Op por tu ni ties and con straints for re con struct ing palaeoenvironments from sta ble light iso tope ra tios in fos sils
Julia A. LEE-THORP and Matt SPONHEIMER
Lee-Thorp J. A. and Sponheimer M. (2005) — Op por tu ni ties and con straints for re con struct ing palaeoenvironments from sta ble light iso tope ra tios in fos sils. Geol. Quart., 49 (2): 195–204. Warszawa.
Sta ble light iso tope ra tios (13C/12Cand 18O/16O) in fos sil teeth pro vide key ar chives for un der stand ing ecol ogy of past fau nal com mu ni ties and the evo lu tion of en vi ron ments dur ing the Plio-Pleis to cene. Given the in ev i ta ble pro cesses of diagenesis dur ing fos sili sa tion, the in - teg rity of iso to pic in for ma tion and the de gree of de tailed in for ma tion that can be ex tracted, re main im por tant is sues in all fos sil stud ies.
The most ap pro pri ate tests are those in trin sic to iso to pic abun dances in eco sys tems. They are eas ier to de velop for 13C/12C in sa vanna en - vi ron ments where large 13C/12Cdif fer ences ex ist be tween C4 trop i cal grasses and C3 trees and shrubs. Val i dat ing 18O/16O ra tios in fos sil car bon ate or phos phate is more dif fi cult, but pat terned vari abil ity, mainly track ing wa ter-re lated be hav iour, within mod ern fau nal com - mu ni ties has been rep li cated in sev eral fos sil as sem blages. The iden ti fi ca tion of sea sonal vari a tion in 13C/12C and 18O/16O along the growth axis of a tooth crown, also ap pli ca ble in ar eas com posed solely of C3 plants, fills a dual role as a test and for pro vid ing data on sea - sonal am pli tude. The re sults of stud ies from low- and mid-lat i tude Af ri can sites sug gest that iso to pic vari a tion in rain fall on short timescales and eco log i cal dif fer ences amongst an i mals, dom i nate over smaller dif fer ences in 18O/16O com po si tion due to tem per a ture.
Julia A. Lee-Thorp, De part ment of Ar chae o log i cal Sci ences, Uni ver sity of Brad ford, Brad ford, UK-BD7 1DP, e-mail:
j.a.lee-thorp@brad ford.ac.uk; Matt Sponheimer, De part ment of An thro pol ogy, Uni ver sity of Col o rado, Boul der, USA, e-mail:
msponheimer@ya hoo.com (re ceived: De cem ber 23, 2004; ac cepted: April 11, 2005).
Key words: sta ble light iso topes, fos sils, ver te brates, diagenesis, en vi ron men tal re con struc tion.
INTRODUCTION
Un der ap pro pri ate con di tions the min eral com po nent of ver te brate bones and teeth sur vives for mil lions of years. Fos - sil ised bones and teeth pre serve in for ma tion about many of the pro cesses and con di tions to which that an i mal was sub ject dur - ing its life time, and which can be ac cessed via chem i cal in di ca - tors, most no ta bly, the sta ble light iso tope com po si tion. One ma jor con straint is that the ef fects of al ter ations in duced by de - cay and fos sili sa tion pro cesses on these bio chem i cal in di ca tors must not ob scure the orig i nal com po si tion, at least, not to the ex tent that it can no lon ger be in ter preted. The ex tent to which these nat u ral pro cesses might al ter biogenic isotopic compositions is a source of long-standing con tro versy.
Re cently it has be come ap par ent that iso to pic deg ra da tion does not nec es sar ily fol low physico-chem i cal al ter ation (Stu -
art-Wil liams et al., 1996; Lee-Thorp and Sponheimer, 2003).
In other words, pro cesses such as recrystallisation that are, af ter all, bound to oc cur to some de gree dur ing fos sili sa tion, do not in ev i ta bly lead to poor iso to pic in teg rity. The iso to pic data from fos sils are of ten sur pris ingly ro bust. There are lim its to this gen er ali sa tion, how ever, some iso to pic ap pli ca tions (most no ta bly in palaeoclimatology) re quire a great deal of pre ci sion for mean ing ful in ter pre ta tion. The ques tion then be comes what are the lim its im posed by diagenesis and nat u ral vari abil ity in those cases re quir ing pre ci sion in iso to pic data? In or der to as - sess this, it is im por tant to un der stand (1) the con straints im - posed by the na ture of bi o log i cal min er als on fos sili sa tion path - ways and (2) nat u ral vari abil ity in eco sys tems, both past and pres ent. In this pa per, we as sess cur rent un der stand ing of fos sil cal ci fied tis sue chem is try in the light of these prob lems and go on to ad dress the lim i ta tions and ad van tages that nat u ral vari - abil ity im poses on our in ter pre ta tions of past en vi ron ments.
NATURE OF BIOLOGICAL MINERALS AND DIAGENESIS
Bones and teeth can sur vive for so long pre cisely be cause the min eral com po nent of these tis sues — a se ries of cal cium phos phate-based min er als called bi o log i cal apatites (or bioapatites) — is sta bi lised by post mor tem chem i cal changes.
These changes in te gral to the fos sili sa tion pro cess, al ter the min eral struc ture im me di ately post mor tem and there af ter; al - ter ation or diagenesis, oc curs whether or not burial takes place (Tu ross et al., 1989; Per son et al., 1995).
The min er als in mam ma lian skel e tal tis sues are col lec tively known as bi o log i cal apatites: hex ag o nal cal cium phos phates, re sem bling (but not iden ti cal to) hydroxyapatite [Ca10(PO4)6(OH)2] in struc ture. Dif fer ences com pared to syn - thetic or geo log i cally oc cur ring apatites in clude non-stoichio - metry, high iso mor phic sub sti tu tion and ad sorp tion, lat tice dis - tor tions, and small crys tal size (LeGeros, 1991). Bi o log i cal apatites are a con tin uum of struc tures dif fer ing in kind and de - gree of sub sti tu tions, func tion and crystallinity1. All of these fac tors af fect the prop er ties and func tion of the min eral in some de gree, both in life and in death.
The lo ca tion of phos phate within the apatitic struc ture is well-un der stood, but car bon ate ions may oc cur in sev eral lo ca - tions, both as ad sorbed ions on crys tal sur faces, and within the unit cell sub sti tuted mainly in the phos phate po si tion (some - times called struc tural or “B” car bon ate) and to a lesser ex tent in the hydroxyl po si tion (“A” car bon ate) (Rey et al., 1991).
Most CO32-is sub sti tuted for PO43-, (usu ally in com bi na tion with Na+ sub sti tuted for Ca2+ to pre serve net charge), while mi - nor amounts of HPO3-, car bon ates and bi car bon ates are ad - sorbed on hydration lay ers in bone ap a tite.
As a re sult of en hanced sur face area/mass ra tios due to small crys tal size, re ac tive sur faces and ad sorbed ions, and high num ber of sub sti tu tions that raise sol u bil ity, bone ap a tite is highly re ac tive (Driessens et al., 1978; LeGeros, 1991).
Enamel ap a tite dif fers from bone and dentine: it has fewer sub - sti tu tions, less dis tor tion, greater long-range or der and crys tals are about an or der of mag ni tude larger (LeGeros, 1991). About half the amount of sub sti tuted CO32-(~3%) is pres ent com pared to bone ap a tite (~6%).
