Revisiting Weir et al. 2002

In a study published in Science in 2002, Alex Weir, Jackie Chappell and Alex Kacelnik demonstrated that New Caledonian crows can bend wires into the shape of hooks to access food. This study was probably the first report of an animal purposefully modifying an object to use as a tool. Fifteen years after the paper was published, I spoke to Alex Kacelnik about the observations that motivated this study, the writing of this paper, and what we have learnt since about tool use and cognition in New Caledonian crows. 

Citation: Weir, A. A., Chappell, J., & Kacelnik, A. (2002). Shaping of hooks in New Caledonian crows. Science, 297(5583), 981-981.

Date of interview: 12th August 2016 (on Skype)


Hari Sridhar: In the beginning of the paper you say this work was “inspired by the observation that a captive female spontaneously bent a piece of straight wire into a hook and successfully used it to lift a bucket containing food from a vertical pipe”. Could you tell us a little more about this incident?

Alex Kacelnik: We knew that these animals use tools extensively because of the work of Gavin Hunt and others; Russell Gray in particular. When we started to work on them our goal was to see how flexible this tool use was, and how sensitive it was to the needs of the task. In order to do that we placed two wires, one which was bent and one which was straight, on top of a vertical tube which contained a bucket in the bottom. We were trying to see if they would pick up the bent one to collect the bucket. In the experiment we had two individuals close together – a male and a female. The male was bigger and dominant over the female. What happened was the male picked the bent wire and took it away. The female, Betty, was left with only the straight wire with which she tried to retrieve the bucket but failed. What she then did was to basically jam the wire against the base of the tube and bend it. The reason we call this behaviour spontaneous is because there was no intervention from any of the experimenters. And then she used the bent wire to retrieve the bucket.

On other occasions – because she did it reliably afterwards – every time, she made the hook in different ways. Sometimes she made it by holding the wire and bending the tip. In that case when she bends the tip she has to go and grab it on the other side with the beak to use it, not to twist it. This also implies some sensitivity to the functionality of the shape of the instrument. That was the first experiment, but of course that was 15 years ago, and ever since we have done many different variations and different studies that show different capabilities.


HS: What happened to the experiment that you were doing at the time when you discovered this behavior? Was that what was published later in Animal Cognition?

AK: No. We never published that because, in a sense, it became irrelevant. All we were trying to see in that experiment was whether crows picked a straight or a hooked piece when necessary. But once we had seen that they can bend a straight one when required, then it became clear that the answer was: yes, they can do that. The new question becomes: “In what sense can it be said that they ‘understand’ the tasks they solve?”  How does their genetic inheritance articulate, with both learning by trial and error and reasoning, so as to let them achieve the solutions they find? What we published later in Animal cognition was a contribution to this. We showed that they could not just make a hook, but also the opposite, that is unbend a hook when that was required to elongate a piece of wire and reach farther. This excludes some kinds of insensitive tool use, but is only one step towards mapping the limits of their minds.


HS: Do you remember the date when you made this observation of the crow bending the wire for the first time?

AK: No. I know that it was in 2002, and I know that the first observation was made by my student Alex Weir, who became the first author of that report. Alex Weir had started his PhD six months earlier and so this was practically his first experiment. He was being co-supervised by someone called Jackie Chappell and myself. Jackie was my post-doc and we had been studying the sensitivity of crows to the length of tools. When Alex started his PhD, one of the first experiments we suggested to him was to test the functionality of different shapes of tools. So that was the context in which he started to do the original experiment. Then one day he came to us with this video. He said: well, my experiment didn’t work as planned because one of the crows took away the bent wire! Then this is what happened. And he showed us the video of Betty bending the wire. When I saw that I was completely bowled over. It was unbelievable.


HS: I would like to step back a bit and ask you how you got interested in studying New Caledonian crows.

