Self-Compatibility and Reproductive Success of Oenothera drummondii subsp. drummondii: Is It Similar between Native and Non-Native Populations?
Round 1
Reviewer 1 Report
General:
The manuscript is interesting and very meticulously written.
It analyses the differences of native and not native Oenothera drummondii subsp. Drummondii populations. Moreover, a new tool in the form of a comprehensive integral reproductive success index (IRSI) was built in the study. This can help in other studies and other plant species. The results are not revolutionary but it is clear that a success of populations is not easily explained so the results and conclusions are versatile.
The authors have recently published another paper (Gallego-Fernández and García-Franco. 2021. Gallego-Floral traits variation in Oenothera drummondii subsp. drummondii across a wide latitudinal about treating different elements. The authors conclude that floral traits variation of O. drummondii throughout its native and non-native distribution is the result of the latitudinal gradient and the local population environment. The same conclusion is related to the present manuscript. In the present manuscript this should be connected and stated.
Specific comments:
Introduction
How the author identified the species from all locations as Oenothera drummondii subsp. drummondii Hook. and not Oenothera drummondii Hook?
Materials and Methods
Since the most of the results and conclusions are related to different latitude it would be of help to have a map with locations (six native populations and four non-native populations - Spain, Israel, South Africa, Australia).
Line 493 non-naive, correct in non-native
Author Response
Reviewer 1.
The manuscript is interesting and very meticulously written.
Reviewer 1. It analyses the differences of native and not native Oenothera drummondii subsp. Drummondii populations. Moreover, a new tool in the form of a comprehensive integral reproductive success index (IRSI) was built in the study. This can help in other studies and other plant species. The results are not revolutionary but it is clear that the success of populations is not easily explained so the results and conclusions are versatile.
The authors have recently published another paper (Gallego-Fernández and García-Franco. 2021. Gallego-Floral traits variation in Oenothera drummondii subsp. drummondii across a wide latitudinal about treating different elements. The authors conclude that floral traits variation of O. drummondii throughout its native and non-native distribution is the result of the latitudinal gradient and the local population environment. The same conclusion is related to the present manuscript. In the present manuscript, this should be connected and stated.
Response: Thanks for the suggestion, we have included some lines where the relationship of the floral characteristics of native and non-native populations with latitude is mentioned (L403-407).
Reviewer 1. Introduction, How the author identified the species from all locations as Oenothera drummondii subsp. drummondii Hook. and not Oenothera drummondii Hook?
Response: Several authors point out that O. drummondii subsp. drummondii is naturally distributed along the coasts of the Gulf of Mexico, from North Carolina, United States, to the state of Campeche, Mex. (eg Dietrich et al 1998). Based on this, and bibliographic and herbarium studies (see references 24, 25, 26, 27) we identified several native and non-native populations of the study species. Later, with the help of colleagues from other countries, seeds were obtained from some native (SPA, BOL) and non-native (LEZ, MAH, MUI) populations. In the acknowledgments section, the names of the colleagues who helped in the collection of the seeds are indicated.
Reviewer 1. Since most of the results and conclusions are related to different latitude it would be of help to have a map with locations (six native populations and four non-native populations - Spain, Israel, South Africa, Australia).
Response. Thank you for the suggestion. We have already included the map (Figure 1)
Reviewer 1. Line 493 non-native, correct in non-native
Response: Done. Sorry for the mistake.
Reviewer 2 Report
Diversity 1344678
The above paper examines pollination in native and non-native cultivars of Oenothera drummondii growing in a greenhouse in Spain. The species is native to southern USA and Mexico. The main objective of the study was to determine whether the native and non-native populations have similar or different modes of pollination. Specifically, are the non-natives mainly self-pollinated due to the absence of the insect pollinators found in the Americas (and no suitable alternatives).
