Isotopic paleoecology of the Pleistocene megamammals from the Brazilian Intertropical Region: Feeding ecology (δ13C), niche breadth and overlap

Quaternary Science Reviews 170 (2017) 152e163

Contents lists available at ScienceDirect

Quaternary Science Reviews journal homepage: www.elsevier.com/locate/quascirev

Isotopic paleoecology of the Pleistocene megamammals from the Brazilian Intertropical Region: Feeding ecology (d13C), niche breadth and overlap
rio Andre
Trindade Dantas a, *, Alexander Cherkinsky b, Herve
Bocherens c, Ma d e f Morgana Drefahl , Camila Bernardes , Lucas de Melo França
rio de Ecologia e Geoci^ Laborato encias, Instituto Multidisciplinar Em Saúde, Universidade Federal da Bahia e Campus Anísio Teixeira, Rua Rio de Contas,
ria da Conquista, BA, Brazil 58, Candeias, cep. 45029-094, Vito Center for Applied Isotope Studies, University of Georgia, Athens, GA 30602, USA c €t Tübingen, Ho €lderlinstr. Biogeology, Department of Geosciences and Senckenberg Center for Human Evolution and Palaeoenvironment (HEP), Universita 12, 72074 Tübingen, Germany d Grupo de Estudos de Paleovertebrados, Universidade Federal da Bahia, Salvador, BA, Brazil e Universidade Federal Fluminense, Rio de Janeiro, RJ, Brazil f ~o Em Ecologia e Conservaça ~o, Universidade Federal de Sergipe, Sa ~o Cristova ~o, SE, Brazil
s-graduaça Programa de Po a

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 19 January 2017 Received in revised form 7 June 2017 Accepted 28 June 2017

The extinct megamammals Eremotherium laurillardi (weight 6550 kg), Notiomastodon platensis (w ¼ 6000 kg), Toxodon platensis (w ¼ 3090 kg), Valgipes bucklandi (w ¼ 980 kg) and Equus (Amerhippus) neogaeus (w ¼ 370 kg) are recorded for the late Pleistocene of the Brazilian Intertropical Region. In order to evaluate the isotopic paleoecology (feeding diet, niche breadth and overlap) of these species, 14C dates, d13C and d18O analyzes were performed. Our results suggest that E. laurillardi (md13C ¼ 4.35 ± 2.87‰; mBA ¼ 0.77 ± 0.25), T. platensis (md13C ¼ 5.74 ± 4.80‰; mBA ¼ 0.57 ± 0.40) and N. platensis (md13C ¼ 1.17 ± 2.76‰; mBA ¼ 0.56 ± 0.20) were mixed feeders with a wide niche breadth, while E. (A.) neogaeus (md13C ¼ 0.73 ± 1.19‰; mBA ¼ 0.38 ± 0.22) was a grazer, and V. bucklandi (d13C ¼ 10.17‰; BA ¼ 0.13) was a specialist browser. A narrow niche overlap occurred between V. bucklandi and the species that fed principally on C4 plants (>70%; O ¼ 0.24e0.43). In contrast, there was a high niche overlap between E. neogaeus and N. platensis (O ¼ 0.75) and between E. laurillardi and T. platensis (O ¼ 0.86). Therefore, E. laurillardi was probably a key species in this Pleistocene community due to its high body weight and wide niche breadth, suggesting that E. laurillardi was a great competitor for resources in the BIR. © 2017 Elsevier Ltd. All rights reserved.

1. Introduction Compared to other continents, our knowledge regarding the paleoecology of the South American Pleistocene megafauna is still scarce. It was not until the past two decades that there was a substantial increase of published studies regarding the paleodiet of the South American Quaternary megafauna (MacFadden et al.,
nchez et al., 2004; MacFadden, 2005; Dantas et al., 1999; Sa 2013a,b; Lopes et al., 2013; França et al., 2014; Bocherens et al., 2016, 2017).

* Corresponding author. E-mail address: [email protected] (M.A.T. Dantas). http://dx.doi.org/10.1016/j.quascirev.2017.06.030 0277-3791/© 2017 Elsevier Ltd. All rights reserved.

The majority of South American Pleistocene megafauna species are extinct and have no modern close relatives or analogs, which makes it difficult to reconstruct their ecology, such as niche breadth, habitat and diet. There are few or no information based on isotopic analyses about their ecological roles, and so about their sensitivity to environmental changes, a key issue when trying to understand their extinction. Thus, an effort to yield better information about the paleoecology of the Pleistocene megafauna is needed. Lately, the Brazilian Intertropical Region (BIR hereafter, Fig. 1) has been an area of paleontological interest in South America. This zoogeographic region was defined by Cartelle (1999) based on the occurrence of endemic mammal species in the Brazilian states of

M.A.T. Dantas et al. / Quaternary Science Reviews 170 (2017) 152e163

153

Fig. 1. Brazilian Intertropical Region (BIR) map. Studied localities: (1) Felipe Guerra/RN, (2) Currais Novos/RN, (3) Rui Barbosa/RN, (4) Barcelona/RN, (5) Maravilha/AL, (6) Poço
/BA, (10) Quinjingue/BA, (11) Pedra Vermelha/BA, (12) Ourol^andia/BA and (13) Vito
ria da Conquista/BA. States that Redondo/SE, (7) Gararu/SE, (8) Canhoba/SE, (9) Coronel Jo~ao Sa
ias; MG - Minas Gerais; RJ - Rio de Janeiro; ES - Espírito Santo; BA - Bahia; SE - Sergipe; AL - Alagoas; PE - Pernambuco; PB - Paraíba; RN - Rio Grande do comprise the BIR: GO - Go Norte; CE - Cear
a; PI - Piauí (sensu Cartelle, 1999).

Goi
as (GO), Minas Gerais (MG), Rio de Janeiro (RJ), Espírito Santo (ES), Bahia (BA), Sergipe (SE), Alagoas (AL), Pernambuco (PE), Rio
(CE), and Piauí (PI). In the Grande do Norte (RN), Paraíba (PB), Ceara BIR, the records of fossil mammals are found mainly in caves (e.g. Hubbe and Auler, 2012) or in “tanks” (e.g. Araujo Jr. et al., 2013), which are natural depressions in Precambrian crystalline rocks filled with Quaternary sediments (Ximenes, 2008). In both types of deposit fossils of medium size (weight between 10 kg and 100 kg), large size (weight between 100 kg and 1000 kg), and megamammals (weight > 1000 kg) can be found, while fossils of smallsized mammals (weight < 10 kg) are mainly found in caves. These differences are explained by the particular conditions that allow fossilization in each depositional environment (Araujo Jr. et al., 2013). The current knowledge about the BIR Pleistocene megafauna is mainly based on the studies conducted by Cartelle (1999), and subsequent studies have been mostly relying on them. Most ecomorphological data are not yet published (only as gray literature). However, some paleoecological informations of BIR megamammals based on analyses of stable isotopes have been published (Dantas et al., 2013a,b; França et al., 2014). Hitherto, the published d13C values for BIR mammals encompass five taxa: the giant ground sloths Eremotherium laurillardi (Lund, 1842) and Valgipes bucklandi (Lund, 1839); the gomphothere Notiomastodon platensis (Ameghino, 1888); the “hippo-like” species Toxodon platensis Owen, 1840; and the horse Equus (Amerhippus)
nchez et al., 2004; neogeus Lund, 1840 (MacFadden et al., 1999; Sa MacFadden, 2005; Dantas et al., 2013a,b; Pereira et al., 2013; França et al., 2014). An interesting pattern for this region was observed by Dantas et al. (2013a), where the species E. laurillardi, N. platensis, and T. platensis presented a generalist feeding habit according to carbon isotopic values, and therefore could have lived in a variety of habitats during the late Pleistocene. Together with the 14C dates and d18O values, the d13C values of these BIR species

