding the planification of the use and categorization of both tools as a tool set: we do not know whether Goffin’s have the capacity to identify a tool set or if they build and use specific tools for individually perceived sub-steps of the problem.

Goffin's cockatoos, unlike other tool-using birds, are neither specialized on nor morphologically adapted to tool use.Like in primates, their innovative capacity thus seems to largely depend on domain general cognitionand on knowledge actively acquired through exploration and playful object combinations.

Goffin's cockatoos are extremely opportunistic extractive foragers from a small archipelago in the Molucca region of Indonesia. Captive studies show that they are highly capable of innovating solutions to physical problems including toolsat the level of specialized tool users such as New Caledonian crows or the great apes, innovating complex associative tool use such as composite tools.We recently also learned that they can manufacture and use a complex set of up to three different tools specifically crafted out of wood for three different functions (a sturdy tool for wedging, a slim tool for cutting, and a long, broad tool for spooning). The tools are flexibly used in sequence to access the seed content of a local fruit stone.

After years of research in the southern region of Noubalé-Ndoki National Park (Goualougo Triangle), Crickette Sanz and David Morgan, together with Richard Byrne himself, rejected the previous argument after the former observed that the chimpanzees not only transported both tools at the same time but also did so in a flexible manner, according to present need: chimpanzees carried only the fishing probe or both the fishing probe and the perforating tool depending on the conditions of the termite mounds, omitting redundant and unnecessary steps.

This finding sparked a decade-long debate on action planning abilities involved in chimpanzee termite fishing. Importantly, in 2004, Richard Byrne acknowledged that the N’doki chimpanzees may not identify the use of a tool set as a solution to a single problem but may instead perceive each tool as an independent solution to a different, unrelated problem (a termite mount being blocked; entering an open mount through slim channels).

In 1995, Suzuki et al. found that chimpanzees in the N’doki forest (northern Congo) used a set of tools to fish for termites—a cultural innovation that facilitates access to the termite nest, subterranean or epigeal.The tools used by these Congo chimpanzees have at least two complementary characteristics and functions: a perforating stick (short and rigid) used to pound a hole in the termite mound, and a fishing probe (long and flexible), to access the deeps of the mound and extract aggressive soldier termites.

Chimpanzees stand out for their use of tools in terms of variety, diversity, and cognitive-cultural dependence on tool use.They have demonstrated a great capacity to innovate solutions to physical problems through the use of tools, which differ between geographical regions, resulting in a remarkable cultural richness.

A particularly remarkable form of associative tool use is the use of two or more different kinds of tools of different functions on the same goal, traditionally referred to as a tool set.Only two non-human species have been described to use tool sets in the wild beyond the anecdotal,chimpanzees and, as we only very recently learned, Goffin’s cockatoos.

The most sophisticated types of animal tool innovations recorded to date are those that involve more than one tool to achieve a single goal (associative tool use).Within associative tool use, complexity arises for a variety of reasons, such as different tools having complementary functions, each tool requiring different movement patterns, a higher total number of spatial relationships to consider, or even a need for sophisticated action planning.

Tool innovations are a prime repository for the evolution of technology across species. Nevertheless, although rare overall, tool innovations can still occur along a considerable scale of relational and sequential action complexity across animals, with the vast majority of examples being simple, primary, or even self-directed object interactions.

Pipin continued to carry both tools every time he encountered the tool set box (22/22) and also transported both tools occasionally to the single box, but on half as many occasions (11/22). Kiwi, having transported for the first time in the last trial of the previous phase, from session five of this phase onward, transported both tools on all occasions when confronted with the tool set box (17/17) and only very rarely when confronted with the single-tool box (5/16). Figaro transported both tools together in session 8 (facing the tool set box), and from session 9, he transported whenever he encountered the tool set box (9/9 tool set transportation for session 10 onward but also in 7 out of 9 trials in the single-tool condition). Fini transported both tools together on only one occasion during the fourth session when she was facing the tool set box.

All of the flying cockatoos used various techniques to minimize effort, both on the way up and on the way down, which indicates that flying was costly.

During the third and last phase of the experiment (vertical flight condition for 3 birds—even greater tool distance for Pipin; see STAR Methods for details) all 4 individuals transported both tools at once in flight eventually.

Pipin continued transporting the tool set, but now he always carried both tools in the tool set condition (5/5 times) and less often when the single-tool box was present (4/7).

During the second phase of the experiment, the cockatoos had to fly horizontally to reach the platform (Pipin still relied on the ladder—for him, the tools were placed farther apart to increase his energetic investment; see STAR Methods for details).