Overall, the experimental work and data analyses appear sound. However, the paper is not suitable for publication in its present form. The presentation and interpretation of the data are confusing at times. It is difficult to determine whether there are significant differences in the mode of pollination between the two groups of plants. In my mind, the two groups seem to be similar.
There is a lot of data presented, and probably too much discussion. These both distract from the key messages and findings of the study.
Some specific comments are provided below.
Is there a lack of potential pollinators (bees, etc.) across Europe for the species? There are numerous pollinators in this environment for wild and crop plants. Is there something specific about this species and its pollination and fertility?
The authors fertility index, IRSI makes no sense to me and should probably be deleted. Indicate fertility can be assessed using different methods. Probably fruit set (%), total seed mass and seed germination (%) could be used.
Some of the data seem secondary and could probably be deleted (fruit length, fruit width, first germination, Timson index, empty seeds (%) and non-germinated seeds (%)). This will help the authors/readers focus on the key issues of the study.
Figures 1-3 seem to be duplicates of the data presented in Tables 1-3 and should be deleted. I didn’t understand the relationship between fertility and latitude in these figures, especially when no data on latitude were presented.
Figure 4 shows the key data for the study. Cross pollination improves fruit weight in the natives but not in the non-natives. This is what the authors were predicting (natives better after crossing). Seed set, seed mass and individual seed weight are similar with selfing and crossing in the two groups. Final germination shows a different trend. Germination is promoted by crossing in the non-natives but not in the natives. There are different responses (selfing versus crossing) depending on the data collected. No data are presented on fruit set (%) for the two groups of plants in Figure 4.
Figure 5 should be deleted. IRSI is a confusing term. In any case, there are no significant relationships here (P > 0.05).
The ‘Conclusion’ is very weak and really does not answer the key question asked by the authors. It is apparent that the need for cross pollination in the two groups of plants (native versus non-native) depends on the type of data collected. Maybe this is the main finding of the study.
Author Response
Reviewer 2
Reviewer 2. The above paper examines pollination in native and non-native cultivars of Oenothera drummondii growing in a greenhouse in Spain. The species is native to southern USA and Mexico. The main objective of the study was to determine whether the native and non-native populations have similar or different modes of pollination. Specifically, are the non-natives mainly self-pollinated due to the absence of the insect pollinators found in the Americas (and no suitable alternatives).
Overall, the experimental work and data analyses appear sound. However, the paper is not suitable for publication in its present form. The presentation and interpretation of the data are confusing at times. It is difficult to determine whether there are significant differences in the mode of pollination between the two groups of plants. In my mind, the two groups seem to be similar.
There is a lot of data presented, and probably too much discussion. These both distract from the key messages and findings of the study.
Response: Thank you very much for the comment. Sorry if the interpretation of the results was confusing for you. We think that by presenting the different characteristics of fruits and seeds obtained in the two pollination treatments and in each population, it would be clearer.
As the reviewer says, many of the characteristics studied do not show differences between native populations. However, it is very clear that the final germination and the germination rate (Timson's index) are different among native populations, and that they also had a positive relationship with the altitudinal increase. We consider that in this way the questions 1, 2, and 3 are answered. While the graphs in Figure 4 (now fig. 5) address questions 4 and 5 and show that the characteristics of the fruits and seeds obtained by the two treatments vary between the set of native and non-native populations. This figure also shows that the final germination and the germination rate are little variables in non-native populations, and they have a tendency to be higher in cross-pollination. And finally, Figure 5 (now fig. 6) presents both the results of the self-compatibility index calculated for each of the treatments and population, and the proposed ISRI that integrates the results of the set of fruits, set of seeds, and germination. With this, we answer question 6. The figure shows that the self-compatibility (SCI) of native populations, resulting from the different pollination treatments, increases with the latitudinal increase, and also shown that the non-native populations do not follow any pattern. While the IRSI has a slight tendency to decrease with the latitudinal increase in the native populations, non-native populations show an IRSI similar to the native populations located in the higher latitudes.
