Ontogeny of the skull of the blind snake Amerotyphlops brongersmianus (Serpentes: Typhlopidae) brings new insights on snake cranial evolution

Abstract

Blind snakes represent the most basal group of extant snakes and include fossorial species with unusual skeletal traits. Despite their known phylogenetic position, little is known about their ontogeny and what it might reveal about the origin of their skull anatomy. Here we describe for the first time the ontogenetic transformations of the skull of a blind snake, the typhlopid Amerotyphlops brongersmianus, including embryos and postnatal individuals. Furthermore, we provide data on the size changes relative to skull growth of the main elements of the gnathic complex. We observed that the skull of this blind snake undergoes considerable morphological change during late ontogeny. Additionally, we detected delayed development of some traits (closure of the skull roof, opisthotic-exoccipital suture, ossification of the posterior trabeculae) simultaneously with clearly peramorphic traits (development of the crista circumfenestralis, growth of the pterygoid bar). Our analysis suggests that the unique skull anatomy of blind snakes displays plesiomorphic and highly autapomorphic features, as an outcome of heterochronic processes and miniaturization, and is shaped by functional constraints related to a highly specialized feeding mechanism under the selective pressures of a fossorial lifestyle.

development, fossorial snakes, growth, heterochrony, micro-CT, miniaturization, skull morphology

INTRODUCTION

Blind snakes (Anomalepididae, Typhlopoidea and Leptotyphlopidae) represent the first extant branches in the evolutionary radiation of modern snakes (Miralles et al., 2018; Burbrink et al., 2020). They are fossorial forms with tubular bodies, smooth scales and a substantial reduction of eyes and head scalation (Vitt & Caldwell, 2009). They feed on social insects, such as ants, termites and their larvae (Cundall & Greene, 2000), using a singular intraoral prey transport mechanism called mandibular or maxillary raking (Kley, 2001). All elements of the gnathic complex (i.e. palatomaxillary bars, suspensorium and lower jaw) are involved during feeding, and toothed elements of the jaws (maxilla or dentary) are used to ratchet small prey into and through the mouth (Cundall & Greene, 2000; Kley, 2001). Thus, the overall skull morphology of blind snakes is conditioned by their unique feeding system and the mechanical demands imposed by excavation (Cundall & Irish, 2008), deviating significantly from the general trends of squamates (Evans, 1955; Kley, 2006; Palci et al., 2016; Da Silva et al., 2018; Chretien et al., 2019; Strong et al., 2021).

The skull anatomy along with other particular traits constitute a set of characters called ‘scolecoidy’ (sensuMiralles et al., 2018), an ecomorphotype shared by the three major clades of blind snakes. These clades represent successful groups of fossorial snakes, with more than 450 species currently recognized, and several new species (e.g. Koch et al., 2015, 2016, 2019; Shea, 2015; Kraus, 2017; Dehling et al., 2018) and genera (Martins et al., 2019) described each year. They are distributed on all continents except Antarctica, with a long evolutionary history that probably pre-dates the Upper Cretaceous (Vidal et al., 2010; Miralles et al., 2018).

Ontogenetic studies are relevant to understand the processes responsible for morphological diversity and result in a powerful source of information about evolutionary patterns in the skeleton of squamates (e.g. Rieppel, 1992a, 1994; Maisano, 2001; Bhullar, 2012; Roscito & Rodriguez, 2012; Werneburg et al., 2015; Da Silva et al., 2018; Ollonen et al., 2018; Khannoon et al., 2020, among others). Regarding snakes, numerous studies have pointed to heterochronic processes as responsible to a large extent for morphological evolution of the group (Rieppel, 1988; Irish, 1989; Werneburg & Sánchez-Villagra, 2014), but they were solely based on adult specimens and no ontogenetic evidence was included to support their hypotheses. Additionally, the study of ontogenetic sequences may bring data on functionally-relevant transient structures (De Beer, 1940) that may be helpful to elucidate evolutionary aspects of different snake clades. Moreover, in spite of descriptions of skeletal ontogeny in snakes having increased and gaining relevance in recent years (Boback et al., 2012; Polachowski & Werneburg, 2013; Khannoon & Evans, 2015; Palci et al., 2016; Scanferla, 2016; Sheverdyukova, 2017; Al Mohammadi et al., 2020), studies have been focused on the most widely known group of extant snakes (i.e. Alethinophidia).

To date, the cranial ontogeny of blind snakes remains largely unknown due to their secretive habits, which make them rarely encountered in the field, and scarce in museum collections (Koch et al., 2019) with the exception of a few locally abundant species (McDiarmid et al., 1999). Consequently, the study of ontogenetic series of these elusive snakes represents a challenge, and only a few observations regarding the postnatal ontogeny of the skull of certain species are available (Cundall & Irish, 2008; Palci et al., 2016; Scanferla, 2016). However, there is a renewed interest in blind snake morphology, fuelled by new phylogenetic perspectives and the introduction of micro-CT technology to uncover the anatomy of these small-sized, often minute snakes (Bell et al., 2021).

Amerotyphlops brongersmianus (Vanzolini, 1976) is a relatively large (snout-vent length = 300 mm on average) blind snake belonging to the family Typhlopidae. This oviparous species has recently been the focus of different studies, among which stands the detailed description of its adult osteology (Lira & Martins, 2021). Furthermore, it represents the first blind snake for which embryonic staging based on external morphology has been described (Sandoval et al., 2020). In this line, the present study provides the first detailed description of the skull changes during embryonic and postnatal ontogeny of the blind snake A. brongersmianus through 3D reconstructions based on micro-CT data. This is presented within a comparative approach incorporating skulls of embryos of alethinophidian species and lizards. In light of the relevance of developmental processes for understanding the evolution of blind snakes, this research represents a reliable source to test previous hypotheses of heterochronic processes occurring in this clade (e.g. Rieppel, 1988; Irish, 1989; Werneburg & Sánchez-Villagra, 2014; Da Silva et al., 2018; Strong et al., 2020).

MATERIAL AND METHODS

Specimens

The ontogeny of the skull of A. brongersmianus was studied through embryos and postnatal specimens housed in the Herpetological Collection of Universidad Nacional del Nordeste (UNNEC), Corrientes, Argentina (Table 1). Three post-ovipositional embryos were selected based on the staging criteria of Sandoval et al. (2020), and four postnatal specimens were selected based on their snout-vent length (SVL). The smallest individual of the postnatal ontogenetic sequence was considered a hatchling due its SVL being lower than that of the hatchling reported by Sandoval et al. (2020). We recognized two stages between hatching and adult stage, designated as juvenile and subadult only for descriptive purposes since data on gonadal development of these specimens was not available.

Table 1.

