Fossil feathers from the Colli Albani volcanic complex (Late Pleistocene, Central Italy) preserved in zeolites Open Access

Fossils preserving evidence of soft tissues are essential for our understanding of the evolution of life and ancient biodiversity. These fossils are usually associated with fine-grained sediments deposited in lacustrine and marine environments or are preserved as inclusions in amber. Recent reports show the importance of volcanic material, e.g., fine-grained surge and ash deposits, for soft tissue preservation in many Konservat-Lagerstätten (El Albani et al., 2024; Yang et al., 2024). Many of these volcanic ashes were deposited in aquatic environments, such as lakes (Yang et al., 2024) or shallow marine waters (El Albani et al., 2024). In subaerial environments, volcaniclastic sediments can facilitate the preservation of skeletal remains (e.g., Pompeii and Herculaneum, Italy [Petrone, 2019]; Jehol Biota [China] preservation type B sensu [Pan et al., 2013]) and plant remains (Góis-Marques et al., 2019). Preservation of animal soft tissues in subaerial volcanic sediments is rare and poorly understood (Rößler et al., 2012).

The Late Pleistocene Peperino albano pyroclastic deposit (Colli Albani, Rome, Italy) hosts a remarkable specimen of the griffon vulture Gyps fulvus with extensive preserved plumage (Marra et al., 2022; Meli and Maineri, 1889). This specimen offers a rare opportunity to investigate the taphonomy of soft tissues in volcanic sediments. The taphonomy of the plumage has not been investigated, and the mode and quality of preservation of the feathers is unknown. Previous studies of the external mold of the head and neck hypothesized that the specimen was buried by a low-temperature (<100–250 °C; Iurino et al., 2014) volcanic deposit, more specifically, either a moderately turbulent pyroclastic surge or a secondary volcaniclastic flow (e.g., a mud flow or lahar).

Here, we investigate the anatomical and chemical preservation of the plumage of the griffon vulture using scanning electron microscopy (SEM) coupled with energy dispersive X-ray spectroscopy (EDS), nano-X-ray fluorescence analysis (n-XRF), micro-Raman spectroscopy (µ-Raman), and micro-Fourier transform infrared spectroscopy (µ-FTIR) (see Supplemental Material¹). Our results reveal that fossil melanosomes and feather cortex are mineralized in nanocrystalline zeolite. This represents a new mode of preservation that has the potential to apply to various fossil soft tissues and taxa hosted in fine-grained, ash-rich sediments in terrestrial pyroclastic deposits.

The fossil feathers are yellow to rust-orange in color compared to the surrounding gray-colored zeolitized rock matrix (Supplemental Material). Most are preserved in dorsal view and parallel to bedding (Figs. 1 and S1–S9). All preserved feathers comprise only the middle portion of the central shaft (i.e., rachis) and the plumed part, including barbs and barbules (i.e., vane); the most distal and proximal portions of the feathers (including the calami) are not preserved. The rachi are preserved as external molds that are typically beige-orange in color and often show a longitudinal ridge (Fig. 1C); the latter corresponds to the position of the axial groove, common in feather rachi in extant birds (Chang et al., 2019). The rachis medulla and ridges are not preserved. Visual assessment in hand specimen and thin sections suggests that the barbs and barbules are preserved in three dimensions (Figs. 1 and S1–S10).

SEM images reveal variations in the fidelity of ultrastructural preservation of the feathers (Figs. 2, S11, and S12). Barbs and barbules are usually well preserved (Figs. 2A, 2B, and S10); the latter can show hamuli (hooklets; Fig. 2C). Both fractured and polished sections often show details of internal structures and a thin homogenous nanocrystalline outer layer (~10 µm thick in barbs and ~1 µm thick in barbules) (Figs. 2F, 2G, and S12). This corresponds to the cortex in the feathers of extant birds (Fig. S11). Internal to this outer layer, the barbules contain densely packed, round microbodies (Figs. 2F, 2G, 3, S12, and S13) (0.5–1 µm long). These are observed exclusively within the barbules. The size and shape of the microbodies, plus their location and packing within barbules, are consistent with melanosomes (Shawkey and D’Alba, 2017). A bacterial or abiotic origin of the microbodies is excluded, as bacteria and abiotic precipitates would be expected to infill the barb central void and to be widespread in the sedimentary matrix.

