Understanding geographic range expansions in human-dominated landscapes: does behavioral flexibility modulate flexibility in foraging and social behavior?

Dear Corina and co-authors,

Many thanks for sending this revised version of your pre-registration. The two reviewers and I found that such version is highly improved and that you did a great job on clarifying many of the points that we highlighted. We appreciate your dedication and effort to improve your pre-registration and to answer our comments. Reviewer 2 has still some concerns about your predictions and the way you will measure your response variables associated with social behaviour (see below). Dr. Sebastián-González just pointed out minor comments. I agree with all comments.

Please, try to assess the points that were highlighted and provide a rebuttal letter explaining your decisions.

Best

Julia

My specific minor comments 1. Abstract. I found that your abstract is highly improved. Many thanks four your job! 2. About the populations you choose. Many thanks for providing such detailed information about the discussions you had related to this point. Now I have more elements to evaluate your design (time and financial limitations, permissions) I think that the one you are proposing is the most adequate. I particularly found interesting your partial solution (running a flexibility/individual differences test on groups of unmarked grackles at additional sites, which could really help with the issue that you bring up). 3. Species ecology. Thanks for adding so detailed information about previous knowledge on the species. 4. Social network analyses. Thank you for clarifying your analyses.

Reviewer 1 This is my second review of the pre-registration by Dr. Logan et al. This is a much-improved version of the project where most of my concerns and especially my questions on the fieldwork of the study have been addressed. I think the project is highly relevant and that it will help to go one step further in the understanding of behavioral flexibility. This pre-registration is very well designed, it has been thought and written with care and many details have been already taken care of. For example, authors include the R code they will use in their analyses, which is very advisable. The project is ambitious, combining both experimental and fieldwork data gathering, but seems overall feasible. I only have some minor suggestions to improve clarity in the text. I am looking forward to see the results of this study! Minor comments: • In the second sentence of H1, you use “they” to refer to the grackles (“where they must learn…”). I suggest using directly grackles to make this clear. Same in the first sentence of P3 (“the proportion of their diet”). • The first 7 sentences of H1 are written in a confusing way, with a lot of information in bracket and a continuation of the sentence that starts in line 1 only in line 5. I suggest dividing this in two sentences. • In methods, P4 you first cite the “strength of the maximum bond” which will be calculated as the “half-weighted index”, however, it is not until the analyses section when you explain this in detail. I suggest either moving the description of how this will be measured here, or to cite that this explanation is located in the analyses section of the proposal. • When you explain the Independent variables for P1-P4 and P6, you cite the “Dominance rank”. In your response you say that you describe how you measure “dominance”, but not what you mean with “dominance rank” Is this just the rank of the species ordered by dominance? If so, maybe it is good to explicitly say as other may get confused such as me.

Reviewer 2 (Dr. Esther Sebastián-González) Here, I have reviewed a revised pre-registration manuscript title ‘Is behavioural flexibility related to foraging and social behaviour in a rapidly expanding species?’. I reviewed this pre-registration previously. I have to admit that I find this revision is in a much better shape than the previous version (great job!). Authors have addressed each comment in details – thank you for doing this! Overall, I still find experimentally inducing varied level of behavioural flexibility interesting. Here are some comments authors may want to take into consideration during their final revision/write up: 1) P1 alternative 2: if this prediction turns out to be true, it would be interesting to analyse the type of food. In particular, the contents of the chosen food type by an individual. This may provide information about the relationship between behavioural flexibility and foraging 'success' (e.g. more behaviourally flexible individuals may chose particular food type that has relatively more fat, high calories etc). 2) In this revised version, authors have provided more ecological information about the study species than the previous one (great job!). However, I still find the information about grackles’ sociality missing – this will be related to your H2 hypothesis regarding social behaviours. While I assume grackles are social species, I prefer to read a clear statement/ fact about this species (if possible). 3) I apologise for not spotting this in the earlier version. The predictions for social bonding (both in quantity and quality) suggest that records will be taken during the out-of-breeding season. However, the examples of social behaviours that authors provide in their hypothesis suggests otherwise (i.e. data will be collected during the breeding season). I wonder whether the authors are planning to conduct observation during the breeding season? If so, would authors make similar predictions? Perhaps authors could clarify this further. From my opinion, on the one hand, taking records during the breeding season may add an addition confounding variable and complicate the understanding between behavioural flexibility and social behaviours - In this case, authors shall conduct observation during the non-breeding season. On the other hand, conducting observation during the breeding season may provide interesting data to understand the relationship between the level of behavioural flexibility and mating strategies (e.g. more flexible individuals may use several strategies to gain access to females than less flexible individuals). Perhaps authors could clarify this further, too? 5) H3 is actually an interesting hypothesis – it relates behavioural flexibility and habitats choice. The two variables, foraging choices and microhabitat types, are indeed tightly correlated. Authors shall pay particular attention when testing the hypothesis. 6) Last and a more trivial comment: In ‘C. Hypothesis’, using ‘behavioural flexibility’ would not be helpful. Perhaps adding what behavioural flexibility would relate to may be useful, for example, H1: Behavioural flexibility and foraging behaviour (or in foraging context, something alike); H2: Behavioural flexibility and social behaviour (or in social context, something alike).