AK: Well, it was an accidental meeting with Gavin Hunt. Gavin had been working on these animals before and is the real pioneer of research on tool use by New Caledonian crows in the wild. Although there were reports before he started, he really introduced them to science and did the first detailed analysis. By that time, he had watched the birds and had studied the tools, mostly the Pandanus tools, and mostly through the templates they leave on the Pandanus trees rather than directly observing tool use. I met him 20 years ago in Vienna, at an International Ethological conference. But anyway, at that point, Gavin was coming to the end of a contract, and he wondered whether I would be interested in working with him. I was, and I applied for a grant to have him and my other post-doc working at that time, Jackie Chappell, who was also running out of contract. So, I applied for a Leverhulme Trust grant for two post-docs. As the application was being processed, Gavin got funding to work with Russell gray in New Zealand. He said he was going to stay there, but we also won our grant and so Jackie and I also started work on the crows. Jackie went to New Caledonia, returned with two crows that she managed to trap, and we started to work on their cognition in the laboratory. As time went by, obviously, the interests of the different groups expanded, and there has been some overlap. But that was the story.


HS: Were the crows that Jackie trapped the ones that were used in this experiment?

AK: Yes. One was Betty and the other – the male – was called Abel. The name was a play on Homo habilis, a little joke on being able to do things. Betty was named after the woman in the house where Jackie stayed when she was trapping the birds.


HS: How were these birds trapped? Did you use mist nets?

AK: Not mist nets, we used Cannon nets. It is very difficult to trap these crows because they are very responsive and very sensitive. It’s almost impossible to trap them twice in the same area. You have to set a Canon trap with net, bait an area repeatedly, and wait in a hide for a long time, till you attract, hopefully, a group of crows, at which point you shoot the cannon net and run and get them. There are other ways to trap them, but that’s what we did.


HS: It is interesting that there is this difference between males and females in the making of tools, at least in this experiment.

AK: I have been repeatedly asked that but that is not the case. We haven’t, so far, with the many other Crows we have studied, detected a significant difference between sexes. In this case, we had only one male and one female, and it just so happened that the female was more responsive. She was also younger. She was trapped as a younger animal and was more co-operative. At no point have we ever documented a good difference between sexes, but that doesn’t mean that it doesn’t exist.


HS: Did you use both males and females in the subsequent experiments?

AK: Yes.


HS: And they both make tools?

AK: Yes, they both do it, but quantitatively I can’t say if there is a bias.


HS: Do you remember how long the writing of this paper took?

AK: Ah well, it couldn’t have taken too long because it was published in the same year in which we made the observations. Science has a relatively fast processing system, so that the interval between making the discovery and publishing the paper was much shorter than usual, at least for us.


HS: This is related to what I wanted to ask you next. This paper appears in a section called Brevia in the journal. Was there a reason why you wanted to quickly publish a short communication and not wait till more experiments were done?

AK: We didn’t submit it as a Brevia. I believe that we submitted it as a full article but the editor said that they are willing to send the paper, but as a Brevia. So, our original submission was longer than this, had more detail, but we considered that a Brevia was still sufficient because it would give the basis of the finding we were reporting. After all, without pretending there is any similarity, Watson and Crick’s paper on the DNA was a very brief one and so was Koch’s report on the tuberculosis bacterium!

I think we felt that it was the kind of observation where the information was sufficient for the message we could give at that point. Obviously, experimentation would follow and did follow in many respects. If you discovered a single pig flying then you just report that pigs can fly. Statistically, you’re not making a statement about how well they fly, how many can fly etc.; you just say this is a possibility.

Of course, the field of animal cognition is evolving very rapidly and accumulating all sorts of evidence, so, today, writing a paper that says that an animal can make a hook, though still interesting if it hasn’t been observed before in a species, is less of a shock to the system, than when you observe it the first time. You have to go back to Jane Goodall‘s study of tool use by chimpanzees. When she first reported this, the reaction of people like [Louis] Leakey was that we now have to redefine what it means to be human. Today, we know that that were an overblown response, that it wasn’t justified, but he was shocked by the finding then. There was an established consensus that that time that only human used tools. Here we showed that another species does it. I am not saying that it had the same kind of effect, the same impact, but it is that kind of phenomenon.


