Trait-Based Selection of Seeds Ingested and Dispersed by North American Waterfowl

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The authors Almeida et al. present a meta-analysis of published data examining the dietary preferences of waterfowl (differentiated by foraging guilds) in wetland ecosystems in North America. With a particular focus on seed consumption and its implications for potential seed dispersal. This study offers valuable insights into the selective interactions between waterfowl and plant traits—a topic of significant ecological and applied interest.

I found the manuscript to be well-written, clearly organized, and makes a strong contribution by synthesizing a broad dataset to highlight key trait-based preferences. The authors reveal interesting patterns, notably the tendency of waterfowl to select seeds of intermediate size, and differences among foraging guilds. These findings have potential implications for wetland restoration and plant biogeography under shifting climatic and land-use scenarios. I find this work to be both insightful and relevant for applications in wetland management and ecological restoration. The authors have compiled a robust and valuable dataset that can serve as a springboard for future work—particularly studies incorporating direct measures of movement and gut passage.

I fully support its publication, pending minor revisions to address the following points:

1. Because the data are inherently correlative, it is important to interpret these patterns as proxies for potential dispersal rather than definitive mechanistic explanations. This distinction is acknowledged in the manuscript, but might benefit from a more explicit statement in the introduction.
2. The introduction could benefit from a concise overview of relevant feeding ecology for each foraging guild. For example, whistling ducks are more likely to forage in trees, which may explain their different patterns. This context would help readers interpret the trait-based findings more holistically.
3. The authors mention in the conclusions that small seeds might be underrepresented due to identification challenges. Given the importance of this potential bias, I recommend bringing this up earlier—perhaps in the methods or discussion—to emphasize how it might shape interpretations of seed size preferences.
4. To connect these dietary preferences to actual dispersal and LDD, it would be helpful to briefly discuss gut retention time, movement behavior, and seed survival during gut passage. These processes are key for translating ingestion data to real-world dispersal outcomes and should be at least acknowledged as crucial next steps.
5. There is a well-known allometric relationship between body size, gut retention time, and movement capacity in animals. Including the data on the average body size for each guild could strengthen the link between the observed dietary patterns and the potential for LDD.
6. One inherent challenge with meta-analyses is that each included study has its own microhabitat, sampling methods, and conditions—factors that can strongly influence food selection and thus apparent patterns. It would be helpful for the authors to acknowledge this explicitly as a potential limitation of their approach.
7. Finally, some additional clarity on the dataset would be helpful. Specifically:
   1. How many plant species were recorded per foraging guild?
   2. Did the authors detect significant spatial or taxonomic overlap across studies?

Thank you for this interesting work.

Author Response

Comments: The authors Almeida et al. present a meta-analysis of published data examining the dietary preferences of waterfowl (differentiated by foraging guilds) in wetland ecosystems in North America. With a particular focus on seed consumption and its implications for potential seed dispersal. This study offers valuable insights into the selective interactions between waterfowl and plant traits—a topic of significant ecological and applied interest.

I found the manuscript to be well-written, clearly organized, and makes a strong contribution by synthesizing a broad dataset to highlight key trait-based preferences. The authors reveal interesting patterns, notably the tendency of waterfowl to select seeds of intermediate size, and differences among foraging guilds. These findings have potential implications for wetland restoration and plant biogeography under shifting climatic and land-use scenarios. I find this work to be both insightful and relevant for applications in wetland management and ecological restoration. The authors have compiled a robust and valuable dataset that can serve as a springboard for future work—particularly studies incorporating direct measures of movement and gut passage.

I fully support its publication, pending minor revisions to address the following points:

 

Response:  Thank you for these positive comments.

 

 

Comment: Because the data are inherently correlative, it is important to interpret these patterns as proxies for potential dispersal rather than definitive mechanistic explanations. This distinction is acknowledged in the manuscript, but might benefit from a more explicit statement in the introduction.

