Modeling avian eggshell microstructure to predict ontogenetic age and reveal patterns of human-avifauna interaction

Highlights

• Eggshell is underutilized and can clarify past human-avifauna interaction.

• We demonstrate the utility of capturing high-resolution 3D surfaces of eggshell.

• We characterize changes in eggshell microstructure linked to embryonic development.

• We developed statistical models to predict eggshell ontogenetic age.

• Assessment of eggshell ontogenetic age can be used to infer human exploitation.

Abstract

Archaeological eggshell is a commonly recorded, yet underutilized material for understanding human-environment interaction in the past. In addition to the use of archaeological eggshell as a paleoenvironmental proxy, archaeologists have innovated important new approaches to the study of archaeological avian eggshell, including the application of scanning electron microscopy (SEM) for the characterization of eggshell microstructures. These studies have demonstrated the importance of eggshell for understanding how ancient communities exploited and interacted with avifauna. In this paper, we build on these methodological advances by testing new approaches to image avian eggshell and characterizing complex eggshell surface microstructures. We demonstrate the utility of capturing high-resolution 3-dimensional (3D) eggshell surfaces using advanced imaging modalities (optical profilometry, scanning electron microscopy, digital microscopy, computed tomography), to model changes in eggshell microstructures that are correlated with ontogeny. Using the Common Ostrich (Struthio camelus) as our model system, we introduce a statistical modeling approach to predict the ontogenetic age of ratite eggshell using roughness measurements of 3D features. Successful prediction of ontogenetic age has great potential to reveal archaeological patterns of human exploitation of avian eggs. These findings further illustrate the importance of archaeological eggshell for investigating human-environment interactions, emphasizing the need for archaeologists to use field methods (hand/trowel excavation and fine mesh screen) that facilitate eggshell recovery.

Introduction

The archaeological record is replete with evidence of the importance of birds to human communities, extending to at least the Middle Stone Age in Africa when ostrich eggshell was engraved (Assefa et al., 2018; Texier et al., 2013; Texier et al., 2010) and the Upper Paleolithic in Europe when bird bones were transformed into some of the earliest known flutes (Conard et al., 2009). Birds are integral to human interaction, subsistence, ritual, migration and navigation, and the recovery and analysis of bird remains from archaeological sites yields crucial insights into the human past (Dirrigl et al., 2020; Serjeantson, 2009). Bird remains in archaeological sites include bones, gizzard stones, pellets, feathers, and eggshell. Despite the importance of birds to human societies and the abundance of bird remains in many archaeological sites around the world, several scholars have noted that avian eggshell remains an understudied component of the archaeological record (Beacham and Durand, 2007; Stewart et al., 2013a).

This paper describes approaches for high-resolution 3D imaging of avian eggshell and a statistical model for predicting eggshell ontogenetic age built using a time series of ostrich eggs (Struthio camelus). The aim of developing these methods is to increase archaeologists’ ability to derive information about human-avifauna interactions from archaeological eggshell. Eggshell is a commonly recorded material in archaeological deposits and has been used successfully as a paleoenvironmental proxy and dating material through the analysis of eggshell geochemistry (Brooks et al., 1990; Donaire and López-Martínez, 2009; Ecker et al., 2015; Freundlich et al., 1989; Higham, 1994; Janz et al., 2009; Johnson et al., 1998; Clarke et al., 2007; Long et al., 1983; Vogel et al., 2001). Molecular techniques, including ancient DNA analysis (aDNA) and proteomics of archaeological avian eggshell elucidate bird population dynamics and taxonomy (Stewart et al., 2013a; Jacomb et al., 2014; Allentoft et al., 2014; Stewart et al., 2013b; Oskam et al., 2012; Oskam et al., 2010).

Relative to most other classes of archaeofaunal remains, however, morphological analysis of avian eggshell has rarely been used to understand human-avifauna interactions, including egg harvesting and husbandry practices. The first effort to encourage the morphological study of archaeological avian eggshell resulted in a manual for ootaxonomic identification, covering a range of avian families and species, based on Scanning Electron Microscopy (SEM) imagery (Sidell, 1993). Although ootaxonomic identification through morphological analysis of avian eggshell has proven challenging (Buss and Keiss, 2009), morphological analysis of anthropogenic modification of eggshell (esp. in the production of artifacts like eggshell beads and liquid containers) has enabled the study of human symbolic expression, craft production, interaction and exchange (Texier et al., 2013; Kandel and Conard, 2005; Wei et al., 2017; Jacobson, 1987). Others have distinguished anthropogenic modification in the production of liquid containers from similar patterns produced by carnivore predation (Kandel, 2004). Additional modifications on archaeological avian eggshell result from intentional and unintentional heat treatment (Miller et al., 2016; Taivalkoski and Holt, 2016).