Dear Dr. Corina Logan and co-authors,

Thank you for submitting your preregistration “Is behavioral flexibility related to foraging and social behavior in a rapidly expanding species? ” to PCI Ecology. My sincere apologies for the delay in giving you a first answer.

Dr. Esther Sebastián-González, an anonymous reviewer and myself have now reviewed your preregistration. You will find our general and specific comments below. I hope that they will help to improve your project. We all agree that this is an interesting, original and timely project. However, Dr. Sebastián González have some concerns on your hypotheses and the way you will carry out fieldwork related to the characterization of foraging and social behavior. Reviewer 2 suggests that (1) you should include general background information supporting your hypotheses, (2) reformulate your hypotheses to include the alternative predictions, and (3) rethink about how you will measure and characterize some of your variables. I agree with all their comments and suggestions, and I have some additional ones in the same line that are explained below.

I would like to invite you to submit a revised version of this pre-registration considering all comments and suggestions.

Sincerely

Julia

General comments

About the populations you will sample My main concern is related with the design of your study. You decided to conduct your study in three populations and consider them as blocks in your statistical analyses since you recognize that different relationships among variables can be expected in each population, which I agree. However, I am not sure that the populations you chose will be the most informative for your general question concerning the role of behavior flexibility on rapid range expansion. I mean you will use one population from the center, one from the middle and one from the edge of the geographic range of the species. If you are trying to understand the role of behavior flexibility in the expansion of a species range, why focusing on a population from the center of the species distribution and not different populations from the edge (i.e. trying to represent different edge conditions)? It is widely recognized that the different edges of the geographic range of species will impose different limitations for range expansion. Moreover, I expect geographic range expansion be more linked with populations in the edge of the geographic distribution of a given species. Then I think it will be more informative to perform your study in different populations from the edge of the geographic range of grackles, having central populations as controls.

I also suggest that you provide more information on the ecology of the species and the expansion of its range, which certainly may help understand your decisions about the populations that will be sampled.

In regard with your study design, have you thought about the possibility of comparing the association between behavioral variation and foraging and social behavior across species with contrasting geographic ranges or that expanded their range at different speeds? Have you considered performing such comparative analyses in the future?

Data analyses Social network analyses are comparatively less explained than other analyses. Please, include more information on which variables you will measure and how you will do it, how you will construct random networks and how you will compare your observed social network with randomized ones.

Specific comments

Abstract “However, behavioral flexibility is rarely directly tested at the individual level, thus limiting our ability to determine how it relates to other traits, which limits the power of predictions about a species' ability to adapt behavior to new environments”. This sentence is not clear to me. Which other traits? Other traits that are already know to relate to rapid geographic range expansion? What do you mean by rapid geographic range expansion? Please, try to be as specific as you are in the following sentence: “Results will allow us to determine whether (…) individual-level variation in flexibility is linked with diet breadth, foraging proficiency, social interactions, habitat use, and movement into new geographic areas”.

