Faunal migration in late-glacial central Italy: implications for human resource exploitation

Faunal Migration in Late Glacial Central Italy and implications for Human

Resource Exploitation

Maura Pellegrini,1* Randolph E. Donahue,2* Carolyn Chenery,3 Jane Evans,3 Julia

Lee-Thorp,2 Janet Montgomery,2 and Margherita Mussi4

1 Istituto di Biologia Agro-Ambientale e Forestale (CNR) via Marconi, 2 Porano,

Terni, 05010, Italy

2 Division of Archaeological, Geographical and Environmental Sciences, University

of Bradford, Bradford BD7 1DP, UK

3 NERC Isotope Geosciences Laboratory, Keyworth, Nottingham NG12 5GG, UK

4 Dip. di Scienze Storiche, Archeologiche e Antropologiche dell'Antichità, Università

di Roma, via Palestro, 63, Roma, 00195, Italy

* Corresponding authors: e-mail: [email protected]; [email protected]

Maura Pellegrini, present address: Max Planck Institute for Evolutionary Anthropology, Department of Human Evolution, Deutscher Platz 6, D-04103, Leipzig, Germany

1

Abstract

The hunter-gatherer transhumance model presents foragers as specialised

hunters of migratory ungulates, which moved seasonally between coastal lowlands

and interior uplands. We studied six animal teeth of horse (Equus hydruntinus) and

red deer (Cervus elaphus) from four different archaeological sites: the Grotta di Vado

all’Arancio, Grotta di Settecannelle, Grotta Polesini and Grotta di Pozzo, in central

Italy to test whether migratory patterns and seasonal variations recorded in their teeth

were consistent with expectations of the transhumance model for this region during

the late Upper Palaeolithic. Sequential subsamples of enamel were analysed from

each tooth for oxygen, carbon and strontium isotope ratios to reconstruct mobility and

yearly seasonal variations. Results show little evidence that these animals were

moving over different geological terrains throughout the year, although small

variations in Sr isotope ratios and concentrations were detected that corresponded to

likely seasonal variations as shown by variability in oxygen isotope sequences. The

results do, however, demonstrate that Cervus elaphus and Equus hydruntinus had

different ranging behaviours, with the former moving over wider areas compared to

the latter. This methodology produces results appropriate to assess animal migratory

behaviour and, in turn, to test the consistency of proposed models of hunter-gatherer

subsistence and mobility strategies.

Keywords: 87Sr/86Sr, Bioapatite δ18O and δ13C, Epigravettian, Hunter-gatherers,

Migration, Late-glacial, Italy

2

1.0 Introduction

Eurasian hunter-gatherer populations experienced rapid and extreme

environmental changes from the Last Glacial Maximum to the early Holocene

(c.18,000 to 8,000 BP).e.g., 1,2 During this period, profound socio-economic changes

and technological developments took place culminating in the transition to

agriculture. Various models have been applied to explore the relationship between the

environment and hunter-gatherer behaviour during the Late Glacial in Mediterranean

Europe. The hunter-gatherer transhumance model, a variation of the seasonal

mobility model originally developed by Higgs,3,4 presents foragers as specialised

hunters of migratory ungulates that moved seasonally between coastal lowlands and

interior uplands. In Italy, Barker hypothesised that hunter-gatherers of the late

Pleistocene followed seasonally migratory herd species including Cervus elaphus and

Equus hydruntinus between the Apennine uplands and the coastal plains.5,6 These two

locations offer seasonally distinct opportunities for grazing species: in winter the

mountains are inhospitable and cold while the coastal plains remain grass-covered. In

summer, instead, the lowlands are dry with little vegetation and the mountains remain

vegetated. These distinctions would have been intensified in the late Pleistocene, and

Barker argued that the seasonal climatic variations of this Mediterranean landscape

would have selected for wild species of grazing herbivores that could migrate

between these two regions.6:114

Although the hunter-gatherer transhumance model is attractive for various

reasons, attempts to test it based on archaeological evidence have brought forth

ambiguous results. Since the model is predicated on seasonal migration between the

highlands of the Apennines and the Tyrrhenian coast of central Italy, we tested its

central precept by applying sequential combined oxygen (18O/16O), carbon (13C/12C)

3

and strontium (87Sr/86Sr) isotope measurements in tooth enamel crowns to reconstruct

seasonal environmental variations as well as animal mobility patterns. Skeletal

mineral (bioapatite) 13C/12C and 18O/16O reflect the isotopic chemistry of food and

drink that the animal acquired in vivo. In the case of herbivores, 13C/12C reflects the

composition of the ingested vegetation, primarily C3 versus C4 or, at a lower level of

variability, a number of factors including the level of humidity or aridity in the

environment, solar radiation, and the concentration of CO2.e.g.,7 The 18O/16O ratio in

the phosphate and carbonate components of fossil bioapatite in individuals reflects

the isotopic composition of the total ingested water in food and drink8,9 and, in turn,

climate variability via the effects of temperature and Rayleigh distillation on the

isotopic composition of precipitation environmental water. Sequential intra-tooth

measurements of 18O/16O and 13C/12C along the direction of growth are useful to

assess the compositional variations of these parameters along a tooth mineralization

trajectory, as they reflect seasonal environmental variations during this period.10-13

Mobility (or residence) patterns can be reflected in 87Sr/86Sr from tooth enamel, since

Sr deposited in the skeletal tissues from food and drink links the animal to the

geographical region from where it sourced its food and drink. 11 The ability to detect

movement across a landscape, or landscapes, requires that the underlying geologies

across which the animal traverses have differing 87Sr/86Sr compositions. Strontium

varies in different geological formations because of variations in their age and/or

initial amount of Rb.

Recent studies of combined intra-tooth and intra-jaw enamel stable and

radiogenic isotope variations have proved to be useful tools for tracing prehistoric

animal and human mobility.10-15 On this basis, we performed a detailed isotopic

4

investigation on animal dental remains of ungulates from four archaeological sites in

central Italy.

2.0 The geology and 87Sr/86Sr biosphere variations in Italy

The geology of Italy (Fig. 1) is dominated by the creation of the Alps when

the European tectonic plate converged with the Adriatic part of the African Plate at

around 100 Ma. The geology of the north is dominated by this Alpine Belt, formed by

the stacking of sedimentary and igneous sequences, and the large basins of

sedimentary debris along the flanks of the Alpine range16,17 Central and southern Italy

comprises mostly of limestones and other sedimentary rocks of predominantly

Cretaceous age. In western Tuscany older, Palaeozoic basement rocks of possible

Later Cambrian to Devonian Age are exposed, whereas southern Tuscany and

Latium, in central-western Italy, is dominated by Cenozoic volcanism.

These geological variations also need to be interpreted in terms of possible

87Sr/86Sr variations in the biosphere, which may not perfectly reflect the bulk isotopic

composition of the underlying rocks or overlying soils, as biologically available

strontium is likely to be dominated by the more soluble carbonate fraction.18-24 In

addition, the 87Sr/86Sr composition of animals from a specific locality is influenced by

non-geologic sources, such as rainwater.25-27

There is, however, limited 87Sr/86Sr biosphere data currently available for the

study region, and we have, therefore, largely relied on predictions based on the

geological literature and studies from comparable geological regions.

