q 2003 The Paleontological Society. All rights reserved. 0094-8373/00/2904-0005/$1.00 Paleobiology, 29(4), 2003, pp. 506–519 A paleoecological paradox: the habitat and dietary preferences of the extinct tethythere Desmostylus, inferred from stable isotope analysis Mark T. Clementz, Kathryn A. Hoppe, and Paul L. Koch Abstract.—The Desmostylia, an extinct order of mammals related to sirenians and proboscideans, are known from the late Oligocene to late Miocene of the North Pacific. Though often categorized as marine mammals on the basis of fossil occurrences in nearshore deposits, reconstructions of desmostylian habitat and dietary preferences have been somewhat speculative because morpho- logical and sedimentological information is limited. We analyzed the carbon, oxygen, and stron- tium isotope compositions of enamel from Desmostylus and co-occurring terrestrial and marine taxa from middle Miocene sites in California to address the debate surrounding desmostylian ecol- ogy. The d 13 C value of tooth enamel can be used as a proxy for diet. Desmostylus had much higher d 13 C values than coeval terrestrial or marine mammals, suggesting a unique diet that most likely consisted of aquatic vegetation. Modern aquatic mammals tend to exhibit lower variability in d 18 O values than terrestrial mammals. Both fossil marine mammals and Desmostylus exhibited low d 18 O variability, suggesting that Desmostylus spent a large amount of time in water. Finally, the Sr isotope composition of marine organisms reflects that of the ocean and is relatively invariant when com- pared with values for animals from land. Sr isotope values for Desmostylus were similar to those for terrestrial, rather than marine, mammals, suggesting Desmostylus was spending time in estu- arine or freshwater environments. Together, isotopic data suggest that Desmostylus was an aquatic herbivore that spent a considerable portion of its life foraging in estuarine and freshwater ecosys- tems. Mark T. Clementz. Department of Earth Sciences, University of California, Santa Cruz, California 95064. E-mail: [email protected]Kathryn A. Hoppe. Department of Environmental Science, Policy & Management, University of California, Berkeley, California 94720. E-mail: [email protected]Paul L. Koch. Department of Earth Sciences, University of California, Santa Cruz, Califorina 95064. E-mail: [email protected]Accepted: 5 March 2003 Introduction Mammalian herbivores are a minor com- ponent of the total fauna in modern marine and coastal ecosystems, limited to just four si- renian species. Nonetheless, these herbivores are thought to play critical roles in structuring the species richness and productivity of these ecosystems (Bowen 1997; Peterken and Con- acher 1997). These effects may have been even greater in the past, when the diversity of large-bodied, mammalian herbivores was higher (Domning and Furusawa 1992; Aran- da-Manteca et al. 1994; Domning 2001). Other groups of mammals may have exploited the abundant vegetation in shallow coastal waters that were widespread from the Eocene through the Miocene, such as the Desmostylia, an extinct group of hippo-sized mammals re- lated to sirenians. The coexistence of several species of large-bodied mammalian herbi- vores in coastal ecosystems has no modern an- alog, so the dynamics of ancient coastal eco- systems may have been very different from those today. The feeding ecology and habitat preferences of desmostylians are not well understood, but before we can begin to explore what part, if any, desmostylians played in past coastal eco- systems, we need such basic autecological data. Here, we reconstruct the ecology of one genus, Desmostylus, through stable isotope analysis of tooth enamel. Using carbon iso- topes, we assessed feeding preferences, ex- ploring levels of dependence on terrestrial versus aquatic food sources. To assess the ad- aptation of Desmostylus to aquatic habitats, we used a method that relies on contrasts in ox- ygen isotope variability between terrestrial and aquatic species. Finally, we explored dif- ferences in habitation of marine, estuarine,
Transcript
q 2003 The Paleontological Society. All rights reserved. 0094-8373/00/2904-0005/$1.00
Paleobiology, 29(4), 2003, pp. 506–519
A paleoecological paradox: the habitat and dietary preferences ofthe extinct tethythere Desmostylus, inferred from stableisotope analysis
Mark T. Clementz, Kathryn A. Hoppe, and Paul L. Koch
Abstract.—The Desmostylia, an extinct order of mammals related to sirenians and proboscideans,are known from the late Oligocene to late Miocene of the North Pacific. Though often categorizedas marine mammals on the basis of fossil occurrences in nearshore deposits, reconstructions ofdesmostylian habitat and dietary preferences have been somewhat speculative because morpho-logical and sedimentological information is limited. We analyzed the carbon, oxygen, and stron-tium isotope compositions of enamel from Desmostylus and co-occurring terrestrial and marinetaxa from middle Miocene sites in California to address the debate surrounding desmostylian ecol-ogy. The d13C value of tooth enamel can be used as a proxy for diet. Desmostylus had much higherd13C values than coeval terrestrial or marine mammals, suggesting a unique diet that most likelyconsisted of aquatic vegetation. Modern aquatic mammals tend to exhibit lower variability in d18Ovalues than terrestrial mammals. Both fossil marine mammals and Desmostylus exhibited low d18Ovariability, suggesting that Desmostylus spent a large amount of time in water. Finally, the Sr isotopecomposition of marine organisms reflects that of the ocean and is relatively invariant when com-pared with values for animals from land. Sr isotope values for Desmostylus were similar to thosefor terrestrial, rather than marine, mammals, suggesting Desmostylus was spending time in estu-arine or freshwater environments. Together, isotopic data suggest that Desmostylus was an aquaticherbivore that spent a considerable portion of its life foraging in estuarine and freshwater ecosys-tems.
Mark T. Clementz. Department of Earth Sciences, University of California, Santa Cruz, California 95064.E-mail: [email protected]
Kathryn A. Hoppe. Department of Environmental Science, Policy & Management, University of California,Berkeley, California 94720. E-mail: [email protected]
Paul L. Koch. Department of Earth Sciences, University of California, Santa Cruz, Califorina 95064.E-mail: [email protected]
Accepted: 5 March 2003
Introduction
Mammalian herbivores are a minor com-ponent of the total fauna in modern marineand coastal ecosystems, limited to just four si-renian species. Nonetheless, these herbivoresare thought to play critical roles in structuringthe species richness and productivity of theseecosystems (Bowen 1997; Peterken and Con-acher 1997). These effects may have been evengreater in the past, when the diversity oflarge-bodied, mammalian herbivores washigher (Domning and Furusawa 1992; Aran-da-Manteca et al. 1994; Domning 2001). Othergroups of mammals may have exploited theabundant vegetation in shallow coastal watersthat were widespread from the Eocenethrough the Miocene, such as the Desmostylia,an extinct group of hippo-sized mammals re-lated to sirenians. The coexistence of severalspecies of large-bodied mammalian herbi-
vores in coastal ecosystems has no modern an-alog, so the dynamics of ancient coastal eco-systems may have been very different fromthose today.
The feeding ecology and habitat preferencesof desmostylians are not well understood, butbefore we can begin to explore what part, ifany, desmostylians played in past coastal eco-systems, we need such basic autecologicaldata. Here, we reconstruct the ecology of onegenus, Desmostylus, through stable isotopeanalysis of tooth enamel. Using carbon iso-topes, we assessed feeding preferences, ex-ploring levels of dependence on terrestrialversus aquatic food sources. To assess the ad-aptation of Desmostylus to aquatic habitats, weused a method that relies on contrasts in ox-ygen isotope variability between terrestrialand aquatic species. Finally, we explored dif-ferences in habitation of marine, estuarine,
507HABITAT AND DIETARY PREFERENCES OF THE EXTINCT TETHYTHERE DESMOSTYLUS
FIGURE 1. Proposed cladogram for the mirorder Tethytheria modified after Inuzuka et al. 1995.
