Two Late Cretaceous sauropods reveal titanosaurian dispersal across South America

South American titanosaurians have been central to the study of the evolution of Cretaceous sauropod dinosaurs. Despite their remarkable diversity, the fragmentary condition of several taxa and the scarcity of records outside Patagonia and southwestern Brazil have hindered the study of continental-scale paleobiogeographic relationships. We describe two new Late Cretaceous titanosaurians from Quebrada de Santo Domingo (La Rioja, Argentina), which help to fill a gap between these main areas of the continent. Our phylogenetic analysis recovers both new species, and several Brazilian taxa, within Rinconsauria. The data suggest that, towards the end of the Cretaceous, this clade spread throughout southern South America. At the same locality, we discovered numerous accumulations of titanosaurian eggs, likely related to the new taxa. With eggs distributed in three levels along three kilometres, the new site is one of the largest ever found and provides further evidence of nesting site philopatry among Titanosauria.

INTRODUCTION 2/ The introduction of the manuscript must be improved. Now, it provides a general view of the distribution of titanosaurs in South America, and pay special attention to the absence of saltarurids in Brazil; but nothing is mentioned about the clade that the authors focus on: Colossosauria. Some lines explaining the distribution, abundance or relationship with saltasaurids would be appreciated.
Systematic descriptions are appropriated and arguments to erect new taxa are convincing. Since I'm not an expert in titanosaurian taxonomy I have very little to say in this section: 7/ Line 75: If authors want to be consistent along the text and use American English, then they should use paleontology instead of palaeontology.
8/ Line 144: The figure 2 caption indicates that "circled number correspond to synapomorphies in the text". However, the number 1 is marked in Fig2a, which corresponds to a cervical vertebra (C12), while the first synapomorphy, according to the text, is referred to middle dorsal vertebra (D6-D7) as correctly indicated in Fig2d. The miss-indication on figure should be emended. 9/ Line 223: "centra is spongy" change for "cantra are spongy" 10/ Lines 295-302: The division of the articular condyle of the quadrate is a very peculiar character indeed. I do not doubt of the validity of this character, but it is difficult to tell just from a picture; could you provide a better (perhaps more contrasted) picture? Can you ensure that this feature is not produced by taphonomic means? When comparing this character, what is the condition of the articular process in South American titanosaruians Tapuisaurs and Sarmientosaurus? 11/ Lines 302-307: I am unable to recognize the quadratojugal in Fig.3a.
Is it like I tried to indicate in the figure?
Even if the quadratojugal extends more backward, how can the authors infer the projection of the anterior process of the bone? This region is not preserved in the figured specimen.
12/ Lines 374-375: In addition to the current comparison, the deltopectoral crest is also markedly expanded distally in Dreadnoughtus and Opisthocoelicaudia, just if authors want to include this information to the description. 13/ According to Fig. 1 both Punatitan and Bravasaurus preserve rib elements. One significant synapomorphy among titanosaurians is the occurrence of pneumatic foramen at the proximal end of the dorsal rib (see Mocho et al 2019). Is this feature present in the new taxa? Mocho, P., Pérez-García, A., Martín Jiménez, M., & Ortega, F. (2018). New remains from the Spanish Cenomanian shade light on the Gondwanan origin of European Early Cretaceous titanosaurs. Cretaceous Research. doi:10.1016/j.cretres.2018.09.016 14/ Lines 389-394: Assuming that BM estimation is correct, 2.9 tones is a really low value for a titanosaure. The purported "dwarfs" titnosaurians Ampelosaurus from France and Lirainosaurus from Spain are estimated to be 2.5 and 1.8 tones respectively (see Sup. Excel file of Bensot et al. 2012). I realize that it is not the main scope of the paper, but I think that, apart of solely mentioning that Bravasaurus is the smallest colossosaurian yet recovered (lines 479-480), authors should propose some hypothetical scenario that helps to explain the occurrence of such small "titano" among colossal beast.