In the Alpine region to the north, relatively high Rb content and/or age of the

rocks provide relatively high (radiogenic) 87Sr/86Sr values, as shown by values of

about 0.72 recorded in tooth enamel of modern inhabitants of the Italian-Austrian

5

border region.28 The tectonic window within Tuscany that exposes Palaeozoic schists

should also generate radiogenic biosphere values, and comparable rocks in Britain

give values within the range 0.711-0.714.27 The Cenozoic volcanic rocks around

Rome are alkaline29 with a possible wide range of 87Sr/86Sr depending upon the

specific lithology. Some 87Sr/86Sr data from Elephas (palaeoloxodon) antiquus teeth

from two late mid Pleistocene are available to provide an estimate for the region;30

dentine values are 0.7099±0.0002 (2sd, n=8), while tooth enamel gives

0.7100±0.0002 (2sd, n=29). Since dentine is poorly crystalline and reacts with the

burial environment (unlike enamel which is far more resistant to diagenesis)31 it

provides a good estimate of the local biosphere value. Palombo et al.30 concluded that

the animals had grazed the region where their remains were found. Limestone

87Sr/86Sr values can be estimated from seawater curves.32 For Cretaceous deposits

pure limestone has relatively low values between 0.7078 and 0.7071. This may be

moderated by clay content, and other dietary contributions, but it will be the main

source of Sr for herbivores grazing pasture on limestone.

In summary, the geological complexity of Italy potentially provides a wide

range of 87Sr/86Sr biosphere values. The Po basin and the limestone regions will

likely supply the least variable signatures. The Alps should show the highest diversity

of rock types and thus of biosphere Sr isotope compositions. The volcanic area

appears to have a distinct signature and the areas of older Palaeozoic and basement

rock (parts of Tuscany and southern Italy) should generate more radiogenic (i.e.

higher 87Sr/86Sr) values. Since this is a pilot project, we relied of necessity on limited

sources of information, but a more thorough study would require us to sample all

major grazing areas in order to obtain a representative strontium isotope map of the

biosphere.

6

3.0 The distribution of oxygen isotopes in meteoric precipitation

The distribution and isotope compositions of meteoric precipitation in Italy

have been extensively described.33,34 These studies show that the Italian peninsula

and its mountain ranges act as barriers to separate the Mediterranean basin

longitudinally into western and eastern zones. The western coast of Italy is

principally affected by precipitation originating from air masses from the west, while

the east coast is also subject to precipitation originating from water masses carried by

south-easterly or dry north-easterly winds from the easternmost regions of the

Mediterranean basin. There seems to be a contribution of the easterly water vapour

masses from the Adriatic Sea over the Po valley, up to the Piedmont area. The

rainwater isotopic composition in Italy shows no latitudinal gradient along the

Tyrrhenian coast. Seasonal isotopic variation is normally significant, though highly

variable throughout the territory, and is likely due to the typical Mediterranean

seasonality pattern, with most rain falling in the cool (winter) season while summers

are hot with strongly evaporative conditions. The relationship between the mean δ18O

and δD shows shifts from the global meteoric water line35 in many of the Italian sites.

These shifts differ in northern, central and southern Italy, with values changing north

to south according to the continental or Mediterranean climatic features of the area.

In line with the trend observed in many Mediterranean regions, mean temperatures

and isotopic values are often poorly correlated, and this is mostly due to the “amount

effect”.36 Importantly for our purposes, however, there is a mean vertical (altitudinal)

gradient, which has been calculated at -0.2‰/100m.

Rainfall isotopic composition at a particular altitude and place, which should carry

the climate signature, however, may not be perfectly reflected in the animal

7

bioapatite. Environmental water that is drunk by animals, for example in streams, can

be a mixture from different altitudes. Moreover, plant water may be strongly affected

by relative humidity due to evapo-transpiration effects. Finally, the “climate signal”

from local waters may be modulated by the drinking behaviour of the herbivore.

Given this complexity, oxygen isotope compositions of fossil remains should be

interpreted on a site-by-site basis along with an assessment of the contribution of

seasonal and the altitudinal effects at each site.

4.0 The Palaeolithic Sites

We analysed teeth from the collections of four Epigravettian archaeological

sites: Grotta Polesini (PL), Grotta di Pozzo (PZ), Grotta di Settecannelle (ST) and

Grotta di Vado all’Arancio (VA). The locations of these sites are presented in Fig.1.

Grotta Polesini is a cave in a travertine deposit at 70 m asl, adjacent to the

present Aniene River between Rome and Tivoli. It was excavated to 5 m below

datum in 1953-1956.37 There is a single radiocarbon date, 10,090±80 BP from level

7, in the middle part of the stratigraphic sequence.38 Both the lithic industry,

identified as Epigravettiano finale, and the faunal remains were extremely abundant.

Red deer predominates, but bovids, equids, caprids, and a variety of carnivores were

also discovered. Human bones are also present.

Grotta di Pozzo, at 710 m asl, is a cave in the limestone of the central

Apennines, 100 km east of Rome.39 It opens onto the Fucino basin, a basin of tectonic

origins surrounded by mountains which reach 1700-1800 m within a few km. The

basin is currently an upland plain, but a shallow lake intermittently developed during

the Upper Pleistocene, extending 150 km2 in historic times before land reclamation.

The archaeological sequence starts with a lithic industry of Epigravettiano antico a

8

cran (undated), and continues with archaeological layers of Late Pleistocene age

(15,790±90 BP; 14,100±70 BP; 12,590±40 BP; 12,320±50 BP, unpublished data).

The fauna, at c. 12,000 BP, is dominated by chamois and red deer, accompanied by

ibex, wild boar, and E. hydruntinus. I. Fiore, personal communication

Grotta di Settecannelle, is located in the tuffs of the volcanic district of Monti

Volsini, 10 km east of Lake Bolsena and 100 km north of Rome at 215 m asl,. The

archaeological sequence starts with level 17 and 16, dated respectively at 16,620±210

and 16,200±200 BP. A lithic industry of the Epigravettiano antico a cran and a

faunal assemblage dominated by E. hydruntinus, E. caballus and Bos primigenius

were found up to level 12. On top of level 12, which is sterile, archaeological levels

10 (12,700±270, 12,500±140 and 12,050±150 BP) and 8 (10,570±260 BP) include

remains of red deer, wild boar and roe deer with lithic industry of the Epigravettiano

finale.40

Grotta di Vado all’Arancio is located in travertine deposits, 15 km from the

modern seashore of Tuscany, at about 300 m asl.41,42 The archaeological evidence is

restricted to a late phase of the Upper Palaeolithic, with a lithic industry of the

Epigravettiano finale dated to 11,330±50 BP. The animal species include bovids,

Equus hydruntinus, Equus caballus, red deer, roe deer and wild boar. Two burials (an

adult male and a child) were also discovered.