FIGURE 2. Three different skeletal reconstructions ofdesmostylians based on morphological data and pro-posed lifestyle. A, Cole and Domning (1998) proposedthat Paleoparadoxia exhibited a posture and joint me-chanics similar to those of extinct ground sloths and theextant hippopotamus. B, Repenning’s (1965) reconstruc-tion of Desmostylus, with a posture similar to that ofmodern otariid pinnipeds, was based on interpretingthe fore- and hindlimbs as modified into flippers. C,Inuzuka (1984) proposed a completely new style of pos-ture for Desmostylus, termed herpetiform, which wasthought to enhance stability in high-wave-energy envi-ronments.
and terrestrial systems by identifying differ-ences in strontium isotope compositionamong taxa.
Background
The Desmostylia. Desmostylians are large,hippo-sized mammals known to have inhab-ited the Pacific coast of North America andAsia from the late Oligocene (;28 Ma) to latemiddle Miocene (;10 Ma) (Barnes et al. 1985;Inuzuka et al. 1995; Inuzuka 2000). Severalcharacters of the skull suggest that desmos-tylians share a common ancestry with sireni-ans and proboscideans, which are classifiedwith desmostylians in the mirorder Tethyth-
eria (Domning et al. 1986; Novacek and Wyss1987) (Fig. 1). The Desmostylia is composed oftwo families, the Paleoparadoxidae, whichpossessed low-crowned or bunodont molars,and the Desmostylidae, which had high-crowned or hypsodont molars (Inuzuka et al.1995; Inuzuka 2000). Both families coexistedalong the northern Pacific coast of Asia andNorth America until the late middle Miocene.
Desmostylian ecology is currently a mys-tery. Though most desmostylian fossils havebeen found in nearshore marine deposits,there are occasional reports of desmostyliansassociated with terrestrial mammals at siteswhere the depositional environment is un-clear. Desmostylian morphology has also gen-erated conflicting ecological interpretations.Early reconstructions portrayed desmostyli-ans with a pinniped-like posture (Fig. 2A),suggesting they were fully aquatic swimmers(Repenning 1965), but recent studies compar-ing desmostylians with extant aquatic and ter-restrial mammals have challenged this idea.Cole and Domning (1998) argued that des-mostylian postcranial elements were mostsimilar to those of modern terrestrial ungu-lates or extinct ground sloths, and they sug-gested a mode of locomotion and a lifestylesimilar to those of the semiaquatic hippopot-amus. A more radical interpretation was sug-gested by Inuzuka (1984). Unlike other mam-mals, which hold their limbs directly underthe body, Inuzaka’s reconstructed desmosty-lians had a ‘‘herpetiform’’ or sprawling pos-ture similar to lizards, which would have pro-
508 MARK T. CLEMENTZ ET AL.
vided a high degree of stability against lateralforces, such as those produced by waves innearshore ecosystems. It is unclear which (ifany) interpretation is correct.
Most interpretations imply some use ofaquatic habitats by desmostylians, but the ex-tent to which they foraged within these eco-systems is uncertain. Cranial and dental mor-phology, as well as their phylogenetic posi-tion, suggests that desmostylians were prin-cipally herbivores, though at least one studyhas proposed that they consumed mollusks(McLeod and Barnes 1984). Their northerngeographical distribution and the deposition-al settings of their occurrences led Domninget al. (1986) to conclude that desmostyliansprimarily ate marine vegetation (e.g., macroal-gae, seagrass) in cold, marginal marine envi-ronments. However, if desmostylians were ca-pable of terrestrial locomotion (Cole andDomning 1998), they could have come ashoreto feed on terrestrial plants, filling an ecolog-ical niche similar to that of the hippopotamus.Thus, even the dietary preferences of this ex-tinct group of mammals are unresolved.
Carbon Isotopes and Foraging Preferences.Analysis of carbon isotopes in biogenic sub-strates (e.g., bone or enamel carbonate, colla-gen) has proven useful for paleodietary recon-structions (see reviews in Schwarcz andSchoeninger 1991; Koch et al. 1995). Toothenamel carbonate is the preferred substratefor reconstructions more than 100,000 yearsold because it is less susceptible to diageneticalteration (Lee-Thorp 2000; Wang and Cerling1994; Koch et al. 1997). The d13C value1 of car-bonate in tooth enamel apatite is labeled bythe d13C of an animal’s diet. For ungulate her-bivores, enamel d13C values are controlled bythe d13C value of the vegetation the animalconsumes with a small physiological fraction-ation of ;112–14.1‰ for wild populations(Lee-Thorp et al. 1989; Cerling and Harris1999).
The d13C values of primary producers at thebase of food webs vary as a result of differ-
1 d13C 5 [(13C/12Csample 4 13C/12Cstandard) 21) * 1000], wherethe standard is V2PDB. d18O follows the same conven-tions, where the ratios are 18O/16O and the standard isV2SMOW. Units are parts per thousand (‰).
ences in photosynthetic physiology, sources offixed carbon, and environmental conditions.In terrestrial ecosystems, differences are re-lated to photosynthetic pathway (O’Leary1988), resulting in relatively high d13C values(213 6 2‰) for plants using C4 photosynthe-sis (e.g., warm-climate grasses), low d13C val-ues (227 6 3‰) for C3 photosynthesizers(e.g., trees, most shrubs, herbs, and cool-cli-mate grasses), and d13C values that can be any-where between these extremes for plants us-ing CAM photosynthesis (e.g., most succu-lents). In aquatic systems, environmental con-ditions (dissolved [CO2], mixing of the watercolumn, nutrient supply, etc.) have a strongerinfluence on primary-producer d13C values,creating differences in mean d13C values forkelp (217 6 4‰), seagrass (210 6 3‰), near-shore and offshore marine phytoplankton(220 6 2‰, 223 6 2‰, respectively), andfreshwater vegetation (226 6 7‰) (Osmondet al. 1981; Boon and Bunn 1994; Hemmingaand Mateo 1996; Rau et al. 2001; Ravens et al.2002).
Many studies have exploited these differ-ences in primary-producer d13C values to tracethe foraging habits of terrestrial and marineconsumers (Cerling and Harris 1999; Cle-mentz and Koch 2001) (see Table 1). Withinmarine habitats, consumer d13C values typi-cally increase toward shore, with the highestd13C values reported for consumers withinkelp or seagrass beds. Onshore, consumerd13C values vary among C4 grazers (highd13C), C3 browsers and grazers (low d13C), andfreshwater foragers (very low d13C values).Overall, enamel d13C values provide strong ev-idence about the types of food mammals con-sume and from which food webs that foodcame. We used this approach to determine theecosystems in which Desmostylus foraged.
