The manuscript, "Tracking sulfate, magnesium, phosphorus and amorphous phases in Rosalina-like benthic foraminifera", by Paris et al., presents results of correlative TEM and NanoSIMS imaging to track the distribution of sulfur, phosphorus, calcium and magnesium in cultured foraminiferal cells. The foraminifera were cultured within ³³S-labeled seawater to track the incorporation of inorganic sulfate into the test walls.

The paper is generally well-written and the results are novel and exciting. Nevertheless, there are some serious issues with the interpretation of the results. While the interpretation of the sulfur distribution is generally fine, I have problems to agree with the interpretations about the phosphorus distribution. Many studies showed that inorganic phosphorus is completely discriminated in foraminiferal calcite, which is different to corals (see Science paper by Boyle, 2006). Most of the P in foraminiferal calcite is associated with the organic matrix and efficiently removed by bleaching of the tests and in my opinion the results of the study by Paris et al. show exactly this. In addition, there are certain other major points of revision that are in described in detail below. In my opinion, the paper is definitely suited for Biogeosciences, but it needs substantial major revisions before publication.

Below you can find my major points of revision:

1.: As mentioned above, the interpretation of the P distributions in the test walls are problematic and not really supported by the results of this study. Previous studies showed that P associated with foraminiferal calcite either is located in inorganic coatings (Boyle, 2006) or sits in the organic matrix of the test walls (Geerken et al., 2019; Glock et al., 2019). Nevertheless, the authors interpret their results as evidence for P that is bound in the calcite lattice. This is questionable, especially by looking at the plots in figure 4. Figure 4b shows the plot of P vs. CN. CN tracks the distribution of nitrogen in the test and is basically a proxy for organic matter in the test. The plot shows that there are two or three slightly different phosphorus bearing phases in the test but in all these phases P is strongly correlated with organic matter. While this is similar to the correlation between S and CN (Fig. 4a), the intercept with the CN axis for all three different phases is zero or close to zero for the correlation between P and CN (Fig. 4b). This is not the case for the correlation between S and CN where the intercept indicates always significant S concentrations even when nitrogen is absent. In addition, the plot in Fig. 4g indicates, that S and P certainly are sitting in different phases, when normalized to CN. To really show that there is inorganic P within the calcite, I would have recommended to bleach the shells before the analyses to remove the organic matter in the walls. As it is now, I really think the authors have to do some major revisions to their interpretation here. Also, I have a bit of doubt about the interpretation of amorphous calcium phosphate (ACP) phases in the test walls.

Then, the authors found some P-hotspots within the cells, which are partly associated with Ca and Mg. These have been interpreted as acidocalcisomes that have recently been found in foraminifera. In this context, those results could potentially be exciting. Nevertheless, there are some issues with the methodology, since classical drying for the preparation of thin sections for TEM, that the authors used, usually removes soluble content and dissolved ions from the cells (Ayache et al., 2010). This is also stated by the authors in line 174, although there is a reference missing there. That is also the reason why traditional fixation is not recommended, if acidocalcisomes are supposed to be visualized (Goodenough et al., 2019). Goodenough et al. (2019) recommend cryo-fixation to preserve acidocalcisomes and dissolved content. The only study that found evidence for acidocalcisomes in foraminifera so far also used cryo-fixation (Glock et al., 2025). In addition, the authors used a phosphate buffer solution solution for the cell fixation (line 171), which potentially could create artificial P hotspots in the foraminifera. In line 372 it is written: “More phosphorus and sulfur are observed in the organic matter located close to the test compared to the center of the cell.” Could this maybe be an indication that the P originates from the buffer but did not intrude deeper into the cell? Nevertheless, there could be a slight possibility that the P hotspots in the cell are no artifacts, since they have a similar size and shape as the recently found P-bearing structures in foraminifera. This would have been incredibly lucky since acidocalcisomes in foraminifera have not been found yet by using classical TEM-fixation methods. Therefore, I would recommend to tone down the interpretation of the P hotspots in the cell and recommend to add a discussion of the possible artifacts, related to the preparation method of the thin sections.

The next point is not that severe but should maybe be considered by the authors: I was a bit puzzled by the use of the term “Rosalina-like foraminifera”. Either the species that the authors cultured is a Rosalina or not. I can imagine, where this is coming from: Most likely a foraminiferal species that has been optically identified as a Rosalina species has been cultured and later genetic analyses revealed that it is actually not a Rosalina. In this case, I would recommend to give this species (and genus) a name, rather then to use the term “Rosalina-like foraminifera”, which is objectively not correct and misleading.