Equus hydruntinus and Cervus elaphus were selected for this research

because they were named by Barker6 as the species that migrated seasonally between

lowlands and uplands. They are also extremely suitable for this investigation as they

are found in many sites in central Italy during the Late Pleistocene and because they

have sufficiently large and slow-growing teeth. One specimen of each species was

analysed from the four sites where possible.

9

5.0 Cervid and equid tooth mineralization

In this study the approach was to use transverse serial sampling of enamel along a

tooth crown from occlusal edge to cervix, in order to construct a sequence. This

approach assumes that enamel mineralization is sequential from upper to lower

regions, which is likely only partially correct.

Few studies have investigated the duration and spatial progression of enamel

mineralization, as opposed to tooth eruption, in horses and red deer. Hoppe et al.43

demonstrated that horse tooth enamel continues to mineralize after the tooth has

erupted and that the input period represented by a transverse enamel section is several

months. Kohn44 estimated that enamel maturation continued for 5 months in equids

and for 6 months in large cervids, resulting in an estimated damping factor of 40-50%

in equids and cervids, and, for oxygen isotopes specifically, residence time of the

isotope in the animal can lead to further dampening of 10%.44 Notwithstanding the

dampening effect, several studies have obtained seasonal oxygen isotope variation

from serial sections of enamel e.g., 44-47 that allow palaeoclimate and seasonal variation

to be assessed. In order to control for possible differences in attenuation resulting

from the comparison of teeth with different enamel maturation times, the same tooth

from each species was selected where possible. For Cervus elaphus, M2 molars were

sampled, and for Equus hydruntinus M3 molars. The third mandibular molar crowns

of modern horses commence mineralization at approximately 18 months of age and

take around three years to complete.43 The second mandibular molar crowns of

modern red deer commence mineralization around birth and are complete by the end

of the first year of life.48 Loss of enamel from the cusps through post-eruptive wear,

however, will reduce the length of time represented by individual teeth. The sample

10

chosen were therefore minimally worn. Where some wear has occurred, we estimate

that less that 2 mm of tooth growth was lost.

6.0 Sampling strategy and methodological approach

A longitudinal section of enamel (Fig. 2) was cut from each tooth along the

growth direction (i.e., from cusp to cervix) using a flexible diamond-edged rotary

dental saw. These sections were first sonicated to remove soil residues and then

carefully cleaned of cementum and dentine to a depth of more than 100 µm with a

tungsten carbide burr, to produce a sample of clean core enamel. A sample of dentine

from each tooth was retained for analysis. Sequential enamel subsamples of 2-3 mm

length (sections 1-9 in Fig. 2) were cut transversely from the enamel and then

bisected to produce one sample for δ18O in bioapatite phosphate and one sample for

δ18O and δ13C in bioapatite carbonate, and 87Sr/86Sr. All transverse subsamples were

analysed for δ18O in bioapatite phosphate (δ18Op), whereas only five subsamples per

tooth were tested for 87Sr/86Sr and δ18Oc and δ13C in the carbonate fraction. The

precise number of transverse subsamples obtained was dependent on tooth size:

Equus hydruntinus produced fifteen to thirty-six subsamples and Cervus elaphus six

to ten subsamples. In addition, 87Sr/86Sr and strontium concentrations were obtained

for each dentine sample to provide an indication of diagenetic strontium present at the

burial site.26

7.0 Analytical methods

For phosphate oxygen isotope ratio measurements, the enamel was reduced to a

fine powder, chemically treated to extract PO4 radicals, and precipitated as silver

11

phosphate.49 Approximately 10 to 15 mg of enamel bioapatite was treated with

peroxide at 80°C to eliminate organic material, dissolved in 2M nitrate acid, and

treated with 2 ml KOH and 2 ml HF. The solution obtained was separated by

centrifuge from the CaF2 precipitate, mixed with 15 ml of silver amine solution, and

put over a hot plate at 70°C. Silver phosphate crystals precipitated following a

decrease in pH as NH3 was discharged from the solution.

Analyses of the oxygen isotope ratios were performed in a TC/EA device with

combustion temperature set at 1400° C, coupled to a Delta Plus XL mass

spectrometer. Samples and reference materials (NBS 120c and Aldrich

Hydroxylapatite), also precipitated as silver phosphate, were weighed and sealed into

silver capsules. NBS 120c and Aldrich Hydroxylapatite, after repeated

measurements, were assigned δ18O values of +21.7 and +14.0, respectively. The

reference material NBS 120c, calibrated against certified reference material NBS 127

(assuming δ18O of NBS 127 = +9.3‰ versus SMOW (Ref = IAEA, 2004), has an

accepted value of 21.70‰. Each sample was analysed in triplicate. The

reproducibility of NBS 120c during this set of analyses was 21.71±0.14‰ (1σ,

n=19).

To measure the oxygen and carbon isotope ratios in the carbonate fraction of

biogenic apatite, about 3 mg of untreated powder was reacted with 100% phosphoric

acid at 90°C into an automated VG Isocarb device, coupled to a VG Optima mass

spectrometer. Analytical reproducibility for the laboratory standard calibrated against

NBS standards is 0.07‰ for both δ13C and δ18O. The fractionation factor between the

acid and the biogenic hydroxyapatite was assumed to be the same as that between

acid and calcite. Samples were not pre-treated to eliminate secondary carbonate

because of the very small size of the subsamples, but they were double-checked with

12

treated hydroxyapatite when possible. In these cases, the enamel powder was soaked

in 35% H2O2 for 30 min, then filtered and reacted with 0.1M acetic acid for 10 min,

adapted from the method outlined in Sponheimer and Lee-Thorp.50 The dried powder

was weighed into glass vials and reacted in the above device along with the untreated

samples. No significant differences between the treated and untreated samples were

detectable.

To measure strontium concentrations and isotope ratios, enamel samples were

cleaned ultrasonically in MQ H2O and weighed into Teflon beakers along with a

“spike” of known 84Sr concentration. About 1 ml of Teflon distilled 16N nitric acid

was added to each sample and evaporated to dryness on the hotplate. Ionic Sr was

separated from other ions by Eichrom Sr-spec resin using a separation method

adapted from the method of Horwitz et al.51 and Deniel and Pen.52 Sr isotope

composition and concentration were determined by means of a Finnigan Mat 262

multi-collector mass spectrometer (TIMS) on single rhenium filament and TaF as

activator. The international standard for 87Sr/86Sr, NBS 987, gave a value of

0.710261±10 (2σ, n=8) for static analysis. All strontium ratios have been corrected to

a value for the standard of 0.710250. Blank values were in the region of 150 pg.

8.0 Results

The results obtained from the isotopic assays are listed in Table 1.

8.1 Sr isotope Results

Sr isotopic ratios and elemental compositions are plotted in Fig. 3.