Phylogenetic analysis and biogeographic distribution
The phylogenetic results are convincing and help to depict a remarkable paleobiogeographic pattern. It is noteworthy the huge effort that the author made in re-codding several taxa, which sure help improved the final results. However, I have some comments for this section: 15/ Although Bremer support index is provided in de Supplementary Material, I miss the report of other important phylogenetic indices, such as RI and CI. This information is useful to give an idea of the "robustness" of the results. 16/ Fig.4 I recommend to somehow highlighting the position of both Punatitan and Bravasaurus in the phylogeny, just to make visually easier its location.
17/ Since TNT is unable to establish time calibrations, to me, it remains unclear how authors established the time of divergence (nodes) between groups/taxa in the phylogeny. Did you consider any specific ratio of speciation/evolution? Did you place them according previous studies? This part should be improved given that it should affect parts of the discussion (Lines 481-493).
18/ I have to admit that I am not familiar with the paleogeographic characteristics of South America at the end of the Cretaceous (Campanian-Maastrichtian), and therefore I have some questions about this topic and how author established the criteria to conduct the paleobiogeographic analyses.
19/ If I am right, authors establish a "current" latitudinal gradient in order to define the location of different titanosaure taxa. But, does the geographic position (latitude/longitude) of South America have remained invariable for the last 80-66 Ma? If the answer is yes (or almost negligible) then I agree that the current location of the fossil sites can be assumed as the same as that from 70 Ma. Otherwise, if the South American continent experienced any significant displacement on the lithosphere during the last millions of years, then the geographic position should be corrected and the paleogeography analysis re-run considering past latitude/longitude locations. QSD Nesting Site 20/ Despite its remarkable significance, it seems that this section is somehow off place and disconnected from the main porpoise of the study, which -as far as I could understand -seems to be focused on provide evidence of the wide geographic distribution of the clade Colossosauria. 21/ However, since dinosaur eggs and reproduction is my main topic of study, I have some comments for the authors about this section.
22/ By contrasting the information provided in both the main text and SI, I found some problems to set a clear picture of the occurrence of titanosaure eggs in terms of geographic and stratigraphic distribution. Following I list the three main issues that I found: 1) Stratigraphic distribution 23/ The distribution of the 3 egg-bearing strata needs some clarification. In the main text it is mentioned that they outcrop within a 30 m thick sequence, which I suspect that it refers to the "grey interval" compressed between 45 and 70-75m above the base of Ciénaga del Río Hueco Fm. (according to Fig S1b), although no indication is provided in this regard. On the other hand, in the Supplementary Information it is said that they are distributed in a 5 m thick interval, that according to the text it is located 50-65 m above the base of the formation. First, that is actually 15 m interval not 5m, and second the interval is located around 60-65m from the base according to Fig. S1b.
So, what is the real stratigraphic distribution of the 3 egg-bearing levels? Please, improve this part.
24/ In lines 429-430 it is stated that eggs and eggshells occur at different (stratigraphic?) positions within the floodplain depositional sequence and, therefore, it is likely to interpreted that they represent several nesting events. However, when the authors say "several nesting events", would they like to mean "three", as the number of strata? How the authors can establish the number of nesting events? Additionally, could it be that eggs with different positions within the profile could represent eggs of the same clutch but in a distinct internal location (in line 441 it is mentioned that eggs can occur in two-row arrangement) rather than different events?
2) Lateral extension of the nesting site 25/ The way in which the information is provided in the main text (lines 425 to 427) seems to suggest that the accumulations of eggs can be traced laterally over three kilometers, which would imply a humongous nesting area. However, information from Supplementary Information seems to point that it is the stratigraphic interval (lithology/strata) the one that can be traced for more than 3 km, which leads a total different interpretation.
Thus, I suggest clarifying this point: Where egg remains recovered along 3 km? Or were they found in a single site (spot)?
26/ In addition of all of the above, it would be useful to provide the number of clutches, eggs, and eggshells discovered in the new nesting site, just to give a clear idea of the abundance and richness of the locality.
3) Time occurrence 27/ Authors seem to be somehow conditioned for the Auca Mahuevo nesting site, and omit many other evidences of "nesting site fidelity" in titanosaure sauropods. Additional examples of nesting recurrence in a single site can be found in both South America (see Salgado et al., 2007Salgado et al., , 2009 and Europe (Sellés et al., 2013, 2017. I recommend including those additional references. Salgado L, Coria RA, Ribeiro CM, Garrido A, Rogers R, Simón ME, Arcucci AB, Rogers KC, Carabajal AP, Apesteguía A, et al. 2007 28/ Line 432-433: It seems that "Auca Mahuevo" is repeated twice within the same phrase. "as reported for Auca Mahuevo the world-famous Auca Mahuevo nesting site, in the Argentine Patagonia" It should be: "as reported for the world-famous Auca Mahuevo nesting site, in the Argentine Patagonia" 29/ Line 440: Authors stars the identification of QSD eggs mentioning that they share many features with those discovered in the Auca Mahuevo nesting site. But in my opinion the description of the QSD eggshell is so general that such features also agrees with several egg-types from India, Spain, France, Romania, Brazil, and several site from Argentina (i. ex. Salitral Rosa, Salitral Oje de Agua, Salitral Moreno, or Neuquen City). Thus, I suggest developing this section a little more. Apart from the shell thickness, shell unit shape, and pore canal system, authors should focus the new description and discussion on two additional features: 1) shape and contact between adjacent shell units, and 2) H:W ratio of the shell units.
1) shape and contact between adjacent shell units. According to the most recent reviews on megaloolithid eggshells (see Fernández and Koshla, 2017 for further details) the way in which adjacent shell units growth allows to distinguish to main groups of megaloolithid eggs: Megaloolithus-with well-defined and straight edges; and Fusioolithus-with fused edged. Eggs from Auca Mauevo have been included in the later group (Fernández and Koshla, 2017), while those from QSD site seems to posse well-defined shell margins (personal interpretation based on Fig.5d). I recommend the authors to take a look at the samples from the studies of Salgado et al (2007).
2) H:W ratio of the shell units. This feature is commonly used to discriminate between "oospecies" of megaloolithid eggs (see Fernández and Koshla, 2017). It seems that in QSD shells this values is about up to 2.5:1 or so, which is much closer to the specimens from Neuquen City (Megaloolithus jabalparuensis) than that of Auca Mahuevo (Fusioolithus bahensis). I believe that, by following my recommendations and processing a little bit the oological information, author should be able to extract some additional paleobiogeographic conclusions that may help to support their hypotheses.
Moving to the discussion section reserved for QSD eggs (starting from Line 494), this part must be largely improved. 30/ Line 496: As aforementioned, the features exhibited by QSD eggs are common for several sites from South America (Peru, Brazil, Argentina, Uruguay), Europe (Spain, France), India, and even Africa (Morocco, Tanzania); so to restrict the comparison to Auca Mauevo and Totesti seems too poor. 31/ Line 497: I do not see the porpoise to compare the eggs from QSD to those from Los Llanos region, especially because they belong to a different type of sauropod eggs. Those from La Rioja belong to the group of Faveoloolithus (extremely think eggshells with complex reticulate pore system), while it seems that those from QSD are attributable to Megaloolithus (more thin eggshell with simple pore system). 32/ Lines 500-504: The size of the egg is not a valid criterion for comparison. Eggs of the same eggshell-type can greatly vary in size (see Vianey-Liaud et al., 2003, Fernández andKhosla, 2014). Although the size of eggs from both QSD and Totesti can be similar, this is not enough to use this criterion to support the small body size of new taxa. Eggs of 14 cm in diameter are also reported from Auca Meuevo and several localities from India, and coeval titanosaurian taxa are mid-to large-size. Reviewer #3 (Remarks to the Author): The paper is interesting and the materials are important. All comments are made via track changes in the attached document. I combined the supplemental info and main text into one document. More justification needs to be present in the main text; as it stands, so much is relegated to the supplemental information, the main paper doesn't stand on its own. The phylogenetic analysis unfortunately doesn't appear to have been very carefully done. I did not go through each and every character (let alone each character scoring), but looking over the list even briefly, I found a duplicate (copied and pasted) character, typos, and potential overlap among characters put forth as disparate. These issues will not give the reader confidence in the results of the phylogenetic analysis, if it is left in. I don't think a phylogenetic analysis is necessary, you could just note synapomorphies present in the material based on previous analyses. This would free up room for describing the provenance, age, and taphonomy of the specimens, all of which are severely underreported in the main text (but present in nice detail in the supplemental information). In sum, I think these are very important fossils and congratulate the authors for the huge undertaking involved in bringing them to light, but I feel that there are substantial issues that prevent publication of the manuscript in its present form.

