1. Skip to Menu
  2. Skip to Content
  3. Skip to Footer

The mental maps of bees


Randolf Menzel is Professor Emeritus of Neurobiology at the Free University of Berlin in Germany. Since he first started investigating insects almost 50 years ago, Professor Menzel has studied how honeybees learn colours, how their memory is organized, and how memory is implemented in the neural network of the bee brain. More recently he has been looking at how bees navigate to different locations and how they communicate about locations. Here he discusses his recent paper published in Proceedings of the National Academy of Sciences about a proof for the existence of mental maps in honeybees.1  

Credit: Kino Lorber Inc.


What got you started in studying the way honeybees "think" and behave ?

One of the questions that motivated me to study biology, chemistry and physics at university was how animals with rather small brains like worms, crustaceans and insects learn and use memory for adapting their behavior to the conditions of a changing world. At high school, I observed plankton under the microscope. Tiny little animals, rotifers, which belong to the group of nematodes, particularly impressed me. They have a clear glass body, and I could see their eyes, muscles and brain under a microscope when they maneuvered through the water. My first scientific publication describes the ecology of a fresh water pond close to my hometown on the Rhine River. Martin Lindauer, a Professor at the University of Frankfurt and a co-worker of Karl von Frisch, accepted me as a graduate student without any exams, and allowed me to address the question of how bees learn colours. Martin Lindauer was an enthusiastic zoologist. He continued to do experimental work even when he had to run the zoology department. He and Franz Huber, an inspiring Professor at the University of Tübingen, were role models for me throughout my academic career.


One of your recent projects is bee navigation and how it depends on mental maps. How do you define a mental map ?

Animals, including humans, use many senses and many behavioral strategies to navigate from one place to the next. Both innate mechanisms and acquired information about the world are used for navigation. The question for behavioral biologists and neuroscientists is how these various forms of information about the environment are integrated to organize navigation. Neuroscientists have good reasons to believe that mammals, for example the laboratory rat, navigate by a map-like representation of the explored environment – a cognitive map in the hippocampus of the brain. Such forms of memory about the environment store the spatial relations of landmarks and the meaning of locations for behavioral control. Behavioral biologists find it difficult to accept the concept of a cognitive map because they believe that the same behavioral phenomena can be explained by multiple and separate sensory-motor routines. Since there is no neural correlate of navigation in insects it is rather difficult to prove the necessity or existence of a cognitive map in insects. I define a mental or cognitive map as a memory structure that stores spatial and temporal relations of landmarks in such a way that behavioral routines like expectation, planning and travel shortcuts are possible. Shortcutting is the essential proof of a metric mental map, because it requires self-localization and goal localization if other, more simple mechanisms, like steering towards a beacon or matching with the panorama, can be excluded. I view the mental map of animals, including the honeybee, as an action memory of spatial relations rather than a sensory representation as we humans experience by introspection. Action means not only expressed motor behavior but also "internal doing" or "thinking" that leads to expectation and planning.


Do we know what type of molecular or cellular events underlie the formation of a mental map in an insect's brain ?

Unfortunately no. Together with my co-workers we are working hard to search for neural correlates in the honeybee brain that code for spatial relations of objects that possibly control navigation.


On a practical note, how difficult is it to track the flight paths of bees ?

Navigation cannot be studied in the lab or in the vicinity of the hive or a feeding place. In my view, the large amount of data collected under such restricted conditions are rather irrelevant for navigational studies. Bees cover distances of kilometers. Therefore, we use a special radar system, a harmonic radar, that tracks them over 1–2 km. The bee has to be equipped with a transponder, a passive radar antenna that weighs 20 mg and measures 12 mm in length. It can easily be carried by a bee and does not disturb its flight. The data are highly informative, but the method is not easy because the harmonic radar is a sensitive device that is exposed to the rough conditions of the environment, for example rain, strong wind, thunderstorms. I would not have been able to collect such data if I had not had the luck to collaborate with Uwe Greggers, a highly gifted engineer, for more than 40 years.


You reported your recent results in Proceedings of the National Academy of Sciences. Have you been surprised by the interest in the paper ?

Bees are small animals with a small brain. Navigation according to a mental map has been considered to be the realm of big brains. Documenting a mental map for navigation in bees comes as a surprise for some and arouses skepticism and opposition. We had proposed the existence of a mental map in bees earlier as one of several mechanisms of their navigation.2 This interpretation was met with skepticism and alternative interpretations were published. We therefore searched for experimental approaches that could decide between alternative interpretations.


Some researchers say bees can navigate without a mental map and that your study does not exclude this possibility. What is the reasoning and how do you respond to this ?