The most striking feature of the enamel dataset is that all 87Sr/86Sr values are

below 0.7092, the value of modern seawater, and are consistent with marine

13

carbonate sediments.32 There is little data in the literature on normal Sr isotope ranges

in co-habiting animal groups, but compared to the range observed in archaeological

human communities inhabiting regions of sedimentary marine carbonates,41 the

isotopic range amongst animals in this study is small (i.e., 0.70861 to 0.70904),

despite the considerable geographic distances between the cave sites (Fig.1) and even

though the geologic formations of these sites are often all marine sediments of

different ages.

The enamel strontium ratios are actually consistent with sedimentary marine

geologic formations, which would include the limestones where some of the

archaeological sites in this study are located. It is possible that the strontium ratios

result from mixing rather than being a direct reflection of a single input source,26

however, given the small isotopic range observed, it is difficult to identify suitable

end-members other than marine carbonates to support this conclusion for these

particular samples. Fig. 3 shows that the ratios and concentrations obtained from four

of the six dentine samples are quite different from the corresponding enamel. Dentine

values are often being used as an indicator of extent of diagenesis, and this is because

dentine is extremely permeable to elements such as strontium after burial, whereas in

vivo levels of strontium in all calcified tissues are relatively low and also relatively

similar. Strontium is distributed quite homogeneously in the skeleton and

concentrations of Sr in skeletal tissues from a single individual are very similar, with

bone and dentine containing slightly more than enamel.53:406, 54:446, 55:202, 56:243

Reported values for modern human skeletal and dental tissues are typically 50-300

ppm.54:445/6, 57:260, 58:175, 59:545. Modern animal tissues exhibit a similar range, although

herbivores tend to have higher concentrations than carnivores because plants are Sr-

14

rich and meat Sr-poor.60:784, 61:661 Nevertheless, it is extremely rare for any

mammalian tissues to exceed 1000 ppm.62:283

Diagenesis in enamel can therefore be discounted from the observation that the

concentration of Sr remains low in this tissue (for example up to a maximum of 434

ppm), while it is much higher (up to 2190 ppm) in the corresponding dentine samples

of the same teeth (sample STCE, Table 1). There is considerable overlap in the

mineralization periods of enamel and primary crown dentine and neither tissue

subsequently remodels or re-grows.63,64 Consequently, in modern individuals one can

expect the Sr concentration and isotope ratio of these two tissues to remain very

similar and this does appear to be the case even in individuals who have subsequently

been exposed to different sources of Sr.64,65 Similarly, archaeological teeth that have

not been buried in soil also show this pattern.26 In teeth excavated from buried

sediments, however, dentine usually contains more Sr than enamel and this increase

is normally coupled with 87Sr/86Sr ratios intermediate between the enamel value and

that of the mobile strontium in the burial soil.26,66,67

In most of our samples, dentine shows an elevated Sr concentration and a

decrease in isotope ratio suggesting this tissue underwent secondary modifications

during burial and the mobile soil strontium was less radiogenic than that preserved in

the enamel. Exceptions to this occurrence are the dentine samples of the two teeth

from G. di Pozzo, which retain low Sr contents. The enamel in these teeth, however,

shows a very consistent intra-tooth variation which is difficult to ascribe to diagenetic

effects. For example, in both teeth (PZCE and PZEH), the isotope ratio of the

sequential sections becomes more radiogenic (i.e., moves away from the dentine

ratio) with increasing strontium concentration, which is the opposite of what would

be predicted if the enamel were subject to post-mortem alteration.

15

The Cervus elaphus tooth from G. Settecannelle shows the highest radiogenic

strontium ratios, consistent with the values of the high-potassium Roman Magmatic

Province that characterizes the area around the site.68 The dentine of the same tooth

lies, however, within the range of the other samples.

8.2 Oxygen isotope results

Sequential phosphate oxygen isotope results (δ18Op) in red deer and horse,

respectively, are plotted in Figs. 4a and 4b, versus the position in the tooth crown.

Higher numbers along the x axis indicate earlier stages of tooth development. Fig. 4a

depicts the δ18Op values measured in Cervus elaphus enamel from G. Polesini, G. di

Pozzo and G. di Settecannelle; the first two specimens show a remarkably similar

trend in terms of pattern (linear) and isotopic values. Both teeth register the transition

between a cold season (lower δ18O values) and a warmer season (higher δ18O). Given

these isotopic patterns, and considering that the teeth are not significantly worn, we

estimate that mineralization took place over about nine months, between winter and

late summer. The widest compositional range is registered in the Cervus elaphus

from G. Polesini with 3.8‰ difference between the cusp and the cervix. For M2s, it is

likely that nursing introduces a further filter to the δ18O composition of tooth enamel

since δ18O of milk is higher in 18O than ambient drinking water by about 1-2‰ (since

it has been "filtered" through the maternal system).69 and references therein This implies that

the lower δ18Op values observed during early M2 mineralization of the two animals

could have been even lower than those registered.

The third deer sample, from G. di Settecannelle, has an apparent opposite trend:

the δ18O recorded values suggest that the mineralization commenced in a warmer

16

season and finished in colder conditions. The seasonal isotopic range is only ~1.5‰,

thus smaller compared to the other deer specimens investigated, while the mean δ18O

is higher (i.e.18O enriched). The tooth in this case shows a higher level of wear

compared to the other deer teeth. However, we estimate that the enamel loss was no

more than a couple of mm, corresponding to one or two lost sub-samples. In this

sample, the weaning effect, if any, would have enhanced the warm season signature,

suggesting that the registered seasonal δ18O variation during mineralization could be

even smaller. Because of the much narrower isotopic variation registered in this tooth,

we suggest that this animal obtained most of its water from a source which changed

little throughout the year, for example, a lake. The strontium isotope data for this

individual also suggest that it lived on a more radiogenic terrain with respect to the

other animals in this study and, as it happens, a large volcanic lake, Lake Bolsena, is

located about 10 km from the G. di Settecannelle.

Fig. 4b shows the intra-tooth δ18Op variation determined in horse teeth from the

sites of G. di Polesini, G. di Pozzo, G. di Settecannelle and G. di Vado all’Arancio.

The profiles observed in these specimens are all sinusoidal, and imply a longer

amelogenesis compared to the red deer specimens. In some cases (ST-EH and VA-

EH), amelogenesis may have been as long as two years or more. The tooth from G. di

Pozzo provides the shortest profile but this tooth crown was broken towards the root,

and we cannot assess the amount of information lost. Unfortunately, this was the only

Equus hydruntinus tooth specimen available from the site. The above sample also has

the lowest δ18Op values among all the teeth and shows the seasonal variation of

almost one year. Such low oxygen isotope values are certainly consistent with the

altitude of the Fucino basin, sited in the Apennine altiplano in Abruzzo. Unlike the

two Cervus elaphus samples, Equus hydruntinus specimens from G. Polesini and G.

17

di Pozzo are clearly distinct in terms of their oxygen isotope composition, with the G.

