Rule et al. provide an interesting manuscript arguing that a fundamental pattern in baleen whale evolution may have a geographic spin: that extremely large body size first originated in the Southern Hemisphere, and that the presumed pattern is biased from a more northernly weighted record. This suggestion is novel, but I think there are major and minor concerns that the authors needs to address first.

Major concerns:

First, I need to raise the false conflict flag because the issue sits at the core of the paper's relevance for PRSB. The authors present the rise of mysticete gigantism as a debate, when I'm not sure that's a fair claim: Slate et al. 2017 (the primary analysis underpinning this argument)'s hypothesis of a "delayed rise" is a fundamentally an argument about a clade-wide shift towards extremely large body size (i.e., >20 m) ALONG with the extinction of smaller body sizes. Refs. 3, 7, 8 present singleton records that aren't the same thing as presenting an analysis of clade-wide trait evolution; in fact, Ref. 8 presents a specimen that falls cleanly in a confidence interval of Slater et al., hardly an outlier. More importantly, Slater et al. (2017) explicitly tested for the appearance of large sizes mysticetes (10 m, not 20 m) as early as the late Eocene (e.g., Llanocetus in Ref. 7), and this potential early appearance still did not falsify a "delayed rise" clade-wide pattern scenario. Moreover, it's hard to say that the authors differ that much from Slater et al. (2017) reading their own description of their results on ll. 104-106: "Mysticetes were mostly small for the first half of their evolutionary history and then gradually became larger after ~16 Ma (Figure 3a,b). Small baleen whales abruptly decline around 4.5 Ma." How are these findings (e.g., Figs. 2a, 3a) in opposition to Slater et al. (2017)? They seem to confirm it, largely.

Second, the geographic issue balances on a key assumption: the high endemism in the fossil record that is not a result of sampling bias (e.g., geologic megabias or otherwise). On l. 123 the authors explicitly make this endemism an assumption, attributing it to "a result of greater regional productivity." That's fine, but is the high endemism in the fossil record (i.e., exclusively southern fossil species) really true? The authors should do a comparison of fossil vs. extant N/S comparison, rather than an explicitly fossil-based one. Why? Extant mysticete genera are effectively global in distribution, with endemism only in monotypic Balaena and Caperea; also, we only tend to see hemispheric separation in sub-species groups for larger rorquals (Balaenoptera spp.), although there are plenty of documented transequatorial individuals. The extant distribution matters because it's the only basis for making a claim about the endemism documented in the fossil record, which is largely monotypic and centered on a generic level for analyses. I wouldn't be surprised if we don't see cosmopolitan taxa until the Pleistocene, but you'd need to eliminate another source of bias (next item).

I think the authors need to go beyond PBDB record and lineage-counting metrics to evaluate geologic megabiases in the rock record. For example, many stem mysticetes appear highly restricted to specific hemispheres (e.g., Llanocetidae in the south, Aetiocetidae in the north -- the authors may wish to point this out), but the amount of rock outcrop (by area or volume) in the Oligocene is effectively the same in both hemispheres. By the Neogene, there is substantially more rock outcrop in the Northern Hemispheres, and I suspect that's why there's a heavy Northern bias: more rocks, more collecting, more time spent by more researchers. Those biases are well documented (including by one of the co-authors), and I think it needs to be addressed in an explicit way: show me that rock, collecting, and researcher effort biases aren't equally possible drivers of the discrepancy between northern and southern fossil mysticete sample sizes.

Lastly, agency and mode: the authors are arguing that an onset of size increase in the Early Miocene was driven by the onset of the ACC some 10 million years prior -- that's nonsensical, and just doesn't track with what we know about earth history and ocean change over the Cenozoic. Looking at Figure 2, yes there are geographic differences, but invoking ACC as a driver of these changes doesn't align with other facts, namely the onset of much larger body sizes in fossil mysticetes subsequent to the Early Miocene. The conclusion's emphasis of "gigantism in the Plio-Pleistocene, but only for the Northern Hemisphere" ignores that the gigantic taxa of this time were already global -- without the kind of N/S fossil-modern taxa comparison, I don't see how this distinction is anything but arbitrary. Also, it's worth noting that Slater et al. (2017) proposed a mechanism grounded in Northern glaciation because the onset of the Ice Ages by Milankovitch cycles are explicitly tied to the larger available landmasses in this hemisphere, which fueled positive feedback loops.

