This is a very nice contribution proposing the first growth model for trilobites. The manuscript also nicely illustrates how model parameters can be estimated from empirical data. More importantly, the model is used to test the relative impact of varying particular parameters on body size as well as assess how estimates differ when subsampling the representation of ontogenetic stage and/or specimen number per stage (or more generally). I would like to recommend it, but i had some minor points which I would like to see implemented in the final version.

The main points are:

1) Introduction: I agree that the Raup papers were major landmarks, but the method is not so well-suited to analyze growth (Urdy, 2015). In this context, it would be worth pointing out here that many studies have focused on invertebrates with accretionary growth and developed newer models which better capture particular growth parameters (Okamoto, 1988a, b; Ackerly, 1989; Urdy et al., 2010a, b).

2) Additional figures. It is appropriate to discuss the different schedules in trilobite (hypoprotomeric, synanthromeric, proarthrous), but it might be worth considering to illustrate the differences graphically and how these would be implemented in the model. Also I feel it might help to have a schematic trilobite in various ontogenetic stages showing/illustrating the main parameters used in the model.

3) One additional species: I understand that for showing and assess your model the dataset of Aulocopleura koninckii is the largest of its kind. However, I do feel that further testing your model would be aided by showing it also works for at least one additional species. This is probably problematic for disarticulated specimens, but is there no datasets suitable to at least try? Your model seems to be quite robust once a certain amount of specimens and ontogenetic stages are represented – at least in the case of Aulocopleura koninckii. This is at least alluded to in text (p. 6, line 6, “the growth rate for both species” – you are only analyzing one?).

4) Meraspid growth gradient. I understand the choice to only fit curves to juvenile stages 9-17, but it might be worth considering to test what would happen with parameters if you would somehow include earlier stages (do they contain additional information or not?)

5) Table 2: in my opinion it would also be worthwhile to tabulate the differences in variance in this or an additional table for the model and empirical data.

6) Limit total number of moulds: you mention N < 31, but before you mention 20 for the holaspis stage (so maybe explicit add maximum to the N to avoid confusion with the median N)

7) Figure 3: I feel it might be worth to directly compare total body length both based on the model results with those measured in the empirical dataset.

These and additional comments can be found in the annotated pdf.

Looking forward to seeing these implemented.

Kenneth De Baets

Suggested references:

Ackerly, S. C., 1989, Kinematics of Accretionary Shell Growth, with Examples from Brachiopods and Molluscs: Paleobiology, v. 15, no. 2, p. 147-164.

Okamoto, T., 1988a, Developmental Regulation and Morphological Saltation in the Heteromorph Ammonite Nipponites: Paleobiology, v. 14, no. 3, p. 272-286.

Okamoto, T., 1988b, Analysis of heteromorph ammonoids by differential geometry: Palaeontology, v. 31, no. 1, p. 35-52.

Urdy, S., 2015, Theoretical Modelling of the Molluscan Shell: What has been Learned From the Comparison Among Molluscan Taxa?, Ammonoid Paleobiology: From anatomy to ecology, Springer, p. 207-251.

Urdy, S., Goudemand, N., Bucher, H., and Chirat, R., 2010a, Allometries and the morphogenesis of the molluscan shell: A quantitative and theoretical model: Journal of Experimental Zoology Part B: Molecular and Developmental Evolution, v. 314 B, no. 4, p. 280-302.

Urdy, S., 2010b, Growth-dependent phenotypic variation of molluscan shells: Implications for allometric data interpretation: Journal of Experimental Zoology Part B: Molecular and Developmental Evolution, v. 314 B, no. 4, p. 303-326.

This manuscript describes and discus a generative mode of trilobite growth, with respect to its segmentation and size changes, based on the Silurian species Aulacpleura koninckii. The results of the manuscript are, however, crucial to our understanding of trilobite body patterning and evolution, in particular to developmental modifications resulting in different trilobite morphotypes. The model generated here shows that nine (or twelve) parameters are needed to model trilobite growth and segmentation. This is important for several reasons. Firstly, it shows that strikingly different morphotypes of the trilobite body can result from changing only a few of these parameters and thus suggesting only simple developmental modifications might be responsible for the evolution of different morphotypes. Secondly, it shows that the trilobite growth is regulated by an array of independent factors and we should be careful when interpreting subjectively intuitive relationships in trilobite ontogeny (i.e. saying that big juveniles must result into big adults and small juveniles into small adults). Moreover, the methods and results have the potential to be useful for generating similar models for other hemianamorhic arthropods. For these reasons, I believe that the manuscript is worth a publication in a high profile journal as it is of interest to a broad readership.

Despite the high quality of the manuscript, I have two minor comments that I believe should be addressed prior to publication:

1) There are few simplifications of the trilobite development which I understand are here in order to generate a simple model. I, however, believe they should be discussed a bit more in the text. These comprise, for example, the statement that one tergite is released in one molt during meraspid period; ignoring largely the protaspid growth; or just briefly mentioning the variability in the holaspid segment no. of the model species.

2) I understand this is a manuscript describing and discussing mostly the particular model. Some of the morphologies generated by changing relative timing of growth gradient or no. of tregites (figs. 4-7) are resembling morphologies of other trilobites, such as illaenids, phacopids, olenellids, etc. I think, that it is a bit pity not mentioning this a bit more in the manuscript in order to stress the importance of the presented model for trilobite evolution.

Sincerely,

Lukas Laibl