EDS analysis shows that the feather cortex and sedimentary matrix are enriched in Si, O, and Al; melanosomes are enriched in Fe relative to the surrounding sedimentary matrix and feather cortex (Figs. 3A, 3B, and S13). C is not enriched in any feather structure. In the sedimentary matrix, certain individual mineral grains are enriched in Na, Ca, and K (Fig. 3). Nano-XRF analysis supports the EDS data and shows Ti, Cr, Mn, Zn, and Cu (Figs. 3D and S14) associated with the barbules. Collectively, these results indicate that the feathers have been replicated in three dimensions in an aluminosilicate mineral phase.

Micro-Raman mapping of several regions (n = 5) of thin sections (n = 2) reveals the presence of large crystals of augite, garnet, and leucite, whereas the fine matrix is formed primarily of analcime and hematite (Figs. S15 and S16); rare metal oxides are also observed (Fig. S16). Notably, laser damage is observed on the barbule surface but not in the surrounding matrix, suggesting differences in crystallinity (Figs. S17–S19).

Raman point analysis of regions of the barbule surface that lack laser damage provides a Raman signature that is consistent with a mineral composition (Fig. 4A). The Raman signal shows a strong, sharp peak at ~482 cm⁻¹ and a weaker peak at ~393 cm⁻¹. The first peak is consistent with the vibrational mode of T-O-T bonding (T denotes Si or Al atoms; Finocchiaro et al., 2022). The weak peak at ~393 cm⁻¹ can be tentatively assigned to the vibration modes of M-O bonds (M denotes a metal atom) (Finocchiaro et al., 2022). Point analysis of the white crystalline mineral phase adjacent to the barbules shows similar major peaks (Fig. 4B). These peaks are consistent with those for certain zeolite minerals such as analcime (Tsai et al., 2021). Moreover, the orange-colored surface of the barbules shows sharp peaks at ~222, 290, 406, and 610 cm⁻¹. These peaks can be assigned to the Fe-O bond stretching vibrations typical of hematite (Mohammed et al., 2018).

Micro-FTIR analyses of barbules (n = 4) in thin section show only two major peaks: a strong peak at ~1000 cm⁻¹ and a sharp, moderately strong peak at ~1640 cm⁻¹ (Fig. 4C); these peaks are assigned to the stretching vibration of Si-O and O-H bonds, respectively (Mozgawa, 2001). This FTIR signature is consistent with that of nanocrystalline zeolite (Taufiqurrahmi et al., 2011). FTIR analysis did not recover peaks that can be assigned to preserved melanin and/or keratin (or their diagenetically altered molecular products) (Slater, et al., 2023).

Our SEM investigations reveal a remarkably high fidelity of anatomical preservation of feather ultrastructure, including barbule hamuli, barbule cortex, and melanosomes. In addition, our chemical investigations using EDS, n-XRF, µ-FTIR, and µ-Raman strongly suggest that the feather tissues are replicated in three dimensions by nanocrystalline zeolite. The orange coloration visible in hand specimen is due to the later precipitation of iron oxides/oxyhydroxides, such as hematite.

Three-dimensional preservation of vertebrate soft tissues with a high level of ultrastructural fidelity is usually associated with replication of tissue structure in calcium phosphate (i.e., phosphatization; McNamara et al., 2016). A similarly high level of anatomical fidelity can also be achieved via replication in silica (i.e., silicification [Yang et al., 2024] and organic preservation [McNamara et al., 2010]), especially in amber (Xing et al., 2016). Preservation of soft tissues via clay templating (i.e., aluminosilicification) has been reported for soft tissues in diverse fossils (Cai et al., 2012; Gabbott, 1998); in these cases, however, the tissues are organically preserved and not replicated by clay minerals. Replication of fossil soft tissues in zeolites has not been reported previously; clearly, this new mode of soft tissue preservation has the capacity for a remarkable quality of preservation, extending to the microscale and three dimensions.