Minor concerns:

The metaphor of a dark age is a bad one: it's a historical misconception that doesn't have a basis in Medieval scholarship. I think it's cartoonish trope, and the authors don't serve themselves well to use the term in quotes, as if it's a euphemism. One of the authors referred to the Early Miocene baleen whale gap -- I think that's a more useful set of words, rather than a term freighted to poor understanding of a historical period in human history.

ll. 18 and 41 should explicitly state that the authors present new fossil material referable only to a clade-level diagnosis -- the lines seem to hint at a new taxon, when that's not really the case.

Ll. 152-160 is highly speculative and a bit confusing: the whale-krill feedback loop is dependent on krill, which do not have a fossil record, and at best have a Mid Miocene antiquity based only on molecular divergence estimates. I also think it's a bad idea to discuss population densities of "smaller animals like NMV P218462," which confuses a single, incomplete specimen of low diagnostic value with taxonomic variation and population density for a feeding rated-based phenomenon (i.e., Savoca et al. 2022 make a case not on body size but on feeding energetics related to population sizes; body size is just a window for a return on investment). The reader of PRSB deserves more clarity.

In light of the major concerns, I think the conclusion section needs some revision and attention to diction. First, I think "evolve" is a better word than "grew," (l. 162) especially for readers of PRSB. The argument about ACC doesn't really track (see above), arguing for a "Northern Hemisphere only" scope of agency for cosmopolitan taxa is a bit arbitrary (see above). I agree that more collecting would be warranted, but isn't that always the case? If the authors are going to make this call, then they should probably make sure to run an analysis on collecting effort to back it up.

I think the paper is mostly solid, and well written; some points are worth clarifying in more detail, and the pitch of the arguments should be refined.

This is the second time I have reviewed this manuscript by Rule et al. The authors have revised the manuscript, including the addition of new data, that makes their point clear, and improved the word choices and nuances for the context of their results. Is whale gigantism a debate? I don't think so, but it is clearly a topic of inquiry for which new data should be welcomed. The authors propose a novel spin to the interpretation of mysticete gigantism, and I think that's worth publishing. Please see some minor points that I think require consideration, below.

Overall there was lots of overlap between the two referee reports; in the interest of getting to the point, I'll use the authors' reply to my (Referee 2's) queries as a framework:

"Accounting for these specimens does notably shift the timing of, and weakens support for, the proposed mode shift. Another recent analysis (Burin et al. 2023) failed to identify a Plio-Pleistocene rate shift in mysticetes and instead revealed a Late Miocene trend towards large size in rorquals. Overall, we thus maintain that the evolution of mysticete gigantism remains debated, especially considering the still-patchy southern fossil record."

Again, the main point of Slater et al. (2017)'s argument is a clade-level shift towards extremely large body size (20 m and larger) and the extinction of smaller bodied forms. The specimens identified by Bianucci et al. (2019) don't actually change the modal shift and they don't exceed the confidence intervals. I think it's disingenuous to say "weakens" support for a modal shift.

I encourage the authors to read Burin et al. 2023 more carefully to understand the kind of analyses that they undertook and why they are different from the ones that Slater et al. (2017) conducted, and also to see that they did recover large size shifts for both rorquals and balaenids.

While the timing seems to have a broader stratigraphic window based on new data (i.e., late Miocene through Pliocene), the word "debated" creates considerable confusion that I think is unwarranted. What are the alternative hypotheses or scenarios? This paper proposes a hidden pattern, but the authors are not playing fair by calling mysticete gigantism a debate. Consider, for example, alternate hypotheses for mass extinction events, and other phenomena. Debates have differences in explanations, less so the differences in data types and sources.

I think some of the confusion does arise from ref. 8 (Marx and Fordyce), which focused on a large mysticete (8-10 m) from the Eocene lacking a filter-feeding apparatus, but Llanocetus isn't anywhere near the size of today's rorquals. The Bianucci and Burin references are relevant, and I suppose the revised parsing is OK.

What about Bianucci et al.'s Perucetus paper from this summer? If the authors are right about Perucetus, those size estimates certainly change the picture. I think citing that paper is highly relevant.

"... l. 144-145."

OK on that phrasing.

"...We now clarify this in l.154-156"

OK.

"...a second pie chart in Figure 2b to visualise the discrepancy in the available rock record."

The inclusion of these new data are helpful and does contribute to the authors' case. It fits with their other data, so this is now compelling.

"...We now clarify this in l. 156-160."

OK.

"...clarify this in l. 144-146...and l. 171-174"

OK, thank you for clarifying, this is much better.

"...we allude to it in l. 144-145"

OK.

On the minor concerns, OK.