Nucleation and Initial Growth of Garnet in Low-Grade Metamorphic Rocks of the Sanbagawa Metamorphic Belt, Kanto Mountains, Japan

Round 1

Reviewer 1 Report

Dear authors,

You will find my comments on the file attached. even if it is just semi-quantitative point analysis that you have done on the garnet, it will be interested to have for each samples an average composition of the garnet for each type (small or coarser). Do you have any data on WR composition of the rocks? This will be interesting in your discussion about the chemical control.

Coudl you also put the sample localities on a map to see the distribution of the samples?

Sincerely yours

Comments for author File: Comments.pdf

Author Response

Thank you for the comments. This is the answer letter to the reviewer

> even if it is just semi-quantitative point analysis that you have done on the garnet, it will be interested to have for each samples an average composition of the garnet for each type (small or coarser).

I was encouraged to show the data so I not only added a table but
performed the whole analysis again for the figures and the table in this
manuscript.

Machine setting and target element was changed so some of the plots
may differ from the first manuscript. Target element was manually fixed
this time so the new data are more consistent.

Figure 7 (profile) and Table 1 (quantitative data) are newly added.
Figure 5, 6, 8 are replace using new data.

> Do you have any data on WR composition of rocks? This will be interesting in your discussion about the chemical control.

I have tried several whole rock chemical composition analyses (XRF)
but I do not think I am ready to publish the data yet. I would like to publish
this manuscript without the whole rock composition data.

The problem is the sample size, I think. The "bulk rock control" I mention
in this manuscript is acturally the "chemical composition of a certain micaceous layer".
The analyses so far is performed using samples with 2~3 cm3 which inevitably
contain "several micaceousd layers".
I have nothing to discuss yet about the whole rock data at the moment.

> Coudl you also put the sample localities on a map to see the distribution of the samples?

I have put the smaller scale sample locality map in Figure 1.
The locality is mentioned in the discussion section concerning
the locality of the larger detrital fragments.

Below are the answer to the reviewer1's comments given in the pdf file.
■line 31 citations in order to illustrate your presentation

I added a reference following reviewer's comment.

■line 49 In the litterature and so far I know, the Sambagawa belt is well-know for its HP metamorphic conditions (related to the subduction), and the eclogite facies that you mentioned approve it. So, why do you write medium?

I revised following reviewer's comment. It was said to be "medium P/T" a little while ago and I just made mistake and wrote that in my old habit.

■line 68 "and characteristic of greenschist facies conditions"

I revised following the reviewer's comment. Besides, I noticed my mistake on the previously estimated P-T condition and revised it. I also did not show reference about it so added reference.

■figure 1 samples distribution on a map will be also of interest

I added a new locality map to show the distribution of the analyzed samples on the outcrop.

■line 101 Ok, but where are used the mineral abbreviations in your text?

The reviewer was surely right, but I added a new graphic in this revision and abbreviations are used in the graphic so left this as it is.

■figure 2 Are organic matter also present on this photo?

Yes. I added black arrows to show the black area where carbonaceous material are. (I have not analyzed them but empirically)

■figure 3 Indicate on a map the sample localities. This will help to see the distribution

I added the sample locality map in Figure 1 so the locality is indicated.

■line 149,168 I will also integrate a table with the average composition of the garnet in the samples fonction of their size. Even if the analysis are just semi-quantitative, this will helph to see the difference in composition of the different part of the garnet.

Thank you so much. I was so encouraged I performed new analyses with new machine settings and all data (Figure5, 6, 7, 8, and Table 1) are added or replaced.

■figure 5 Do you have the same BSE image for the garnet from figure6? It will be interesting also to have it.

I performed new analyses so the same grains were measured for figures 5 and 6.

■line 257,263 Do you have whole rock chemical composition of the rock where you analysed the Grt composition? Do you have any correlation on that? Especially with the next assumption.

As written above, I have attempted whole rock chemical analyses but it is not yet enough to tell us something. I hope to discuss the whole rock data in my next research.

■line 261 Not quantitative, but semi-quantitative as you mentioned it. So you can have a trend.

I added several new graphic to show the chemical trend.

This is all.

Sincerely yours,

Mutsuko

Reviewer 2 Report

The manuscript is aimed to explain nucleation and initial growing process of garnet in respect to the size distribution and compositional zoning in low-grade quartz-rich schist. The idea is fine, and the rocks are suitable to do such experiment, but the data presented are very weak to support any interpretation. The authors claim that small garnet grains are mostly zoned with homogenous core rich in Mn. It is demonstrated by compositional maps of one garnet. However, compositional profiles to support the homogenous distribution of main components in the core is not shown and it is not clear, what number of garnets were analysed. Most data from the literature on garnet with preserved prograde zoning indicate a bill shaped compositional profile for Mn distribution. Gaidies et al (2017) analysed high number of garnets from garnet zone and the amplitude of this zoning profile increases with the decrease of the garnet radius. The homogeneous core composition is theoretically possible, if the bulk rock fractionation is taken in account (Faryad and Jezek, 2019), but this was not supported by observation of measured profile in natural garnet yet.