Other dif fer ences are in higher-or der struc tures and in the or ganic ma trix. De po si tion and ori en ta tion of bone and dentine ap a tite is reg u lated by col la gen fi bril pe ri od ic ity, whereas enamel forms as rods on tem plates of tu bules (Boyde, 1967).
The tu bu lar or ganic ma trix, made up of phos pho pro teins and amelogenins, de creases dur ing enamel mat u ra tion to neg li gi ble amounts (<1%), whereas the pro por tion of col la gen re mains high (~20–30%) in bone and dentine. For ma tion and turn over times dif fer; bone is con stantly re mod elled dur ing life whereas enamel forms dur ing a (rel a tively) dis crete, early pe riod in the in di vid ual’s life. Hence iso to pic ra tios in phos phate or car bon - ate com part ments of bone ap a tite re flect con di tions in te grated over many years, whereas those in enamel re flect con di tions at the time of min er ali sa tion of a par tic u lar tooth. It is im por tant to note, how ever, that enamel ma tures over sev eral months — the
first phase of rel a tively rapid min er ali sa tion of tu bules is fol - lowed by a lon ger pe riod of mat u ra tion of sev eral months du ra - tion (Boyde, 1967; Balasse, 2003). There fore, iso tope val ues ob tained from even ex tremely tiny sam pling in cre ments can not re flect dis crete, short ep i sodes, no mat ter how small the sam - pling in ter vals or at tempts to sam ple along enamel prisms (for a re view see Balasse, 2003). The iso to pic sig nal is in ev i ta bly mixed and damp ened, al though ef forts are cur rently un der way to dis en tan gle the smear ing by ap pli ca tion of math e mat i cal al - go rithms to, ini tially, enamel formed un der con trolled con di - tions (Passey and Cerling, 2002; Passey et al., 2003).
It is widely re cog nised that enamel pre serves both chem i cal struc ture and iso to pic com po si tion far better than bone min eral (Lee-Thorp and van der Merwe, 1987), a phe nom e non that can be as cribed to its higher crystallinity and less po rous micro - structure. In liv ing bone, the min eral is ac tively main tained in a paracrystalline state, be cause one of its key func tions is to pro - vide a res er voir of Ca2+ and CO32-ions readily ac ces si ble to the blood sup ply. In the ab sence of these in vivo mol e cules and in the pres ence of a ready sup ply of the raw ma te ri als for growth (prin ci pally Ca2+ and PO43-ions), bone ap a tite will “grow” by pro cesses of recrystallisation and Ostwald “rip en ing” (scav - eng ing of smaller by larger crys tal lites; Eanes and Posner, 1970). Es sen tially the tra jec tory is to wards sta bili sa tion and greater crystallinity (Tu ross et al., 1989; Per son et al., 1995;
Sponheimer and Lee-Thorp, 1999a). In the pro cess num bers of for eign ions com pat i ble with the ap a tite struc ture may be added or ex changed from the sur round ings. Con trary to ear lier be liefs by phos phate prac ti tio ners (e.g. Longinelli, 1984; Luz and Kolodny, 1985), these pro cesses can in cor po rate new PO43-ions. Enamel is al ready rel a tively sta ble and es sen tially pre con di tioned, there fore pos si bil i ties for recry stallisation and crys tal growth are mini mised. Ionic or iso to pic ex change pro - cesses, how ever, can con tinue in both tis sues.
These prop er ties need also to be con sid ered in the chem i cal pre treat ment meth ods com monly used to elim i nate in tru sive ma - te rial or to mini mise al ter ation in fos sils. For in stance, use of too strong acid so lu tions or too lengthy im mer sions in even weaker acid so lu tions, will in duce recrystallisation and iso to pic ef fects (Lee-Thorp and van der Merwe, 1991; Koch et al., 1997;
Sponheimer, 1999; Balasse et al., 2002). The path ways of al ter - ation — ad di tion of ex tra ne ous ma te rial, recrystalli sation and rip en ing — are com mon to both phos phate- and car bon ate-based stud ies. The main ex cep tion is that ox y gen bound to phos phate is not eas ily ex change able, whereas ox y gen bound to car bon ate is.
How ever, the phos phate bond is not im mune to post-mor tem at - tack by en zy mat i cally-catalysed mi cro bial at tack (Blake et al., 2001) and al ter ation of iso to pic ra tios in biogenic phos phate has been dem on strated (Sharp et al., 2000). There seems to be lit tle ev i dence from Fou rier Trans form In fra red stud ies that car bon ate it self is any more or less mo bile that phos phate (Sponheimer and Lee-Thorp, 1999a; Lee-Thorp and Sponheimer, 2003); both ions un dergo sub tle re or gani sa tion dur ing fos sili sa tion. In spite of these sub tle changes, how ever, con sis tent off sets be tween
18O/16O com po si tion in these two ions sug gest that iso to pic com - po si tions are ro bust (Bryant et al., 1996; Iacumin et al., 1996;
Sharp et al., 2000).
Hav ing es tab lished the frame work im posed by the na ture of the min er als, we can pro ceed to con sider some of the lim its and
1Crystallinity de notes both crys tal size and a lack of de fects or dis tor tions.
pos si bil i ties of car bon and ox y gen iso tope-based tools of fer for palaeo ec ol ogy and palaeoenvironmental re con struc tion.
CARBON ISOTOPE ABUNDANCES IN FOSSIL FOODWEBS
The car bon ate sub sti tuted within bi o log i cal apatites is de - rived from blood bi car bon ate, which is in turn de rived from di - etary car bon. The ex pected en rich ment in d13C of about 10‰ in the sys tem CO2-HCO3--CO32- at equi lib rium is matched by iso to pic dif fer ences of ca. 10–14‰ be tween di etary car bon and ap a tite car bon ate2 (Lee-Thorp et al., 1989; Cerling and Har ris, 1999). Early scep ti cism about the use of min eral (e.g.
Schoeninger and DeNiro, 1982) has been largely over come es - pe cially in the case of enamel (for a re view see Lee-Thorp, 2000). A cru cial dem on stra tion showed that ex pected dif fer - ences in d13C of brows ers and graz ers eat ing iso to pi cally dis - tinct C3 and C4 plants were main tained for mil lions of years (Lee-Thorp and van der Merwe, 1987). Typ i cal val ues for mod ern brows ers and graz ers are mod i fied slightly in fos sils; a small pos i tive d13C shift of ~2–3‰ is partly at trib ut able to diagenesis and partly to the “fos sil fuel ef fect” that has seen d13C of en tire mod ern eco sys tems de crease. None the less, val - ues for fos sil fauna are clearly iden ti fi able to C3, C4 or mixed di ets, even at mil lions of years re move.
This find ing opened the way for a pleth ora of palaeodietary and palaeoenvironmental ap pli ca tions rang ing widely in age.
The most straight for ward ap pli ca tion is in biomes where the large dis tinc tion be tween C3 and C4 feed ers pro vides both the means to in ter ro gate the en vi ron ment, and an in trin sic test for re li abil ity. Other kinds of biomes are also ame na ble, but pub - lished ap pli ca tions are fewer.