INTRODUCTION
Lines 41 and 44_These are rather old references. Reply: We added most of the suggested references that improved the manuscript.
Line 53_Yamanasaurus now as well -cited in the supplementary information but not here, for some reason.

Reply:
We added Yamanasaurus plus the respective reference.

RESULTS
Line 116_The centrum is depressed, with… R#1: In what way? Do you mean it is shorter dorsoventrally than it is wide transversely?
Reply: We modified the text accordingly. Now it is: The centrum is shorter dorsoventrally than it is wide transversely, with… Line 129_spinoprezygapophyseal fossa (sprf). R#1: Following Wilson et al. (2011), this is the sprf, not the sprlf.

Reply:
The reviewer is right. We slightly modified the sentence: As in most titanosaurians, these caudal vertebrae have strongly procoelous centra .
Lines 236-238_The haemal arches are similar to those reported for other derived titanosaurians.
R#1: So…? I presume they're open dorsally, with blades that are not expanded anteriorly or posteriorly. But how can I know? Perhaps describe at least a little.

Reply:
The haemal arches are not the most diagnostic bones and we have a word limit for this submission. We added some information about them: The available haemal arches are opened Y-shaped, with no expanded pedicels, as those reported for other derived titanosaurians (Otero et al., 2011).
Lines 267-271_The holotype of Bravasaurus (Fig. 3), as well as the referred specimen, indicates a small-sized titanosaurian, much smaller than Punatitan (Fig. 1c Reply: We agree with the reviewer. We added the following sentence: Considering that both specimens could be adults (see below), they would be similar to Neuquensaurus or Magyarosaurus.
Lines 271-273_Cranial elements include partial right quadrate and quadratojugal ( Fig.  3a,b). The quadrate is anteroventrally directed and bears part of the quadrate fossa. R#1: It is hard to make much of either of these elements from the figure. Please annotate and/or provide more views. Maybe dash in the outline of the complete element.

Reply:
We modified the Fig. 3. Now the lateral and ventral views are enlarged and we added an interpretative drawing for the lateral view and more labelling.

Reply:
We deleted "as in Titanosauria", as it is already stated that procoelous caudals are frequent in this clade in lines 234-235. Now it reads: The preserved anterior and middle caudal vertebrae of Bravasaurus are procoelous.
Lines 337-338_The neural arches are on the anterior portion of the centra, as in most titanosaurians. R#1: This character is more widespread within Sauropoda -for example, Wintonotitan shows it.

Reply:
We totally agree with the reviewer. However, as space is limited, we restricted the comparisons to the titanosaurian in-group. We slightly modified the sentence: The neural arches are on the anterior portion of the centra, as in most titanosaurians, and some other titanosauriforms (e.g. Wintonotitan).