On the other hand, the reviewer's comment is appreciated, where he mentioned that the data is repeated in some tables and figures. We have removed Tables 2-3 and put them in Appendix B. In this way information, such as the comparison between the characteristics of fruits and seeds obtained by the two pollination treatments, in each of the populations, remains available. Also, Table 1 was removed, as we accepted the other reviewer's suggestion to present the geographic distribution of the studied populations on a map.
The reviewer considers there is a lot of text in the discussion section. However, we do not have any clear guidance on the section or sections where specifically the discussion is repetitive or out of place. We have reviewed the text of the discussion, and we could not identify any particular point that could be modified or that could be deleted with which the text is clearer.
Reviewer 2. Some specific comments are provided below.
Reviewer 2. Is there a lack of potential pollinators (bees, etc.) across Europe for the species? There are numerous pollinators in this environment for wild and crop plants. Is there something specific about this species and its pollination and fertility?
Response: In the discussion (L424-434) it was mentioned that the authors had the opportunity to observe both bees and sphingids visiting the flowers of the population studied in Spain (DIQ) and that bees have also been recorded in the population of Israel (LEZ). To date, we do not specifically know the identity of the floral visitors in the populations studied. That is why in this work we point the floral visitors as large groups, Hymenoptera and Sphingidae. Also, we propose that it is necessary to carry out specific work both in native and non-native populations to be able to know exactly the floral visitors, to understand their contribution both in self-pollination or cross-pollination.
Reviewer 2. The authors fertility index, IRSI makes no sense to me and should probably be deleted. Indicate fertility can be assessed using different methods. Probably fruit set (%), total seed mass, and seed germination (%) could be used.
Response: As we mentioned before, we are sorry if the reviewer feels that the explanation for the IRSI has not been clear. We have cited several studies in the text (52,53,54,55,56) where only some isolated indices are considered as reproductive success in plants. We think that plant reproduction can be seen as a set of isolated events, but all together shape up the reproductive event. That is why we raised the IRSI.
Reviewer 2. Some of the data seem secondary and could probably be deleted (fruit length, fruit width, first germination, Timson index, empty seeds (%) and non-germinated seeds (%)). This will help the authors/readers focus on the key issues of the study.
Response: We include the characteristics of fruits and seeds since each of them integrates the reproduction of plants. On the other hand, the response to manual pollinations could be reflected in some of the selected parameters. For example, response to incompatibility can be shown in smaller fruits, fewer seeds, or even less weight in fruits and seed mass. On the other hand, the selfing or crossing can also have an effect on the final germination, the germination rate, or the number of non-germinated seeds (but with embryo). Changes in these parameters may be differential responses of pollination treatments in each population. That is why we consider the information on these factors is important to include.
Reviewer 2. Figures 1-3 seem to be duplicates of the data presented in Tables 1-3 and should be deleted. I didn’t understand the relationship between fertility and latitude in these figures, especially when no data on latitude were presented.
Response: We appreciate the reviewer's comment. The information on the latitudinal location of the populations was presented in Table 1, but following the suggestion for reviewer 1, we have included a map that shows the geographic distributions of populations. On the other hand, in lines 209-211, we have mentioned that the charts include the names of the populations in order of latitudinal increment, following our previous work [34] in which the same latitudinal relationship scheme was handled.
On the other hand, we accept the suggestion of reviewer 1 to include a map, which can help to visualize the geographical distribution of the populations studied. Also, previously we described that the reviewer's comment was accepted, and Tables 2 and 3 were removed from the body of the text, and they were included in Appendix B.
Reviewer 2. Figure 4 shows the key data for the study. Cross-pollination improves fruit weight in the natives but not in the non- natives. This is what the authors were predicting (natives better after crossing). Seed set, seed mass, and individual seed weight are similar with selfing and crossing in the two groups. Final germination shows a different trend. Germination is promoted by crossing in the non-natives but not in the natives. There are different responses (selfing versus crossing) depending on the data collected. No data are presented on the fruit set (%) for the two groups of plants in Figure 4.