Notably, during this and the subsequent phases, Pipin readjusted his tool choices several times before and during the climbing ( Video S2 ).

During this first phase of the experiment, one individual (Pipin, an adult male) began carrying both tools in both conditions. As other individuals would do in later phases, he inserted the short tool into the long one (halved straw), transporting them as a compound object ( Figure S4 Video S3 ). Importantly, Pipin was missing some primary feathers and had to actively climb each step of the ladder. No other cockatoos transported both tools simultaneously during this phase.

During the first phase (climbing a ladder) of the experiment, all individuals started transporting the tools individually, with a single wing flap jumping from the table to the perch on the platform.

All three individuals transported both tools significantly above chance expectation when faced with the tool set box. By contrast, when faced with the single box, only Kiwi differed from chance expectation by transporting the single tool more often, whereas the other two individuals did not differ from chance level ( Table 3 ).

To evaluate if individuals transport the appropriate tools to the corresponding box, i.e., tool set to box “d” and single tool to box “s.”

In this last experiment, once again, the 5 individuals who completed the first experiment participated. There were 4 of them that eventually transported both tools together, and 3 of them did it recurrently (Figaro, Pipin, and Kiwi).

Overall, we found a significant effect of the test predictors on the probability of using the correct first tool (full-null model comparison: χ= 16.477, df = 8, p = 0.036). However, the three-way interaction between box type, switching, and trial number appeared non-significant ( Table S3 ). After the removal of this and other non-significant interactions, we found clear significant effects of trial number and box type. We found a significant improvement in the probability of choosing the correct tool with increasing trial number ( Figure S2 Table 2 ) and a difference between box types, with the probability of choosing the correct tool being higher when confronted with the single-tool box ( Figure S3 Table 2 ). Neither time nor switching tools significantly affected the probability of choosing the correct tool.

Estimates, together with standard errors, confidence intervals, test results, and minimum and maximum of model estimates obtained after dropping levels of random effects one at a time. See also Figures S2 and S3

Trial number and time were z-transformed to a mean of zero and a standard deviation of one; mean (SD) of trial number was 15.5 (8.7); mean (SD) of time was 53.0 (73.9)

In this experiment in which the box requiring a tool set was semi-randomly alternated with a box requiring only the long tool (single-tool box), subjects performed 128 correct insertions of the first tool and 22 incorrect ones of which only 4 occurred with the single-tool box.

Those 5 individuals who solved the first experiment, and were given the full 15 sessions, had the opportunity to participate in the second experiment (Dolittle had dropped out due to a lack of motivation).

We also analyzed what influenced switching behavior. The full-null model comparison revealed a significant influence of the test predictors (χ= 6.781, df = 2, p = 0.0336) and found that the time required to solve the task did not have an effect on switching events (GLMM: β = −0.029, SE = 0.243, χ= 0.015, p = 0.902); however, significantly more switches occurred in later trials (GLMM: β = 0.943, SE = 0.302, χ= 5.862, p = 0.0155; Figure S1 Table S2 ).

When analyzing which variables (trial, switching, time until solving the task) were associated with the probability to choose the correct tool, we found an overall significant effect of the fixed effects predictors when testing the reduced model against the null model indicated (likelihood ratio test comparing full and null model: χ= 9.617, df = 4, p = 0.047). However, the two-interaction between the trial number and switching was not significant ( Table S1 ). After removing this, we found that the time required to complete a trial was not associated with the probability of choosing the “correct” tool, but individuals tended to improve performance over trials ( Figure 3 ), and the probability of choosing correctly increased if a switch of tools occurred before employing the tool on the task ( Figure 4 Table 1 ).

Trial number and time were z-transformed to a mean of zero and a standard deviation of one; mean (SD) of trial number was 22.0 (12.5); mean (SD) of time was 100.2 (91.5)

Violin plots indicate the distribution of data and the size of gray dots indicate the number of individuals switching; asterisks indicate fitted model means and error bars represent 95% CI based on 1,000 bootstraps.

Solid line indicates the fitted model, and gray areas represent 95% confidence intervals derived by 1,000 bootstraps; circle sizes indicate the number of individuals choosing correctly/incorrectly for each trial.

Although with differences between individuals, all solvers started showing switching behaviors between the two tools, grabbing and releasing and alternating them multiple times before the first insertion.

After solving their first session, most birds did not fail a single trial again throughout the whole experiment (only one bird did, Dolittle).

Figaro and Fini stand out among the 6 solvers: both have repeatably demonstrated innovative competence in a tool-using task in previous studies.Both solved this task on their first trial, reaching criterion (3 consecutive successful sessions) in the earliest possible time, and never failed to obtain the reward. On his first trial, Figaro needed only 31 s to explore the box, test both tools, and find the solution ( Video S1 ). Fini needed 34 s.

