The study aims to test whether wolf cranial form differs between historical and modern wolf populations of Fennoscandia. The manuscript presents an extensive background of genetic processes that lead to population growth and population changes, a variety of good discussion points, and a small but valuable dataset. I found the discussion of landmarking methods particularly interesting and useful. A correlation is established whereby the historical wolf crania appear to differ from the modern populations, but strong conclusions are drawn that are not (yet) supported by the analyses presented. I think with some work, the paper can address their goals better.
There are some minor technical edits needed, i.e. syntax, clarification of language. I made some suggestions throughout but especially for the Introduction. In several instances references are missing and needed, and I would recommend a closer review of this aspect overall.
However, there are key items in the Methods section that need clarification, explanation. Namely, the description of the distinction between historical and modern samples is confusing, and the purpose of some groups (i.e., undated wolves) is not entirely clear. These should be easy fixes, but important ones, as they strongly impact the results due to the small sample sizes in general.
The more pressing concerns relate to the main analyses. Namely, the influence of size has not been sufficiently addressed, and this weakens the interpretations the authors have made about temporal changes. Size as a likely driver in the cranial form differences observed in the PCA; the authors have already shown that size is significantly different for the Scandinavian sample. There is certainly a correlation with the temporal differences, but I am not certain the evidence is there to support it being the main driver/explanation. I have suggested analyses that I hope will help clarify the allometric effects.

Lastly, I would recommend the conclusions and goals be considered more carefully. For conclusions: the study reports skull form differences correlated with (according to Methods) the lowest population numbers and/or the onset of protection laws for wolves. The phrasing that the study provides evidence of morphological changes “resulting from genetic replacement” makes it sound like genetic analyses were performed, and may come across as a bit misleading, even though the authors provide a good background for genetic processes of population growth in the Introduction. For goals (lines 94-96), I think correlations are tested, not causal relationships, for example.

Specifically:
1. Last Introductory paragraph (lines 94-98): the grammar and syntax in the list of goals is off. It would make the paper much easier to follow if these were made clear from the start.

1. In line with above, I suggest rephrasing the goals to avoid claims of testing “causal” relationships. I think there is a strong case for testing correlations with bottlenecks/protection laws, but not causal relationships.

2. Methods: the description of the cutoff between modern and historical samples in unclear. Perhaps it is better to remove/edit line 114-115: “the dataset was divided into groups based on the year when wolf protection was implemented”, since there are different parameters for the two samples: for the Finnish wolves, it is 1923 = “lowest population numbers”. For Scandinavians it is 1983 (?), when wolf protection began or when their “comeback” began?

3. Results: due to the small sample sizes overall, I would encourage the inclusion of sample sizes in the boxplots.

4. Results: I recommend clarifying the descriptions of variation along the PCs. It may seem repetitive, but it is necessary to repeat which axis is being described and at which extreme. Ex.: line 205: it is not enough to refer to “…these individuals…”. The term “taller” is not something I have seen for cranial descriptions, perhaps “higher” would be more traditional/accurate. “Slope of the orbits” (line 204) is unclear, it would be helpful to describe the plane you are referring to. Line 209: “smaller” along a PC is often referred to as “smaller in magnitude”. These may seem like tedious edits, but I think they will make this section of your study more understandable.

5. The most significant concern with the analyses and results is that PC1 almost always correlates with size (allometry). A quick review of 3DGMM studies will confirm this. PC1 captures 32% of the total variation—this is a significant proportion of the variation that is likely attributed to allometry, especially since (a) within the larger sample (Scandinavian wolves), centroid size differs significantly between historical and modern populations; and (b) the statistical models. The influence of size needs to be discussed as likely driver in the cranial form differences observed. There is certainly a correlation with the temporal differences, but I am not certain the evidence is there to support it being the main driver. Therefore, interpretation of the results needs to be tempered to reflect this.

One way to clarify this further would be to perform a PCA on the residuals of the linear regression between the shape scores and centroid size. This would remove the size effect (CS) and you can then test if the resultant PCA separates the temporal populations.

Since there are differences in the size of different cranial bones between the historical and modern populations of both geographical regions, as seen in Figure 5, it would be informative to test if the allometry of the skull is the same (or different) for both regions. This can be done by, once again, regressing shape scores against CS and plotting the allometry for both regions separately, checking statistical differences in the slope of the allometries. This would help to understand whether the cranial shape differences are headed in the same direction for both countries, or not. How does this impact the paper? If allometry differs for the two regions, one can argue that the maintenance of allometry in the analysis is indeed informative.

Minor edits/suggestions:
Although describing the process of genetic change and population growth is important for the study, in my opinion, the introduction takes a bit too long to introduce the subject of the study: the grey wolf. Simply in the interest of capturing and focusing the reader, I recommend introducing wolves earlier. However, this is simply a suggestion.
Lines 52-56. In line with above, an entire introductory paragraph is devoted to the “founder effect”, but it is unclear how this is important/relevant to the rest of rest of the study. Perhaps it should be incorporated into the previous paragraph (?), if it is not meant to be a main focus of the study.
Lines 76 to79: The prior paragraph (L. 76- 78) describes “increased immigration”; the following paragraph begins with the fact that “few new immigrants arrived”. Perhaps the timeline needs to be clarified or specified, to avoid losing the reader.
Methods: Is there a reference to be cited regarding the establishment of Russian wolves as the source of migrant populations to Scandinavia? Re: line 31: “These specimens were
relevant in establishing the source population of the current Fennoscandian wolves...”. A citation for the GPA is also missing.

Line 89: “A previous study observed some differences between extinct and extant...”. It would be helpful to refer to the type of differences, which here is likely “morphological” (?).
Line 147: “brain cavity”. Most often, this is referred to as the endocranial cavity.
Lines 70 and 157: underscore after citations.
Line 188-189 Results: “Variance in and between the means of the two temporal groups was further visualized using the rnfelice/hot.dots function”...it should be clarified that is variance in the “mean shapes”...if that is the case.

Dear Authors,
Your manuscript has improved much since the last round. It is more focused and organized, thank you. However, there are many opportunities for improvement, some small and some more significant. I have done my best to highlight these areas, line by line. Overall, I think you have provided many tests/analyses, but for a very small dataset, and thus your conclusions need to be attenuated. I provide a list of all edits, small and larger. Best wishes,