The First Discovery of a Fish Fossil (Phareodus sp.) from Paleogene Fluvial Deposits in Western Washington State, USA

Round 1

Reviewer 1 Report

In this paper, the author presents a brief description of a partial fish fossil representing the caudal and dorsal fins, as well as what appears to be an impression of the posterior fourth of an articulated fish specimen, along with a more fragmentary counterpart of the dorsal fin area. The specimen was recovered in the Chuckanut Fm, representing Eocene deposits of Washington state, which constitutes its main significance, given the paucity of articulated Cenozoic fish fossils having been recovered from floodplain settings. I agree with the identification of the specimen, and I support the publication of this article, although I do have several comments and suggested changes in order to emphasize the relevance of the article.

The main issue I see with the identification of the specimen is that it is based on few discreet characters, but rather mainly on broad correspondence: fin insertion points, caudal fin shape, relative sizes of the fins compared to the caudal peduncle, relative size of scales, etc., and that such broadly corresponding morphologies are not easy to support in an objective way—even though I concur with the identification. As it stands, I almost agree with the author’s attribution of the specimen to “cf. Phareodus”, as he puts it himself in the conclusion. The caudal fin ray counts constitute some of the best evidence for the attribution of the specimen to Phareodus, the only other evidence being the side-to-side presentation of their fossil with that of a more complete specimen of Phareodus with a highlighted area corresponding to the part of the fish the author believes was recovered. As a side note, the figure where a better-preserved specimen is presented, highlighting the part purported to correspond to the part of the fish preserved in the Chuckanut Fm specimen is misleadingly captioned “Chuckanut Formation fossil fish fragment shown in relation to Phareodus testis…”, when the figure does not appear to contain the Chuckanut Fm specimen at all. The Chuckanut Fm does not seem to preserve the distal margin of the dorsal fin or the morphology of vertebral centra, contrary to what figure 7 seems to suggest. If I am misunderstanding this and the Chuckanut specimen does preserve these morphologies, the specimen should be better illustrated in order to document these morphologies.

In an effort to better support the identification of the fossil, I would recommend more space be spent on the examination of the scales, given that those of Phareodus have long been recognized as large, cycloid, with distinctive reticulations (see, for instance, L. Grande’s Green River monograph from 1984 for a description and an illustration). Figure 10 is meant to illustrate a detailed view of the fossilized scales of the Chuckanut specimen, but is difficult to interpret, and seems to show ridged patterns on the scales that are not characteristic of Phareodus. That said, much of the photograph is not in sharp focus, and I may be misreading the features shown in the photograph. Figure 1 is not detailed enough to determine whether additional meristics could be used to further support the attribution, although, given the poor preservation of the fossil, traditional scale counts would not be possible, and the counts that could potentially be taken (e.g., scale count at the peduncle) would not be easy to use. In sum, I agree with the attribution, although it remains somewhat unreliable without supporting it by more diagnostic characters.

In addition to the description of the specimen, the manuscript presents some palaeoenvironmental considerations based on the recovery of Phareodus in the formation and based on its sedimentology. This is an interesting section and a welcome inclusion, although it remains somewhat underdeveloped in some aspects. For instance, although the taxonomic diversities of the fish faunas recovered from a number of Eocene localities are compared in table 2, this is only mentioned in the text with no further discussion, and the main ecological conclusion drawn from the presence of Phareodus is that other smaller taxa must have been present to serve as prey, as supported by the rarity of Phareodus compared to other taxa in Green River deposits. This is a reasonable conclusion, but more could probably be drawn from this. Whether the low diversity of the Chuckanut is due to low preservation potential of the formation or to a relative lack of diversity of the local ichthyofauna at time of deposition could be discussed, for instance. The Chuckanut Fm is also occasionally referred to as “riverine”, even though a wider variety of environments seems to be represented in the formation, as illustrated in figure 11, and referring to it as a floodplain deposit might be more accurate. If this is the case, and given the lamination of the sandy siltstone where the specimen was found (cf. Figure 13), I doubt that the specimen was deposited in an active channel, but either in very slow-moving backwaters or in oxbow lake settings; faster flows would have resulted in massive bedding instead. If this is the case, the recovery of Phareodus may be an indication of hypoxic conditions, given that modern osteoglossids are tolerant of such conditions (see Welcomme, 1979), and may explain the scarcity of fish material in the Chuckanut Fm (compare with the palaeoenvironmental reconstruction suggested for a Bridger Fm locality in Divay & Murray, 2016, which also includes floodplain disarticulated Phareodus material). With that said, osteoglossine material (perhaps Phareodus or indistinguishably close relatives) has been described from conclusively riverine deposits, preserved as disarticulated material (see the Belly River Group ichthyofauna published by Brinkman, 2019, and the Maastrichtian non-marine teleost material published by Brinkman et al., 2020). More background about this osteoglossine material may provide some useful context to readers, although the author’s approach of solely focusing on formally described Phareodus species is another valid option. Either way, if the Chuckanut specimen is indeed from active channel deposits, this would additionally distinguish it as the only articulated Phareodus specimen recovered from such settings, to my knowledge.