HS: Was this the first example from a non-mammal?

AK: I think it was the first example of a non-human after the chimpanzees.  The chimpanzees modify objects to make tools, but they don’t modify in terms of bending a pliable substance to make a hook. They can refine a tool, and at that time it was known that they could pick a twig and remove the excess leaves to improve its functionality. We knew that the crows did that as well, but we didn’t know that they could make something by bending and shaping pliable material.


HS: Have other examples been discovered since?

AK: Yes, a group in Cambridge with two authors called Bird and Emery have found that rooks, another species of the crow family, when faced with a similar problem to Betty’s, arrive at a similar solution. What they do is not exactly the same but close to it. What the rooks did was to insert the wire into the vertical tube and pull over the edge to bend it. One of the differences is that Betty did it in many different ways and produced much better products. But more significantly, she took the wire away from the target and did it elsewhere and then came back to use it, which suggest that it involves some form of working memory. So, the animal is psychologically engaged with the task even if it moves away from the scene and if it takes some time to use the tool.

The rook is one interesting case of animals that do not use tools in the wild, but have the capability of doing it in some circumstances. Now there are also other examples. We also know keep discovering tools of greater complexity being made and used by other species. For example, chimpanzees can gather many leaves and make sponges. In doing that they are clearly modifying the substrate and manufacturing something. It is not a stable tool but it is still something that is constructed from many components. An interesting area that is understudied for the moment is the manufacturing of compound tools, i.e. tools made by the combination of more than one part. That is still in its baby days.


HS: Did this paper receive a lot of attention, when published, among peers and in the popular press?

AK: Yes indeed. It appeared in virtually every media across the world you could think of. It was on the cover of Le Monde, on the cover of the New York Times, in the BBC, in all the newspapers in the UK. I remember there was a lovely cartoon on the cover of the Daily Telegraph, one of the large British papers. There are two Crows, a couple, and one of them was trying to open the car with a hook and the other is saying: If you are so intelligent how come you forgot the keys inside the car? It was everywhere. And really, if anything, we had to engage in an exercise of press control, because they wanted to claim things that we hadn’t shown. Like the claim that this was a genius animal, or that we had shown that the animal understood the physics of the problem and things like that. In our paper, we were very careful to describe the facts and extract the inferences that we could. It is funny because, even today, even in the scientific press, some people say that we showed that Betty understood the physics, but those words are not present in our paper.


HS: What about attention within academia? Was it controversial when it was published?

AK: No. We didn’t have any controversy because it was just a finding. The videos we had make it completely clear. Sometimes we do have controversies, when the inferences depend on statistical analysis. The statistics are always subject to assumptions, and if the phenomenon is not obvious, you have to debate about the assumptions made in the analysis, but in the case of Betty’s finding there was not much to question. We didn’t face much debate; like I said, only in the media. Some of them were funny. One reporter insisted in that we had shown that females were more intelligent than males, which of course we hadn’t, with only one subject of each sex and only one task. There were all sorts of excesses and misunderstandings.


HS: What kind of impact do you think this paper had on the future course of your research? Did it open a new line of research for you?

AK: Yes, very much so. We had published some work on the New Caledonian crows before then, but I did not know whether this project was going to be a long-term one, or just over the life of a two-year grant. The fact that they did so many interesting things as soon as we brought them to our lab in Oxford meant that I stayed with the topic for many years. In the years that followed my main collaborator in the crow work became Auguste von Bayern, who is now based in Germany. All our Oxford crows were moved to Germany, where we did many other studies in Auguste’s laboratory.


HS: Did you continue this particular experiment – providing material and observing the shaping of tools – in a more detailed way, maybe with a larger sample size, after this paper was published?

