INTRODUCTION
Land-snail shells are widely distributed in Quaternary deposits and frequently found on archaeological sites (Prendergast and Stevens Reference Prendergast and Stevens2014). Initial investigations (Goodfriend and Hood Reference Goodfriend and Hood1983; Goodfriend Reference Goodfriend1987; Pigati et al. Reference Pigati, Rech and Nekola2010; Hill Reference Hill2015; Hunt et al. Reference Hunt, Gilbertson, Hill and Simpson2015) suggest that in suitable circumstances it is amenable to radiocarbon (14C) dating, opening up the possibility of extending 14C chronologies into sedimentary facies containing no other source of biotic carbon. For instance, in calcareous oxidized sediments in semiarid and arid environments, land-snail shell is often well preserved but plant macrofossils and wood are not always available and bone collagen is often not preserved (e.g. Hunt et al. Reference Hunt, Gilbertson and Donohue1992; Anketell et al. Reference Anketell, Ghellali, Gilbertson and Hunt1995; Hunt and Garrard Reference Hunt and Garrard2013). Land-snail shells may offer the only possibility for 14C dating such sites.
Previous studies (e.g. Goodfriend Reference Goodfriend1987; Romanellio et al. 2008; Pigati et al. Reference Pigati, Rech and Nekola2010; Hill Reference Hill2015) have shown that the ecological pathways and uptake mechanisms undergone by 14C as it is incorporated into the shell of land snails are not straightforward. Some species are detritivores and others graze on the algae encrusting carbonate rocks. In both, old carbon may be consumed and incorporated into shell, causing a well-documented impact on 14C values, and the shells to have 14C age values significantly older than their actual age. It is also possible, although thus far unproven, that some taxa may fractionate carbon in their metabolic processes (McConnaughy and Gillikin Reference McConnaughey and Gillikin2008). An understanding of the 14C ecology at a generic, if not specific, level is therefore required before a species can be used for dating.
This paper presents the 14C ecology of the Helicid land snail Helix melanostoma (Draparnaud Reference Draparnaud1801) in northeastern Libya. The work was undertaken to provide a baseline for 14C dating of this species in Quaternary archaeological deposits throughout the Mediterranean and in particular in the Haua Fteah archaeological site lying approximately 1 km south of the Mediterranean coast in northeastern Libya, some 7.5 km east of Susa at 22º03′06′′E, 32º54′01′′N (Barker et al. Reference Barker, Hunt and Reynolds2007, Reference Barker, Basell, Brooks, Burn, Cartwright, Cole, Davison, Farr, Grun, Hamilton, Hunt, Inglis, Jacobs, Leitch, Morales, Morley, Morley, Pawley, Stimpson, Twati and Van der Veen2008, 2009, Reference Barker, Antoniadou, Armitage, Brooks, Candy, Connell, Douka, Farr, Hill, Hunt, Inglis, Jones, Lane, Lucarini, Meneely, Morales, Mutri, Prendergast, Rabett, Reade, Reynolds, Russell, Simpson, Smith, Stimpson, Twati and White2010, Reference Barker, Bennett, Farr, Hill, Hunt, Lucarini, Morales, Mutri, Prendergast, Pryor, Rabett, Reynolds, Spry-Marques and Twati2012; Rabett et al. Reference Rabett, Farr, Hill, Hunt, Lane, Moseley, Stimpson and Barker2012; Farr et al. Reference Farr, Lane, Abdulazeez, Bennett, Holman, Marasi, Prendergast, Al-Zweyi and Barker2014).
Background: Radiocarbon Ecology of Land Mollusks
The mechanisms by which 14C is incorporated into the shells of land snails are complex and interconnected (Figure 1). Three key pathways can be clearly defined by which carbon moves through the metabolism of a snail: respiration, diet, and hydration. Of these, diet accounts for the vast majority of carbon laid down in the shell (e.g. Stott Reference Stott2002; McConnaughy and Gillikin Reference McConnaughey and Gillikin2008; Xu et al. Reference Xu, Gu, Han, Hao, Lu, Wang, Wu and Peng2011). Dietary carbon can be from living vegetation, which will be broadly in 14C equilibrium with atmospheric levels. Dietary carbon is also likely to include ancient carbon from a variety of sources, as mentioned above. Similarly, hydration may include precipitation water condensed from the atmosphere, which will be in equilibrium with atmospheric 14C values and also water that has been in contact with, and dissolved, ancient carbon; including atmospheric mineral dust, limestone, and dead organic matter in soils.
Figure 1 Model for the radiocarbon ecology of land snails after Goodfriend (Reference Goodfriend1987) with modifications based on Stott (Reference Stott2002) and Pigati et al. (Reference Pigati, Quade, Shahanan and Haynes2004, Reference Pigati, Rech and Nekola2010).
The role of δ13C in interpreting land snail ecology has been widely explored. The relationship between this isotope and 14C dating land snails has long been an integral element in assessing the impact of “old carbon.” More recent detailed investigations of the 14C dating of land snails (e.g. Pigati et al. Reference Pigati, Rech and Nekola2010) have begun to demonstrate that while δ13C is important for fractionation correction in the measurement of 14C values it is not particularly useful in determining “old carbon” impact (e.g. Romaniello et al. Reference Romaniello, Quarta, Mastronuzzi, D’Elia and Calcagnile2008; Xu et al. Reference Xu, Gu, Han, Hao, Lu, Wang, Wu and Peng2011).
Understanding the relationship between the three isotopes of carbon within the biochemistry of the land snail is integral to both understanding why δ13C is not as important for identifying old carbon and also why a clear understanding of the life processes of the snail in acquiring carbon are vital to this point (Figure 1). Research has shown that atmospheric sources of carbon tend to have proportionally less impact on the processes involved in shell formation in land snails in comparison to dietary sources (e.g. McConnaughy and Gillikin 2009 and references therein). Thus, it has been argued, based on the biological principles underlying the carbon cycle in mollusks, that CO2 derived from respiration does not, in normal circumstances, have a significant impact on the 14C/12C ratio of landsnails (McConnaughy and Gillikin 2009).
The influence of dissolved inorganic carbonate (DIC) in water is understood only for a very few species, e.g. Succinidae, which are generally atypical of land snails (Pigati et al. Reference Pigati, Quade, Shahanan and Haynes2004, Reference Pigati, Rech and Nekola2010). Inference based on biochemistry of similar organisms does suggest that DIC should not in normal circumstances account for a significant portion of land-snail shell carbonate (McConnaughy and Gillikin 2009).
The ingestion of old carbon from various sources has been the subject of a limited number of studies (e.g. De Jorge and Haeser Reference De Jorge and Haeser1968; Tamers Reference Tamers1970; Magaritz et al. Reference Magaritz, Heller and Volokita1981; Goodfriend and Hood Reference Goodfriend and Hood1983; Goodfriend Reference Goodfriend1987; Goodfriend and Ellis Reference Goodfriend and Ellis2002; Stott Reference Stott2002; Balakrishnan and Yapp Reference Balakrishnan and Yapp2004; Pigati et al. Reference Pigati, Quade, Shahanan and Haynes2004; Baldini et al. Reference Baldini, Walker, Bruce, Baldini and Crowe2007; Yanes et al. Reference Yanes, Delgado, Castillo, Alonso, Ibanez, De la Nuez and Kowalewski2008; Romaniello et al. Reference Romaniello, Quarta, Mastronuzzi, D’Elia and Calcagnile2008; Pigati et al. Reference Pigati, Rech and Nekola2010; Xu et al. Reference Xu, Gu, Han, Liu, Pei, Lu, Wu and Chen2010, Reference Xu, Gu, Han, Hao, Lu, Wang, Wu and Peng2011; Rakovan et al. Reference Rakovan, Rech, Pigati, Nekola and Wiles2013). Review of these papers demonstrates that there is little replication or overlap in species analyzed or geographical regions sampled. Only in North America is there any consequential overlap; for Europe, North Africa and elsewhere there is no significant replication. A number of studies examining the use of terrestrial shell carbonate for 14C dating have also indicated, using 14C analysis, that old carbon can account for as much as 30–40% of carbonate in a shell (e.g. Goodfriend and Stipp 1983; Goodfriend and Ellis Reference Goodfriend and Ellis2002; Pigati et al. Reference Pigati, Quade, Shahanan and Haynes2004, Reference Pigati, Rech and Nekola2010). However, the quantities ingested appear to hinge on a wide variety of influencing factors in local environments and are taxon-dependent.
The mass balance approach pioneered by Goodfriend and Hood (Reference Goodfriend and Hood1983) for modeling the isotopic fractions of carbon in terrestrial land-snail shells has been heavily utilized. It has been found to be problematic because it is not possible to separate the sources of 14C (Figure 1) to the degree required to sufficiently isolate the component parts of the diet for the mass balance equation to be properly calculated (e.g. Romanellio et al. 2008; Pigati et al. Reference Pigati, Rech and Nekola2010; Xu et al. Reference Xu, Gu, Han, Hao, Lu, Wang, Wu and Peng2011; Hill Reference Hill2015). These issues further reinforce the need to build a clear understanding of the modern ecology, as well as the physiology of any species, before the 14C ages can be defined with confidence. This is particularly the case for the current work in Cyrenaica because previous studies of 14C ecology in the Mediterranean (e.g. Goodfriend Reference Goodfriend1987; Romanellio et al. 2008) have not examined terrestrial gastropods common in this region.
METHODS AND MATERIALS
Study Region
Sampling for modern snail shells and vegetation took place in northeast Libya (Figure 2). Here the Gebel Akhdar massif rises to over 800 m in three great escarpments. The region is an inverted Mesozoic-Tertiary basin, geologically characterized by thick limestones and thin calcareous mudstones (El Hawat and Abdelsamad Reference El-Hawat and Abdulsamad2004). Rainfall on the Gebel Akhdar is significantly higher than elsewhere along the northeast African littoral (Sen and Eljadid Reference Şen and Eljadid1999) and the region is a “habitat island” of Mediterranean vegetation among the generally arid lowlands to the west, south and east. The region bordering the coastal plain with richer soils and less limestone outcropping has noticeably more lush vegetation cover compared with the rising slopes of the Gebel Akhdar, which are heavily grazed. The slope of the first escarpment of the Gebel Akhdar is dominated by juniper scrub with large limestone outcroppings and shallow soils. The higher regions of the Gebel were originally covered in Cupressus-Oak forests but are now largely cleared for agriculture. The southern slopes are in rain shadow and largely covered with steppe, grading southward into true desert (Barker 2007; Hegazy et al. Reference Hegazy, Boulos, Kabiel and Sharashy2011).
Figure 2 Sample locations in Cyrenaica for this study. [A]=Haua Fteah (HF_Eco), [B]=Wadi Chartopolis (MH12-6), [C]=Gebel Akhdar 1 (MH12-9), [D]=Gebel Akhdar 2(MH12-10), [E]=Gebel Akhdar 3(MH12-11), [F]=Shahat project Resthouse (RH12), and [G]=Apollonia (AP12).
Ecological Patterning of H. melanostoma
Very little information is available on the ecology and distribution of land snails in the Gebel Akhdar. The most informative publication (Brandt Reference Brandt1959), although primarily taxonomic, provides some very short ecological descriptions of the family Helicidae. All Helicidae in the region are xerophytic to varying degrees and the species are found in habitats ranging from Mediterranean woodland to semi-desert.
In the study area, our surveys and conversations with local specialists suggest, that H. melanostoma spends most of the year aestivating in shallow, damp shaded soils beneath scrub or small trees such as Ceratonia siliqua, Pistacia lentiscus, and Rhus tripartita in areas of limestone pavement, scrub on slopes with shallow soils, and scrub-forest on deep soils (Hill Reference Hill2010). It is inferred that low-lying, mostly evergreen, vegetation forms the local primary food source of H. melanostoma. These findings were further reinforced by the stable isotope study (Prendergast et al. Reference Prendergast, Stevens, Hill, Barker, Hunt and O’Connell2014, Reference Prendergast, Stevens, Barker and O’Connell2015) which showed that δ13C in H. melanostoma shell carbonate was primarily a function of the animals’ diet, which is the local available vegetation. Therefore, changes in shell δ13C can be used to detect changes in vegetation composition, particularly in relation to the abundance of C3 and C4 vegetation (Prendergast et al. Reference Prendergast, Stevens, Hill, Barker, Hunt and O’Connell2014). Our environmental observations concerning this species broadly support ecological findings from southern France (e.g. Kerney et al. Reference Kerney, Cameron and Jungbluth1983; Pfleger Reference Pfleger1984) for H. melanostoma.
Sample Sites and Methodology
An initial 14C evaluation of a number of surface collected mollusk shells, presumed to be recently dead, carried out in 2010, of the four archaeologically important land snail species known to be abundant in the Haua Fteah archaeological site is shown in Table 1. This evaluation suggests that Helix melanostoma has less of an age offset, and less variability than Trochoidea cretica or Sphincterochila spp., the other ubiquitous species present both in the modern landscape and in the archaeological assemblages and sediments in the Haua Fteah. Rumina cf. decollata was ruled out based on ecological knowledge of diet and because no modern living specimens could be found in follow-up sampling. Therefore, based on its ubiquitous presence in archaeological sediments, and modern ages in this study H. melanostoma was selected for intensive work. The outlying sample, UBA 15321, appears to be archaeological material that was present on the landscape at the point of collection, based on comparisons of ages of material from within the Haua Sequence (Hill, Reference Hill2015) reinforcing the importance of live-collected specimens for studies like this.
Table 1 Results of pilot study evaluating F14C variation of land snails in northeastern Libya at site CPP07 1587. 14C ages are not rounded here to avoid introduction of errors in calculations.
The sampling of live collected modern specimens was carried out during 2010 and 2012 field seasons (Hill Reference Hill2010; Barker et al. Reference Barker, Bennett, Farr, Hill, Hunt, Lucarini, Morales, Mutri, Prendergast, Pryor, Rabett, Reynolds, Spry-Marques and Twati2012; Rabett et al. Reference Rabett, Farr, Hill, Hunt, Lane, Moseley, Stimpson and Barker2012; Prendergast et al. Reference Prendergast, Stevens, Barker and O’Connell2015). The sampling strategy was designed to encompass the widest possible range of modern environments within the study area but focused on five key locations given the constraints of time and security (Figure 2; Table 2). Site sample identifications for sites have been shortened in all tables and figures as follows: HF12_Eco refers to sites around the Haua Fteah; AP12 to Apollonia; RH12 to the Project Resthouse, Shahat; MH12 to Gebel Akhdar massif sample sites.
Table 2 Details of sample sites for modern sampling for mollusk and vegetation.*
* PL=Pistacia lentiscus; PH=Phlomis sp.; JO=Juniperus oxycedrus; RT=Rhus tripartia; CS=Ceratonia silaqua; SS=Sarcopoterium spinosum; AS=Artemisia sp.; P=Poacae; Z=Zizyphus sp.
At these localities transect sampling found live aestivated specimens in leaf litter and soil beneath vegetation and active mollusks on vegetation after very rare rainfall events. Ethnographic observations suggest that all of these sites are extremely heavily grazed by domesticated animals, mostly goats and sheep and vegetation levels are consequently lower than would be expected in habitats grazed by native animals. It is also likely that a number of sites were impacted by the relatively intense anthropogenic activity on the Gebel, near Shahat, Susa and Albayda including mass building work, cement works and atmospheric pollution.
A total of 37 H. melanostoma were selected for 14C analysis with a minimum of three specimens analyzed from each sample location. Vegetation and soil samples were gathered from every sample site where live mollusks were collected. Where possible, samples of vegetation were gathered in association with live mollusks. In selecting samples for atmospheric and dietary 14C values species favored by H. melanostoma were selected in accordance with existing ecological knowledge of the species (e.g. Brandt Reference Brandt1959) and observations made in the field. Where such species (Table 2) were unavailable, dominant vegetation was sampled. Once collected, mollusks were killed by freezing. Freeze-dried bodies were separated from the shell using forceps.
Radiocarbon
Whole shells were carefully broken apart using forceps or a Dremel tool with a circular carborundum cutting head. An aperture sample was obtained from every adult H. melanostoma that was dated (Figure 3). In small specimens with less than one full season of growth the entire shell was sampled and considered as an apex because the aperture had not thickened and so was too insubstantial for effective analysis.
Figure 3 Diagram of H. melanostoma showing parts selected for sampling. The stippling shows the dark brown areas on the shell after which the species is named.
Many previous radiocarbon studies, which were concerned primarily with predicting the “limestone effect,” did not assess lifetime changes in apparent 14C age (e.g. Quarta et al. Reference Quarta, Romaniello, D’Elia, Mastronuzzi and Calcagnile2007; Romanellio et al. 2008; Pigati et al. Reference Pigati, Rech and Nekola2010; Xu et al. Reference Xu, Gu, Han, Hao, Lu, Wang, Wu and Peng2011). Investigations of parts of the shell were carried out to evaluate changes in apparent 14C age over the lifetime of the individual snail (Figure 3). At least two paired aperture/apex dates were obtained from adult specimens for every location sampled except for the environment immediately outside the Haua Fteah where all the apices were used in a related δ13C shell study (Prendergast Reference Prendergast2013; Prendergast et al. Reference Prendergast, Stevens, Barker and O’Connell2015), looking at the suitability of H. melanostoma for climatic reconstruction. In some instances, whorl segments were also selected to provide further detail.
Vegetation samples for 14C analysis were rinsed in distilled water, dried, weighed into a precombusted quartz tube with an excess of copper oxide (CuO), sealed under vacuum and combusted to carbon dioxide (CO2) and converted to graphite using the hydrogen reduction method (Vogel et al. Reference Vogel, Southon and Nelson1987). Shell samples were pretreated using 1% HCl to remove any surface contamination and then rinsed in distilled water, dried and weighed. The samples were then placed in to septa seal containers and converted into CO2 using 80% phosphoric acid. The resulting CO2 was converted to graphite on an iron catalyst using the zinc reduction method (Slota et al. Reference Slota, Jull, Linick and Toolin1987). The 14C/12C and 13C/12C ratios were measured by accelerator mass spectrometry (AMS) at the 14CHRONO Centre, Queen’s University Belfast. The 14C/12C ratio of the vegetation and shell samples were background corrected using measurements of anthracite and Icelandic Spar Calcite, respectively, and normalized to the HOXII standard (SRM 4990C; National Institute of Standards and Technology). The 14C ages were corrected for isotope fractionation using the AMS measured 13C/12C, which accounts for both natural and machine fractionation. The 14C age and one standard deviation were calculated using the Libby half-life of 5568 yr following the methods of Stuiver and Polach (Reference Stuiver and Polach1977). 14C data for modern samples are presented as F14C values (Reimer et al. 2004). The 1σ uncertainties are the maximum of the measurement statistics and the variance of seven 2-min runs for each sample. Throughout this paper, 14C ages reported in tables are not rounded as this may introduce errors in calculations (Millard Reference Millard2014).
Stable Isotopes
Stable carbon analysis of vegetation and soil was carried out at the 14CHRONO Centre at Queen’s University Belfast. Stable isotopes (δ13C and δ15N), % carbon, and % nitrogen from vegetation and soil samples were measured on a Thermo Fisher Delta V Advantage with elemental analyzer together with Iso-Analytical Laboratory Standard IA-R041 L-Alanine (δ15NAIR=5.56 ‰; δ13CVPDB=23.33 ‰) which was measured 11–13 times throughout the runs, bracketing between 6 and 8 samples.
Modeling and Data Analysis
Data analysis was carried out using Microsoft Excel 2013 and SPSS version 19. Simple and exponential linear regression and chi square (χi 2) were used to investigate relationships between variables. Weighted averages of multiple samples and chi square tests were calculated using a VBA coded macro in Excel based on Bevington (Reference Bevington1969). The basic relationships between 14C results from different sites were evaluated using chi-squared (χi 2,0.5) analysis and analysis of correlations and sample scatter (standard deviation).
F14C values for shells, where apertures, apices, and whorl fragments were measured, and where required, combined for certain analyses of the 14C ecology of H. melanostoma. This was done by calculating the weighted averages of the F14C values to simulate analysis of the whole shell (Table 6). Because the aperture and apex F14C values should represent maximum range of F14C values in a land snail across its growth cycles, the resultant average F14C and its aggregated error should represent a reasonable average F14C value for a sample across its whole life span. This was further tested by the measurement of a number of whorl fragments to determine intermediate values.
Predicted δ13C values were calculated using the method set out in Stott (Reference Stott2002). The Stott equation is based on a linear regression between snail body δ13C and snail shell δ13C. The values used in this study were first calculated in Prendergast (Reference Prendergast2013) and snail body δ13C and shell δ13C for samples comes from Prendergast et al. (Reference Prendergast, Stevens, Hill, Barker, Hunt and O’Connell2014). While the equation was originally created for H. aspersa rather than H. melanostoma, the two species occupy similar niches in different areas around the Mediterranean and ecological observations suggest that H. melanostoma is locally being out-competed by introduced H. aspersa, particularly in urban areas. It is important to note that there may be further species-specific offsets that may impact detailed comparison with such values.
Calculation of the difference between F14C of mollusk shells and F14C of preferred vegetation diet for H. melanostoma, Equation (1), was used to model the potential impact of old carbon on individual shells (e.g. Pigati et al. Reference Pigati, Rech and Nekola2010)
$${\rm F}^{{14}} {\rm C}_{{{\rm Shell}\,{\rm Average}}} -{\rm F}^{{14}} {\rm C}_{{{\rm diet}}} {\equals}{\rm F}^{{14}} {\rm C}\,{\rm from}\,{\rm nondietary}\,{\rm sources}$$
F14CShellAverage represents the average F14C value across the lifetime of the shell (Table 6). F14Cdiet is derived using F14C atmospheric values for vegetation consumed by H. melanostoma gathered from sample sites (Table 5). In instances where more than one value exists for F14Catmosphere from a site, a weighted average was used for the value of F14Cdiet.
This method follows the understanding that the total quantity of F14C present in a mollusk shell can only derive from three overall sources, dietary carbon, respiration and hydration. Of these, only dietary carbon is likely to be significant in the vast majority of cases where snails pass most of their lives on or above the soil surface. Dietary carbon will have two components; “young,” derived from living tissue, and “old,” derived from dead tissue or mineral sources. If the F14C of the vegetation eaten by the snail can be established, then the remaining difference between the F14C of the shell and the F14C of the diet should represent the quantity that the shell’s F14C value has been reduced through the incorporation of older carbon, even if the exact source of carbon contributing to the difference cannot be fully known.
The age offset in 14C years, the “old carbon effect” as discussed in this study, where used, was calculated using the following Equation (2)
$${\rm Shell }\,{\rm reservoir }\,{\rm offset {\equals} }{\minus}\!{\rm 8033}\,{\rm {\asterisk}}\,ln\left( {{{F^{\rm 14}Cshell} \over {F^{\rm 14}Catm}}} \right)$$
This equation was defined after Keaveney and Reimer (Reference Keaveney and Reimer2012), F14Cshell=F14C value derived from the land-snail shell. The −8033 is derived from the Libby half-life of 14C (T1/2=5568 yr). Atmosphere values; F14Catm , in this study are based primarily on vegetation samples gathered at sample sites for live land snails in the year of collection (2010 and 2012) as the Northern Hemisphere Zone 2 dataset (Hua et al. Reference Hua, Barbetti and Rakowski2013) is only current until late 2009 and Levin et al. (Reference Levin, Kromer and Hammer2013) does not have any data points with sufficient geographic proximity.
The uncertainty that is associated with each reservoir offset is calculated using the uncertainty associated with the 14C measurement (F14Cshell) from each modern shell sample and its associated uncertainty F14Csigma using the Equation (3) below (Soulet et al. Reference Soulet, Skinner, Beaupre and Galy2016).
$${\rm sigma}{\equals}8033\,{\asterisk}\,\sqrt {\left( {{{F{}_{{}}^{{14}} C \,Shell\,Sigma} \over {F{}_{{}}^{{14}} C \,Shell\,measured}}} \right)^{2} {\plus}\left( {{{F{}_{{}}^{{14}} C \,atm \,sigma} \over {F{}_{{}}^{{14}} C \,atm\,measured}}} \right)^{2} } $$
RESULTS
Fifty-nine 14C analyses were made from 32 live H. melanostoma over the course of the field sampling exercises. The results of the δ13C on vegetation (Table 3), 14C of vegetation (Table 4), 14C of shells (Table 5), and modeling of 14C shell data and old carbon impact (Table 6) are presented here.
Table 3 δ13C of vegetation and soil samples from across Cyrenaica.
* Sample 1546 was gathered from outside the Haua Fteah.
Table 4 Radiocarbon results of vegetation samples taken from across Cyrenaica as a proxy for atmospheric 14C and dietary F14C for land snails. Average dietary F14C of suggested mollusk diet is based on the per site values presented in this table.
Table 5 Radiocarbon and stable isotope results and analysis on apertures, whorl fragments, apices and aggregated ages of H. melanostoma from the Cyrenaica region of Libya.
Table 6 Table showing the limestone effect observed in sampled mollusks in the form of differences in F14C values between dietary sources and shells across the Gebel Akhdar.
* F14C shell average is calculated based on the median F14C value for samples taken from specific shells. Estimation of old carbon content was based on the methods set out in Pigati et al. (Reference Pigati, Rech and Nekola2010). Data paired with δ13C Shell and δ13CBody values measured by IRMS (not machine measurement from the AMS) on selected samples (Prendergast Reference Prendergast2013) and δ13C predicted was calculated using the method described in Stott (Reference Stott2002).
Biometrics
Biometrics were gathered for all H. melanostoma specimens sampled for 14C dating. Mean basal diameter (MBD) height and growth checks were measured for all samples. The mean MBD for all specimens was 21.3 mm, Mean height was 18 mm and the average number of growth checks was 2.4. The largest MBD was 28.7 mm, whereas the smallest was 7.3 mm. Those with the smallest MBD and Height were all found in Apollonia (AP12).
Within-Shell Radiocarbon Variation
There was a substantial but predictable “age” difference between apices and apertures (Figure 4), with apices being “older” than apertures. This is significant for 14C dating as it suggests that juvenile specimens may have a higher propensity towards old carbon intake than adult specimens. It is also important for archaeologists selecting dating samples because apices and aperture fragments are easily identifiable and preserve well. Whorl fragments, where analyzed, were found to more closely reflect values of apertures (Table 5) rather than more broadly intermediate ages. In ecological terms, it is likely that young snails require calcium (Ca) to start building a shell, obtaining it from small limestone particles in the soil, from outcrops, or even dead snail shell. They may also start life grazing algal biofilms on the soil/rocks. Generally, growth banding suggests lifespans of 2–5 growth seasons. In adults, the shell may grow less rapidly and the snail seems to get a proportionally larger quantity of their calcium through grazing vegetation. More homogenous and higher F14C values in aperture and whorl samples could reflect the high ecological mobility of adult specimens (Kerney et al. Reference Kerney, Cameron and Jungbluth1983) and would also explain why no H. melanostoma were collected in proximity to Juniperus, since gymnosperms such as this notoriously acidify soils (Augusto et al. Reference Augusto, De Schriver, Vesterdal, Smolander, Prescott and Ranger2015) and such environments are not preferred by land snails. This may also account for older average “ages” and wider variation in H. melanostoma specimens gathered in some parts of the Gebel Akhdar transect where the primary vegetation is Juniperus rather than Ceratonia or Pistacia, which would not acidify the soil in the same way. In these regions, there is also significantly less vegetation cover and poorer overall diversity potentially driving land snails such as H. melanostoma to seek carbonates from non-vegetative sources and exacerbating the “old carbon effect” where it occurs.
Figure 4 Boxplot of F14C of H. melanostoma apices and apertures by site: ML12_6,9,10: wt. average of aperture and apices values (n=9) from sample sites on the Gebel Akhdar. AP_12: Apollonia (Susa) (n=2); RH_12: Project Resthouse (Shahat) (n=4). Individual data points are displayed within the box plot as “+”.
Regional Trends in the Radiocarbon Ecology of Helix melanostoma
Regional trends in the data are present. Comparison of the differences between F14Capertures and F14Capices values with levels of total vegetation finds weak but significant correlations for both apertures and apices (Table S1). This relation was found to improve noticeably if F14C values were compared specifically with arboreal cover (Table 4). There is an even more pronounced association between vegetation cover and variations in F14Cwholeshell values (Figure 5a) that also show a weak but significant correlation and demonstrates a clear positive trend between increased total vegetation cover and higher F14C values. This correlation is further improved when arboreal cover is compared directly to F14Cwholeshell values (Figure 5b). The observed relationship between F14C values with vegetation cover suggests that the underlying ecological behavior of juvenile and adult H. melanostoma is quite similar in terms of the vegetation they consume across all the sample sites. It is interesting that this correlation is noticeably stronger in the combined F14Cshell values (F14Cwholeshell) analysis than for either apertures or apices individually. This is likely to be due to the similarity of the datasets amplifying the observed trend. It is also probable that this approach filters out a significant portion of the local noise reflected in F14C values at individual sites. The stronger correlations between F14C values and tree and shrub cover are almost certainly a reflection of the dietary preferences of H. melanostoma, indicating that they prefer taller arboreal vegetation over grasses and low-lying herbaceous vegetation.The availability of such vegetation would appear to be a key variable that molds the 14C ecology of H. melanostoma in the study region where the vegetation is largely limited by grazing pressure since rainfall levels are always sufficient to support trees. Therefore, 14C ecology is responding to grazing pressure from domesticated animals and does not reflect a truly natural situation. This conforms well to existing research (e.g. Stott Reference Stott2002) and could be an important influence if dating shells of this species found in association with Neolithic or later archaeological material where grazing of the local landscape would have been possible.
Figure 5 (a) Relationship between F14Cwholeshell and total vegetation cover; (b) relationship between F14Cwholeshell and arboreal vegetation cover.
These findings also strongly support the hypothesis raised in other studies (e.g. Pigati et al. Reference Pigati, Rech and Nekola2010) that calcium availability could be a contributory factor to the incorporation of old carbon into the diet. In almost all the sites, limestone availability was observed to be essentially inversely proportional to vegetation cover, so the positive trend between vegetation and higher F14C values is also a negative trend between limestone availability and lower F14C values. The most striking example of this is at Apollonia, which has very poor vegetation cover, and F14C values for apices and apertures that are correspondingly much lower (Table 5). It may be, however, that this is a function not of limestone availability but of vegetation availability, which enforces grazing of endolithic algae, with an associated old carbon uptake from the limestone.
Previous study has demonstrated that H. melanostoma in the sample region have a diet of C3 vegetation (Prendergast et al. Reference Prendergast, Stevens, Hill, Barker, Hunt and O’Connell2014, Reference Prendergast, Stevens, Barker and O’Connell2015). As part of that study a modeling exercise calculating the dietary δ13C values for H. melanostoma if their diet was 100% vegetation was carried out. The results of this exercise together with the 14C age offset are included in Table 6. The measured δ13C shell values, when compared with their modeled values, show both a significant observable trend in measured versus predicted δ13C enrichment in H. melanostoma shells (Figure 6a) across the study areas (Table S1). Measured samples from the Haua Fteah are mostly within the margin of error (±1‰) when compared. However, samples, from other sites have much greater variation. There is a significant, positive relationship, with a moderate correlation between δ13Cshell enrichment/depletion and arboreal vegetation cover (Figure 6b). Further comparison of δ13Cshell variations against other potential variables (Table S1) found that there were no significant trends across the sites sampled and no significant observable relationship between δ13Cshell and F14Cwholeshell values when compared directly across sampled sites (Table S1).
Figure 6 (a) Measured against predicted enrichment δ13C across the study area; (b) δ13C enrichment and depletion against arboreal cover across the study area. Dashed lines represent the lines of best fit.
These findings are not surprising from an ecological or stable isotope perspective since there is a large body of past work that has found that the δ13C of land-snail shells has a strong relationship with diet (e.g. Stott Reference Stott2002; Metref et al. Reference Metref, Rousseau, Bentaleb, Labonne and Vianey-Liaud2003; Colonese et al. Reference Colonese, Zanchetta, Perlès, Drysdale, Manganelli, Baneschi, Dotsika and Valladas2013; Reference Colonese, Zanchetta, Fallick, Manganelli, Lo Cascio, Haussman, Baneschi and Regattieri2014; Yanes et al. Reference Yanes, Delgado, Castillo, Alonso, Ibanez, De la Nuez and Kowalewski2008; Reference Yanes, Izeta, Cattaneo, Costa and Gordillo2014). The changes in enrichment and depletion and their correlation towards F14C values is interesting for 14C dating and the issue of “old carbon.” Sites that present the lowest F14Cshell values e.g. Apollonia, Shahat (Table 6), also present the most substantial enrichment in δ13C (>2‰) values compared with what would be expected if their diet was entirely vegetation. The Haua Fteah, which has good vegetation cover and higher F14C values, has less enrichment and are in some cases marginal depletion of δ13C values (±1‰). The trend shown in Figure 6b suggests that the magnitude of enrichment of samples may specifically relate to the availability of suitable dietary vegetation at sample sites, e.g. Ceratonia siliqua or Pistacia lentiscus. Areas lacking in arboreal vegetation in northeastern Libya also tend to have more significant outcroppings of limestone, which would account for lower F14C values for samples from those regions as H. melanostoma could change its diet to compensate for the lack of vegetation. When considering the suitability of a species such as H. melanostoma for 14C dating this demonstrates the impact of dietary pressure on the F14C value of the shell very clearly in a way that was not picked up by direct comparison of the δ13C and F14C values. This relationship suggests that such a modeling exercise could be a useful tool in the future to flag specimens that may have an old carbon effect without directly dating 14C dating a sample.
Relationships between F14Cshell and F14Cdiet
The relationship between F14Cshell and F14Cdiet was also investigated for H. melanostoma. There is only a slight correlation between changing F14C dietary values and F14C shell values (Table 3). This is broadly expected, as there is considerable variation within individual sample sites in F14Cshell and the sample size is not particularly large for some locations. In order to further compare F14C shell and F14C diet, age corrected Δ14C (Figure 7) values were calculated, and the proportion of old carbon in the diet derived following the methods set out in Pigati et al. (Reference Pigati, Rech and Nekola2010) after Stuiver and Polach (Reference Stuiver and Polach1977) to facilitate easier comparison to previous studies.
Figure 7 F14Cshell and F14Cdiet plotted as age corrected Δ14C, demonstrating the variation and regional clustering of the old carbon effect for H. melanostoma in the study region.
Findings imply that most specimens are incorporating a proportion of old carbon into their diets as a matter of course. What is significant is that there is a clear delineation that sits well with the earlier observations about vegetation availability and calcium dependency as modulating factors for shell carbonate composition. While the total range of values measured (Table 5) is large; (UBA 18082) 109±38 14C yr to (UBA 21758) 2001±43 14C yr with a weighted mean of 659±76 14C yr, the majority of samples from non-urbanized sample locations have only a small old carbon effect with a weighted mean of 476±48 14C yr, which approximately relates to between 1% and 9% of their dietary F14C coming from non-organic carbonate sources (Table 6) where this has been measured. The clear exceptions are samples from Apollonia and a single sample from Shahat; the urbanized sample locations manifest old carbon effects of >1000 14C yr and skew the averages accordingly. These more extreme carbonate environments are most likely to represent the modern extremities of the old carbon effect on H. melanostoma. Such environments are highly unlikely to impact archaeological ages significantly since palynological evidence (Hunt et al. Reference Hunt, Gilbertson, Hill and Simpson2015; Simpson Reference Simpson2016) shows relatively stable vegetation cover in the region throughout the later Quaternary.