In terms of organization, there are no sections that are missing or superfluous, although the paper could probably be edited more tightly. The focus of the manuscript is somewhat loose, and the text occasionally goes on tangents that don’t add much overall (e.g., repetitions in paragraphs dealing with the sedimentology of the formation in introductory, geological and discussion sections), some of the paper’s goals announced in the abstract and introduction (e.g., “this information improves our understanding of the range of depositional environments within the Chuckanut Formation, and provides clues for searching for additional specimens”) are never referred to again, and some of the included figures aren’t overly useful to the points the paper is trying to make or could be combined into a single one to save space (e.g., a figure exploring the parasitic larval stage of unionid mussels that is interesting although not vital to the point being made, the inclusion of 3 different maps of the same area...). Other figures are presented but never specifically discussed in the text (e.g., the photomicrographs in Figure 13), which—in my opinion—should either be discussed in the text, or removed altogether. Ultimately, because this doesn’t actively detract from the article, the editor is in the best position to decide what should be done about these issues, whether the article should be more tightly focused to save space or not, etc.

This review probably appears overly negative, but ultimately, I agree with the essential points: this specimen deserves publication and most likely represents Phareodus. I would argue that some additional work is required to present the article in its best light, to document/identify the fossil as fully as it can be, and to make the article as useful to readers as possible. It is debatable whether this required additional work should be considered extensive or not, but given that another round of reviews would be beneficial, I opted to recommend "major" revisions.

Language-wise, the article has no major problems, with a few slight issues that I could spot. There are several typos (e.g., “distal” fin for dorsal fin, l. 27, “British Columba” for British Columbia, and “vertebrae” for vertebrate, both on l. 207), missing words (e.g., “represent the extinct genus, an open-water carnivore that…”, l. 9-10), Phareodus isn’t italicized in the abstract, the use of the antiquated term “Tertiary” (e.g., l. 205), the inappropriate capitalization of ‘early’ and ‘late’ in “early Eocene” and “late Eocene” (which aren’t formally defined time intervals and therefore shouldn’t be capitalized), etc.

Chronological and stratigraphical characterizations are occasionally used interchangeably (e.g., sometimes “early Eocene” is used, sometimes “lower Eocene” is used)—I would just recommend harmonizing this and using either chronological terms or stratigraphical terms throughout. I suppose that this is another nitpick, but Phareodus is also known from Wyoming Bridger Fm outcrops (contra “and in the Green River and Bridger Formations in Utah”, l. 160-161). Finally, I don’t support the use of italics for any term that is not a genus or species name (cf. “littoral”, “limnetic”, l. 176), but I suppose that this is an editorial decision for the journal to take.

Author Response

Thanks for the many detailed comments, and for your constructive approach as a reviewer.

During the time when the manuscript was in review I was able to find some more fragments from the original slab, resulting in a counterpart impression that provides additional anatomical detail. I have been able to provide a better description of the specimen. The specimen can be seen to include most of the anal fin, and part of the vertebral column. On close inspection, the mottled areas that were first assumed to represent scale patterns turn out be preservational artifacts where the carbon film has been locally damaged.

I have followed your advice to expand the discussion of the sedimentology Because the specimen was found in a “float” specimen along the rocky stream bed, the geologic setting for the specimen is a  matter for conjecture. However, I am inclined to belief that the host sediment may represent an oxbow lake. The absence of plant remains suggests the siltstone did not originate in an overbank deposit; a hallmark for Chuckanut overbank deposits is the abundance of plant fossils that range from large palm fronds to an abundance of plant fragments.