Information about the specimens analysed. Abbreviations: UNNEC, Herpetological Collection of the Universidad Nacional del Nordeste, Corrientes (Argentina); ZFMK, Herpetological Collection of the Alexander Koenig Museum, Bonn (Germany)

Species	Specimen	SVL (mm)	Ontogenetic stage	Stage (author)
Amerotyphlops brongersmianus	UNNEC 10326	-	Posovipositional embryo	33 (Sandoval et al., 2020)
Amerotyphlops brongersmianus	UNNEC 10356	-	Posovipositional embryo	34 (Sandoval et al., 2020)
Amerotyphlops brongersmianus	UNNEC 11433	-	Posovipositional embryo	36 (Sandoval et al., 2020)
Amerotyphlops brongersmianus	UNNEC 12783	84	Hatchling	-
Amerotyphlops brongersmianus	UNNEC 12831	120.5	Juvenile	-
Amerotyphlops brongersmianus	UNNEC 12796	190.6	Subadult	-
Amerotyphlops brongersmianus	UNNEC 12792	277.7	Adult	-
Lanthanotus borneensis	ZFMK 97200	58.7	Posovipositional embryo	13 (Werneburg et al., 2015)
Liasis mackloti	ZFMK 41572	397	Posovipositional embryo	10 (Boughner et al., 2007)
Candoia bibroni	ZFMK 41185	299	Posovipositional embryo	10 (Boughner et al., 2007)
Naja oxiana	ZFMK 55961	216	Posovipositional embryo	8 (Khannoon & Evans, 2015) 10 (Jackson, 2002)

Embryos of the anguimorph lizard Lanthanotus borneensis Steindachner, 1878 and the alethinophidian snake species Liasis mackloti Duméril & Bibron, 1844, Candoia bibroni (Duméril & Bibron, 1844) and Naja oxiana (Eichwald, 1831) were studied to make comparisons with structures of interest. Information about all specimens analysed is given in Table 1. Skulls of adult specimens of squamates available in the online repository MorphoSource (www.morphosource.org) were also examined for comparative purposes.

Micro-CT

The head of each specimen was scanned on a Bruker SkyScan 1173 Micro-CT scanner at the Zoological Research Museum Koenig (ZFMK), Bonn, Germany. The scan parameters for each specimen are detailed in the Supporting Information (Table S1). Micro-CT data sets were reconstructed using N-Recon software (Bruker Micro-CT) and rendered in three dimensions through the aid of CTVox for Windows 64 v.2.6 (Bruker Micro-CT). We followed the terminology of Haluska & Alberch (1983), Cundall & Irish (2008) and Rieppel et al. (2009) for skull embryonic ontogeny and osteology.

Morphometric analysis

In order to compare growth trajectories of the skeletal elements that constitute the gnathic complex (palatomaxillary bar, suspensorium and lower jaw), we measured selected bones and bony structures on 3D reconstructions of A. brongersmianus postnatal specimens (Table 2). The same measurements were taken on dry and cleared and stained skulls of postnatal ontogenetic sequences of representatives of the lizard outgroup Anguimorpha (Ophiodes intermedius Boulenger, 1894) and of the alethinophidian snake group (Philodryas psammophidea Günther, 1872). This alethinophidian snake species represents herein the macrostomous condition (i.e. ingestion of large prey with a large cross-sectional area in relation to the head dimensions of the snake; Cundall & Greene, 2000; Scanferla, 2016). Measurements were taken with a dial calliper under a binocular microscope to the nearest 0.1 mm and are provided in Table 2. The skull length was measured from the anterior tip of the snout to the posterior end of the occipital condyle in dorsal view, while the lower jaws were measured from the anterior tip of the dentary to the posterior end of the retroarticular process in lateral view. Measurements of selected structures of the gnathic complex (pterygoid, quadrate, dentary and lower jaw) were simply quantified as a ratio of their length against skull length. Then, growth trajectories were evaluated by plotting this ratio against successive ontogenetic stages. Data analyses and visualization were conducted using R v.4.0.2 (R CoreTeam, 2020).

Table 2.

Measurements of the bony elements and structures of the gnathic complex throughout postnatal ontogeny of the anguid lizard Ophiodes intermedius, the blind snake Amerotyphlops brongersmianus and the alethinophidian snake Philodryas psammophidea. All measurements are given in mm. Abbreviations: DL, dentary length; LjL, lower jaw length; PtL, pterygoid length; QL, quadrate length; SkL, skull length; SVL, snout-vent length; MCN, Museo de Ciencias Naturales de Salta, Salta (Argentina); UNNEC, Herpetological Collection of the Universidad Nacional del Nordeste, Corrientes (Argentina)

Species	Specimen	SkL	PtL	QL	DL	LjL	SVL
Ophiodes intermedius	MCN 4443	6.7	2.8	1.3	3.2	5.1	32.0
Ophiodes intermedius	MCN 4444	8.1	4.3	1.6	4.2	7.6	56.0
Ophiodes intermedius	MCN 4445	11.8	5.9	2.0	6.1	10.9	114.0
Ophiodes intermedius	MCN 4446	18.9	9.1	3.2	9.9	17.8	237.0
Amerotyphlops brongersmianus	UNNEC 12783	5.5	3.1	1.9	0.5	2.9	84.0
Amerotyphlops brongersmianus	UNNEC 12831	6.5	3.7	2.1	0.6	3.9	120.5
Amerotyphlops brongersmianus	UNNEC 12796	7.3	4.3	2.1	0.8	4.2	190.6
Amerotyphlops brongersmianus	UNNEC 12792	9.5	6.3	2.4	0.8	5.6	277.7
Philodryas psammophidea	MCN 190	11.4	7.3	2.3	6.2	11.6	213.0
Philodryas psammophidea	MCN 133	16.4	10.5	4.3	9.2	16.7	473.0
Philodryas psammophidea	MCN 187	18.5	12.9	5.7	10.4	20.7	631.0
Philodryas psammophidea	MCN 4447	22.4	18.5	7.8	13.3	27.9	1020.0

RESULTS

Snout complex

The snout region in A. brongersmianus is formed by nasals, a premaxilla, septomaxillae and vomers (Figs 1–3). In the earliest embryo available (Stage 33) all these elements are fully differentiated and ossified. During embryonic ontogeny, this region is slightly narrower than the braincase and goes through a marked shift upwards (Fig. 1A-C). The foramina typical of the snout of blind snakes are already present at Stage 33, distributed ventrally in the premaxilla and in the dorsal lamina of the nasals (Figs 2A, 3A). The premaxilla also exhibits a prominent anteromedial carina and three pointed processes: a single posteriorly oriented vomerine process and paired laterally oriented septomaxillary processes (Fig. 3). The ventral closure of the snout is completed by the septomaxillae and vomers, which jointly outline the fenestra vomeronasalis (Fig. 3). The borders of the fenestra are remodelled through ontogeny due to the growing ossification of the posterior border of the septomaxilla (Fig. 3A-F). As in all blind snakes, the prefrontals are incorporated into the snout region in A. brongersmianus and laterally limit the external narial opening (Fig. 1).

Figure 1.

Lateral view of the skull of Amerotyphlops brongersmianus throughout embryonic and postnatal ontogeny: A, embryo at Stage 33; B, embryo at Stage 34; C, embryo at Stage 36; D, hatchling; E, juvenile; F, adult. Abbreviations: a, angular; bo, basioccipital; cb, compound bone; co, coronoid; d, dentary; et, egg tooth; fr, frontal; mx, maxilla; n, nasal; ot, otooccipital; p, parietal; pa, palatine; pbs, parabasisphenoid; pmx, premaxilla; po, prootic; pp, postorbital process; prf, prefrontal; pt, pterygoid; q, quadrate; smx, septomaxilla; so, supraoccipital; sp, splenial; II, foramen for the optic nerve; V, foramen for both rami of the trigeminal nerve. Scale bars equal to 1 mm.

Figure 2.