will allow us to make a better interpretation of their ecological niche and what role did these species play in the community structure of this region during the Late Pleistocene. Thus, the main objectives of this paper are: (i) to further elucidate the paleoecology, especially the feeding diet, niche breadth and niche overlap, of the species E. laurillardi, N. platensis, T. platensis, V. bucklandi and E. (A.) neogaeus, based on d13C and d18O analyzes associated to 14C dates; (ii) to compare the paleoecology of the studied species and the paleoenvironmental conditions in which they lived with one of the world's last remaining savanna megafauna, in Amboseli, Kenya (Bocherens et al., 1996); and (iii) to investigate how key species contributed to the structure of the BIR community. 2. Materials and methods 2.1. Description of studied fossil material Eighteen skeletal samples of adult individuals of Eremotherium laurilardi (Lund, 1842), Notiomastodon platensis (Ameghino, 1888), Toxodon platensis (Owen, 1840), and Valgipes bucklandi (Lund, 1839) from Bahia (BA), Sergipe (SE) and Rio Grande do Norte (RN) e which were previously analyzed for 13C isotopic ratios (Dantas et al., 2013a; Pereira et al., 2013) e were reanalyzed to obtain the carbon and oxygen isotopic composition from the structural carbonate of their bone, dentin and enamel, associated to 14C AMS datings (Fig. 1; Table S1). The isotopic composition of hydroxyapatite can be preserved with minimal or no significant diagenetic
nchez et al., 2004). alteration (Bocherens et al., 1996; Sa Samples were collected from fossil specimens housed at the
rio de Geologia/Universidade following institutions: Laborato
rio de Estadual do Sudoeste da Bahia (Bahia, Brazil); Laborato Paleontologia/Universidade Federal de Sergipe (Sergipe, Brazil); Memorial de Sergipe/Universidade Tiradentes (Sergipe, Brazil); and

154

M.A.T. Dantas et al. / Quaternary Science Reviews 170 (2017) 152e163

^mara Cascudo/Universidade Federal do Rio Grande do Museu Ca Norte (Rio Grande do Norte, Brazil), and all permissions for study these materials were given. 2.2. Choice of tissues The hydroxyapatite of N. platensis and T. platensis was extracted from enamel samples, the only exception was the T. platensis sample UGAMS 09444, which was extracted from dentine. The hydroxyapatite of E. laurillardi was extracted mostly from outer dentine (six samples), the only exception was the sample UGAMS 06136 that was sampled from bone. The one sample from V. bucklandi was analyzed in bone as well. The stable isotope analyzes were performed at the Center for Applied Isotopes Studies at the University of Georgia (United States of America). We used the isotopic values found in enamel and dentin to reconstruct the paleodiet of these species. Using Rare Earth Elements analysis (REE), MacFadden et al. (2010) has confirmed that enamel is less susceptible to diagenesis and that dentine falls within acceptable limits (REE index values < 0.35). This confirmation allows us to state that this tissue may not have suffered from diagenesis as well. Although we are aware that bone is more susceptible to diagenesis, we used the isotopic values found in this tissue for the sample UGAMS 06136 of E. laurillardi because this value is very similar to the other found in dentine (Table S1), as well as for V. bucklandi because this is the only sample available at the moment. There are different options regarding reliability of d13C values in Xenarthran dentine, and bone, and dentine in general. Although the early publications (Nelson et al., 1986) have suggested that hydroxyapatite was exposed to diagenetic transformations, most recent studies (Harrison and Katzenberg, 2003; Lee-Thorp and
rez-Crespo et al., 2012; Dantas et al., 2013a; Sponheimer, 2003; Pe França et al., 2014) have shown that dentine and bone may be reliable for paleodietary reconstructions. However, preservation of these materials depends on factors specific to each site, and remaining collagen may protect the mineral fraction from extensive diagenetic alteration (e.g., Person et al., 1996; Sillen and Parkington, 1996; Tütken et al., 2008; Bocherens et al., 2017). 2.3. Additional published datae In order to complement our results, we included (Table S1) previously published isotopic data (d13C and d18O, of which most have 14C AMS datings) of E. laurillardi (Viana et al., 2011; Dantas et al., 2014; França et al., 2014), N. platensis (S
anchez et al., 2004; Viana et al., 2011; França et al., 2014), T. platensis (MacFadden, 2005; França et al., 2014), and Equus (Amerhippus) neogaeus (MacFadden et al., 1999). Furthermore, we also compared our results with the data of the following extant African megamammals from Amboseli, Kenya (Bocherens et al., 1996): Loxodonta africana (Blumenbach, 1797) (African elephant, mixed feeding diet), Hippopotamus amphibius Linnaeus, 1758 (hippopotamus, aquatic diet), Equus quagga Boddaert, 1785 (¼ E. burchelli; zebra, grazing diet, consumption of more than 80% of C4 plants), Connochaetes taurinus (Burchell, 1823) (blue wildebeest, grazing diet), Syncerus caffer Sparrman, 1779 (African buffalo, grazing diet), and Dicerus bicornis (Linnaeus, 1758) (black rhinoceros, browsing diet, consumption of more than 75% of C3 plants) (Figs. 2e3). The carbon isotopic values from the extant mammals from Amboseli were corrected by adding þ2‰ to take into account the Suess Effect (Keeling, 1979), allowing us to compare their carbon

isotopic data with those from the extinct mammals from BIR. The oxygen isotopic values from Bocherens et al. (1996) were presented in V-SMOW values, therefore, we converted these values to V-PDB values by using the equation proposed by Friedman and O'Neil (1977): dVSMOW ¼ 1.03086 * dVPDB þ 30.86, in order to compare these values with our data. 2.4. Interpretation of d13C and d18O data Taking in account the Suess Effect (þ2‰), the interpretation of the d13C results was based on the known average values of d13C around 25 ± 3‰ for C3 plants; the C4 plants have higher d13C values, averaging 11 ± 2‰; while CAM plants present intermediate d13C values (MacFadden et al., 1999; MacFadden, 2005; Domingo et al., 2012). Dietary inferences were based on the fact that modern medium-to large-bodied herbivorous mammals record the d13C values of the ingested vegetation with an enrichment of d13C values on the hydroxyapatite. In medium-to large-sized animals d13C can be enriched by 12‰e14‰ (Cerling and Harris, 1999). We used a value of 14‰, therefore, d13C values lower than 11‰ are typical of animals with a diet consisting exclusively of C3 plants, while d13C values higher than þ3‰ are consistent with a diet based on C4 plants. Intermediate d13C values (between 11‰ and þ3‰) indicate a mixed diet of C3 and C4 plants or a diet based on CAM plants (MacFadden et al., 1999; MacFadden, 2005). Due to the large range of carbon isotopic values of our specimens (Table S1), we also calculated the 25th-75th Interquartile Ranges (IQR) of their diets in order to minimize the effects caused by different time ranges and lack of diet information of the analyzed species. The IQR of the Amboseli megamammals were also calculated for purposes of comparison. The d18O results were interpreted in the following way: grazers (consumers of C4 plants) typically show higher d18O values than browsers and mixed-feeders because C4 plants are more 18Oenriched than C3 plants. Thus, grazers have higher d18O and d13C values than browsers (Kohn et al., 1996; Helliker and Ehleringer, 2000). Both d13C and d18O analyzes were performed on hydroxyapatite, since fossils from tropical regions such as the BIR rarely preserve collagen, a protein easily lost through natural diagenetic processes (Holmes et al., 2005, 2006). All samples were cleaned by ultrasonic bath with distilled water and then left to dry naturally. The samples were then crushed into smaller fragments to be treated with diluted 1 N acetic acid to remove surface absorbed and secondary carbonates. Periodic evacuation ensured that evolved carbon dioxide was removed from the interior of the sample fragments, and that fresh acid was allowed to reach even the interior microsurfaces. The chemically cleaned samples were then reacted under vacuum with 100% phosphoric acid to dissolve the bone/dentine/ enamel mineral and release carbon dioxide from hydroxyapatite. The resulting carbon dioxide was cryogenically purified from other reaction products and catalytically converted to graphite (Cherkinsky et al., 2013). Graphite 14C/13C ratios were measured using a CAIS 0.5 MeV accelerator mass spectrometer. The sample ratios were compared to the ratio measured from the Oxalic Acid I (NBS SRM 4990). The 13 12 C/ C ratios were measured separately using a stable isotope ratio mass spectrometer with respect to PDB. All results are reported using the delta notation, d ¼ [(Rsample/ Rstandard  1)*1000] (Coplen, 1994). The reference for carbon isotope values (R ¼ 13C/12C), and oxygen isotope values (R ¼ 18O/16O), is V-PDB.

