Ectoparasitism and infections in the exoskeletons of large fossil cingulates

Studies on paleopathological alterations in fossil vertebrates, including damages caused by infections and ectoparasites, are important because they are potential sources of paleoecological information. Analyzing exoskeleton material (isolated osteoderms, carapace and caudal tube fragments) from fossil cingulates of the Brazilian Quaternary Megafauna, we identified damages that were attributed to attacks by fleas and dermic infections. The former were compatible with alterations produced by one species of flea of the genus Tunga, which generates well-delimited circular perforations with a patterned distribution along the carapace; the latter were attributable to pathogenic microorganisms, likely bacteria or fungi that removed the ornamentation of osteoderms and, in certain cases, generated craters or pittings. Certain bone alterations observed in this study represent the first record of flea attack and pitting in two species of large glyptodonts (Panochthus and Glyptotherium) and in a non-glyptodontid large cingulate (Pachyarmatherium) from the Quaternary of the Brazilian Intertropical Region. These new occurrences widen the geographic distribution of those diseases during the Cenozoic and provide more evidence for the co-evolutionary interaction between cingulates and parasites registered to date only for a small number of other extinct and extant species.


Introduction
Among the representatives of the South American Pleistocene Megafauna, the cingulates stand out as one of the most diverse and peculiar clades. In this group, three main lineages are traditionally recognized: Dasypodids, likely a paraphyletic group [1], with fossil and recent species, and the extinct pampatheres and glyptodonts [2].
Paleopathological studies about cingulates, as for other taxa, can provide paleoecological and evolutionary insights concerning organism diseases from past to present. Cingulates show a complex exoskeleton overlaying their head, back and tail [3] formed by the fusion or articulation of dermal ossifications (osteoderms), whose external ornamentation varies considerably among species and are abundantly represented in the fossil record. However, few works in paleopathology have focused on exoskeleton diseases, although the inclusion of this structure a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 In the BIR, several types of fossil deposits occur [14,15]. The fossils analyzed in this work were collected from sedimentary infillings of limestone caves and natural tanks and in alluvial deposits. Most available absolute dates place the megafauna remains from tank and cave deposits from the BIR in the Late Pleistocene [15,16], with few exceptions assigned middle Pleistocene [17] and early Holocene ages [18,19]. Therefore, we considered the material studied herein as belonging to the Late Pleistocene. The specimens included the following: isolated osteoderms and carapace fragments of P. brasiliense and Glyptotherium sp. from a limestone cave in the Lajedo da Escada (5˚14´31"S, 37˚44'20"W), Baraúna Municipality, Rio Grande do Norte State; isolated osteoderms of P. brasiliense collected in a tank in the Fazenda Nova locality (8˚11'2"S, 36˚10'01"W), Brejo da Madre de Deus Municipality, Pernambuco State; one partial carapace and isolated osteoderms of the genus Panochthus collected in a tank in the Lagoa do Santo locality (19˚61'46"S, 44˚03'09" W), Currais Novos Municipality, Rio Grande do Norte State; two fragments from the lateral portion of a caudal tube of Panochthus collected in a fluvial deposit in Barra do São Miguel Municipality, Paraíba State; and isolated osteoderms belonging to this last genus collected in a muddy layer near a river [20] in the Povoado Caboclo locality (08˚30'54"S, 41˚00'18"W), Afrânio Municipality, Pernambuco State. Natural tanks are common fossiliferous deposits in the BIR and correspond to natural depressions caused by erosion of fractures in basement rocks by physical-chemical weathering, forming small pluvial water reservoirs [21,22,23]. These tanks are filled with clasts and bioclasts deposited primarily by hydraulic transport under a debris flow regime [23].
The exoskeleton elements of Glyptotherium collected in the Lajedo da Escada site were found in association with several postcranial bones that might belong to a single individual [24]. The partial carapace of Panochthus in the Lagoa do Santo site were also found in association with isolated osteoderms and with several postcranial bones [25].