 

Response:  in the introduction we have now added the following words: “(as a proxy for which seeds they disperse)”

 

 

Comment: The introduction could benefit from a concise overview of relevant feeding ecology for each foraging guild. For example, whistling ducks are more likely to forage in trees, which may explain their different patterns. This context would help readers interpret the trait-based findings more holistically.

 

Response: we have now added the following text: “We tested these hypotheses for different guilds, since they have differences in feeding ecology which were expected to influence the traits of the plants whose seeds they ingest (Baldassarre & Bolen 2006). Dabbling ducks feed mainly at or close to the surface in and around the shoreline. Diving ducks feed mainly by diving to depths from 40cm to several metres. Sea ducks are largely marine carnivores that also dive, but often breed in freshwa-ters and may disperse seeds attached to or inside their prey (van Leeuwen et al. 2017). Geese are herbivores that graze on land or grub in sediments. Whistling ducks are limited to the southernmost parts of North America, and are relatively understudied.”

We add that we are not sure it is correct to say that whistling ducks forage mainly in trees. Personally, I have only seen them in open wetland habitats and never in trees. Many speices from other guilds can nest in trees.

 

 

Comment: The authors mention in the conclusions that small seeds might be underrepresented due to identification challenges. Given the importance of this potential bias, I recommend bringing this up earlier—perhaps in the methods or discussion—to emphasize how it might shape interpretations of seed size preferences.

 

Response: we have now added the following text in the “Wetland plant list and plant traits” part of the Methods section: “Furthermore, we cannot rule out the possibility that the smallest seeds were sometimes overlooked in waterfowl diet studies, with further potential for bias”

 

 

Comment:  To connect these dietary preferences to actual dispersal and LDD, it would be helpful to briefly discuss gut retention time, movement behavior, and seed survival during gut passage. These processes are key for translating ingestion data to real-world dispersal outcomes and should be at least acknowledged as crucial next steps.

 

Response: We feel that this is important, but was already covered to some extent in the Conclusions section. However, we have now extended the “Limitations of our study” section, by adding the following new text: “Furthermore, we have used seed ingestion as a proxy for seed dispersal, and have ignored the importance of plant traits in determining seed survival during gut passage, as well as gut retention time which is also influenced by diet (see Green et al. 2016, 2023 for review). For example, ….. LDD events are favoured by migratory behaviour, and by longer re-tention times (e.g. for larger seeds, Lovas-Kiss et al. 2023b).”

 

 

Comment: There is a well-known allometric relationship between body size, gut retention time, and movement capacity in animals. Including the data on the average body size for each guild could strengthen the link between the observed dietary patterns and the potential for LDD.

 

Response: In fact, there is no consensus on such an allometric relationship for waterfowl. Whereas it is generally assumed for animals that larger animals move more, and have longer retention times (e.g. Yoshikawa et al. 2019), for waterbirds Viana et al. (2013) found precisely the opposite (i.e. they found that smaller birds move farther, and have longer retention times). Given this uncertainty, and how data on body sizes and movement patterns of North American waterfowl are readily available in previous literature, we suggest that it is best not to enter into much speculation in the current paper about how differences between guilds in body sizes will influence dispersal processes.

Viana, D. S., Santamaria, L., Michot, T. C., & Figuerola, J. (2013). Allometric Scaling of Long-Distance Seed Dispersal by Migratory Birds. American Naturalist, 181(5), 649–662.

Yoshikawa, T., Kawakami, K., & Masaki, T. (2019). Allometric scaling of seed retention time in seed dispersers and its application to estimation of seed dispersal potentials of theropod dinosaurs. Oikos, 128(6), 836–844. https://doi.org/10.1111/oik.05827

 

 

Comment: One inherent challenge with meta-analyses is that each included study has its own microhabitat, sampling methods, and conditions—factors that can strongly influence food selection and thus apparent patterns. It would be helpful for the authors to acknowledge this explicitly as a potential limitation of their approach.

