Paleoecological inferences about the Late Quaternary giant sloths

The sloths that dispersed into Central and North America were most likely derived from tropical taxa, as indicated by their high occlusal surface area (OSA) in comparison with body mass, as in South American tropical species. Relative muzzle width and hypsodonty indices were used to infer the dietary adaptation of several Late Pleistocene/Early Holocene giant ground sloths from the Americas. Nothrotheriidae and Megatheridae taxa were adapted to a browser or browsing mixedfeeder diet, Mylodontidae taxa to a variety of dietary types, and Megalonychidae taxa to a browsing mixedfeeder diet. However, these are generalized dietary adaptations and do not necessarily reflect the consumption of specific food resources. In general, Megalonychidae and Nothrotheriidae were climbing species, Mylodontidae were digging species, and Megatheriidae were fully terrestrial species.

While many studies on the paleoecology of these taxa have been conducted; the paleoecological characteristics of several taxa (mainly recently described species, many of which are only known from the holotype) still remain unknown, limiting our understanding of niche partitioning or potential ecological competition of taxa with overlapping

Body mass estimation
In this study, we estimated the body mass of eleven extinct giant sloths from the Americas using circumference measurements from the humerus and femur, following the regression for xenarthrans presented by Dantas (2022; Table 1 and Online Resource 1: Table S2).These measurements were acquired from collections or estimated through figures in publications (Hoffstetter 1952;Mathew and Paula Couto 1959;MacPhee et al. 2000;De Iuliis et al. 2009;McDonald and Lundelius 2009;Stinnesbeck et al. 2021;McAfee et al. 2021; Online Resource 1: Table S2).
For Diabolotherium nordenskioldi, and Glossotherium tropicorum only a single humerus from a single adult individual is known (Pujos et al. 2007;De Iuliis et al. 2017).Therefore, the circumference (C) of the femur was estimated using the humerus: femur proportion (1:1.35 for D. nordenskioldi; 1:1.40 for G. tropicorum).The relationship between the circumference of the femur and humerus was extrapolated from Ahytherium aureum, Australonyx aquae, Glossotherium wegneri, and Paramylodon harlani for which both bones are available (Online Resource 1: Table S2).
These growth stage of the taxa was inferred based on the stage of epiphysis-diaphysis fusion in the humerus and femur as juveniles (clear separation between the epiphysis and diaphysis), subadults (a visible scar between the epiphysis and diaphysis), or adults (epiphysis-diaphysis completely fused).

Occlusal surface area
The cheek teeth occlusal surface area (OSA; the total surface used for mastication considering all cheek teeth) is larger in grazing ungulates than in browsers because of physiological adaptation to the nutritional value of consumed food, which led these taxa to adopt different dietary strategies (Janis 1995).Grazers consume lower-quality and highly abrasive grasses and process large amounts of food orally, leading to lower fermentation requirements.In contrast, browsers consume succulent and nutritionally high-quality leaves and fruits (Vizcaino et al. 2006, and references therein).The same pattern has been observed in both the extant arboreal sloths (Reis et al. 2011) and extinct giant sloths (Dantas and Santos 2022, and references therein).Therefore, OSA values relative to body mass provides a mean to determine if an extinct giant sloth taxon had a physiological adaptation similar to browsing or grazing ungulates or extant Xenarthra (Vizcaíno et al. 2006).
The occlusal surface area (OSA in mm 2 ) of the upper dentition of each extinct giant sloth was estimated only for species for which body mass could also be estimated (i.e., taxa for which both the humerus and femur are known; Online Resource 1: Table S2) adapted from the technique proposed by Vizcaíno et al. (2006).OSA was calculated using ImageJ software (Abràmoff et al. 2004), utilizing freehand selection to determine the exact area of each molariform.Only the skulls of adult individuals were measured (except for Xibalbaonyx oviceps, which is known only from a single immature skull).When the molariform was absent, measurements were made from either the tooth alveolus, or on one side when a single tooth was present and multiplied by two to estimate the total OSA of the studied individual.All OSA measurements are presented in Online Resource 1: Table S1.
The OSA values were plotted against body mass (kg) of each sloth taxon (Table 1 and Online Resource 1: Table S2; Dantas 2022).The two variables (OSA and body mass) of each extinct giant sloth examined were log-transformed and compared with the OSA data presented by Vizcaíno et al. (2006) and Dantas and Santos (2022).