GLOBAL C4 GRASS EXPANSION
One im por tant ap pli ca tion to in ves ti gate veg e ta tion and cli - mate shifts in older ma te rial, is to the ma jor global ex pan sion of C4 grass sys tems be tween ~8 and 5 mil lion years ago (Cerling et al., 1997). This phe nom e non had been first de tected in soil iso tope stud ies (Quade et al., 1989; Cerling, 1993) but the pat - terns are clearer and less am big u ous in fos sil fauna, al low ing de tailed track ing of the ex pan sion across lat i tudes (Cerling et al., 1997). C4 grasses first be gan their ex pan sion in lower lat i - tudes, spread ing over the next few mil lion years to mid-lat i - tudes in Af rica, North and South Amer ica, Aus tra lia and parts of Eur asia (Cerling et al., 1997). The ex cep tions are also in - struc tive; some mid-lat i tude re gions re mained res o lutely C3. Lack of any ev i dence for C4 plants in Greece (Quade et al., 1994) and Tur key (Quade et al., 1995) and the south west ern
Cape of South Af rica (Franz-Odendaal et al., 2002) strongly sug gests that win ter-rain fall, Med i ter ra nean-type eco sys tems were al ready in place in these re gions in the late Mio cene and early Plio cene.
The near-syn chro nous ex pan sion of C4 grass lands across large parts of the world must have a global driver, but the ex act causes have re mained cu ri ously elu sive. A pop u lar and plau si - ble hy poth e sis (Cerling et al., 1997) pro posed that plum met ing CO2 lev els to wards the end of the Mio cene fa voured C4 plants, since they are adapted to cope with lower pCO2 lev els (Ehleringer et al., 1997). But more re cent ev i dence from alkenones in ma rine cores sug gests that no fall in pCO2 lev els oc curred at that time, since they were al ready low (Pagani et al., 1999), has led to a re-eval u a tion of all the ev i dence. Since new con ti nen tal-based car bon iso tope ev i dence sup ports the ma rine ev i dence (Segalen et al., 2002), the ma jor forc ing mech a nisms may re side in a com bi na tion of tec tonic and so lar in so la tion driv ers. Res o lu tion of the prob lem likely re quires a good deal more de tailed iso to pic and other ev i dence of the tran - si tion pe riod in var i ous re gions.
EVOLUTION OF OPEN HABITATS IN SOUTHERN AFRICA
A good deal of work on Plio- and Pleis to cene fau nas has been done in East and South Af rica, and more re cently a large-scale pro ject in Chad (Zazzo et al., 2000). The re sults showed that sig nif i cant pro por tions of C4 grasses were pres ent in the ear li est pe ri ods rep re sented by the South Af ri can sites, from about 3.3 Ma on wards (Lee-Thorp and van der Merwe, 1987; Sponheimer and Lee-Thorp, 1999b). Later stud ies were able to re fine the pic ture for var i ous pe ri ods, but this data merely shows pres ence or ab sence of C4. In the Chad case, this was very use ful, since the sites spanned ma jor changes from the late Mio cene to the Plio cene (more woody veg e ta tion) to the more open hab i tats of the Plio cene (Zazzo et al., 2000).
In cases where good data for a large cross-sec tion of the fau nal as sem blage ex ists, it has been pos si ble to de velop a
“C3/C4” in dex from the fau nal d13C data. The in dex, es sen tially cal cu lated from the num bers of gen era or in di vid ual spec i mens fall ing into one of a grazer, mixed feeder or browser, cat e gory, pro vides an in di ca tor of how closed or open the land scape was.
Im por tantly the in dex makes no as sump tions about the di etary hab its of the fauna, it is based rather on the as sump tion that in a closed hab i tat more brows ers will find ap pro pri ate for age, while graz ers will be fa voured in open grassy land scapes (Sponheimer and Lee-Thorp, 2003).
The in dex was ap plied to a se ries of sites, dif fer ing in age, in or der to con struct a view of the longterm evo lu tion of open hab i tats on the cen tral in land pla teau of south ern Af rica (Luyt and Lee-Thorp, 2003). The fau nal d13C pro vides a more di rect re flec tion of ac tual diet rather than one as sumed from tax o - nomic con sid er ations, as a re sult, a new view of land scape evo - lu tion emerges that dif fers from ear lier sce nar ios in one im por - tant re spect. It shows that the larg est and most sig nif i cant flo ral change to wards an open grassy land scape oc curred about 1.7–1.8 Ma, rather than in stepwise fash ion from 3 Ma ago as ear lier sug gested (Fig. 1). This re sult is con cor dant with the emer gence of very open grassy Serengeti-like eco sys tems in East Af rica at about this time.
2Sta ble light iso tope abun dances are by con ven tion ex pressed in the d for - mat in re la tion to an in ter na tional stan dard, in parts per mil. For 13C/12C the ex - pres sion is: d13C (‰) = (Rsam ple – Rstd)/Rstd ´ 1000 where: R = 13C/12C. The stan dard is Peedee Bel em nite car bon ate (VPDB). The same ex pres sion ap plies to 18O/16O but the stan dards are VPDB or Stan dard Mean Ocean Wa ter (VSMOW). The for mer is used through out this pa per.
OXYGEN ISOTOPE VARIABILITY IN ECOSYSTEMS
SYSTEMATICS OF 18O/16O IN CALCIFIED TISSUE MINERALS
Un til quite re cently ap pli ca tions of ox y gen iso topes in cal - ci fied tis sues have been based mostly on ap a tite phos phate rather than ap a tite car bon ate. This was be cause of the ex cep - tional strength of the P–O bond, which is un likely to ex change ox y gen at oms with wa ter, in con trast to car bon ate. Hence it was gen er ally as sumed that phos phate ox y gen iso tope com po si tion (dOp) is sta ble and that “the re cord is nearly per fectly re tained af ter death” (Luz et al., 1984, p. 256). One prob lem with this as ser tion is that it ne glects a ba sic fea ture of cal ci fied tis sue chem is try, the ten dency to recrystallise, and iso to pic al ter ation of phos phate has been dem on strated (see above).
Phos phate-based d18O stud ies have from the start fo cused on ex trac tion of palaeoclimate-re lated in for ma tion, and in par - tic u lar, on iso to pic com po si tion of palaeowaters and the es ti - ma tion of palaeotemperatures in con ti nen tal con texts. The link to palaeotemperature de rives from the ma jor ef fect of lat i tude on me te oric rain fall ox y gen iso tope ra tios, val ues be come in - creas ingly de pleted at higher cooler lat i tudes (Dansgaard, 1964). How ever, the re la tion ship is in fact rather com plex and sub ject to a whole range of in flu ences which may dom i nate on a re gional ba sis.
d18Op (and like wise d18Ocarb) is formed in equi lib rium with blood plasma (d18Obw), which is in turn de ter mined by the iso - to pic mass bal ance for the body (in put vs. out put) and strongly
in flu enced by the iso to pic com po si tion of en vi ron men tal wa - ters (d18Ow) (Longinelli, 1984). Luz and col leagues de rived flux mod els for ox y gen through the mam ma lian body, sup port - ing their mod els by ex per i men tal data on rats (Luz and Kolodny, 1985) and d18Obw, and d18Op ob ser va tions from mod - ern and his tor i cal hu mans re spec tively (Luz et al., 1984). They found sim ple lin ear re la tion ships be tween d18Ow, d18Obw and d18Op, but that de vi a tions de pended pri mar ily on the ra tio be - tween rate of drink ing to rates of drink ing and res pi ra tion, and pro duc tion of met a bolic wa ter. A third pre dic tion, that en vi ron - men tal ef fects other than iso to pic com po si tion of wa ter sources were small, was later re vised (Luz and Kolodny, 1989) fol low - ing ob ser va tions that where sig nif i cant amounts of body wa ter come from leaf wa ter, arid ity caused fur ther en rich ment in d18Op via leafwater evapotranspiration iso tope ef fects. Luz and Kolodny (1985, 1989) pro posed that those spe cies whose wa ter con sump tion was large rel a tive to en ergy ex pen di ture were more suit able in di ca tors of en vi ron men tal d18Ow. This is the model that has been fol lowed in palaeotemperature stud ies, with a fo cus on data from “well-be haved” spe cies (ac cord ing to Luz and Kolodny’s cri te ria), such as equids.
