Reconstructing freshwater fishing seasonality in a neotropical savanna: First application of swamp eel (Synbranchus marmoratus) sclerochronology to a pre-Columbian Amazonian site (Loma Salvatierra, Bolivia)

Highlights

• Reconstruction of season of occupation of archaeological sites based on the study of fish growth structures.

• The first time that sclerochronology has been applied to a freshwater species in archaeology.

• The seasonal growth of the fish is related to the hydrological cycle where fast growth period coincides with rainfall in the region.

• Archaeological vertebrae demonstrate that fishing of swamp eels might have occurred all year-round.

Abstract

Sclerochronology is a method used to estimate the season of death (season of capture) of archaeological individuals based on a modern growth model. This method has been increasingly accepted in South America and has mainly been applied to coastal archaeological sites (on the Atlantic and Pacific Oceans). This is the first time that this method has been applied to a freshwater species, the marbled swamp eel (Synbranchus marmoratus), in archaeology. Excavations undertaken at Loma Salvatierra, a human-built platform located in the Bolivian Amazon and occupied from 500 until 1400 CE, have yielded 111 zooarchaeological vertebrae of the marbled swamp eel, which is one of the most widely distributed species recovered in South American continental archaeological sites. In order to estimate the fishing season for these archaeological individuals, we developed a modern osteological reference collection, made up of 61 specimens with known capture dates sampled monthly over a one-year period, about 60 km from Loma Salvatierra. The vertebrae present periodic growth patterns with a succession of dark and light bands alternately. Consequently, the vertebrae are a reliable basis for the estimation of the marbled swamp eel fishing season. The analysis of the marginal increments of vertebrae in present-day fish allowed us to elaborate a modern growth model showing that the seasonal growth of the marbled swamp eel is related to the hydrological cycle, whereby the fast growth period coincides with the onset of rainfall in the region. On the basis of this modern-based model, the analysis of zooarchaeological vertebrae demonstrates that fish were captured over several seasons. Demonstrating that human groups occupied villages year-round does not mean that they were not mobile but shows year-round fishing in the savanna. This year-round fishing practice raises questions concerning the generalized idea of fishing as an exclusively dry-season activity. As wild resources are generally seasonal, the evidence of the year-round fishing of swamp eels might suggest year-round fishing at Loma Salvatierra and contributes to the understanding of late-Holocene mobility patterns in pre-Columbian times.

Introduction

One of the most accurate methods to estimate the occupation season of riverine or coastal archaeological sites is the study of fish or mollusk growth structures, which allows for an estimation of the ontogenetic age, growth and season of death (capture period) of archaeological individuals (Casteel, 1974, Casteel, 1972, Desse and Desse-Berset, 1992, Guillaud et al., 2017, Van Neer, 1993). In American archaeology, such studies have been increasingly applied to coastal environments (Hales and Reitz, 1992, Scartascini et al., 2015, Torres, 2016, Torres et al., 2020). However, very few studies have attempted to apply this approach to freshwater contexts (Cahiza, 2003, Svoboda, 2013). Despite the increasing interest in age determination and seasonality markers of modern freshwater tropical species, the analysis of growth-increment structures has been underexploited in South American archaeological contexts (Prestes-Carneiro et al., 2019, Prestes-Carneiro et al., 2020).

Fishing was claimed to be a seasonal activity by the first scholars who studied the subsistence strategies adopted by pre-Columbian Amazonian groups (Moran, 1993, Roosevelt, 1980). Due to the scarcity of seasonality indicators, archaeological research carried out in the Amazon region frequently uses indirect evidence to indicate the seasonality of occupations. For example, Schaan (2008) carried out investigations within sites of the Amazon River estuary on Marajo Island that revealed a system of artificial temporary ponds that were filled by higher water levels during the rainy season. Associated overbank flooding led the author to suggest that fishing was mainly a dry-season activity. Similarly, in the Venezuelan Llanos, where Garson (1980) described systems of water-retention dikes and ponds across extensive savannas, dry-season fishing was suggested by aspects of fish ecology, such as the size of recovered fish communities and the dominance of drought-tolerant fish. In the absence of biological evidence, the dry season character of fishing in both examples was only supported by indirect corroboration of periodicity.

In the 1950́s and 1960́s, the “sedentary versus nomadic” character of the strategies adopted by pre-Columbian Amazonian groups fueled a lively debate among the first archaeologists working in the region (Lathrap, 1968, Meggers, 1954). Around CE 1000 several regions of the Amazon were occupied by large societies with villages. There is evidence of monumentality in the form of earthworks, such as the roads that connected villages in the Xingú region, the mounds, defensive ditches, and circular villages in the central Amazon, the great extensions of black earth in the Santarém region, and the monumental platforms in the southwestern Amazon (Heckenberger et al., 2003, Neves et al., 2004, Prümers, 2004, Moraes and Neves, 2012). It is, however, difficult to understand the seasonality of resource use and occupations of these groups since the seasonality of past occupations is commonly inferred from density and extent of the archaeological settlements.

