Author response:
The following is the authors’ response to the original reviews.
eLife assessment
The authors present 16 new well-preserved specimens from the early Cambrian Chengjiang biota. These specimens potentially represent a new taxon which could be useful in sorting out the problematic topology of artiopodan arthropods - a topic of interest to specialists in Cambrian arthropods. Because the anatomic features in the new specimens were neither properly revealed nor correctly interpreted, the evidence for several conclusions is inadequate.
We thank the Senior Editor, Reviewing Editor and three reviewers for their work, and for their comments aimed at improving this project and manuscript. We have engaged with all the comments in detail, in order to strengthen our work. This includes adding additional data to support that all Acanthomeridion specimens belong to a single species, running further phylogenetic analyses including more trilobite terminals to test the specific hypothesis and interpretation raised by Reviewer 2, and visualising our results in treespace in order to determine support for the different interpretations of the ventral structures and their implications for the evolution of Artiopoda. We have also greatly expanded the introduction, which we feel adds clarity to areas misunderstood by some reviewers in the previous version of the manuscript.
Our point-by-point response to the public reviews of the reviewers are outlined below. We have also made changes resulting from the additional suggestions which are not public, which we have not reproduced below. We submit a new version of the main text, and can provide a tracked changes version if required. The new main text includes 9 figures and is 8624 words including captions and reference list.
Public Reviews:
Reviewer #1 (Public Review):
Summary:
Du et al. report 16 new well-preserved specimens of atiopodan arthropods from the Chengjiang biota, which demonstrate both dorsal and ventral anatomies of a potential new taxon of artipodeans that are closely related to trilobites. Authors assigned their specimens to Acanthomeridion serratum and proposed A. anacanthus as a junior subjective synonym of Acanthomeridion serratum. Critically, the presence of ventral plates (interpreted as cephalic liberigenae), together with phylogenic results, lead authors to conclude that the cephalic sutures originated multiple times within the Artiopoda.
We thank Reviewer 1 for their comments on the strengths and weaknesses of the previous version of the manuscript. We hope that the revised version strengthens our conclusions that Acanthomeridion anacanthus is a junior synonym of A. serratum.
Strengths:
New specimens are highly qualified and informative. The morphology of the dorsal exoskeleton, except for the supposed free cheek, was well illustrated and described in detail, which provides a wealth of information for taxonomic and phylogenic analyses.
Weaknesses:
The weaknesses of this work are obvious in a number of aspects. Technically, ventral morphology is less well revealed and is poorly illustrated. Additional diagrams are necessary to show the trunk appendages and suture lines. Taxonomically, I am not convinced by the authors' placement. The specimens are markedly different from either Acanthomeridion serratum Hou et al. 1989 or A. anacanthus Hou et al. 2017. The ontogenetic description is extremely weak and the morpholical continuity is not established. Geometric and morphometric analyses might be helpful to resolve the taxonomic and ontogenic uncertainties.
We appreciate that the reviewer was not convinced by our synonimisation in the first version of the manuscript. The recommendation of the reviewer to provide linear morphometric support for our synonymisation was much appreciated. We have provided measurements of the length and width of the thorax (Figure 6 in the new version), visualising the position of specimens previously assigned to A. anacanthus, to show this morphological continuity. These act as a complement to Figure 5, which shows the fossils in an ontogenetic trend.
I am confused by the author's description of the free cheek (libragena) and ventral plate. Are they the same object? How do they connect with other parts of the cephalic shield, e.g. hypostome, and fixgena? Critically, the homology of cephalic slits (eye slits, eye notch, dorsal suture, facial suture) is not extensively discussed either morphologically or functionally.
We appreciate that the brevity of the introduction in the previous version led to some misunderstandings and some confusion. We have provided a greatly expanded introduction, including a new Figure 1, which outlines the possible homologies of the ventral plates and the three hypotheses considered in this study. The function of the cephalic and dorsal suture are now discussed in more detail both in introduction and discussion.
Finally, the authors claimed that phylogenic results support two separate origins rather than a deep origin. However, the results in Figure 4 can explain a deep homology of the cephalic suture at molecular level and multiple co-options within the Atiopoda.
A deep molecular origin is difficult to demonstrate using solely fossil material from an extinct group such as Artiopoda. Thus our study focuses on morphological origins. The number of losses required for a deep morphological origin means that we favour multiple independent morphological origins.