Response. Thanks for the comment. As mentioned by the reviewer, Figure 4 shows the comparison of pollination treatments between groups of populations with different origins. We do not present the data of the fruit set by their origin (native and non-native), because in both cases of pollination it was very high, close to 100%, so there was no point in presenting these data in a graph. Please see Table A1.
Reviewer 2. Figure 5 should be deleted. IRSI is a confusing term. In any case, there are no significant relationships here (P > 0.05).
Response. Thanks for the comment. We understand that the approach to this index, being new, can be a bit complicated. We try to explain it as clearly as possible by pointing out two examples showing how values can change as a result of the changes in fruit set, seed set, or final germination (L220-236). On another hand, the reviewer is right, the IRSI was not related to the latitudinal increment. The interesting thing about the values is that they are similar between the populations (not the same) despite they are structured with different values of fruit set, seed set, and final germination in each population. In light of our index, all populations have the same overall reproductive success, but with a different contribution from fruits and seeds.
Reviewer 2. The ‘Conclusion’ is very weak and really does not answer the key question asked by the authors. It is apparent that the need for cross-pollination in the two groups of plants (native versus non-native) depends on the type of data collected. Maybe this is the main finding of the study.
Response: Thanks for your comment. We have rewritten some parts of the conclusions to highlight the most important results of our work.
Round 2
Reviewer 2 Report
Diversity 1344678 Revised
The revised paper is much improved and many thanks to the authors.
I still have a few issues with the paper which I would the authors to address if possible.
In the ‘Abstract’, what is the regression with latitude that the authors are referring to?
Information on fruit length and fruit width in Figures 2 and 5 is secondary and not needed to calculate reproductive success. These data should be deleted. Success is likely to be more closely related to fruit weight.
I still do not understand the regressions shown in Figures 3 and 4 (individual seed weight versus latitude and final germination and Timson index versus latitude). It is not possible to run a regression between a variable such as seed weight (y-axis) and categorial values (x-axis as populations of a species). There is no indication of the latitude of each of the populations. These data would need to be shown in a separate figure.
The earlier version of the paper indicated that there were only small differences in self-compatibility index (SCI) across the populations. I don’t see why the authors want to determine the relationship between this parameter and latitude (Figure 6). As the data show, there is no significant relationship between SCI and latitude (P = 0.12). I recommend that these data/regressions be deleted.
Information needs to be provided on Integrated Reproductive Success Indices for the different populations and the two groups (natives versus non-natives), as displayed in Figures 2, 3, 4 and 5.
The relationships between IRSI and latitude in the self- and cross-pollinated treatments (Figure 6) need to be further explored using the whole data set. The data with the populations from 30 latitude or higher need to be integrated with those from 30 latitude and lower (new models if possible).
Author Response
The revised paper is much improved and many thanks to the authors.
I still have a few issues with the paper which I would the authors to address if possible.
Reviewer 2. In the ‘Abstract’, what is the regression with latitude that the authors are referring to?
Answer: Thank you for the observation. Sorry for the mistake, the sentence is incomplete. We refer to the positive and non-significant trends of the characteristics of fruits and seeds with the latitudinal increase. We have already corrected the error (L16-17)
Reviewer 2. Information on fruit length and fruit width in Figures 2 and 5 is secondary and not needed to calculate reproductive success. These data should be deleted. Success is likely to be more closely related to fruit weight.
