Reversal learning of visual cues in Heliconiini butterflies

This article has been Reviewed by the following groups

Read the full article See related articles


No abstract available

Article activity feed

  1. This Zenodo record is a permanently preserved version of a PREreview. You can view the complete PREreview at

    Review of the preprint- Reversal learning of visual cues in Heliconiini butterflies DOI:

    Discussion led by: Irena Kleckova (28.04.2023)

    Notes taken by: Irena Kleckova (5.05.2023)

    Review written by: Journal Club (Prapti Gohil, Sakshi Ravindra Tripathi, Daniel Linke, Mar Repulles, Phil Honle, Petr Klimes, Pavel Matos, Irena Kleckova, Pedro Ribeiro, Karolina Hruba, Kristian Grell, Alena Suchackova)


    The manuscript deals with testing the cognitive abilities of 6 representatives of the butterfly tribe Heliconiini - three Heliconius species and three species of other genera (Agraulis, Dryadula, Dryas). The cognitive skills are then attempted to be explained by differences in brain structure (mushroom bodies) and feeding strategy (pollen feeding in Heliconius). The study is based on the hypothesis that Heliconius butterflies with comparatively larger mushroom bodies will have better cognitive abilities than other butterfly genera. The methods include training the butterflies to associate a food reward (sugar-protein)/aversiveness (quinine solution) with a visual cue and then comparing their ability for reversal learning. All the six species showed reversal learning but no significant difference was detected among Heliconius and non-Heliconius species. Overall, the study is interesting, well-designed and tackled a single, clearly stated hypothesis, but is missing some important details about methods and the overall framing of the concept (more focused on mushroom bodies, thus, expecting to be a physiological type of study) and discussion (more focused on ecology) do not fully correspond with the objectives of the study and results. 

    Positive feedback

    The study is overall nicely structured and easy to read. The experimental design of the study with butterflies is logical and easy to follow. The introduction contains a broad background. Also, the discussion contains a nice summary of results with appropriate background. We believe that the manuscript is strong and below we provide suggestions to further strengthen its clarity.  



    The title corresponds well with the content of this study.

    Abstract and introduction 

    In general, the abstract and introduction may lead readers to misinterpret the main hypothesis  in the study. The study tested reversal learning abilities of several butterfly species and found no significant differences between species. However, the abstract and introduction mainly focus on the link between mushroom body sizes and reversal learning, but this is not formally being tested in the study by quantifying the mushroom body sizes of the tested individuals. Although this hypothesis may be interesting and is directly connected to the obtained results, it might be better to add it during the discussion, but not as the main objective/focus of the study.

    The last paragraph of the introduction states, that 6 different Heliconiini species have been tested for their cognitive reversal abilities, this should be extended to include the prediction, that true Heliconius species will show higher reversal learning capabilities compared to the other genera (Agraulis, Dryadula, Dryas).


    Experimental protocol:

    • It was somehow unclear whether the authors also recorded the learning experiments using a GoPro camera. If so, it might be possible to record the feeding attempts through time, thus, to estimate butterfly learning rates.

    • How was the proportion of feeding attempts calculated? Is it proportion per flower or proportion per individual? 

    • How many individuals were put in each cage during the test period?

    • Training took 4 days - How long was the feeding activity of each species per day?

    • If recorded, what was the weather and temperature during the testing days - might be important.

    • Write the total time recorded by cameras and a basic overview of the raw data


    • It would be useful to add an extra figure with the experimental protocol workflow and perhaps photos of the experimental settings.

    • The number of tested butterfly individuals are only stated in Figure 1, but they can be also mentioned in the text.

    • What do the points in Figure 1 represent? Are these feeding attempts by individual butterflies, or are they feeding attempts on each particular flower?

    • In the ethical note it is mentioned that individuals were marked, thus, is it possible to quantify butterfly individual variability in feeding attempts?


    It is somehow unclear that the first paragraph of the results focuses on a significantly better performance of Heliconius than non-Heliconius in the tests, while in the abstract it states "we found no evidence that Heliconius performed better than the other Heliconiini species". Also, an interaction effect among clades is mentioned, but not explained and/or visualised. This is confusing, and perhaps came from differences in main effects vs pairwise comparison, and contra the results among trials at species level in the second paragraph. Those shall be better explained and a consistent conclusion made.


    • The discussion about vertebrate brain size does not seem relevant. Perhaps adding discussion on learning abilities of other pollinators would be better (e.g bees).

    • Expand the discussion on why Dryadula was different in its cognitive skills – smaller mushroom bodies?

    • Perhaps breaking the third paragraph would add to the clarity and fluidity of the text. It starts with a general claim about "adding to the evidence" and then it changes to "brain size" and the two things do not seem connected in this paragraph.


    • Line 211: big letter in species name Melpomene

    • Line 225: "a widely used a measure…" - I believe there is an extra "a" there.

    • Line 227: update instead of updating?

    Lines 229-230: Maybe "better reversal learners" would read more smoothly than "more capable reversal learners"

    Competing interests

    The author declares that they have no competing interests.