Swallowtail Butterflies Use Multiple Visual Cues to Select Oviposition Sites
Round 1
Reviewer 1 Report
The authors have undertaken a study to investigate the visual cues that guide oviposition selection in Papilio xuthus. To test this, they carried out behavioural experiments and characterised various features of leaves to determine which cues were potentially attractive for egg laying. Their finding that green reflectance, brightness and low specular reflectance correlate with egg laying is complementary to the observation that Papilio favour younger leaves. Further the authors posit that Papilio may be repelled by vertically polarized light, which is an interesting idea. I found this study to be interesting and mostly well written. I suggest however some additions to the methods and clarifications in the discussion.
General comments
For completeness of the discussion, it could be useful to consider how important the role of UV light may be in the choices of leaves by Papilio, as there does not appear to be much UV in the experimental illumination (Fig 1. C). Although leaves predominantly reflect in the green, there is some evidence at least in birds, that UV contrast could assist in the detection of leaves (e.g. see https://doi.org/10.1038/s41467-018-08142-5). Since Papilio has UV sensitivity, is there any evidence that leaf characteristics in the ultraviolet may provide cues for leaf selection in Papilio?
L109 It is unclear to me how the eggs were removed, either simply mechanically or if there was a procedure to clean the leaves. Is it possible that the removal of eggs between trials affects the measured properties of the leaves (e.g. DoP or brightness?). Also, is it possible that there were residual olfactory cues from previously deposited eggs that acted as a deterrent/attractant to visiting females? Were there more failed attempts at ovipositing with increased numbers of trials with the same tree?
L143 It would be informative to include more detail about the calculations used for degree and angle of polarization using the three images at each angle of polarizer rotation. Were these images aligned for comparisons per pixel across the three polarizer rotation angles? If this method has been outlined more comprehensively elsewhere then the authors could reference that source.
Minor comments
I think the abstract could be improved by replacing the majority of the semicolons with commas.
L31 Formatting error, extra “.”
L36 Suggest adding “compound” before “eyes”
L37 “sensitive color vision” - do the authors mean in terms of acute spectral discrimination? For those not familiar with butterfly visual systems, it may also be useful to briefly outline the spectral sensitivity of Papilio xuthus. A brief introduction here to polarized light and in particular, polarization sensitivity (e.g. what its behavioural relevance for butterflies/insects is) would be useful, to introduce readers to the topic.
L42 It could be useful here to point out how Pierid butterflies are related to Papilio e.g. by adding Papilio (family: Papilionidae) and Pierid butterflies (family: Pieridae) or something similar, for those without knowledge of how these groups are related. There are also times where it is hard to determine if Papilio (e.g. L39 and L51) is in reference to the genus more generally or the species being studied.
L55 This may be better described as “significant correlations”
L65 It is unclear why females were housed in dimly-lit boxes. Is there any evidence that this could have affected the development of their visual system?
L75 Is this a particular Citrus species?
Figure 1 Missing T: Citrus tree in the legend.
Figure 2 I think “(upper)” and “(lower)” be “left” and “right”. Also, I think “(upper right)” should be “(lower right)”?
L118 It is not clear to me how soon the leaves were measured after the behavioural experiments began and therefore whether we would expect to see any of the features affected by age to change. In addition, if leaves were cleared of eggs between experiments, how were total egg numbers tracked between trials?
L122 Later the authors test height of each leaf as a potential factor driving egg laying but would this not be affected by the choice of “unfavored” leaves in similar positions to the “favoured” leaves?
L149 I think “POL” should be defined as polarization before its first use.
L157 Brightness and DoP were measured using the same angle of elevation (E60). I am not sure why this angle was chosen of the three. Is E60 equivalent to the approximate viewing angle of the butterflies?
L162 Were these three regions sampled randomly across the length of the leaf?
L187 Was detailed analysis only carried out only for Tree A as the general findings were only significant for this tree, as seen in Fig. 3 C?
L211 It is unclear to me which statistical test this is referring to and what level of significance is being used here.
L273 “indicat” typo
L294 I think Fig. 2B should be Fig. 2D.
L296 It’s not clear to me what “top left” is referring to here. Is there a missing inset showing all reflectance spectra of leaves? It would be informative to include a figure showing the spectra of all leaves, not just a few, to indicate the spectral variation across the leaves sampled.
L316 “This cue might help to identify target leaves against the background”. Due to the negative correlation this would assume that the background is more highly polarized than the leaves themselves. I am unsure if this would be the case in a natural scene. But, the idea that vertically polarized light might act as a repellent, rather than an attractive cue is compelling.
Author Response
Our answers is in the attachment file.
Author Response File: 
Author Response.docx
Reviewer 2 Report
This manuscript describes the results from probing the swallowtail butterfly Papilio xuthus for preferences in leaf choice for egg deposition. The question is clear, relevant, and interesting. The approach taken is reasonable and its execution meticulous. The author’s efforts to document methods and discuss findings in a rigorous manner are commendable. While this topic is a bit off my personal research radar, I enjoyed learning about this study and reading the manuscript. Overall, I find myself persuaded by their carefully drawn conclusions.