Methods
We restricted our analyses to osteoderms with a preservation level of 50% or more, which allowed a reliable analysis of pathologies and more secure taxonomical identifications. We conducted a macroscopic inspection [26], useful to detect alterations in bone surfaces. Additionally, we also used a stereoscopic magnifying glass, a digital microscope Dino-Lite Basic with DinoCapture 2.0 Software and digital caliper with accuracy of 0.01 mm to scale out the lesions. We based our diagnoses on comparisons with previously reported cases in the literature [10,8,27]. We used taphonomically and pathologically unaltered osteoderms of Panochthus and Glyptotherium to compare with the affected ones. The unaltered osteoderm of Panochthus (132-V-UERN) was collected in a natural tank from Taperoá Municipality, Paraíba State, Brazil, and is deposited in the collection of the Laboratório de Paleontologia of Universidade do Estado do Rio Grande do Norte (LABPALEO-UERN), Mossoró city, Rio Grande do Norte State, Brazil. The osteoderm of Glyptotherium was collected in the Lajedo da Escada site, as was the remaining material of this genus analyzed here, and deposited in MCC.
In this work, we adopted particular anatomical terminologies for different parts of the exoskeleton. For the carapace as a whole and caudal tube, we used the terminology proposed by Porpino et al. [28] (see Fig 2 of this work). For description of osteoderms (isolated or in carapace fragments), we adopted the anatomical terms commonly used in the literature [29,14,28,30] regarding the ornamentation of their external surface (Fig 2). In this context, we used the term main figure for the largest figures in osteoderms instead of the term central figure. We considered as fragments of the carapace the elements comprising two or more fused osteoderms.
For the purposes of the present work, which concerned ante mortem alterations, we treated pitting and perforation as terms referring to different types of bone bioerosion. Pitting or cratering was cavities generated horizontally by erosions in the external cortical surface and with the potential to penetrate into trabecular bone (See Fig 3A in [27]). By contrast, a perforation is a vertical cavity that penetrated the bone that was also created by erosion, although showing well-delimited circular edges. Glyptodonts of the genus Panochthus have osteoderms that show a homogenous ornamentation along most of the carapace. The osteoderms next to lateral edge of the carapace have differentiated central and peripheral figures in contrast to the rest of lateral and dorsal regions, which, in most species, have only small figures [31,28] likely homologous to the peripheral ones. We observed alterations in isolated osteoderms and in different points of the partial carapace MCC 1603-V, which in this case, varied according to the ornamentation.

Panochthus
Description of the lesions. Missing was most of the anterior region of the partial carapace MCC 1603-V. Consequently, we could not evaluate whether this section was pathologically affected. Among the preserved parts, the posterodorsal and posterolateral lateral regions were affected by pitting associated or not with loss of ornamentation and bone response in some points.
The right posterolateral and posterodorsal regions showed loss of the ornamentation, which conferred to the affected areas an irregular and rough aspect due to the exposition of the trabecular tissue. The osteoderms of the edge of right posterolateral region showed a high frequency of microcavities in the main figures. Additionally, on this region, we observed multilocated loss of ornamentation, primarily in the lateral edges. Next to the posterodorsal region, we verified a crater surrounded by a crest created by bone response (Fig 3A). In the left posterolateral region, the ornamentation showed few signs of erosion; depressions formed cavities, although they were less frequent and larger than the similar features on the right posterolateral region (Fig 3C). We considered these cavities as pitting.
We observed a modified area with approximately 10 cm extending from the posterodorsal to the right posterolateral region, showing bone response with a contorted aspect, suggesting calcium deposition (Fig 3B). The ornamentation around this new bone growth remained unaltered, with a low frequency of pits. Inside this area, pits were associated with exposure of trabecular bone and microfractures. By contrast, in the margins of this area, the frequency of pits was high, characterizing a multifocal distribution of pits.
All injured isolated osteoderms from Currais Novos and Brejo da Madre de Deus sites had the external surface eroded resulting in soft to severe exposure of trabecular bone by pitting or other erosive process.