Oxygen Isotopes and Aquatic Habitat Use.The d18O value of biogenic apatite in bonesand teeth is a function of the d18O of body wa-ter, plus a temperature-dependent fraction-ation that is constant in homeothermic mam-mals (Longinelli 1984; Luz et al. 1984). Unlikecarbon in enamel, which has just one source(i.e., diet), body water has multiple sources ofoxygen, and both environmental and physio-logical factors influence its d18O value. Physi-
509HABITAT AND DIETARY PREFERENCES OF THE EXTINCT TETHYTHERE DESMOSTYLUS
TABLE 1. Mean d13C and d18O values for modern taxa collected from central California and Amboseli Park in Kenya.Values are reported 6 1s.
* Data from Clementz and Koch 2001. All collections are from geographically constrained modern populations, chiefly in central and southern Cal-ifornia. Terrestrial plants in this region are dominantly C3.
‡ Data for populations from Amboseli Park, Kenya, discussed in Bocherens et al. 1996. Grasses in Amboseli are C4, whereas nearly all trees, shrubs,and herbs are C3. Grant’s gazelle and rhinoceros are browsers on C3 plants, wildebeest and zebra are grazers on C4 plants, and elephants are mixedfeeders.
ology affects body water d18O values by con-trolling the magnitude of oxygen fluxes andisotopic fractionation that occurs as oxygenpasses into and out of the body (Luz and Ko-lodny 1985; Bryant and Froelich 1995; Kohn1996). The d18O values of environmental oxy-gen sources provide the baseline from whichmammalian body water d18O values can de-viate via physiological effects.
The mean d18O value of mammalian toothenamel has been proposed as a monitor of ad-aptation to freshwater or marine systems(Bocherens et al. 1996; Roe et al. 1998). How-ever, mean d18O values may not always be di-agnostic of habitat preferences (Clementz andKoch 2001), because physiological differencesamong species also affect mean values. Differ-ences in the population-level standard devia-tion of enamel d18O values, on the other hand,can allow discrimination of aquatic/semi-aquatic from terrestrial species (Table 1). Ter-restrial mammals experience greater physio-logical and environmental variability thanaquatic/semiaquatic mammals, resulting inhigher variability in their body water andtooth enamel d18O values.
Fully aquatic and semiaquatic species typi-
cally yield d18O standard deviations of 0.5‰or less, whereas terrestrial mammals typicallyhave values .1‰ (Table 1). However, thereare exceptions to this pattern. For terrestrialmammals, body size and environmental con-ditions may cause d18O variation of popula-tions to be lower than predicted. Large mam-mals (.1000 kg) obtain more of their oxygenfrom drinking water (;60%) than do smallermammals (;20%) (Bryant and Froelich 1995).If drinking-water sources are isotopically ho-mogeneous, large-mammal d18O valuesshould show less variability among individ-uals (Amboseli elephants; Table 1). Likewise,for animals living in humid environmentswith low evaporative water loss, d18O valueswithin a population may be less variable.Among aquatic species, differences in vari-ability can result from d18O variation of oxy-gen sources, either by movement between wa-ters of different isotopic composition (e.g.,marine vs. fresh water—sea otters, estuarineriver otters, manatees) or by ingestion of otheroxygen sources with distinct d18O values (e.g.,terrestrial, 18O-enriched plant water—Ambo-seli hippopotamus). Though these exceptionsdo generate overlap in 1s values for aquatic
510 MARK T. CLEMENTZ ET AL.
and terrestrial mammals, only fully aquatic,strictly marine (e.g., most pinnipeds and ce-taceans), and semiaquatic mammals that for-age aquatically (e.g., river otters) exhibit 1sd18O values #0.5‰, providing a clear meansof differentiating these habitat preferencesfrom fully terrestrial populations. Here, wefirst test this approach on fossil taxa withknown habitat preferences, and then use it toassess the extent of aquatic habitat use by Des-mostylus.
Strontium Isotopes and Marine versus Terres-trial Ecosystems. The ratio of 87Sr to 86Sr inbiogenic materials has proven to be a valuabletool for extracting ecological information frommodern and fossil organisms (Koch et al. 1992,1995; Kennedy et al. 1997; Hoppe et al. 1999;Ingram and Weber 1999). Because strontium isnot fractionated measurably when it is incor-porated, the 87Sr/86Sr ratio of biogenic mate-rial is identical to that of the source of Sr andis passed up the food web without modifica-tion (Capo et al. 1998).
The Sr sources for an organism are the wa-ter it drinks or inhabits and the plant or ani-mal food it ingests. On land, the 87Sr/86Sr ra-tios of rivers and plants are controlled by rockand soil compositions (Miller et al. 1993). Ra-tios of terrestrial rocks are highly variable anddepend on rock type and age (Capo et al.1998). River and soil 87Sr/86Sr ratios are con-trolled both by bedrock inputs and by atmo-spheric input of Sr as dust and precipitation(Capo et al. 1994; Kennedy et al. 1998). Atcoastal sites, precipitation has a 87Sr/86Sr ratiosimilar to that of the ocean, and if rainfall issignificant, it can dominate the 87Sr/86Sr ratioof soils (Kennedy et al. 1998); dust inputs,however, can affect regional soil isotope com-positions far from dust sources (Muhs et al.1990).
Today, the 87Sr/86Sr ratio of seawater is ho-mogeneous globally, because the residencetime of oceanic Sr is several million years, or-ders of magnitude longer than the timescalefor oceanic mixing (;1000 years). The 87Sr/86Sr ratio of seawater is controlled by the mag-nitude and isotopic ratios of Sr influxes, in-cluding weathering of terrestrial rocks andoceanic basalt alteration, and by Sr removalthrough deposition of marine carbonates. Sea-
water 87Sr/86Sr ratios show long timescalefluctuations due to shifts in these fluxes(Armstrong 1971; Capo and DePaolo 1990).
The 87Sr/86Sr ratio of water in estuaries iscontrolled by mixing. Freshwater has a muchlower [Sr] (0.006–2.94 ppm) than seawater (8.0ppm) (Capo et al. 1998). Mixing of fresh andmarine water produces a correlation of 87Sr/86Sr ratio with salinity, but differences in [Sr]cause the marine 87Sr/86Sr signal to dominate(Bryant et al. 1995).
Modern animals foraging in marine sys-tems have 87Sr/86Sr ratios that closely matchthose of seawater (Schmitz et al. 1997; Venne-mann et al. 2001). Modern land and freshwa-ter animals, in contrast, exhibit a greater rangeof ratios that depend on the geologically con-trolled differences at the base of the food web(Nelson et al. 1986; Koch et al. 1995; Ingramand Weber 1999). Animals foraging in estu-aries have not received extensive study butwould be expected to show a wide range in87Sr/86Sr ratios among individuals in a popu-lation, depending on the salinity of the waterthey frequent. Sr isotope ratios will be our pri-mary tool for determining whether Desmos-tylus inhabited marine, estuarine, or fully ter-restrial ecosystems.
Diagenetic Monitoring via Control Taxa. Pri-or work has shown that tooth enamel is resis-tant to diagenetic alteration of stable isotoperatios (Wang and Cerling 1994; Bocherens etal. 1996; Koch et al. 1997). Still, we will usecontrol taxa to test for alteration wheneverpossible. For example, we expect population-level d18O variability to be higher in obviouslyterrestrial animals (e.g., horses) than in obvi-ously marine mammals (e.g., whales and dol-phins). Because diagenetic alteration wouldtend to homogenize isotopic signals amongspecimens from a single locality, retention ofconsistent differences in mean and variabilityin control taxa provides support for the as-sumption that isotopic patterns in Desmostylusare preserved as well.