M.A.T. Dantas et al. / Quaternary Science Reviews 170 (2017) 152e163

155

Fig. 2. d18O and d13C values of the extant megamammals from East Africa (Bocherens et al., 1996).

Fig. 3. d18O and d13C values of the megamammal species through the Brazilian Intertropical Region. Labels. Inferences of the limits of a (B) browser, (G) grazer and (H) hippopotamus based in extant megamammals from Amboseli National Park - Kenya (Bocherens et al., 1996).

2.5. Radiocarbon dating The quoted uncalibrated dates are given in radiocarbon years before 1950 (years BP), using the 14C half-life of 5.568 years. The error is quoted as one standard deviation and reflects both statistical and experimental errors. The date has been corrected for

isotope fractionation. The reliability of the applied technique for purification of hydroxyapatite was previously demonstrated by a comparison of 14C dating of both collagen and hydroxyapatite fractions, which yielded comparable results, and, therefore suggesting none or insignificant isotopic exchange between the biogenic hydroxyapatite and the

156

M.A.T. Dantas et al. / Quaternary Science Reviews 170 (2017) 152e163

depositional environment (S
anchez et al., 2004; Cherkinsky, 2009). The radiocarbon data were calibrated into calendar ages before present using CALIB 7.1 program (Reimer et al., 2013) and reported in the parentless after measured ages in the supplementary Table 1.

2.6. Analytical methods The analysis of variance (ANOVA; 1 factor, a ¼ 0.05) and nonparametric KruskaleWallis tests were used to examine if the differences between the isotopic values (d13C, d18O) were statistically significant for the species E. laurillardi, N. platensis, T. platensis, and E. (A.) neogaeus (Table S1). The Tukey test was used to identify which species presented similar isotopic signature (d13C, d18O). A correlation test between the d13C and d18O values was made as well. All statistical analyzes were performed with the PAST 3.11 software. The use of stable isotopes to estimate the niche breadth and overlap is a relatively new approach, which consists in transforming the d-values in p-values, thus assessing the proportions of dietary sources consumed by the animal (Bearhop et al., 2004; Newsome et al., 2007). The proportions of food sources (f1, f2) reflected by the isotopic composition of the sampled tissues were estimated using a single isotope mathematical mixing model (Phillips, 2012). We used the d13C values (subscript mix) of the BIR fossil megamammals (Table S1), plus the C3 plants (d13C1 ¼ 11‰) and C4 plants (d13C2 ¼ þ3‰) enriched values (see previous topic), and applied it in the following equations:

f1 ¼

d13 Cmix  d13 C2 d13 C1  d13 C2

f2 ¼ 1  f1

(1)

(2)

Isotope niche breadth (B) was calculated using Levins', 1968 measure:

B¼

1 Sp2i

(3)

Where pi is the relative proportion of individuals in isotope bin i. This measure was then standardized (BA) from 0 to 1 following the equation:

BA ¼

B1 N1

(4)

Where n is the total number of isotope bins available. B and BA are more traditional measures of niche breadth used in Ecology. Hence, we also calculated average niche overlaps (O) through Pianka's (1973) index:

Spij :Spik Ojk ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Sp2ij :Sp2ik

(5)

Where pi is the relative proportion of individuals in bin i. A result of 0 represents zero niche overlap, and 1 represents complete overlap. To enhance the discussion we calculated the estimate weight for the megafauna species that lived in the BIR (Table 1Table S2) using the following regression (Anderson et al., 1985) 2;73 W ¼ 0:078Cðhþf Þ

(6)

Where W is the weight (g), C is the minimum circumference of the humerus and femur diaphysis (mm). We calculated the average

value of the circumferences based on the information available in articles and thesis (Cartelle, 1992; Cartelle et al., 2009; Molena, 2012; Guerin and Faure, 2013; Pereira et al., 2013; Oliveira et al., 2015) and in some collections that were accessible. When the circumference information was not available we estimated it using the minimum width of the humerus and femur diaphysis as a diameter (d) measure (Table S2), through the circumference estimation: C ¼ dp. For the femur of giant ground sloths we used a different circumference estimation, based on the proportions of specimens
rio de Paleontologia” of the from the collection of the “Laborato Federal University of Sergipe: C ¼ d2.4. These megamammals have flat femur with a high circumference of the diaphysis values, leading to an overestimation of the weight if using the standard method. To avoid this problem we multiplied their femur circumference by 0.5, trying to acquire a more realistic weight estimation. For Equus (Amerhippus) neogaeus and the extant African megafauna used here, the estimate weights are from the available articles (Coe et al., 1976; Prado and Alberdi, 1994). 3. Results 3.1. Radiocarbon dating Nine new 14C dates are from the late Pleistocene, showing ecological data for a period ranging from 11 to 21 ka (Fig. 4; Table S1). For the giant ground sloths, five 14C dates are available: one for Valgipes bucklandi in Felipe Guerra/RN (12,114e12,537 cal yr BP; 10,440 ± 35 years BP), and four for Eremotherium laurillardi, two for Poço Redondo/SE (11,597e13,474 cal yr BP; 10,140 ± 40 years BP and 11,540 ± 40 years BP), one for Gararu/SE (13,272e13,474 cal yr BP; 11,540 ± 40 years BP) and Barcelona/RN (11,324e11,807 cal yr BP; 10,050 ± 35 years BP). There are two 14C datings for the gomphothere Notiomastodon platensis, one for Poço Redondo/SE (16,644e17,147 cal yr BP; ~o Sa
/BA 13,950 ± 40 years BP) and another one for Coronel Joa (18,321e18,636 cal yr BP; 15,210 ± 40 years BP). Finally, there is one ~o Sa
/BA dating available for Toxodon platensis from Coronel Joa (14,563e15,176 cal yr BP; 12,580 ± 40 years BP). 3.2. The giant ground sloth Valgipes bucklandi The only specimen of V. bucklandi (weight 980 kg; Table 1) is from Felipe Guerra/RN (05 490 S; Table S1). Its diet was mainly composed of C3 plants from open landscapes (94%; d13C ¼ 10.2‰; d18O ¼ 1.7‰), which reflects a narrow niche breadth (BA ¼ 0.13; Table 1; Fig. 5a), although this conclusion is preliminary and should be confirmed with more isotopic data. 3.3. The giant ground sloth Eremotherium laurillardi The d13C values found for E. laurillardi (weight 6650 Kg; Table 1) are between 7.2‰ and 0.9‰ (Table S1). This interval increases to 9.2‰ to 0.9‰ - a mean value of 4.35 ± 2.87‰ (Table S1, Table 1) - when the data published by other authors (Viana et al., 2011; França et al., 2014; Dantas et al., 2014) are included. Such a large range covers pure C3 (browser), mixed feeders and pure C4 (grazer) consumers. Their Interquartil Range (IQR) was 2.85, confirming a wide niche breadth (mean BA ¼ 0.77 ± 0.25) for the E. laurillardi that lived in the BIR (Table 1; Fig. 5a). The d18O values ranged from 3.2‰ to 0.5‰ (Table S1). When including previously published data, the interval increases to 3.2‰ to 2.5‰ (Viana et al., 2011; França et al., 2014). The mean

M.A.T. Dantas et al. / Quaternary Science Reviews 170 (2017) 152e163

157

Table 1 Mean values of proportional contributions of food sources (pi C3 plants, pi C4 plants), d13C (‰), standardized isotopic niche breadth (BA) and isotopic niche overlap (O) for the Pleistocene megamammals from Brazilian Intertropical Region and extant megamammals from Kenya, Africa. Species

D. bicornis (Db) E. quagga (Eq) C. taurinus (Ct) S. caffer (Sc) L. africana (La) H. amphibius (Ha) Species

E. (A.) neogaeus (En) V. bucklandi (Vb) N. platensis (Np) T. platensis (Tp) E. laurillardi (El) a b c

mass (ton)

1.00a 0.29a 0.22a 0.66a 4.99a 1.40a mass (ton)