Relative muzzle width
To infer the diet of the Late Quaternary giant ground sloths, we also used the muzzle width technique for giant ground sloths proposed by Bargo et al. (2006b).For all sloths, the maximum muzzle width (MMW) was measured on the maxilla at the premaxillo-maxillary suture.The palatal width (PW) was based on the mean width between the first molariform (M1) and last molariform (M3 in Megalonychidae and Nothrotheriidae, M4 in Mylodontidae and Megatheriidae; Fig. 1).
The relative muzzle width (RMW) index is the ratio of MMW/PW.We suggest modifying the interpretation of Bargo et al. (2006b) using the following intervals: animals with an RMW index from 0 to 0.30 and a wide muzzle are interpreted as being adapted to feed primarily as a grazer;    T, terrestrial when this index is between 0.31 and 0.70, and with an intermediate muzzle, taxa with values between 0.31 and 0.50 are considered grazing mixedfeeders, while those with an RMW index between 0.51 and 0.70 are considered browsing mixedfeeders browsers.Finally, taxa with RWM values between 0.71 and 1.00 and with a narrow muzzle are considered obligatory browsers (Table 1 and Online Resource  1: Table S1).

Hypsodonty index
The degree of hypsodonty for several extinct giant sloth taxa in the Americas (Table 1 and Online Resource 1: Table S2) was estimated following the criteria of Bargo et al. (2006a): depth of the mandible (DM) measured at the level of the third molariform tooth divided by the length of the molariform tooth row (LTR).All measurements were taken using ImageJ software (Abràmoff et al. 2004).
It should be noted that the lower tooth rows of megalonychids and nothrotheres include only three molariforms, while in megatheres and mylodonts, there are four molariforms.Calculations of the relationship between mandible depth and the lower tooth row length do not significantly alter the hypsodonty index (Fig. 2).

Index of dorsal olecranon projection (IDO)
The index of dorsal olecranon projection (IDO) is calculated by dividing the length of the dorsal extent of the olecranon process by the length of the trochlear notch from the most proximal, anterior point to the junction of the trochlear and radial notches (Fig. 3a; Table 1 and Online Resource 1: Table S1).
Here, we use the proposition of Santos et al. (2023), which uses the body mass (BM; transformed to logarithm values at base 10) associated with IDO.This could separate  Those taxa with a log-transformed body mass value higher than 2.40 (BM > 250 kg) should be considered terrestrial.

Index of fossorial ability (IFA)
The index of fossorial ability (IFA), developed by Vizcaíno et al. (1999), is calculated by dividing the length of the olecranon (Ol) by the difference between the total length of the animals into suspensory, climbing, and terrestrial modes of locomotion.
Based on data from primates and xenarthras, animals with suspensory or climbing habits present IDO values varying between 0.52 and 1.24, and separated as suspensory based  which is slightly lower than the 463 kg estimate of McDonald (2005) based on femur length.
For the Mylodontidae, the body mass of the South American Glossotherium tropicorum (Ecuador) was 610 kg and Glossotherium wegneri (Ecuador) was 439 kg.For the North American Paramylodon harlani, BM was 798 kg, which is less than the estimated 1,392 kg reported by McDonald (2005).For Megatheriidae taxa, the body mass of the South American Megatherium tarijense (Bolivia) was 947 kg, and for an individual of Eremotherium laurillardi from North America, the body mass was 1,763 kg (Table 1 and S2).

Relative muzzle width
The RMW values (Online Resource 1: Table S1 S1; Fig. 3b).In this way, it is possible to calculate the movement capacity and strength of the triceps muscle, which is mainly responsible for the extensor movement of the forearm and is widely used by mammals that dig with the movement of the entire arm.Burrowing animals generally have a higher index than non-digging animals (Hildebrand 1985).
Here, we used the adaptation of Bargo et al. (2000) made by Santos et al. (2023).To estimate the percentage of IFA (%IFA), the IFA of each giant sloth is divided by the value of G. robustum (IFA/0.56*100),which has the highest value of the Late Pleistocene giant sloths from South America.Thus, values lower than 50% indicate non-digging mammals, from 51 to 70% diggers, values from 71 to 95% are highly specialized diggers, and values from 96 to 100% indicate burrowers.

Analytical methods
We used Student's t-test (p = 0.05) and ANOVA (p = 0.05) to compare similarities among the RMW, OSA, and HI values of the studied taxa.Tukey's test was used to identify species that presented similar values.These analyses were performed using PAST 3.11 (Hammer et al. 2001).