Moderate points of revision:

The introduction appears to be a bit long and detailed. It almost reads like a review. I have the feeling that the introduction can be shortened quite a bit and/or some parts could be moved to appropriate sections in the discussion. This is not completely obligatory, since Biogeosciences does not have a length limit but keep in mind that the length could deter some readers to read the text in detail.

Minor points of revision:

Line 62: “CO₂“ instead of “CO2”.

Line 70: Add “a” before sulfate reduction pathway.

Line 113: See Boyle (2006) regarding phosphorus discrimination in foraminiferal calcite in contrast to incorporation of P in aragonite of corals.

Line 171: If the authors used phosphate buffer for cell fixation: Couldn´t P distribution in the cell be partly artificial? P in acidocalcisomes is usually lost during cell fixation, based on dehydration that removes free ions in the cell and cryo-fixation is the preferred method (see longer comment above).

Line 220: This transparency regarding the problem of god standard materials is really appreciated. This is often a problem. I think it is indeed very unlikely that the strong fractionation of the S incorporated into the test is caused by artifacts from machine drifts.

Line 244-245: “The laterally heterogeneous lightness or darkness of the grey in the outer calcitic layers is intriguing.” In my opinion this sentence is a bit redundant and intriguing is a bit to pictorial for a scientific paper.

Line 262: “chemically lighter” sounds a bit strange. Do the authors mean that the material is “composed out of lighter elements” or that it is “isotopically lighter”?

Line 264: “show” instead of “shows” (plural)

Line 271: Again “chemically lighter”. See above. Just “lighter” would be also enough, I guess.

Line 302-303: I would write “to the strongest P, S and Mg enrichments in relation to CN” instead of “to the strongest P, S and Mg enrichments compared to CN”.

Line 331: I don´t see that those distributions are very similar in fig. 5. Actually at the edges the distributions are very different. P gets really high at the edges and Mg not. S is in between. Also the minima and maxima in the inner parts are sometimes more or sometimes less pronounced in the different elements. I would suggest to show a correlation plot and test if the correlation is significant. This could very well be the case but it is very hard to see in fig. 5.

Line 351 and 352: Replace “foraminifera” with “foraminifer” in both cases, since it is used for singular here and not for plural.

Line 396: The cited study doesn´t really show “Ca-phosphate” but most likely polyphosphate granules and gels that are associated with Ca in addition to dissolved pyrophosphate.

Line 443: Maybe use “primarily” instead of “first”?

Line 444: Replace “foraminifera” with “foraminifer”, since it is used for singular here and not for plural.

Line 460: “but also to…”. To what? I have the feeling a word is missing in this sentence.

Section 4.1.1: There is a difference between precipitation of calcite and ion exchange between the lattice of already formed calcite in the surrounding medium. I have the feeling that both processes are a bit mixed up in this section.

Sections 4.1.1-4.1.3: I don´t really understand, why these three previously proposed biomineralization mechanisms have to be discussed and reviewed in such a detail in this section. They are partly not really related to the authors results.

Line 510: I think this sentence could be formulated a bit better.

Line 512 and following: Could the mixing of the labeled with the unlabeled P be related to mixture of unlabeled organic S with the labeled inorganic S that is incorporated into the test walls? Would the percentage fit to the plot in fig. 4a? For example in phase 1 it looks like up to 30-50% of the sulfur would sit in the organics. This would roughly fit to your 1/3 of non-labeled S.

Line 514: There is a word missing before “than”. More than? Less than?

Line 528-530: What would that mean? This section is a bit confusing.

Line 567: The use of “on the other hand” needs also “on the one hand”.

 Line 581: “Locally dark local”? Please reformulate.

Line 581-591: See my long comment above about the association of P to organic matter in foraminiferal tests and the discrimination of inorganic P in the calcite lattice. I really do not understand why the authors interpret their results as an evidence for inorganic phosphorus in the test. In my opinion all their results show the opposite and former literature is completely ignored. Especially the older works of Boyle (2006) and the studies that are cited within…

Line 606: Why co-distribution? P and S are not even clearly correlated when they are normalized to CN (the organics). They are only co-correlated, if they are not normalized to CN, because both elements are at least partly associated to organic matter (P more and S to a lesser content).