The large garnet shown is based on the interpretation contains old detrital garnet. But it is not explained why this rim zone is richer in Mn and Ca then mantle part overgrowing the old garnet. If detrital garnet was present, then it was preferably used for further growing of new garnet. This is not the best example taken by the authors to support their idea. The authors statement about formation of homogenous garnet at constant pressure and temperature is not explained.

It is not discussed if the converted 2D crystal size distribution int 3D reflect natural examples. There are many examples of crystal size distribution in three dimensional spaces by x-ray tomography to obtain (see Whitney et al, 2008; Gaidies et al., 2017), which are not discussed.  Even, it is no clarified, how this 3D calculations of CSD help to the problem they are dealing with.

The author concluded that the local occurrence of garnet grains in the Nagatoro area records the beginning of garnet recrystallization controlled by the local chemistry in the thin section scale. This is clear without any CSD, because the rest of the rocks quartzite and garnet formed in mica rich layers. But how the author size distribution experiment contributes to explain this process. There are many other statements, which are not well explained and have no links with the experiment. Therefore, it was hard in many places to understand what they want to tell.

Based on the above comments, the manuscript in the present state is not suitable for publication, but needs new analytical data and clear presentation of the data, used method and interpretations.

Author Response

Thank you for the comments.
This is the answer to the reviewer.

> The manuscript is aimed to explain nucleation and initial growing process of garnet in respect to the size distribution and compositional zoning in low-grade quartz-rich schist. The idea is fine, and the rocks are suitable to do such experiment, but the data presented are very weak to support any interpretation.
> The authors claim that small garnet grains are mostly zoned with homogenous core rich in Mn. It is demonstrated by compositional maps of one garnet. However, compositional profiles to support the homogenous distribution of main components in the core is not shown and it is not clear, what number of garnets were analysed.

Answer:
We had the semi-quantitative EDS data and a lot of BEIs to support our idea but
most of them were not shown. Encouraged by your comment, we performed the whole analyses for the figures in this manuscript again with new settings and target elements to make it as accurate as we can on the EDS.

Figures 5, 6, and 8 are replaced. Figure7 (line profile) and a table is added.
Figure 7 shows the homogeneous profile of Mn-rich core.

As for the number of analyses, garnet grains on 12 thin sections were took BEIs,
BEIs for 30 to 100+ grains on each 4 thin sections (the samples used for the CSD), and BEIs for 10 to 20+ grains on each of the 8 thin sections (used for Figure 9)
Semi-quantitative analyses were newly performed for 6 grains on one thin section (Figures 7, 8)

> Most data from the literature on garnet with preserved prograde zoning indicate a bill shaped compositional profile for Mn distribution. Gaidies et al (2017) analysed high number of garnets from garnet zone and the amplitude of this zoning profile increases with the decrease of the garnet radius. The homogeneous core composition is theoretically possible, if the bulk rock fractionation is taken in account (Faryad and Jezek, 2019), but this was not supported by observation of measured profile in natural garnet yet.

Answer:
Thank you so much about the Faryad and Jezek 2019. This is exactly what we believe we see in this study. Homogeneous core, and jump of garnet formation stage.
(I do not have access to the journal at my office so I did not know).
I added to the reference and referred to it in the discussion.

Most of the prograde zoning in garnet in the Sanbagawa metamorphic belt also have bell shaped zoning of Mn, as you say. However, the BEIs and the profiles shown in this study clearly have plane homogeneous core in the center.
(It is not so clear in the chemical mapping by EDS due to the resolution limit.
The grains are too small. That's why we wanted to show the BEIs)

> The large garnet shown is based on the interpretation contains old detrital garnet. But it is not explained why this rim zone is richer in Mn and Ca then mantle part overgrowing the old garnet. If detrital garnet was present, then it was preferably used for further growing of new garnet. This is not the best example taken by the authors to support their idea. The authors statement about formation of homogenous garnet at constant pressure and temperature is not explained.

Answer:
Our interpretation is that detrital garnet was present but was unstable (not in equilibrium).
When at least the environment reached the garnet stability field, new Mn-rich garnet was formed consuming the existing chlorite. The detrital garnet was surrounded and overgrown.
I think it never had chance to take part in the metamorphic recrystallization.

We added some text in the discussion to explain my view.