0.37b 0.98c 6.00c 3.09c 6.55c

d18O (‰)

n

1.33 ± 1.30 0.51 ± 1.47 2.38 ± 0.87 0.28 ± 1.40 1.08 ± 0.90 3.89 ± 1.54

08 09 08 04 06 07

pi C3

pi C4

mean ± s

IQR

BA

0.79 0.12 0.09 0.05 0.67 0.37

0.21 0.88 0.91 0.95 0.33 0.63

8.05 ± 1.91 1.35 ± 1.23 1.85 ± 1.37 2.85 ± 1.58 6.42 ± 2.14 2.24 ± 1.60

1.65 1.57 1.85 1.95 3.38 2.40

0.50 0.27 0.20 0.11 0.71 0.82

O

± ± ± ± ± ±

0.31 0.21 0.21 0.15 0.29 0.15

Db

Eq

Ct

Sc

La

Ha

e 0.38 0.35 0.31 0.94 0.61

0.38 e 0.89 0.89 0.54 0.84

0.35 0.89 e 0.91 0.52 0.85

0.31 0.89 0.91 e 0.48 0.85

0.94 0.54 0.52 0.48 e 0.66

0.61 0.84 0.85 0.85 0.66 e

d13C (‰)

n

03 01 12 08 14

d13C (‰)

d18O (‰)

pi C3

pi C4

mean±s

IQR

BA

O En

Vb

Np

Tp

El

0.67±2.68 1.74 0.76 ± 1.95 1.28 ± 2.58 1.45 ± 1.44

0.16 0.94 0.30 0.60 0.52

0.84 0.06 0.70 0.40 0.48

0.73±1.19 10.17 1.17 ± 2.76 5.74 ± 4.80 4.35 ± 2.87

1.15 e 1.43 6.69 2.85

0.38±0.22 0.13 0.56 ± 0.20 0.57 ± 0.40 0.77 ± 0.25

e 0.24 0.75 0.51 0.57

0.24 e 0.43 0.85 0.77

0.75 0.43 e 0.76 0.67

0.51 0.85 0.76 e 0.86

0.57 0.77 0.67 0.86 e

Mass presented by Coe et al. (1976). Estimated mass proposed by Prado and Alberdi (1994). Our data (S2 Table).

value for d18O was 1.45 ± 1.44‰ (Table S1). The correlation between d13C and d18O (Fig. 6a) presented a weak correlation (R2 ¼ 0.24, p < 0.05). 3.4. The gomphothere Notiomastodon platensis For this species (weight 6000 Kg; Table 1), the d13C values ranged from 1.5‰ to 1.0‰ (Table S1). By including the previously
nchez published data, this interval increases to 8.2‰ to 1.3‰ (Sa et al., 2004; Viana et al., 2011; França et al., 2014). In the localities of Barcelona/RN (05 570 S), Poço Redondo/SE (09 550 S), Canhoba/SE
/BA (10170 S) this species presented a (10 050 S) and Coronel Jo~ ao Sa grazing diet, while in Ourolandia/BA (10 550 S) it had a mixed feeding diet. The IQR was 1.43, the mean d13C value was 1.17 ± 2.76‰, which represents a niche breadth (BA) of 0.56 ± 0.20 (Table 1; Fig. 5a). The d18O values measured in the current study ranged from 1.7‰ to 0.6‰. When considering previously published data, the interval increases, from 1.7‰ to 3.6‰ (Table S1). The carbon isotopic values showed differences between some localities. In Poço Redondo/SE and Maravilha/AL, d18O values are more positive (0.58e3.60‰; Table S1), while in the other localities they are more negative (1.90 to 0.11‰). The mean value of d18O in BIR was 0.76 ± 1.95‰ (Table 1). The correlation between d13C and d18O is weaker than the correlation for E. laurillardi (R2 ¼ 0.16, p < 0.05; Fig. 6b). 3.5. The “hippo-like” species Toxodon platensis The d13C values of T. platensis (weight 3090 Kg; Table 1) ranged between 12.3‰ and 0.12‰ (Table S1; UGAMS 9444 is an outsider value, not computed in the analysis). This interval slightly increases to 12.6‰ to 0.12‰ when the values from MacFadden (2005) are included. The mean d13C value for the BIR specimens was 5.74 ± 4.80‰, and the interquartil range (IQR) was 6.45. In the localities of Rui Barbosa/RN (05 520 S), Poço Redondo/SE ~o S
(09 550 S) and Coronel Joa a/BA (10170 S) this taxon presented a
ria da grazing diet, while in Ourolandia/BA (10 550 S) and Vito Conquista/BA (14 460 S) its diet was characterized by an important consumption of C3 plants (~60%; sample U-96-148 and UGAMS

9445). Its niche breadth (BA) was 0.57 ± 0.40 (Table 1; Fig. 5a). The d18O values ranged from 5.7‰ to 2.1‰ (Table S1), with a mean value of 1.28 ± 2.58‰ for this species. The correlation between d13C and d18O (Fig. 6c) for this species was good (R2 ¼ 0.43, p < 0.05). 3.6. The horse species Equus (Amerhippus) neogaeus The d13C and d18O values for E. (A.) neogaeus (weight 370 Kg; Prado and Alberdi, 1994) in the Brazilian Intertropical Region were published by MacFadden et al. (1999), and were available only for Ourolandia/BA (Table S1). The mean d13C value for the BIR specimens was 0.73 ± 1.19‰, the interquartil range (IQR) was 1.15, and its niche breadth (BA) was 0.38 ± 0.22 (Table 1; Fig. 5a). The d18O mean value was 0.67 ± 2.68‰. 4. Discussion 4.1. Megamammal chronology in the Brazilian Intertropical Region The gomphothere N. platensis has the highest number of dated specimens (n ¼ 21) by 14C (AMS) and Electron Spin Resonance - ESR (Table S1). The results show that this taxon lived, at least, between 120 ka to 12 ka (Fig. 4). Among the giant ground sloths, Valgipes bucklandi has only one dating (~12 ka, Table S1, Fig. 4), while E. laurillardi has 13 direct datings (14C AMS; Table S1) and one indirect dating (U-series, Table S1), which showed a time range between 27 and 11 ka (Fig. 4). Although most data are from localities in Sergipe, the occurrence of similar dates in localities from Rio Grande do Norte and Bahia (Table S1) allows us to make this interpretation. The occurrence of E. laurillardi in the Middle Pleistocene of the BIR, at 205 ka and 295 ka, is suggested by two indirect datings (Useries, Table S1) made on unidentified mammal bones found in the same layers that yielded the E. laurillardi fossils in the “tank” from Central/BA (Table S1, Fig. 4). Therefore, this time range is also a possibility to this taxon. The absence of more data is due to the fact that the majority of the dates were made by 14C (AMS), which has a limit of about 60 ka. Thus, an effort to make new age determinations with U-series (mainly with cave fossils) is

158

M.A.T. Dantas et al. / Quaternary Science Reviews 170 (2017) 152e163

Fig. 4. Chronology (14C, ESR, U-series) for four extinct Pleistocene megamammal species from the Brazilian Intertropical Region. Abbreviations. (Vb) Valgipes bucklandi; (El) Eremotherium laurilardi; (Np) Notiomastodon platensis; (Tp) Toxodon platensis.

recommended, to generate new data for this taxon. Eight specimens of the “hippo-like” T. platensis were dated (14C and ESR) and showed a time range of 50 ka to 9 ka, suggesting that this taxon lived during the Late Pleistocene and Early Holocene (Fig. 4). The new dates, as well as the review of the published age determinations presented here, expand the chronological data known for the BIR megamammals, and confirm that these species were contemporaneous in the studied localities. Moreover, it suggests the occurrence of these megamammals mainly in the late Pleistocene (Fig. 4). 4.2. Late Pleistocene BIR megamammals: feeding diet (d13C, d18O) and niche breadth All data (Table S1) were interpreted together as belonging to a unique and wide region, the Brazilian Intertropical Region (Fig. 1). The data are from different chronological periods (Table S1),