Answer: Thank you very much for the comment. Sorry if we couldn't explain why we think the measurements on the fruits are important. The reviewer is right the size of the fruits (length and width) is not necessary to calculate the reproductive success of our study species. However, the size of the fruits may be related to the number of seeds and their weight. It has been seen that larger fruits may contain a greater number of seeds, or they may have greater weight (Primark 1987; among others). On the other hand, the width of the fruits was one of the variables that had a positive relationship with the altitudinal increase. In Fig 2, it can be seen that the width of the fruits increased with the latitudinal increase in the native range of distribution. Whereas figure 5b shows that the length of the fruit was significantly greater in the two pollination treatments in the native populations; while in fig 5c, it can be seen that the width of the fruits was greater in non-native populations. However, data about fruit size have been removed.
On the other hand, the number of seeds and their weight, can also determine, in a certain way, the success of germination. The hope is that seeds with greater weight have a greater probability of germination (Leas-Aráujo et al 2020), what directly influences the species fitness. In our case, the weight of the fruits had a good relationship (r> 0.5) with the other characteristics of the fruit (size) and the seeds. That is why we think that showing the size of the fruits, as well as the number of seeds and their weight, allow us to observe the chances of success resulting from each pollination manipulation. However, these data were also removed in the new version.
Please note, that removing several charts, the figure number has change. Also, we remove these data from the Tables.
Reviewer 2. I still do not understand the regressions shown in Figures 3 and 4 (individual seed weight versus latitude and final germination and Timson index versus latitude). It is not possible to run a regression between a variable such as seed weight (y-axis) and categorial values (x-axis as populations of a species). There is no indication of the latitude of each of the populations. These data would need to be shown in a separate figure.
Answer: We are sorry if we were unable to communicate the idea of how the graphs that include the data from the regression analyzes were structured. We try to explain it on lines 216-218 (Round 1). In a previous work (Gallego-Fernández & García-Franco, Flora, 2021), we studied the clinal variation of the floral characteristics of native and non-native populations of O. drummondii. To show this relationship, we substitute the geographic location of each population (latitude) for the population acronyms. The justification for doing this was that each population have different geographic location, and by putting the acronyms in the graphs following the latitudinal increase, the explanation of these results would be clearer. Based on the above, we propose the same construction of the graphs. We have rewritten this explanation again in the new version of the manuscript (L215-218), we hope that the explanation will allow us to understand the construction of the graphs.
L216- . “Since each native population represents a latitudinal position along the gradient of distribution of the study species (Figure 1), the charts presented in the result section include the population acronyms in instead of the latitudinal data. Acronyms follow the order of latitudinal increase (see Table A1).”
On the other hand, the significant results of the regression analyzes were included in the same graphs where the characteristics of fruits and seeds are compared. We did this to simplifying the results, and thus not having to include new graphs with the regressions, which can take up a lot of space. To clarify what is presented in the graphs we rewrite a part of the figure captions.
“Population acronyms follow the order of latitudinal increase (see Table A1). Dotted lines representing significant regression analyses but are not in scale with the latitudinal increase.”
Finally, considering the new comment of the reviewer, we have again included the Table with the geographic location of the populations in Appendix B. We hope that this information will answer the reviewer's concern.
Reviewer: The earlier version of the paper indicated that there were only small differences in self-compatibility index (SCI) across the populations. I don’t see why the authors want to determine the relationship between this parameter and latitude (Figure 6). As the data show, there is no significant relationship between SCI and latitude (P = 0.12). I recommend that these data/regressions be deleted.
Answer: Thank you very much for your comment. In Figure 6 we include the self-compatibility index to show that the reviewer mentioned. On the one hand, the SCI is not related to the latitudinal increase, and it is similar in all populations along the latitudinal gradient, both native and non-native. That is, all populations are highly self-compatibility, but reproductive success changes. The IRSI of each of the pollination manipulations in native populations shows a tendency to increase with latitudinal increment. Low-altitude populations (OJO and SON) are those with the least integral reproductive success, whereas the population located at a higher latitude has the highest success, both as a result of both self- and as cross-pollination. In addition, Figure 6 also highlights that non-native populations also presented high IRSI values.