AK: Yes. In one follow-up study we questioned whether they would be able to make other transformations of tools, and whether they showed high competence for solving tasks that did not involve tools. We also hand reared some crows to exclude the role of social learning, and identify heritable competences.


HS: Have you been to New Caledonia to see these crows in the wild?

AK: Yes, I have, just to accompany members of my group at the time. I had one post-doc and one graduate student – Christian Rutz and Lucas Bluff, who were working in the field. We worked as a team, but I was based in Oxford.


HS: When did you go to the field site?

AK: That was very long ago, in 2006.


HS: Has this behaviour – the shaping of objects to use as tools – also been documented in the field?

AK: Not exactly, but some similar behaviour has. In fact, their making of tools by sculpting twigs and branches was known before our work. It has recently been reported that in the field they also bend twigs to some extent, but the making of a functional tool by bending a pliable material has not yet been seen. However, this is not a major issue: if they can do it in the lab they surely can do it in the field. We should expect sooner or later to see that some crow exposed to the same problem in the field will reach the same solution. Since we did not train the animals, but they discover it on their own, it would be very strange if they could not do it in the wild.

Remember that these animals can make tools of many shapes by cutting and bending, can use multiple tools in a sequence, can select the kind of tool according to the need, can use tools to explore objects which they deem dangerous, can correct the diameter of a tool when it is too thick etc. In addition, we have shown, in a paper published in Nature, evidence for genetic propensity, that if you raised them in isolation they start showing some stereotypic tool related actions. So, there is this genetic propensity, but the many different experiments also show that there is creativity and flexibility. A similar story took place with capuchin monkeys. Capuchin monkeys were known to use tools in captivity but not in the wild. For a long time, people said that this was very strange. Why did they show this capability? And some people reached conclusions – wrong conclusions – about the lack of validity or interest of the tool-use observations in the lab. Some years ago, people discovered that capuchin monkeys used tools in the wild, and used them in a very flexible and creative way as well. The behaviour of every organism, including humans, has a genetic component, a learnt component, a social one and an element of creativity. The last is the least understood – How does creativity come about?


HS: Towards the end of the paper you make some suggestions for future lines of research. You say, “Comparisons between New Caledonian crows and their relatives, as well as between other cognitively exceptional birds and their relatives (11), offer a unique natural experiment to examine hypotheses about the ecological and neural preconditions for complex cognition to evolve.” Have any such comparative studies been carried out?

AK: Yes. For example, in our own work, we did a comparison of new Caledonian crows and Keas, the New Zealand parrot, for the solution of four different tasks. We found that, although it is only two species and four tasks, while the crows were better at using tools, the Kea seem to show greater cognitive flexibility in problem solving. I would say that the field is open. I do not think that the New Caledonian crows have exceptional intelligence in all dimensions. Every organism, including us, has genetic predispositions that allow good performance in certain areas of behaviour. For New Caledonian crows, the use of physical materials as tools is exceptionally well developed.


HS: That is also something I wanted to ask you about. In the last sentence of your paper you say, “It is not yet known if New Caledonian crows are also exceptional in cognitively demanding tasks not involving tools.” I am guessing that after this study you tested the crows for many other tasks as well?

AK: Yes. What we wrote in another paper is that we think that tool use is particularly revealing and makes it easier to demonstrate capabilities, rather than being particularly demanding in terms of intelligence. For this reason, we have shown certain capabilities in New Caledonian crows which are not known in other species, but that doesn’t mean that others don’t have it. It means that it is more difficult to demonstrate it in others. For example, ravens have given many indications of being exceptionally smart in many respects – the word exceptional here means more intelligent than what people expected. The philosopher [Daniel] Dennett says that we have to look for different ‘kinds of minds’ rather than whether there is mind or not, and I agree. We have to configure an overall picture of how the mind of the crow works, which we know is flexible, which we know is highly responsive to demands, and creative, but we are still very far from understanding.