This mode of preservation is also unusual in the context of the tissue type involved. Fossil feathers are almost always preserved as dark carbon-rich films (Cincotta et al., 2020); this mode of preservation is typical in lacustrine and restricted marine sedimentary deposits. Feathers can also be preserved in three dimensions as inclusions in amber (Xing et al., 2016) and rarely as impressions (i.e., an external mold lacking organic components; Wellnhofer, 2010), but examples of mineralized feathers are particularly limited in the fossil record. One such example is a specimen of the American coot (Fulica americana) from a Holocene hot spring deposit (Wyoming, USA; Channing et al., 2005). Similarly to G. fulvus, the body of the F. americana specimen is preserved as an external mold, and part of the plumage is preserved in three dimensions; unlike G. fulvus, however, the feathers are preserved in silica and with a low fidelity of preservation at the microscopic level (Channing et al., 2005). Mineralization of feathers in calcium phosphate and gypsum has also been reported, but a detailed investigation was not conducted (Davis and Briggs, 1995).

Zeolites are a class of aluminosilicate tectosilicates that occur as amorphous to crystalline, primary, or secondary minerals in volcanic (i.e., lava and pyroclastic deposits), volcaniclastic, and sedimentary rocks. Zeolitization occurs naturally post-deposition in pyroclastic deposits due to the interaction between glassy components (i.e., glass shards and pumice) and water. This process is controlled by fluid chemistry (e.g., pH, Eh), lithology (including the abundance of glassy components), and physical parameters such as temperature, pressure, porosity, and the duration of the reaction between fluid and glass (Chipera and Apps, 2001; Cicerali et al., 2020). Hydration of volcanic glass and ash will mobilize Si⁺³, Al⁺⁴, and metal ions, which, in turn, will precipitate as hydroaluminosilicate precursors that transform subsequently to zeolites (Cicerali et al., 2020). Zeolites have complex physico-chemical properties and have been induced to form synthetic nanocrystalline structures in laboratory experiments using various temperature and pressure conditions, including room conditions (Rani and Srivastava, 2020; Valtchev et al., 2005). Zeolitization clearly does not require burial, but could occur subaerially within days (Valtchev et al., 2005). Intriguingly, laboratory experiments have demonstrated that synthetic nanocrystalline zeolites can nucleate on biological tissues (e.g., the membrane of eggs [Valtchev et al., 2008] and plant material [Valtchev et al., 2003]), resulting in the replication of tissue microstructure.

The incomplete anatomy of the specimen and the unknown stratigraphic position limit our interpretation of the depositional setting (Supplemental Material). Despite this, our study provides important new insights that are the basis of a taphonomic model. The vulture carcass was buried in a low-temperature, ash-rich flow that lithified relatively quickly. The early lithification of the sediment promoted the retention of the three-dimensional shape of the body and feather structures. The initiation of zeolitization of the ash component of the rock matrix might have promoted the mineralization of the plumage. The local pH (i.e., surrounding the decaying carcass; Clements et al., 2022) favored the interaction of Al⁺³ and Si⁺⁴ ions with the protein-rich feather cortex. Mineralization of melanosomes in the barbules, but not the barbs, was likely facilitated by the thin barbule cortex. The lack of internal structure in the rachi and barbs, e.g., medulla and barb melanosomes, indicates that these latter internal structures decayed before completion of the mineralization process. Laboratory experiments have demonstrated that rapid and early mineral precipitation can lead to micro- and nanoscale soft tissue preservation within days to weeks post-mortem (Naimark et al., 2016; Slagter et al., 2021). Although many aspects of the zeolitization process are unknown, some speculation on the nucleation process is possible. It is plausible to infer that—as per mineralization in silica—certain functional groups common in decaying protein-rich tissues (in particular OH, NH, and CO) actively bind aluminosilicate ion species via hydrogen bonding, electrostatic interactions, and/or cation bridging (Slagter et al., 2021).

By identifying zeolitization as a new mode of soft tissue preservation that can preserve anatomy at the microscale, our findings represent an important advance in the study of fossil soft tissues. Terrestrial volcanic settings clearly have the capacity for remarkably high-fidelity preservation of tissue ultrastructure and should be actively targeted in future studies as loci for soft tissue preservation.

We thank two anonymous reviewers and M. Benton for constructive criticism that strengthened the manuscript, and M. Albano, T. Ruspandini, D. Mannetta, S. Ronca, A. Somogyi, P. Chung, C. Maragoni, M. Gattabria, and A. Reverbi for support. Rossi was supported by a Sylvester-Bradley Award (PA-SB202003; Paleontological Association) and a Norman Newell Early Career Grant (Paleontological Society). Rossi, Slater, Carazo del Hoyo, and McNamara are supported by an ERC (European Research Council) Consolidator grant H2020-ERC-CoG-1010003293-PALAEOCHEM awarded to McNamara.