Polesini sample showing higher δ18Op values, again consistent with the lower altitude

of the site. Seasonal variations registered in the specimens from G. Polesini, G. di

Pozzo and G. di Vado all’Arancio range from 2.5 to 2.9‰, and are of lower

amplitude than the Cervus elaphus from these sites. As with the red deer from

Settecannelle, the horse tooth from the same site shows a reduced seasonal isotopic

amplitude when compared to the other horse teeth (Fig. 4a), with a variation of about

1‰. Although there is no Sr isotope data for this sample, it is of note that both the

horse and red deer samples from Settecannelle display a narrow oxygen isotope

variation of 1.0% and 1.5‰, respectively, which (when translated in water δ18Ow

with the appropriate equation) represents the same isotope variation of 1.4‰ of the

environmental water. This may suggest exploitation of a similar water source. The

offset of about 0.8‰ observed in the mean δ18Ow values of the two animals is likely

due to inter-annual variations as the two animals may easily not have been exactly

coeval.

We interpreted the δ18O values in terms of the modern meteoric values available

in literature. δ18Ow values of the environmental water were calculated from the

δ18Op using the calibration lines built upon deer70 and horse.71 Results are consistent

with the amount weighted mean oxygen values of the pluviometric sites reported in

Longinelli and Selmo.34 For example, the mean weighted value for Piombino, the

nearest pluviometer from G. di Vado all’Arancio, is -6‰, and the mean δ18Ow value

from the horse tooth is -5.9‰; the mean weighted value for Nepi, the nearest

pluviometer from G. Polesini, is -6.3‰, while the horse and red deer teeth from that

site, both give a mean value of -7‰. The mean meteoric value for L’Aquila,

18

comparable to G. di Pozzo, though sited at a much higher elevation, is -7.1‰, while

the horse from that site shows an average value δ18Ow of -7.4‰ and the red deer of -

9.5‰. It must be highlighted in this case that L’Aquila may register very negative

δ18O values (down to -13‰, or less, during the winter due to its peculiar geographic

location).

The nearest pluviometric stations to our sites from which monthly data were

available are: Piombino, Canino, San Gemini (representative of G. Polesini and of the

lower mountains of the Apennines) and L’Aquila. Data from these sites are from

Longinelli and Selmo.34, personal communication We cautiously compared the δ18O meteoric

variations observed within one year with the seasonal variations registered in the

teeth, mindful that annual temperatures in the late Pleistocene are likely to have been

lower and seasonal variability higher, and that we have very few samples. Piombino

shows a seasonal variation of about 6-7‰, Canino of about 4‰, San Gemini of 6-

7‰, and L’Aquila of up to 10-11‰. If we use a conservative dampening effect

during amelogenesis of 50%,44 we can estimate that the expected within-tooth

seasonal variation would vary from 2 to 5‰, according to sites, which is similar to

the range observed in the investigated samples. Consistently, intra-tooth δ18O

variations of the same order of amplitude (2-5 per mil) have also been observed in the

carbonate fractions of prehistoric bovid tooth enamel from a Mediterranean climate

region in South Africa.10

The horse from G. di Pozzo, however, shows relatively small within-tooth δ18Op

variations, but in this case, we take into account that the Fucino basin during the

Upper Palaeolithic was hosting a shallow lake that could have been accessed by the

19

animals roaming the area, thus explaining, like was the case of the animals at

Settecannelle, the lower than expected seasonal variation.

The basic assumption of Barker’s model implies that the animals migrated from

lowlands, where they would spend the cold seasons, to uplands, where they could

find extensive meadows during the summer. These movements may have influenced

the enamel δ18Op composition: if we assume, for example, that the animals moved

seasonally between an elevation of 0 m asl, to an elevation of 2000 m asl, the vertical

isotopic gradient (0.2‰/100m, following Longinelli and Selmo34 would lead to 4‰

difference between the two sites. This altitude effect is quite small, and similar to the

seasonal range observed and predicted. Therefore, we cannot at this stage reliably

separate out the two effects without statistically larger sample sizes.

8.3 Carbon isotope results

All specimens have mean δ13C compositions of approximately -11‰ or lower,

entirely consistent with consumption of C3 vegetation. Within-tooth variation ranges

from 0.4 to 1.4‰, with the exception of a horse from G. di Pozzo where the variation

is 3.3‰. The lower mean value (12.4±1.3) for this particular animal might suggest

cooler and/or higher humidity or moisture conditions in the area it roamed. Carbon

isotope values in all specimens are consistent with the patterns shown by the δ18O

values in terms of inferred seasonality. These variations are most likely ascribed to

isotopic shifts between winter (cold, wet with lower δ18O and δ13C) and summer

(warm, dry with higher δ18O and δ13C) conditions. Seasonal δ13C shifts in plants from

Mediterranean habitats are to be expected as plants generally tend to be more 13C-

enriched in hot and dry conditions, as is the case of Mediterranean summers, and 13C-

depleted in cool wet conditions occurring in winters, as reported in numerous studies.

20

e.g., 72-75 Seasonal within-tooth variations may also be associated with shifts to

different plant sources. Grasses and herbs tend to be slightly depleted in 13C

compared to trees.73,76 In Mediterranean biomes, the availability of various plant

resources differs seasonally, and could also contribute to some of the observed

systematic intra-tooth δ13C variations.

9.0 Discussion

The direct comparison of the δ18O and the 87Sr/86Sr trends within the teeth is

shown in Figs. 5a-f. Given the very small variation in Sr composition shown by all

the investigated specimens, one would be tempted to assert that these animals did not

move over long distances or even did not move at all during tooth mineralization. Sr

compositions, however, do change systematically in all the specimens and these

changes are co-incident with δ18Op trends.

Because samples obtained from sites at some distance from each other all

show surprisingly similar 87Sr/86Sr ratios, we cannot rule out the possibility that these

animals moved within similar geologies. To investigate this matter further, it would

be necessary to collect numbers of environmental samples (including rocks, soils,

vegetation and water) along the possible transhumant routes to determine their Sr

isotope compositions and the real variations in the ecosystem.

Fig. 6 shows the results of oxygen and Sr isotopes in Cervus elaphus and

Equus hydruntinus from G. Polesini and G. di Pozzo. This provides a direct

comparison of the behaviours of these two species. In both sites, red deer seem to

have moved over wider areas compared to horses, which may have roamed areas of

more homogeneous Sr isotope composition.

21

10.0 Conclusion

The application of combined sequential intra-tooth strontium, oxygen and carbon

isotope measurements certainly provides an appropriate method to test the underlying

assumption of long distance migration of herbivores in Barker’s model of hunter-

gatherer transhumance and for similar models.

The results obtained from this study suggest that the two species sampled each

show consistent, but distinct, patterns of movement in the landscape, (Figs. 6a and

6b). Cervus elaphus appears to have been more mobile while Equus hydruntinus was

more residential. While not conclusive, our results do not show strong evidence for

altitudinal seasonal mobility by either the equids or cervids. Our results, however, are

confined to the movements of animals and say little about the possible movements of

the hunter-gatherers other than that they were unlikely to have been in response to

(altitudinally) migrating herds. We expect applications of these methods will advance

understanding of migratory behaviour of extinct and extant species and establish the

origins of long distance pastoral transhumance in Mediterranean Europe.

