We thank the Muséum national d’Histoire naturelle, Paris, in particular Violaine Colin, and the Natural History Museum, London, in particular curator Roberto Portela Miguez, for granting access to their reference collections and all the subsequent help. Financial support for these missions was provided by UNIARQ through FCT (Fundação para a Ciência e Tecnologia) project UIDB/00698/2020, to Cleia Detry, and by ICArEHB (Interdisciplinary Center for Archaeology and Evolution of Human Behaviour, University of the Algarve), to João Luís Cardoso. The archaeological excavation and study of Figueira Brava cave was supported by FCT grants to João Zilhão (PTDC/HIS-ARQ/098164/2008 and PTDC/HAR‑ARQ/30413/2017). Miguel Telles Antunes prepared the fossil, José Paulo Ruas took the specimen’s photos in fig. 3, and Gloria Cuenca-Bescós helped with references and for the very helpful comments and advice.

1 - Introduction

1Hystricidae (Fischer, 1817), a family of large rodents known as Old World porcupines by contraposition to the New World porcupines of the family Erethizodontidae, The Hystricidae have a line of fossil representatives stretching back to the Early Miocene, c. 20 MA (millions of years) ago (Barbière & Marivaux, 2015). Their scarcity in the fossil record is a characteristic of the whole family, represented today by only three genera: Hystrix, Atherurus and Trichys. In Europe, the area of distribution of porcupines declined from the Pliocene to the Pleistocene, probably due to the increasingly colder and unstable climactic conditions (Montoya, 1993). The family’s present-day distribution includes all of Africa and tropical and sub-tropical Asia, with a small nucleus in Italy (Rosevear, 1969; Monadjem et al., 2015).

2Because of marked intra-specific variability, namely in size and shape, the taxonomy of this family, and especially so that of the Hystrix genus, remains a focus of intense debate (Sallari & Sardela, 2009). This is the more so with regards to the family’s fossil representatives, despite the phylogenetic studies already undertaken (Rovie-Ryan et al., 2017). Of the three genera recognized among the Hystricidae, the one of relevance here, Hystrix, features eight extant species: H. cristata, H. africaeaustralis, H. indica, H. brachyura, H. javanica, H. crassispinis, H. pumila and H. sumatrae (Wilson & Reeder, 2005).

3In Europe, most Pleistocene remains of Hystrix have been assigned to H. primigenia or H. major, the former apparently older and smaller than the latter (Koliadimou & Koufos, 1991; Montoya, 1993). However, a large number of other taxa have been named, namely H. vinogradovi, a Middle Pleistocene form found in forested contexts and thought to be significantly smaller than H. cristata (Gallindo-Pelicenna et al., 2011). In Iberia, all late Middle or Upper Pleistocene fossil remains of Hystrix have been assigned to H. vinogradovi, H. brachyura or even H. brachyura vinogradovi. The single exception is a specimen from the Mousterian levels of Devil’s Tower (Gibraltar), which Garrod et al. (1928) identified as H. cristata.

4The specimen described here is the first record of H. cristata (Linnaeus, 1758), the crested porcupine, in Portugal. It was found at Figueira Brava cave (38°28’14”N, 8°59’10”W; WGS84 datum), in a Middle Palaeolithic archaeological context dated to the later part of the Last Interglacial period (figs. 1 & 2).

5Nowadays, H. cristata is found mainly in Africa, where it prefers hot and dry climates. It is spread across the northern coasts of Morocco, Algeria, and Tunisia, with a small nucleus in the coast of Libya. It also occurs in shrublands, grasslands and forests along a strip bordering the southern part of the Sahara, from the Atlantic to Ethiopia (Amori & Smet, 2016; Aulagnier et al., 2018). An European population exists in southern Italy and Sicily as a result of human introduction in Late Antiquity or Early Medieval times (Santini, 1980; Riquelme Cantal & Morales Muñiz, 1997; Amori & Angelici, 1999;

6Masseti et al., 2010; Aulagnier et al., 2018), with a recent expansion to northern Italy having been noted (Mori et al., 2013). Garrod et al. (1928) mention earlier reports of the species’ presence in southern Spain during the xix^th century but so far no material evidence to that effect (e.g., osteological remains and field reports) has been forthcoming.

2 - Materials and methods

7Within all species of the genus Hystrix, dental wear introduces much inter-individual variation in tooth morphology, there exists a lack of studies addressing that variation, and enamel patterns have not changed much since the Neogene (van Weers, 2005; Tong, 2008). These factors complicate species differentiation in the fossil record (Sulimski, 1960; Fistani et al., 1996), where osteometrics, and in particular odontometry, are considered to provide the best discriminators (Tong, 2008). For instance, H. vinogradovi is defined as significantly smaller than H. cristata, and the latter as smaller than H. primigenia, with no overlapping (Fistani et al., 1996).