Reply: Modified. Its robustness is high (RI(Wilson and Upchurch, 2003) =0.35), as in
Opisthocoelicaudia (Borsuk-Białynicka, 1977), Diamantinasaurus (Poropat et al., 2015), and Savannasaurus (Poropat et al., 2016), much more than in Reply: Yes indeed, Atacamatitan is not far from NW Argentina but it is not conclusive that it is a colossosaurian. In lines 458-461, we already provided more information regarding this Chilean species and other fragmentary findings in NW Argentina, which, as best, can be referred as non-saltasaurine titanosaurians. In addition, Yamanasaurus is considered a saltasaurine, and its record is far away from NW Argentina. Distance between Ecuador and NW Argentina (c. 3000 km) almost doubles the distance between the latter and Patagonia (c. 1500 km). Reply: We agree with R#1. We modified the figure accordingly, by adding a picture of the cervical vertebra in anterior view (Fig. 2a).

SUPPLEMENTARY INFORMATION
Lines 38-47_These ages are much older than Maastrichtian… is that not problematic? Reply: The age of Ciénaga del Río Huaco is controversial. The outcrops at QSD seem to correlate with the uppermost sections of southern exposures in La Rioja and San Juan provinces. We modified the text between lines 38 and 47, to clarify on this. 1/ I do not pretend to be fussy, but (like some reviewers already did to me) I have a general observation in the usage of the term "titanosaur" or "titanosaurs". Despite being widely (commonly and colloquially) used, the term titanosaur -referring to members of the clade Titanosauria -is incorrectly used, and should be avoided.

Supplementary
Reasons for such proposal are base on the fact that the word "titanosaur" derives from the genus Titanosaurus, so it makes reference sensu stricto to an individual of this specific genus. However, Wilson and Upchurch (2003) considered Titanosaurus as nomen dubium, and consequently the derived term also loosed its validity.
In this regard, according to the ICZN, the most similar and valid nomenclatural term is: Titanosauria (Bonaparte and Coria, 1993), and as such the derived term should be titanosaurian or titanosaurians.

INTRODUCTION 2/
The introduction of the manuscript must be improved. Now, it provides a general view of the distribution of titanosaurs in South America, and pay special attention to the absence of saltarurids in Brazil; but nothing is mentioned about the clade that the authors focus on: Colossosauria. Some lines explaining the distribution, abundance or relationship with saltasaurids would be appreciated.

Reply:
We introduced three phrases in order to provide information about the clade Colossosauria (lines 58-64).
3/ Although it can be deduced, there is no direct reference of the institution where the Punatitan and Bravasaurus type material is housed. This observation leads me to realize that there is no specific "Institutional abbreviation" section, so the acronym CRILAR is only resolved in the authors affiliation. Therefore, I suggest including a line solving all these questions: All martial descried in this study is housed in the collections of Centro Regional de Investigaciones Científicas y Transferencia Tecnológica de La Rioja (CRILAR; La Rioja, Argentina). Reply: We clarified this by adding "Paleovertebrate Collection of Centro Regional de Investigaciones Científicas y Transferencia Tecnológica de La Rioja, La Rioja, Argentina" next to the first reference of CRILAR-Pv (lines 85-87). Also added a sentence in Methods section (lines 523-524).
4/ However, although it significance, I do not see the purpose to include the report of the new nesting site given that it seems to have no special relation with the main goal of the study. If authors want to keep this part, they should find a way to incorporate the occurrence of new egg remains to the general discussion of titanosaure geographic distribution or, if any prove, to link the eggs to any of the two new species.
I added some specific comments regarding this section in the comments for the authors.
Reply: Yes, we want to keep this part and some sections of the manuscript were modified to address this issue.

Systematic paleontology
Systematic descriptions are appropriated and arguments to erect new taxa are convincing. Since I'm not an expert in titanosaurian taxonomy I have very little to say in this section: 7/ Line 75: If authors want to be consistent along the text and use American English, then they should use paleontology instead of palaeontology.

Reply:
We have to keep it in British English.
8/ Line 144: The figure 2 caption indicates that "circled number correspond to synapomorphies in the text". However, the number 1 is marked in Fig2a, which corresponds to a cervical vertebra (C12), while the first synapomorphy, according to the text, is referred to middle dorsal vertebra (D6-D7) as correctly indicated in Fig2d. The miss-indication on figure should be emended.

Reply:
We emended it. 9/ Line 223: "centra is spongy" change for "cantra are spongy" Reply: The subject is the tissue and not the centra (line 218).
10/ Lines 295-302: The division of the articular condyle of the quadrate is a very peculiar character indeed. I do not doubt of the validity of this character, but it is difficult to tell just from a picture; could you provide a better (perhaps more contrasted) picture? Can you ensure that this feature is not produced by taphonomic means? When comparing this character, what is the condition of the articular process in South American titanosaruians Tapuisaurs and Sarmientosaurus?
Reply: We addressed this issue by introducing the following modifications on the original Fig. 3: a) We split the pictures in two new figures, one for the axial skeleton and one for the appendicular skeleton; b) we enlarged the pictures of the cranial bones in both lateral and ventral views; c) we added an interpretive drawing, since the preservation makes it difficult to observe the material. Regarding Tapuiasaurus and Sarmientosaurus, both are articulated. On the other hand, there is no specific information on this character in the publications and we have not made direct observations on these taxa. Even if the quadratojugal extends more backward, how can the authors infer the projection of the anterior process of the bone? This region is not preserved in the figured specimen.