PALAEOCLIMATE APPLICATIONS BASED ON PHOSPHATE ANALYSES
Sub se quently, num bers of stud ies have ex plored and en - larged on the palaeotemperature/palaeocli mate theme. Sev eral aimed to es tab lish en vi ron men tal d18Ow and hence tem per a ture se quences from large her biv o rous mam mals such as equids (e.g. Sanchez-Chillon et al., 1994; Bryant et al., 1996) in con - texts rang ing in age from Pleis to cene to Mio cene or older.
Ayliffe and Chivas (1990) showed that, even if not re li able in - di ca tors of d18Ow are avail able, d18Op of drought-tol er ant macro pods could be used to in fer rel a tive hu mid ity in the past.
Fricke et al. (1995) ex am ined cli mate shifts from d18Op in Vi - king set tlers to de ter mine likely causes of their dis ap pear ance from Green land dur ing the 14th cen tury AD.
With wider ap pli ca tion came the reali sa tion that rea son able en vi ron men tal wa ter d18Ow or tem per a ture re sults could not be ex tracted in many cases, or the re sults were in ex pli ca bly vari - able (see Sanchez-Chillon et al., 1994), or the cal cu la tions re - quired as sump tions about d18Ow to de rive tem per a tures, a prac - tice prob lem atic even in the re cent (Ho lo cene) ma te ri als be ing stud ied (Stephan, 2000). The pos si bil ity of diagenesis was raised se ri ously for the first time. Ayliffe et al. (1994) dem on - strated that enamel d18Op from Pleis to cene el e phant ma te rial was in tact but bone and dentine d18Op from the same in di vid ual was not. How ever, emerg ing com plex intra-as sem blage and intra-in di vid ual vari abil ity were all too eas ily at trib uted to diagenesis, to the ne glect of other pos si ble in flu ences. In the face of un ex plained vari abil ity, the lack of an ob vi ous, in trin sic test for ox y gen iso to pic fi del ity meant that diagenesis could not be ruled out as a ma jor in flu ence. It has, how ever, tran spired that the trou ble some vari abil ity, once better un der stood, holds a key to these tests and may pro vide new in for ma tion about eco - sys tems and their nat u ral vari abil ity.
0 10 20 30 40 50 60 70 80
Mbr 4 Mbr 5a Mbr 5b
grazers mixed feeders browsers
Fig. 1. His to gram show ing strong shifts in the rel a tive pro por tions of graz ers, mixed feed ers and brows ers de rived from tooth enamel
d13C data in 3 fos sil fau nal as sem blages
The fauna are from mem bers 4, 5a and 5b at Sterkfontein Cave, Gauteng Prov ince, South Af rica and rep re sent pe ri ods at ap prox i mately 2.5, 2, and 1.7 Ma, re spec tively. Plot ted in this way, it is clear that the land scape be - comes sig nif i cantly more open with time. The larg est shift oc curs with the tran si tion to 5b at about 1.7 Ma; data from Luyt and Lee-Thorp (2003)
ISSUES OF VARIABILITY — INTERSPECIFIC
A com par a tive study of a small fau - nal as sem blage from Turkana, Kenya, sug gested a strong re la tion ship be tween d18Op and diet, and phys i ol ogy in ad di - tion to drink ing be hav iours (Kohn, 1996; Kohn et al., 1996). A hand ful of com par a tive stud ies us ing d18Ocarb have shown that the dis tri bu tion of ox y gen iso topes, in one place, and at one time, is strongly dif fer en ti ated. One ad van tage of car bon ate over phos phate anal y ses is that larger num bers of sam ples can be eas ily ana lysed, lead ing to more ro bust data sets from a greater di ver sity of en vi - ron ments. The Amboseli study, which fo cussed on large-bod ied her bi vores to meet the Luz and Kolodny cri te rion, found large dif fer ences of up to 6‰ in the means (not in di vid ual val ues, the vari a tion amongst which dif fers by up to10‰) and the means for large-bod ied graz ers alone dif fered by ~3‰ (Fig.
2A). Sim i larly large dif fer ences were ob served in two mod ern as sem blages from south ern Af rica, Takatshwane in Bot swana and the Morea Es tate in Klaserie, in east ern South Af rica (Fig.
2B, C). Takat shwane is lo cated in the Cen tral Kalahari near Ghanzi and there - fore rep re sents a warm open desert en vi - ron ment with sparse Kalahari thornveld veg e ta tion. Klaserie lies im me di ately to the east of the Kruger Na tional Park, near the es carp ment and rep re sents a warm, rel a tively moist, wood land en vi ron ment.
Ex pla na tions of fered for these large dif - fer ences in voke wa ter-re lated and cool - ing be hav iours, al though some of the de - tails may dif fer. There seems lit tle ar gu - ment that con sis tently low d18Ocarb val ues for Hip po pot a mus amphibius (hip po pot - a mus) are linked to their hab its of im mer - sion un der wa ter dur ing the day and for - ag ing at night. This be hav iour means that they do not need to cool down by mech a - nisms such as pant ing, and plant wa ters are de pleted in 18O at night (Bocherens et
al., 1996). The dif fer ences amongst the graz ers, which are mostly ob li gate drink ers, likely re flect sub tle dif fer ences in feed ing and drink ing be hav iour. For in stance, graz ers that pre fer to live near open wa ter sources, such as Kobus ellipsoprymnus (water buck) are con sis tently de pleted in 18O com pared to other graz ers. In ter - est ingly, the mean value for the equids (the “well-be haved”
grazer) is ~1‰ higher than an other large-bod ied grazer, Synce rus caffer (Af ri can buf falo) at Amboseli.
Ox y gen iso tope dis tri bu tion, un ex pect edly, also re flects trophic be hav iour. In Takatshwane and Klaserie, the
faunivores, Otocyon megalotis (bat-eared fox), Crocuta crocuta (spot ted hy ena) and Orycteropus afer (aard vark), are sig nif i cantly de pleted in 18O com pared to the her bi vores.
(p = 0.0004; ANOVA and t test). Low val ues for faunivores are likely re lated to higher di etary pro tein lev els, since pro teins are rel a tively de pleted in 18O com pared to car bo hy drates (Kohn, 1996; Sponheimer and Lee-Thorp, 2001). There are other pat - terned dif fer ences of in ter est, for in stance, suids and most pri - mates have rel a tively lower d18O, al though not so low as the faunivores (Fig. 2C).