Materials and Methods
Locality and Specimen Information. Speci-mens were obtained from five sites in centraland southern California (Fig. 3). Four siteswere in a restricted, shallow marine basin that
511HABITAT AND DIETARY PREFERENCES OF THE EXTINCT TETHYTHERE DESMOSTYLUS
FIGURE 3. A, Map of modern-day central California highlighting the location of sampling sites included in thisproject. B, Central California as it appeared at ;13–16 Ma, during the time that our fossil specimens were deposited(modified from Bartow 1987).
inundated the interior of California during theMiocene; one site was exposed to open-ma-rine conditions. Site 1292 is the famous SharkTooth Hill deposit within the Round Moun-tain Silt Formation, which has an extensive ac-cumulation of marine mammals, shark’s teeth,and rare terrestrial mammals deposited dur-ing the Barstovian Land Mammal Age (;13.5to 15.9 Ma) (Barnes 1976). We obtained severalDesmostylus molars and teeth from marinemammals, including the early pinniped Allo-desmus and small and large odontocetes(toothed whales). At site 2124, Desmostylusteeth and isolated molars from the early horseMerychippus and the proboscidean Gomphoth-erium were obtained from the Temblor For-mation, which is estimated to be coeval withthe Shark Tooth Hill deposit. At site V5555,Desmostylus from the Santa Margarita For-mation (late Miocene, Clarendonian Age, ;10Ma) were found in association with marine(e.g., Allodesmus, small odontocetes, sirenians)and terrestrial mammals (e.g., the equid Hip-parion). Site V3301 is from a reef bed depositin the Temblor Formation, which has yieldeda large number of Desmostylus molar frag-ments of Barstovian age, but no control taxa.
Control taxa were also absent at the open-ma-rine site V3215, which was located north of theother sites in the Barstovian-aged Briones For-mation.
Sampling Protocol and Analyses. Stable iso-tope analysis was conducted on the carbonatewithin tooth enamel biological apatite. Ap-proximately 5 to 10 mg of powder was drilledfrom each tooth after the surface had beenabraded to remove possible contamination.Following the protocol in Koch et al. (1997), allpowders were soaked for 24 hours in ;2%NaOCl to oxidize organic matter, rinsed fivetimes with distilled water, soaked for 24 hoursin 1 M calcium acetate-buffered/acetic acid toremove contaminating carbonate in non-lat-tice sites, rinsed five times with distilled wa-ter, and then freeze-dried. Because extendedexposure time may alter enamel isotope val-ues, all samples were treated for the samelength of time (24 hours) to ensure the com-parability of isotope values and reduce therisk of lab-induced variation in isotope values(Koch et al. 1997).
Approximately 1 mg of pretreated powderwas analyzed using an Isocarb automated car-bonate analysis system interfaced with a Mi-
512 MARK T. CLEMENTZ ET AL.
TABLE 2. Mean d13C and d18O, and Sr isotope values for taxa from each site. All values are reported 6 1s.
Locality Taxon nMean
d13C 6 1sMean
d18O 6 1sMean
87Sr/86Sr 6 1s
1292129212921292
DesmostylusAllodesmusOdontocete, largeOdontocete, small
cromass Optima gas source mass spectrome-ter in the Deptartments of Earth and OceanSciences, University of California, Santa Cruz.Samples were dissolved in 100% phosphoricacid at 908C, with concurrent cryogenic trap-ping of CO2 and H2O. The CO2 was then ad-mitted to the mass spectrometer for analysis.The standards used in this study were CarreraMarble and NBS 19 and values are reportedrelative to V-PDB (for carbon) and V-SMOW(for oxygen). Precision, determined by repeat-ed concurrent analysis (n 5 19) of a modernelephant enamel standard, was 0.1‰ for d13Cand 0.2‰ for d18O.
For 87Sr/86Sr analysis, ;1 mg of powder wascollected from each sample, soaked in 0.5 mlof 1 N acetic acid for 20 minutes, rinsed in dis-tilled water, and then repeated four moretimes. Samples were then dissolved overnightin 2.5 N HCl, dried down, redissolved in 0.5ml of 2.5 N HCl, and injected onto a columnfilled with a cation exchange resin to isolate Sr.After washing with 20 ml of 2.5 N HCl, an ad-ditional 7 ml of 2.5 N HCl was passed throughthe column and collected for analysis. Sam-ples were dried down overnight then dis-solved in 1 ml of 10% nitric acid and placed onrhenium filaments. To conduct the isotopeanalyses we used the VG 354 Thermal Ioni-zation Mass Spectrometer located in the Dep-tartment of Earth Sciences at the University ofCalifornia, Santa Cruz. Precision, determinedby repeated measurement of the NBS 987 Srstandard, was 6 0.00003.
Data Analysis. We assessed the statistical
significance of mean values by using a Stu-dent’s t-test for comparisons between twopopulations or a parametric, one-factor anal-ysis of variance (ANOVA) for multiple popu-lations. If a significant difference was detectedamong populations, we applied a pairwisecomparison (post-hoc Tukey test) to identifypopulations that were statistically distinct. Touse these tests, however, sample populationsmust be normally distributed and equivalentin variance. For sample populations that didn’tmeet these criteria, we assessed statisticallysignificant differences in median values usinga nonparametric Kruskal-Wallis one-factoranalysis of variance followed by pairwisecomparison using the post-hoc Dunn’s meth-od. We note which method of comparison wasused in the ‘‘Results’’ section. For compari-sons of variance between populations, a sim-ple F-test was used. Spearman’s rank correla-tion method was used to assess significance ofcorrelation of different isotopic values amongsamples. Either Sigmastat 2.03 or MicrosoftExcel 2000 was used for all calculations.
Results
Samples were grouped into three categoriesfor statistical comparisons: unambiguouslymarine mammals, unambiguously terrestrialmammals, and Desmostylus. Desmostylus hadthe highest mean d13C value (61s) (25.2 61.9‰), followed by marine mammals (28.1 61.4‰), then terrestrial mammals (29.5 61.2‰) (Table 2, Fig. 4A). Median d13C valuesdiffered significantly among groups (Kruskal-
513HABITAT AND DIETARY PREFERENCES OF THE EXTINCT TETHYTHERE DESMOSTYLUS
FIGURE 4. Carbon (a), oxygen (b), and strontium (c) isotope data collected for all taxa. Along the horizontal axis,terrestrial (squares) and marine (triangles) faunas are grouped on either side of the data for Desmostylus (circles)from all five sites. A 5 Allodesmus; G 5 Gomphotherium; H 5 Hipparion; LO 5 large odontocete; M 5 Merychippus;SO 5 small odontocete. Enlarged, closed symbols represent the mean value for each taxon or group, and all othervalues are plotted as open symbols. Vertical bars represent 61s. Estimated range in middle Miocene seawater 87Sr/86Sr values is based on Hodell et al. 1991.
514 MARK T. CLEMENTZ ET AL.
Wallis: H 5 53.927, p , 0.01). Pairwise com-parisons revealed that Desmostylus was signif-icantly different from both marine and terres-trial mammals, but that marine and terrestrialmammals were not different from each other(Dunn’s method). Differences in d13C varianceamong groups were only significant betweenDesmostylus and terrestrial mammals (F-test: p, 0.01). Within marine mammals, mean d13Cvalues differed significantly between Allodes-mus and all cetaceans (t-test: t 5 2.413, d.f. 520, p 5 0.026). Mean d13C values also differedsignificantly among terrestrial mammals(one-factor ANOVA: F 5 39.591, p , 0.01);pairwise comparison revealed statistically sig-nificant differences among all three groups(Tukey test: p , 0.01).