Reviewer #2 (Public Review):
Overall: This paper describes new material of Acanthomeridion serratum that the authors claim supports its synonymy with Acanthomeridion anacanthus. The material is important and the description is acceptable after some modification. In addition, the paper offers thoughts and some exploration of the possibility of multiple origins of the dorsal facial suture among artiopods, at least once within Trilobita and also among other non-trilobite artiopods. Although this possibility is real and apparently correct, the suggestions presented in this paper are both surprising and, in my opinion, unlikely to be true because the potential homologies proposed with regard to Acanthomeridion and trilobite-free cheeks are unconventional and poorly supported.
What to do? I can see two possibilities. One, which I recommend, is to concentrate on improving the descriptive part of the paper and omit discussion and phylogenetic analysis of dorsal facial suture distribution, leaving that for more comprehensive consideration elsewhere. The other is to seek to improve both simultaneously. That may be possible but will require extensive effort.
We thank the reviewer for their detailed comments and suggestions for multiple ways in which we might revise the manuscript. We have taken the option that is more effort, but we hope more reward, in interrogating the larger question alongside improving the descriptive part of the paper. This has taken a long time and incorporation of new techniques, but has in our opinion greatly strengthened the work.
Major concerns
Concern 1 - Ventral sclerites as free cheek homolog, marginal sutures, and the trilobite doublure
Firstly, a couple of observations that bear on the arguments presented - the eyes of A. serratum are almost marginal and it is not clear whether a) there is a circumocular suture in this animal and b) if there was, whether it merged with the marginal suture. These observations are important because this animal is not one in which an impressive dorsal facial suture has been demonstrated - with eyes that near marginal it simply cannot do so. Accordingly, the key argument of this paper is not quite what one would expect. That expectation would be that a non-trilobite artiopod, such as A. serratum, shows a clear dorsal facial suture. But that is not the case, at least with A. serratum, because of its marginal eyes. Rather, the argument made is that the ventral doublure of A. serratum is the homolog of the dorsal free cheeks of trilobites. This opens up a series of issues.
We appreciate that the reviewer disagrees with both interpretations we offered for the ventral plates, and has offered a third interpretation for the homology of this feature with the doublure of trilobites. Support for our original interpretation comes from the position of the eye stalks in Acanthomeridion, which fall very close to the suture between ventral plate rest of the cephalon. However, we appreciate that the reviewer has a valid interpretation, that the ventral plates might be homologues of the doublure alone.
To clarify the (two, now three) hypotheses of homology for the ventral plates considered in this study, we provide a new summary figure (Figure 1). In addition, the introduction has been greatly lengthened with further discussion of the different suture types in trilobites, their importance for trilobite classification schemes, and extensive references to older literature are now included. Further, we add background to the hypotheses around the origins of dorsal ecdysial sutures.
We add that the interpretation of A. serratum as having features homologous to the dorsal sutures of trilobites is already present in the literature, and so while the reviewer may disagree with it, it is certainly a hypothesis that requires testing.
The paper's chief claim in this regard is that the "teardrop" shaped ventral, lateral cephalic plates in Acanthomeridion serratum are potential homologs of the "free cheeks" of those trilobites with a dorsal facial suture. There is no mention of the possibility that these ventral plates in A. serratum could be homologs of the lateral cephalic doublure of olenelloid trilobites, which is bound by an operative marginal suture or, in those trilobites with a dorsal facial suture, that it is a homolog of only the doublure portions of the free cheeks and not with their dorsal components.
We include this third possibility in our revised analyses and manuscript. To test this properly required adding in an olenelloid trilobite to our matrix, as we needed a terminal that had both a marginal and circumoral suture, but not fused. We chose Olenellus getzi for this purpose, as it is the only Olenellus with some appendages known (the antennae). We also added further characters to the morphological matrix, and additional trilobites from which soft tissues are known, in order to better resolve this part of the tree. Trilobites in the final analyses were: Anacheirurus adserai, Cryptolithus tesselatus, Eoredlichia intermedia, Olenoides serratus, Olenellus getzi, Triarthrus eatoni.
However, addition of these trilobites added a further complication. Under unconstrained analysis, Olenellus getzi was resolved with Eoredlichia intermediata as a clade sister to all other trilobites.