however, the feeding ecology information for the species E. laurillardi (~11e13 ka x ~18e27 ka; ANOVA, Fobs ¼ 1.107, P ¼ 0.3153; non-parametric KruskaleWallis tests, Hobs ¼ 0.2579, P ¼ 0.6116), N. platensis (~12e18 ka x ~19e21 ka; ANOVA, Fobs ¼ 0.07208, P ¼ 0.7961; non-parametric KruskaleWallis tests, Hobs ¼ 0.0667, P ¼ 0.7963) and T. platensis (~11e12 ka x ~14e15 ka; ANOVA, Fobs ¼ 0.8204, P ¼ 0.4607; non-parametric KruskaleWallis tests, Hobs ¼ 0.6, P ¼ 0.4386) are limited, but similar, and thus were not separated. It is important to note that between 11 and 27 ka, some climatic oscillations occurred in South America. However, as noted by ~o Jose
farm locality (Poço Redondo/SE), França et al. (2014) for the Sa it looks like E. laurillardi and N. platensis did not change their diet through time, suggesting either a stability of the local environmental, or the inability for E. laurillardi and N. platensis to change their diet through a rapid shift of the vegetation during the climatic oscillation. As these species represent members of an extinct assemblage, we will compare their isotopic ecology information with extant megamammals from Kenya, Africa, which live in an open environment, corresponding to the most similar modern environment to the one where the extinct megamammals from BIR probably lived (Cartelle, 1999). We used the available diet isotopic values (d13C, d18O) for six megamammal species from the Amboseli park, in Kenya, Africa (Phacochoerus aethiopicus was not used because there is only one value; Bocherens et al., 1996) as representing the expected interval for grazers (Equus quagga; Connochaetes taurinus; Syncerus caffer), browsers (Dicerus bicornis) and mixed feeders (Loxodonta africana and Hippopotamus amphibius; Fig. 2). To increase the information given by these isotopic data and discuss the data more in-depth, we calculated the niche breadth for each species (Table 1; Fig. 5b). The extant grazers Equus quagga (Zebra; md13C ¼ 1.35 ± 1.25‰; IQR ¼ 1.57), Connochaetes taurinus (Blue Wildebeest; md13C ¼ 1.85 ± 1.37‰; IQR ¼ 1.85), and Syncerus caffer (African buffalo; md13C ¼ 2.85 ± 1.58‰; IQR ¼ 1.95) have a diet composed predominantly of C4 plants, but they consume C3 plants as well (Table 1). E. quagga (weight 290 kg) presents a mean niche breadth of 0.27 ± 0.21, which is larger than C. taurinus (mBA ¼ 0.20 ± 0.21; weight 220 kg) and S. caffer (weight 660 kg), which is virtually a specialist in C4 plants with a mBA ¼ 0.11 ± 0.15. As grazers, they present high d18O values (Table 1), and, for the Zebra and African buffalo there is virtually no correlation between d13C and d18O (both, R2 ¼ 0.01, p < 0.05), indicating that the d18O signature of these species reflects the ingested local water rather than the water from ingested food. In the BIR both N. platensis and E. (A.) neogaeus (ANOVA, Fobs ¼ 1.312, P ¼ 0.2726; non-parametric KruskaleWallis tests, Hobs ¼ 1.718, P ¼ 0.1899) had similar isotopic values. E. (A.) neogaeus (mBA ¼ 0.38 ± 0.22; Table 1) feeding on high content of C4 plants (md13C ¼ 0.73 ± 1.19‰; IQR ¼ 1.15; Table S1; Figs. 3e4), however, their niche breadth is larger compared to E. quagga, C. taurinus and S. caffer, which could be due the limited number of analyzed samples. The correlation between d13C and d18O to E. (A.) neogaeus was surprisingly high (R2 ¼ 0.80, p < 0.05; Fig. 6d) considering that Equus shows a C4 grazing diet. When including data of E. neogaeus from Peru, Colombia and Equador (MacFadden et al., 1999), this correlation drops to R2 ¼ 0.009 (p < 0.05), indicating that the South American Equus was an obligatory drinker. In almost all studied localities, N. platensis shows a mixed feeder diet, with a high proportion of consumed C4 plants (~70%; md13C ¼ 1.17 ± 2.76‰; IQR ¼ 1.43; Table S1; Figs. 3e4). The only ^ndia/BA that showed a high exception was the sample from Ourola

M.A.T. Dantas et al. / Quaternary Science Reviews 170 (2017) 152e163

159

Fig. 5. Comparison of standardized isotopic niche breadth (BA) and mean isotopic niche overlap (O) for the Pleistocene megamammals from the Brazilian Intertropical Region (a and c) and extant megamammals from Kenya, Africa (b and d; Bocherens et al., 1996).

proportion of consumed C3 plants (~70%; d13C ¼ 5.50 and 8.20‰; Table S1). This is corroborated by its moderate niche breadth value (mBA ¼ 0.56 ± 0.20). As E. neogaeus, N. platensis was an obligatory drinker (md18O ¼ 0.76 ± 1.95‰; R2 ¼ 0.16, p < 0.05; Fig. 6b). Our results for the feeding behavior of N. platensis from the BIR are corroborated by dental microwear and dental calculus analyzes, which estimate the diet of several individuals found in different localities of the BIR (Silva, 2015). The most interesting results are from the same samples analyzed here, which are from Canhoba/SE ~o Sa
/BA (UGAMS 09438). Those (UGAMS 09439) and Coronel Joa specimens also showed a C4 plant-based diet. The authors also found results similar to ours for fossils from Paraíba (PB), Pernambuco (PE), Piauí (PI) and Alagoas (AL), reinforcing our hypothesis. The N. platensis enriched d18O values found in Poço Redondo/SE are consistent with what is expected for a grazer, since C4 plants have high evaporation rates (Lopes et al., 2013). The values found in

other localities could reflect variations in the isotopic composition of meteoric water, which could explain the low oxygen isotope values, which is unexpected for grazing animals. The browser guild in Amboseli is represented by Dicerus bicornis (black rhinoceros), which has a moderate niche breadth (mBA ¼ 0.50 ± 0.31), feeding on almost 79% of C3 plants (d13C ¼ 10.05 ± 1.91‰; IQR ¼ 1.65; Table 1; Fig. 5b). Analogously, V. bucklandi shows a browsing diet in the BIR (d13C ¼ - 10.2‰; Table S1; Figs. 3e4) with a diet based on 94% of C3 plants. Its niche breadth was narrower (BA ¼ 0.13; Table 1) to that of D. bicornis. However, as we only have one measured sample so far, and considering the very large range of d13C values for the other megamammals in the region, it is too early to draw firm conclusions. H. amphibius fed on C4 plants (63%) and aquatic C3 plants (37%), presenting a mean d13C value of 4.24 ± 1.60‰, and a IQR of 2.40. However, the d18O values of both species are different, being more positive in the African elephant (md18O ¼ 1.08 ± 0.90‰) than in the hippopotamus (md18O ¼ 3.89 ± 1.54‰). This probably reflects

160

M.A.T. Dantas et al. / Quaternary Science Reviews 170 (2017) 152e163

Fig. 6. Correlation between d13C and d18O values of the megamammal species in Brazilian Intertropical Region. (a) Eremotherium laurillardi; (b) Notiomastodon platensis; (c) Toxodon platensis; (d) Equus (Amerhippus) neogaeus.

the consumption of aquatic C3 plants as well as their amphibious behavior that limits the loss of water by transpiration, causing the depletion of 18O (Bocherens et al., 1996). The African elephant (L. africana) and the hippopotamus (H. amphibius) from Amboseli are mixed feeders (Table 1), where L. africana has 67% of C3 plants and 33% of C4 plants in its diet. Their niche breadth is large (mBA ¼ 0.71 ± 0.29), but narrower than the suggested niche breadth of the hippopotamus (mBA ¼ 0.82 ± 0.15; Table 1; Fig. 5b). The correlation between d13C and d18O values in L. africana is similar to grazer species (R2 ¼ 0.13, p < 0.05), suggesting that this animal was also an obligatory drinker. The available isotopic values (d13C) for the feeding behavior of E. laurillardi and T. platensis allow us to infer that both species had similar feeding strategies (mixed feeding diet). Their niche breadth (T. platensis, mBA ¼ 0.57 ± 0.40; E. laurillardi, mBA ¼ 0.77 ± 0.25) is comparable to the ones of L. africana and H. amphibius, however, the correlation between the d13C and d18O values of T. platensis (R2 ¼ 0.43, p < 0.05; Fig. 6c) indicate that a great part of the water consumed by them was from their food. Yet, E. laurillardi (R2 ¼ 0.24, p < 0.05; Fig. 6a) presents a value close to values of L. africana. There are few data available for T. platensis, but in general it shows a mixed-feeding diet with high proportion of consumed C4 plants (~60%; Table S1; Fig. 3). The oxygen isotope data of T. platensis show significant differences between these localities. In ~o Sa
/BA and Pedra Vermelha/BA the oxPoço Redondo/SE, Cel. Joa ygen isotope values correspond to C4 plant feeders, while in Our
ria da Conquista/BA it coincide with C3 plant ol^ andia/BA and Vito feeders. The only exceptions are the data from Rui Barbosa/RN, which could be explained by the results for N. platensis from Bar~o Sa
/BA, which represent the celona/RN, Canhoba/SE and Cel. Joa diet of a moister period/rainy season. Generally T. platensis is called as a “hippo-like” animal, however, its oxygen values were not similar to those found for Hippopotamus amphibius, as in this taxon the oxygen isotope values are much more negative than the values of the terrestrial fauna (Bocherens et al., 1996, Tables 2; Table S1; Figs. 3e4). A similar result was found for Toxodon fossil samples from Rio Grande do Sul (Lopes et al., 2013). These animals were therefore probably terrestrial