Since the Italian territory appears more homogeneous than expected in terms

of Sr isotopes, results may also indicate that the investigated animals moved over

landscapes characterised by similar geology or Sr isotope composition. In order to

strengthen the basis of our study, we would need to increase the environmental

information from the investigated territory, to test for the isotopic composition of

biologically available Sr, and compare results from transhumant animals from

modern contexts with those from stadial and interstadial layers, to asses if migratory

habits changed with climate change.

11.0 Acknowledgments

22

The authors would like to thank Hillary Sloane, Claudio Sorrentino, Fabio Martini,

Ivana Fiore, Zelia De Giuseppe and Holly Crawford for their assistance. This project

was funded by the Hunter-Gatherer Research Laboratory of the University of

Bradford.

12.0 References

1. Blockley SPE, Lowe JJ, Walker MJC, Asioli A, Trincardi F, Pollard AM,

Donahue RE, Coope GR. J. Quat. Sci. 2004; 19: 159.

2. Prasad S, Brauer A, Rein B, Negendank JFW. Holocene 2006; 16: 153.

3. Higgs ES, Vita-Finzi C, Harris DR, Fagg AE. Proc. Prehist. Soc. 1967; 33: 1.

4. Higgs ES, Jarman MR. Palaeoeconomy. Higgs ES (ed). Cambridge University

Press: Cambridge, 1975; 1.

5. Barker GWW. Palaeoeconomy. Higgs ES (ed). Cambridge University Press:

Cambridge, 1975; 111.

6. Barker GWW, Landscape and society: Prehistoric Central Italy. Academic

Press, London, 1981.

7. Tieszen LL. J. of Archeol. Sci. 1991; 18: 227.

8. Luz B, Kolodny Y. Earth and Planetary Sci. Letters 1985; 75: 29.

9. Fricke HC, O'Neil JR. Palaeogeogr., Palaeoclimatol., Palaeoecol. 1996; 126:

91.

10. Balasse M, Ambrose SH, Smith AB, Price TD. J. Archeol. Sci. 2002; 29: 917.

11. Bentley RA, Knipper C. Antiquity 2005; 79.

12. Hoppe KA, Koch PL, Carlson RW, Webb SD. Geology 1999; 27: 439.

13. Sykes NJ, White J, Hayes TE, Palmer MR. Antiquity 2006; 80: 948.

23

14. Montgomery J, Cooper RE, Evans JA. From Stonehenge to the Baltic. Living

with cultural diversity in the third millennium BC. Larsson M, Parker Pearson

M (eds). International Series BAR 1692, Archaeopress, Oxford 2007; 65.

15. Richards MP, Harvati K, Grimes V, Smith C, Smith T, Hublin J-J, Karkanas

P, Panagopoulou E. J. Archeol. Sci. in press.

16. Lemoine M. Geological atlas of alpine Europe and adjoining alpine areas.

Lemoine M (ed). Elsevier: New York, 1978.

17. Pfiffner OA. The Pre-Mesozoic Geology in the Alps. V RJF, Neubauer F (ed).

Springer Verlag: Berlin. 1993.

18. Åberg G. Water, Air and Soil Pollution 1995; 79: 309.

19. Bau M, Alexander B, Chesley JT, Dulski P, Brantley SL. Geochim. et

Cosmochim. Acta 2004; 68: 1199.

20. Blum JD, Erel Y, Brown K. Geochim. et Cosmochim. Acta 1993; 57: 5019.

21. Evans JA, Tatham S. Forensic Geoscience: Principles Techniques and

Applications. Pye K, Croft DJ (eds). Geological society of London Special

Publication: London, 2004.

22. Gallet S, Jahn BM, Torii M. Chem. Geol. 1996; 133: 67.

23. Jacobson AD, Blum JD. Geology 2000; 28: 463.

24. Sillen A, Hall G, Richardson S, Armstrong R. Geochim. et Cosmochim. Acta

1998; 62: 2463.

25. Bentley RA. J. Archeol. Method and Theory 2006; 13: 135.

26. Montgomery J, Evans JA, Cooper RE. Appl. Geochem. 2007; 22: 1502.

27. Montgomery J, Evans JA, Wildman G. Appl. Geochem. 2006; 21: 1626.

28. Muller W, Fricke H, C., Halliday AN, McCulloch MT, Wartho JA. Science

2003; 862.

24

29. Serri G, Innocenti F, Manetti P. Tectonophysics 1993; 223: 117.

30. Palombo MR, Filippi ML, Iacumin P, Longinelli A, Barbieri M, Maras A.

Quat. Int. 2005; 126-128: 153.

31. Hoppe KA, Koch PL, Furutani TT. Int. J. Osteoarchaeol. 2003; 13: 20.

32. McArthur JM, Howarth RJ, Bailey TR. J. Geology 2001; 109: 155.

33. Longinelli A, Anglesio E, Flora O, Iacumin P, Selmo E. J. Hydrol. 2006; 329:

471.

34. Longinelli A, Selmo E. J. Hydrol. 2003; 270: 75.

35. Craig H. Science 1961; 133: 1702.

36. Rozanski K, Araguas-Araguas L, Gonfiantini R. Climate change in

Continental Isotopic Records. Swart PK, Lohmann KC, MacKenzie J (eds).

1993; 1.

37. Radmilli AM. Gli scavi nella Grotta Polesini a Ponte Lucano di Tivoli e la più

antica arte nel Lazio. Sansoni: Firenze, 1974.

38. Belluomini G. Geografia Fisica e Dinamica Quaternaria 1980; 3: 25.

39. Mussi M, Melis RT, Mazzella G. Atti della XXXVI° Riunione scientifica

dell'Istituto italiano di Preistoria e Protostoria 2003: 65.

40. Ucelli Gnesutta P. XXII Valcamonica Symposium 2007: 193.

41. Minellono F. Rivista di Scienze Preistoriche 1985-1986; 40: 115.

42. Minellono F, Pardini E, Fornaciari G. Rivista di Scienze Preisoriche 1980; 35:

3.

43. Hoppe KA, Stover SM, Pascoe JR, Amundson R. Palaeogeogr.,

Palaeoclimatol., Palaeoecol. 2004; 206: 355.

44. Kohn MJ. Geochim. et Cosmochim. Acta. 2004; 68: 403.

25

45. Fricke H, C., Clyde WC, O'Neil JR. Geochim. et Cosmochim. Acta 1998; 62:

1839.

46. Kohn MJ, Schoeninger MJ, Valley JW. Chem. Geology 1998; 152: 97.

47. Zazzo A, Mariotti A, Lecuyer C, Heintz E. Paleogeogr., Paleoclimatol.,

Paleoecol. 2002; 186: 145.

48. Brown JD, Chapman HG. J. Zoology 1991; 224: 85.

49. O'Neil JR, Roe LJ, Reinhard E, Blake RE. Israel J. of Earth Sci. 1994; 43:

203.

50. Sponheimer M, Lee-Thorp J, A. Handbook of Paleoanthropology. Henke W,

Rothe H, Tattersall I (eds). Springer Verlag: Berlin, 2007; 555.