Reply:
We addressed this issue in the last point.

12/ Lines 374-375:
In addition to the current comparison, the deltopectoral crest is also markedly expanded distally in Dreadnoughtus and Opisthocoelicaudia, just if authors want to include this information to the description.
Reply: We added the information (now line 354). Reply: Ribs, as well as haemal arches are not particularly relevant for ingroup comparisons and we have a word limit for this paper. We provide several characters that support the titanosaurian affinities of Punatitan and Bravasaurus. Considering the limited space available we will publish that information elsewhere along with detailed osteology of both new taxa.

13/
14/ Lines 389-394: Assuming that BM estimation is correct, 2.9 tones is a really low value for a titanosaure. The purported "dwarfs" titnosaurians Ampelosaurus from France and Lirainosaurus from Spain are estimated to be 2.5 and 1.8 tones respectively (see Sup. Excel file of Bensot et al. 2012). I realize that it is not the main scope of the paper, but I think that, apart of solely mentioning that Bravasaurus is the smallest colossosaurian yet recovered (lines 479-480), authors should propose some hypothetical scenario that helps to explain the occurrence of such small "titano" among colossal beast.
Reply: Yes, indeed, it is a low BM value. Although the information provided by the vertebrae is convincing regarding the adult size of Bravasaurus, we consider that it would be a bit premature to venture possible scenarios that explain its size. In addition, the text would extend into something that, as the reviewer clarifies, is far from the main scope of the paper. We added few sentences comparing Bravasaurus with dwarf sauropods from Europe (lines 372-375 and 466-468).

Phylogenetic analysis and biogeographic distribution
The phylogenetic results are convincing and help to depict a remarkable paleobiogeographic pattern. It is noteworthy the huge effort that the author made in recodding several taxa, which sure help improved the final results. However, I have some comments for this section: 15/ Although Bremer support index is provided in de Supplementary Material, I miss the report of other important phylogenetic indices, such as RI and CI. This information is useful to give an idea of the "robustness" of the results.

Reply:
We added a new paragraph in the "Phylogenetic analysis" section of the Supplementary Information. It includes RI and CI values. Fig.4 I recommend to somehow highlighting the position of both Punatitan and Bravasaurus in the phylogeny, just to make visually easier its location. Reply: Done.

16/
17/ Since TNT is unable to establish time calibrations, to me, it remains unclear how authors established the time of divergence (nodes) between groups/taxa in the phylogeny. Did you consider any specific ratio of speciation/evolution? Did you place them according previous studies? This part should be improved given that it should affect parts of the discussion (Lines 481-493).
Reply: It seems that the term "time calibrated phylogeny" led to a wrong interpretation. We did not establish times of divergence between groups. The phylogenetic tree only contains occurrences of each taxon, according to published data. We modified the figure caption.
18/ I have to admit that I am not familiar with the paleogeographic characteristics of South America at the end of the Cretaceous (Campanian-Maastrichtian), and therefore I have some questions about this topic and how author established the criteria to conduct the paleobiogeographic analyses.

Reply:
We have not carried out any specific palaeobiogeographic analysis. According to the palaeogeographic maps from Scotese and Wright (2018; see Fig. 1a  American continent experienced any significant displacement on the lithosphere during the last millions of years, then the geographic position should be corrected and the paleogeography analysis re-run considering past latitude/longitude locations.

Reply:
The reviewer is right; we used a modern latitudinal gradient. Using palaeolatitudes does not represent a significant difference because South America moved more or less homogeneously since the Cretaceous. Much of that continental drift was due to the opening of the Atlantic Ocean, and therefore, the movement was mainly longitudinal. The latitudinal data barely vary between 3 and 5 degrees of latitude (data now included in Supplementary Information, Table S4). We modified the data and former Fig. 4 (now Fig. 5) so that they fit the palaeolatitude values. As you could see below, at this scale, the colour differences between the previous and new figure are minimal. Palaeolatitudes were calculated using the open source software GPlates (Seton et al., 2012;Müller et al., 2018).
As already stated, we didn't run a specific palaeobiogeographic analysis. The phylogenetic tree shows the occurrence of each taxon using a colour code referred to their palaeolatitudinal position ( Supplementary Information, Table S4). The colour coding facilitates the visualization of latitudinal differences between clades of derived titanosaurians. Thus, at a glance, it is possible to identify that the Lognkosauria, for example, is a strictly Patagonian clade, while Bravasaurus has closer affinities with taxa that inhabited lower latitudes in Brazil.
We modified the "Latitudinal position of South American taxa" section of the Supplementary Information accordingly.
QSD Nesting Site 20/ Despite its remarkable significance, it seems that this section is somehow off place and disconnected from the main porpoise of the study, which -as far as I could understand -seems to be focused on provide evidence of the wide geographic distribution of the clade Colossosauria.