8Our identification is based on the morphological comparison of the Figueira Brava specimen with extant Hystrix from the reference collection of the Muséum national d’histoire naturelle (Paris). In addition, we measured the teeth of 28 mandibles from that collection identified as H. cristata and H. javanica, and of 99 mandibles from the Natural History Museum (London) identified as H. cristata (60), H. africaeautralis (16), H. brachyura (8) and H. indica (39). To reduce inter-observer variability, all tooth measurements, including those of the Figueira Brava fossil, were made by the same person. We then compared our data with published measurements of reference and fossil material (Sulimski, 1960; Montoya, 1993; Van Weers, 1994; Cuenca-Bescós et al., 1999; Baryshnikov, 2003; Van Weers, 2005; López-García et al., 2008; Salari & Sardella, 2009).

9Using a Mitutoyo Vernier calliper with a precision of 0.01 mm, we took two measurements: length of the mandibular PM4, and length of the mandible’s tooth series. The latter was taken from PM/4 to M/3. In the Figueira Brava specimen, however, the M/3 is missing, so in this case our measurement for the series was taken from PM/4 to the limit of the alveolar space. The potential source of error introduced by this resort must be borne in mind when assessing results drawn from the comparison of complete mandibles.

10Reference specimens measured for this study are deposited in the following institutions: Muséum National d’Histoire Naturelle (MNHN), Paris, France; Natural History Museum (NHM), London, United Kingdom. The Figueira Brava specimen is currently in storage at UNIARQ – Centro de Arqueologia da Universidade de Lisboa, Portugal.

3 - Description and context

11Our specimen bears inventory number FB-2013-838 and is a well preserved left hemimandible with PM/4, M/1, and M/2 (fig. 3). It clearly belongs to a rodent, one bearing hypsodont teeth with roots, as in all Hystricidae. The teeth have buccal and lingual infoldings that, with wearing, become infibulate (small enamel islets), producing a particular pattern typical of the genus Hystrix (Hillson, 2005).

12The fossil was found in unit IH8 of the stratigraphic succession excavated in Area F of Figueira Brava. IH8 is a variably cemented sandy loam at the base of the site’s FB4 phase, which is comprised of units IH2-IH8. Bayesian modelling of the U-Th dating results for flowstone and other stratigraphically associated speleothems allows us to constrain the succession with much precision and indicates that the accumulation of FB4 took place during GI (Greenland Interstadial) 22, i.e. 87.6-90.0 ka ago (Rasmussen et al., 2014). This dating is consistent with the paleonvironmental information provided by the remains of land vertebrates (including the tortoise, Testudo hermanii) and the plant charcoal assemblage, which is dominated by the stone pine, Pinus pinea, and where the Angiosperm component reveals a low-elevation, Mediterranean, limestone bedrock ecosystem with Olea europaea, Quercus sp., Prunus sp., Rhamnus-Phillyrea, Ficus carica, and Vitis vinifera. This evidence suggests climate and environmental conditions largely similar to those extant in the Holocene. For a detailed description of context, see Zilhão et al. (2020).

13Area F corresponds to the back of the original cave, where the deposit is protected by a speleothem-cumbreccia barrier situated at the interface with the nowadays external parts of the site. Most of the latter’s original sedimentary infill has been lost to marine erosion. IH8 accumulated atop flowstone unit IL1, which formed in the early part of a hiatus in sedimentation and in human occupation during which the site was used as a den by carnivores, namely bear and lynx. Porcupines also den in deep burrows and in caves (Amori & De Smet, 2016), where their presence is often revealed, if only indirectly, by the gnawing marks left on bones previously accumulated by carnivores, e.g. hyenas (MacDonald & Barrett, 1993; Diedrich, 2009). Once sedimentation resumed, the skeletal remains produced by the animal activity that took place at Figueira Brava during the hiatus became incorporated in the IH8 deposit, commingled with the stone tools, the charcoal and the anthropogenic faunal remains denoting human repossession of the site.

14The stratigraphic context described above implies that the porcupine mandible must have entered the site during the time interval calculated for the boundary between IL1 and IH8. Based on the Bayesian modelling of the sequence of dating results, that interval corresponds to the two millennia between 89.8 and 91.7 ka, towards the end of the ten millennia long GI 23 (90.1-104.0 ka) and at the very beginning of MIS (Marine Isotope Stage) 5b.