Fig. 2. Plots of d18O ver sus d13C (both from tooth enamel car bon ate) in 3 mod ern fau nal as sem blages from Amboseli, Kenya (A), Takatshwane, Bot swana (B) and Klaserie,
South Af rica (C)
Takatshwane is lo cated in the Cen tral Kalahari, at ap prox i mately 22°S, 22°E, while Klaserie is lo - cated far to the East, at ap prox i mately 23°S, 31°E. These two sites form part of the same warm-sea son rain fall sys tem sourced in the In dian Ocean. d13C val ues dis tin guish C3 from C4 feed ers in a pre dict - able way; d18O shows high interspecific and even intra-spe cific vari abil ity as dis cussed in the text;
data from Bocherens et al. (1996, Amboseli), Sponheimer and Lee-Thorp (2001, Klaserie) and Lee-Thorp and Sponheimer (unpubl. data, Takatshwane)
These pat terns are es sen tially rep li cated in fos sil fau nal as - sem blages (Bocherens et al., 1996; Sponheimer and Lee-Thorp, 1999c; Luyt et al., 2000; Cerling, pers. comm.). Giraffids in Mio cene sites were con sis tently en riched, an ob ser va tion as - cribed to feed ing in the up per can opy where leaf-wa ter val ues are high (Cerling et al., 1997). Bocherens et al. (1996) showed that hip pos re tained their pre dict able low d18Ocarb val ues in Pleis - to cene sites, they fur ther sug gested that the large dif fer ences be - tween most her bi vores and hip pos could be used as tests for in - teg rity of d18O from enamel car bon ate. The idea was tested at the 5 Ma site of Langebaanweg on the south west ern coast of South Af rica and inter-spe cific in deed the ex tinct hippopotamids were sig nif i cantly lower than the other her bi vore fauna. This ob ser va - tion was par tic u larly valu able at this purely C3 site, since car bon iso tope dis tinc tions could not be used to test for iso to pic in teg rity (Franz-Odendaal, 2002).
Given these high lev els of inter-spe cific vari a tion at any one site, rep re sent ing one time and place, can use ful palaeoclimate in for ma tion be de rived? Some in for ma tion about pre vail ing, over all d18Ow is avail able from a bulk view of each dataset. Fig ure 2 shows that each of these sites has a dif - fer ent av er age value for the en tire fauna. The high est val ues, on av er age, are found at the in te rior Kalahari site of Takatshwane.
The re sults sug gest that evap o ra tion plays a strong role in rain - fall events and/or on en vi ron men tal wa ter sup plies avail able to an i mals. Given the site’s lo ca tion, the av er age d18Ow val ues at this site are likely to be a com pos ite of two iso tope ef fects work ing in op po si tion: the con ti nen tal ef fect (since the vapour source is in the In dian Ocean far to the east) and evap o ra tion.
Av er age val ues for Klaserie are more neg a tive, al though this site is nearer the vapour source. Rain fall at both of these sites (and likely Amboseli as well) show high co ef fi cients of vari a - tion on interannual and de cad al scales, so one might ex pect that lon ger term fau nal col lec tions would show even higher vari - abil ity (these were all short-term col lec tions over no more than a few years). High res o lu tion sta lag mite re cords from the Makapans Val ley in the same gen eral re gion as Klaserie, dem - on strate very high an nual-, de cad al- and cen ten nial-scale vari - abil ity in rain fall d18Ow (Lee-Thorp et al., 2001; Holmgren et al., 2003). Se rial iso tope data from el e phant ivory from the Kruger Na tional Park nearby sug gests that d18Ow and rain fall amount vari abil ity at sub- and de cad al-scales are re flected in ivory d18Ocarb (Codron, unpubl. data).
Inter-spe cific vari abil ity amongst the graz ers, al though pre - sent ing pos si ble sources of in for ma tion about palaeo ec ol ogy of an cient and ex tinct spe cies, clearly rep re sents a co nun drum for those wish ing to ex tract de tailed palaeoclimate in for ma tion and es pe cially for cal cu lat ing palaeotemperatures. The is sue, even when con fronted with a mod ern as sem blage of an i mals, is sim - ply one of which well-be haved spe cies to choose? This is even more prob lem atic when con fronted with fos sil fau nal as sem - blages, where a high pro por tion of the fau nal as sem blage are ex tinct and of (largely) un known hab its.
ISSUES OF VARIABILITY — INTRA-INDIVIDUAL
Sev eral phos phate- and car bon ate-based stud ies have shown that inter- and intra-tooth vari a tion could be used to in -
fer vari ables such as sea sonal cli ma tic am pli tude, sea son of birth and other bi o log i cal vari ables (e.g. Bryant et al., 1996;
Fricke and O’Neil, 1996; Sharp and Cerling, 1998; Stu art-Wil - liams and Schwarcz, 1998; Balasse et al., 2002;
Franz-Odendaal et al., 2003). The ex is tence of sea son ally dis - tinct pat terns at the same time pro vides a means for as sess ing whether the orig i nal ox y gen iso tope sig nals were pre served or not. This was one of the tac tics used in ad dress ing the is sue at the early Plio cene site (5 Ma) of Langebaanweg, where it was un clear that biogenic could be ex tracted from ma te rial of that age (us ing enamel car bon ate; Franz-Odendaal et al., 2002).
Intra-in di vid ual vari abil ity can pro vide short time se ries of se quen tial in for ma tion about range and am pli tude of sea son al - ity. This may be ac com plished in sev eral ways. Some stud ies have con cen trated on con tin u ously grow ing teeth such as bea - ver in ci sors (e.g. Stu art-Wil liams and Schwarcz, 1998), the chal lenge here be ing the meth ods re quired to deal with phos - phate anal y ses of very small sam ples. Oth ers have fo cussed on tooth rows and/or mul ti ple anal y ses of hypsodont or large, ma - ture teeth (e.g. Balasse et al., 2002; Fig. 3).
Fig ure 3 shows the re sults of sev eral stud ies doc u ment ing sea sonal pat terns from the West ern Cape, South Af rica. It is a coastal re gion with a win ter-rain fall, Med i ter ra nean-type cli mate and the rain fall sources are lo cated in the SE At lan tic. Ev i dence sug gests that this cli mate re gime has been in ex is tence for mil - lions of years (Franz-Odendaal, 2002), there fore com par i son across large pe ri ods of time re mains rea son ably ap pro pri ate. The se quences (com posed of the 2nd and 3rd mo lars) for the mod ern Raphicerus campestris (steen bok), a small mixed feeder, show a range in d18Ocarb of 4–5‰ across a nearly com plete sea sonal cy - cle (Fig. 3A). This is a rea son able rep re sen ta tion of the sea sonal rain fall cy cle, al though val ues are un doubt edly damp ened and av er aged be cause of the lim i ta tions im posed by the na ture of min er ali sa tion and sam pling (Balasse, 2003). Low points cor re - spond to low d13Ccarb, in di cat ing that low est points for d18Ocarb re - flect win ter. Data for the ap prox i mately 1000-year old sheep 3rd mo lars (Fig. 3B), from the pastoralist site of Kasteelberg, show a lower sea sonal range (~3‰) with val ues slightly pos i tively off - set. This “shrink ing” could be taken as ev i dence of some diagenesis, ex cept that there ap pears to be a con sid er able range in the sea son al ity in di ca tors of dif fer ent in di vid u als from the ar - chae o log i cal site (Balasse et al., 2002). More over, the do mes tic an i mals were man aged and there fore might not al ways show the ex tremes in di cated by wild an i mals. The Plio cene sivatheres (Fig. 3C) show a sim i lar sea sonal range (2–3‰) to the ar chae o - log i cal sheep, but the av er age val ues re sem ble the mod ern steen - bok more closely. These sam ples were sam pled more crudely (lower sam ple num bers per tooth) that those in Fig ure 3A and B, so it is likely that more time-av er ag ing oc curred. The small pos i - tive off set for the ar chae o log i cal sheep might in di cate slightly drier or warmer con di tions at this time, but clearly there is a great deal of intra- and inter-in di vid ual vari abil ity, likely rep re sent ing chang ing con di tions over the time pe riod of ac cu mu la tion.