Reply:
We regret that the section on the nesting site has seemed disconnected from the central purpose of the paper. This might be related to the difficulty in assigning the eggs to a clade within Titanosauria. We made some modifications to the text to better contextualize the results. 21/ However, since dinosaur eggs and reproduction is my main topic of study, I have some comments for the authors about this section.
22/ By contrasting the information provided in both the main text and SI, I found some problems to set a clear picture of the occurrence of titanosaure eggs in terms of geographic and stratigraphic distribution. Following I list the three main issues that I found: 1) Stratigraphic distribution 23/ The distribution of the 3 egg-bearing strata needs some clarification. In the main text it is mentioned that they outcrop within a 30 m thick sequence, which I suspect that it refers to the "grey interval" compressed between 45 and 70-75m above the base of Ciénaga del Río Hueco Fm. (according to Fig S1b), although no indication is provided in this regard. On the other hand, in the Supplementary Information it is said that they are distributed in a 5 m thick interval, that according to the text it is located 50-65 m above the base of the formation. First, that is actually 15 m interval not 5m, and second the interval is located around 60-65m from the base according to Fig. S1b. So, what is the real stratigraphic distribution of the 3 egg-bearing levels? Please, improve this part.

Reply:
The 30 m thick sequence referred in the main text is the stratigraphic interval of laminated siltstones and mudstones that contains the titanosaurian eggs. The three egglevels, clearly separated stratigraphically, occur in a 5 m interval within this sequence. The reviewer is right; there was a typing error in the Supplementary Information, which we have corrected. To avoid confusion we modified the main text and Supplementary Information: Main text (lines 404-406): The egg-clutches and eggshells are included in an interval of floodplain deposits in at least three distinct but closely spaced horizons at 59.2, 62.8 and 63.9 m above the base of the unit (Supplementary Fig. 1b).
Supplementary Information (292-302): Egg shells and egg clutches were found in at least three distinct but closely spaced horizons in a 5 m thick interval located 59-64 m above the base of Ciénaga del Río Huaco Formation (Supplementary Fig.  1b).

24/
In lines 429-430 it is stated that eggs and eggshells occur at different (stratigraphic?) positions within the floodplain depositional sequence and, therefore, it is likely to interpreted that they represent several nesting events. However, when the authors say "several nesting events", would they like to mean "three", as the number of strata? How the authors can establish the number of nesting events? Additionally, could it be that eggs with different positions within the profile could represent eggs of the same clutch but in a distinct internal location (in line 441 it is mentioned that eggs can occur in two-row arrangement) rather than different events?
Reply: The egg-bearing horizons at QSD are 59.2, 62.8 and 63.9 m above the base of Ciénaga del Río Huaco Formation. Within each level there are small vertical variations between different egg accumulations, and the eggshells often appear scattered at intervals several tens of centimetres thick. The floodplain sediments usually show evidences of pedogenesis. Soft sediment deformation and dislocation are frequent and we have documented them in QSD. As already documented for Auca Mahuevo (Jackson et al., 2013), the three egg bearing levels at QSD cannot be interpreted as three "nesting seasons". Each one could constitute a time-averaged assemblage. We modified the text to clarify this.
2) Lateral extension of the nesting site 25/ The way in which the information is provided in the main text (lines 425 to 427) seems to suggest that the accumulations of eggs can be traced laterally over three kilometers, which would imply a humongous nesting area. However, information from Supplementary Information seems to point that it is the stratigraphic interval (lithology/strata) the one that can be traced for more than 3 km, which leads a total different interpretation. Thus, I suggest clarifying this point: Where egg remains recovered along 3 km? Or were they found in a single site (spot)?
Reply: The egg bearing strata are exposed along 3 km, and eggshells have been regularly found over more than 2 km of this stratum so far. We modified the main text (lines 410-413), see answer to the next comment (n°26).

26/
In addition of all of the above, it would be useful to provide the number of clutches, eggs, and eggshells discovered in the new nesting site, just to give a clear idea of the abundance and richness of the locality.