Thus the topology of Paterson et al. 2019 (PNAS) was not recovered, and so the hypothesis of Reviewer 2 could not be robustly tested. In order to achieve a topology comparable to Paterson et al., we ran a further three analyses, where we constrained a clade of all trilobites except for O. getzi. This recovered a topology where the earliest diverging trilobites had unfused sutures, and thus one suitable for considering the role of Acanthomeridion serratum ventral plates as homologues of the doublure of trilobites.
Unfortunately, for these analyses (both constrained and unconstrained), Acanthomeridion was not resolved as sister to trilobites, but instead elsewhere in the tree (see Table 1 in main text, Fig. 9, and SFig 9). Thus our analyses do not find support for the reviewer’s hypothesis as multiple origins of this feature are still required.
It was still an excellent point that we should consider this hypothesis, and we have retained it, and discussion surrounding it, in our manuscript.
The introduction to the paper does not inform the reader that all olenelloids had a marginal suture - a circumcephalic suture that was operative in their molting and that this is quite different from the situation in, say, "Cedaria" woosteri in which the only operative cephalic exoskeletal suture was circumocular. The conservative position would be that the olenelloid marginal suture is the homolog of the marginal suture in A. serratum: the ventral plates thus being homolog of the trilobite cephalic doublure, not only potential homolog to the entire or dorsal only part of the free cheeks of trilobites with a dorsal facial suture. As the authors of this paper decline to discuss the doublure of trilobites (there is a sole mention of the word in the MS, in a figure caption) and do not mention the olenelloid marginal suture, they give the reader no opportunity to assess support for this alternative.
At times the paper reads as if the authors are suggesting that olenelloids, which had a marginal cephalic suture broadly akin to that in Limulus, actually lacked a suture that permitted anterior egression during molting. The authors are right to stress the origin of the dorsal cephalic suture in more derived trilobites as a character seemingly of taxonomic significance but lines such as 56 and 67 may be taken by the non-specialist to imply that olenelloids lacked a forward egressionpermiting suture. There is a notable difference between not knowing whether sutures existed (a condition apparently quite common among soft-bodied artiopods) and the well-known marginal suture of olenelloids, but as the MS currently reads most readers will not understand this because it remains unexplained in the MS.
As noted in response to a previous point (above) we now have a greatly expanded introduction which should give the reader an opportunity to assess support for this alternative hypothesis. We now include Olenellus getzi in our analyses, and have added characters to the morphological matrix to make this clear.
A reference to the case of ‘Cedaria’ woosteri is made in the introduction to highlight further the variability of trilobites, as is a reference to Foote’s analysis of cranidial shapes and support this provides for a single origin of the dorsal suture.
With that in mind, it is also worth further stressing that the primary function of the dorsal sutures in those which have them is essentially similar to the olenelloid/limulid marginal suture mentioned above. It is notable that the course of this suture migrated dorsally up from the margin onto the dorsal shield and merged with the circumocular suture, but this innovation does not seem to have had an impact on its primary function - to permit molting by forward egression. Other trilobites completely surrendered the ability to molt by forward egression, and there are even examples of this occurring ontogenetically within species, suggesting a significant intraspecific shift in suture functionality and molting pattern. The authors mention some of this when questioning the unique origin of the dorsal facial suture of trilobites, although I don't understand their argument: why should the history of subsequent evolutionary modification of a character bear on whether its origin was unique in the group?
We include reference to evolutionary modification and loss of this character as it is important to stress that if a character is known to have been lost multiple times it is possible that it had a deeper root (in an earlier diverging member of Artiopoda than Trilobita) and was lost in olenelloids. This is the question that we seek to address in our manuscript.
The bottom line here is that for the ventral plates of A. serratum to be strict homologs of only the dorsal portion of the dorsal free cheeks, there would be no homolog of the trilobite doublure in A. serratum. The conventional view, in contrast, would be that the ventral plates are a homolog of the ventral doublure in all trilobites and ventral plates in artiopods. I do not think that this paper provides a convincing basis for preferring their interpretation, nor do I feel that it does an adequate job of explaining issues that are central to the subject.
We stress that our interpretations – that the ventral plates are not homologous to any artiopodan feature or that they are homologous to the free cheeks of trilobites – have both been raised in the literature before. Whereas we could not find mention of the reviewer’s ‘conventional view’ relating to Acanthomeridion. We appreciate that this view is still valid and worth investigating, which we have done in the further analyses conducted. However, we did not find support for it. Instead we find some support for both ventral plates as homologues of free cheeks, and as unique structures within Artiopoda.