Body mass
The estimated body mass of continental Megalonychidae from the Americas was 67 kg for Diabolotherium nordenskioldi (Peru) and 189 kg for a juvenile individual of Xibalbaonyx oviceps (Mexico), which is the only known specimen of this taxon to date.The body mass of X. oviceps calculated in this study approximates the 200 kg estimate of Stinnesbeck et al. (2021) based on femur length.For Megalonyx jeffersonii (USA), based on the bones of several adults from three localities, we suggest a body mass of 527-668 kg (mean value 588 ± 71 kg; Table 1 and Online Resource 1: Table S2), which is lower than the estimate of 1,090 kg of McDonald (2005) based on femur length.

Metabolism inferences
The pioneering work of Vizcaíno et al. (2006) was based on the analysis of five giant sloth species representing two families from the Pampean Region (Argentina and Uruguay), four mylodontids (Glossotherium robustum, Lestodon armatus, Mylodon darwini, and Scelidotherium leptocephalum) and one megatheriid (Megatherium americanum).Mylodontids presented lower OSA values, whereas Me. americanum had high OSA values.Vizcaíno et al. (2006) suggest that the lower values of the mylodontids could indicate a higher level of fermentation in the digestive system and a possible lower metabolic requirement, while Me.americanum could present similar metabolic requirement to epitherian taxa of the same size.

Hypsodonty index
The hypsodonty index (HI) of the extinct giant sloths was not statistically different among the four families (ANOVA, F = 28.96,p < 0.05), varying from 0.48 to 1.32, with low values indicative of a relatively shallow jaw relative to the tooth row and high values indicative of greater jaw depth.These data could be compared between the giant sloth families due to the isometric increase in the length of the tooth row associated with the depth of the dentary (R 2 = 0.81; Fig. 2).The lowest HI values were observed in members of the Mylodontidae (0.62 ± 0.06), followed by the Megatheriidae (0.75 ± 0.08); Megalonychidae (0.91 ± 0.22) and Nothrotheriidae (1.02 ± 0.18) had the highest values.
In terms of time to adapt to differences in the North American vegetation, Megalonyx would have had the greatest amount of time, as it first appears in the Late Miocene (Hemphillian), Paramylodon and Eremotherium arrived later in Blancan, and Nothrotheriops was the last to reach North America and had the least amount of time to adapt; however, all of these taxa are well represented in faunas at temperate latitudes in North America.Because Nohochichak xibalbahkah and Xibalbaonyx oviceps were restricted to in tropical habitats in North America, they do not appear to have adapted to different types of vegetation in their diet that differed significantly from those utilized by their ancestral taxa living in the tropics of South America.