The mean d18O value for terrestrial mam-mals (28.3 6 1.4‰) was higher than for eithermarine mammals (27.6 6 0.6‰) or Desmos-tylus (27.7 6 0.7‰) (Table 2, Fig. 4B). Mediand18O values did not differ among the threegroups of mammals (Kruskal-Wallis: H 53.823, p 5 0.148). Desmostylus and marinemammals exhibited significant differences ind18O variance from terrestrial mammals (F-test: p , 0.01), but not between each other (F-test: p 5 0.496). Within marine mammals, nosignificant difference was detected betweenmeans for Allodesmus and all cetaceans (t-test:t 5 21.029, p 5 0.316). Mean d18O values diddiffer significantly among terrestrial mam-mals (one-factor ANOVA: F 5 4.974, p 50.014); pairwise comparison revealed thatonly Merychippus and Gomphotherium were sig-nificantly different (Tukey test: p 5 0.031).
Mean 87Sr/86Sr ratios were highest for ma-rine mammals (0.70854 6 0.00014), interme-diate for Desmostylus (0.70826 6 0.00048), andlowest for terrestrial mammals (0.70783 60.00067) (Table 2, Fig. 4C). Median 87Sr/86Srvalues differed significantly among groups(Kruskal-Wallis: H 5 10.462, p , 0.01), butpairwise comparison showed that only terres-trial and marine mammals were statisticallydistinct (Dunn’s method: p , 0.05). Differenc-es in 87Sr/86Sr variance were statistically sig-nificant for Desmostylus versus marine mam-mals (F-test: p , 0.01) and marine mammalsversus terrestrial mammals (F-test: p , 0.01),but not for Desmostylus versus terrestrial
mammals (F-test: p 5 0.237). Comparison ofmean 87Sr/86Sr ratios among marine mammals(t-test: t 5 1.358, p 5 0.202) and terrestrialmammals (one-factor ANOVA: F 5 0.011, p 50.989) revealed no statistically significant dif-ferences.
Desmostylus is the only taxon that occurs atenough sites to warrant examination of geo-graphic differences. Mean d13C values weresignificantly different among sites (one-factorANOVA: F 5 6.09, p , 0.05) with the highestvalues reported at northern sites V3301 andV3215. Mean d18O values were, likewise, sig-nificantly different among sites (one-factorANOVA: F 5 7.73, p , 0.05). Pairwise com-parisons revealed that the mean for Desmos-tylus from site 3215 was significantly lowerthan mean d18O values from all other sites ex-cept site 2124, and that mean values for sitesV5555 and 2124 were significantly different(Tukey test: p , 0.05). d18O variance at sites1292 and V5555 were significantly differentfrom variance at sites 2124 and V3301. No sta-tistically significant differences were detectedfor mean 87Sr/86Sr values (Kruskal-Wallis: H 55.67, p 5 0.129).
Discussion
How Committed Was Desmostylus to AquaticHabitats? Most ecological interpretations forDesmostylus favor a connection with aquatichabitats, but the amount of time that thismammal actually spent in the water remainscontested. Possible scenarios include that Des-mostylus was fully aquatic, Desmostylus wassemiaquatic and foraged onshore, or Desmos-tylus was semiaquatic and foraged in the wa-ter. Our proxy for extent of adaptation toaquatic life is population-level variability ind18O values, which we used after validation byanalysis of fully aquatic and fully terrestrialfossil taxa.
As expected from studies of modern spe-cies, the standard deviation of d18O values forfossil fully aquatic mammal populations (ofsufficient sample size, i.e., n $ 5) was #0.5‰,significantly lower than the values for fossilfully terrestrial mammal populations (1s $0.8‰) (Table 2). The 1s values for fossil fullyaquatic and fully terrestrial mammals werecomparable to those for modern mammal
515HABITAT AND DIETARY PREFERENCES OF THE EXTINCT TETHYTHERE DESMOSTYLUS
populations (Bocherens et al. 1996; Clementzand Koch 2001). We conclude that diageneticalteration has not erased the in vivo differenc-es in d18O values among mammals at thesesites. Because Desmostylus has thicker enamelthan any of the marine and terrestrial mam-mals, it is even less likely to be subject to dia-genetic homogenization of d18O values.
Desmostylus populations had 1s valuesranging from 0.2‰ to 0.7‰ (Table 2, Fig. 4B).These values are lower than those for any ofthe fossil terrestrial mammals and for all butone of the modern terrestrial mammals (Table1). Furthermore, these 1s values are signifi-cantly lower than values for semiaquatic, ter-restrial-foraging hippopotamus from Ambo-seli (Table 1), suggesting that though desmos-tylians and hippopotamids were similar inbody size and basic morphology, the nichesoccupied by these taxa were quite distinct.Low d18O variability, particularly 1s valuesnear 0.2 or 0.3‰, provides strong support forthe conclusion that Desmostylus was either asemiaquatic mammal feeding in water, or afully aquatic mammal.
However, one population (V5555) had a 1svalue higher than that for modern fully ter-restrial Amboseli elephants, which suggeststhat the low d18O variability in Desmostylusmay simply be a result of large body size. Thishypothesis is unlikely for two reasons. First,Gomphotherium is larger than Desmostylus, yetat the site where these two species co-occur,the difference in d18O variation is extreme (Ta-ble 2). An additional factor must be generatingthe low variability in Desmostylus. Further-more no modern fully terrestrial mammal hasyielded d18O variability as low as that reportedfor Desmostylus at the majority of sites. Thus,we stand by the conclusion that Desmostyluswas either semiaquatic and feeding in water,or fully aquatic. The slightly higher 1s valueat site V5555 could therefore imply use of anestuary with significant d18O variability.
What Kinds of Aquatic Habitats Did Desmos-tylus Frequent? Though the occurrence ofDesmostylus remains in nearshore marine de-posits suggests an affinity for marine ecosys-tems, the possibility of transport prior to de-position means that we can’t rule out otheraquatic environments (i.e., estuarine and
freshwater ecosystems) as potential habitats.Patterns in mean and variability in d18O and87Sr/86Sr values can be useful for discriminat-ing among these alternatives, even though wewere unable to include clear isotopic end-members of freshwater and estuarine fossilaquatic mammals in our study. For d18O val-ues, we would expect marine species to havehigh mean values and low variability, estua-rine environments to have lower mean valuesand higher variability, and freshwater envi-ronments to yield the lowest mean values andlow variability. For 87Sr/86Sr values, we wouldexpect values for marine species to clusternear values calculated for middle Mioceneseawater (Fig. 4C), whereas values for fresh-water and estuarine species should be morevariable.
However, mean d18O values in modern eco-systems were not always diagnostic of fresh-water, let alone marine or estuarine, habitats(Table 1), and we observed a similar problemfor our fossil samples (Table 2, Fig. 4B), whichonly yielded a 2‰ range in mean values forall taxa. In addition, the mean d18O values formarine mammals from our sample sites arehigher than expected from estimated middleMiocene seawater d18O values (0.5‰ belowmodern seawater d18O (J. Zachos personalcommunication); Fig. 4B). However, the highd18O values we have reported for our marinemammals may reflect the regional hydrologicconditions of the basin in which these animalslived (Fig. 3). The fossils we sampled were alldeposited within a restricted basin, whichmay have experienced significant evaporationand limited exchange with open-ocean wa-ters. If so, d18O values for marine waters in thisbasin may have been enriched relative to meanseawater and could account for the high d18Ovalues we have reported.
From the 87Sr/86Sr values of the tooth enam-el, we were able to detect a clearer separationbetween terrestrial and marine mammals.Marine taxa had little variation in 87Sr/86Srvalues (1s , 0.0002‰) when compared withterrestrial mammals (1s . 0.0006‰) (Table 2,Fig. 4A). As with the d18O data, the 87Sr/86Srvalues we collected for marine mammals arenot expected, according to the calculatedrange in 87Sr/86Sr values for seawater at this
516 MARK T. CLEMENTZ ET AL.
time (0.70876 to 0.70877) (Hodell et al. 1991)(Fig. 4C). Again, the hydrology of the basinmay explain these low 87Sr/86Sr values. Withrestricted flow to and from the ocean, terres-trial inputs of 87Sr/86Sr from rivers could havelowered the mean 87Sr/86Sr value of waterswithin the basin.
Of most significance, however, is the largedifference in range of values between terres-trial and marine mammals. Desmostylus 87Sr/86Sr variation is greater than that calculated forany marine taxon and similar to the degree ofvariation observed for terrestrial taxa (Fig.4C). High variation in 87Sr/86Sr values wouldbe unlikely if Desmostylus was foraging onlywithin nearshore environments and suggeststhat it was spending a considerable amount oftime in estuarine or freshwater environments.
What Did Desmostylus Eat? From our pre-vious analyses, we have concluded that Des-mostylus was a fully aquatic or semiaquaticforager that was not restricted to marine hab-itats but foraged also in freshwater and estu-arine habitats. Our next step is to identify thefood webs within which Desmostylus was feed-ing—were they nearshore marine, kelp beds,seagrass beds, or freshwater food webs? To ex-plore this question, we used mean d13C valuesof tooth enamel as a proxy for dietary pref-erences, basing our interpretations on valuesfrom both modern and fossil taxa (Tables 1, 2,Fig. 4A).
Fossil terrestrial taxa have d13C values sim-ilar to those reported for middle Miocenemammals that fed on C3 vegetation elsewherein North and South America (MacFadden et al.1994; MacFadden and Cerling 1996). Our ma-rine species, including the early otariid Allo-desmus and two size classes of odontocetes,have been interpreted as nearshore, marineforagers on the basis of sedimentological andmorphological criteria (Barnes 1972; Dupras1985). As expected, mean d13C values for thesespecies were substantially more enriched in13C than values for modern offshore foragers(Table 1), but they were also slightly more en-riched than values for modern nearshore for-agers (Table 1). This offset from modern near-shore foragers may reflect ocean d13C valuesduring the middle Miocene, which were;1.5‰ higher than modern seawater values
(Zachos et al. 2001). With known marine andterrestrial end-members, we can now deter-mine the food webs contributing carbon toDesmostylus (Fig. 4A).
Mean d13C values for Desmostylus are signif-icantly higher than would be expected for aterrestrial C3 consumer and confirm that Des-mostylus must have been foraging within otherfood webs (Fig. 4A), a conclusion supportedby our d18O and 87Sr/86Sr evidence. Severalaquatic habitats are possible alternatives: ma-rine (nearshore, kelp, seagrass), estuarine(seagrass), and freshwater ecosystems. Iso-tope analysis of modern marine ecosystemshas found that nearshore consumers oftenhave higher d13C values than consumers lim-ited to terrestrial C3 resources (Hobson 1987;Bearhop et al. 1999). Seagrass, in particular,has d13C values that are often higher thanmodern C4 vegetation, with a mean d13C valueof ;211‰ (Hemminga and Mateo 1996). An-other potential resource would be marine al-gae, which also can have high d13C values (Ra-vens et al. 2002). In particular, the kelp groupof marine algae (Family Laminaria), which isbelieved to have arisen along the Pacific Coastat this time (Estes and Steinberg 1988), yieldsmean d13C values of ;217‰ (Raven et al.2002).
Each of these marine resources could pro-duce the high d13C values observed in Desmos-tylus, either via direct consumption or indi-rectly via consumption of other consumers inthese food webs. Incorporation of a multi-iso-tope analysis and inclusion of ecological con-trol taxa limit the possible scenarios (Fig.4B,C).
First, d13C values for Desmostylus are typi-cally higher than values reported for near-shore foraging marine mammals, suggestingthat Desmostylus was not foraging within thisecosystem. This interpretation is also con-firmed by the high 87Sr/86Sr variation reportedfor Desmostylus, which suggests it was notspending much time in a strictly marine en-vironment (Fig. 4C). The lack of a marine sig-nal in 87Sr/86Sr values for Desmostylus also al-lows us to exclude kelp beds as a potential di-etary resource for Desmostylus, given that kelpis an exclusively marine macrophyte.
Seagrass, on the other hand, grows within
517HABITAT AND DIETARY PREFERENCES OF THE EXTINCT TETHYTHERE DESMOSTYLUS
estuaries and can tolerate low salinity condi-tions (.20 ppt). Also, seagrasses growing inlow-salinity environments often exhibit ex-tremely low d13C values that are .1s from themean (,211‰). In these habitats, Desmosty-lus could also have foraged on other types ofestuarine or marsh vegetation (includingsome species of macroalgae, such as Ulva),which can exhibit a large range in d13C values.If Desmostylus was consuming a mixture offreshwater and estuarine vegetation or forag-ing on other consumers within these foodwebs, then the d13C and 87Sr/86Sr values seemmore reasonable.
Conclusion
By applying a multi-isotope approach, wehave been able to identify significant ecologi-cal differences between Desmostylus and bothterrestrial and marine mammals in terms ofaquatic affinity and dietary preferences, re-spectively. The high d13C values suggest thatDesmostylus foraged on seagrasses, probablyrooting up the vegetation to consume the car-bohydrate-enriched rhizomes that would haveformed dense mats within the shallow la-goons and estuaries along the Pacific coastline.In addition, Desmostylus was not limited toforaging on seagrasses but likely incorporateda wide range of freshwater and estuarineaquatic vegetation into its diet. In this way, theecology of Desmostylus was most similar tothat of modern manatees in Florida, whichseasonally forage on aquatic vegetation infreshwater and marine ecosystems. Unlikemanatees, Desmostylus was fully capable ofterrestrial locomotion and was probably sim-ilar to the modern hippopotamus in terms ofaquatic and terrestrial habits.
Thus, desmostylians were a unique groupof mammals with no real modern analogs interms of habitat preferences, diet, or locomo-tor capabilities. As the largest mammalianconsumers of coastal aquatic vegetation at thetime, their impact on coastal ecosystems musthave been substantial. In addition, becausethey are believed to be the sister group to pro-boscideans, their aquatic preferences raise in-teresting evolutionary questions. Were theaquatic habits of this group characteristic ofthe basal members of the clade to which both
desmostylians and proboscideans belong, ordid desmostylians evolve aquatic habits short-ly after their divergence? This issue may beaddressed by applying similar techniques tosamples of early proboscideans and basaltethytheres (e.g., anthracobunids).
Our results highlight the necessity of usingmultiple proxies to answer questions aboutthe ecology of extinct taxa. If we had limitedour interpretations to only one isotopic sys-tem, we would have generated substantiallydifferent conclusions about the ecology of Des-mostylus. In future work, we will incorporatemicrowear analysis into studies of other des-mostylian species to identify differences inforaging preferences among co-occurringtaxa, and to develop a method for testing ourinterpretations of the trophic level of these ex-tinct animals.
Acknowledgments
We would like to thank P. Holroyd at theUniversity of California Museum of Paleon-tology, L. Barnes at the Natural History Mu-seum of Los Angeles County, and T. Demereat the San Diego Natural History Museum forproviding access to specimens for analysis. Aspecial thank you goes to P. Holden for assist-ing in analysis of Sr samples and to D. Domn-ing for providing information on desmostyli-an phylogenetics. M.T.C. was supported by aNational Science Foundation (NSF) Predoctor-al Fellowship and Achievement Rewards forCollege Scientists Fellowship when much ofthis research was conducted. Analytical andtravel costs were covered by NSF grants EAR-9725854 and 0087742.
Literature Cited
Armstrong, R. L. 1971. Glacial erosion and the variable isotopiccomposition of strontium in seawater. Nature 230:132–133.
Aranda-Manteca, F. J., D. P. Domning, and L. G. Barnes. 1994. Anew middle Miocene sirenian of the genus Metaxytheriumfrom Baja California and California, relationships and paleo-biogeographic implications. In A. Berta and T. A. Demere,eds. Contributions in marine mammal paleontology honoringFrank C. Whitmore, Jr. Proceedings of the San Diego Societyof Natural History 9:191–204.
Barnes, L. G. 1972. Miocene Desmatophocinae (Mammalia: Car-nivora) from California. University of California Publicationsin Geological Sciences 89:1–68.
———. 1976. Outline of eastern North Pacific fossil cetacean as-semblages. Systematic Zoology 25:321–343.
Barnes, L. G., D. P. Domning, and C. E. Ray. 1985. Status of stud-
518 MARK T. CLEMENTZ ET AL.
ies on fossil marine mammals. Marine Mammal Science 1:15–53.
Bartow, J. A. 1987. Cenozoic evolution of the San Joaquin Valley,California. U.S. Geological Survey Report OF 87-0581.
Bearhop, S., D. R. Thompson, S. Waldron, I. C. Russell, G. Al-exander, and R. W. Furness. 1999. Stable isotopes indicate theextent of freshwater feeding by cormorants Phalacrocoraxcarbo shot at inland fisheries in England. Journal of AppliedEcology 36:75–84.
Bocherens, H., P. L. Koch, A. Mariotti, D. Geraads, and J. Jaeger.1996. Isotopic biogeochemistry (13C, 18O) of mammalianenamel from African Pleistocene hominid sites. Palaios 11:306–318.
Boon, P. L., and S. E. Bunn. 1994. Variations in the stable isotopecomposition of aquatic plants and their implications for foodweb analysis. Aquatic Botany 48:99–108.
Bowen, W. D. 1997. Role of marine mammals in aquatic ecosys-tems. Marine Ecology Progress Series 158:267–274.
Bryant, J. D., and P. N. Froelich. 1995. A model of oxygen isotopefractionation in body water of large mammals. Geochimica etCosmochimica Acta 59:4523–4537.
Bryant, J. D., D. S. Jones, and P. A. Mueller. 1995. Influence offreshwater flux on 87Sr/86Sr chronostratigraphy in marginalmarine environments and dating vertebrate and invertebratefaunas. Journal of Paleontology 69:1–6.
Capo, R. C., and D. J. DePaolo. 1990. Seawater strontium isotopicvariations from 2.5 million years ago to the present. Science249:51–55.
Capo, R. C., O. A. Chadwick, and D. M. Hendricks. 1994. Con-straining atmospheric inputs and in situ weathering in soilsdeveloped along a climate gradient using Sr isotopes. In M.A. Lanphere, G. B. Dalrymple, and B. D. Turin, eds. Abstractsof the Eighth International Conference on Geochronology,Cosmochronology, and Isotope Geology. U.S. Geological Sur-vey Circular 1107:1–47. Denver, Colo.
Capo, R. C., B. W. Stewart, and O. A. Chadwick. 1998. Strontiumisotopes as tracers of ecosystem processes: theory and meth-ods. Geoderma 82:197–225.
Cerling, T. E., and J. M. Harris. 1999. Carbon isotope fraction-ation between diet and bioapatite in ungulate mammals andimplications for ecological and paleoecological studies. Oec-ologia 120:347–363.
Clementz, M. T., and P. L. Koch. 2001. Differentiating aquaticmammal habitat and foraging ecology with stable isotopes intooth enamel. Oecologia 129:461–472.
Cole, M., and D. P. Domning. 1998. Locomotor and positionaladaptations in the desmostylian genera Desmostylus and Pa-leoparadoxia. Journal of Vertebrate Paleontology 18(Suppl. toNo. 3):35A.
Domning, D. P. 2001. Sirenians, seagrasses, and Cenozoic eco-logical change in the Caribbean. Palaeogeography, Palaeocli-matology, Palaeoecology 166:27–50.
Domning, D. P., C. E. Ray, and M. C. McKenna. 1986. Two newOligocene desmostylians and a discussion of Tethytheriansystematics. Smithsonian Contributions to Paleontology 59:1–56.
Domning, D. P., and H. Furusawa. 1992. Summary of taxa anddistribution of Sirenia in the North Pacific Ocean. Island Arc3:506–512.
Dupras, D. L. 1985. Life beneath the Temblor Sea: SharktoothHill, Kern County, California. California Geology 38:147–154.
Estes, J. A., and P. D. Steinberg. 1988. Predation, herbivory, andkelp evolution. Paleobiology 14:19–36.
Hemminga, M. A., and M. A. Mateo. 1996. Stable carbon iso-topes in seagrasses: variability in ratios and use in ecologicalstudies. Marine Ecology Progress Series 140:285–298.
Hobson, K. A. 1987. Use of stable-carbon isotope analysis to es-
timate marine and terrestrial protein content in gull diets. Ca-nadian Journal of Zoology 65:1210–1213.
Hodell, D. A., P. A. Mueller, and J. R. Garrido. 1991. Variationsin the strontium isotopic composition of seawater during theNeogene. Geology 19:24–27.
Hoppe, K. A., P. L. Koch, R. W. Carlson, and S. D. Webb. 1999.Tracking mammoths and mastodons: reconstruction of mi-gratory behavior using strontium isotope ratios. Geology 27:439–442.
Ingram, B. L., and P. K. Weber, 1999. Salmon origin in Califor-nia’s Sacramento-San Joaquin river system as determined byotolith strontium isotopic composition. Geology 27:851–854.
Inuzuka, N. 1984. Skeletal restoration of the desmostylians: her-petiform mammals. Memoirs of the Faculty of Science, KyotoUniversity, Series of Biology IX:157–253.
Inuzuka, N. 2000. Primitive Late Oligocene desmostylians fromJapan and phylogeny of the Desmostylia. Bulletin of AshoroMuseum of Paleontology: 91–123.
Inuzuka, N., D. P. Domning, and C. E. Ray. 1995. Summary oftaxa and morphological adaptations of the Desmostylia. Is-land Arc 3:522–537.
Kennedy, B. P., C. L. Folt, J. D. Blum, and C. P. Chamberlain.1997. Natural isotope markers in salmon. Nature 387:766–767.
Kennedy, M. J., O. A. Chadwick, P. M. Vitousek, L. A. Derry, andD. M. Hendricks. 1998. Changing sources of base cations dur-ing ecosystem development, Hawaiian Islands. Geology 28:1015–1018.
Koch, P. L., A. N. Halliday, L. M. Walter, R. F. Stearly, T. J. Hus-ton, and G. R. Smith. 1992. Sr isotopic composition of hy-droxyapatite from recent and fossil salmon: the record of life-time migration and diagenesis. Earth and Planetary ScienceLetters 108:277–287.
Koch, P. L., J. Heisinger, C. Moss, R. W. Carlson, M. L. Fogel, andA. K. Behrensmeyer. 1995. Isotopic tracking of change in dietand habitat use in African Elephants. Science 267:1340–1343.
Koch, P. L., N. Tuross, and M. L. Fogel. 1997. The effects of sam-ple treatment and diagenesis on the isotopic integrity of car-bonate in biogenic hydroxylapatite. Journal of ArchaeologicalScience 24:417–429.
Kohn, M. J. 1996. Predicting animal d18O: accounting for dietand physiological adaptation. Geochimica et CosmochimicaActa 60:4811–4829.
Lee-Thorp, J. A. 2000. Preservation of biogenic carbon isotopicsignals in Plio- Pleistocene bone and tooth mineral. In S. H.Ambrose and K. A. Katzenberg, eds. Biogeochemical ap-proaches to paleodietary analysis. Advances in archaeologicalScience 5:89–116. Plenum, New York.
Lee-Thorp, J. A., J. C. Sealy, and N. J. van der Merwe. 1989. Stablecarbon isotope ratio differences between bone collagen andbone apatite, and their relationship to diet. Journal of Ar-chaeological Science 16:585–599.
Longinelli, A. 1984. Oxygen isotopes in mammal bone phos-phate; a new tool for paleohydrological and paleoclimatolog-ical research? Geochimica et Cosmochimica Acta 48:385–390.
Luz, B., and Y. Kolodny. 1985. Oxygen isotope variations inphosphate of biogenic apatites. IV. Mammal teeth and bones.Earth and Planetary Science Letters 75:29–36.
Luz, B., Y. Kolodny, and M. Horowitz. 1984. Fractionation of ox-ygen isotopes between mammalian bone-phosphate and en-vironmental drinking water. Geochimica et CosmochimicaActa 48:1689–1693.
MacFadden, B. J., and T. E. Cerling. 1996. Mammalian herbivorecommunities, ancient feeding ecology, and carbon isotopes: a10 million-year sequence from the Neogene of Florida. Journalof Vertebrate Paleontology 16:103–115.
MacFadden, B. J., Y. Wang, T. E. Cerling, and F. Anaya. 1994.South American fossil mammals and carbon isotopes: a 25
519HABITAT AND DIETARY PREFERENCES OF THE EXTINCT TETHYTHERE DESMOSTYLUS
million-year sequence from the Bolivian Andes. Palaeogeog-raphy Palaeoclimatology Palaeoecology 107:257–268.
McLeod, S. A., and L. G. Barnes. 1984. Fossil desmostylians. Pp.39–44 in B. Butler, J. Grant, and C. J. Stadum, eds. The naturalscience of Orange County. Memoirs of the Natural HistoryFoundation of Orange County. Natural History Foundation ofOrange County, Newport Beach, Calif.
Miller, E. K., J. D. Blum, and A. J. Friedland. 1993. Determinationof soil exchangeable-cation loss rates using Sr isotopes. Na-ture 362:438–441.
Muhs, D. R., C. A. Bush, R. R. Tracy, and C. C. Stewart. 1990.Geochemical evidence of Saharan dust parent material forsoils developed on Quaternary limestones of Caribbean andwestern Atlantic islands. Quaternary Research 33:157–177.
Nelson, B. K., M. J. DeNiro, M. J. Schoeninger, D. J. DePaolo, andP. E. Hare. 1986. Effects of diagenesis on strontium, carbon,nitrogen, and oxygen concentration and isotopic compositionof bone. Geochimica et Cosmochimica Acta 50:1941–1949.
Novacek, M. J., and A. R. Wyss. 1987. Selected features of thedesmostylian skeleton and their phylogenetic implications.American Museum of Novitates 2870:1–8.
O’Leary, M. H. 1988. Carbon isotopes in photosynthesis. Bio-science 38:328–336.
Osmond, C. B., N. Valaane, S. M. Haslam, P. Uotila, and Z. Rok-sandic. 1981. Comparisons of d13C values in leaves of aquaticmacrophytes from different habitats in Britain and Finland:some implications for photosynthetic processes in aquaticplants. Oecologia 50:117–124.
Peterken, C. J., and C. A. Conacher. 1997. Seed germination andrecolonisation of Zostera capricorni after grazing by dugongs.Aquatic Botany 59:333–340.
Rau, G. H., F. P. Chavez, and G. E. Friederich. 2001. Plankton13C/12C variations in Monterey Bay, California: evidence of
non-diffusive inorganic carbon uptake by phytoplankton inan upwelling environment. Deep-Sea Research I 48:79–94.
Ravens, J. A., A. M. Johnston, J. E. Kubler, R. Korb, S. G. McInroy,L. L. Handley, C. M. Scrimgeour, D. I. Walker, J. Beardall, M.Vanderklift, S. Fredriksen, and K. H. Dunton. 2002. Mecha-nistic interpretation of carbon isotope discrimination by ma-rine macroalgae and seagrasses. Functional Plant Biology 29:355–378.
Repenning, C. A. 1965. Stanford fossil; studied by U.S.G.S. Min-eral Information Service 18:124–125.
Roe, L. J., J. G. M. Thewissen, J. Quade, J. R. O’Neil, S. Bajpai, A.Sahmi, and S. T. Hussain. 1998. Isotopic approaches to un-derstanding the terrestrial-to-marine transition of the earliestcetaceans. Pp. 399–422 in J. G. M. Thewissen, ed. The emer-gence of whales. Plenum, New York.
Schmitz, B., S. L. Ingram, D. T. Dockery III, and G. Aberg. 1997.Testing 87Sr/86Sr as a paleosalinity indicator on mixed marine,brackish-water and terrestrial vertebrate skeletal apatite inlate Paleocene–early Eocene near-coastal sediments, Missis-sippi. Chemical Geology 140:275–287.
Schwarcz, H. P., and M. J. Schoeninger. 1991. Stable isotope anal-yses in human nutritional ecology. Yearbook of Physical An-thropology 34:283–322.
Vennemann, T. W., E. Hegner, G. Cliff, and G. W. Benz. 2001. Iso-topic composition of recent shark teeth as a proxy for envi-ronmental conditions. Geochimica et Cosmochimica Acta 65:1583–1599.
Wang, Y., and T. E. Cerling. 1994. A model of fossil tooth andbone diagenesis: implications for paleodiet reconstructionfrom stable isotopes. Palaeogeography Palaeoclimatology Pa-laeoecology 107:281–289.
Zachos, J., M. Pagani, L. Sloan, E. Thomas, and K. Billups. 2001.Trends, rhythms, and aberrations in global climate 65 Ma topresent. Science 292:686–693.