herbivores being more similar to a rhinoceros than to a hippopotamus, thus, in a comparison with the megamammals from Africa, would be better call them “rhino-like” than “hyppo-like” animals. The results for E. laurillardi showed a mixed feeding diet in all studied localities (Table S1; Fig. 3). The d13C and d18O values suggest an opportunistic behavior with a contribution of C4 plants and leaves and fruits of C3 plants. Based in the carbon isotopic information for the extant megamammals from Amboseli, Africa, we noticed that specialists (grazers and browsers) present a narrow carbon isotopic variation (IQR) in comparison with generalists (mixed diet). This pattern was observed in the carbon isotopic signature for the megamammal species from the BIR as well (Fig. 7). 4.3. Late Pleistocene BIR megamammals: niche overlap and assemblage structure The assemblage structure of the African megamammals from Amboseli is controlled primarily by the abundance of food resources (bottom-up controlled, Sinclair, 1975), with African elephant (Loxodonta africana) being a key species in the structuring of savannah ecosystems, acting in the modeling of the environment, facilitating the access to resources by other species (e.g. middle sized mammals, Makhabu et al., 2006), and limiting the abundance of other megamammal species through the competition for resources (Fritz et al., 2002). In our analysis, L. africana exhibited the widest niche breadth (mBA ¼ 0.71 ± 0.29; weight 4990 Kg) among the terrestrial herbivorous species of Amboseli (Table 1; Fig. 5b). This niche breadth value plus its high weight can be used as criterion to identify a key species in the structure of an extinct ecosystem, as we have examples that L. africana influenced directly the dynamic of other species in the African savanna ecosystem, such as D. bicornis (mBA ¼ 0.50 ± 0.31; weight 1.000 Kg) and E. quagga (mBA ¼ 0.27 ± 0.21; weight 290 Kg). The competition between L. africana and D. bicornis is weak during the rainy season, because they have a clear separation in their diet. Therefore, their niche overlap only occurs during the dry

M.A.T. Dantas et al. / Quaternary Science Reviews 170 (2017) 152e163

161

Fig. 7. Cluster analysis of the carbon isotopic variation (Interquartile Range - IQR) of megamammals from Amboseli, Africa and from the Brazilian Intertropical Region.

season, when both species browse and share the same type of resources (Landman et al., 2013). Our analyses on niche overlap suggest a high degree of competition between these species (O ¼ 0.94) due the high content of C3 plants (Table 1; Fig. 5d). African elephants compete directly with zebras for grass as well, having an increase on the L. africana population coeval with a decrease of the zebra population (Boer et al., 2015). This interspecific competition probably occurs mainly during the dry season. Our results suggest a moderate overlap of 0.54 between African elephants and Zebras (Table 1; Fig. 5d), which is lower than the overlap between African elephants and black rhinoceroses. This could be due to the low proportion of C4 plants consumed by the African elephant (~33%), in comparison with C3 plant intake (~67%; Table 1). The overlap between D. bicornis and E. quagga is moderate (O ¼ 0.38; Table 1; Fig. 5d), maybe due their opposite feeding strategies. Hippopotamus presented a moderate overlap between L. africana (O ¼ 0.66), D. bicornis (O ¼ 0.61), but a high overlap with C. taurinus (O ¼ 0.85), S. caffer (O ¼ 0.85) and E. quagga (O ¼ 0.84). However, the C3 plants consumed by Hippopotamus are aquatic (lower values of 18O), and, thus, it does not represent a true competition with the terrestrial herbivorous species (Bocherens et al., 1996). The isotopic values of the extant megamammals from Amboseli (Kenya, Africa; Bocherens et al., 1996) do not reflect seasonal differences in their diets, as it is the case for our data for the BIR megamammals, preventing us to present a more precise interpretation of seasonal differences. However, we noticed that since the competition between L. africana, D. bicornis and E. quagga is high during the dry seasons, the overlap values observed for these taxa could reflect the high competition in this season. Thus, using the overlap values found for the BIR megamammals we will derive assumptions about resources competitions in dry seasons, as well. By analyzing the niche overlaps of the studied species, we noticed that Valgipes bucklandi (weight 980 Kg; browser diet; BA ¼ 0.13) had low overlap, and therefore low competition for resources, with Equus (Amerhippus) neogaeus (weight 370 Kg; C4

grazer; O ¼ 0.24; Table 1; Fig. 5c) and Notiomastodon platensis (weight 6000 Kg; mixed feeder diet, high consumption of C4 grazer; O ¼ 0.43; Table 1; Fig. 5c), due to the opposite feeding diet strategies. The overlap with E. laurillardi (weight 6550 Kg; O ¼ 0.77) and T. platensis (weight 3090 Kg; O ¼ 0.85) was high due to the consumption of C3 plants. The gomphothere N. platensis (mBA ¼ 0.56 ± 0.20) presents high overlap between E. laurillardi (O ¼ 0.67) and T. platensis (O ¼ 0.76), due their high consumption of C4 plants. Yet, N. platensis has a higher overlap with E. neogaeus (O ¼ 0.75; Table 1; Fig. 5c), which could be interpreted as competitive exclusion. It is important to notice that N. platensis presented a moderate niche breadth (mBA ¼ 0.56 ± 0.20), being more restricted than L. africana (mBA ¼ 0.71 ± 0.29), and although N. platensis had similar body weight (6000 Kg), it apparently did not had the same role in the structure of the ecosystem as L. africana does. The species T. platensis and E. laurillardi presented moderate niche overlaps (O ¼ 0.51 and 0.57; Table 1; Fig. 5c) with E. neogaeus. This probably does not mean that a competitive exclusion occurred between these species. The isotopic results do not allow us to refine what kind of plants they fed on. Thus, this overlap shows the plant species richness of the community in which they lived, showing that these and other species (extant and extinct) shared the same resources. T. platensis (mBA ¼ 0.57 ± 0.40) has a narrow niche breadth in comparison with L. africana. It seems that E. laurillardi (mBA ¼ 0.77 ± 0.25) was the main species influencing the dynamics of this ecosystem. There is a high overlap between T. platensis and E. laurillardi (O ¼ 0.86) that reflects the similar food content of their diets. However, due the wide niche breadth and high body weight of E. laurillardi, we believe that this species could be a better competitor for resources, and thus may have limited directly the growth of the T. platensis population. A similar pattern could be suggested for N. platensis and E. neogaeus, in which the gomphothere could have had an advantage over Equus.