51. Horwitz EP, Chiarizia R, Dietz ML. Solvent Extraction and Ion Exchange

1992; 10: 313.

52. Deniel C, Pin C. Analytica Chimica Acta 2001; 426: 95.

53. Turekian KK, Kulp JL. Science 1956; 124: 405.

54. Underwood EJ. Trace elements in human and animal nutrition. Academic

Press: London, 1977.

55. Parker RB, Toots H. Fossils in the Making: Vertebrate Taphonomy and

Paleoecology. Behrensmeyer AK, Hill AP (eds). University of Chicago Press:

London, 1980; 197.

56. Aufderheide AC. Reconstruction of life from the skeleton. Iscan MY, Kennedy

KAR (eds). Alan R. Liss, Inc.: New York, 1989.

57. Brudevold F, Söremark R. Structural and Chemical Organization of Teeth.

Miles AEW (ed).Academic Press: London, 1967; 247.

58. Hancock RGV, Grynpas MD, Pritzker KPH. Archaeometry 1989; 31: 169.

26

59. Elliott TA, Grime GW. Nuclear Instruments and Methods in Physics

Research Section B: Beam Interactions with Materials and Atoms 1993; 77:

537.

60. Bocherens H, Brinkman DB, Dauphin Y, Mariotti A. Canadian J. of Earth

Sci. 1994; 31: 783.

61. Tuross N, Behrensmeyer AK, Eanes ED. J. Archeol. Sci. 1989; 16: 661.

62. Radosevich SC. Investigations of Ancient Human Tissue: Chemical Analyses

in Anthropology. Sandford MK (ed). Gordon and Breach: Amsterdam, 1993;

269.

63. Tsalev DL, Atomic absorption spectrometry in occupational and

environmental health practice. (CRC Press Inc, Drexel University, 1984).

64. Boyde A. Handbook of microscopic anatomy: teeth. Berkovitz BKB (ed).

Springer Verlag: Berlin, 1989.

65. Montgomery J, University of Bradford (2002).

66. Budd P, Montgomery J, Barreiro B, Thomas RG. Appl. Geochem. 2000; 15:

687.

67. Trickett MA, Budd P, Montgomery J, Evans JA. Appl. Geochem. 2003; 18:

653.

68. Conticelli S, D'Antonio M, Pinarelli L, Civetta L. Mineral. and Petrol. 2002;

74: 189.

69. White C, Longstaffe FJ, Law KR. J. Archaeol. Sci. 2004; 31: 233.

70. D'Angela D, Longinelli A. Chem. Geol. 1990; 86: 75.

71. Delgado-Huertas A, Iacumin P, Stenni B, Chillon BS, Longinelli A. Geochim.

et Cosmochim. Acta 1995; 59: 4299.

27

72. Brandes EN, Kodama K, Whittaker C, Weston H, Rennenberg C, Keitel M,

Adams A, Gessler A. Global Change Biol. 2006; 12: 1922.

73. Brugnoli E, Farquhar GE. The Netherlands Leegod RC, Sharkey TD, von

Caemmerer S (eds). Kluwer Academic: New York, 2000; 399.

74. Helle G, Schleser GH. Plant and Cell Environment 2004; 27: 367.

75. Lauteri M, Brugnoli E, Spaccino L. Stable isotopes and plant carbon-water

relations. Ehleringer JR, Hall AE, Farquhar GE (eds). Academic Press:

London, 1993; 93.

76. Lloyd J, Syvertsen LJ, Kriedemann PE, Farquhar GD. Plant, Cell &

Environment 1992; 15: 873.

Table 1. Results of sequential O and Sr isotopic analyses of the investigated specimens. The δ18Ow of the environmental water was calculated according to the calibration line proposed by D’Angela and Longinelli70 for deer, and by Delgado-Huertas et al.71 for horses. (δ18Op= phosphate oxygen; δ18Cca= bioapatite carbon; M= molar; P=premolar).

Figure 1. Schematic geological map of central Italy. Different shades of grey

represent different formations in terms of age and potential Sr isotopic variation. Black dots indicate the location of the various archaeological sites.

Figure 2. Dental enamel sampling strategy of the incremental sub-samples. Sub-

samples were then divided in two parts for the various isotope determinations. Figure 3. The Sr isotopic composition are plotted against the Sr content (ppm) for the

five specimens. The different symbol shapes represent different animals. White symbols represent enamel samples and grey symbols represent the dentine sample from the same tooth.

Figure 4. Within-tooth changes in δ18Op along the growing direction: a) Cervus

elaphus tooth specimens from G. Polesini, G. Pozzo and G. Settecannelle; b) Equus hydruntinus specimens from G. Polesini, G. Pozzo, G. Settecannelle and G. Vado all’Arancio. Higher numbers in the abscissa represent earlier stages of tooth development.

28

Figure 5. Plots (a-f) show intra-tooth isotope values of 87Sr/86Sr (black symbols) and δ18O (open symbols). Oxygen values, which provide the seasonal climatic variations, are followed by consistent, though small, Sr isotope variations in almost all the animals. This could suggest animal mobility throughout the year, either within limited areas or across similar geologies. Further study is required.

Figure 6. Comparison of Cervus elaphus (open symbols) versus Equus hydruntinus

(black symbols) oxygen and Sr isotopic patterns at G. Polesini (a) and G. Pozzo (b). In both sites the migratory behaviour of the two animals seems to differ, with deer more likely to move over wider areas compared to horses.

Table 1

29

Site Species Item Sub sample

Distance from root

(mm) δ18Op (V-

SMOW)

δ13Cca (V-PDB1)

87Sr/86Sr Sr (ppm)

δ18Ow (V-SMOW)

Grotta Polesini Cervus elaphus M2 PLCE 1 2.0 19.9 -5.0

" " " PLCE 2 3.9 19.6 -10.2 0.708613 302 -5.3 " " " PLCE 3 6.0 18.6 -6.2 " " " PLCE 4 8.5 18.1 -10.8 0.708730 328 -6.6 " " " PLCE 5 11.0 17.0 -7.6 " " " PLCE 6 13.0 16.8 -11.0 0.708837 365 -7.7 " " " PLCE 7 15.1 16.6 -11.4 0.708756 377 -8.0 " " " PLCE 8 17.1 16.3 -8.2