We think that these results show the change in the reproductive success of O. drummondii in its native range and high similar success in non-native populations. The latter, which may confer O. drummondii more reproductive and colonization possibilities in non-native sites.
Reviewer 2. Information needs to be provided on Integrated Reproductive Success Indices for the different populations and the two groups (natives versus non-natives), as displayed in Figures 2, 3, 4 and 5.
Answer: Thank you for the comment, we have included the IRSI values of each population in Table A3, so that data will be available to readers. On the other hand, thank you very much for the suggestion, we built a figure where we compare the IRSI values between native and non-native groups.
Reviewer 2. The relationships between IRSI and latitude in the self- and cross-pollinated treatments (Figure 6) need to be further explored using the whole data set. The data with the populations from 30 latitude or higher need to be integrated with those from 30 latitude and lower (new models if possible).
Answer: Thanks for the comment. We are sorry but we do not get clear about your suggestion. We understand that the reviewer suggests the analysis and comparison between the populations grouped by origin, you can answer this comment. We have carried out this analysis, and now a chart is included, which shows that the IRSI is higher in non-native populations. With these data, the discussion is reinforced where it was mentioned that O. drummondii presents reproductive characteristics that may favor it in non-native sites.
Final comment
Please note that the document has changed. By eliminating the data on fruit size (length and width), the weight of the seeds and the number of non-germinated seeds suggested by the reviewer, the text was reduced a little in the method and results. The figures also changed when removing the fruit and seed data. However, the discussion did not change a lot, because the deleted data does not support it.
Round 3
Reviewer 2 Report
Diversity 1344678 R2
Many thanks to the authors for their response to the reviewers’ comments.
The paper is close to acceptance. I provide two suggestions.
For Figures 3a and 3c, provide a second x-axis showing the latitude of the different populations.
The start and end points for the regressions shown in Figure 5 seem rather arbitrary and do not include all the data. For the black data points, show an analysis/regression for all the data points – the slope of the regression will be close to zero (no relationship, no increase or decrease). For the orange data points, there are two linear relationships – one for 1 to 7 (increases) and another for 7 to 10 (decreases). For the blue data, there are two linear relationships – one for 1 to 8 (increases) and another for 8 to 10 (decreases).
Can the title be amended to indicate the main result of the paper? The current one just asks a question.
Author Response
Reviewer. For Figures 3a and 3c, provide a second x-axis showing the latitude of the different populations.
Answer: Thanks for the suggestion. In the figures, we have included below the acronyms of the populations the latitude data of each population. We consider it inappropriate to put it on a secondary axis, since, as we have explained previously, the populations follow a positive latitudinal gradient, and a secondary axis with latitude is likely to cause confusion.
Reviewer: The start and end points for the regressions shown in Figure 5 seem rather arbitrary and do not include all the data. For the black data points, show an analysis/regression for all the data points – the slope of the regression will be close to zero (no relationship, no increase or decrease). For the orange data points, there are two linear relationships – one for 1 to 7 (increases) and another for 7 to 10 (decreases). For the blue data, there are two linear relationships – one for 1 to 8 (increases) and another for 8 to 10 (decreases).
Answer: We appreciate the observation. We had not included those regression lines previously, as they are not significant and could possibly cause confusion. However, following the reviewer's recommendation, we have included the regression trend lines in the figure and the regression values in the boxes. In addition, we modified the wording of the results to include this data.
“On another hand, in the non-native populations a non-significant negative trend with the latitudinal increase; while the IRSI of both hand-pollination treatments showed and the SCI of all non-native populations were as high as of the two native population located at the highest latitude (SPA and BOL).”
Reviewer: Can the title be amended to indicate the main result of the paper? The current one just asks a question.
Answer: Thanks a lot for the suggestion. However, we stated the title as a question because our main aim was; to know if native and non-native populations of Oenothera drummondii have the same mating systems and reproductive success. Also, this title reflects the objectives of the work.
Author Response File: 
Author Response.pdf