Concern 2. Varieties of dorsal sutures and the coexistence of dorsal and marginal sutures
The authors do not clarify or discuss connections between the circumocular sutures (a form of dorsal suture that separates the visual surface from the rest of the dorsal shield) and the marginal suture that facilitates forward egression upon molting. Both structures can exist independently in the same animal - in olenelloids for example. Olenelloids had both a suture that facilitated forward egression in molting (their marginal suture) and a dorsal suture (their circumocular suture). The condition in trilobites with a dorsal facial suture is that these two independent sutures merged - the formerly marginal suture migrating up the dorsal pleural surface to become confluent with the circumocular suture. (There are also interesting examples of the expansion of the circumocular suture across the pleural fixigena.) The form of the dorsal facial suture has long figured in attempts at higher-level trilobite taxonomy, with a number of character states that commonly relate to the proximity of the eye to the margin of the cephalic shield. The form of the dorsal facial suture that they illustrate in Xanderella, which is barely a strip crossing the dorsal pleural surface linking marginal and circumocular suture, is comparable to that in the trilobites Loganopeltoides and Entomapsis but that is a rare condition in that clade as a whole. The paper would benefit from a clear discussion of these issues at the beginning - the dorsal facial suture that they are referring to is a merged circumcephalic suture and circumocular suture - it is not simply the presence of a molt-related suture on the dorsal side of the cephalon.
We have added in an expanded introduction where these points are covered in detail. We appreciate that this was not clear in the earlier version, and this suggestion has greatly improved our work.
Concern 3. Phylogenetics
While I appreciate that the phylogenetic database is a little modified from those of other recent authors, still I was surprised not to find a character matrix in the supplementary information (unless it was included in some way I overlooked), which I would consider a basic requirement of any paper presenting phylogenetic trees - after all, there's no a space limit. It is not possible for a reviewer to understand the details of their arguments without seeing the character states and the matrix of state assignments.
A link to a morphobank project was included in the first submission. This project has been updated for the current submission, including an additional matrix to treat the reviewer’s hypothesis for the ventral plates. Morphobank Project #P4290. Email address: P4290, reviewer password:
Acanthomeridion2023, accessible at morphobank.org. We have added in additional details for the reviewer and others to help them access the project:
The project can be accessed at morphobank.org, using the below credentials to log in: Email address: P4290, Password: Acanthomeridion 2023.
The section "phylogenetic analyses" provides a description of how tree topology changes depending on whether sutures are considered homologous or not using the now standard application of both parsimony and maximum likelihood approaches but, considering that the broader implications of this paper rest of the phylogenetic interpretation, I also found the absence of detailed discussion of the meaning and implications of these trees to be surprising, because I anticipated that this was the main reason for conducting these analysis. The trees are presented and briefly described but not considered in detail. I am troubled by "Circles indicate presence of cephalic ecdysial sutures" because it seems that in "independent origin of sutures" trilobites are considered to have two origins (brown color dot) of cephalic ecdysial sutures - this may be further evidence that the team does not appreciate that olenelloids have cephalic ecdysial sutures, as the basal condition in all trilobites. Perhaps I'm misunderstanding their views, but from what's presented it's not possible to know that. Similarly, in the "sutures homologous" analyses why would there be two independent green dots for both Acanthomeridion and Trilobita, rather than at the base of the clade containing them both, as cephalic ecdysial sutures are basal to both of them? Here again, we appear to see evidence that the team considers dorsal facial sutures and cephalic ecdysial sutures to be synonymous - which is incorrect.
We appreciate that the reviewer misunderstood the meaning of the dots, leading to confusion. The dots indicated how features were coded in the phylogenetic analysis. In our revised version of this figure (Figure 8 in the new version), these dots are now clearly labelled as indicating ‘coding in phylogenetic matrix’. Further, with the revised character list, we now can provide additional detail for the types of sutures (relevant as we now include more trilobite terminals).
This point aside, and at a minimum, that team needs to do a more thorough job of characterizing and considering the variety of conditions of dorsal sutures among artiopods, their relationships to the marginal suture and to the circumocular suture, the number, and form of their branches, etc.
We thank the reviewer for this summary, and appreciate their concerns and thorough review. Our revised version takes into account all these points raised, and they have greatly improved the clarity, scope and thoroughness of the work.