Among the isolated osteoderms and carapace fragment from Currais Novos alterations led to partial loss of the ornamentation in wide extensions of their external surface and the formation of crests by irregular and differential erosion (MCC 1412-V; Fig 4C and   4D). Few isolated osteoderms from the dorsal region showed pitting. As for Glyptotherium (next section), in Panochthus, few osteoderms showed bone growth in response to the lesions, and those responses, when present, conferred to the affected surface a contorted aspect (MCC 1653-V; Fig 5A). We identified some cases in which the loss of the ornamentation exposed the spongy bone (MCC 1632-V; Fig 5D). In digital microscopical view, we observed a twisted and contorted pattern of filaments overlapping the spongy bone (Fig 5B and 5C). Moreover, the digital microscopy of the eroded area without bone response showed the typical porous aspect of the spongy bone, but without sign of bone response ( Fig 5E). We also observed pitting in the main figure of posterior region osteoderms, which modified the normal aspect of the main figure by producing a depression with a midline crest (MCC 1653-V; Fig 4B); this feature was also observed in some osteoderms in the carapace MCC 603-V (Fig 3).
Regarding several osteoderms from Brejo da Madre de Deus, we noted abrasion affecting only the edges, causing an exposition of trabecular bone with a polished aspect (Fig 6A and  6B), but these marks were likely produced by taphonomic processes (i.e., post mortem). However, we observed pitting and irregular wearing of the external surface and bone response in osteoderms of this locality, configuring ante-mortem reactions ( Fig 6C).
In the two fragments from the latero-proximal region of the uncatalogued caudal tube of Panochthus sp. (Fig 7), only the main and lateral figures from the dorsal region and the main, peripheral and lateral figures from the ventral region were preserved. The primary alteration found in these fragments was the formation of cavities around the lateral figures and in some adjacent peripheral figures (Fig 7A-7C). The alterations were distributed on both lateral regions of the tube, not necessarily affecting the lateral figures. We observed severe damage in the distolateral left region, with the main figure showing a large and deep cavity, measuring approximately two centimeters wide, with microcavities inside that reached the trabecular bone ( Fig 7A). Additionally, we noted additional smaller cavities, with trabecular bone exposition and another isolated cavities in the same region, but in a more ventral portion ( Fig  7B). However, these cavities did not appear to be generated by horizontal erosion and, therefore were not considered pittings. The other fragment from the lateral region had a depression,  hair follicle pits placed in the intersection of the central and radial sulci. Field records report that these osteoderms were collected in association with endoskeleton elements, which included various paired elements showing anatomic and ontogenetic correlation (e.g., left and right tibia-fibulae) and might belong to a single individual [24]. Different types of arthritis were identified in several bones of this presumed individual [24].
Description of the lesions. Pittings of variable sizes were the primary alterations on the osteoderms of this taxon, bearing irregular and/or regular furrows in or outside them. Several osteoderms showed multifocal lesions. Notably, in osteoderms severely affected by pitting, a notable growth of the cavity occurred toward a point at which it became deeper, generating a slope, as observed in MCC 2221-V (Fig 8A). In some cases, the erosion on the surface of osteoderms did not necessarily create a cavity, whereas in others, the erosion produced a large damage that obliterated the surface completely. Some osteoderms showed signs of bone response in the form of calcium deposition in places with severe spongy bone exposition (MCC 2375-V; Fig 9A). In digital microscopical view, we observed the same twisted and contorted pattern of filaments overlapping the spongy bone, as observed in specimen MCC 1653-V (Panochthus; Fig 3).The pittings observed in the examined osteoderms were in initial, intermediary and advanced stages, as exemplified by MCC 668-V (Fig 8B), MCC 631-V ( Fig 8C) and MCC 2229-V (Fig 8D), respectively, resulting in partial or total loss of the ornamentation. The areas on the external surface of osteoderms with the ornamentation completely obliterated showed spongy bone expositions with varied extensions.
Several osteoderms showed well-delimited circular orifices that penetrate the cortical bone layer reaching the spongy bone. Moreover, the internal surfaces of these orifices had a polished aspect, as observed in MCC 2565-V and MCC 1198-V (Fig 9B and 9C, respectively), some coinciding with the enlargement of hair follicle pits (Fig 9B). We considered that these orifices were not pittings because they did not cause a uniform wearing of the surface, such as a cratering. Another peculiar characteristic was that they were not always associated with loss of the ornamentation in the osteoderms in which they occurred (MCC 1198-V). Digital microscopy images of the perforation in MCC 1198-V showed no bone growth in the perforation or on its edges, but with some furrows and other small cavities inside, and although clearly penetrating the spongy bone area, the perforation showed no exposition of trabeculae in its internal walls (Fig 9E), which had a polished aspect. As in some specimens of Panochthus (MCC 1653-V and the posterodorsal region of the carapace), the portion covered by the calcium deposition had a denser aspect than that of the adjacent surfaces (Fig 9D). Compared with the glyptodonts previously described, P. brasiliense had smaller hexagonal osteoderms, some proportionally thicker, with smooth surface, main figure displaced posteriorly and polygonal peripheral figures [13]. We analyzed 33 osteoderms from Fazenda Nova, of which nine presented alterations, and 22 from Lajedo Escada, of which only one showed alterations.