Diet inferences
The relative muzzle width and hypsodonty indices provide a means to suggest the different feeding adaptations of extinct giant sloths within the environment in which they lived (since the Neogene; Bargo et al. 2006a, b), allowing an interpretation of their dietary and habitat preferences.These indices may allow for the interpretation of niche partitioning when multiple sloths are associated in a fauna, as well as how they may have avoided competition with other sympatric megaherbivores, both of South American origin (e.g., by McDonald et al. (2013) for Megistonyx oreobios.The exceptions were D. nordenskioldi and Mx.jeffersonii, which fell on the Xenarthra regression, probably because of larger body mass relative to OSA, which could have a direct effect, causing a high energy cost, probably lowering metabolism.
The OSA × body mass of Nothrotheriops shastensis suggests that members of the Nothotheriidae probably had low oral food processing and metabolic requirements, together with D. nordenskioldi, Megalonyx jeffersonii, Ocnotherium giganteum, and the mylodontid species from the Pampean Region (Fig. 4).This hypothesis for Nh.shastensis is supported by the size of plant fragments in its dung, which are large and not fully chewed.A similar pattern was seen in plant fragments inferred to be from the dung of E. laurillardi (Lindsey et al. 2020) and may be related to long transit time in the gut permitting the animal to absorb a proportion of nutrients in their food.Dantas and Santos (2022) proposed that giant sloths from temperate regions would have adapted to a lower oral food processing capability, which was associated with a higher degree of fermentation.However, data from the North American temperate zones allowed us to reject this hypothesis (Table 1; Fig. 4).This could reflect the evolution of North American taxa from tropical forms (McDonald 2006;McDonald et al. 2017;Varela et al. 2019), and feeding on new types of vegetation that necessitated different Unfortunately, only a limited number of other techniques can provide inferences on the diet of extinct species.However, dental microwear analyses have indicated that Acratocnus (Ac.odontrigonus) could have had a mixedfeeder "browser" diet, while Megalonyx (Mx.wheatleyi) was a grazer (Resar et al. 2013).
Both the North American and South American nothrotheres, Nothrotheriops shastensis and Nothrotherium maquinense, had RMW values indicative of adaptation as browsers (RMW = 0.66-0.74;Fig. 5).The HI values suggest an adaptation to a less abrasive diet for N. maquinense toxodonts and litopterns) or from North America, such as equids, camelids, and gomphotheres.
The hypsodonty index values (Table 1) did not differ between the extinct giant sloth families (ANOVA, F = 28.96,p < 0.05), suggesting adaptation to a similar degree of tooth wear between all taxa.As all sloths lack enamel on their teeth, and the teeth are composed of similar types of dentine (Kalthoff 2011).They are also hypselodont (ever-growing), and the rate of growth to compensate for tooth wear during mastication may be essentially the same in all sloths.Therefore, the height of the teeth, as indicated by depth of the mandible, may not be a critical factor.
When all available δ 13 C collagen data for giant sloths from the Americas were interpreted together, a moderate correlation between δ 13 C collagen and body mass was observed (n = 7; R 2 = 0.48; Fig. 6b), similar to that observed by Dantas (2022).However, δ 13 C bioapatite showed no correlation with body mass (n = 10; R 2 = 0.04; Fig. 6c), with results opposite to those reported by Dantas (2022).This occurs because of the discordant carbon isotopic data acquired through bioapatite found for Mx.jeffersonii (δ 13 C = 1.7 ± 0.04‰; Table 1) and P. harlani (δ 13 C = 3.3 ± 4.2‰; Table 1), which differ strongly from the carbon isotopic data found in collagen.Without these values, the correlation coefficient increased to 0.62.

Niche differentiation
The niche differentiation of the Late Quaternary giant sloths has been studied, and currently we have a view that most members of Megalonychidae and Nothrotheriidae, particularly the smaller taxa, could be mainly climbers, the Mylodontidae could be mainly diggers, while the Megatheriidae were fully terrestrial animals (Bargo et al. 2000;Vizcaíno et al. 2006;Santos et al. 2023).Our results suggest that within the Megalonychidae, Neocnus dousman and Acratocnus ye could be suspensory, while Parocnus domenicanus, Mesocnus rodens, Xibalbaonyx oviceps, and Diabolotherium nordenskioldi were climbers (Fig. 7), with Pa.domenicanus probably also being a digger (Table 1).The results for D. nordenskioldi are in accordance with the interpretation made by Pujos et al. (2007).Unfortunately, Megistonyx oreobios and Nohochichak xibalbahkah do not have preserved ulnae, which prevents us from proposing their ecological niche based on this bone.
As in the case of Mx. jeffersonii, the larger body mass of Nothrotheriops shastensis probably prevented it from being considered to be a climber unlike its close relative, N. maquinense (Fig. 7; Santos et al. 2023), and it was also not a digger (Table 1).It is possible that immature juveniles of Megalonyx could climb until they reached a critical size, as was proposed by Grass (2019) based on the scapula, but the paucity of juvenile skeletons at different growth stages (HI = 0.77; Table 1), but Nh. shastensis had a more abrasive diet (HI = 1.09;Table 1), confirmed by direct evidence of the diversity of plants in preserved coprolites (Hansen 1978;Thompson et al. 1980).
The species of Mylodontidae examined had the widest RMW in comparison with the other families (RMW = 0.28-0.58;Table 1; Fig. 5), indicative of a mixedfeeder diet, in agreement with the results presented by Naples (1989).The only exception was L. armatus, which was adapted to a grazing diet (RMW = 0.28; Table 1).The HI values reinforce this interpretation, suggesting an adaptation to less abrasive food items (HI = 0.48-0.91;Bargo et al. 2006a; Table 1).
Currently, the only taxa in the family for which carbon isotopic data acquired from collagen and bioapatite is available are Glossotherium robustum, Catonyx cuvieri (from Brazil), Valgipes bucklandi (Brazil), Paramylodon harlani (United States), and Lestodon armatus (Uruguay).The data for these taxa indicate a mixedfeeder "browser" diet for G. robustum and C. cuvieri and a browser diet (p i C 3 > 80%) for P. harlani, V. bucklandi, and L. armatus (Table 1).As occurred in Mx. jeffersonii, the carbon isotopic data acquired through bioapatite for Paramylodon harlani indicate a diet that is opposite of that indicated by the collagen analysis data (Ruez 2005;Pérez-Crespo et al. 2014;DeSantis et al. 2019; Table 1).