4 - Results and discussion

15Our results are presented in tabs. 1 & 2 and figs. 4 & 5. Fig. 4 shows that the Figueira Brava premolar is significantly larger than the few published measurements for H. vinogradovi (= H. brachyura vinogradovi) and is consistent with H. cristata. It also shows that the variation in lower premolar size documented among extant H. cristata encompasses the corresponding H. vinogradovi range even though both have premolars that, on average, are smaller than those of H. primigenia and H. major.

16One female H. cristata from Senegal curated at the NHM has a lower premolar (6.52 mm) smaller than most H. vinogradovi specimens (with a single exception, a 6.0 mm tooth found among the latter). These results suggest a wider range of variation in the cranial size of H. cristata than hitherto assumed. Indeed, most values for the extant and almost all values for the fossil lower premolars that we measured fall within the interval bracketed by the minimum and maximum values for H. cristata given in fig. 4.

17In fig. 5, we compare the size of the tooth series in extant species with four fossil specimens: two H. vinogradovi and two H. cristata, one from Figueira Brava (30.23 mm), the other from Devil’s Tower (30.0 mm). The two fossil H. cristata are in the range of extant H. cristata, H. africaeaustralis and H. indica. The two H. vinogradovi fall just below the range of extant H. cristata (tab. 2).

18Based on these comparisons, we conclude that the Figueira Brava mandible belongs in H. cristata. We further suggest that, on the grounds of odontometrics alone, little reason exists to go on accepting that the specimens so far assigned to H. vinogradovi represent a distinct taxon; in all likelihood, they are but small H. cristata.

19Even though attempts to phylogenetically differentiate the several Hystrix species have been made, the only exemplar of H. africaeaustralis studied by Rovie-Ryan et al. (2017) does not separate completely from H. cristata. These two species hybridize in captivity, but they do not breed in the wild, not even when they are sympatric (Barthelmess, 2006). Nevertheless, it remains possible that all recognized extant species of Hystrix are in fact but different populations of a single species, or at least very close species that diverged not long ago. With time, the rise of geographic or ecological barriers may have led to divergence, but the variation in size seen across both time and space could well be largely explained by adaptation to climate.

20According to Cuenca-Bescós et al. (1999), H. cristata is an indicator of warmer phases because of its current African distribution. Therefore, the presence of the genus in Middle and Upper Pleistocene contexts of Europe could stem from warm-phase immigration. During the coldest phases, porcupines would have persisted in Iberia only, because of its more southerly latitude and comparatively milder climate.

21No confirmed record of H. cristata exists in the Holocene of Iberia. The taxon is also entirely absent from archaeological and paleontological contexts of the peninsula’s Pleistocene that securely post-date the end of the Last Interglacial. In this respect, the porcupine contrasts with other warm-adapted taxa that would seem to have persisted well into MIS 3 and have natural, non-reintroduced extant populations — namely, the tortoise, Testudo hermanii (Jiménez-Fuentes et al., 1998; Nabais & Zilhão, 2019). Unlike cave lion, leopard, hyena, and other taxa that, nowadays, only exist in Africa but persisted in Europe until the end of the Ice Age (Cardoso, 1993; Davis, 2002), porcupines would therefore seem to have become extinct in Iberia with the onset of the first glacial maximum of the Upper Pleistocene, ca. 72 ka ago.

5 - Conclusion

22The morphology of teeth and mandible, as well as the size of the lower PM/4 and the length of the mandible’s tooth series place the Figueira Brava specimen well within the variation of extant and paleontological specimens of Hystrix cristata. It is therefore to this taxon, not to H. vinogradovi, the other species hitherto recognized in the Upper Pleistocene of Iberia, that we have assigned our fossil. Available H. vinogradovi measurements, however, fall within the lower range of extant H. cristata, and so the possibility must be entertained that the corresponding material represents smaller-sized individuals, or populations, of H. cristata; previously found in Gibraltar (at Devil’s Tower), this taxon’s Upper Pleistocene existence in Iberia is now corroborated in Portugal (at Figueira Brava). The secure and precise dating of the Figueira Brava specimen to the 89.8-91.7 ka ago interval shows that Hystrix persisted in the area until the end of the Last Interglacial. No reliably dated fossil occurrences are known thereafter, and so porcupines probably became regionally extinct with the onset of last Ice Age. Human reintroduction in Late Antiquity explains the extant Italian population as well as, in all likelihood, the unconfirmed reports that porcupines were present in southern Spain during the xix^th century.