CONCLUSIONS
The ecological observations from all the sampled sites suggest that H. melanostoma has a very homogenous 14C ecology in its natural habitats. There is a similar level of variation within F14C values across the majority of sample sites within the study area. Only in localities with drastically anthropogenically altered habitats (Apollonia, Shahat) does H. melanostoma exhibit F14C values that are significantly distant from values observed at less impacted sites. This significant deviation from predicted levels of F14C for a species that preferentially consumes fresh, leafy vegetation is attributed to both the lack of choice vegetation and the influence of anthropogenic factors upon available vegetation (e.g. cement works, air pollution). These issues are unlikely to be significant in an archaeological context prior to the Roman Period and therefore these two localities may be regarded as clear outliers. It is significant that F14C values of apices are consistently lower than apertures across all sample sites. The scatter of F14C values for shell apertures, the most widely comparable F14C dataset, is very constrained. There is also a demonstrable trend of enrichment of δ13C at coastal and urbanized sample sites, decreasing availability of choice vegetation such as Ceratonia siliqua and Pistacia lentiscus and lower F14C values. This may suggest that enrichment in δ13C is a proxy indicator for the old carbon effect on H. melanostoma under certain circumstances.
The most significant of these ecological observations are the very similar, and thus predictable, differences between apertures and apices at all the sites and the predictable behavior of old carbon incorporation, both of which appear to be heavily modulated by the availability of choice vegetation such as Ceratonia siliqua and Pistacia lentiscus. The scale of differences between apertures and apices are very similar despite the distance between the sample sites, some variation in numbers of growth seasons and size of specimens. This suggests that the 14C ecology of juvenile H. melanostoma may differ from that of adults where incorporation of carbonates is concerned. The lack of a significant difference between adult and juvenile δ13C, even taking into consideration the small number of samples involved (n=5), opens up the possibility that the factor causing this variation is not vegetation availability. The investigation of modern H. melanostoma further suggests weaknesses within the existing methods used in the 14C dating of land mollusks, particularly the use of δ13C to estimate the impact of the “limestone effect” within mollusk shells. In this instance this was found to over-emphasize the signature of dietary vegetation.
The similarity of the relationships between F14C values of shell apices and apertures to total vegetation cover suggests that both juvenile and adult H. melanostoma consume the same dietary vegetation, and that both are equally affected by this variable. Thus, the increased F14C values of H. melanostoma apices should originate in the non-organic carbonate component of diet and has the largest impact on the F14C values of the apices which is the longest growing part of the animal. It may also relate to subtle differences in vital effects due to differing fractionation chemistry between adult and juvenile animals, although only a limited number of juvenile snails were available to test this.
This is further supported by a marked increase in “age” difference between the F14C of apices and apertures in Apollonia and those from the Gebel Akhdar. The landscape around Apollonia is limestone-dominated and has poor vegetation availability, whereas the Gebel Akhdar is comparatively well vegetated with much less outcropping limestone. This reinforces the conclusion that the significantly decreased F14C values at Apollonia are probably an extreme response to dietary pressure. It therefore seems reasonable to infer that dietary calcium availability could be the controlling factor for the incorporation of limestone into the diet of H. melanostoma. The difference in the F14C values could be an indicator that juveniles simply consume more calcium carbonate than adults in order to construct their shells quickly during their first season of growth. There is much scope for further work on this hypothesis. Moreover, many of the findings observed for H. melanostoma should also be observable for other species with similar ecological behavior and should therefore be replicable for other taxa, elsewhere in the Mediterranean and further afield.
This study demonstrates that, in circumstances where there is a relatively complete vegetation cover including at least some trees and shrubs, the 14C ecology of H. melanostoma is stable and predictable. The observed instabilities in the 14C ecology of this species in Cyrenaica can all be attributed to the results of intense human activity which would not have been present before the Roman Period. Therefore, this study opens up the use of H. melanostoma and species with similar ecological behavior for 14C dating of archaeological and geological deposits, with the use of a suitable offset to correct for the predictable effects of diet observed here.
ACKNOWLEDGMENTS
We thank the Department of Antiquities of Libya in Tripoli and in Shahat for permissions and continued support from the inception of the project. We acknowledge the financial support of the Society for Libyan Studies, the Leakey Foundation, the University of Cambridge, and the European Research Council (ERC Advanced Investigator Grant 230421 to Graeme Barker). We thank our many Libyan and European colleagues for help, support, and discussion over several field seasons. We thank the staff of the 14CHRONO Centre, Queen’s University Belfast including Stephen Hoper, Jim McDonald, Ron Reimer, Julia Simpson, and Michelle Thompson for their expertise and advice. We also are indebted to constructive comments by Jeff Pigati and an anonymous reviewer.
SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit https://doi.org/10.1017/RDC.2017.49