Firmer con clu sions would re quire much larger sam ple num bers.
In spite of the dif fer ences in ranges of sea sonal am pli tude rep re sented at these sites, ob tain ing in for ma tion about sea sonal range would seem to be a more at tain able (and pos si bly, use ful) goal than de ter min ing large-scale se quences of d18O of me te - oric wa ter, which may vary on short or lon ger timescales for
any one of a num ber of rea sons. Intra-in di vid ual, inter-in di vid - ual and inter-spe cific vari abil ity is an area that prom ises to re - veal in for ma tion about both sea sonal cli ma tic am pli tude and
an i mal or hu man be hav iour. It’s not at all clear that at tempt ing to de rive d18Oof en vi ron men tal wa ters and tem per a tures, un der the cir cum stances de scribed above where high an nual or sub-de cad al vari abil ity ex ists in rain fall pat tern ing, and con se - quently, in d18O of en vi ron men tal wa ters. Es ti mat ing palaeo - temperatures from these kinds of highly vari able data is even more prob lem atic. There are sev eral im ped i ments, they in clude firstly, iden ti fy ing spe cies that best rep re sent am bi ent en vi ron - men tal wa ter con di tions, in the ab sence of fil ters, and sec ondly, high iso to pic rain fall vari abil ity where short-term shifts in d18Ow are high and would pro vide vastly over-in flated es ti - mates of tem per a ture shifts.
CONCLUSIONS
d13C and d18O from ei ther phos phate or car bon ate in fos sil tooth enamel, pro vide man i fold in di ca tors of past en vi ron men - tal and cli mate con di tions. d13C data is in for ma tive about dom i - nant flo ral sys tems in past en vi ron ments and where ap pro pri - ate, the strong dis tinc tion be tween C3 and C4 plants dou ble-up as in di ca tors of iso to pic in teg rity. Just two ex am ples of ap pli ca - tions have been dealt with in this pa per, many oth ers ex ist. The wide range of palaeoenvironmental ap pli ca tions is greatly as - sisted by the ease in ob tain ing large amounts of data and the level of de tail or an a lyt i cal pre ci sion re quired to es tab lish the nec es sary in for ma tion is not ex ces sively high. On the other hand, the level of pre ci sion re quired for palaeotemperature es ti - ma tions would seem to be fre quently un at tain able in the face of high en vi ron men tal and bi o log i cal vari abil ity. Vari abil ity as - cribed to dif fer ent an i mal be hav iours, while pre sent ing al ter na - tive means for test ing iso to pic in teg rity, pres ents im ped i ments to the goals of ex tract ing d18Ow and palaeotemperature in for - ma tion. Fur ther vari abil ity across short timescales sug gests strong in flu ences of highly vari able rain fall d18O at the low- to mid-lat i tude sites dis cussed in this pa per.
We have dis cussed inter- or intra-tooth shifts in d18O, rep re - sent ing sea sonal changes in d18Ow, largely for the pur poses of high light ing the prob lems pre sented by vari abil ity. Nev er the - less, it is ap pro pri ate to stress the pos i tive as pects as well.
Firstly, ob tain ing se quen tial in for ma tion about the range and am pli tude of sea son al ity seems to pres ent a more eas ily at tain - able, and per haps ul ti mately more use ful, goal for palaeo - climate re search. Sec ondly, pat terned vari abil ity, whether re - flect ing bi o log i cal spe cies dif fer ences or sea sonal shifts, of fers a ro bust means for test ing the re li abil ity of iso tope val ues from ei ther the car bon ate or phos phate com part ments. The is sues of vari abil ity raised here also clearly dem on strate the need for larger-scale anal y ses of fau nal ma te rial and in a greater va ri ety of set tings, than has pre vi ously been car ried out.
Ac knowl edge ments. We thank A.-V. Bojar and H. P.
Bojar for or gan is ing ESIR VII and F. J. Longstaffe, Z. Sharp, and an anon y mous re viewer, for their help ful com ments on the manu script. Fund ing sup port was pro vided by the Na tional Re - search Foun da tion (South Af rica), the Na tional Sci ence Foun - da tion (USA), the Leakey Foun da tion and the Uni ver sity of Cape Town.
Fig. 3. Intra-tooth vari abil ity and sea son al ity in mod ern, ar chae o log i cal and fos sil (Plio cene) teeth in the South west ern Cape A — mod ern steen bok, B — sheep from the ar chae o log i cal herder site of Kasteelberg( KBD), C — sivatheres from Plio cene age Langebaanweg (LBW). Two ex am ples are given in each case to show that sim i lar intra-tooth pat terns pre vail. d18O val ues are ex pressed rel a tive to VPDB.
The range of vari a tion in d18Ocarb is broadly sim i lar for all three sets, but it can be seen that intra- or inter-tooth vari a tion in one an i mal (mod ern or an - cient) is very large. The M2’s and M3’s for the steen bok, which have small teeth, pro vide part of a sea sonal se quence, while the M3’s of the larger an i - mals rep re sent a lon ger time pe riod, a year in the case of the sheep and >1 year for the sivatheres. Data are from Balasse et al. (2002) and Franz-Odendaal (2002) for Langebaanweg
REFERENCES
AYLIFFE L. K. and CHIVAS A. R. (1990) — Ox y gen iso tope com po si tion of the bone phos phate of Aus tra lian Kan ga roos: po ten tial as a palaeoenvironmental re corder. Geochim. Cosmochim. Acta, 54:
2603–2609.
AYLIFFE L. K., CHIVAS A. R. and LEAKEY M. G. (1994) — The re ten - tion of pri mary ox y gen iso tope com po si tions of fos sil el e phant skel e - tal phos phate. Geochim. Cosmochim. Acta, 58: 5291–5298.
BALASSE M. (2003) — Re con struct ing di etary and en vi ron men tal his tory from enamel iso to pic anal y sis: time res o lu tion of intra-tooth se quen - tial sam pling. In tern. J. Osteoarchaeology, 12: 155–165.
BALASSE M., AMBROSE S. A., SMITH A. B. and PRICE T. D. (2002)
— The sea sonal mo bil ity model for pre his toric herd ers in the south-west ern Cape of South Af rica as sessed by iso to pic anal y sis of sheep tooth enamel. J. Archaeol. Sc., 29: 917–932.
BLAKE R. E., ALT J. C. and MARTINI A. M. (2001) — Ox y gen iso tope ra tios of PO4: an in or ganic in di ca tor of en zy matic ac tiv ity and P me - tab o lism and a new biomarker in the search for life. PNAS, 98 (5):
2148–2153.
BOCHERENS H., KOCH P. L., MARIOTTI A., GERAADS D. and JAEGER J.-J. (1996) — Iso to pic biogeochemistry (13C, 18O) of mam - ma lian enamel from Af ri can Pleis to cene hominid sites. Palaios, 11:
306–318.