162

M.A.T. Dantas et al. / Quaternary Science Reviews 170 (2017) 152e163

5. Final remarks The new datings, together with the revision of the published data (14C dating; ESR; U-series), allow us to suggest that all mammal specimens analyzed here lived in the Brazilian Intertropical Region during the Middle and late Pleistocene, at least between 300 ka to 10 ka. During the late Pleistocene, there were species belonging to different guilds that lived in this region: grazers, Equus (Amerhippus) neogaeus (md13C ¼ 0.73 ± 1.19‰) and Notiomastodon platensis (md13C ¼ 1.17 ± 2.76‰); browsers, Valgipes bucklandi (d13C ¼ 10.17‰); and mixed feeders, the “rhino-like” Toxodon platensis (md13C ¼ 5.74 ± 4.80‰) and the ground giant sloth Eremotherium laurillardi (md13C ¼ 4.35 ± 2.87‰). The niche breadth (BA) of the megamammal species from the BIR suggests a narrow niche breadth for V. bucklandi (BA ¼ 0.13) and E. neogaeus (mBA ¼ 0.38 ± 0.22), a moderate niche breadth for N. platensis (mBA ¼ 0.56 ± 0.20) and T. platensis (mBA ¼ 0.57 ± 0.40), while E. laurillardi had the largest niche breadth (mBA ¼ 0.77 ± 0.25). The comparison between E. neogaeus (mBA ¼ 0.38 ± 0.22) and the grazing species of Amboseli indicates that its niche had a larger breadth than S. caffer (mBA ¼ 0.11 ± 0.15), C. taurinus (mBA ¼ 0.20 ± 0.21) and E. quagga (mBA ¼ 0.27 ± 0.21), probably due the low number of samples for E. (A.) neogaeus or the relatively high consumption of C3 plants (~16%) for a grazer. The browsing species V. bucklandi (BA ¼ 0.13) had a narrower value of niche breadth compared to D. bicornis (mBA ¼ 0.50 ± 0.31) due the higher consumption of C3 plants, which could suggest that it was a specialist species. In addition E. laurillardi (mBA ¼ 0.77 ± 0.25) showed similar niche breadth values than L. africana (mBA ¼ 0.71 ± 0.29). The sampled megamammal species from the BIR showed a larger Interquartile Range of stable isotope values when compared to the Amboseli species. The Pleistocene megamammal community from the BIR had similar niche overlaps than the extant megamammal assemblage from Kenya, Africa. These results further elucidate how this megamammal community was structured, and suggest that E. laurillardi played an important ecological role in the BIR during the Pleistocene. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.quascirev.2017.06.030. References Anderson, J.F., Hall Martin, A., Russel, D.A., 1985. Long-bone circumference and weight in mammals, birds and dinosaurs. J. Zool. 207, 53e61. Araújo Jr, HI de, Porpino, K. de O., Ximenes, C.L., Bergqvist, L.P., 2013. Unveiling the taphonomy of elusive natural tank deposits: a study case in the Pleistocene of northeastern Brazil. Palaeogeogr. Palaeoclimatol. Palaeoecol. 378, 52e74. http:// dx.doi.org/10.1016/j.palaeo.2013.04.001. Bearhop, S., Adams, C.E., Waldron, S., Fuller, R.A., MacLeod, H., 2004. Determining trophic niche width: a novel approach using stable isotope analysis. J. Anim. Ecol. 73 (5), 1007e1012. Bocherens, H., Cotte, M., Bonini, R., Scian, D., Straccia, P., Soibelzon, L., Prevosti, F.J., 2016. Paleobiology of sabretooth cat Smilodon populator in the pampean region (buenos aires province, Argentina) around the last glacial maximum: insights from carbon and nitrogen stable isotopes in bone collagen. Palaeogeogr. Palaeoclimatol. Palaeoecol. 449, 463e474. Bocherens, H., Cotte, M., Bonini, R., Scian, D., Straccia, P., Soibelzon, L., Prevosti, F.J., 2017. Isotopic insight on paleodiet of extinct Pleistocene megafaunal xenarthrans from Argentina. Gondwana Res. 48, 7e14. Bocherens, H., Koch, P.L., Mariotti, A., Geraards, D., Jaeger, J.J., 1996. Isotopic biogeochemistry (13C, 18O) of mammalian enamel from African Pleistocene hominid sites. Palaios 11, 306e318. Boer, W.F., Van Oort, J.W.A., Grover, M., Peel, M.J.S., 2015. Elephant-mediated habitat modifications and changes in herbivore species assemblages in Sabi Sand, South Africa. Eur. J. Wildl. Res. 61, 491e503. http://dx.doi.org/10.1007/s10344-

015-0919-3. Cartelle, C., De Iuliis, G., Ferreira, R.L., 2009. Systematic revision of tropical brazilian Scelidotheriine sloths (Xenarthra, Mylodontoidea). J. Ver. Paleo 29 (2), 555e566. Cartelle, C., 1992. Edentata e megamamíferos herbívoros extintos da toca dos ossos ^ndia, BA). Tese Doutorado, Univ. Fed. Minas Gerais. (Ourola Cartelle, C., 1999. Pleistocene mammals of the cerrado and caatinga of Brazil. In: Eisenberg, J.B., Redford, K.H. (Eds.), Mammals of the Neotropics. The Central Neotropics. University of Chicago Press, Chicago, pp. 27e46. Cerling, T.E., Harris, J.M., 1999. Carbon isotope fractionation between diet and bioapatite in ungulate mammals and implications for ecological and paleoecological studies. Oecologia 120, 347e363. Cherkinsky, A., Ravi Prasad, C.V., Dvoracek, D., 2013. AMS measurement of samples smaller than 300 lg at center for applied isotope studies, University of Georgia. Nucl. Instrum. Methods Phys. Res. A 294, 87e90. Cherkinsky, A., 2009. Can we get a good radiocarbon age from “bad bone”? determining the reliability of radiocarbon age from bioapatite. Radiocarbon 51, 647e655. Coe, M.J., Cumming, D.H., Phillipson, J., 1976. Biomass and production of large African herbivores in relation to rainfall and primary production. Oecologia 22, 341e354. Coplen, T.B., 1994. Reporting of stable hydrogen, carbon, and oxygen isotopic abundances. Pure Appl. Chem. 66, 273e276. Dantas, M.A.T., Dutra, R.P., Cherkinsky, A., Fortier, D.C., Kamino, L.H.Y., Cozzuol, M.A., Ribeiro, A.S., Silva, F.V., 2013a. Paleoecology and radiocarbon dating of the Pleistocene megafauna of the brazilian intertropical region. Quat. Res. 79, 61e65. http://dx.doi.org/10.1016/j.yqres.2012.09.006.
Dantas, M.A.T., Santos, D.B., Liparini, A., Queiroz, A.N., Carvalho, O.A., Castro, E.S.V., Cherkinsky, A., 2014. A dated evidence of the interaction between humans and megafauna in the late Pleistocene of Sergipe state, northeastern Brazil. Quat. Int. 352, 197e199. http://dx.doi.org/10.1016/j.quaint.2014.09.040. Dantas, M.A.T., Xavier, M.C.T., França, L.M., Cozzuol, M.A., Ribeiro, A.S., Figueiredo, A.M.G., Kinoshita, A., Baffa, O., 2013b. A review of the time scale and potential geographic distribution of Notiomastodon platensis (Ameghino, 1888) in the late Pleistocene of South America. Quat. Int. 317, 73e79. http://dx.doi.org/ 10.1016/j.quaint.2013.06.031. Domingo, L., Prado, J.L., Alberdi, M.T., 2012. The effect of paleoecology and paleobiogeography on stable isotopes of Quaternary mammals from South America. Quat. Sci. Rev. 55, 103e113. http://dx.doi.org/10.1016/j.quascirev.2012.08.017. França, L.M., Dantas, M.A.T., Bocchiglieri, A., Cherkinsky, A., Ribeiro, A.S., Bocherens, H., 2014. Chronology and ancient feeding ecology of two upper Pleistocene megamammals from the Brazilian Intertropical Region. Quat. Sci. Rev. 99, 78e83. http://dx.doi.org/10.1016/j.quascirev.2014.04.028. Friedman, I., O'Neil, J.R., 1977. Compilation of stable isotope fractionation factors of geochemical interest. USGPO. Fritz, H., Duncan, P., Gordon, I.J., Illius, A.W., 2002. Megaherbivores influence trophic guilds structure in African ungulate communities. Oecologia 131, 620e625.
rin, C., Faure, M., 2013. Un nouveau Toxodontidae (Mammalia, Notoungulata) du Gue
istoce ne supe
rieur du Nordeste du Bre
sil. Geodiversitas 35 (1), 155e205. Ple http://dx.doi.org/10.5252/g2013n1a7. Harrison, R.G., Katzenberg, M.A., 2003. Paleodiet studies using stable carbon isotopes from bone apatite and collagen: examples from southern Ontario and San Nicolas Island, California. J. Anthropol. Archaeol. 22, 227e244. Helliker, B.R., Ehleringer, J.R., 2000. Establishing a grassland signature in veins: 18O in the leaf water of C3 and C4 grasses. Proc. Natl. Acad. Sci. U. S. A. 97, 7894e7898. Holmes, K.M., Brown, K.A.R., Oates, W.P., Collins, M.J., 2005. Assessing the distribution of African Palaeolithic sites: a predictive model of collagen degradation. J. Archaeol. Sci. 32, 157e166. Holmes, K.M., Brown, K.A.R., Oates, W.P., Collins, M.J., 2006. Assessing the distribution of Asian Palaeolithic sites: a predictive model of collagen degradation. J. Archaeol. Sci. 33, 971e986. Hubbe, A., Auler, A.S., 2012. A large Cervidae Holocene accumulation in Eastern Brazil: an example of extreme taphonomical control in a cave environment. Int. J. Speleol. 41 (2), 299e307. http://dx.doi.org/10.5038/1827-806X.41.2.15. Keeling, C.D., 1979. The Suess effect: 13Carbon-14Carbon interrelations. Environ. Int. 2, 229e300. Kohn, M.J., Schoeninger, M.J., Valley, J.W., 1996. Herbivore tooth oxygen isotope compositions: effects of diet and physiology. Geochim. Cosmochim. Acta 60, 3889e3896. Landman, M., Schoeman, D.S., Kerley, G.I.H., 2013. Shift in black rhinoceros diet in the presence of elephant: evidence for competition? PLoS One 8 (7), e69771. http://dx.doi.org/10.1371/journal.pone.0069771. Lee-Thorp, J., Sponheimer, M., 2003. Three case studies used to reassess the reliability of fossil bone and enamel isotope signals for paleodietary studies. J. Anthropol. Archaeol. 22, 208e216. Levins, R., 1968. Evolution in Changing Environments. Princeton University Press, New Jersey. IXþ120pp. Lopes, R.P., Ribeiro, A.M., Dillenburg, S.R., Schultz, C.L., 2013. Late middle to late Pleistocene paleoecology and paleoenvironments in the coastal plain of Rio Grande do Sul State, Southern Brazil, from stable isotopes in fossils of Toxodon and Stegomastodon. Palaeogeogr. Palaeoclimatol. Palaeoecol. 369, 385e394. MacFadden, B.J., Cerling, T.E., Harris, J.M., Prado, J., 1999. Ancient latitudinal gradients of C3/C4 grasses interpreted from stable isotopes of New World Pleistocene horse (Equus) teeth. Glob. Ecol. Biogeogr. 8, 137e149.