Dorsal view of the skull of Amerotyphlops brongersmianus throughout embryonic and postnatal ontogeny: A, embryo at Stage 33; B, embryo at Stage 34; C, embryo at Stage 36; D, hatchling; E, juvenile; F, adult. Abbreviations: bo, basioccipital; fr, frontal; mx, maxilla; n, nasal; ot, otooccipital; p, parietal; pa, palatine; pbs; parabasisphenoid; pmx; premaxilla; po, prootic; pp, postorbital process; prf, prefrontal; q, quadrate; so, supraoccipital. Scale bars equal to 1 mm.

Figure 3.

Ventral view of the skull of Amerotyphlops brongersmianus throughout embryonic and postnatal ontogeny. Lower jaw as well as the maxilla, palatine and pterygoid on the left side were digitally removed: A, embryo at Stage 33; B, embryo at Stage 34; C, embryo at Stage 36; D, hatchling; E, juvenile; F, adult. Abbreviations: bo, basioccipital; cs, crista sellaris; et, egg tooth; fr, frontal; fv, fenestra vomeronasalis; hf, hypophysial fenestra; mx, maxilla; ot, otooccipital; p, parietal; pa, palatine; pbs, parabasisphenoid; pmx, premaxilla; po, prootic; pov, posterior opening of the Vidian canal; pfr, prefrontal; pt, pterygoid; q, quadrate; s, sulcus of the embryonic parabasisphenoid; sav, secondary anterior opening of the Vidian canal; set, scar of the egg tooth; smx, septomaxilla; ut, unossified trabecula; II, foramen for the optic nerve; V, foramen for maxillary and mandibular rami of the trigeminal nerve. Scale bars equal to 1 mm.

A fully differentiated egg tooth was observed in embryos of A. brongersmianus (Fig. 1A-C), which shows a contrasting morphology with respect to the egg tooth of other squamates examined (Fig. 4). It is attached at the base of the posterior process of the premaxilla, markedly displaced posteroventrally (Fig. 4B). The length of the egg tooth is almost half the height of the snout region in lateral view, it has a sigmoidal shape and projects downwards (Fig. 4B). In contrast, the egg teeth of the anguimorph lizard La. borneensis and alethinophidian snakes Li. mackloti and Naja oxiana are attached to the anterior edge of the premaxilla (Fig. 4A, C, D). The egg tooth of these species is short and markedly curved forward, thus it protrudes from the tip of the snout (Fig. 4A, C, D). In terms of morphology, the egg tooth of A. brongersmianus is conical with pointed tip, and has a rounded central cavity that narrows proximodistally (Fig. 4B). In this regard, the egg teeth of the lizard La. borneensis and both alethinophidian snake embryos are compressed anteroposteriorly and their width slightly decreases distally. In La. borneensis, the egg tooth’s distal third narrows, ending in a rounded tip and its central cavity is circular in cross section (Fig. 4A), whereas the egg teeth of Li. mackloti and N. oxiana have a truncated end and their central cavity is a horizontal ellipse (Fig. 4C, D).

Figure 4.

Lateral (left), anterior (centre) and ventral (right) views of the egg tooth of squamate embryos at equivalent pre-hatching stages: A, Lanthanotus borneensis; B, Amerotyphlops brongersmianus; C, Liasis mackloti; D, Naja oxiana. The premaxilla is highlighted in light grey. The ventral view is a horizontal cutaway showing the section of the egg tooth near its base. Scale bars equal to 1 mm.

Braincase roof

Paired frontals, a single parietal and paired supraoccipitals constitute the braincase roof of A. brongersmianus. Prootics, otooccipitals and stapes form the posterolateral (otic) region of the braincase, and the parabasisphenoid and basioccipital form the braincase floor (i.e. basicranium). At Stage 33, the supraoccipital, prootic, otooccipital and basioccipital are well ossified, but all these elements remain separated from one another by narrow unossified zones (Figs 1A, 2A, 3A).

The lateral descending flange of the frontal bone is the first portion to ossify during embryonic development, then the ossification progresses dorsally towards the dorsal lamina. In the late embryo (Stage 36), the dorsal lamina is well developed anteriorly but its posteromedial region is still unossified (Fig. 2C). The dorsal lamina of the frontal grows forwards and backwards during postnatal ontogeny. Therefore, an expansion of its anterior border between the nasals is observed and the frontoparietal suture is shifted posteriorly (Fig. 2D-F). Furthermore, there is a lateromedial compression of the frontals from juvenile to adult stages and a posterior process abutting against the postorbital process of the parietal develops at the adult stage (Fig. 2F).

The parietal is the only component of the braincase starting its ossification as a paired element and subsequently merging as an azygous bone. The ossification of the parietal begins in the descensus parietalis, and as the development progresses, it spreads dorsally towards the midline, constituting paired dorsal laminae which ultimately fuse. A wide gap still separates both dorsal laminae in the late embryo (Stage 36; Fig. 2C), but at the hatching stage only a small suture remains posteriorly in the midline (Fig. 2D). Furthermore, the ossification of the dorsal laminae of the parietals is notably delayed when compared to the ossification of the frontals at all embryonic stages (Fig. 2A-C). As with frontals, there is a remarkable dorsal remodelling of the parietal during postnatal ontogeny. It changes from rounded at the hatchling stage to a compressed configuration in adulthood (Fig. 2D-F). The anterolateral crest of the dorsal laminae projecting laterally in the contact zone with the frontals, and the posterolateral corner partially covering the suture between the prootic and supraoccipital (named the supratemporal process) are constituted late during postnatal ontogeny (Fig. 2E-F). The descensus parietalis approaches the parabasisphenoid along its lateral edge, although there is a persistent gap along almost the entire ontogenetic sequence (Fig. 3A-E).

There is a mid-dorsal gap between both supraoccipitals during embryonic ontogeny but they approach each other at the midline at the hatching stage (Fig. 2A-D). The supraoccipital also undergoes shape changes from hatching stages to adulthood. Its anterior margin changes from rounded to straight in the dorsal view, while its lateral margin narrows. Both changes are tightly linked to shape change of the parietal described above, mostly related to growth of the supratemporal process.

Otic region

The bony elements belonging to the otic region are differentiated and in an advanced stage of ossification at Stage 33 (Figs 1–3). The prootic occupies the posterolateral surface of the skull, and it expands in an anteroposterior direction during postnatal ontogeny (Fig. 1D-F). The sole trigeminal foramen is located in the contact zone of the prootic, the descensus parietalis and the parabasisphenoid (Fig. 3). The contribution of the parabasisphenoid to the margins of this foramen may vary, since the prootic seems to exclude the parabasisphenoid at some stages during postnatal ontogeny (Fig. 3D, E).

The two elements that constitute the otooccipital (i.e. opisthotic and exoccipital) are distinguishable in the embryo at Stage 33. The dorsomedial and ventral parts of the otooccipital correspond to the exoccipital contribution, while the anterolateral portion bears on the opisthotic contribution (Fig. 5A). The suture between the exoccipital and opisthotic is still present in the late embryo (Stage 36) and remains at the hatching stage as a small gap restricted to the dorsal region of the bone (Fig. 5B). In the examined embryo of the anguimorph lizard, La. borneensis, both elements are unfused, whereas in the embryos of the alethinophidian snakes (Candoia Gray, 1842, Liasis Gray, 1842 and Naja Laurenti, 1768), the fusion between elements is advanced and only a small suture in the dorsal region of the otoccipital is observed.

Figure 5.