Major issues:
1.
How are the leaf preferences related to offspring (caterpillar) viability/survival? Would it have been a reasonable prediction that leaf size might be a top relevant feature with the rationale being that this is essential to assure food provision in close reach for the first instar caterpillar? This is addressed at the very end of the discussion (“ Pieris rapae preferentially select leaves with higher nutrient 350 content [21]. Caterpillars of P. aegeus develop better on young leaves [8].”). I think this information should be moved into the introduction to develop the full scope of the significance and motivation for this study.
2.
“To identify which visual cues female Papilio used for selecting oviposition sites, we
119 measured several features of leaves from trees on which more than three individuals each
120 laid more than twenty eggs.”
What was the rationale for this threshold?
How many trees ended up being below and above this threshold?
Would it be worthwhile to explore visual differences of above and below threshold trees?
Minor issues:
1.
“particularly well studied [1] [2]”
particularly well studied [1,2]
2.
“…from flowers[3-5]. . Diurnal butterfly…”
from flowers[3-5]. Diurnal butterfly
3.
…six classes of spectral receptor, each…
…six classes of spectral receptors, each…
4.
“Papilio’s vision employs at
35 least six classes of spectral receptor, each of which has a certain polarization sensitivity,
36 arranged in one of three combinations to form the ommatidia that comprise the eyes [14].”
I would break this up into one statement describing what ommatidia and their subtypes are, followed by the differential deployment of the different receptors.
5.
“olfaction and taste [17-19]”
olfaction and gustation [17-19]
6.
“, we hand-paired newly emerged females”
hand-paired or hand-mated?
7.
“We observed a female laying eggs on a potted Citrus tree in a small cage (105 x 60 x
76 70 cm, Fig. 1A).”
Since it is more than one observation:
We observed a femaleS laying eggs on a potted Citrus treeS in small cageS (105 x 60 x
76 70 cm, Fig. 1A).
8.
“Figure 1. Experimental setup. A. Photograph of the set-up. B. Top view of the set-up”
Replace setup with observation cage?
9.
“described in Stewart et al, 2015”
described in Stewart et al. (2015)
10.
“we mapped these values to the range 0~1;”
range of 0 to 1
11.
“range 0~180Ëš”
range of 0 to 180Ëš
12.
“Unlike our measure of green reflectance, which was concerned with the material properties of each leaf, these measurements aim to describe the leaf’s appearance in situ on the tree, under the lighting conditions of the experiment.”
Note the inconsistent use of past tense (was concerned) and present tense (aim) in the sentence. This applies to the methods and results section in general. I recommend putting all in past tense bc they describe unique events that are now in the past. But that may just be my personal preference.
13.
“We applied Wilcoxon signed-rank tests “
Add reference
14.
“We used Wilcoxon’s rank sum test to determine whether the various measures of leaf
175 features differed between the favored and unfavored groups.”
Add reference
15.
“Pearson’s correlation coefficient”
Reference
16.
Cite program(s) used for the statistical analyses
17.
“We analyzed in detail the oviposition behavior of the eight individuals
187 tested on Tree A (Fig. 3A,B), but when investigating which visual features were associated
188 with oviposition, we pooled data from 23 individuals across the five trees (Fig. 3C, Suppl.
189 2)”
I would break this sentence up into the 2 approaches. Give the specific rationale and results for each.
18.
“Ovipositing behavior of Papilio xuthus consists of five steps:..”
2 things:
The ovipositing behavior of P. xuthus consists of five steps:
19.
Consistently abbreviated P. xuthus after first introduction in the intro section
20.
“. These observation indicat that..”
Plural:
. These observationS indicatE that
Author Response
Our answers are in the attachment file,
.
Author Response File: Author Response.docx
Reviewer 3 Report
Oviposition is a behaviour rarely dissected at the level of individual butterflies and considering the sequence of actions. This paper presents interesting observations which in my opinion are very novel and exciting in that they track individual butterflies when selecting an oviposition site on a citrus tree. This offers an opportunity to gain better insight into the potential impact of visual cues on oviposition under natural conditions.
The authors report consistency in choice for a few of the many leaves and investigate what visual cues might guide the butterflies. Although data are mainly correlational, it is an interesting exploration of four visual cues that can influence the appearance of the leaves to the butterfly eye. I have a few comments which I hope the authors can easily address.
The authors acknowledge in the Introduction that leaves present chemical cues. In my view, it would be helpful to present the reader with a more detailed explanation as to what visual cues can offer that chemical cues do not although being so prominent in plant-insect interactions.