The individuals who solved the test showed notable differences in their ripping technique after the membrane had been punctured. Accessing the reward behind the membrane using a flexible tool requires some skill in tearing the membrane in a way that the nut can be reached in a straight line through the window. For most individuals, the tearing ultimately traced horizontal or diagonal lines. Through these horizontal or diagonal holes, the second tool was introduced and, by means of a sweeping movement, the nut was hit.

Out of 10 cockatoos, 7 innovated the use of a tool set eventually and 6 of them reached the proposed criterion of 9 consecutive successful trials. There were 2 cockatoos, the adult male Figaro and the adult female Fini, that solved the task on their first trial with very short trial times (see below). There were 2 more cockatoos that were able to solve it in their second session and 2 more in their third. There was 1 cockatoo, the adult male Muki, that solved it later in the experiment but only on a single trial.

Discussion

15 Byrne R.W.

Sanz C.M.

Morgan D.B. Chimpanzees plan their tool use. In our first and second experiments, we provide the first controlled evidence that the majority of Goffin’s spontaneously innovated tool set use under controlled experimental conditions, without social facilitation, and learned to apply it flexibly according to need. Furthermore, our third experiment suggests that the tool set is more than just the use of tools in sequence (as historically suggested for chimpanzees before the flexibility of their tool set transport was observed; see introduction): four Goffins were observed to transport two tools simultaneously, and two Goffin’s were able to not only transport their tool set together but even showed some flexibility depending on the task requirements. This suggests that, like in chimpanzees,two tools may be categorized as a tool set.

1 Shumaker R.W.

Walkup K.R.

Beck B.B. Animal Tool Behavior: The Use and Manufacture of Tools by Animals. 1 Shumaker R.W.

Walkup K.R.

Beck B.B. Animal Tool Behavior: The Use and Manufacture of Tools by Animals. , 26 Davidson I.

McGrew W.C. Stone tools and the uniqueness of human culture. 6 O’Hara M.

Mioduszewska B.

Mundry R.

Yohanna

H. T.

Rachmatika R.

Prawiradilaga D.M.

Huber L.

Auersperg A.M.I. Wild Goffin’s cockatoos flexibly manufacture and use tool sets. 27 Toth N.

Schick K.D.

Savage-Rumbaugh E.S.

Sevcik R.A.

Rumbaugh D.M. Pan the Tool-Maker: investigations into the stone tool-making and tool-using capabilities of a bonobo (Pan paniscus). , 28 Motes-Rodrigo A.

McPherron S.P.

Archer W.

Hernandez-Aguilar R.A.

Tennie C. Experimental investigation of orangutans’ lithic percussive and sharp stone tool behaviours. 29 Westergaard G.C.

Suomi S.J. A simple stone-tool technology in monkeys. 12 Sanz C.M.

Morgan D.B. Chimpanzee tool technology in the Goualougo Triangle, Republic of Congo. In the first experiment, subjects innovated the puncturing and tearing of a membrane, a yet unreported mode of avian tool use.There are no known cases of cutting tools in either wild primates or birdsbeyond perhaps the second tool in the Goffin’s original tool set study on Tanimbar.However, in captivity, cases of cutting tool use have been documented in great apesand in capuchin monkeys.Sticks with puncturing function are indeed the first commonly used tool type in the Goualougo Triangle (Congo) to open epigeal or subterranean termite mounds before applying another tool within a tool set.

30 Auersperg A.M.I.

Köck C.

O’Hara M.

Huber L. Tool making cockatoos adjust the lengths but not the widths of their tools to function. The way certain subjects initially solved the task, dropping the short tool right after having used it to tear the membrane, is reminiscent of previous experimental results where cockatoos would drop a tool that was too short upon sight of the apparatus without using it and would manufacture a longer one.

14 Byrne R.W. The manual skills and cognition that lie behind hominid tool use. The way in which they solved these first trials does not yet require the categorization of both tools as a tool set. As Byrne suggested following observations of tool set use in chimpanzees, they could be using several tools with complementary functions in sequence in which the use of the second tool of a tool set could only be the result of the outcome of the use of the previous tool.For example, during Figaro’s first interaction with the box ( Video S1 ), the use of the first tool is not immediately succeeded by the use of the second tool: it is only after walking around and exploring the box and after picking up and discarding the first tool again that the second tool is picked up and used.