Response: We have now added the following text to begin our section “Limitations of our study”: “Each study identified from our literature search involves a particular study area, sampling and seed identification methods, and other specific conditions, which can all influence the data included in our analysis, and hence the patterns we observed.”

 

Comment:    Finally, some additional clarity on the dataset would be helpful. Specifically:

        How many plant species were recorded per foraging guild?

        Did the authors detect significant spatial or taxonomic overlap across studies?

Response: Please bear in mind that this is a sister publication to the previously published Almeida et al. 2025, where many such details can be found. We would think it is not necessary to repeat the same information in this new paper. For the earlier paper, the supplementary material (specifically Table S1) gives summaries of the number of studies per bird species, and the number of plant species, separated by guilds.

This Table S1 shows clearly that E.g. there is a maximum 36 studies for each bird species (mallards in this case, which also has the maximum (280) number of plant species). In terms of the median number of plant species found in a member of each guild, the order (decreasing) would be:

Dabbling ducks > Diving ducks > Whistling ducks > Sea ducks > Geese

For taxonomic overlap for plant species between studies, we can see that up to 26 vectors have been recorded for the same plant species (maximum for Ruppia maritima). Since all the dataset are provided in Almeida et al. 2024, there it can be seen e.g. that these data for Ruppia maritima come from 24 separate studies.

In terms of spatial overlap, in part that comes from the many studies that looked at different bird species in the same area (e.g. Landers et al. 1976 looked at 11 bird species in one site in South Carolina).

At another level, there are many studies from the same state, e.g. 13 studies from Texas, 8 from California plus two more than come from an area spanning California and Oregon. This does not include review publications that already combine data from many states.

 

Comment: Thank you for this interesting work.

 

Response: Thank you for your time and insightful comments.

Reviewer 2 Report

Comments and Suggestions for Authors

I have read the paper " Waterfowl-plant interactions in North America: comparing  traits of plants whose seeds are ingested and dispersed by ducks and geese with those of the wetland flora" by Green et al.

This is a well written and insightful paper. I only have a few comments for the authors to consider:

1) I think you cannot properly talk about "selection" when ingestion of seeds can be accidental or unintentional.  Selection derives from intentional non-random actions. I thus suggest more care with the use, and possible using instead "frequently ingested" and similar variations. I have made comments in the PDF attached. 

2) The statistical analyses using chi-square tests are minimally adequate. I think the authors could have used a more robust GLM for frequency data (binomial errors) that allow for an integrated (as opposed to fragmnetary) model that can include all factors and their relevant interactions. A consequence of the current approach is the overestimation of "significant" effects

Please see my comments in the uploaded df.

sincerely,

Tomás A. Carlo

(signed review)

Comments for author File: Comments.pdf

Author Response

Comment: I have read the paper " Waterfowl-plant interactions in North America: comparing  traits of plants whose seeds are ingested and dispersed by ducks and geese with those of the wetland flora" by Green et al. This is a well written and insightful paper.

Response: Thank you for your comments and insightful, thought provoking feedback.

 

 

 

Here are details of the comments in the uploaded pdf, and their responses:

 

Comment: "more than" - compared to what? smaller and larger seeds? Clarify

Response: words have been added to provide further details.

 

Comment: this could be expanded a bit to say something about the "connectivity" provided, gene flow, etc. briefly I mean. within the same sentence.

Response: the sentence has now been expanded to read “Through seed dispersal, waterfowl in North America provide a major ecosystem service, ensuring connectivity and gene flow between plant populations”

 

 

Comment: could add that they exceed the average dispersal distance of most fruigivorous landbirds as well!

Response: We have extended the sentence with the words “exceed those of most frugivorous terrestrial birds and”

 

 

Comment: didn't Janzen coined this term? Janzen, D. H. (1984). Dispersal of small seeds by big herbivores: foliage is the fruit. The American Naturalist, 123(3), 338-353.