M.A.T. Dantas et al. / Quaternary Science Reviews 170 (2017) 152e163 MacFadden, B.J., DeSantis, L.R.G., Hochstein, J.L., Kamenov, G.D., 2010. Physical properties, geochemistry, and diagenesis of xenarthran teeth: prospects for interpreting the paleoeology of extinct species. Palaeogeogr. Palaeoclimatol. Palaeoecol. 291, 180e189. MacFadden, B.J., 2005. Diet and habitat of toxodont megaherbivores (mammalia, notoungulata) from the late quaternary of South and Central America. Quat. Res. 64, 113e124. Makhabu, S.W., Skarpe, C., Hytteborn, H., 2006. Elephant impact on shoot distribution on trees and re-browsing by smaller browsers. Acta Oecol 30, 136e146. Molena, F.P., 2012. Variaç~ ao individual no esqueleto apendicular de mastodontes ~o Bento do Una, Pernam(Proboscidea: Gomphotheriidae) provenientes de Sa ~o Paulo, S~ buco, Brasil. Dissertaç~ ao de Mestrado. Universidade de Sa ao Paulo. Nelson, B.K., DeNiro, M.J., Schoeninger, M.J., De Paolo, D.J., 1986. Effects of diagenesis on strontium, carbon, nitrogen and oxygen concentration and isotopic composition of bone. Geochimica Cosmochimica Acta 50, 1941e1949. Newsome, S.D., Rio, C.M., Bearhop, S., Phillips, D.L., 2007. A niche for isotopic ecology. Front. Ecol. Environ. 5 (8), 429e436. http://dx.doi.org/10.1890/ 060150.01. ^ncia de GomphoOliveira, G.C., Viana, M.S.S., Oliveira, E.V., 2015. Primeira ocorre theriidae no município de Irauçuba, Cear
a, Brasil. Rev. Bras. Paleo 18 (2), 339e342. http://dx.doi.org/10.4072/rbp.2015.02.12. Pereira, I.C.S., Dantas, M.A.T., Ferreira, R.L., 2013. Record of the giant sloth Valgipes bucklandi (Lund, 1839) (Tardigrada, Scelidotheriinae) in Rio Grande do Norte state, Brazil, with notes on taphonomy and paleoecology. J. South Am. Earth Sci. 43, 42e45. http://dx.doi.org/10.1016/j.jsames.2012.11.004.
rez-Crespo, V.A., Arroyo-Cabrales, J., Alva-Valdivia, L.M., Morales-Puente, P., Pe Cienfuegos-Alvarado, E., 2012. Diet and habitat definitions for Mexican glyp
xico) based on stable isotope analysis. todonts from Cedral (San Luis Potosí, Me Geol. Mag. 149, 153e157. Person, A., Bocherens, H., Mariotti, A., Renard, M., 1996. Diagenetic evolution and experimental heating of bone phosphate. Palaeogeogr. Palaeoclimatol. Palaeoecol. 126, 135e150. Phillips, D.L., 2012. Converting isotope values to diet composition: the use of mixing

163

models. J. Mammal. 93 (2), 342e352. http://dx.doi.org/10.1644/11-MAMM-S158.1. Pianka, E.R., 1973. The structure of lizard communities. Annu. Rev. Ecol. Syst. 4, 53e74. Prado, J., Alberdi, M.T., 1994. A quantitative review of the horse Equus from South America. Palaeontology 37, 459e481. Reimer, P.J., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Bronk Ramsey, C., Buck, C.E., Cheng, H., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P.,
, C., Heaton, T.J., Hogg, A.G., Hughen, K.A., Haflidason, H., Hajdas, I., Hatte Kaiser, K.F., Kromer, B., Manning, S.W., Niu, M., Reimer, R.W., Richards, D.A., Scott, E.M., Southon, J.R., Turney, C.S.M., Van der Plicht, J., 2013. IntCal13 and MARINE13 radiocarbon age calibration curves 0-50000 years calBP. Radiocarbon 55 (4), 1869e1887. http://dx.doi.org/10.2458/azu_js_rc.55.16947. S
anchez, B., Prado, J.L., Alberdi, M.T., 2004. Feeding ecology, dispersal, and extinction of South american Pleistocene gomphotheres (gomphotheriidae, proboscidea). Paleobiology 30, 146e161. Sillen, A., Parkington, J., 1996. Diagenesis of bones from Eland's bay cave. J. Archaeol. Sci. 23, 535e542.
rios (proboscidea: mammalia) Silva, L.A., 2015. Paleoecologia alimentar dos gonfote ^nicos da Ame
rica do Sul. Diss. Mestr. Univ. Fed. do Estado do Rio J. pleistoce Sinclair, A.R.E., 1975. The resource limitation of trophic levels in tropical grassland ecosystems. J. Anim. Ecol. 44, 497e520. Tütken, T., Vennemann, T.W., Pfretzschner, H.-U., 2008. Early diagenesis of bone and tooth apatite in fluvial and marine settings: constraints from combined oxygen isotope, nitrogen and REE analysis. Palaeogeogr. Palaeoclimatol. Palaeoecol. 266, 254e268. Viana, M.S.S., Silva, J.L.L., Oliveira, P.V., Juli~ ao, M.S.S., 2011. H
abitos alimentares em ^nica do Nordeste do Brasil. Estud. Geol. 21 herbívoros da megafauna pleistoce (2), 89e95.
rios Ximenes, C.L., 2008. Tanques Fossilíferos de Itapipoca, CE - bebedouros e cemite
-histo
rica. In: Winge, M., Schobbenhaus, C., Souza, C.R.G., de megafauna pre
gicos e Fernandes, A.C.S., Berbert-Born, M., Queiroz, E.T. (Eds.), Sítios Geolo
gicos do Brasil. SIGEP, Brasília, pp. 1e13. Paleontolo