The birds had to innovate the destruction of the membrane that required first a stout piercing push followed by a forceful wagging type of movement with the short stout tool. The destruction of the membrane would then allow for the full sight of the food reward and thereby possibly trigger the switch to the probing tool. Although they had a pre-experience session in which they learned that the membrane can be destroyed without the use of tools, it is reasonable that the tearing of the membrane was facilitated by the extensive experience in using sticks as tools of these subjects.

16 Auersperg A.M.I.

Szabo B.

von Bayern A.M.P.

Kacelnik A. Spontaneous innovation in tool manufacture and use in a Goffin’s cockatoo. , 17 Auersperg A.M.I.

von Bayern A.M.I.

Weber S.

Szabadvari A.

Bugnyar T.

Kacelnik A. Social transmission of tool use and tool manufacture in Goffin cockatoos (Cacatua goffini). 17 Auersperg A.M.I.

von Bayern A.M.I.

Weber S.

Szabadvari A.

Bugnyar T.

Kacelnik A. Social transmission of tool use and tool manufacture in Goffin cockatoos (Cacatua goffini). , 20 Osuna-Mascaró A.J.

Mundry R.

Tebbich S.

Beck S.R.

Auersperg A.M.I. Innovative composite tool use by Goffin’s cockatoos (Cacatua goffiniana). Figaro stood out: unlike the rest, he ripped through the membrane vertically, and throughout the experiment, he perfected his technique until he barely needed a mere closely centered pinhole. Through this small hole, he directed the longer tool with extreme precision toward the nut ( Video S1 ). Figaro is the most experienced tool user in the sample,using tools regularly outside of experiments as a means to explore and play (personal observation). His efficiency and effectiveness are probably the result of two non-exclusive factors, his particular technique for inserting the sticksand a great deal of practice in different contexts.

Over the 15 sessions of experiment 1, the birds showed learning, gradually improving on the choice of the correct tool order to solve the tool set box ( Figure 3 ). In the process, we observed a lot of switching behavior between the two tools ( Figures 4 and S1 ). It is likely that picking the short tool initially required some level of impulse control as only the long tool had a direct reward association ( Video S2 ).

Switching behavior did not disappear throughout experiment 1 but increased. Interestingly, the probability of correctly choosing the tool was higher when the birds showed switching behavior prior to insertion ( Figure 4 ). This may have several non-exclusive reasons, such as birds that use more haptic exploration being able to improve their performance. Additionally, a bird may impulsively pick up the food-associated tool (long tool) and drop it to pick up the short, stout tool upon seeing the intact membrane.

31 Lambert P.J.

Stiegler A.

Rössler T.

Lambert M.L.

Auersperg A.M.I. Goffin’s cockatoos discriminate objects based on weight alone. 20 Osuna-Mascaró A.J.

Mundry R.

Tebbich S.

Beck S.R.

Auersperg A.M.I. Innovative composite tool use by Goffin’s cockatoos (Cacatua goffiniana). 32 Allritz M.

McEwen E.S.

Call J. Chimpanzees (Pan troglodytes) show subtle signs of uncertainty when choices are more difficult. , 33 Smith T.R.

Parrish A.E.

Creamer C.

Rossettie M.

Beran M.J. Capuchin monkeys (sometimes) go when they know: confidence movements in Sapajus apella. Switching behavior toward the correct option has been observed in previous experiments in which the Goffin’s cockatoos had to choose between identical objects of different weights,as well as when selecting among different tools to solve a problem.Recently, movement during decision-making has been used to study metacognition through uncertainty and confidence in chimpanzees and capuchin monkeys, respectively.Switching behavior in these birds could thus serve as a future means of studying their metacognition.

In the second experiment, cockatoos were confronted with both single-tool and tool set tool boxes in a semi-randomized fashion. This part of the study was designed to enable the animals to differentiate the operational properties of our two experimental apparatuses to thereby set a certain level of pre-experience for the tool transport phase.

The five individuals who proceeded into this experiment chose the correct tool above chance expectations from the start ( Figure S2 ). Nevertheless, they still improved as the experiment progressed.

15 Byrne R.W.

Sanz C.M.

Morgan D.B. Chimpanzees plan their tool use. 14 Byrne R.W. The manual skills and cognition that lie behind hominid tool use. The aim of the third experimental design was to test whether Goffin’s cockatoos would categorize both tools as a tool setor whether the use of a second tool would only be the outcome of the use of the previous tool.

15 Byrne R.W.

Sanz C.M.

Morgan D.B. Chimpanzees plan their tool use. Goualougo Triangle chimpanzees transport a single tool or a tool set (together) to catch termites in a somewhat flexible way depending on the circumstances.As mentioned earlier, it has been argued that this cannot be explained by two isolated tool uses triggering one another.