These suggestions may be refined by direct study of archaeological remains (such as biorecorders) that contain markers of seasonality. Sclerochronology addresses ontogenetic age estimation, which refers to the time elapsed after birth, and growth, and can be used to explore a range of other themes such as the impact of human action on local ecology. This term is derived from the Greek, sclêros “hard”, chronos “time”, and logos “study” that together, mean the estimation of time by way of the growth marks on organisms’ calcified structures. Sclerochronology is analogous to dendrochronology (the study of tree rings) in postulating that, as an animal grows, skeletal hard tissue records changes that are linked to environmental conditions and physiological processes. Thus, sclerochronology permits investigation of life-histories as well as environmental and climatic change across space and time. In calcified tissues, growth patterns can be seen under a reflected light, as a succession of alternately dark and light bands. A light band (“zone”) represents a fast-growing period and a dark band (“annulus”) represents a slowdown in growth (see Fig. 6). In many taxa, a couplet of one light and one dark band typically represents one annual growth cycle, although this may differ between tissue types (e.g., bone, otoliths, teeth, scales, vertebrae) (Beamish and McFarlane, 1983, Quitmyer et al., 1997, Van Neer et al., 1999, Andrus, 2011, Meunier, 2012, Vitale et al., 2019).

The presence of an annulus does not imply that an animal stopped growing, but it indicates a decrease in growth rate and metabolic activity (Quitmyer et al., 1997, Panfili et al., 2002). The deposition of calcified tissue (increments) is commonly driven by factors related to temperature, such as seasonal contrasts, and the availability of water (Andrus, 2011). There are, however, many other potential factors related to internal rhythms (reproduction, sex reversal, migration, maturation) or external conditions (lack of food or water, water quality, salinity) that can affect growth and thus the increments formed. As such, sclerochronology can be used to study anthropogenic or naturally occurring impacts on natural resources and anthropogenic impacts on the local ecology (Campana, 1990, Panfili et al., 2002, Schone, 2008, Vitale et al., 2019).

Since the vast majority of fish are poikilothermic (organisms that do not control their internal temperatures), they are sensitive to external environmental variations, and are therefore potential bio-indicators of seasonal fluctuations. The increment for the last period of growth is used to estimate the season of death. Its measurement is made from the last annulus to the margin of the bone or otolith, which shows how much the individual has grown and lived since the last formed annulus (Casteel, 1972, Mahé et al., 2009, Torres et al., 2020).

As each type of calcified structure has a particular chemical composition and biomineralization process, specific studies are necessary to observe how each calcified structure records growth (Castanet et al., 1992, Panfili et al., 2002). The great majority of studies applying the sclerochronological approach to archaeological contexts are based on bivalve shells and fish otoliths (Andrus, 2011, Carré et al., 2005, Casteel, 1974, Hales and Reitz, 1992, Van Neer et al., 2004). Sclerochronological studies based on scale, pectoral spine, opercle, cleithrum, and vertebra are also possible, depending on the species and on the condition that increments are visible, regular and readable, although these studies are less common (Brewer, 1987, Desse and Desse-Berset, 1992, Morey, 1983, Guillaud et al., 2017).

Over the past decades, great efforts have been made to document the age and growth of South American freshwater fish based on calcified structures (Boujard et al., 1991, Cutrim and Batista, 2005, Fabré and Saint-Paul, 1998, Martins et al., 2009, Mateus and Petrere, 2004, Santos and Barbieri, 1993, Ponton et al., 2001). Baseline or proxy studies of modern living taxa need to be conducted to establish a model of growth relevant to the location of the archaeological site (Deith, 1983, Quitmyer et al., 1997, Butler et al., 2019). Archaeological studies also require good preservation of the bone material, exhaustive sampling of archaeological remains using fine screen recovery, and modern comparative collections (Casteel, 1972, Rojo, 1988, Deith, 1983). The large majority of sclerochronological studies carried out with modern and archaeological samples have been developed in temperate latitudes, where the formation of growth increments is strongly related to winter versus summer contrasts (variations of temperature and photoperiodicity), even though the hard tissue formation is a complex phenomenon and not necessarily temperature-driven (Guillaud et al., 2017, Panfili et al., 2002). In tropical areas, where temperature gaps are less disparate, factors driving growth increment formation are complex and, for this reason, the impact of seasonality in archaeological assemblages requires careful modeling (Meunier, 2012).

Archaeological taxa with large sample sizes generate more accurate sclerochronological estimates of seasonality. Therefore, here we focus on the South American marbled swamp eel (Synbranchus marmoratus), which is one of the most abundant and widely distributed synbranchid species in Central and South America and a common taxon recovered in archaeological assemblages in the Amazon Basin, Pantanal wetlands and estuarine zones of Rio Grande do Sul (Prestes-Carneiro and Béarez, 2017, Prestes-Carneiro et al., 2018, Rosa, 2000, Rosa, 2006). In south-western Amazonia, S. marmoratus comprises more than 70% of the fish remains recovered at Loma Salvatierra (Trinidad, Bolivia), a site was occupied from AD 500 to CE 1400 (Béarez and Prümers, 2005, Von den Driesch and Hutterer, 2012). Loma Salvatierra is an earthen platform mound that extends over 2 ha and reaches a height of 20 m. In the Llanos de Mojos there are more than 100 human-built earthen mounds (Jaimes Betancourt et al., 2012). These sites are frequently associated with other earth works, such as canals, ponds, and kilometers of causeways that radiate from the mounds and connect the sites together. The monumentality of these earth works raises questions about mobility and the interconnections between other groups, although no specific studies on mobility patterns or seasonality have been carried out. In order to address seasonality, we developed a modern reference collection of 61 S. marmoratus skeletons collected monthly over a year. We established a growth increment model for the species based on the vertebrae, and applied it to the 111 vertebrae recovered at Loma Salvatierra. Here, we present the possible factors driving the formation of the annulus and use the model to estimate the seasonality of fishing activities and mobility in pre-Columbian times.