“We use great-tailed grackles (a bird species) as a model to investigate this question because they have rapidly expanded their range into North America over the past 140 years ((Wehtje 2003), (Peer 2011)) (Fig. 1)”. Here I suggest you provide more information about the natural history explaining the expansion of the geographic range of this species, in order to link it with the mechanisms you are proposing to evaluate in this project. Does this rapid expansion include new biomes? Or is the species just moving on the same kind of environments (i.e. expanding more by modifying dispersal ability and not new habitat use?). Are grackles using different vegetation cover types?

“This investigation: In this piece of the long-term project, we will assess whether performance in experiments that assess behavioral flexibility relates to variation in ecological and social behavior in the natural environment”. Are you referring to individual performance and individual variation? Or is it individual performance and across-individual (population) variation in ecological and social behavior? By being more specific in the formulation of these questions you may allow readers to evaluate their match with the methodology you are proposing. From this abstract is difficult to see if you will evaluate variation in flexibility, ecological and social behavior just at the individual level.

Hypotheses “H1: Individuals that are more behaviorally flexible (measured by reversal learning and switching between options on a multi-access box) will differ in their foraging behavior in the wild (measured with focal follows)”. I suggest you rewrite your hypothesis as: Behavioral flexibility is related to/modulates/influences foraging behavior. “Prediction 1: (…) validating the cross-species correlational finding that technique breadth (Overington et al. (2009)) and diet breadth (Ducatez, Clavel, and Lefebvre (2015)) indicate flexibility”. I suggest you change “indicate” by “are associated to”.

“P1 alternative 1: If there is no correlation, this suggests that flexibility is an independent trait from the number of foods eaten and foraging techniques used.” Why you use a statistical term in your prediction? I suggest you use “relationship” or association. On the other hand, since flexibility is the factor of analysis in fact this will suggest that the number of foods eaten and foraging techniques used are independent from behavioral flexibility.

“P1 alternative 1: (…) Flexibility is not necessarily associated with diet and foraging technique breadth because flexibility could be constrained in a foraging context due to social competition (e.g., subordinates are outcompeted while foraging and thus try new foods and techniques) or ecological limitations (e.g., constrained by what is available)”. In the same line of my previous comment and following the logic you presented previously, you should state that foraging techniques and items might be constrained by other factors such as… For me, this will constitute an alternative hypothesis and not an alternative prediction.

“P3: The more flexible individuals eat more human food, potentially due to A) having stayed in their parent's home range (i.e., they eat human food because it happens to be more prevalent in their home range than in other home ranges; local specialization) or B) because these individuals move around to seek out such opportunities (potentially seeking out habitat edges within their population). Foods eaten will be recorded during focal follows”. What does it mean “more human food”? More quantity? More items? Eating human food will be considered an expansion of their diet (i.e. adaptation to new environments) or not? How innovative techniques will function in this case? Do you have predictions on these techniques? Thinking on the populations you chose (see general comments): Do the populations that you propose to sample differ on their proximity to urban areas? How may this proximity affect the relationships you expect to find? “H3: Individuals that are more behaviorally flexible (measured by reversal learning and switching between options on a multi-access box) will use a wider range of habitats (measured with GPS point for each focal follow).” Can you be more specific about what you will consider different habitats and therefore how you will measure the range of habitats? In P5 and P6 you use diversity of habitats so I think it will help to have a more specific definition of habitat here. Will you use different vegetation cover types? Or compare wild, semi-natural and urban habitats?

METHODS P6: flexible = wider range of habitats 1.Evenness in the proportion of time spent in each habitat type (grass, gravel and other natural substrate, cement, cafe, dumpster). Comment: Why you use one measure of habitat diversity? I suggest you use different qualitative and quantitative measures to fully understand habitat use or provide a strong argument to just use evenness (which I found will provide you a highly constrained definition of habitat use).

About the use of G tests It is not clear for me why you present those G tests if (1) you cannot include the exact variance structure of your future models, which will modify the sample size you will need; (2) you are planning to use the Bayesian approach.