Response: Yes, this citation was now added (but since Janzen focused on mammals, we have retained Green et al. 2022 which treats its application in waterfowl).

 

 

Comment: you could have used a GLMM with a binomial error distribution for a more robust analysis - this would allow you to to analyze the effects of different types of waterfowl too, and the plant traits directly in one model. This woiuld include interaction terms betweeen waterfowl types and plant traits.

 

Response: unfortunately, there are too many data gaps to apply such a GLMM approach. We were missing plant trait data for ~42% of NWPL species, and many species found in the waterfowl diets. When a plant species has data for one trait, it does not always have data for the other trait. Furthermore, many plant species occur in only one guild’s diet. This explains our approach of doing separate tests for each trait (since the gaps for each trait do not align). However, we do aim to work to filling data gaps, and hope to be able to pursue a GLMM approach at some point in the future

 

 

Comment: this is very coarse, but given the scale of analysis it is a good first step.  Of course, for selection to occur relative abundance data would be more appropriate to have a more informative result.

Response: we totally agree, but unfortunately as explained the abundance data provided in the studies we have used are not suitable or comparable. We hope our study can stimulate more attention to seed abundance in waterfowl in the future. Until now, the emphasis in diet studies has often been to quantify total seed biomass instead of seed abundance. This is essentially because of the auto-ecological focus of these studies, and greater interest in providing food resources for waterfowl populations that are major hunting quarry.

 

 

Comment: graminoids tend to be more abundant, so in the list they are diminished. This is the type of result that would change if relative availability data is considered.

 

Response: we are not sure if the reviewer is saying that graminoids have fewer species, but they are more abundant in nature (i.e. in the environment, not in the birds)? In any case, we agree that relative availability data would be preferable to the qualitative data we have been forced to use.

 

 

Comment: almost all parameters are significant, which makes me suspicious of the test. This may be a result of using simple chi-squared tests. The analyses and conclusions would be more robust if authors use a linear model for proportion data (GLMM)

 

Response: Most global tests are significant, meaning that the trait distributions differ between waterfowl and the NWPL list. This is not surprising, given the large sample sizes in most cases. However, when we get to the finer level of asking whether there are differences for specific trait categories, only a minority of tests are significant. The supplementary material and post-hoc tests shows that only 24 of 75 tests (32%) are significant.

 

 

Comment: rewrite - "... but aquatic species are more frequent"

 

Response: we have rewritten the subheading as “Terrestrial species dominate, but aquatic species are ingested more often than expected”. This is because aquatic species are not more frequent (i.e. they are <50% of species ingested), although they are ingested more often than expected from the NWPL list.

 

 

Comment: I suggest caution in the use of selection. THe fact they occur inside the bodies of ducks is not in itself evidence of "selection". Here you mean "more frequent" -  as the intentions of the ducks to ingest them cannot be determined in many of your cases. For selection to occur it has to be deliberate, not accidental.  Also, selection is evident in scenarios where the relative availability is measured at some relevant scale. our paper  on fruit selection can be a good reference for this process: https://doi.org/10.1098/rstb.2023.0128

 

Response: Upon first use of “selection” in this context, in the introduction, we have now added the sentence “(note, by “selection” we refer to any process of non-random association, which is not necessarily “intentional”)” in order to avoid giving the impression that we are implying a conscious, deliberate process. We suggest this is sufficient clarification, as we think our usage of the word “select” elsewhere is standard practice, and that in many cases there is no better synonym that we are aware of.

We would add that it is standard practice in ecology to talk about “selection” without providing proof that these non-random associations are “deliberate” in a meaningful way. Also we are dealing with a scenario of relative availability of different traits among plant species in the wetland environment in North America, and how this relates to the relative frequency of those traits in species recorded as ingested by waterfowl. We can see that the paper on density dependence by Carlo et al. is a good example of how an ideal paper on seed selection by waterfowl could be studied, in principle (and we have cited this paper further below in the text), but for now we only have the possibility of doing the coarse scale study that is represented by our manuscript. But our paper is at least a start….