Pachyarmatherium brasiliense
Description of the lesions. The osteoderms from the two localities showed similar alterations to one another and to those of Glyptotherium sp. Most alterations corresponded to welldelimited circular cavities on the osteoderm surface, with some, as observed for Glyptotherium (see above), representing enlargements of the hair follicle pits. For DEGEO-UFPE 7266, DEGEO-UFPE 7410 and MCC 1515-V (Fig 10), measurements indicated that these enlargements were as large as those observed in Glyptotherium (Figs 8 and 9). As in Glyptotherium, loss of ornamentation did not necessarily occur in association with the circular perforations. The small erosions on the main figure were a second type of alteration observed, which were not as extensive as in Glyptotherium, showed punctual expositions of trabecular bone tissue by excavation, representing probable pittings (Fig 10A).

Evidences produced by flea infection
In all taxa, we observed marks consistent with a flea infection based on previous works [33,10,8]. Fleas are obligatory parasites of many groups of mammals and birds [34]. Among the studied materials, these marks occurred in isolated osteoderms of Glyptotherium and P. brasiliense, and caudal tube fragments of Panochthus. They consisted of perforations that penetrated into the bone structure of osteoderms.
In the lateral figures of caudal the tube of Panochthus, the flea infection marks were circular cavities, with minor cavities inside of them or in other points of the tube, similar to those figured for late Miocene armadillos (Chasicotatus ameghinoi and Vetelia perforata), which were attributed to fleas of the genus Tunga (See Fig 2E in[8]). This genus includes an extant species with proven ability to produce bioerosion in osteoderms of cingulates (Tunga perforans; see [35]). In isolated osteoderms of Glyptotherium and P. brasiliense, we also observed alterations comparable with those described for C. ameghinoi and V. perforata [8] and particularly for the extant armadillo Zaedyus pichiy, for which infections by fleas of the genus Tunga are also reported (See Fig 2 in [9]). In the osteoderms of Glyptotherium and P. brasiliense, these alterations consisted of well-delimited circular perforations, with most representing the enlargement of hair follicle pits.
A comparison among the perforations in exoskeleton elements of Panochthus sp., Glyptotherium sp. and P. brasiliense attributed to fleas and similar lesions described in previous work [8] revealed that in all cases, the perforations: (i) were consistently circular; (ii) had the diameter of the external opening broader than that of the internal, forming a tapered orifice; and (iii) were mostly isolated and, in few cases, associated (i.e., with more than one per osteoderm). In some osteoderms of living armadillos the perforations reach the internal surface, that is, the bioerosion crossed the osteoderm thick entirely [33]. By contrast, the perforations observed in the osteoderms analyzed in this study did not reach the internal surface of the specimens; instead, they ended in chambers inside the osteoderms, as in some cases also reported for living [33] and extinct armadillos [8]. This condition might be explained by differences in osteoderm thickness, because both Glyptotherium and P. brasiliense have osteoderms much thicker than those in the dasypodids analyzed in previous studies [33,10,8].
In this study, the flea perforations in the osteoderms of Glyptotherium and P. brasiliense ranged from one to five. This condition contrasts with the range registered for the extant armadillo Z. pichiy (one or two per osteoderm; [10]). At first glance, the size of the osteoderm offered a reasonable explanation for such distinct frequency per osteoderm: the Glyptotherium osteoderms have a greater superficial area and are thicker than those of Z. pichiy and therefore could offer a much larger area for the infesting fleas. The same explanation applies to the osteoderms of P. brasiliense, which although smaller than those of Glyptotherium, are considerably larger than those of armadillos for which similar alterations are described (C. ameghinoi, Z. pichiy).