A-C, posterolateral views of the otic region of Amerotyphlops brongersmianus throughout ontogeny: A, embryo at Stage 33; B, hatchling; C, adult. Black arrowheads indicate the suture between the exoccipital and opisthotic. D, endocranial view of the otic region in an embryo of A. brongersmianus at Stage 34 showing the crista interfenestralis concealing the crista tuberalis. E-F, endocranial and lateral views of the otic region of the embryo of Candoia bibroni showing the type 3 configuration of the crista circumfenestralis. G, lateral view of the otic region of the embryo of Naja oxiana showing the type 4 configuration of the crista circumfenestralis. Abbreviations: bo, basioccipital; ci, crista interfenestralis; cop, commissura praefacialis; cp, crista prootica; ct, crista tuberalis; ct+ci, fused crista tuberalis and crista interfenestralis; dp, descensus parietalis; jf, jugular foramen; ot, otooccipital; p, parietal; pbs, parabasisphenoid; po, prootic; pov, posterior opening of the Vidian canal; so, supraoccipital; st, stapes; V, foramen for the maxillary and mandibular rami of the trigeminal nerve; V₂, foramen for the maxillary ramus of the trigeminal nerve; V₃, foramen for the mandibular ramus of the trigeminal nerve; VII, foramen for the facial nerve. Scale bars equal to 1 mm.

As in most blind snakes, in the adult stage of A. brongersmianus the stapedial footplate is almost concealed by the crista circumfenestralis and only the short stapedial shaft emerges from the juxtastapedial recess (Fig. 5C). Additionally, the crista tuberalis is expanded and has incorporated the crista interfenestralis. This morphology corresponds with the type 4 configuration of the crista circumfenestralis described by Palci & Caldwell (2014) for adult blind snakes and most colubroids. The available ontogenetic sequence shows a progressive growth of the crista prootica and crista tuberalis until reaching adult configuration (Fig. 5A-C). Embryos of A. brongersmianus show an extreme anterodorsal development of the crista tuberalis, which in turn, laterally conceals the small crista interfenestralis. Then, the crista interfenestralis is not visible as a discrete element laterally during embryonic development (Fig. 5A). Instead, the contribution of both cristae (crista tuberalis and crista interfenestralis) can be seen medially in the otic region of A. brongersmianus embryos (Fig. 5D). The alethinophidian C. bibroni shows a contrasting morphology of the crista circumfenestralis (Fig. 5E, F), bearing a type 3 configuration (sensuPalci & Caldwell, 2014). The embryo of the colubroid Naja oxiana also bears the type 4 configuration of the crista circumfenestralis (Fig. 5G), although unlike A. brongersmianus, the crista interfenestralis seems to reach the lateral wall of the skull, and then fuse with the crista tuberalis. Hence, despite the type 4 configuration being present in adults of both species, their development follows different pathways.

Basicranium

The compound parabasisphenoid is in advanced stage of ossification in the examined Stage 33 of A. brongersmianus (Fig. 3A), thus it is not possible to discern the extent of the contribution of each former element—the membranous parasphenoid and the chondral basisphenoid—to adult bone. The subcircular hypophysial fenestra, located in the centre of the parabasisphenoid, is well developed in early embryos (Stages 33–34; Fig. 3A, B), but closed in the late embryo (Stage 36) and the hatching stage (Fig. 3C, D). Likewise, a small remnant of the basicranial fenestra is present between the posterior border of the parabasisphenoid and the anterior border of the basioccipital at Stage 36, but it is fully closed at the hatching stage (Fig. 3C, D). The posterior tip of the basioccipital, jointly with ventromedial projections of the otoccipital, form the occipital condyle in the adult stage (Figs 3F, 5C).

The cartilaginous regions corresponding to the trabeculae cranii and the crista sellaris (derived from the acrochordal cartilages) are seen as zones in light grey or even as empty spaces in Micro-CT images of embryos and hatchlings of A. brongersmianus (Fig. 3A-D), thus outlining part of the chondrocranium. These cartilages establish a triangular structure that acts as a scaffold from where most of the parabasisphenoid ossifies (Fig. 6A, B). In A. brongersmianus, the trabeculae diverge immediately posterior to the trabecula communis, in contrast to alethinophidians where the trabeculae run almost parallel most of their length (Fig. 6A-C). Furthermore, in adult alethinophidians there is a ridge named the crista trabecularis in the transition zone between cartilaginous and ossified trabecula in the parabasisphenoid, lateral to the base of the parasphenoidal rostrum. This structure is not observed in the adult of A. brongersmianus (Fig. 6B). Interestingly, the posterior portion of the trabeculae remain cartilaginous (at least their centre) until the juvenile stage in A. brongersmianus (Fig. 6D), while in alethinophidians they are ossified at embryonic stages.

Figure 6.

A-B, internal (endocranial) view of the parabasisphenoid bone in Amerotyphlops brongersmianus throughout ontogeny showing the conformation of the Vidian canal: A, embryo at Stage 33; B, adult. C, endocranial view of the parabasisphenoid bone of the embryo of Liasis mackloti. D, frontal cutaway posterior to the hypophysial fenestra of embryo of A. brongersmianus at Stage 33 showing the cartilaginous nuclei of the trabeculae cranii. E-F, internal views of a sagittal cutaway of the skull of A. brongersmianus throughout ontogeny showing different endocranial apertures of the Vidian canal and presence of the lateral wing of the parabasisphenoid: E, embryo at Stage 33; F, adult. G, lateroventral view of the adult skull of A. brongersmianus showing external apertures of the Vidian canal. H-I, internal views of a sagittal cutaway of the skull of Candoia bibroni throughout ontogeny showing different endocranial apertures of the Vidian canal and the lateral wing of the parabasisphenoid: H, embryo; I, adult. J, lateroventral view of the adult skull of C. bibroni showing external apertures of the Vidian canal. Abbreviations: bo, basioccipital; cbc, cerebral carotid artery; ccf, cerebral carotid foramen; fr, frontal; lo, lateral ossification of the posterior trabecula; ot, otooccipital; p, parietal; pan, palatine nerve; par, palatine artery; pav, primary anterior opening of the Vidian canal; pbs; parabasisphenoid; po; prootic; pov, posterior opening of the Vidian canal; sav, secondary anterior opening of the Vidian canal; t, trabecula cranii; tc, trabecula communis; V, foramen for the maxillary and mandibular rami of the trigeminal nerve; V₄, cid nerve; VII_p, palatine ramus of the facial nerve. Scale bars equal to 1 mm.

There is an important bone outgrowth in the posterolateral region of the embryonic parabasisphenoid, probably originated as perichondral ossification from the trabecula, the crista sellaris, the basal plate or a combination of them (Fig. 6A). In this region the Vidian canal is formed, a complex structure for the passage of nerves and vessels. In ventral view, embryos show a sulcus and a foramen (i.e. posterior Vidian opening) in the posterolateral region of the parabasisphenoid (Fig. 3A, C). This sulcus is gradually floored during late embryonic and early postnatal stages, thus defining a duct named the closed Vidian canal (Figs 3C-D, 6A, B). The medial opening of this duct forms the dorsolateral margin of the cerebral carotid foramen as in alethinophidians (Fig. 6B, C, E-F, H-I). The lateral opening of the duct is the primary anterior opening of the Vidian canal and leads to an open groove—the Vidian groove—on the dorsal surface of the parabasisphenoid (Fig. 6B, F). The Vidian groove runs forwards and ends in an opening near the lateral rim of the parabasisphenoid, named the secondary anterior opening of the Vidian canal (Fig. 6B, F). This opening can be fully incorporated to the parabasisphenoid or be formed by this element and the parietal, showing asymmetrical variation in the same individual (Fig. 6F-G). Late embryos of Candoia (Fig. 6H) and Liasis show a similar intracranial morphology of the Vidian canal, bearing a short passage through the parabasisphenoid (i.e. closed Vidian canal) and a primary anterior opening at the level of the cerebral carotid foramen (Fig. 6H).