HS: In the acknowledgements you thank Joan Silk for comments on the manuscript. Could you tell us a little more about how you knew her and how she helped?

AK: She was, at the time, present in the Wissenschaftskolleg in Berlin, where I was also spending time. I spent a lot of time discussing these matters with her and her husband Robert Boyd, and I found her comments very helpful. They helped my thinking in a general way. I wasn’t, at that time, familiar with the literature on primatology at all. Today I know a little bit more but then I was very new. And it was important to make the comparison with primates. Joan was very helpful in that respect.


HS: You also thank someone called D. Wilson for technical assistance. Could you tell us a little more about this person?

AK: Yes. David Wilson was a technician who helped us in handling the animals. He was extremely caring and understanding, and had a good relation with the animals. In fact, there is room for a funny story here: There were, in fact, two technicians, both called David Wilson. One of them (Old David) retired, and Young David took over. Both were, at different times exceptionally helpful and competent.


HS: You said Betty died 10 years ago. How did it happen?

AK: She got a fungal infection called Aspergillosis, which affected her breathing. We and specialist vets did everything we could but she didn’t survive. In the end, I was called by young David Wilson. I went to the aviary and picked her up. She was looking poorly, gasping and finding it difficult to breathe. I had her in my hands and eventually she couldn’t cope.


HS: What about Abel?

AK: Abel died earlier. We didn’t know his age at the time of trapping. While Betty was a young animal – she was only about 6 years old when she died – Abel was probably older. He also died of some natural cause.


HS: Were both these animals being used for experiments till the time they died?

AK: Yes. They were treated almost like pets because they were so responsive. Able was always a little bit more nervous, but Betty was very responsive.


HS: Do they also feature in later papers describing other experiments?

AK: Yes.


HS: What do you feed the crows?

AK: They eat fruit, nuts, and meat. They really like eating meat, such as pieces of cow heart or lung, as well as balanced dog and cat food. We try to give them a very varied diet, because we don’t really know their diet very well.


HS: I remember watching a BBC film in which they are shown eating these big white grubs.

AK: Yes, those are the ones that they extract with the twig. We have another paper studying the chemical properties of the grubs they eat. They are the larvae of Cerambycid beetles.


HS: Have you ever read this paper after it was published?

AK: Good question. I don’t think I have! Maybe, in the context of specific work, I might have looked at it partially. I should do it, to see how our thoughts have matured since. I would hope that we have learnt something since publication of that paper 15 years ago. Surely, certain things said in that paper will now be written in a different form.


HS: Would you count this paper as one of your favourites among all the papers you have published?

AK: Yes, it is among them. I wouldn’t say that it is the favourite, but it gave me enormous pleasure. I don’t think this is one of the papers where I felt I was – how can I put it- maximally challenged from the point of view of analysis. It was reporting an observation, so, in that sense, it was easy. There are other papers, which of course the media don’t touch, where we spent a very long time producing a model and trying to think about the problem and eventually getting a statistically meaningful result. Such work gives me enormous satisfaction intellectually but it does not necessarily reflect in the impact that it has.


HS: What would you say to a student who is about to read this paper today? What would you tell him or her are the main takeaways from this paper written 15 years ago?

AK: Be open to observation. That is the main thing. Be prepared to capitalise on the unexpected. This paper taught us, not that we didn’t know it already, but it convinced us even further, that you cannot anticipate what observing real animals will tell you. You have to be open to what they are telling you with their behaviour. That is crucial. At the same time, you have to think theoretically as well. It is not a matter of just describing what you see, like if you were just a camera. You have to interpret. This is my feeling.

The other thing is to tell them that the world is full of things that we still don’t understand, which makes biology an incredibly fascinating field. And this seems to only increase with time. We don’t know how the mind works in any species, including humans. And in my opinion, one of the most exciting questions for biology in the 21st century is to understand the relation between biology and mind. Many scientists are working on it, but we are still very far.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s