We documented in earlier publications that bees are able to perform shortcutting flights thus fulfilling the requirement of a mental map if more elementary solutions could be excluded. One elementary solution of shortcutting could not be excluded in our former studies which is discussed in an earlier paper.3 It was therefore helpful that Cruse and Wehner pointed out that bees could solve the kind of shortcutting we demonstrated by an algorithmic procedure, namely vector addition of two stored vectors.4


The essence of the argument put forward by Cruse and Wehner was that the directional component of the respective vectors is related to the sun compass. We therefore searched for a method that allows us to shift the sun compass. We found a solution by anesthetizing the animals for six hours, which is a long time. This stops their inner clock, which in turns leads to a shift in their sun compass. We found that these timeshifted animals initially fly in the wrong direction. In this case, further to the east because the inner clock gave them an earlier time. However they returned home as fast as the control bees that had not had their internal clocks shifted in time.1 Since we excluded more elementary solutions we concluded that the bees referred to a representation of the ground structure equivalent to a mental map. This conclusion was questioned on two grounds: First, bees may have adjusted their inner clock within minutes during the vector and search flights, and second bees may have referred to the skyline of the panorama.5 Both of these suggestions can be rebutted because adjustment of the time shift after 6 hours of anesthesia requires days and the angular modulation of the skyline was below a 2° visual angle in the test area. Bees do not have a spatial resolution by their compound eye better than a 2° visual angle. An improvement of spatial resolution as suggested by Cheung and co-workers, namely a modulation of the brightness received by a single ommatidium, a single visual pixel in the compound eye, has not been shown in any insect and is not supported by any experimental evidence in the honeybee.5 We therefore maintain our conclusion that vector addition is excluded as a possible mechanism for homing in these experiments.


Is it a bold statement to say that your experiments prove the use of mental maps in bees ?

There are two components to the concept of a mental map: the relational representation of landmarks and the meaning of locations to the animal. Indeed, we believe that we have documented beyond doubt that bees refer in their navigation to a metric representation of the environment. However, we have only little evidence that they assign meaning to the experienced locations. It will be a topic of further research to address this question and honeybees provide us with the opportunity for an answer. Bees communicate about locations with a symbolic form of information transfer – the waggle dance. Relating waggle dance communication and navigation has allowed us to postulate that they refer to the same kind of spatial memory in their own navigation and in their dance communication.3 Further studies will help us to test whether they make decisions about the value of the dance-indicated goal based on their own experience with that goal.


It is true that there is a distinct difference between the way ants, for example, navigate on the ground and bees navigate in flight ?

Indeed there is a distinct difference between earth bound animals like ants and airborne animals like bees as they experience and use spatial information about the environment. Ants see the panorama as a 360° circle of skyline profile, bees experience the ground structure as a geometric map at the outset combining these sequentially experienced maps to an integrated map. Generalizing from ants to bees appears to me as inappropriate.


Science is often more about differences in opinion that many people think and it is important to reach a consensus. What is the best way to do this ?

Disagreement is a highly productive resource in science and should not be considered as a nuisance. The opinion of the minority can open the field to significant new discoveries and fundamentally new directions. Science is not about democratic consensus building. Experimental evidence and conclusive arguments are the backbone of science. The controversy about the mental map in the bee is an intense debate going on right now in behavioral biology and neuroscience. Traditional behavioral biology tends to ignore the richness of even small brains and tries to avoid making any assumptions about their capacity to probe the potential outcomes of future behavioral actions. This "inner doing" is at the realm of neuroscience. In this sense, neuroscience has followed the cognitive turn much more radically than behavioral biology. My thinking is governed by both disciplines but I notice that behavioral biology lags behind evidence and sticks to traditional concepts that become increasingly irrelevant when we ask how the brain performs all these wonderful tasks.


You plan to do experiments with bees in mazes. What will this show ?

We need to find neural correlates of navigation in insects. There is no hippocampus in insects but the necessary convergence site between the multiple sensory modalities involved in navigation and the high order control of behavioral acts is already quite well known. Insects are too small to mount amplifiers and wireless transmitting electronic units on their body. We recently succeeded to record multiple high order interneurons in the bee brain when the animal is freely moving within a colony. We also found it possible to record from such neurons when walking bumble bees navigate in a small maze. I say navigate but this kind of navigation in such a maze is rather simple. It is close to the conditions rats are tested when their hippocampus is recorded. These experiments are a beginning. We are confident we shall find ways to record high order interneurons in the bee brain even when they fly and explore the environment.



  1. Cheeseman et al. (2014) Way-finding in displaced clockshifted bees proves bees use a cognitive map. Proceedings of the National Academy of Sciences USA 111(24): 8949–8954.
  2. Menzel et al. (2005) Honey bees navigate according to a map-like spatial memory. Proceedings of the National Academy of Sciences USA 102(8): 3040–3045.
  3. Menzel et al. (2011) A common frame of reference for learned and communicated vectors in honeybee navigation (2011) Current Biology 21(8): 645–650.
  4. Cruse H and Wehner R (2011) No need for a cognitive map: decentralized memory for insect navigation. PLoS Computational Biology 7(3): e1002009.
  5. Cheung et al. (2014) Still no convincing evidence for cognitive map use by honeybees. Proceedings of the National Academy of Sciences USA. doi: 10.1073/ pnas.1413581111: .






This email address is being protected from spambots. You need JavaScript enabled to view it.


Tilmann KießlingTilmann Kießling
Head, Communications
T. + 49 160 9019 3839