Author's personal copyJournal of Archaeological Science 37 (2010) 317–327Contents lists available at ScienceDirectJournal of Archaeological Sciencejournal homepage: http://www.elsevier.com/locate/jasThe fish of Lake Titicaca: implications for archaeology and changing ecologythrough stable isotope analysisMelanie J. Miller a, José M. Capriles b, Christine A. Hastorf a, *abDepartment of Anthropology, University of California at Berkeley, 232 Kroeber Hall, Berkeley, CA 94720-3710, USADepartment of Anthropology, Washington University in St. Louis, One Brookings Drive C.B. 1114, St. Louis, MO 63130, USAa r t i c l e i n f oa b s t r a c tArticle history:Received 23 May 2009Received in revised form18 September 2009Accepted 22 September 2009Research on past human diets in the southern Lake Titicaca Basin has directed us to investigate thecarbon and nitrogen stable isotopes of an important dietary element, fish. By completing a range ofanalyses on modern and archaeological fish remains, we contribute to two related issues regarding theapplication of stable isotope analysis of archaeological fish remains and in turn their place within humandiet. The first issue is the potential carbon and nitrogen isotope values of prehistoric fish (and how thesewould impact human dietary isotopic data), and the second is the observed changes in the fish isotopesthrough time. Out of this work we provide quantitative isotope relationships between fish tissues withand without lipid extraction, and a qualitative analysis of the isotopic relationships between fish tissues,allowing archaeologists to understand these relationships and how these values can be applied in futureresearch. We test a mathematical lipid normalization equation to examine whether future researcherswill need to perform lipid extraction procedures for Lake Titicaca fish. We also analyze a number ofaquatic plants to better understand the range of isotopic signatures of the Lake Titicaca ecosystem. Weuse these data to better understand prehistoric human diet and the role that fish may have played in thepast as well as potential changes in local lake ecology through time.! 2009 Published by Elsevier Ltd.Keywords:Prehistoric fish usePaleoecology of Lake TiticacaSouth AmericaCarbonNitrogenStable isotopesArchaeologyLipid extraction1. IntroductionThis study grew out of the rich yet complex study of prehistoricdiets, including stable isotope data from cooking pots, plants,animals and human teeth that have been collected by the TaracoArchaeological Project working in the Titicaca Basin of Bolivia. Herewe present stable isotope analysis of the archaeological fishsamples to understand their role in the diet of the Formative Periodinhabitants of the southern Lake Titicaca Basin. Carbon andnitrogen stable isotope analysis has become very useful in a varietyof dietary studies (Ambrose, 1993; Finlay et al., 2002; Schoeningerand DeNiro, 1984), and ancient human diets and their changethrough time (Ambrose et al., 1998; Hastorf, 1991; Katzenberg et al.,1995; Prowse et al., 2004; Richards et al., 2003; Rutgers et al., 2009;Schwarcz et al., 1985; Van Klinken et al., 2000; White et al., 1999).However, in order to place the fish in human history, it is necessaryto learn about the relationships and ecology of the fish from theirmuscle, bone and scales, since muscle is rarely preserved inarchaeological contexts, whereas bone and scales are. We* Corresponding author.E-mail address: [email protected] (C.A. Hastorf).0305-4403/ – see front matter ! 2009 Published by Elsevier Ltd.doi:10.1016/j.jas.2009.09.043investigate modern fish specimens from Lake Titicaca to comparewith our archaeological analysis. Studying fish through isotopicanalysis introduces an additional layer, as recent research hashighlighted the important role that lipid extraction plays inretrieving correct d13C signatures from samples that contain lipids(Kojadinovic et al., 2008; Mintenbeck et al., 2008; Post et al., 2007;Sotiropoulos et al., 2004; Sweeting et al., 2006).In this paper we contribute to the discussion of two relatedissues regarding the application of stable isotope analysis toarchaeological fish remains and in turn their place within humandiet. The first issue is the interpretation of the carbon and nitrogenisotopic values of prehistoric fish (and how these impact humandietary values), and the second is the lake–fish ecological relationship. By understanding the isotopic compositions of the fishwithin the larger ecology of Lake Titicaca we can better understandthe human interactions in this rich and diverse region as well as theroles that fish played in human life on the Taraco Peninsula duringtheir first 1000 years of living in settled communities. To addressthese two questions, we first examine modern fish and the effectsof lipid extraction on carbon and nitrogen isotope values. Then weturn to the isotopic relationships between the muscle, bone, andscale tissues of modern fish to learn of their variance. The fish andaquatic plants of Lake Titicaca have not previously been analyzed