Previous work to estimate the ontogenetic age of archaeological eggshell investigated potential husbandry of turkey (Meleagris gallopavo) in the ancient American Southwest and suggested that changes in eggshell microstructures were time-dependent and correlated with stages of embryonic growth (Beacham and Durand, 2007). During the course of incubation, developing avian embryos derive a substantial portion of the calcium they require for growth (>80%) by resorbing calcium carbonate from the surrounding eggshell (Carey, 1983). This process of resorption is enabled by the interaction between eggshell mammillary conesmicrostructures on the interior surface of avian eggshell (Fig. 1) — and the shell membranes (Burley and Vadehra, 1989). As a result of this process of resorption, the surface structure of eggshell mammillary cones changes over the course of incubation (Fig. 2). Furthermore, because resorption of eggshell calcium carbonate is linked to the development of the avian embryo (Blom and Lilja, 2004), changes in eggshell mammillary cone surface structure are hypothesized to be time-dependent, and thus indicative of the ontogenetic age of the embryo (Beacham and Durand, 2007).

Using two-dimensional (2D) SEM images of eggshell, Beacham and Durand qualitatively assessed changes in eggshell microstructures and demonstrated the possibility of identifying eggshell from hatched versus un-hatched turkey eggs in archaeological contexts (Beacham and Durand, 2007). Their work built upon previous documentation of the presence of turkey eggshell in Chaco Canyon deposits (Windes, 1987); they interpreted the high proportion of archaeological eggshell showing signs of significant resorption as evidence that ancient Chacoan communities were hatching turkey eggs, in order to raise chicks, as opposed to harvesting and consuming the egg contents (Beacham and Durand, 2007). Following this innovative approach, further studies were conducted on archaeological eggshell from the American Southwest (M. gallopavo), Oaxaca (M. gallopavo), and eastern Europe (Gallus gallus), further elucidating processes of avian domestication, the importance of bird husbandry, particularly during periods of climate downturn, and the ritual significance of fertilized eggs as grave goods (Conrad et al., 2016; Lapham et al., 2016; Jonuks et al., 2017).

Successful prediction of eggshell ontogenetic age has great potential to reveal archaeological patterns of human exploitation of avian taxa. As indicated by the studies referenced above, a primary question for archaeologists is whether we can use eggshell morphology to successfully distinguish between eggs from which a chick hatched naturally and eggs for which normal chick development was halted at some ontogenetic stage prior to hatching, potentially as a consequence of harvesting by people. Further questions regarding human behavior can also be posed, if our analytical methods allow us to identify more precisely the ontogenetic timing of egg harvesting. Evidence for the harvesting of fertilized eggs during later stages of embryogenesis, for example, may indicate dietary preferences (e.g. a preference for consuming a more developed chick embryo, as opposed to yolk and albumen).

In this paper, we develop a highly resolved ontogenetic time series of Common Ostrich (Struthio camelus) eggshell and a statistical model to predict the ontogenetic age of unknown eggshell samples. Although chicken is the classic model species in bird developmental biology, we selected ostrich as our model species to expand analytical possibilities with regard to ratites. Several ratite taxa have gone extinct over the course of the Holocene, including New Zealand's moa (Dinornithiformes) and Madagascar's elephant birds (Aepyornithidae). These extinctions often followed human dispersals to islands, and we know little regarding the potential contribution of ratite egg harvesting to the success of human communities in colonizing new environments, or to subsequent ratite population declines. Our work furthers the study of archaeological eggshell in three important ways:

• We demonstrate the utility of capturing high-resolution three-dimensional (3D) eggshell surfaces using advanced imaging modalities (Optical Profilometry [OP], SEM, digital microscopy, Micro-Computed Tomography [micro-CT]), in order to quantitatively model changes in eggshell microstructures that are correlated with ontogeny. This allows us to compare different imaging instruments and techniques, and evaluate the reliability of qualitative visual inspection of 2D images, as compared to a statistical modeling approach using 3D data.

• We test whether our statistical model, derived from an ostrich time series, successfully predicts the ontogenetic age of eggshell of other avian taxa and identify limitations of this proxy approach. The ability to use ostrich as a model oospecies would greatly expand the utility of these methods to the study of extinct ratite species and extant species for which high-resolution eggshell ontogenetic time series are not available. Generating an ontogenetic time series of eggshell requires sacrificing fertilized eggs with developing embryos and known lay dates. As such, creating the necessary comparative samples is often neither feasible nor ethical for the particular species of interest.

• We conduct a sensitivity analysis to test the robustness of our model, providing an approach to the issue of archaeological eggshell diagenesis. This is particularly important in archaeological contexts where taphonomic factors lead to poor preservation of eggshell remains.