Reply:
We completed the text as follows: QSD Nesting Site. We documented three egg-bearing levels in the lower section of Ciénaga del Río Huaco Formation at QSD. The egg-clutches and eggshells are included in an interval of floodplain deposits in at least three distinct but closely spaced horizons at 59.2, 62.8 and 63.9 m above the base of the unit (Supplementary Fig. 1b). Fossil-bearing rocks are siltstones and sandy siltstones with horizontal lamination and graded and massive bedding that form thin tabular sheets, extending for tens to hundreds of meters. Some layers contain calcium carbonate as root encrustations and small soil nodules. This fossiliferous layer is laterally traced over more than three kilometres and the egg-clutches and eggshells (CRILAR-Pv 620-621) are exposed regularly all along it. Nineteen egg clutches were spotted, one with up to 15 sub-spherical eggs, arranged in two superposed rows.
3) Time occurrence 27/ Authors seem to be somehow conditioned for the Auca Mahuevo nesting site, and omit many other evidences of "nesting site fidelity" in titanosaure sauropods. Additional examples of nesting recurrence in a single site can be found in both South America (see Salgado et al., 2007, 2009) and Europe (Sellés et al., 2013, 2017. I recommend including those additional references. Salgado L, Coria RA, Ribeiro CM, Garrido A, Rogers R, Simón ME, Arcucci AB, Rogers KC, Carabajal AP, Apesteguía A, et al. 2007 The term nest site fidelity is frequently used to describe nesting of modern reptile and bird species. In the literature, it is often used as a synonym for breeding-site philopatry and refers to the return of an individual or group of individuals of a species to the same location to breed. As the reviewer clarifies, there are numerous sites in the world with multiple eggbearing levels. However, in most cases, systematic allocation relies on parataxonomic comparisons. This nomenclatural system can obscure phylogenetic relationships (e.g., Grellet-Tinner et al., 2012;Jackson et al., 2013;Mikhailov 2019). In contrast, only a few sites such as Auca Mahuevo (Argentina), Toteşti (Romania) or Dholi Dungri (India) preserve embryonic remains in ovo, which confirm their assignment to Titanosauria. These sites represent, therefore, the most reliable source for morphological comparison within the framework of phylogenetic systematics. At Auca Mahuevo there are four egg-layers, three of which could correspond to the same species of titanosaurids (Hechenleitner et al 2015). We appreciate the suggestion to incorporate more citations. However, since we have already exceeded the maximum references for this journal, we try to keep those that could be more relevant for comparison. sauropod dinosaurs. However, among such diversity, several parameters such as the shell thickness, the diameter of the ornamentation, the size of the eggs and the morphology of the pore canals, show remarkable similarities with the eggs from layers 1-3 of Auca Mahuevo. In the present version of the manuscript we extended the comparisons with confirmed records of titanosaurian eggs from India and Romania (e.g., Wilson et al., 2010;Grellet-Tinner et al., 2012) as well as some records that provide consistent information, such as those from Spain. The Brazilian record is fragmentary, but previous comparisons show strong similarities with the Auca Mahuevo site (Grellet-Tinner and Zaher, 2007). We added this information to the discussion. We avoid the description and comparisons of the characters suggested by the reviewer because of the following reasons: 1) Shape of the eggshell units: It can vary between samples of the same site (personal observations EMH and LL). Some of these variations are related to diagenetic processes, both dissolution and recrystallization (e.g., Grellet-tinner et al., 2010;Mikhailov, 2019). Regarding the apparent differences between the parataxonomic families mentioned by the reviewer, Fernández and Khosla (2014, p. 10) state: "The shell units [of Fusioolithidae] are fan shaped similar to the eggs of oofamily Megaloolithidae but it differ in the nature of the eggshell units in which they are partially fused." In addition to vagueness in definition, the authors do not give nor suggest reasons that could explain such variations in terms of biological adaptation. As mentioned above, parataxonomy does not bring an appropriate framework for taxonomic identification, which is the main purpose here.
2) The H:W ratio of the shell units does not provide new relevant information. As we mentioned in the main text, titanosaurian eggshells are mono-layered, which means that the shell units make up the total thickness of the eggshell itself. They even shape the characteristic ornamentation, known as nodular. It would be risky to measure the width of the shell units in a radial section (as the reviewer suggests), since this section would hardly show a perfect radial cut of the units themselves. Furthermore, the thin section would have to provide enough perfect radial sections to be able to obtain a statistically significant value (Fig. 6d of the main text, illustrates the possible size variations for the shell units). Tangential sections are needed to measure the diameter of (many) shell units, and the solution for this is a bit tricky. As the diameter of the shell units increases outward, Fernández and Khosla (2014), for example, measure it on the external ornamentation. In other words, H equals shell thickness and W equals node diameter. Both data are available in the main text. 30/ Line 440 and 494: I believe that, by following my recommendations and processing a little bit the oological information, author should be able to extract some additional paleobiogeographic conclusions that may help to support their hypotheses. Moving to the discussion section reserved for QSD eggs (starting from Line 494), this part must be largely improved.

GENERAL COMMENT
The paper is interesting and the materials are important. All comments are made via track changes in the attached document. I combined the supplemental info and main text into one document. More justification needs to be present in the main text; as it stands, so much is relegated to the supplemental information, the main paper doesn't stand on its own. The phylogenetic analysis unfortunately doesn't appear to have been very carefully done. I did not go through each and every character (let alone each character scoring), but looking over the list even briefly, I found a duplicate (copied and pasted) character, typos, and potential overlap among characters put forth as disparate. These issues will not give the reader confidence in the results of the phylogenetic analysis, if it is left in. I don't think a phylogenetic analysis is necessary, you could just note synapomorphies present in the material based on previous analyses. This would free up room for describing the provenance, age, and taphonomy of the specimens, all of which are severely underreported in the main text (but present in nice detail in the supplemental information). In sum, I think these are very important fossils and congratulate the authors for the huge undertaking involved in bringing them to light, but I feel that there are substantial issues that prevent publication of the manuscript in its present form.

Reply:
We tried to address all the comments from the reviewer, but several of them pointed for a different kind of manuscript, which is not the case we have here presented. Most of the palaeontological and geological information in the present manuscript, including two titanosaurian sauropod species and a nesting site, are new. This means there is a lot to be said for contextualizing the reader in a short article. Therefore, we must transfer a lot of relevant data to the Supplementary Information. We conducted a careful phylogenetic analysis, which based using one of the most recent and complete phylogenetic data sets available (published in the Journal of Systematic Palaeontology). We provide detailed explanations regarding possible mistakes and duplicate characters below. We ran the phylogenetic analysis to understand the position of both new species in the context of the derived Titanosauria of South America, not as taxonomic support. Geological and taphonomic data are also essential for any new palaeontological record, but they are not central to the present investigation. We provide more geological data within the "nesting site" section of the Results.