You discuss nicely in the opening paragraphs of the paper that butteflies are pollinators and their sensory capabilities well studied in the context of flower visits. However, when selecting plants and leaves for oviposition, the butterfly is also a herbivore. It makes important food choices for its larvae. It needs to find a protected site for the eggs, with a non-toxic surface to which it can securely attach its eggs. It would be helpful if the authors could expand these parts of the Introduction and Discussion a little, perhaps also elaborating on the potential benefits of visual cues for the different behavioural choices made when searching and selecting the substrate for oviposition at the host plant. For instance, the relationship between waxiness and polarisation of the leaf surface, its spectral cues, surface structure of the leaves that might generate tactile but also visual cues etc. There might be some relevant literature on this, but if not it could be interesting for the reader to be presented with more detailed explanations. It would be great if the authors could cover how butterflies might benefit from their sensory abilities, how cues from different modalities could inform the buttefly's choices once it has found the host plant, starting from visual and olfactory detection to contact with the surface and final decision to deposit the egg. An interesting paper is by Abe et al. 1981 Appl Entomol Zool 16(4), 493-496.
Many researchers have studied the relationship between leaf properties and herbivory which is likely to affect oviposition. Relevant here seems Bergstrom et al. Chemoecology 1994 reports how butterflies might be able to distinguish young leaves from leaf volatiles (including Citrus), Stefanescu et al. Eur J Entomol 2006 looks at greenness. Besides tactile and chemical cues there is also water evaporation in leaves from transpiration and thermal irradiation which insects can perceive and which could change between leaves of different ages or positions on the tree (e.g. Gates 1968 Ann Rev Plant Physiol, Dakhiya & Green 2019 New Phytologist, Syversen Annals of Botany 1982, Hirano 1931 Botanical Gazette).
I was slightly confused by the repeated statement "different butterflies selected the same leaves". Please could you state in the Methods that five trees were used, how they were exchanged between tests (though this is later mentioned in the Discussion, it should be very clearly described early on, in the Methods, to help the reader understand the paper). Given the discussion (2nd para), I wonder why the authors have not included tree and preceding visitors as a random factor in the statistical analysis. As mentioned, the preference ranking (Figure 3C) relates to just one tree but the conclusions drawn are quite broad (first para Discussion). This set of data (Figure 3C) should be interpreted with more caution, also in view that previous visits and/or egg depositions by other butterflies could have left chemical cues that the next butterfly might sense prior to or during landing. Removing eggs might not be sufficient. For instance, it is known that bees sense footprints of other insects and will avoid landing on previously visited flowers (Eltz J Chem Ecol 2006). I would suggest that the authors discuss this limitation of their study more critically. Typically, behavioural studies deploying natural stimuli will have a range of these to prevent pseudoreplication. If the number of replications (number of trees) is low, like here, it might not reflect properly the assembly of leaves that might exist in the natural population of Citrus trees. Whilst I don't think this is a great problem, it is a limitation of the study nevertheless that should be acknowledged more clearly.
Discussion 4.1, para 2 - This should be moved to the Methods section, and the conclusion in the first paragraph toned down.
Figure 3A - It would be innovative and valuable to provide quantitative data showing how the butterflies utilised the area of the cage, how large the space was they covered. The sequence of successful and unsuccessful visits could also reveal whether potentially the butterflies had a spatial strategy for concentrating on particular area of the tree. Such analysis would help to understand whether butterflies viewed and approached the tree from many different directions and to show that a limited capacity of flying around the tree was unlikely to be a determinant of leaf choices.
Discussion 4.1, para 3 - It's not clear what you mean by "descended rather directly". The arguments presented in the third paragraph of 4.1 on the use of olfactory information strike me as quite speculative. As suggested above, the quantitative analysis of the flight data (e.g. descent speed and duration of leaf visits) would not only enhance the novelty and contributions from this work but also provide better support for these conclusions. It would be very valuable to understand how little time a butterfly has to assess sensory information during a successful visit. There is some work that shows how quickly insects can process olfactory information. Perhaps you could refer to these when discussing the lower likelihood of making the choice for a particular leaf during the descent rather than prior to it, which seems to be what you suggest.
Figure 4A: Height could potentially be confounded with temperature and evaporation, as these leaves are not only closest to the butterfly when it flies around the ceiling but also closer to the lights, and presumably it is warmer up there.
Figure 4C: Flatness is really interesting one, and the authors discuss this well in 4.2 (first para). I could not quite follow how flatness might affect visual cues, or whether it is completely unrelated (as suggested in the last para of the Discussion). Perhaps a sentence or two could be added in 4.2 to explain the rationale for measuring leaf properties and visual cues.
P. 1, line 40: It would be useful to mention already in the Introduction that the development of the larvae will depend on the quality of leaves the butterfly selects (bringing forward the sentence and reference from 4.3).
Figure/Table captions - please could you add sample sizes.
Author Response
Please see the attachment file.
Author Response File: Author Response.docx
Round 2
Reviewer 3 Report
The revised version has clarified many points, thank you.