The diameter of the cavity generated by fleas represents additional evidence to reinforce our diagnostic. One study registered diameters ranging from 1.97 to 2.92 mm [10]; similarly, perforations from 1.1 to 3.55 mm were found for C. ameghinoi and V. perforata [8]. Regarding the perforations in the osteoderms of P. brasiliense, we found diameters between 2.1 and 4.3 mm (DEGEO-UFPE 7266 specimen, Fig 10A), which are values partly consistent with those registered by the mentioned authors. For Glyptotherium, we found some perforations with diameters similar to those noted above, but we also found some perforations with much greater diameters. For example, in the specimens MCC 2565-V and MCC 1198-V (Fig 9B and  9C), for instance, we measured the greatest diameters among all infected osteoderms, ranging from 8 to 9 mm.
Following the ichnotaxobases classification proposed by [36], the process of bone removal creates a cavity in the formation of the chamber (see Fig 8B). The formation of the chamber by the flea is caused by neosomy in its reproductive cycle. The neosomes are organisms, in this case, fleas, that suffer a radical change during development acquiring a new morphological structure in the metamorphosis process [37]. Initially, the pregnant female penetrates into the host through bone perforation and initiates the neosomy [35]; the chemical or physical bioerosive mechanism by which the flea perforates the bone is unknown. After laying eggs, with the development of the neosome, the flea dies [38].
Traditionally, the identification of ante-mortem alterations requires the observation of a bone response [39]; however, no previous authors who registered bioerosion by fleas mentioned evidence of such process. Additionally, the figures in these works [33,10,8] do not show any obvious evidences of bone response. Similarly, we did not observe traces of bone response associated specifically with the lesions attributed to fleas. In fact, the living species T. perforans was described based on its ability of bone perforation [35], although associated bone response was not detected; at the end of the process of neosomy, the appearance of sequels, such as healing or infections, can occur or not [40].
The absence of bone response associated with the perforations on the external surface that were attributed to fleas might cast doubt on the differentiation between ante-and post-mortem alterations. Therefore, some additional comments on the alterations that we attributed to fleas are necessary. Several works cite post-mortem alterations in bone that, such as perforations generated by necrophagic coleopteran larvae in fossil vertebrate bone (See [41,35,42]), which at first glance, might be confused with ante-mortem marks, such as those attributed to fleas in this study. However, their internal morphology diverges, because necrophagic larvae construct chambers with a diameter twice the diameter of the opening. Moreover, the cavity has an ellipsoidal format (Cubiculum levis; see Fig 4 in [43]). By contrast, fleas produce cavities with the diameter equal or minor to the opening with a circular outline, as observed in the material in this study (Fig 9B). A possible reason for this difference is the finality of the excavation. The necrophagic larvae reach for the bone medullary channel to feed on the bone marrow, and consequently, the perforations reach deep portions of the bone and are interconnected (see Fig  2 in [42]). By contrast, fleas are less invasive because the bone is used as a reproduction site and for feeding on blood, and therefore, the cavity created is not required to reach deeper regions of the bone.
Another notable character that discriminates ante mortem bioerosion by fleas from post mortem bioerosion by necrophagic larvae is the relative number of cavities per osteoderm and their frequency along the carapace. Many cavities in a short area and with small spacing among them would suggest that many larvae actuated in the process. For fleas, few osteoderms show more than one cavity, which may result from competition avoidance behavior. Accordingly, few osteoderms among all analyzed material of Glyptotherium and P. brasiliense had perforations attributable to fleas, which is similar to the conditions previously reported for dasypodids. In fact, a notable detail regarding infection by fleas on dasypodids is the scanty and isolated nature of the perforations produced by these infestations in carapaces of a specimen of the living armadillo Z. pichiy (see Fig 5 in [9]). Similar lesions are observed in extant armadillos in a recent paper [11], but which does not discuss their distribution over the carapace. However, assessing the figures of the mentioned work (Figs 4G, 5B and 6F in [11]), we note that the lesions have a similar pattern of distribution. Assuming that a similar scanty distribution was also common in fossil cingulates infested with fleas, the low frequency of osteoderms showing this type of perforation in the analyzed material would be explained, with 10 of 55 osteoderms of P. brasiliense and only nine of 1436 of Glyptotherium. Additionally, note that all perforations attributed to fleas in this study occurred only on the external surface and were not observed in the underside or in the lateral regions of osteoderms, which is an unlikely condition had they been caused by necrophagic larvae or other post mortem process.