Palatomaxillary bar

The palatomaxillary bar is formed by the maxilla, the palatine and the pterygoid rod (Figs 1, 3). The wide posterior half of the maxilla bears four tooth positions from embryonic to subadult stages, and the final five tooth count is reached in adult specimens. The pterygoid is a rod-like element, anteriorly bifurcated, that runs along the ventrolateral border of the skull (Fig. 3F). During embryonic development, the pterygoid changes from lateroventrally curved to straight, and it never surpasses the posterior limit of the prootic in lateral view (Figs 1A-C, 3A-C). In contrast, during postnatal ontogeny it doubles its length until reaching adult proportions, its posterior tip being projected beyond the occipital condyle (Fig. 1D-F).

Suspensorium and lower jaw

The quadrates suspend the lower jaw from the skull and have three main processes: an anterodorsal cephalic process, an anteroventral mandibular process and a posterior suprastapedial process (Fig. 1F). During embryonic development, the anterior region of the bone surpasses the anterior border of the prootic, reaching the posterior half of the parietal (Fig. 1A-C). However, in the adult stage the quadrate does not surpass the prootic anteriorly (Fig. 1F; see below for further explanation). In addition, the quadrate bone does not experience rotation along embryonic development, but it progressively opens laterally during postnatal ontogeny (Fig. 3E, F) as was also described by Palci et al. (2016) for the typhlopid Anilios bicolor (Peters, 1858).

The lower jaw in adult specimens of A. brongersmianus is formed by the dentary, the splenial, the angular, the coronoid and the compound bone (Fig. 7C). The embryo at Stage 33 exhibits these elements as well differentiated and in an advanced stage of ossification (Fig. 7A). There are no traces of tooth sockets nor teeth along the lower jaw of the embryo at Stage 33 (Fig. 7A). The Meckelian canal is open medially between the dentary and splenial, and in the anterior region of the compound bone (Fig. 7A). The ossification of these bones increases during embryonic ontogeny closing the canal medially (Fig. 7B). Additionally, in the embryo at Stage 33, the anterior portion of the compound bone exhibits a longitudinal ventral gap which separates a major ossification spreading on the dorsolateral face of the mandible from a ventromedial ossification. They may correspond to the prearticular and the surangular contributions to the compound bone, respectively.

Figure 7.

Lateral and medial views of the lower jaw of Amerotyphlops brongersmianus throughout ontogeny: A, embryo at Stage 33; B, hatchling; C, adult. Abbreviations: a, angular; co, coronoid; cb, compound bone; d, dentary; gc, glenoid cavity; mf, mandibular fossa; rp, retroarticular process; sp, splenial. Scale bars equal to 1 mm.

Growth of the gnathic complex

The size changes of the bones of the gnathic complex—relative to linear skull growth—in A. brongersmianus involve an overall allometric growth, except for the dentary (Figs 7, 8). Notably, the pterygoid experiences a positive allometric growth during postnatal ontogeny as is also the case for the alethinophidian snake Philodryas psammophidea (Fig. 8). The relative size of the quadrate decreases with increasing ontogenetic stage in A. brongersmianus and in the anguimorph lizard Ophiodes intermedius, opposite to what occurs in macrostomous alethinophidian snakes, as exemplified herein by P. psammophidea (Fig. 8). This negative allometric growth of the quadrate, along with its postnatal lateral aperture, lead to the postnatal change of position of the anterior region of the quadrate with respect to the braincase mentioned above.

Figure 8.

Ratio between the length the bones of the gnathic complex and linear skull length plotted against postnatal ontogenetic stages of the anguimorph lizard Ophiodes intermedius, the blind snake Amerotyphlops brongersmianus and the alethinophidian snake Philodryas psammophidea. The successive postnatal ontogenetic stages were ordered according to SVL and named from 1 to 4 corresponding to hatching, juvenile, subadult and adult stages respectively.

Growth of the lower jaw shows similar trends across examined species, with a relative increase of the size occurring throughout ontogeny although less markedly pronounced in A. brongersmianus (Fig. 8). The relative growth of the dentary of A. brongersmianus, graphically represented as a flat line almost parallel to the x-axis, is noteworthy (Fig. 8). This bone grows isometrically along the ontogeny, suggesting that its small size in adult forms is the result of proportions established during early embryonic stages (Fig. 7A). Therefore, the compound bone is likely the main element contributing to the overall length of the lower jaw in A. brongersmianus.

DISCUSSION

The study of the skull ontogeny of A. brongersmianus allowed the description of transient structures (the egg tooth), the identification of some traits (closure of the skull roof, fusion of elements of otoccipital bone, ossification of the trabeculae cranii) for which development is delayed compared to alethinophidian snakes, and of phylogenetically relevant characters (the lateral wings of the parabasisphenoid) for blind snakes. Furthermore, the compiled information permitted us to link particular characteristics of the adult skulls with heterochronic processes, and relate them to the evolutionary processes that shaped cranial anatomy of this group (miniaturization and fossoriality).

Egg tooth

The presence of one egg tooth was described as the predominant condition in Squamata, and was employed along with molecular data to define the large clade Unidentata (Vidal & Hedges, 2005, 2009). In spite of the phylogenetic relevance of this character, knowledge about early stages of its development and differentiation is scarce (Hermyt et al., 2017, 2020a; Fons et al., 2019). Previous studies reported the egg tooth rudiments in the Unidentata as either paired or unpaired, and described two different developmental pathways for paired rudiments according to the squamate group (Anan’eva & Orlov, 2013; Hermyt et al., 2020b). In anguimorph lizards and alethinophidian snakes both rudiments merge to form a single egg tooth (Smith et al., 1952; Anan’eva & Orlov, 2013; Fons et al., 2019). This particular developmental pathway results in an egg tooth with a horizontally elongated base (Fons et al., 2019). If the shape of the base of the egg tooth is considered a predictor of its developmental origin, then the subcircular base of the egg tooth of A. brongersmianus may indicate no early convergence between rudiments, and therefore may involve a developmental pathway different from that of alethinophidians. However, histological or molecular tools to study early embryos are necessary to accurately test this hypothesis.

The egg tooth of A. brongersmianus described herein represents the first report of this structure in a blind snake. It remains distinctive to all egg teeth described for squamates up to now due to its unique position and remarkable length and orientation (De Beer, 1949; Smith et al., 1952; Trauth, 1988; Underwood & Lee, 2000; Anan’eva & Orlov, 2013; Hermyt et al., 2017, 2020a, b; Fons et al., 2019). We consider that the main features of the egg tooth of A. brongersmianus respond to a functional compromise with traits of the elements of the gnathic complex and the resulting particular morphology of the mouth. The snout region of blind snakes appears to have rotated ventrally (Cundall & Irish, 2008), shifting the premaxilla over the ventral surface of this region (Fig. 3). This shift may also entail a posterior displacement of the dentigerous zone of this bone and the consequent position of the egg tooth into the mouth cavity. In addition, as a result of the strongly reduced dentary, the lower jaw of typhlopids does not contact the tip of the snout when the mouth is closed (Rieppel et al., 2009). Accordingly, length and orientation of the egg tooth could be related to the ventrally located mouth opening.