Author's personal copy318M.J. Miller et al. / Journal of Archaeological Science 37 (2010) 317–327for their isotope compositions and previous studies of aquaticenvironments show large variation in the isotopic ranges observed(Chisholm et al., 1982; Dufour et al., 1999; Fischer et al., 2007;France, 1995; Katzenberg and Weber, 1999; Mintenbeck et al., 2008;Sotiropoulos et al., 2004; Sweeting et al., 2006). Other studies ofmodern and archaeological fish species from lacustrine environments have reported a large range of carbon values. France (1995)reported pelagic consumers to range from d13C !38 to !26& andlittoral consumers to have d13C values from !32 to !16&. Katzenberg and Weber (1999) report fish from Lake Bikal to have d13Cvalues ranging from !24.6 to !12.9&. Dufour et al. (1999) reportd13C values for 3 European lakes and Lake Bikal to range from !32.5to !19.8&. These carbon values span such a large range that it wasclear to us that Lake Titicaca fish need to be investigated in theirown right. Finally, we compare modern and archaeological organiccarbon and nitrogen isotopes of the fish, providing new information in the foodways scholarship where fish are a potentialcomponent.Lake Titicaca is located at an elevation of 3810 m above sea levelin the south central Andes of South America. The lake coversa surface area of approximately 8400 km2 and is divided in twounequal parts; the northern portion, known as Chucuito is largerand deeper than the southern portion, known as Wiñaymarca(Fig. 1). The Taraco Peninsula, our study area, is located in thesoutheastern portion of Wiñaymarca. Due to its overall shallowness, Wiñaymarca can support higher biomass densities than thenorthern portion, but it is also more vulnerable to climatic andenvironmental changes (Abbott et al., 1997; Baker et al., 2005,2009; Binford et al., 1997; Calaway, 2005).Faunal resources are readily available on the shoreline of theTaraco Peninsula and include over 50 species of small to largeaquatic and terrestrial birds, two dozen species of fish, and a fewspecies of frogs and toads (Dejoux and Iltis, 1992; Kent et al., 1999;Levieil and Orlove, 1990; Orlove 2002; Portugal Loayza, 2002;Steadman and Hastorf, in press;). In addition, readily availablewater from a number of streams and associated wetlands (knownas bofedales) are ideal habitats for terrestrial vertebrates, includingcamelids, deer, and several rodents.Two genera of fish (Orestias and Trichomycterus), comprisingapproximately 26 species, have been documented to live in LakeTiticaca (Lauzanne, 1992; Parenti, 1984; Sarmiento and Barrera,2004). The genus Orestias is composed of a wide variety of speciesmost of which are small, rarely exceeding 5 cm in length, someranging to just above 20 cm of standard length, exhibit high geneticdiversity, and are specialized to specific microhabitats within thelake (Lüssen et al., 2003; Parenti, 1984; Vaux et al., 1988). Twospecies of bottom-dweller filter feeder Trichomycterus have beendescribed for Lake Titicaca: T. dispar and T. rivulatus and they rangein size from 12 to 19 cm standard length (Fernández and Vari, 2009;Sarmiento and Barrera, 2003).Orestias have been known to consume a broad spectrum ofaquatic resources including algae, macrophytes, zooplankton,amphipods, ostracods, insects, and insect larvae (Lauzanne, 1992).Their diet is constrained by a number of factors including species,ontogeny, size, availability, and degree of specialization (Vaux et al.,1988). Lauzanne (1992) among other ichthyologists have suggestedthat Orestias specimens tend to eat more fauna as they get larger.Trichomycterus are filter feeders ingesting organic remains inFig. 1. Map of the region with sites mentioned in text plotted using ArcGIS 9.2. The base image is a Landsat L7 satellite image type Enhanced Thematic Mapper Plus (ETMþ)panchromatic Band 8.
Author's personal copyM.J. Miller et al. / Journal of Archaeological Science 37 (2010) 317–327addition to a number of different autochthonous benthicmacroinvertebrates.In this paper we present a series of analyses that help usinterpret the place of fish in past Lake Titicaca basin human diet.Our analyses demonstrate that a mathematical correction canaccurately predict meaningful modern and archaeological carbonisotope values without completing lipid extraction procedures,saving researchers time and money. Comparison of fish tissues’isotope values provides archaeologists with an important relationship between hard and soft tissue remains, allowing for interpretation of fish muscle’s contribution to human diets from bone orscale. We also confirm that the fish of Lake Titicaca have carbonisotope values that overlap those of maize, which is a significantfactor for archaeologists studying past foodways and diet. Finally,our analyses suggest that there has been a significant shift betweenmodern and archaeological fish carbon isotope values and wesuggest a number of reasons for this finding.2. Archaeological evidence for fish use in Lake TiticacaArchaeologists working in the Lake Titicaca Basin agree thataquatic resources played a significant role in human diet andeconomy as people began settling on the landscape during the EarlyFormative Period (2000–800 BCE). Unfortunately, few studies haveexplicitly focused on building baseline research to understand thecomplex dynamic associated with the exploitation and consumption of aquatic resources (Capriles, 2006).Our recent faunal research based on hundreds of weight andNISP counts from heavy fraction flotation samples recovered fromthe Taraco Archaeological Project excavations suggests fish remainswere consistently important in the economy of the Taraco residentsbetween the Early Formative and Tiwanaku periods with a subtledecrease over time (Bruno, 2008; Capriles et al., 2008; Moore et al.,1999). Fish consumption is also evident in the recovery of severaldozen bone tools used for the manufacture of fishing nets as well asin specialized cooking features recorded as ‘‘fish pits’’ wherethousands of fish bones were systematically processed and/orburied (Moore and Hastorf, 2000; Moore et al., 1999).319a distinctive character of O. luteus. Large and medium (largerthan 5 mm) scales most probably correspond to O. agassii andO. pentlandii. The analyzed operculum bones were specificallyselected due to their better preservation. These bones weremeasured using digital calipers and, using a transfer function,we estimated their live body size to be between 85 and 221 cm.The archaeological fish seem to have been bigger overall thanthe modern specimens, suggesting that the prehistoric fishcaught were more diverse and grew to an older age than modernfishing allows (Capriles, 2006).3.2. Modern fish: collection proceduresModern Orestias and Trichomycterus fish samples were collectedfrom Wiñaymarca Lake by local fishermen from the community ofSan Jose on the peninsula in 2005 and 2008. The Orestias ispisamples were fished from Lake Titicaca and purchased froma market in La Paz in 2005 and 2008. Bones, scales, and musclewere separated and dried in Bolivia and were then mechanicallycleaned and freeze dried using a Labconco in the Center for StableIsotope Biogeochemistry (CSIB) at the University of CaliforniaBerkeley. Samples were homogenized using a Wig-L Bug or anagate mortar and pestle, and were stored in sealed test tubes untilanalyzed or treated.3.3. Modern aquatic plants: collection proceduresModern aquatic plant samples were collected from the lake byfishermen from Santa Rosa on the Taraco Peninsula. The collectedplants were air and sun-dried in Bolivia before export. Additionalimportant Lake Titicaca aquatic plant taxa not obtained by thefisherman were sampled from collections curated in the HerbarioNacional de Bolivia (LPB) in La Paz. Plants were imported to the USAand freeze dried at the CSIB at the University of California-Berkeley.Samples were homogenized using a Wig-L Bug and were stored insealed tubes until analyzed. Carbon and nitrogen isotopes weremeasured together on the same plant sample.3. Materials and methods3.4. Stable isotope analysis and removal of lipids3.1. Archaeological fish: collection proceduresThe stable isotope ratio data are reported in standard delta (d)notation as parts per thousand (per mil, &), with results reportedrelative to the Pee Dee Belemnite standard for d13C, and atmospheric nitrogen for d15N.The presence of lipids in fish tissues alters their isotopiccomposition and causes depleted (more negative) d13C values.Lipids lack nitrogen atoms and are depleted in 13C, compared toother biochemical components such as proteins (DeNiro andEpstein, 1977; Sweeting et al., 2006). The amount a d13C valuealters depends on the lipid content of the tissue and has provento be variable across different animals, with recent studies unableto show a standard shift that can be uniformly applied to all fishsample