We mentioned previously that the Glyptotherium osteoderms in this study were found associated with endoskeleton bones, and these bones were examined with a perspective to identify ante mortem alterations [24]. After we reanalyzed these fossils, post mortem alterations were not found in the endoskeleton elements assigned to Glyptotherium [24] that could be attributed to insect larvae similar to those reported in the works mentioned above. This finding and the morphological differences indicate that such agents did not make the circular perforations on the osteoderms of this genus in this study. Conversely, none of these endoskeleton elements found associated with the osteoderms of Glyptotherium and Pachyarmatherium in the Lajedo da Escada site had perforations similar to those here attributed to fleas. Some of the elements of the exoskeleton of Glyptotherium, in addition to skeletal elements of other Pleistocene taxa from the same site, showed tooth traces likely produced by scavengers and predators [44], but these marks were scratch-like and much different from the perforations and other alterations described here.
Some marks produced by teeth of carnivores in bones bears resemblance to flea perforations, such as those of the ichnospecies Nihilichnus nihilicus [45]. For the cavities produced by the genus Tunga, the cortical bone is reached and the diameters are similar (2-10 mm). However, N. nihilicus is more frequent in long bones than in short bones and as far as we know, not yet reported for osteoderms. Morphologically, N. nihilicus differs from the flea perforations described in this study in having irregular edges and in some cases, associated microfractures (see Fig 6 in [45]). Lastly, the external morphology found in some marks of N. nihilicus is triangular or half-circular, which clearly differs from the well-delimited circular chambers left by the fleas.
Lesions in the carapace of a living armadillo similar to those discussed above were described in an early work [46] and attributed to mites. These marks consisted in the widening of the diameter of hair follicle pits with loss of ornamentation around some enlarged orifices (see Figs 29b, 33 and 36 in [46]). This same work reported similar alterations in a glyptodont (Glyptodon) and inferred that mites also produced these alterations. In fact, mites are among the arthropod parasites living on armadillos, but they do not perforate their carapace [8]. Therefore, we believe that these reported alterations are compatible, in terms of morphology and mode of occurrence, with those described in this study and therefore likely represent an additional case of flea infection in glyptodonts.
Considering the enlargement of the hair follicle pits, these pits were likely a preferred flea entrance place in the osteoderms of large extinct cingulates such as in living cingulates. This enlargement was most like caused by neosomy, although the flea entrance was not exclusively through the hair follicles, because not all cavities observed coincided with these structures; in some cases, the cavity produced by fleas was on the main figure (Fig 9B). In all cases, the perforations produced by fleas were three to fourfold wider than the normal hair follicle pits and were not confined to inside the main sulci, reaching portions of the peripheral and central figures, in contrast to the normal hair follicle pits (see Fig 10A and 10B).
The Tungidae family encompasses fleas that infect cingulates and is the only group that shows neosomy except for some members of the Neotunga genus (Pulicidae), which are restricted parasites of pangolins and phylogenetically distant of Tungidae [34]. Four of the 13 species in the genus Tunga infect living cingulates (see [38], Table 4), with T. penetrans the most common. Until now, flea infections were described only in small dasypodids (extinct and extant); thus, the fossils described here represent the first case of Tunga in glyptodontids and in a non-glyptodontid cingulate with large size (Pachyarmatherium).

Etiology of the pittings
In Panochthus, we noted the formation of pitting in the main figure of isolated osteoderms and on the carapace (MCC 1603-V), with the pitting all located in the osteoderms of the lateral regions (Fig 4), as the primary alteration in these bones. In Glyptotherium, we observed initial (Fig 8B), intermediary ( Fig 8C) and advanced ( Fig 8D) stages of pitting, with spongy bone exposition in all cases. The pittings and the irregular obliteration of the ornamentation associated with the former were not attributable to fleas, although fleas could have fundamental participation in their origin acting as vectors of opportunistic micropathogens that cause dermal lesions, such as Staphylococcus aureus and Clostridium tetani [38]. Additionally, fleas transmit the fungus Paracoccidioides brasiliensis, responsible for some dermatitis and the bacteria Clostridium perfrigens [47]. The erosions of the ornamentation observed in some of the specimens analyzed (MCC 668-V, MCC 631-V and MCC 2229-V) were similar to those attributed to bacterial attack, particularly by Mycobactherium leprae, although leprosy does not cause significant alterations in cingulates [48]. A case of sporotrichosis in living armadillos was reported that destroyed a wide extension of epidermis by the opportunist fungus Sphorothrix scheenckii, which is very common in osteological infections, generating superficial erosion [49]. However, because of the lack of diagnosable features for such infections in osteoderms analyzed in this study, we could not conclude what type of pathogen was responsible for the pitting observed in the fossils.