Regarding its function, Smith et al. (1952) described the egg tooth of some viperids—Sistrurus catenatus (Rafinesque, 1818), Vipera aspis (Linnaeus, 1758) and Vipera berus (Linnaeus, 1758)—as pointing downwards, and suggested this orientation corresponds to a loss of function due to the viviparous condition of the species. However, as we noted herein, the egg tooth of A. brongersmianus in spite of being displaced into the mouth cavity and directed downwards, may still perform its function due to the ventrally directed mouth opening. Therefore, the egg tooth of A. brongersmianus may correspond to a new category of Fioroni’s (1962) classification, where three types of egg teeth were described for snakes according to characters of their morphology and function.

Braincase roof

Several authors recorded skull ossification in alethinophidian snake species and reported the onset of ossification of the parietals in the ventrolateral region of the skull, progressing dorsally and reaching the midline during embryonic development (Haluska & Alberch, 1983; Rieppel & Zaher, 2001; Boughner et al., 2007; Boback et al., 2012; Polachowski & Werneburg, 2013; Sheverdyukova, 2017; Da Silva et al., 2018; Al Mohammadi et al., 2020; Khannoon et al., 2020). In contrast, the fusion of both dorsal laminae of the parietal in A. brongersmianus occurs shortly before hatching, as evidenced by a fissure in the posteromedial region of the parietal in the hatchling (Fig. 2D). Likewise, other authors reported a mid-sagittal fontanelle or paired parietals in juvenile typhlopid snakes—Anilios bicolor and Typhlops jamaicensis (Shaw, 1802)—where left and right counterparts later co-ossify in a single parietal in adults (Evans, 1955; Palci et al., 2016). Therefore, the closure of the skull roof is a delayed process in blind snakes with respect to what occurs in alethinophidians, and may take place during postnatal ontogeny in some species.

Intraspecific variation on the parietal condition of Typhlops pusillus Barbour, 1914 has been inferred, since it was variably reported as paired (List, 1966) or fused (Thomas, 1976). However, List (1966) probably based his observation on an immature specimen, and the paired parietal condition is a misinterpretation of this author. Thus, the few available records on discrepancies of the parietal condition for the same species are not sufficiently sustained to consider this as an intraspecifically variable character. In addition, total length of the individuals, which may be a good proxy of age and hence ontogenetic stage, is not always provided in osteological descriptions. Accordingly, data on size (as total length or snout-vent length) of specimens should be addressed, so that future studies can take this information into account and benefit from it.

The persistence of an embryonic trait such as the parietal fontanelle in juvenile (Anilios bicolor) or adult stages (e.g. Namibiana Hedges, Adalsteinsson & Branch, 2009 and Myriopholis Hedges, Adalsteinsson & Branch, 2009) of blind snake species (Palci et al., 2016; Broadley & Wallach, 2007; Cundall & Irish, 2008; C.K., pers. obs.) may appear counterintuitive for fossorial organisms that are first-head burrowers (Herrel et al., 2021), and has been posed as a non-adaptive ontogenetic constraint (Palci et al., 2016). However, providing an adaptive explanation for every skull structure overlooks the idea that each ontogenetic stage is a fully functional unique organism. In this sense, the absence of a completely ossified braincase roof in early postnatal stages, juveniles or adults of some blind snake species indicates that this is not a prerequisite for burrowing. Furthermore, we hypothesize that a behavioural trade-off during postnatal ontogeny could exist, and those stages with incomplete ossification of the skull roof can either use tunnels and crevices of the nests of social insects or those elaborated by adult congeners, or they can even live in loose substrates.

Otic region

Lira & Martins (2021) reported interspecific variation regarding the participation of the prootic and the parabasisphenoid in the formation of the trigeminal foramen in Amerotyphlops reticulatus (Linnaeus, 1758) and A. brongersmianus. These authors also claimed that this character does not vary intraspecifically within blind snakes, and might be relevant in terms of diagnostic characters for species. However, we observed postnatal ontogenetic variation in this character in A. brongersmianus (Fig. 3E, F). Interestingly, the configuration of the trigeminal foramen in the A. reticulatus specimen of Lira & Martins (2021) is the same as that observed herein in the subadult of A. brongersmianus (Fig. 3E). The ontogenetic variation of a character may hide the presence of interspecific variation or, on the contrary, show false interspecific variation. The latter may be the case of the specimen of A. reticulatus of Lira & Martins (2021), since the character may have not completed its ontogenetic trajectory. In consequence, the taxonomic value of the formation of the trigeminal foramen is challenged, at least for Amerotyphlops species.

In most squamates, the opisthotic and exoccipital fuse during embryonic development to form the otooccipital bone (Greer, 1985; Estes et al., 1988; Maisano, 2001), although there are few records of unfused condition in hatchlings of lizards (Rieppel, 1992a, b, c), and of a suture between both elements in the neonates of Lacerta agilis Linnaeus, 1758 (Rieppel, 1994) and of xantusids (Maisano, 2002). In spite of this character being considered phylogenetically informative for squamates (Estes et al., 1988; Evans, 2008), it has received little attention in ontogenetic studies of snakes. As such, the presence of a suture at an early postnatal stage was only explicitly reported by Rieppel & Zaher (2001) for the caenophidian snake Acrochordus granulatus (Schneider, 1799). Hence, we report here for the first time, evidence of the onset of embryonic fusion of the exoccipital and opisthotic in a blind snake, and persistence of the suture between these bones in the hatchling stage of A. brongersmianus is also indicative of delayed fusion of these bones during embryonic ontogeny in comparison to alethinophidians.

Basicranium

Of the two elements that constitute the adult basicranium in snakes, the parabasisphenoid bone exhibits the most complex ontogenetic trajectory. This bone results from the fusion of the membranous parasphenoid and the chondral basisphenoid (De Beer, 1937; Bellairs & Kamal, 1981). The parasphenoid ossifies in the anterior region of the braincase floor, between the anterior ramus of the trabeculae cranii, while the embryogenesis of the basisphenoid involves the perichondral/endochondral ossification of the trabeculae cranii and crista sellaris, as well as membrane bone outgrowths closely related with these chondrocranial structures (De Beer, 1937; Bellairs & Kamal, 1981; Haluska & Alberch, 1983; Rieppel, 1988).

The alethinophidian parabasisphenoid was traditionally characterized by the development of broad lateral wings corresponding to those ossifications lateral to the trabeculae, and labelled with different names in literature [e.g. lateral bony wing of the parasphenoid, lateral ascending wing of the sphenoid, among others (McDowell, 1967, 2008; Rieppel, 1979a, b, 1988)]. Furthermore, the posterolateral region of the parabasisphenoid bone forms a complex structure for the passage of vessels and nerves—the cerebral carotid artery and the palatine ramus of the facial (VII) nerve—which gained an intracranial course during the evolution of the snake braincase (McDowell, 2008). Some authors claimed that the parabasisphenoid wings were absent in blind snakes, and that this character differentiated them from alethinophidians (McDowell, 1967; Rieppel, 1979a, 1988; Rieppel & Zaher, 2000). These authors also related the absence of this structure with the lack of a closed Vidian canal, a structure largely highlighted as an alethinophidian feature (McDowell, 1967; Rieppel, 1979b; Rieppel & Zaher, 2000). Of note are the available ontogenetic series of A. brongersmianus showing a well-developed ossification laterally to the trabeculae, and a well-defined posterior region of the Vidian canal, which is floored during the ontogeny in the same fashion as observed in alethinophidians (Figs 3, 6). Moreover, the intracranial course of the Vidian canal of A. brongersmianus resembles that of other alethinophidians such as Anilius scytale (Linnaeus, 1758) or the caenophidian Homoroselaps lacteus (Linnaeus, 1758) (Rieppel, 1979b; A.S., pers. obs.). Hence, an ossification lateral to the trabecula and a closed Vidian canal as a passage for the palatine nerve and the cerebral carotid is not an exclusive character of alethinophidians, but seems to be a common feature of extant snakes.