 

 

Comment: yes, but this ican be evidence against "selection" for seeds as food, as they are after soemthing else: unless you can show that plants with fruits and seeds are more attractive (preferred) over the same plant without seeds and fruits - that will be a cool and simple experiment to perform in a controlled scenario!

 

Response: we get the point. However, we now think it is unlikely that geese or other grazing waterfowl are swallowing seeds by accident. Recently, we were shown unpublished experimental work in the Netherlands that showed strong selection by geese for plots of grass with seeds, when mixed with other plots of grass that did not have seeds. We have encouraged the authors to publish this important work, but have no further information. This new evidence would suggest that “the dry fruit is the fruit” rather than “the foliage is the fruit”, but for now we continue to consider Janzen’s hypothesis as a possibility (although we have changed the wording from “our result also supports the “foliage is the fruit” hypothesis” to “our result is also consistent with the “foliage is the fruit” hypothesis”).

 

 

Comment: significant in which direction? avoided? or more frequent

 

Response:  we clarified this by adding “(which ingested fewer tall plants than expected).”

 

 

Comment: you could add: "...  wetland environment to properly test for selection at the local scales (Carlo et al. 2025)."

 

Response: we have added this text as proposed, although we understand the citation is Carlo et al. 2024 (i.e. the paper cited in the earlier comment).

 

 

Comment: like the case of grass you mentioned in the results.

 

Response: we are not clear about what change is suggested here. I don’t think grass seeds are particularly small, compared with e.g. Juncus bufonius.

 

 

Comment: at which scale? I would think that at the scale of the community or ecosystems they inhabit.

 

Response: good question! We have added “at the continental scale across North American wetland ecosystems”

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript presents a novel analysis of plant traits influencing seed dispersal by North American waterfowl. The study is well-structured and addresses an important ecological question. However, several revisions are needed to improve clarity, methodological transparency, and alignment with journal standards. Below are specific recommendations:

1. Title：The title is overly verbose and contains redundant phrasing. "Trait-based selection of seeds ingested and dispersed by North American waterfowl". Maybe, The revised title emphasizes the core focus (trait selection) while removing redundancy and aligning with concise academic standards.

 

2. Abstract:

Lack of quantitative results and vague descriptors (e.g., "intermediate seed size"); Define seed mass categories explicitly; Clarify methodology:"Using chi-squared tests, we compared trait distributions between ingested plants and the National Wetland Plants List (NWPL)."

 

3. Introduction

Insufficient justification for Hypothesis II (plant height does not limit dispersal).

 Repetitive citations (e.g., Green et al., 2022 cited 3x).

 

4. Materials and Methods: Incomplete explanation of data exclusion criteria. Under "Literature search":"Swans were excluded due to insufficient data (n = 3 studies)."*

Missing discussion of trait data gaps (58% of NWPL species lacked traits).Add to "Wetland plant list" (Page 4):*"Trait data were unavailable for 58% of NWPL species, potentially introducing bias toward taxa with better-studied traits."*

Post-hoc test methodology unclear. Specify in "Data analysis" (Page 4):"Post-hoc analyses used Bonferroni-adjusted standardized residuals (α = 0.05)."*

 

5. Results

Figure 1 labels inconsistent with text (e.g., "Water regime" vs. "moisture requirements").

Non-significant results (e.g., plant height) lack context: Add to "Plant height" paragraph (Page 5):"Non-significant trends suggest minimal height-based selection across most guilds, likely due to secondary dispersal mechanisms."

 

6. Discussion

Ambiguity in "terrestrial species dominate."Rewrite (Page 6):"Although terrestrial species dominate the NWPL (68% of taxa), waterfowl disproportionately selected aquatic and semi-aquatic plants, highlighting their role as specialized dispersers."