Line 608: Phospho-aminoacids are not really polyphosphates. Polyphosphates are chained phosphate molecules, which could be completely inorganic or have an organic rest such in ATP.

Line 616: I cannot agree with this hypothesis and don´t see that the data are justifying this hypothesis.

Line 619: See above: The patterns are not similar, if normalized to CN. Thus, P and S are only co-correlated in the organic content.

Line 689: “plays” instead of “play” at the end of the line (singular).

Line 698: How can you design an observation?

References from that review that are not cited in the paper:

Ayache, J., Beaunier, L., Boumendil, J., Ehret, G. & Laub, D. in Sample Preparation Handbook for Transmission Electron Microscopy: Methodology (eds Ayache, J. et al.) 125–170 (Springer, 2010).

Edward A. Boyle: A Direct Proxy for Oceanic Phosphorus? Science 312, 1758-1759 (2006). DOI:10.1126/science.1129723

With great interest I read the manuscript ‘Tracking sulfate, magnesium, phosphorus and amorphous phases in Rosalina-like benthic foraminifera’ by Paris and co-authors. They present interesting data of elemental distribution of foraminifera cultured in controlled conditions enriched in 33S. Although aware of the limitations of their analyses, the work is interesting to the global audience working on biomineralization and element distribution in foraminifera. I do however have some issues that the authors and some inconsistencies that have to be cleared up and fixed before publication, and below I post suggestions of revisions. I would also suggest the authors to present all the maps in the article / supplementary data / data repository, since certain maps are missing from the supplementary data.  

In material and methods, explain why the specimens are called Rosalina-like. Some sentences about the issues with determination would be welcome. Would it be possible to add some SEM photos of the species? I would explain issues with current taxonomy, and I would argue to use the name Rosalina sp. throughout the manuscript. I would remove the species name from the title, and change it for instance into ‘Tracking sulfate, magnesium, phosphorus and amorphous phases in cultured benthic foraminifera’

Line 155: Can you explain the reason for the short duration + the different durations of incubations?

Line 156: this information should be added to table (fed with labelled food)

Line 164: You did not observe it, but it could have been possible?

Fig. 2 - Could the oblique lines indicating additional structures found in the OCL be the edges of cogwheel colums? See Nakajima et al., 2016, van Dijk et al 2020, Cisneros-Lazaro et al., 2022?

Fig. 3 - Panel f, the green lines represent the platinum deposit, but does the top green line not represent the cytoplasm, looking at the 12C14N map?

I have some problems to interpretate figure 4, maybe I am missing something. The figure presents data from 4 profiles on the same chamber. The chamber wall is described in the text (line 294 and onwards) to be an alternation of bands parallel to the chamber wall that have each have a distinct chemical signature ( e.g. high or low ratios). The profiles are taken perpendicular to the foram wall and this banding, should there not in theory be points of each profile in each trend? For example, trend 2 is only observed for profiles 2 and 3. What could cause this heterogeneity?

The foram shown in Figure 3, were not observed to have formed 2 new chamber in the experimental condition – it would be remarkably fast growth rates. Could you derive this from maybe photos taken before and after the incubation? Since there is some 33S enrichment (line 317), could you conclude based on the 33S signal that this n-1 chamber was formed during the experiment? Otherwise, is this enrichment coming from adsorption / isotope exchange? How does it compare to 33S measurements on the A1 specimen – that did calcify for certain.

Line 331-332: ‘similar’ = can you test this?

Line 342+: Since the P/CaO and P/CN profiles show very similar patterns, is this not just proof that you P is coming from organics, especially since your forams did not undergo any cleaning? Could the oblique shape noted in the P/CaO map be caused by being close to a T section of two chambers?

Line 346: when comparing to other studies, caution should be taken, since some of these studies use oxidatively cleaned or uncleaned samples (or both in the case of glock et al., 2019), which

Figure 6: Trend A and trend B are corresponding to profile of another foram? Why not take nanoSIMS profiles and see if the match the profiles of the other specimen used in Fig. 4?)

Line 375: there are white arrows missing in panel 7d and 7i

Line 378: Is the P coming from the chemical fixation of the cell due to the phosphate buffer? Would it no difusse in to the cytoplasm and bind to structures?