Closure of the braincase and burrowing

The descending processes of the frontal and parietal bones reach the parasphenoid rostrum and the parabasisphenoid wings in snakes, performing the lateral closure of the braincase, which represents an exclusive character of snakes among squamates (Bellairs & Underwood, 1951). A completely enclosed braincase, also reinforced by wide sutural contacts between bony elements, has been historically linked to fossoriality as a response to mechanical stress in head-first burrowers (Gans, 1974; Savitzky, 1983). However, whereas most extant snakes exhibit a well-developed suture contact between the floor and lateral wall of the braincase, the highly fossorial blind snakes can display loose contact between braincase elements, including fissures filled with fibrous connective tissue (List, 1966; McDowell, 1967; Cundall & Irish, 2008; Palci et al., 2016). Moreover, the braincase of other active burrowing squamatans such as amphisbaenians fails to fully enclose the endocranial cavity (Gans & Montero, 2008). As such, the relation between a walled and strongly reinforced braincase and burrowing habits does not seem to be straightforward.

Our ontogenetic study indicates that the braincase of the typhlopid A. brongersmianus undergoes gradual reinforcement during postnatal growth, exhibiting fissures or simple contacts between braincase elements in the juvenile and the subadult stages. This condition of early postnatal stages, as well as the paired parietals or poorly ossified skull roofs, was interpreted as ‘maladaptive’ or temporarily non-adaptive for burrowing blind snakes (Palci et al., 2016). Recently, the first comparative survey about burrowing forces in adult blind snakes posited that typhlopid snakes are able to generate higher forces for a given body length compared to other blind snakes and burrowing alethinophidians (Herrel et al., 2021). Considering this observation, and as mentioned above, we hypothesized about this apparent counterintuitive issue by proposing the existence of a behavioural trade-off along postnatal ontogeny. Young blind snakes may use formerly excavated tunnels such as those of insect nests, and probably increase their burrowing capabilities during postnatal advanced stages when skull ossification increases and bone contact is reinforced.

Interestingly, other skull traits traditionally considered crucial for burrowing in snakes have been recently challenged (Deufel, 2017). The shield-nosed cobra Aspidelaps scutatus (Smith, 1849) was described as being able to excavate and construct tunnels in loose substrates with a typical highly kinetic skull present in other surface-dwelling alethinophidians (Deufel, 2017). The apparent eco-morphological mismatch between non-solid/kinetic skulls of burrowing snakes can be interpreted as a case of ‘organic nonoptimal constrained evolution’ (Diogo, 2017). As such, it seems to be plausible that there are different morphologies and trade-offs in burrowing snakes that constitute different mechanisms for burrowing, a complex behaviour that repeatedly appeared in the evolutionary history of snakes.

Growth of the gnathic complex

Since the early works it has been highlighted that the gnathic complex of blind snakes strongly departs from the rest of the squamates (Haas, 1930; Dunn & Tihen, 1944; Tihen, 1945; Evans, 1955). Shortening and simplification of bony elements, lack of teeth in typically toothed bones, and loss of elements of the suspensorium are the most remarkable transformations present in blind snakes (Cundall & Irish, 2008). Additionally, bones of the gnathic complex experience isometric or even negative allometric growth during postnatal ontogeny (Palci et al., 2016; Scanferla, 2016). In this sense, A. brongersmianus displays a distinctive growth pattern of the bones of the gnathic complex (Fig. 8). The length of the quadrate shows a general growth of a negative allometric type, in accordance with the growth pattern described by Palci et al. (2016) for the typhlopid Anilios bicolor. In contrast, the pterygoid bone experiences remarkable longitudinal growth (Fig. 8) during postnatal ontogeny similar to that observed in macrostomous alethinophidian snakes (Fig. 8; Rossman, 1980; Scanferla, 2016). This is an unexpected trait for a non-macrostomous snake, since the pterygoid of macrostomous alethinophidians exhibits positive allometry, whereas most lizards and non-macrostomous snakes show isometric growth (Scanferla, 2016). Notably, the rod-like pterygoid in typhlopids serves as the insertion of the m. protractor pterygoidei, which almost entirely sheathes the bone (Haas, 1930; Iordansky, 1997). This muscle produces the forward displacement of the pterygoid and the consequent erection of the maxilla (Cundall & Rossman, 1993; Iordansky, 1997). As such, elongation of the pterygoid may be promoted by postnatal development of this muscle, whose insertion surface is relevant for muscle contraction during feeding.

Among the characters of the jaw complex of typhlopoids, the relative size reduction of the dentary also stands out (Strong et al., 2021). Our results indicate that this morphology is the outcome of prenatal established proportions as well as postnatal isometric growth. This reduced dentary has consequences for mouth configuration, since the lower jaw is not in contact with the tip of the snout region when it is fully abducted. Therefore, a unique mouth configuration among squamates is set up. In accordance to the morphology of the gnathic complex, a highly specialized mechanism for the rapid ingestion and transport of large numbers of prey named ‘single-axle maxillary raking’ was described for typhlopoids (Iordansky, 1997; Kley, 2001; Strong et al., 2021). The extremely short and toothless dentary of typhlopoids may constitute a functional prerequisite for this type of intraoral prey transport, since lower jaws simply act as passive scoops during feeding (Iordansky, 1997).

Heterochrony, miniaturization and the evolution of the skull of blind snakes

Heterochrony has been proposed as the most relevant developmental phenomenon producing morphological variation (De Beer, 1940; Gould, 1977; McKinney & McNamara, 1991). Therefore, heterochrony has been invoked to explain several osteological novelties of the snake bauplan and evolutionary trends in the group (Rieppel, 1988; Irish, 1989; Werneburg & Sánchez-Villagra, 2014; Da Silva et al., 2018; Strong et al., 2020). In particular, features of the highly modified skull of each blind snake clade have been explained through heterochrony, mostly assigned to paedomorphosis correlated with miniaturization (Rieppel, 1979a, 1988, 1996; Irish, 1989; Kley, 2006; Palci et al., 2016; Strong et al., 2019, 2020; Martins et al., 2021). However, these hypotheses were largely based on adult morphology, which contributes with weak evidence to understanding heterochronic processes (Hanken, 1993). Moreover, the heterochronic change between species must be polarized by outgroup comparison in the context of a phylogenetic hypothesis, in order to identify the heterochronic process between ancestral and descendant ontogenies (Fink, 1982; Reilly et al., 1997).