Limitations section overlooks spatiotemporal variability. Add to "Limitations" (Page 7):

"Pooling data across regions and seasons may obscure local or phenological variation in seed ingestion patterns."

Replace with specifics (Page 8):"Future studies should quantify seed survival rates across traits (e.g., hardness, shape) and assess dispersal quality via germination trials."

Author Response

Comment: The manuscript presents a novel analysis of plant traits influencing seed dispersal by North American waterfowl. The study is well-structured and addresses an important ecological question.

Response:  Thank you for these positive comments.

 

Comment:    Title：The title is overly verbose and contains redundant phrasing. "Trait-based selection of seeds ingested and dispersed by North American waterfowl". Maybe, The revised title emphasizes the core focus (trait selection) while removing redundancy and aligning with concise academic standards.

Response: We have changed the title as suggested.

 

 Comment: Abstract: Lack of quantitative results and vague descriptors (e.g., "intermediate seed size"); Define seed mass categories explicitly; Clarify methodology:"Using chi-squared tests, we compared trait distributions between ingested plants and the National Wetland Plants List (NWPL)."

 

Response: Instructions for authors say “The abstract should be a total of about 200 words maximum”, so our room for details is limited. We have now added “using…..chi-squared tests” and “(1-10mg)” for seed size. With these changes, our abstract is now 276 words.

 

 

 Comment:    Introduction. Insufficient justification for Hypothesis II (plant height does not limit dispersal).  Repetitive citations (e.g., Green et al., 2022 cited 3x).

 Response:  We provided more justification for hypothesis II by adding “birds do not necessarily take seeds from the mother plant” and “(of any height)”. In fact, Green et al. 2022 is only cited twice, as is Green et al. 2023. This is because these are recent review papers, covering many of the key points.

 

 

Comment:    Under "Literature search":  "Swans were excluded due to insufficient data (n = 3 studies)."*

Response: We reworded the text as suggested.

 

 

   

Comment: Add to "Wetland plant list" (Page 4):*"Trait data were unavailable for 58% of NWPL species, potentially introducing bias toward taxa with better-studied traits."*

Response: We added the sentence “Since trait data were unavailable for 58% of NWPL species, there was potentially a bias toward taxa with better-studied traits.”

 

Comment: Post-hoc test methodology unclear. Specify in "Data analysis" (Page 4):"Post-hoc analyses used Bonferroni-adjusted standardized residuals (α = 0.05)."*

Response: We added the sentence as proposed.

 

 

Comment: Figure 1 labels inconsistent with text (e.g., "Water regime" vs. "moisture requirements").

 

Response: We have now modified the label of Figure 1, replacing “water regime” with “moisture requirements” to ensure consistency with the text.

 

 

Comment: Add to "Plant height" paragraph (Page 5):"Non-significant trends suggest minimal height-based selection across most guilds, likely due to secondary dispersal mechanisms."

Response: This sentence has been added.

 

 

Comment:  Discussion. Ambiguity in "terrestrial species dominate."Rewrite (Page 6):"Although terrestrial species dominate the NWPL (68% of taxa), waterfowl disproportionately selected aquatic and semi-aquatic plants, highlighting their role as specialized dispersers."

Response: This text has now been re-written, as proposed.

 

Comment: Limitations section overlooks spatiotemporal variability. Add to "Limitations" (Page 7): "Pooling data across regions and seasons may obscure local or phenological variation in seed ingestion patterns."

Response: We have added this sentence as proposed.

 

Comment: Replace with specifics (Page 8):"Future studies should quantify seed survival rates across traits (e.g., hardness, shape) and assess dispersal quality via germination trials."

Response: The text here has been rewritten and a very similar sentence to that proposed has been included.

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

the authors did good job for revision. I recommend accept for this ms.