BOYDE A. (1967) — The de vel op ment of enamel struc ture. Proc. Royal Soc. Med i cine, 60: 923–933.
BRYANT J. D., FROELICH P. N., SHOWERS W. J. and GENNA B. J.
(1996) — A tale of two quar ries: bi o logic and taphonomic sig na tures in the ox y gen iso tope com po si tion of tooth enamel phos phate from mod ern and Mio cene equids. Palaios, 11: 397–408.
BRYANT J. D., KOCH P., FROELICH P. N., SHOWERS W. J. and GENNA B. J. (1996) — Ox y gen iso tope par ti tion ing be tween phos - phate and car bon ate in mam ma lian ap a tite. Geochim. Cosmochim.
Acta, 60: 5145–5148.
CERLING T. E. (1993) — Ex pan sion of C4 eco sys tems as an in di ca tor of global eco log i cal change in the Late Mio cene. Na ture, 361: 344–345.
CERLING T. E. and HARRIS J. M. (1999) — Car bon iso tope frac tion ation be tween diet and bioapatite in un gu late mam mals and im pli ca tions for eco log i cal and paleoecological stud ies. Oecologia, 120: 347–363.
CERLING T. E., HARRIS J. M., MACFADDEN B. J., LEAKEY M. G., QUADE J., EISENMAN V. and EHLERINGER J. R. (1997) — Global veg e ta tion change through the Mio cene/Plio cene bound ary. Na ture, 389: 153–158.
DANSGAARD W. (1964) — Sta ble iso topes in pre cip i ta tion. Tellus, 16:
436–468.
DRIESSENS F. C. M., VAN DIJK J. W. E. and BORRGREVEN J. M. P. M.
(1978) — Bi o log i cal cal cium phos phates and their role in the phys i ol - ogy of bone and den tal tis sues. 1. Com po si tion and sol u bil ity of cal - cium phos phates. Calcif. Tiss. Res., 26: 127–137.
EANES E. D. and POSNER A. S. (1970) — A note on the crys tal growth of hydroxyapatite pre cip i tated from aque ous so lu tions. Mat. Res. Bull., 5: 377–384.
EHLERINGER J. R., CERLING T. E. and HELLIKER B. R. (1997) — C4
photosyhtesis, at mo spheric CO2, and cli mate. Oecologia, 112: 285–299.
FRANZ-ODENDAAL T. A. (2002) — New anal y sis of den tal pa thol o gies in the Plio cene her bi vores of Langebaanweg and their palaeo - environmental im pli ca tions. Unpubl. Ph.D the sis. Univ. Cape Town.
FRANZ-ODENDAAL T. A., LEE-THORP J. A. and CHINSAMY A.
(2002) — New ev i dence for lack of C4 grass land ex pan sions dur ing the early Plio cene at Langebaanweg, South Af rica. Paleobiology, 28:
378–388.
FRANZ-ODENDAAL T. A., LEE-THORP J. A. and CHINSAMY A.
(2003) — In sights from sta ble light iso topes on enamel de fects and wean ing in Plio cene her bi vores. J. Biosc., 28: 765–773.
FRICKE H. C. and O’NEIL J. R. (1996) — Inter- and intra-tooth vari a tion in the ox y gen iso tope com po si tion of mam ma lian tooth enamel phos - phate: im pli ca tions for palaeoclimatogical and palaeo bio logi cal re - search. Palaeogeogr. Palaeoclimat. Palaeoecol., 126: 91–99.
FRICKE H. C., O’NEIL J. R. and LYNNERUP N. (1995) — Ox y gen iso - tope com po si tion of hu man tooth enamel from me di eval Green land:
link ing cli mate and so ci ety. Ge ol ogy, 23: 869–872.
HOLMGREN K., LEE-THORP J. A., COOPER G. R. J., LUNDBLAD K., PARTRIDGE T. C., SCOTT L., SITHALDEEN R., TALMA A. S. and TYSON P. D. (2003) — Per sis tent mil len nial-scale vari abil ity over the past 25 thou sand years in south ern Af rica. Quatern. Sc. Rev., 22:
2311–2326.
IACUMIN P., BOCHERENS H., MARIOTTI A. and LONGINELLI A.
(1996) — Ox y gen iso tope anal y ses of co-ex ist ing car bon ate and phos - phate in biogenic ap a tite: a way to mon i tor diagenetic al ter ation of bone phos phate? Earth Planet. Sc. Lett., 142: 1–6.
KOCH P. L., TUROSS N. and FOGEL M. L. (1997) — The ef fects of sam - ple treat ment and diagenesis on the iso to pic in teg rity of car bon ate in biogenic hydroxylapatite. J. Archaeol. Sc., 24: 417–429.
KOHN M. J. (1996) — Pre dict ing an i mal d18O: ac count ing for diet and phys i o log i cal ad ap ta tion. Geochim. Cosmochim. Acta, 60: 4811–4829.
KOHN M. J., SCHOENINGER M. J. and VALLEY J. W. (1996) — Her bi - vore tooth ox y gen iso tope com po si tion: ef fects of diet and phys i ol ogy.
Geochim. Cosmochim. Acta, 60: 3889–3896.
LEE-THORP J. A. (2000) — Pres er va tion of biogenic car bon iso tope sig - nals in Plio-Pleis to cene bone and tooth min eral. In: Biogeochemical Ap proaches to Paleodietary Anal y sis (eds. S. H. Ambrose and M. A.
Katzenberg). Ple num Press.
LEE-THORP J. A., HOLMGREN K., LAURITZEN S. E., LINGE H., MOBERG A., PARTRIDGE T. C., STEVENSON C. and TYSON P.
(2001) — Rapid shifts in the South ern Af ri can in te rior through out the mid to late Ho lo cene. Geophys. Res. Lett., 28 (23): 4507–4510.
LEE-THORP J. A., SEALY J. C. and VAN DER MERWE N. J. (1989) — Sta ble car bon iso tope ra tio dif fer ences be tween bone col la gen and bone ap a tite, and their re la tion ship to diet. J. Archaeol. Sc., 16: 585–599.
LEE-THORP J. A. and SPONHEIMER M. (2003) — Three case stud ies used to re as sess the re li abil ity of fos sil bone and enamel iso tope sig nals for paleodietary stud ies. J. Anthropol. Archaeol., 22 (3): 208–216.
LEE-THORP J. A. and VAN DER MERWE N. J. (1987) — Car bon iso tope anal y sis of fos sil bone ap a tite. South Af ri can J. Sc., 83: 71–74.
LEE-THORP J. A. and VAN DER MERWE N. J. (1991) — As pects of the chem is try of mod ern and fos sil bi o log i cal apatites. J. Archaeol. Sc., 18: 343–354.
LEGEROS R. Z. (1991) — Cal cium phos phates in Oral Bi ol ogy and Med i - cine. Karger. Paris.
LONGINELLI A. (1984) — Ox y gen iso topes in mam mal bone phos phate:
a new tool for paleohydrological and paleoclimatological re search?
Geochim. Cosmochim. Acta, 48: 385–390.
LUYT J. and LEE-THORP J. A. (2003) — Car bon iso tope ra tios of Sterkfontein fos sils in di cate a marked shift to open en vi ron ments ca.