Our ontogenetic analysis of the skull of A. brongersmianus shows that development of some traits is clearly delayed regarding the developmental pattern seen in alethinophidians. This is the case of ossification of the parietals and the posterior region of the trabeculae cranii, fusion between opisthotic and exoccipital, and constitution of certain bony processes (e.g. postorbital and supratemporal processes of parietal). Ontogenetic trajectories of these characters do not have consequences on adult morphology, but their record allowed us to corroborate the previous hypothesis of peramorphosis through acceleration of ossification rates in alethinophidian skulls (Da Silva et al., 2018). Furthermore, the ontogenetic data on the anguimorph lizards La. borneensis and Varanus panoptes Storr, 1980 (Werneburg et al., 2015) suggests that the developmental rates of A. brongersmianus resemble those of lizards.

Among the cranial features, the skull roof of blind snakes has been pointed out repeatedly as a paedomorphic trait in the literature (Palci et al., 2016; Da Silva et al., 2018; Lira & Martins, 2021; Martins et al., 2021). Although most species display an azygous parietal in adult forms, others exhibit paired bones in contact with one another, separated by a fissure of variable width, or even a large mid-sagittal fontanelle occupying the skull roof (Broadley & Wallach, 2007; Lira & Martins, 2021 and literature cited therein). Therefore, different configurations of the adult cranial roof in blind snakes can be obtained either via prolongation or truncation of the ossification of the dorsal laminae of the parietals. Since there is no truncation of the skull roof development in A. brongersmianus, this trait cannot be termed as paedomorphic in this species, neither in Typhlopidae with an azygous parietal. However, poorly ossified skull roofs in adults of some blind snake taxa resemble the incompletely ossified skull roofs reported for postnatal stages in several lizard lineages (Maisano, 2001; Hernández-Jaimes et al., 2012; Roscito & Rodriguez, 2012; Werneburg et al., 2015; Skawiński et al., 2021) and do indicate signs of paedomorphosis (Palci et al., 2016; Da Silva et al., 2018). Remarkably, these are forms belonging to the family Leptotyphlopidae, which is known for having a smaller overall size than typhlopids. Thus, it is worth asking if a relationship between adult skull roof morphology and mean adult body size can be established, or if a gradient in parietal ossification corresponding with decrease in adult body size exists along blind snake families?

The anteriorly oriented quadrate of blind snakes is a trait typically attributed to paedomorphosis (Caldwell, 2019; Strong et al., 2020). Our observations of A. brongersmianus skull ontogeny showed that the quadrate has an almost horizontal orientation in embryos and remains the same during postnatal development. Limb-reduced lizards and non-macrostomous alethinophidian snakes have a vertically or slightly anteriorly oriented quadrate, and the lack of rotation along ontogeny was also reported for these groups (Montero et al., 1999; Roscito & Rodriguez, 2012; Werneburg et al., 2015; Scanferla, 2016). In contrast, the quadrate of macrostomous alethinophidians experiences a noteworthy counter-clockwise rotation during ontogeny (Bellairs & Kamal, 1981; Rieppel, 1988; Palci et al., 2016; Scanferla, 2016). In this sense, the anteriorly oriented quadrate of adult blind snakes does not resemble the condition present in embryonic or juvenile stages of lizards and cannot be explained simply by truncation of the ancestral ontogenetic trajectory, as was previously noted by Rieppel (1988).

The most common effect of miniaturization on morphology is skeletal reduction and loss of bony elements (Hanken & Wake, 1993). Furthermore, miniaturization in tetrapods has been hypothesized as being caused mainly by paedomorphosis (Irish, 1989; Rieppel, 1996). However, in spite of miniaturized tetrapods showing a certain degree of paedomorphosis in skull morphology, there is a broad spectrum of effects from heterochronic development (Irish, 1989; Rieppel, 1996). In comparison to lizards and alethinophidian snakes, the skull of blind snakes exhibits strong reduction and loss of some cranial bones (e.g. supratemporal, jugal) and lack of teeth in different toothed bones (e.g. maxilla, dentary; Cundall & Irish, 2008). These absences are not easily attributed to paedomorphosis but probably denote a more fundamental alteration of the processes underlying skeletal morphogenesis and bone differentiation (Hanken, 1993). Additionally, we observed some characters in the skull of A. brongersmianus that can be attributed to peramorphosis. The development of a crista circumfenestralis that completely closes the juxstastapedial recess represents a trait originated by an extension of the ancestral ontogenetic trend. Likewise, the allometric postnatal growth of the pterygoid bar observed in A. brongersmianus, resembling the growth pattern described for macrostomous alethinophidians (Scanferla, 2016) is another peramorphic trait of this blind snake species.

Finally, beyond the heterochronic traits discussed previously, there are further features present in the skull of blind snakes that can be linked to their basal position in the tree of extant snakes. The absence of typical alethinophidian bony structures, such as the medial frontal pillars and the ophidiosphenoid bone, can be better explained as plesiomorphies shared with stem snakes such as Dinilysia patagonica Smith-Woodward, 1901 (Zaher & Scanferla, 2012). Thus, the unique skull anatomy of blind snakes seems to represent a combination of plesiomorphic and highly autapomorphic features, shaped through a complex interplay of heterochronic development, miniaturization, functional demands of fossorial lifestyle and historical contingency.

CONCLUSION

The peculiarity of the skull anatomy of blind snakes appears to be the result of a combination of plesiomorphic traits shared with lizards and stem snakes, along with highly autapomorphic traits shaped via heterochronic processes and miniaturization. These traits have been influenced by functional constraints and selective pressures relative to a fossorial lifestyle. Our ontogenetic analysis of the skull of A. brongersmianus demonstrates that some features such as skull roof morphology, quadrate orientation or absence of some cranial bones cannot be assigned to paedomorphosis, while others can be attributed to peramorphic processes. The ontogenetic series studied herein evidenced that the development of some traits (closure of the skull roof, fusion of elements of otoccipital bone, ossification of the trabeculae cranii) is clearly delayed in comparison to the developmental patterns seen in alethinophidian snakes, and resemble those of anguimorph lizards. Furthermore, the morphology of the egg tooth, and the presence of lateral wings of the parabasisphenoid and the posterior region of the Vidian canal were described for the first time for the group. Finally, we showed that different elements of the gnathic complex can have decoupled and even opposed growth patterns during postnatal ontogeny, and that the braincase of A. brongersmianus undergoes a gradual reinforcement during postnatal growth. Further information on the size and shape ontogenetic changes of the skull of members of the other clades of blind snakes is still needed. Additional studies to test the correlation between size and presence of paedomorphic cranial features, and whether there is any relevant phylogenetic relationship, are also necessary.

SUPPORTING INFORMATION

Additional Supporting Information may be found in the online version of this article at the publisher’s web-site:

Table S1. Micro-CT scanning parameters for each specimen analysed.

ACKNOWLEDGEMENTS

We are indebted to Maria Teresa Sandoval who kindly provided the specimens under her curatorship at UNNEC. We also thank David Blackburn, Matthew Gage, Joseph Martinez, Jennifer Olori, Alan Resetar, Alessandro Palci, Sara Ruane, Greg Schneider and Andrea Villa who generously provided micro-CT data for several squamate species. We would also like to acknowledge the Morphosource and Digimorph teams for supplying valuable data. Financial support was received from Consejo Nacional de Investigaciones Científicas y Técnicas (scholarship to M.C.) and Agencia Nacional de Promoción Científica y Tecnológica (Proyecto de Investigación Científica y Tecnológica 2013-220 and 2020-02443). Finally, we are grateful to the two anonymous reviewers for their helpful comments, which greatly improved this manuscript. The authors have no conflict of interests to declare.

DATA AVAILABILITY

The data underlying this article will be shared on reasonable request to the corresponding author.