Honeybees and bumblebees are remarkable little creatures. Apart from the pollination services they provide for our agricultural industry, they also are extremely good at learning about things. Why are they such good learners? Because bees visit a wide variety of flowers which differ in the quality of food they offer, it makes sense for a bee to be able to rapidly learn which flowers have the best food available. Bees are in fact so good at learning that we use them as ‘models’ for understanding how many animals learn, including humans. In a previous couple of posts (here and here) I discussed the cognitive abilities that these animals have, so I won’t go into too much more detail here suffice it to say that we know an awful lot about bees’ ability to learn about things.

However, even though bees have some pretty impressive learning abilities, nearly all of what we know about this comes from bees’ ability to learn in relation to nectar. Now, this makes sense, given that bees collect nectar from a range of flowers. However, while bees need nectar to stay alive from hour to hour while foraging (like hummingbirds), they also need pollen from flowers as their main source of protein. We could survive on soft drinks and coffee for a while, but eventually we’d need a steak or some beans, and it’s not so different for bees. What’s more, the pollen bees collect is critical for the survival of their offspring, so they pack it on to their legs (in pollen loads) and carry it back to their colony to feed it to developing larvae.

However, despite the importance of pollen to bees’ lives, very few studies have addressed whether they can learn about it as they do nectar. I recently addressed this in a study published in Animal Behaviour.

If bees do learn which flowers have pollen, what might they learn? Since bees collect pollen from the anthers of flowers, anther colour seemed like a likely candidate. However, given that the corolla (petals) of a flower are usually pretty obvious and that bees pay attention to them when collecting nectar, it also seemed plausible that bees might learn about corolla colour when collecting pollen.

To address this, I made artificial flowers with different coloured corollas and anthers. Bees then were allowed to collect pollen from one type of flower, while another type had no pollen on it. I scented the unrewarding flowers to smell like pollen so that bees would still fly to them and investigate, but then leave when they discovered no pollen there.

A first group of bees were trained that flowers with a particular corolla colour had a pollen reward, for example, flowers with blue corollas had pollen while flowers with yellow corollas had no pollen. Other bees were trained to the opposite relationship (i.e. flowers with yellow corollas having pollen while flowers with blue corollas not having pollen). A second group of bees were trained to a particular anther colour (either flowers with blue anthers or flowers with yellow anthers having pollen on them). Finally, a third group of bees were trained to a specific colour combination where either the flower had a yellow corolla with a blue anther, or a blue corolla with a yellow anther.

To keep track of individual bees I stuck little numbers on their backs. This meant that a given bee (e.g. white 10) could be trained by herself. I trained bees in this way rather than lots of individuals at a time because bees copy where other bees get rewards from. I wanted to be sure that everything a bee learned was by herself.

After a bee had had a chance to visit the artificial flowers and collect pollen from them, I then presented her with an array of flowers that looked the same as the flowers she had been trained on, except this time none of the flowers had any pollen on them. If the bee had learned that a particular flower type had pollen then I expected that she would fly to that flower type first to search for pollen there.

This is indeed what I found. In the slowed-down video below, you can see a bee that was trained to find pollen on flowers with yellow corollas. Here she searches on three yellow-corolla flowers. Across all the treatments, the bees learned to visit the flower type they had previously been trained to. What’s more, to test the bees that had been trained to a particular corolla-anther colour combination, I presented them with four different types of flowers: flowers with blue corollas, blue anthers, yellow corollas yellow anthers, blue corollas yellow anthers and yellow corollas blue anthers. If the bees were preferentially learning about either the corolla colour or the anther colour then I expected them to go to, for example, to all the flowers with blue corollas regardless of the anther colour. However, bees went to the specific colour combination they had been trained to, demonstrating that they paid attention to and remembered both the corolla and anther colour when collecting pollen.

Finally, I wanted to see whether the bees could remember which flowers had pollen over a longer period of time. I trained bees as before, but this time tested them 24 hours later, and then seven days later.  Even after seven days, bees searched for pollen on the flower type that they had previously found pollen on.

So, what do these findings tell us? First, that when you see a bumblebee flying around and visiting flowers outside that she isn’t just remembering which flowers had nectar, but also which had pollen. What’s more, she could be remembering pollen she found on those flowers seven days ago, which is a substantial amount time for an animal that only lives a few weeks.

These results are also exciting to me because they give us a new type of reward to work with when looking at learning and memory in this animal. As an individual bumblebee doesn’t only collect nectar or pollen, but rather both at the same time, this means that she is potentially having to keep in mind which flowers offer the best nectar, which offer the best pollen and which have both. As I was interested to see whether individual bees could in fact do this, this lead me to my second experiment, which I talk about in my next blog post.


Muth, F., Papaj, D. R., & Leonard, A. S. (2016). Bees remember flowers for more than one reason: pollen mediates associative learning. Animal Behaviour, 111, 93–100. doi:10.1016/j.anbehav.2015.09.029

A video about this paper can be viewed here: