Environmental implications and evidence of natural products from dental calculi of a Neolithic–Chalcolithic community (central Italy)

Morphological analysis

Food preparation or processing of plant material involve multiple activities and all of them can potentially leave micro-traces in the tartar, together with the environmental components.

Eleven dental calculi showed plant record: starches, pollen grains, one trichome, one sporangium, and tissue fragments (Table 1).

Plant hair

Trichomes are epidermal outgrowths characterized by different structure and function. Although plant hairs are some of the most common findings in the overall particulate matter carried by air (as pollen grains), in literature, only very few examples of trichomes in ancient contexts have been reported^28,29,30. Trichome identification is not a common area in dental calculus research since they do not have a diagnostic morphology. For this reason, the identification of such type of microdebris must be based on realistic criteria, also in accordance with the geographical and historical context, providing all possible interpretative scenarios. The detection of trichomes in ancient tartar may disclose other lines of evidence than nutrition, representing a reliable archaeological environmental proof³¹.

One plant hair was identified in CDD1 sample (Table 1). This remain (Fig. 2L) falls into the general class of dendritic trichomes and its peculiar morphology has been more specifically termed a candelabrum or abietiform³². The overall structure corresponded to non-glandular and pluricellular trichomes with a central uniseriate axis and whorls of unicellular rays emerging at the joints of the axis. Usually, 4 radii from each node occurred perpendicular to the central axis. As exhaustively reported in literature, dendritic trichomes are known in ferns, different groups of modern monocots and basal eudicots, such as Scrophulariaceae and Platanaceae. Although dendritic, trichomes of ferns and monocots were excluded. Indeed, the first ones possess single secondary branches that alternatingly arise at an angle of 70°–120° with respect to the main axis along a single plane³³, while the second ones show morphological features and appearance different from the ancient debris^34,35. Candelabrum-like trichomes have been usually detected in Verbascum L. and Platanus L. species³⁶. For this work, an experimental reference collection of trichomes from these plants was created (Supplementary Information 1). The general aspect of mullein trichomes appears to be capitate, bigger, more elongated, and slenderer than the microremain found in tartar sample. In addition, these trichomes seem to possess a pair of secondary elements per side or single secondary branches, which depart from the nodes, only rarely perpendicular to the central axis^37,38,39. Thanks to the well-preserved morphology, the ancient candelabrum hair was interpreted as a Platanus sp. foliar trichome based on literature⁴⁰ and our experimental reference, although mullein cannot be totally excluded. Dimensions, distance between nodes and the number of tapering secondary branches attached to the central axis of the microremain were like those of all plane species documented in literature^40,41,42.

This finding leads to consider some paleoenvironmental implications. Fossil pollen analysis has demonstrated that, during the Plio-Pleistocene, Platanaceae were present in the Upper Valdarno (Italy)⁴³. For the Holocene, likely as a consequence of Pleistocene glaciations, fragmentary and scarce evidence of plane tree have been found in Spain and French Mediterranean coast; no record of Platanus sp. has hitherto been found in Italy^44,45. This thermophylous taxon has reappeared later as an ornamental tree, providing shade, during Roman times⁴⁶. As we applied rigorous decontamination protocols, the evidence of this ancient trichome, probably accidentally inhaled by CDD1, may testify the presence of Platanus sp. and humid environments in central Italy during the Neo-Chalcolithic period.

Starch granules

More than 70 starches were retrieved from calculus samples (Table 1). Some of them were found in an extraordinary state of preservation, likely due to intentional ingestion and/or accidental inhalation during the processing of starchy foods. These grains were clustered in three different morphological types, based on the morphometric parameters (i.e., shape, size, presence of lamellae and hilum, aggregation level, and other secondary features) evidenced by literature. They were described using the International Code for Starch Nomenclature^47,48.

Morphotype I These starches were consistent with those of Triticeae Dumort. tribe and occurred in almost all samples, as the most copious group (Table 1; Fig. 2A,D). Some grains were still lodged together. The morphotype was characterised by a bimodal distribution, or rather co-presence of large and small granules. Occasionally, the morphology was not completely intact, probably due to chewing as well as grinding and/or cooking procedures. These starch grains were similar to those occurring in caryopses of cereals, such as Hordeum sp. L. and Triticum sp. L. In particular, the diagnostic starches were oval to sub-round in 2D shape (15–43 µm in length; 10–35 µm in width). They had a central and distinct hilum and, sometimes, no visible lamellae. The small granules (≤ 10 μm in diameter) were spherical in shape with a central hilum. Knowledge about the Neo-Eneolithic period in central Italy is characterized by discontinuous data. The archaeobotanical dataset available for Latium is still limited⁴⁹ but information about cultivated and wild-collected plants from Casale del Dolce site exists. In fact, the carpological analysis previously conducted⁵⁰ has identified several caryopses of barley and wheat, supporting our results. The recovery of these starch grains, in almost all samples, suggested that the use of cereals was common and probably frequent for Casale del Dolce people, even if it is quite difficult to correlate presence of plant remains in calculus and quantity of consumed food²⁶. The hypothesis of cereal consumption for this community has been also proposed by stable isotope data. Isotopic values would suggest a subsistence economy based on a great intake of carbohydrates and a lifestyle characterized by a progressive agricultural exploitation, even more evident than other Eneolithic sites of central Italy^6,51. Lastly, Triticeae starches have been also found in dental calculus from Grotta dello Scoglietto (southern Tuscany), for the same pre-historical period⁵².

Morphotype II A low number of starch granules with faceted shape, perpendicular extinction cross and, sometimes, evident central fissures was recovered from dental calculus (Table 1; Fig. 2G,H). The morphology appeared oval to polygon (2D) with centric hilum and fissures radiating from it. The most frequent size distribution length was 14–25 μm in length and 13–17 μm in width. This type of grains exists in seeds of grasses belonging to the Andropogoneae Dumort. and Paniceae R. Br. tribes, as shown in the modern reference material¹⁹. Since an overlap in size and shape occurs among starches of species related to these tribes, the identification of these plant remains is arduous at a lower taxonomic level. Sorghum sp. Moench (sorghum), Setaria sp. P. Beauv. (foxtail millet) and Panicum sp. L. (millet) can be considered as potential candidates. Unfortunately, no phytolith, which would have helped us in distinguishing between the species of Paniceae⁵³, was detected. In addition, the lack of an isotopic signal specific for this type of consumption and the absence of relative carpological remains for the archaeological site of Casale del Dolce might be due to a limited usage of these plants. In fact, although several species of these genera were diffused in Italy, little is known about their employment. The archaeobotanical evidence of millets (i.e., Panicum sp. and Setaria sp.) from the Late Neolithic period has been discussed; however, their cultivation is certain during the Bronze and Iron Ages^52,54,55. Recently, Accelerator Mass Spectrometry-datings of prehistoric charred broomcorn millet grains has pinpointed the earliest occurrence of Panicum miliaceum L. in Europe at the middle of the 2nd millennium BCE (Middle/Late Bronze Age)⁵⁶.

Morphotype III Only one grain contributed to the third type of starch (Table 1; Fig. 2B). It appeared to be consistent with the Fabaceae family, probably Vicieae (Bronn) DC. tribe (e.g., vetches) for its oval to elongated (irregular) shape and kidney-like. The hilum was obscured and sunken, while the lamellae were not fully visible. The size was 42 μm in length and 30 μm in width. Data about pulses are scarce for this period. In northern Italy, a high variety of pulses was already present in the Neolithic^57,58 but this starch grain would seem to be one of the few and unique evidence of consumption in central and southern Italy. As this finding refers to a single individual, certainly, it is not expected to provide an exhaustive image of the use of pulses for the period and region but its presence, together with the carpological remains of Fabaceae^49,50, could attest plant protein consumption.

A single starch granule was not classified because missing diagnostic and distinguishable characteristics. Probably modification events, such as grinding process, cooking procedure in water and/or chewing, and exposure to alfa-amylase, altered its shape.

Pollen grains

Four calculus samples showed the presence of different pollen types (Table 1). In total, 49 grains were found. Three of them were detected in CDD2, 4, and 9 (Fig. 2C,EF), while the remaining ones (46), both in single and in aggregate form, were retrieved from only one individual (CDD7) (e.g., in Fig. 3). All palynomorphs were identified according to morphometric parameters described in literature and evidenced in the Palynological Database⁵⁹ and the names of the pollen types refer to literature^60,61,62.

In this paragraph we describe the pollen grains found in CDD2, 4, and 9 samples.

The ancient microremain embedded in sample CDD2 was apolar and medium in size (63 µm in diameter), showing a morphology which typically occurs in Poaceae^63,64. The stenopalynous nature of such type of pollen (that is, uniform monoporate) makes its systematic identification difficult. Although a low taxonomic determination limits paleoecological inferences, the evidence of Poaceae pollen is usually interpreted as indicative of open grasslands⁶⁵.

One ancient palynomorph displayed morphological traits consistent with Pinaceae (sample CDD4). It appeared as a bisaccate monad with an elliptic corpus and medium reticulation on bladders^59,66,67. Including sacci, the dimension was 56 µm in equatorial view.

A non-saccate Cupressaceae-type pollen, instead, was found in sample CDD9. It appeared spherical (with polar and equatorial axes of 30 µm) and inaperturate at light microscope; the protoplast exhibited itself star-like. Pollen grains produced by several species of Cupressaceae are considered morphologically uniform⁶⁸. Since prehistoric times, Gymnosperm wood has been widely used as raw material and firewood, while needles, nuts and inner bark represented the edible parts of these trees⁶⁹. Noteworthy is that the resins of these plants, possessing adhesive qualities and antibacterial properties, might have been also appreciated by Neanderthal¹⁴. Cupressaceae pollen grain is generally scarce in ancient sediments and one of the most underrepresented palynomorph in archaeological context. Several archaeobotanical studies have demonstrated the use of Juniperus L. species in the Mediterranean basin since the Holocene. In particular, the use of them as a source of aromatic foliage and resins employed for medicinal purposes, wood as fuel and for construction of dwellings, and fresh or dried berries as food has been proposed^70,71,72. Sporadic fossil discoveries of Cupressus sp. L, instead, are rather sparse in the Mediterranean area, although some ancient record has been registered in Italy during the Quaternary⁷³. Thus, the investigated plant microdebris testify the presence of Cupressaceae and provide additional evidence about the possible existence of evergreen Mediterranean forests, during the Neo-Chalcolithic period, in the Sacco River Valley.

Pollen grains in CDD7

CDD7 specimen (Fig. 1B), an adult male affected by severe malocclusion, preserved an interesting set of microparticles at microscopic analysis; therefore, we decided to report and discuss separately the data obtained from his calculus.

Eleven pollen grains out of 46 were not distinguishable due to the lack of diagnostic characteristics. The remaining 35 were found (singly, in pairs, or aggregates; Table 1, Fig. 3) in good or excellent state of conservation. The latter appeared as clusters of Pinaceae pollen (Gymnosperm) and other palynomorphs, including spores. Examples are shown in panels A and B of Fig. 3.

Two Cupressaceae, ten Pinaceae and one Poaceae pollen, presenting the same morphological features described in the previous paragraph, were also found in this sample (e.g., see Fig. 3C,E,F,G).

In addition, pollen grains from four herbaceous plants, namely Cyperaceae, Urticaceae, Trifolium, and Alchemilla species, and from the arboreal genus Corylus L. were detected and aredescribed below. Although pollen morphological variation within Cyperoideae subfamily is notable, one ancient microremain, possessing a pear-shape and a scabrate sculpture on its surface, appeared belonging to the genus Carex^74,75. In equatorial view it was triangular and the polar axis length was 41 µm. A second pollen grain was recognised as Urticaceae-type; it exhibited spheroidal shape (equatorial diameter 23 μm) and scabrate ornamentation. This morphology occurs both in Parietaria sp. and Urtica sp. pollen grains^59,62 and it is very difficult to distinguish them by optical microscope, especially if degraded. The shape of a third ancient monad, attributed to Trifolium-type (Fabaceae), was subprolate in equatorial view (46 μm) with scabrate ornamentation⁷⁶. The Alchemilla-type (Rosaceae) microremain (26 μm equatorial view, Fig. 3I) was radially symmetrical, elliptic and prolate in shape⁷⁷. Finally, another pollen type was found and attributable to Corylus sp. L. (Betulaceae). It was oval in equatorial view (19 μm), smooth, and tripolar with deep oncusis in each pore⁷⁸.

Seven pollen grains were single, prolate, isopolar, and elliptic in equatorial view (polar axis 19–25 µm long). They were tricolpate, with long and narrow colpi. Pores were at times indistinct. Pollen of the different species of Fagaceae shows a high variability in form, size, sculpturing; for this reason, most of them overlap in morphology. The ancient palynomorphs in exam were closely similar to a Quercus-type (examples in Fig. 3A,H)^79,80.

The last 10 grains showed a morphology (3-colpate, reticulate and subprolate) ascribable to Brassicaceae pollen grains (example in Fig. 3D). This is a stenopalynous family in which pollen varies among the genera but rarely in the species under the same genus^81,82.

Intriguingly, pollen findings in sample CDD7 were numerous and deriving also from insect-pollinated plants (e.g., Brassicaceae). This evidence appeared like a honey palynospectrum. This type of assemblage has been never registered in dental calculus deposits and, especially for the aggregates, the hypothesis of accidental inhalation seems implausible. Precisely, the presence of aggregates induced us to reflect upon a common origin of the whole pollen record. However, for single granules, to date, the supposition of aspiration cannot be completely excluded, due to the multiple pathways of inclusion of such type of microparticles²⁷. The high pollen variety could be explained by the presence of residues of natural matrices, as well as honey or beehive products (e.g., wax, propolis), in the calculus sample. To support our hypothesis, we prepared a reference collection based on modern multifloral honey samples (Supplementary Information 1, panel E–J).

Archaeological finds of bee products are quite rare^{83,84,85,86,87,88}. Since the end of the upper Palaeolithic, honey has been employed as sweetener, while beeswax for technological, ritual, cosmetic and medicinal applications^89,90. Regarding the latter, Bernardini et al.⁹¹ have found fascinating traces of a filling with beeswax, highlighting Neolithic dentistry procedures. It is important to recall that bees may also visit non-nectariferous plants (e.g., Poaceae, Betulaceae like Corylus sp.) for collecting pollen as protein source. Moreover, Pinaceae (Pinus sp. L. and Abies sp. Mill.) and Fagaceae (Fagus sp. L. and Quercus sp. L.), among others, emit sweet secretions and may be classified as honeydew producers⁸⁸. Therefore, it is not unlikely to discover pollen grains of pine, hazel, oak, and cereals mixed with melliferous taxa. In fact, similarly, Carboni et al.⁹² have observed a lump of pollen inside an Eneolithic vessel, suggesting the use of a fermented honey-based drink, the mead, for ritual purposes.

According to all this evidence, the pollen record detected in the present ancient calculus could be likely interpreted as direct honey consumption and/or remain of food or beverage including honey as natural sweetener. However, the use of conifer resins as antimicrobial or flavouring agents, mixed to honey or alone, cannot be excluded, together with the hypothesis of inhalation of bisaccate pollen from the immediate environment.

Unfortunately, for the investigated site, no evidence supporting the previous hypotheses exists. Nevertheless, it is possible that the individuals from Casale del Dolce practised bee-keeping culture near woodland pastures, although this interpretation cannot be definitive.

Currently, pollen spectra from beehive products are used to deduce plant biodiversity of the areas visited by insects for nectar collection^93,94. Bearing in mind this indication and the typical habitats of the identified plant taxa, some ecological implications were inferred. A thermophilic broad-leaved forest mainly made up of conifers (such as Pinus) and several deciduous trees (such as Quercus and Corylus), together with wet grasslands (Cyperaceae, Urticaceae, Alchemilla sp.), was outlined by pollen analysis. This hypothesis would seem consistent with Coubray’s work⁵⁰, who has identified the wood charcoals found in the archaeological site of Casale del Dolce as Carpinus L., Quercus, Maloideae, Cornus L., Corylus, Ulmus L., Fraxinus L., and Acer L. remains. In addition, palynological analyses performed in the same region^95,96,97 have detected similar vegetational elements.

Other plant microremains

We detected an unusual range of microparticles, that is, fragments of plant tissues and a sporangium, rarely documented in human dental calculus investigations (Table 1)^69,98,99,100.

Among the first, one microparticle was made up of plant cells associated to a scalariform xylem vessel (Fig. 2I), while another debris showed wood cells with simple pits (Fig. 2J). A brown-yellowish fragment was also photographed (Fig. 2K). As reported in literature⁹⁹, no evidence of charring or burning may be attributed to this type of darkening colouring but, if so, it would suggest an involuntary inhalation of ash particles from trees or shrubs used for fire. Thus, this type of microremain could derive from both non-edible and edible plants. In general, all these fragments retrieved from calculus might be the result of some activities, such as chewing of fresh plant organs, food and/or other uses of bark, oral hygiene procedures with woody dental picks, and/or use of teeth as a third hand^99,101,102.

The second type of uncommon microparticle, found in sample CDD7 (Table 1), appeared morphologically like a sporangium, probably from Monylophyta (Fig. 3J). It was brownish in colour and ovoid in shape. This type of microremain has never been observed in so ancient human dental calculus. A more specific taxonomical identification is very complex and would be risky, since at palaeobotanical and/or archaeological level there is no evidence to support this finding. However, considering that sporangia are typically attached to the abaxial surface of the leaf and that airborne dispersal capability of fern spores into stronger wind currents is rare and improbable^100,103, the recovery of the whole sporangium allowed us to hypothesize a voluntary use of fern leaves.

Biochemical analysis

GC–MS approach revealed the presence of organic compounds derived from the matter ingested and/or inhaled by the individuals. However, the potential of the biomolecular approach on dental calculus is still highly challenging and the capacity to trace the origin of some molecules is still difficult, due to the multifactorial dental calculus’s aetiology^31,104.

In Supplementary Information 2 (SI2), the molecules detected in each sample were listed and clustered in chemical classes. The chromatographic profiles were dominated by a series of C₆ to C₃₀ n-alkenes and n-alkanes, not reported in SI2 because ubiquitous and not taxonomically specific. They could probably come from degradation of oral bacteria or consumed food, representing, for instance, fragments of unsaturated or saturated lipids^{14,105,106,107}.

The typology of residues accumulated in dental calculus and their adsorption capacity determine the lipid profile of this matrix, considering that different foods naturally possess variable lipid composition. For this reason, it is difficult to associate fatty acids to specific dietary sources. The presence of fatty acids (e.g., odd, short, and long chains), ubiquitous components of organic matter, could be considered indicator for consumption of animal fats or plant oils (e.g., oil-rich seeds and fruits)^{104,108,109,110,111,112,113}. Long-chained polyunsaturated fatty acid derivatives (PUFAs; e.g., eicosapentaenoic acid, EPA), abundant in dried fruits¹¹⁴, were detected in some samples. Polyunsaturated omega-3 fatty acids have been rarely identified in archaeological contexts¹¹⁵, due to their highly inclination to oxidative alteration¹¹⁶. However, dental calculus has shown itself conservative for this type of molecules³¹. The consumption of aquatic organisms cannot be excluded, being rich of PUFAs¹¹⁴ and considering the proximity of the ancient settlement to the Sacco River.

Monoterpene derivatives, non-specific compounds with volatile nature, retrieved from some samples, such as citronellol, menthol and pinanol (commonly found in leaves, fruit, and bark of a wide range of plant species), could generically indicate the ingestion of plant materials or waxes¹⁰⁹.

In CDD5 calculus, azulene and coumarin derivatives were also recovered. These secondary metabolites usually occur in species belonging to Apiaceae, Asteraceae, and Rutaceae families, well known medicinal plants possessing a wide range of biological activities^117,118. As suggested by Hardy et al.¹⁴, the plant species rich in such type of bitter-tasting compounds might have been ingested for self-medication.

Two alkaloids were found: trigonelline and hordenine, respectively, in CDD4 and CDD7 specimens. The first one, whose accumulation takes place in various plant species (i.e., Achillea sp. L.) and especially in Fabaceae seeds (e.g., Trigonella sp. L., Trifolium sp. L., and Medicago sp. L.)^119,120, might represents a further proof for consumption of pulses.

Hordenine, which naturally occurs in certain grasses, like cereals (e.g., barley, millet, and sorghum)¹²¹, could demonstrate the ingestion of starchy material, as already testified by the detection of a Triticeae starch granule in the same calculus flakes and the recovery of caryopses at the site⁵⁰.

Source: Ecology - nature.com