" " " PLCE 9 19.4 16.0 -10.9 0.708786 402 -8.4

" " " dentine - 0.708325 1117

Grotta Polesini Equus hydruntinus M or P PLEH 1 3.8 17.4 -7.4

" " " PLEH 2 8.8 17.6 -11.2 0.708788 718 -7.1 " " " PLEH 3 12.0 18.5 -5.8 " " " PLEH 4 14.9 18.5 -5.7 " " " PLEH 5 19.0 18.2 -6.3 " " " PLEH 6 23.5 17.4 -7.4 " " " PLEH 7 28.0 17.1 -11.3 0.708734 775 -7.7 " " " PLEH 8 31.3 16.0 -9.3 " " " PLEH 9 34.7 16.5 -8.5 " " " PLEH 10 37.5 16.8 -8.2 " " " PLEH 11 39.8 17.0 -8.0 " " " PLEH 12 42.0 17.3 -10.8 0.708686 736 -7.5 " " " PLEH 13 44.1 17.3 -7.5 " " " PLEH 14 47.0 17.2 -7.6 " " " PLEH 15 49.1 17.2 -7.6 " " " PLEH 16 52.0 17.9 -6.6 " " " PLEH 17 54.0 18.3 -9.9 0.708743 746 -6.1 " " " PLEH 18 56.3 18.4 -5.9 " " " PLEH 19 58.3 18.9 -5.3 " " " PLEH 20 60.4 17.8 -6.7 " " " PLEH 21 62.8 18.9 -10.8 0.708753 740 -5.2 " " " PLEH 22 65.7 18.4 -6.0

" " " dentine - - 0.708365 1308

Grotta di Pozzo Cervus elaphus M2 PZCE 1 2.9 18.8 -5.9

" " " PZCE 2 4.7 18.8 -10.6 0.708710 197 -6.0 " " " PZCE 3 6.8 18.3 -6.4 " " " PZCE 4 8.5 17.9 -10.8 0.708746 261 -6.7 " " " PZCE 5 10.4 17.3 -7.3 " " " PZCE 6 12.2 16.9 -10.7 0.708759 344 -7.7 " " " PZCE 7 14.6 16.4 -8.1 " " " PZCE 8 16.7 16.1 -10.6 0.708784 414 -8.4 " " " PZCE 9 18.5 15.6 -8.8 " " " PZCE 10 21.0 15.3 -10.4 0.708828 475 -9.1 " " " dentine - - 0.708757 234

30

Site Species Item Sub sample

Distance from root

(mm) δ18Op (V-

SMOW)

δ13Cca (V-PDB1)

87Sr/86Sr Sr (ppm)

δ18Ow (V-SMOW)

Grotta di Pozzo Equus hydruntinus M or P PZEH 1 19.0 15.5 -10.0

" " " PZEH 2 22.2 15.9 -12.1 0.708750 195 -9.4 " " " PZEH 3 25.5 14.9 -10.9 " " " PZEH 4 29.0 14.9 -10.9 " " " PZEH 5 31.6 15.7 -13.5 0.708761 197 -9.8 " " " PZEH 6 33.9 15.5 -10.0 " " " PZEH 7 36.4 15.5 -10.0 " " " PZEH 8 39.5 15.4 -14.0 0.708793 202 -10.2 " " " PZEH 9 42.1 15.7 -9.7 " " " PZEH 10 45.0 16.5 -8.6 " " " PZEH 11 47.9 17.1 -11.8 0.708788 200 -7.7 " " " PZEH 12 50.9 17.3 -7.5 " " " PZEH 13 53.6 16.6 -8.4 " " " PZEH 14 56.7 16.2 -10.7 0.708820 201 -9.0 " " " PZEH 15 60.9 15.5 -10.0 " " " dentine - - 0.708670 213

Grotta Settecannelle

Cervus elaphus M2 STCE 1 3.2 18.2 - - -6.5

" " " STCE 2 5.2 17.9 -11.6 0.708902 422 -6.8 " " " STCE 3 7.8 18.1 -11.6 0.708946 434 -6.6 " " " STCE 4 10.5 18.6 -11.7 0.708989 430 -6.2 " " " STCE 5 12.5 18.8 -11.6 0.708913 422 -5.9 " " " STCE 6 15.2 19.4 -11.3 0.709038 409 -5.4 " " " dentine - - 0.708719 2190

Grotta Settecannelle

Equus hydruntinus M3 STEH 1 4.5 18.3 - - - -6.0

" " " STEH 2 12.0 18.4 - - - -5.9 " " " STEH 3 15.8 18.6 - - - -5.6 " " " STEH 4 20.1 18.6 - - - -5.6 " " " STEH 5 25.5 19.0 - - - -5.0 " " " STEH 6 30.0 19.1 - - - -4.9 " " " STEH 7 35.6 18.4 - - - -6.0 " " " STEH 8 40.9 18.4 - - - -5.9 " " " STEH 9 45.2 18.7 - - - -5.5 " " " STEH 10 50.0 19.2 - - - -4.8 " " " STEH 11 53.2 19.3 - - - -4.6 " " " STEH 12 57.0 18.5 - - - -5.7 " " " dentine - - 0.708962 1353

31

Site Species Item Sub sample

Distance from root

(mm) δ18Op (V-

SMOW)

δ13Cca (V-PDB1)

87Sr/86Sr Sr (ppm)

δ18Ow (V-SMOW)

Vado all'Arancio

Equus hydruntinus M or P VAEH 1 2.2 19.3 - - - -4.6

" " " VAEH 2 4.0 18.2 -6.2 " " " VAEH 3 6.6 17.1 -7.8 " " " VAEH 4 8.8 16.9 -9.7 0.708845 651 -8.0 " " " VAEH 5 11.0 16.8 -8.1 " " " VAEH 6 12.8 17.8 -6.8 " " " VAEH 7 15.3 18.5 -5.8 " " " VAEH 8 18.3 18.8 -5.3 " " " VAEH 9 21.2 18.8 -5.4 " " " VAEH 10 23.7 18.7 -5.5 " " " VAEH 11 25.9 18.7 -11.1 0.708844 590 -5.5 " " " VAEH 12 28.1 18.7 -5.5 " " " VAEH 13 29.9 18.8 -5.3 " " " VAEH 14 32.1 19.0 -5.1 " " " VAEH 15 34.7 19.1 -4.9 " " " VAEH 16 36.5 19.4 -4.5 " " " VAEH 17 39.4 19.5 -4.3 " " " VAEH 18 42.0 19.4 -11.0 0.708896 612 -4.5 " " " VAEH 19 44.2 19.1 -4.9 " " " VAEH 20 46.7 18.9 -5.3 " " " VAEH 21 49.6 18.6 -5.6 " " " VAEH 22 51.8 18.2 -6.1 " " " VAEH 23 54.0 17.7 -6.9 " " " VAEH 24 56.2 17.7 -6.9 " " " VAEH 25 58.8 17.7 -6.9 " " " VAEH 26 61.3 17.8 -6.8 " " " VAEH 27 63.5 17.9 -6.7 " " " VAEH 28 66.4 18.4 -10.3 0.708661 658 -5.9 " " " VAEH 29 69.0 18.3 -6.0 " " " VAEH 30 71.2 18.6 -5.7 " " " VAEH 31 73.0 18.8 -5.3 " " " VAEH 32 75.2 18.8 -5.4 " " " VAEH 33 77.0 18.7 -5.5

" " " VAEH 34 78.8 18.1 -6.4

" " " VAEH 35 80.7 17.7 -10.9 0.708735 687 -6.9 " " " VAEH 36 82.9 17.2 -7.6 " " " dentine - - 0.708613 902

32

Figure 1

Figure 2

33

Figure 3

Figure 4

34

Figure 5

35

Figure 6

36