Reviewer #3 (Public Review):
Summary:
Well-illustrated new material is documented for Acanthomeridion, a formerly incompletely known Cambrian arthropod. The formerly known facial sutures are shown to be associated with ventral plates that the authors very reasonably homologise with the free cheeks of trilobites. A slight update of a phylogenetic dataset developed by Du et al, then refined slightly by Chen et al, then by Schmidt et al, and again here, permits another attempt to optimise the number of origins of dorsal ecdysial sutures in trilobites and their relatives.
Strengths:
Documentation of an ontogenetic series makes a sound case that the proposed diagnostic characters of a second species of Acanthomeridion are variations within a single species. New microtomographic data shed some light on appendage morphology that was not formerly known. The new data on ventral plates and their association with the ecdysial sutures are valuable in underpinning homologies with trilobites.
We thank the Reviewer 3 for their positive comments about the manuscript. We appreciate the constructive comments for improvements, and detailed corrections, which we have incorporated into our revised work.
Weaknesses:
The main conclusion remains clouded in ambiguity because of a poorly resolved Bayesian consensus and is consistent with work led by the lead author in 2019 (thus compromising the novelty of the findings). The Bayesian trees being majority rules consensus trees, optimising characters onto them (Figure 7b, d) is problematic. Optimising on a consensus tree can produce spurious optimisations that inflate tree length or distort other metrics of fit. Line 264 refers to at least three independent origins of cephalic sutures in artiopodans but the fully resolved Figure 7c requires only two origins.
We thank the reviewer for pointing this out. However now the analyses have been re-run we have new results to consider. The results still support multiple origins of sutures. We also note that the dots were indicating how terminals were coded. This is now clearer in the revised version of this figure (Figure 8 in the new version).
We have extended our interrogation of the trees by incorporating treespace analyses. These add support for the nodes of interest (around the base of trilobites), showing that the coding of Acanthomeridion ventral plate homologies impacts its position in the tree, and thus has implications for our understanding of the evolution of sutures in trilobites.
The question of how many times dorsal ecdysial sutures evolved in Artiopoda was addressed by Hou et al (2017), who first documented the facial sutures of Acanthomeridion and optimised them onto a phylogeny to infer multiple origins, as well as in a paper led by the lead author in Cladistics in 2019. Du et al. (2019) presented a phylogeny based on an earlier version of the current dataset wherein they discussed how many times sutures evolved or were lost based on their presence in
Zhiwenia/Protosutura, Acanthomeridion, and Trilobita. To their credit, the authors acknowledge this (lines 62-65). The answer here is slightly different (because some topologies unite Acanthomeridion and trilobites).
The following points are not meant to be "Weaknesses" but rather are refinements:
I recommend changing the title of the paper from "cephalic sutures" to "dorsal ecdysial sutures" to be more precise about the character that is being tracked evolutionarily. Lots of arthropods have cephalic sutures (e.g., the ventral marginal suture of xiphosurans; the Y-shaped dorsomedian ecdysial line in insects). The text might also be updated to change other instances of "cephalic sutures" to a more precise wording.
We appreciate this point and have changed the title as suggested.
The authors have provided (but not explicitly identified) support values for nodes in their Bayesian trees but not in their parsimony ones. Please do the jackknife or bootstrap for the parsimony analyses and make it clear that the Bayesian values are posterior probabilities.
With the addition of further trilobite terminals to our parsimony analyses, the results became poor.
Specifically the internal relationships of trilobites did not conform to any previous study, and Olenellus getzi was not resolved as an early diverging member of the group. This meant that these analyses could not be used for addressing the hypothesis of reviewer two. We decided to exclude reporting parsimony analysis results from this version to avoid confusion.
We have added a note that the values reported at the nodes are posterior probabilities to figures S8, S9 and S10 where we show the full Bayesian results.
In line 65 or somewhere else, it might be noted that a single origin of the dorsal facial sutures in trilobites has itself been called into question. Jell (2003) proposed that separate lineages of Eutrilobita evolved their facial sutures independently from separate sister groups within Olenellina.
We have added this to the introduction (Line 98). Thank you for raising this point.
I have provided minor typographic or terminological corrections to the authors in a list of recommendations that may not be publicly available.
We appreciate the points made by the reviewer and their detailed corrections, which we have corrected in the revised version.