The cave environment of Lajedo da Escada in which the osteoderms of Glyptotherium sp. and some specimens assigned to P. brasiliense were found requires considering the possibility that some alterations observed on these osteoderms were caused by post mortem conditions in such environments. Note that inside limestones basements, the formation of the caves, carbonic acid (H 2 CO 3 ) mediates the formation of the caves by slowly corroding and perforating the limestones rocks [50]. In fact, we noted that the osteoderms of Glyptotherium sp. showed traces of corrosion by acid, such as powder release, as in a chalk.
In this context, we could not discard that carbonic acid contributed to the loss of ornamentation associated with pitting in Glyptotherium sp. osteoderms, although the effect of acid was not primarily responsible. As we noted earlier, some osteoderms collected in Lajedo da Escada showed bone response, which is clear evidence of an ante mortem process (Fig 9A). In these cases, the acid likely magnified the marks left by the pathological process, that is, the injury appeared first, in life, and then, after death, the acid enlarged the diameter of the cavities by corrosion. The fact, mentioned above, that only their external surfaces were affected is important evidence showing that post mortem corrosion by acid was not the primary explanation for the alterations observed in the osteoderms of Glyptotherium sp. from Lajedo da Escada. Although acids can mimic bone infections and create pitted surfaces [51], in such cases, we would expect that all surfaces of the affected osteoderms would show evidence of corrosion, which was not observed on the internal surfaces in the studied specimens.
We must emphasize that the irregular obliterations of the ornamentation along the carapace and some fragments of Panochthus sp. could not be considered pittings, although they were associated with them. These alterations had a spread extension; whereas pitting is a punctuated erosion of bone surface. Another difference was the formation of crests associated with a wide erosion of the ornamentation and holes for pitting.

Paleoecological and evolutionary implications
We recorded new potential cases of disharmonic interspecific interaction (parasitism) between cingulates and fleas for taxa previously not reported for this type of alteration. Fleas of the genus Tunga might have been biogeographically widely distributed and a common parasite of most or all cingulate main linages during the evolution of this group, at least since the late Miocene [8]. All examined taxa (Panochthus sp., Glyptotherium sp. and P. brasiliense) showed perforations attributable to neosomy (flea egg-laying), similarly to the cases involving other extinct (C. ameghinoi and V. perforata) and extant (Chaetophractus villosus) species [8,10]. Moreover, this work expands the distribution area of species of the genus Tunga able to perforate bones, which was restricted to Argentina [10,11,8], reporting their occurrence also in environments of the Brazilian Intertropical Region. We can envisage the success of Tungidae fleas when we correlate flea temporal distribution along the Cenozoic [52] and its singular ability to infect the hosts. The biochron ranges from late Miocene [8] to Holocene [10,11], including the new cases from the Pleistocene described here. This ability to perforate bones could have conferred advantages over other fleas that might have been limited to the dermic surface. This adaptation might have avoided competition for ecological niches, increasing availability of habitat and resources for reproduction.
As stated previously, among the known species of the genus Tunga, only T. perforans shows the bone perforation ability. However, we cannot discard the existence of other species within Tungidae with the same capacity because of the considerable time scale involved in the known interaction between this family and cingulates, whose most ancient record dates from the late Miocene [8]. If we consider T. perforans as the only species infecting cingulates since that period, we must assume a wide evolutionary stasis for a species with a fast life cycle and expansive geographic distribution, which is an unlikely scenario. In short, we consider the hypothesis that the bone perforation ability of Tungidae fleas persisted during this entire time interval but was not necessarily restricted to T. perforans.