1.7 Ma. S. Afr. J. Sc., 99: 271–273.
LUYT J., LEE-THORP J. A. and AVERY G. (2000) — New light on mid-Pleis to cene West coast en vi ron ments from Elandsfontein, West - ern Cape Prov ince, South Af rica. S. Afr. J. Sc., 96: 399–403.
LUZ B. and KOLODNY Y. (1985) — Ox y gen iso topes vari a tions in phos - phate of biogenic apatites, 4. Mam mal teeth and bones. Earth Planet.
Sc. Lett., 75: 29–36.
LUZ B. and KOLODNY Y. (1989) — Ox y gen iso tope vari a tion in bone phos phate. App. Geochem., 4: 317–323.
LUZ B., KOLODNY Y. and HOROWITZ M. (1984) — Frac tion ation of ox y gen iso topes be tween mam ma lian bone-phos phate and en vi ron - men tal drink ing wa ter. Geochim. Cosmochim. Acta, 48: 1689–1693.
PAGANI M., FREEMAN K. H. and ARTHUR M. A. (1999) — Late Mio - cene at mo spheric CO2 con cen tra tions and the ex pan sion of C4 grasses.
Sci ence, 285: 876–879.
PASSEY B. and CERLING T. E. (2002) — Tooth enamel min er al iza tion in ungulates: im pli ca tions for re cov er ing a pri mary iso to pic time-se ries.
Geochim. Cosmochim. Acta, 66 (18): 3225–3234.
PASSEY B., CERLING T. E. and SCHUSTER G. T. (2003) — In ver sion of sta ble iso tope sig nals in tooth enamel: re cov er ing the pri mary iso tope time-se ries. Geol. Soc. Am. Abstr. Progr., 35 (6).
PERSON A., BOCHERENS H., SALIêGE J.-F., PARIS F., ZEITOUN V.
and GÉRARD M. (1995) — Early diagenetic evo lu tion of bone phos - phate: an x-ray diffractometry anal y sis. J. Archaeol. Sc., 22: 211–221.
QUADE J., CERLING T. E., ANDREWS P. and ALPAGUT B. (1995) — Palaeodietary re con struc tion of Mio cene fau nas from Pasalar, Tur key us ing sta ble car bon and ox y gen iso topes of fos sil tooth enamel. J.
Hum. Evol., 28: 373–384.
QUADE J., CERLING T. E. and BOWMAN J. R. (1989) — De vel op ment of the Asian mon soon re vealed by a marked eco log i cal shift dur ing the lat est Mio cene in north ern Pa ki stan. Na ture, 342: 163–166.
QUADE J., SOLOUNIAS N. and CERLING T. E. (1994) — Sta ble iso to - pic ev i dence from paleosol car bon ates and fos sil teeth in Greece for C3
for est or wood lands over the past 11 Ma. Palaeogeogr. Palaeoclim.
Palaeoecol., 108: 41–53.
REY C., RENUGOPALAKRISHNAN V., SHIMIZU M., COL LINS B. and GLIMCHER M. J. (1991) — A res o lu tion en hanced Fou rier trans form in fra red spec tro scopic study of the en vi ron ment of the CO3 ion in the min eral phase of enamel dur ing its for ma tion and mat u ra tion. Calc.
Tiss. Int., 49: 259–268.
SÁNCHEZ-CHILLON B. ALBERDI M. T., LEONE G., BONADONNA F. B., STENNI B. and LONGINELLI A. (1994) — Ox y gen iso to pic com po si tion of fos sil equid tooth and bone phos phate: an ar chive of dif fi cult in ter pre ta tion. Palaeogeogr. Palaeoclimat. Palaeoecol., 107:
317–328.
SCHOENINGER M. J. and DENIRO M. J. (1982) — Car bon iso tope ra tios of ap a tite from bone can not be used to re con struct di ets of an i mals.
Na ture, 297: 577–578.
SÉGALEN L., RENARD M., PICKFORD M., SENUT B., COJAN I., LE CALLONNEC L. and ROGNON P. (2002) — En vi ron men tal and cli - ma tic evo lu tion of the Namib Desert since the Mid dle Mio cene: the con tri bu tion of car bon iso tope ra tios in ratite egg shells. Compt. Rend.
Géosc., 234: 917–924.
SHARP Z., ATUDOREI V. and FURRER H. (2000) — The ef fects of diagenesis on ox y gen iso tope ra tios of biogenic phos phates. Am. J.
Sc., 300: 222–237.
SHARP Z. D. and CERLING T. E. (1998) — Fos sil iso tope re cords of sea - sonal cli mate and ecol ogy: straight from the horse’s mouth. Ge ol ogy, 26: 219–222.
SPONHEIMER M. (1999) — Iso to pic Ecol ogy of the Makapansgat Limeworks Fauna. Unpubl. Ph.D. the sis. Rutgers Univ.
SPONHEIMER M. and LEE-THORP J. A. (1999a) — Al ter ation of enamel car bon ate en vi ron ments dur ing fos sili sa tion. J. Archaeol. Sc., 26: 143–150.
SPONHEIMER M. and LEE-THORP J. A. (1999b) — Iso to pic ev i dence for the diet of an early hominid, Australopithecus africanus. Sci ence, 283: 368–370.
SPONHEIMER M. and LEE-THORP J. A. (1999c) — Ox y gen iso topes in enamel car bon ate and their eco log i cal sig nif i cance. J. Archaeol. Sc., 26: 723–728.
SPONHEIMER M. and LEE-THORP J. A. (2001) — The ox y gen iso tope com po si tion of mam ma lian enamel car bon ate from Morea Es tate, South Af rica. Oecologia, 126: 153–157.
SPONHEIMER M. and LEE-THORP J. A. (2003) — Us ing car bon iso tope data of fos sil bovid com mu ni ties for palaeoenvironmental re con struc - tion. S. Afr. J. Sc., 99: 273–275.
STEPHAN E. (2000) — Ox y gen iso tope anal y sis of an i mal bone phos - phate: method re fine ment, in flu ence of consolidants, and re con struc - tion of palaeotemperatures for Ho lo cene sites. J. Archaeol. Sc., 27:
523–535.
STUART-WILLIAMS H. L. Q. and SCHWARCZ H. P. (1998) — Ox y gen iso to pic de ter mi na tion of cli ma tic vari a tion us ing phos phate from bea - ver bone, tooth enamel and dentine. Geochim. Cosmochim. Acta, 61:
2539–2550.
STUART-WILLIAMS H. L. Q., SCHWARCZ H. P., WHITE C. D. and SPENCE M. W. (1996) — The iso to pic com po si tion and diagenesis of hu man bone from Teotihuacan and Oaxaca, Mex ico. Palaeogeogr.
Palaeoclimat. Palaeoecol., 126: 1–14.
TUROSS N., BEHRENSMEYER A. K., EANES E. D., FISHER L. W. and HARE P. E. (1989) — Mo lec u lar pres er va tion and crys tal lo graphic al - ter ations in a weath er ing se quence of wil de beest bones. App.
Geochem., 4: 261–270.
ZAZZO A., BOCHERENS H., BRUNET M., BEAUVILAIN A., BILLIOU D., MACKAYE H. T., VIGNAUD P. and MARIOTTI A.
(2000) — Her bi vore paleodiet and paleoenvironment changes in Chad dur ing the Plio cene us ing sta ble iso tope ra tios in tooth enamel car bon - ate. Paleobiology, 26: 2994–309.