I have rewritten and reorganized much of the manuscript in the light of your recommendations. I agree that the manuscript could be condensed to have tighter focus, but the journal editor has notified me that Geosciences manuscripts are required to have a minimum word count of 4,000. My submitted manuscript had only about 2,400 words, and after revision the word count is still a bit below the 4,000 target. As a balance, the manuscript contains numerous illustrations, and I lean toward the “a picture is worth a thousand words” philosophy.

Thanks for the careful proofreading. I appreciate your keen editorial eye, I have made the suggested corrections. Among the changes I have eliminated capitalization of ages, e.g., lower Eocene has changed to Lower Eocene. I used to capitalized format to match the style used by Lance Grande in is 1984 “Paleontology of the Green River Formation…” monograph. The Geological Society of America Geologic Time Scale divides the Eocene into E, M, and L subdivisions. The International Chronostratigraphic  Chart  does not make these divisions. Personally, I am happy to go along with whatever format that journal editors prefer.

In closing, I appreciate your expert knowledge of paleoicthyology. As you probably guessed, I am not a “fish guy”. This discovery came to me because have long been the leading authority on the paleontology of the Chuckanut Formation, which underlies the university where I have worked for more than 40 years.

Thanks, George

Reviewer 2 Report

Mustoe paper review

 

Line Comment

 

282  Priscacara, not Prisicacara

 

113  Fig. 8 caption, change (C) to (D)

 

124-125   To my knowledge this is not true. Fish fin-ray counts are fixed early in development and do not change after that. It’s true that ray counts are variable in many species, but they don’t change as the fish grows once they are established very early in development.

 

128 There is a gap in queenslandicus.

129 Insert space before [24]

 

130  The specific epithet is spelled muelleri, after a man named Müller. ü becomes ue because species names can’t have accents, and the single letter i is added to denote that the person is a male (different from the botanical rules).

 

130  no italics for “and the Australian”

 

136-137   You should cite Smith, Stearley, and Badgley here, because this was their original insight about poor preservation in fluvial sediments, although we also know that it can sometimes be overcome. They advocated washing and screening of disarticulated bones, which would certainly work for improving knowledge of diversity, but the poor preservation itself can be overcome by special circumstances as explained below for line 140.

Smith, G. R., R. F. Stearley, and C. E. Badgley. 1988. Taphonomic bias in fish diversity from Cenozoic floodplain environments. Palaeogeography, Palaeoclimatology, Palaeoecology 63: 263–273.

 

140  Overbank deposits can indeed preserve articulated fishes, in special circumstances. For examples, see several papers about fishes (percopsiforms, osmeriforms and osteoglossiforms) and their taphonomy in Paleocene overbank and/or oxbow deposits in Alberta. Some additional examples are of Maastrichtian age also, including osteoglossomorphs and percopsiforms. Look for papers by A. M. Murray and co-authors. It’s exactly in quiet-water overbank deposits and oxbows where the poor preservation of fluvial fishes can be avoided, provided they are soon buried by subsequent deposits, such as repeated overbank flows and crevasse splays.

 

142 spelling of Chuckanut.

 

147 comma after Member

 

148 delete o from oin

 

1.     179       semicolon or period needed after sedimentology

 

1.     18  delete “these”

 

1.     195       I think you mean the J. E. Tynsky Quarry, rather than a Tynish Quarry.

 

1.     213-215 Here I would add, following Smith et al. 1988, that one should focus to the extent possible on the overbank deposits containing quiet-water sediments to find well-preserved fishes.

 

1.     227       Delete “the” before Mia

235-237 correct hanging indent

 

224-226    The specimen(s) should be deposited in a public museum such as the Burke, not in a private research collection.

 

1.     Table 2

2.     I’m not sure Eosalmo has been found at Horsefly.

3.     Priscacara is not spelled correctly.

4.     Delete ] after Canada

5.     Your list of genera is missing Libotonius, which occurs at Princeton and Republic (as two different species).

6.     For completeness you could also include the percomorph fossil from near Liberty.

 

Fig. 2      Please label the city

 

Fig. 3      It’s difficult to spot the HU and MA outcrops, and to know where CK and SW meet or overlap because all the outcrops are the same color. Just a suggestion to use 4 colors instead of just red. If some outcrops are not identified to formation, maybe say so. More importantly, I think it would be helpful to have the fossil fish localities highlighted in this figure: your find, Hesse’s Priscacara, and the Liberty percomorph.