INTRODUCTION
Line 45_...attempted to establish paleobiogeographic links between these regions 6,7 , although there are remarkable faunistic differences… Comment 3: "Since" "because" or "finding" would seem to make more sense to me than "although".
Reply: We disagree. "Since", "because" or "finding" would change the meaning of the sentence.
Comment 4_correct me if I am wrong, but didn't Saltasaurus live after the seaway had mostly retreated?
Reply: According to the palaeogeographic maps from Scotese and Wright (2018; see Fig. 1a of the manuscript), by the Late Campanian, there were large transgressions that covered lowland territories of South America. The seaways started retreating by the Maastrichtian.
Comment 5_it would be good to put here, upfront, the geographic and stratigraphic extent of the new locality. The way the main paper is written it sounds like the two new taxa were found at the same site; instead in the supplemental information it is stated that they are separated by ~150m stratigraphically.

Reply:
The geographical (Quebrada de Santo Domingo locality in the Andes of La Rioja, NW Argentina) and stratigraphic (Upper Cretaceous red beds) locations are clarified in the "horizon and type locality" section of each taxon description (lines 91-93 and 250-252). We do not see that the main text is confusing regarding the origin (locality, quarry) of the two new taxa. The "horizon and type locality" section for each species accounts for their stratigraphic difference, just like what is stated in the Supplementary Information. We added a few words (lines 68-71) to avoid confusion: "We recovered three partial skeletons that belong to two new species of derived titanosaurian dinosaur species (Fig. 1c,d), and a new titanosaur nesting site, in different stratigraphic levels of the Ciénaga del Río Huaco Formation. Moreover, we found Comment 25_here please state that you added two characters from a previous study and created three new ones.
Reply: Sorry, we corrected the phrase and now it reflects what we had already written in the Supplementary Information.
We added five characters (three from previous studies and two new) and modified few scorings.
Comment 26_this seems like a mix of results and methods. Reply: Modified (lines 559-562). We moved the number of steps, along with other information regarding of MPTs at different steps of the analysis, to the Supplementary Information. Fig. 1_ it is confusing to use green and yellow for these silhouettes because green and yellow are used for the percentage regions in part A of the figure. So at first glance it looks like Bravasaurus is from Brazil and Punatitan is from La Rioja. I recommend using different colors.

Reply:
We have changed the colours of the preserved bones of the titanosaurian species. Now both are red.

Fig. 2_part
F only shows these vertebrae slightly larger than they are shown in part E. I would recommend cropping out the first and last vertebra shown in F and enlarging the rest.

Reply:
We modified the figure. Reply: The first version of the manuscript states: Dashed line for Bravasaurus refers to its possibly older age (see Supplementary Information). In the new version we modified the figure. Now Bravasaurus shows the same temporal range than Punatitan, as variations in age cannot be confirmed. See comment 19 from R#2.

SUPPLEMENTARY INFORMATION Comment 27_this image is completely black
Reply: Sorry, there might be a conflict with the figures included in the word text. We submit the Supplementary Information in .pdf format.
Comment 28_again, this image is completely black. with the next round of review please also upload the supplemental info as a pdf.
Reply: Same as comment 27.
Comment 29_I'd expect to see Phylogenetic methods more precisely laid out before results. What are the five new characters? What scoring changes were made, precisely? Why only parsimony?
Reply: We moved part from the method section of the main text to this section (207-213). We added a first paragraph explaining the workflow that we applied. The present paper is not focused on systematics. We introduced minimum changes into the matrix of Carballido et al. (2020), and they do not severely modify their. The authors provide an extensive discussion regarding the methods and limitations of their dataset.
Comment 32_the source for these data needs to be cited. is it from original literature? paleobiology database? elsewhere? also, is it paleolatitude or current latitude?
Reply: We incorporated more data for clarification. See comment 19 from R#2.
Comment 33_there are a very large number of changes, but no discussion or documentation of those changes. at a minimum, for each change you should cite the change's source, i.e., a publication and its figure number or pers. obs. with the author's initials who made the pers. obs.
Reply: For this paper, we made 169 modifications to the dataset provided by Carballido et al. (2020). This value represents 0.4% of the total data (422x98). Of these modifications, 80% were related to missing data that we change to character-states. The new information comes from publications (e.g. Overosaurus, Rapetosaurus) and personal observations (e.g. Trigonosaurus, Baurutitan, 'Aeolosaurus'). We made 32 changes to the original scorings of Carballido et al. (2020); 16 of which represent modifications of the characters' definitions or their states. For instance, we split one state of character 300 into two, which causes variations in the scorings that do not reflect a different interpretation of the observed morphology. Among the other 16, three are changes to ambiguous states (e.g. 0 to 0&1), one to "not-applicable" and another to missing data. We explain the remaining 11 modifications in Supplementary Table 7.