Since our goal here was only to test for the Goffin’s identification of a tool set, we left the box fully visible to the subjects. Nevertheless, derivations of the same setup may serve in future experiments to address other questions such as the mental imagery of future situations.

Three birds continuously transported both tools together in experiment 3 (Kiwi, Figaro, and Pipin). Interestingly, they always transported both tools together as a compound object. However, we must be cautious of how we interpret this action since the combination of the two objects was strongly facilitated by their shape. Future studies will allow us to explore their ability to transfer what has been learned to other object shapes and conditions.

Once they learned to transport both tools together (when they did it more than a single time), they transported every single time that they encountered the tool set box. Less frequently, they also transported both tools to the single box. There were individual differences in the decision to transport one versus two tools when faced with the single-tool box. Figaro transported both tools to the single-tool box nearly as often as he did to the tool set box. Pipin, on the other hand, transported the tool set twice as often when encountering the tool set box than when confronted with the single-tool box. Finally, Kiwi rarely made any errors. The unnecessary tool set transports of Figaro may be due to the consolidation of an inflexible strategy and/or due to a possible trade-off between attention and the likely low energetic cost of transporting both tools together. The two possibilities are not mutually exclusive since transport flexibility can be expected to be influenced by the cost of each wrong choice (and this is dependent on each individual). Unnecessary tool set transport may result from a trade-off between the physical effort invested in transporting an extra tool and the cognitive effort required in making an informed decision on the task at hand relative to the available tools. Moreover, it is noteworthy that an incorrect decision was more costly when a single tool was transported to a box with a membrane than when two tools were transported to a box with no membrane (in the first case, it requires going and coming back with a new tool, whereas in the second case, the extra cost is that of picking up and transporting two tools together).

Throughout the phases of the third experiment, effort may have played a role in the propensity to transport both tools. The early onset of transporting both tools together by Pipin, who was flightless, may be caused by different levels of effort. The effort required to actually climb the ladder (instead of hopping over the ladder from the table to the perch with a single wing flap as other birds did) led him to a more slow-paced approach, also allowing him to make mid-way adjustments of his selected tools to transport ( Video S2 ).

The presumably higher energy cost of vertical flight may have had an effect on the first tool set transport of Figaro and Fini ( Video S3 ), as in the consolidation of Kiwi's transport at that phase. Fini, the only individual that did not continue transporting both tools after having done it once, was also the only female left in the experiment and is considerably weaker and smaller than Figaro and Kiwi.

Notably, experiment 3 may have had a particularly challenging spatial layout. The cockatoos needed to adapt their behavior to the narrow space between the perch and the box on top of the platform ( Video S4 ). This space was an important conditioning factor for two technical aspects: the insertion of the tools (especially the long one) and placing the long tool on the platform for later use.

34 Auersperg A.M.I.

Köck C.

Pledermann A.

O’Hara M.

Huber L. Safekeeping of tools in Goffin’s cockatoos, Cacatua goffiniana. Individuals who transported both tools were obliged to temporarily lay aside one of the tools to be able to use the other; letting the second tool stay on the platform was a challenge in such a limited space, and some mastered it sooner than others. Although the most common strategy here was the release of the long tool very close to the box wall, Figaro developed one in which he held on to the second tool by keeping it elevated with one foot while inserting the first with the beak ( Video S4 ). The development of different techniques supports the innovative nature of this tool-saving behavior.

17 Auersperg A.M.I.

von Bayern A.M.I.

Weber S.

Szabadvari A.

Bugnyar T.

Kacelnik A. Social transmission of tool use and tool manufacture in Goffin cockatoos (Cacatua goffini). , 20 Osuna-Mascaró A.J.

Mundry R.

Tebbich S.

Beck S.R.

Auersperg A.M.I. Innovative composite tool use by Goffin’s cockatoos (Cacatua goffiniana). 35 Rosenbaum D.A.

Marchak F.

Barnes H.J.

Vaughan J.

Slotta J.D.

Jorgensen M.J. Constraints for action selection: overhand versus underhand grips. , 36 Wunsch K.

Henning A.

Aschersleben G.

Weigelt M. A systematic review of the end-state comfort effect in normally developing children and in children with developmental disorders. As documented in previous experiments,the technique used for tool insertion was different among the Goffin’s, and this affected the way both tools were used in the confined space of the platform. For example, when Fini grasped the long tool, she adjusted it before flying so that the long end was facing her left cheek. In this way, when she reached the top of the platform, she was not hindered by the confined space to rotate the tool to her preferred position ( Video S2 ). We hope that future experiments will allow us to investigate whether the gripping behavior of tools by cockatoos includes seeking an end-state comfort effect.