 

Hesse’s Priscacara campi is from within your map area for Fig. 3, as also is a fish you likely refer to in lines 170-171. It is an obscure report (SVP Abstract 1994) about a percomorph fossil from the Swauk Fm near Liberty, WA. Unfortunately if was not identified to family or genus.

 

Fig.  12: Roslyn should be red because of Priscacara campi. Fossil fishes are also known in B.C. from Tranquille, Falkland, and Merritt (and one fragment of Eosalmo from Chu Chua).

Fig. 13    Caption: (B. C) change period to comma

See list of minor corrections above. The English is excellent except for  these typos;

Author Response

Thanks for your detailed review. I have corrected the various typos and also tried to accommodate all of your other recommendations. During the time when the manuscript was in review I was able to find some more fragments from the original slab, resulting in a counterpart impression that provides additional anatomical detail. This has allowed writing a clearer description of the fossil.

Thanks  for telling me about the Smith, Stearley, & Bagley 1988 paper. I found it very helpful. I have expanded the discussion of the depositional environment. As a “float” specimen, the stratigraphic position cannot be ascertained. The specimen is presumed to have come from a nearby outcrop that contains highly-varied sedimentology. The absence of carbonaceous inclusions suggests to me that the fossil may have been preserved in an oxbow lake, but the origin remains a matter of conjecture. I have expanded discussion of this topic, and added some new photos.

You refer to a SVP Abstract 1994 about the Swauk Formation fish fossil. I have the abstracts from that meeting, and I can find no presentation that mentioned that specimen. You refer to the location as “Swauk Formation near Liberty”, but I was contacted by the collector at the time of the discovery, and I recall that he described the location as being near Swauk Pass. There is not much geographic difference between the two locations. However, access to the Swauk Formation outcrops  in the Liberty town area is extensively restricted by numerous mining claims, in contrast to the strata near Swauk Pass. I suspect the “near Liberty” description was because it is the only town ( a very small town) in the general area, and the name appears on highway maps.

I added Tranquille, Falkland, Merritt, and Chu Chua to the fossil fish location map. I have visited all of these sites because of their importance for paleobotany, but I have not seen published mention of fossil fish from these sites. I have seen beautiful fish fossils from the classic BC sites: Princeton, McAbee, Horsefly River, and Driftwood Creek.

In summary, thanks for all the constructive advice. I’ve spent many years studying the paleontology of the Chuckanut Formation, but as you probably guessed, I am not a “fish guy”. I greatly value your expertise.

Best wishes, George

Reviewer 3 Report

While the fossil record of ray-finned fishes is generally of quite high quality, the same can not be said for those found in fluvial deposits. So the discovery of this specimen surprised me with its preservation quality despite the hostile environmental conditions.  In my opinion, the publication of such a find will be highly significant to expanding our understanding of what kind of specimens can be preserved in fluvial deposits and hopefully more thorough exploration of such fossil sites can give us a more compete understanding of the riverine fish fauna of the Washington region during the Early Eocene. Most of my changes to this manuscript concern correcting typos and suggested reformatting of the references section to comply with the standards of Geosciences journal which I have highlighted in the review pdf.

However, there are two issues that I think need to be addressed before publication and those include:

1. The fact that this specimen has no collection number. It is certainly good to know that it is deposited in a collection at the Geology Department of the Western Washington University but for the sake of accessibility to any future researchers who want to study this specimen, a collection number should be mentioned in this manuscript. Especially to such an exceptionally preserved specimen from a locality that is typically hostile to the preservation of any vertebrate material, such official documentation would be invaluable.

2. While the figures are mostly of good quality, I think Figure 2 warrants more explanation in regards to what the map region that is cross checked refers to. Also Figure 4 does not have the time sequences accurately portrayed as the Early Eocene starts too early so the Paleocene needs to move up a bit on the bar showcasing the different time series.

I think once these aforementioned issues and the small changes I highlighted in the review document are resolved, then this paper should be published as it does refer to a important find from an depositional environment whose fish fauna are extremely poorly understood in comparison to lacustrine and marine sediments. Hopefully, such a find being published can encourage other paleontologists and amateur fossil collectors to investigate these fossil deposits more thoroughly to get a better ideas of what life was like in the river systems of early Eocene Washington.

 

Comments for author File: Comments.pdf

Other than a few typos which I highlighted in the review document, the English is satisfactory. Also, Early Tertiary should be changed to Paleogene which is the official name for this time slice.

Author Response

Thanks for the careful review. I have tried to accommodate all of your recommendations. I also appreciate your careful editing in regard to various typos. I have always been somewhat “word blind: when it comes to proofreading my own writing.

During the time when the manuscript was in review I was able to find some more fragments from the original slab, resulting in a counterpart impression that provides additional anatomical detail. This has allowed writing a clearer description of the fossil. I have also expanded the discussion of the sedimentology. As a “float” specimen that has presumably been transported from nearby strata, recognizing the point of origin is complicated by the fact that these strata are complex. My inclination is to believe that the fish was preserved in an oxbow lake, not an overbank deposit. I have discussed this topic  in detail, and added some locality photos

In regard to the collection number, I have curated the specimen into the Western Washington University Geology Museum collection, where it will be placed on public display in an exhibit that includes numerous other Chuckanut Formation fossils. When I retired from WWU in 2014, I donated several thousand specimens to the research archives at the Burke Museum, so I am well acquainted with the curatorial staff there. At WWU, the specimen will be readily available for study.

Thanks for catching the misplaced boundary lines on the stratigraphic column. That’s been fixed. I have extensively revised some parts of the manuscript to try to improve the organization and clarity of presentation.

Best wishes,  George

Reviewer 4 Report

 

There is a considerable discrepancy between the title and the main contents of the ms. Therefore, it should be retitled and reorganized accordingly.

Further suggestions and comments are directly indicated in the text (see attached pdf).

Comments for author File: Comments.pdf

Author Response

Thanks for the useful review. I particularly appreciate your detailed comments on the PDF copy. I have rewritten large sections of the manuscript in an attempt to improve the organization and clarity of presentation. Also, during the time when the manuscript was in review I was able to find some more fragments from the original slab, which has allowed me to as allowed writing a clearer anatomical description of the fossil. In particular, I have tried to strengthen the abstract, introduction, and conclusions to eliminate inconsistencies and weaknesses that you noticed.

Best wishes,George

Round 2

Reviewer 1 Report

This is a much improved version of the MS, and I think it is close to being ready for publication. For instance, the organization of the paper is now fine as-is and the slightly repetitious/unfocused nature of the previous version is no longer an issue. I’m also very thankful to the author for having provided additional explanations as to the preservation of the specimen—it having been initially preserved in an oxbow lake setting before re-entering an active channel to be preserved as a float specimen makes a lot of sense, and explains how it can be described as a floodplain, lake and river specimen, all at once. I also have to commend the author for publishing an incomplete fish specimen; these are too often overlooked given the challenge they represent, so very well done.

It is a little unfortunate that scales don’t appear to have been preserved after all (and I agree with this assessment, given Fig. 11), but the additional slab having been located is excellent news! My main recommendation will actually be to use this additional slab to make more descriptions and to provide more detailed figures (close-up photographs of the vertebral series, especially of the anterior vertebral series, would be extremely helpful). I would also encourage the author to attempt a description of the anterior vertebral centra preserved in that slab, as these will undoubtably confirm the identification of the specimen as an osteoglossid, and therefore as Phareodus. I know that this may seem daunting to a non-fish person (and to a lot of fish folks, too), but osteoglossid vertebral centra have been described from disarticulated specimens in a good level of detail, so the author can use these published descriptions to get an idea of the type of relevant information to be included in such a description. I’ve already mentioned those previous publications in the first round of review, but in case it could help to have them all here, these are the references I mean: Brinkman’s 2019 Teleost abdominal centra from the Belly River Group and Brinkman et al.’s 2020 Systematic reappraisal and quantitative study of non-marine teleosts from the Western Interior both have descriptions of osteoglossid vertebral centra, and Divay and Murray 2016 The Fishes of the Farson cutoff fishbed has extensive descriptions of Phareodus vertebral material, including descriptions of the various morphologies at different vertebral positions, which will probably be useful in further supporting the identification of the Chuckanut fish. I can provide any/all of these, as required. Because I am lead/co-author on two of these publications, I want to clarify that I am not demanding that any of these be cited—I am leaving that decision to the author—just giving reference material on how to describe material that the author may not be familiar with (e.g., describing the condition of the rib articulation, whether parapophyses, neural arches and hemal arches are fused to the centrum or not, etc.). All of this being said, articulated fish specimens are often very compressed and flattened, making a lot of these discrete morphological characters difficult to see in detail, but if any additional character supporting the identification of this fish can be found on the specimen, I believe that these are most likely to be preserved and will be taxonomically informative.