> It is not discussed if the converted 2D crystal size distribution int 3D reflect natural examples. There are many examples of crystal size distribution in three dimensional spaces by x-ray tomography to obtain (see Whitney et al, 2008; Gaidies et al., 2017), which are not discussed. Even, it is no clarified, how this 3D calculations of CSD help to the problem they are dealing with.

Answer:
The CSD analyses in this study do not have the information of the locality of grains, they only have sizes and numbers. Therefore we believe it is not possible to discuss together with the spatial distribution data like the x-raytomography data.

What we call "distribution" is not spatial distribution but it means the shape of the histogram.
The size data measured in 2D tend to count larger grains disproportionately, because the smaller grains have less chance to be cut through by random plane. To avoid this cutting effect, the 2D data were computationally converted into the 3D size histograms.

The 3D size histograms helped us to see if the exceptionally large grains are
part of the log-normal garnet population.

We tried adding more explanation in the text to make this point clearer.

>The author concluded that the local occurrence of garnet grains in the Nagatoro area records the beginning of garnet recrystallization controlled by the local chemistry in the thin section scale. This is clear without any CSD, because the rest of the rocks quartzite and garnet formed in mica rich layers. But how the author size distribution experiment contributes to explain this process. There are many other statements, which are not well explained and have no links with the experiment. Therefore, it was hard in many places to understand what they want to tell.

Answer:
As you say, the CSD does not explain the bulk controlled nucleation we see.
We wanted the CSD data to see if the apparently large garnet grains are part
of the garnet population (formed in the same event).

Following your comment, we tried to make the explanation clearer. In particular,
we restructured the discussion part 5-1 to make the story more straight.

> Based on the above comments, the manuscript in the present state is not suitable for publication, but needs new analytical data and clear presentation of the data, used method and interpretations.

Answer:
We added new (semi)quantitative profile data to show the homogeneous Mn-rich core.
We replaced the BEIs and mapping images to data of the same grains for the readers
to compare. We made effort to make the text explanation more clear, especially
about the purpose of the CSD.

This is all.

Yours sincerely,

Mutsuko

Round 2

Reviewer 2 Report

The manuscript is improved, but as only marked changes version was available, it was slightly hard to read it. It would be nice to indicate position of compositional profile on BSE image, rather than show very long number of samples. The compositional jump between core and rim can be explained by depletion of Mn and stabilize of garnet again at higher temperature.

I have still problem to understand, why these rocks with detrital garnet were selected to investigate the nucleation rate or size distribution during a metamorphic event. What are the results of this size distribution? It is note mentioned in the abstract. On the other hand, they write that garnet nucleation was controlled by the distribution of reactant minerals. This is not shown in the manuscript. Maybe they could add a summary chapter „conclusions“ and tell what are the results of this paper.   

Author Response

Thank you for the comments.
This is the answer to the reviewer.

> The manuscript is improved, but as only marked changes version was available, it was slightly hard to read it.

Answer:
I am sorry, I made a mistake. I attach a clear version and a tracking-text-only version in addition to the tracking-everything version.

> It would be nice to indicate position of compositional profile on BSE image, rather than show very long number of samples.

Answer:

Following your advice, Figure 5 (BEI) and 7 (chemical profile) are merged.
Chemical profiles are shown associated to the BEIs.
Section 4-3 was drastically revised following the change of the order of figures.

> The compositional jump between core and rim can be explained by depletion of Mn and stabilize of garnet again at higher temperature.

Answer:
I am happy that you agree with me. It is discussed in the text around lines 259-260.

> I have still problem to understand, why these rocks with detrital garnet were selected to investigate the nucleation rate or size distribution during a metamorphic event. What are the results of this size distribution? It is note mentioned in the abstract.

Answer:
The truth is, I could not find better samples.
I wanted only to investigate the purely nucleated ones but my students (co-authors) found the larger grains with detrital origin in the same outcrop and I thought it was no good neglecting them.
Therefore I decided to refer to the CSD and found out the larger ones have a
different story. Excluding the larger grains, it is possible to investigate the metamorphic nucleation.
I restructured the abstract, added text to the introduction and added the "conclusion" section. In those sections I tried to make things told in the same order (larger garnet -> smaller garnet).

> On the other hand, they write that garnet nucleation was controlled by the distribution of reactant minerals. This is not shown in the manuscript.

Answer:
I agree I used the term "reactant" rather ambiguously.
As I do not show any observation concerning the reactant mineral, I deleted the term "reactant" from the abstract and added explanation in line 293.

> Maybe they could add a summary chapter „conclusions“ and tell what are the results of this paper.

Answer:
Following your advice, I added the section "conclusion" and tried to explain the aim and result of this study.