Final remarks
The giant sloths that evolved in more temperate southern South America (Varela et al. 2019) probably had higher levels of fermentation/lower metabolic requirements, as shown in the Late Pleistocene taxa found in the Pampean Region (Vizcaíno et al. 2006).As these taxa migrated to tropical regions, they likely evolved higher metabolic requirements, as temperature is one of the main factors for high nucleotide substitution and mass metabolic rates (Gillooly et al. 2005).Body mass is also strongly indicative of higher levels of prevents a resolution of this question at this time based on other parts of the skeleton.
Within the Mylodontidae, Mylodon darwini was probably a digger, while Paramylodon harlani and Glossotherium wegneri, together with O. giganteum, G. phoenesis, L. armatus, and C. cuvieri, were specialized diggers.The Megatheriidae E. laurillardi from USA does not differ from its South American relatives, being a fully terrestrial species, as expected (Fig. 7; Table 1).fermentation/lower metabolic requirements, and we note that the tropical forms of South American giant sloths generally have a higher body mass than those from temperate regions (e.g., Dantas 2022).These tropical taxa subsequently migrated to Central and North America (Varela et al. 2019), which could explain the pattern observed in this study.
There is a paradox here: the more southern giant sloths in South America would have been at higher latitudes and in more temperate zones, such as the Pampas in Buenos Aires Province at 34-38 °S, which is roughly the same latitude as the southern third of the United States.Therefore, sloths in these two regions lived in climatically comparable habitats, which Webb (1991) argued for in his savanna taxa interchange papers and has been reinforced by paleoclimatic models (Brown et al. 2018).We point out that tropical adapted sloths in southern Mexico and Central America did not move into temperate habitats, indicating the range of ecological requirements of these sloth species and other South American mammalian taxa that dispersed into North America.
The relative muzzle width and hypsodonty indices suggest dietary adaptation of several Late Pleistocene/Early Holocene giant sloths in the Americas.Nothrotheriidae and Megatheridae taxa were adapted to a browser or mixedfeeder "browser" diet, Mylodontidae taxa to a variety of diet types, and Megalonychidae taxa to a mixedfeeder to browser diet.The occlusal surface area index together with the body mass allows a better understanding of the paleoecology and evolution of the giant sloths in the Americas.
These dietary adaptations did not necessarily reflect their food niche; the RMW diets for some Mylodontidae taxa were different from what RMW and HI suggested.Similarities among RMW, HI, and isotope/microwear analysis were observed for Megalonychidae, Nothrotheriidae, and Megateridae taxa.
Finally, two indices based on ulna measurements allowed us to differentiate another component of the ecological niche of giant sloths.A probable pattern is that Megalonychidae and Nothrotheriidae could include mainly climber species, while Mylodontidae taxa are mainly diggers, and Megatheriidae, due to their higher body mass, were fully terrestrial species. 157 log-transformed body mass varying from 0.68 to 1.30 (BM = 5-20 kg), and climbing animals with log-transformed body mass varying from 1.70 to 2.40 (BM = 50-250 kg).

Fig. 3
Fig. 3 Diagram of ulna of the extinct giant sloth Nothrotherium maquinense illustrating measurements taken.a index of dorsal extent of the ulna (IDO) was calculated as the length of the dorsal extent of the olecranon process (Deop) by the length of the trochlear notch (Tn); b index of fossorial ability (IFA) was calculated as the length of the olecranon (Ol) divided by the difference between the total length of the ulna (Ul) and the olecranon

Fig. 4
Fig. 4 Regression of log of occlusal surface area (OSA) against log of body mass of the giant sloths of the Late Quaternary in the Americas.The letters E and X on the right indicate the Epitheria and Xenarthra regression lines, respectively, presented by Vizcaíno et al. (2006).Dashed lines above and below the regression lines show the 95% confidence intervals.Light color symbols and lines are from giant sloths from the Brazilian Intertropical Region presented by Dantas and

Table 1
Data for Late Quaternary giant sloths from the Americas.Data are from this study unless indicated otherwise.Column abbreviations: col , carbon isotopes from collagen (‰); p i C 3 , proportion of C 3 plants (%); p