I support the added reference to Gerry Smith’s work on fish taxonomic diversities in different freshwater settings (reference #26) in the Discussion, and I of course agree that the abrasiveness of fluvial sediments and water agitation are the main reason for the poor preservation of riverine fishes. That said, even poorly preserved fluvial remains are often identifiable, so I would argue that our poor knowledge of river fishes through time is mainly due to neglect of these remains in favor of lake fishes that are better preserved and therefore considered more scientifically "valuable". Just a minor observation in passing, but the author might want to add reference to this here—increasing awareness about this issue can only be helpful.

Lines 205-206, the author notes that Phareodus have been recovered in the Green River Formation in Wyoming and in both Green River and Bridger formations in Utah—this is correct but incomplete: Wyoming localities of the Bridger Fm also have Phareodus (see previously mentioned Divay and Murray, 2016). Osteoglossid material is also known from other formations (e.g., Paskapoo and Ravenscrag formations, see Wilson 1980 Oldest known Esox (Pisces: Esocidae), part of a new Paleocene teleost fauna from western Canada), although I don’t believe that they were attributed to genus level. Either way, there is good evidence that Phareodus was much wider-ranging—both geographically and stratigraphically—than specimens formally attributed to the genus may suggest, and this should probably be at least mentioned somewhere.

In case the author wants to add another informative paragraph or so to his MS, some more info about the invertebrate trace fossils and footprints from the formation could be useful (l. 82-83), because it would paint a more detailed paleoenvironmental and paleoecological picture of the formation. A more formal “Systematic Paleontology” section could also be included, specifying the taxonomic level to which the author assigns the specimen, which remains somewhat ambiguous (e.g., the specimen is referred to simply as Phareodus throughout much of the article, but it is referred to as “cf. Phareodus” in the conclusion). Given the nature of the specimen and the broad similarities used to compare it to Phareodus, whether to use “confer” or not is questionable. As mentioned in my previous review, the fin ray counts and scale morphologies were the best lines of evidence positively linking this specimen with Phareodus, the fin insertion points, relative sizes and caudal fin shape serve more to disqualify other possible taxonomic affiliations. With this in mind, and given that scales are no longer useable, I suppose that a “cf.” identification would be more appropriate, unless additional characters can be observed in the vertebral series, which would positively identify the Chuckanut specimen as an osteoglossid. But I may be too strict with this recommendation, given that no other affiliation seems very plausible.

I still don’t see how Figure 8 was obtained from the Chuckanut specimen; it still just looks like a Green River specimen with highlighted parts corresponding to the recovered Chuckanut specimen, although I suppose I could be wrong. Personal preference aside, I agree that it is really the journal’s decision whether all three maps and various other non-essential figures should be kept or not, given that supernumerary figures do not detract from the overall relevance of the article.

All in all, this is good progress, and I do not doubt that this will be publishable with a few minor revisions.

The manuscript only had minor language issues to begin with, so the current state of the article is pretty much excellent. Some older terminology still subsists (e.g., “Tertiary” l. 41, l. 251), although a lot of the terminological issues from the previous versions have been appropriately dealt with, such as the capitalizations of informal time intervals. As a side note for the author, I very enthusiastically recommend using Donald Owen’s 2009 How to use stratigraphic terminology in papers, illustrations, and talks, which is a fantastic resource for such matters (and which is freely available online); it is a life-saver to avoid mixing geochronologic and chronostratigraphic terminology, for instance, or to avoid errors in the capitalization of terms.

Author Response

Thanks once again for a detailed and very constructive review. I have added a brief section on the vertebral centra, and included the references that you so kindly provided. I've also tried to provide a clearer caption for Figure 8.