Line 371, 381-382: The enrichment of S at the outer edge of the cytoplasm, is it located in the cytoplasm, or is it in the calcite? What limit was used to mask the calcite (or cytoplasm) in all the figures? Is this area not heavily impacted by artifacts at the edges? The edges of the test pores and the edges of the test walls show the strongest topography differences on foraminiferal cross-sections after measurements with NanoSIMS (Geerken et al., 2019), which likely impact the ion yield during the NanoSIMS. Caution should be taken when interpretating  measurements .

Line 392-394: I see the vesicle near the IOL, which has points of increased Ca, but are the same pixels that are rich in Ca, also rich in 33S? Or is the rest of the filling of the vesicle (which is low in Ca) enriched? The d33 shows for me a very similar pattern to the vesicle below, were there is no Ca

Table 2: The table is based on very little data, while you have more observations. Would it be worth to expand the calculation to the other specimens / maps? Also the d33 maps are missing in the supplementary materials.

Line 419: reference to wrong figure? Should be 9? Why are only use these maps, and not the maps from previous images?

Line 445: 7h referes to the panel of 31P. you mean 9h? If I look at the test of F2A2 in figure 7 (7j) there seems to be some enrichment?

Line 450-452: This species is calcifying following a laminar model? Why do we not see the more enriched 33S from the n layer of calcite in the n-1 chamber? Also therefore the measurements of n-1 are a mix of n and n -1 calcite ?

452-3: The food is unlabeled, why does the 33S homogeneity proof there is no sufur from food present? The POS is not exclusively made up from organ matter from the food – or if so, add the reference. See Nagai et al., 2018. Also, does the size of your beams allow for the clear distinction of organic banding?

Line 455: There is no map for d33 in figure 5. You mean fig 9?

Line 456 ‘ new seawater sulfate….’ I do not understand the first part of the sentence

Line 456-60 ‘ Rotaliid foraminifera’, sentence incomplete?

4.1.1-4.1.3 This feels very much as a review, and I think you are going a bit far in the interpretation of your limited data in terms of calcification pathways. I think it is interesting, but more to discuss in the light of how open / closed the system is, since you find enriched cytoplasms in forams that did not calcify, and not focus on incorporation, since it is not cryo nanosims and you have no calcification fluids / vesicles left.

Line 468; I do not understand why you would expect labeled S on the outside of the calcite.

Line 470: You mean you see uptake of 33S in pre-experimental chambers ?– this is in contrast what you say in line 456-457

Line 515: why are these specimens not added in table 2?

Line 564-566: For me, the plots in Fig. 4, in which P/CaO and P/CN show almost the exact pattern proofs the opposite...

Line 660: Already commented on this, but there is more evidence that the POS is formed by the cytoplasm/pseudopods, at least for Ammonia: Nagai et al., 2018 https://doi.org/10.5194/bg-. You would therefore not expect to be low in d33, since the maps show that the cytoplasm is enriched. The weaved POS would exhibit tiny gaps (holes within the framework) that may serve as pathways for passive ion exchange, which explains your high 33S in your cytoplasm

Line 673: This 1/3 is referring to the cell, not the test right?

Line 680: replace ‘grew up’ with ‘formed’

Inconsistencies and typos:

Line 145 wed = fed

Line 138: seawater

Line 118: ang = and

Line 210: consisted in suites = consisted of sets?

Line 264: show

Line 285: point

Line 316: double in

Figures and tables: There is inconsistencies when referring to figures, which led to some confusion on my part to which figure was referred to by the authors. Double check the references.

Table 1: are there 2 specimens called A1? One that was in an experiment of 2 days, and one that was in an experiment of 5 days?

Fig. 2: (2)B1 = F5(2)B1?

Fig 4: Trend 1, 2, 3 = trend A, B etc in the text?

Fig 5: add panels, since you refer to them in text? (eg line 455)

Fig. 7: missing element/title for map 7g and 7j + missing magenta arrow in 7j (for the ca rich vesicle near the IOL

The authors talk about ‘trend A’ or ‘trend B’, this correspond to the numbers in figure 4? If so, rename the trends in the figure / text.

Name of the species: ‘Pseudorosalina Banyuls’ in the supplement, ‘Rosalina-like’ or ‘Rosalina like’ throughout the manuscript