Identifying Morphs of the Yellow-Legged Hornet (Vespa velutina) and Other Pests of Quarantine Importance with Geometric Morphometrics

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript is very useful and presents some interesting methods for answering these questions.  It should be published.

This is not necessary for this manuscript to be published, but I would suggest for the authors to consider the following ideas that might help with the overall goals of this work.  

Can you get a significant result using morphometrics only on veins visible "above the fold" of the forewing?  If so, then this would open a whole new world for using citizen science observations in this type of study.  There are nearly 20,000 photos of Vespa velutina available on inaturalist (https://www.inaturalist.org/observations?place_id=any&subview=map&taxon_id=119019), including about 120 from Korea.  Since you crop and rotate the wings for the analysis anyway, you might get at least a few usable results from photos of live individuals. 

Comments for author File: Comments.pdf

Comments on the Quality of English Language

it's mostly fine, a few minor issues.

Special attention needs to be given to formatting and presentation of scientific names. They always need to be italicized and have the genus and species abbreviations when used: "V. v. ..."  This needs to be thoroughly checked in the manuscript.

Author and year also need to be written out for the first mention of any scientific name.

Author Response

Thank you for your comments an suggestions. All of them have been incorporated into the track changes version I am including here as an attachment.

In addition, regarding your question: whether it is feasible to use only landmarks visible in citizen science photos for provenancing Vespa velutina.  Based on our observations, it may be possible to set at least 15 landmarks in the forewing from images of specimens in nature and at rest (vespids fold their wings when not flying) and this may increase the number of images available for analysis, but this would need to be tested carefully to ensure that parallax from the photos did not introduce noise to the geometric morphometric data set and that the subset of landmarks capture enough differentiation to separate populations.  As the reviewer suggests, such a study is beyond the scope of the present manuscript (especially without appropriate samples of photographs from the native ranges of the V. velutina color morphs, but it is an interesting possibility for a future study.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Colleagues,

I have carefully read your manuscript entitled "Identifying morphs of the yellow-legged hornet (Vespa velutina) and other pests of quarantine importance with geometric morphometrics". I believe it presents a reliable technique that can be used to identify individuals of V. velutina and other Hymenoptera of uncertain origin. However, I would like to have your comment on the usability of multivariate ratio analysis (MRA) that involves both LDA and PCA (see Baur and Leuenberger 2011; Baur et al. 2014 etc.), since this technique proved effective for distinguishing between cryptic species of Hymenoptera. In addition, I have a few technical suggestions aimed to improve the manuscript (please see the attached file).

Comments for author File: Comments.pdf

Author Response

Thank you for your comments and suggestions, all of them have been incorporated in the attached version with track changes.

In addition, regarding your comment on MRA, we think Multivariate ratio analysis is a useful technique for certain kinds of data and situations.  It is especially powerful when only two individuals are available for comparison.  A series of measurements can be taken on each individual with each measurement converted into a ratio of individual 1 to individual 2.  These ratios will be > 1 when the first individual has a larger measurement and smaller when individual 2 has a larger one, and that pattern can be analyzed with MRA in the context of other data to determine whether the two individuals are significantly different.  MRA can also be used in cases where the data set itself consists of ratios (e.g., ratio of forewing length to width, ratio of thorax length to width, head length to width, etc.).  Ratios have peculiar statistical properties so MRA is appropriate for such data sets.  Because geometric morphometric coordinates have their own peculiarities related to Procrustes superimposition, which places them in a special coordinate system, they cannot be converted to ratios and so MRA cannot be applied to them.

Author Response File: Author Response.pdf