The distinct synsacrum of knob‐tailed geckos (Gekkota: Carphodactylidae: <i>Nephrurus</i> )

Sacral vertebral numbers in reptiles vary across phylogeny and evolutionary history, but the basal condition for the clade Neodiapsida (Archelosauria, Lepidosauromorpha, and several extinct groups) is two vertebrae (Gauthier et al., 1988; Hoffstetter & Gasc, 1969; Nesbitt et al., 2015; see also Buffa et al., 2022). Additional vertebrae from the thoracic, lumbar, and pygal series are sometimes incorporated to form a synsacrum (Kardong, 2018; Paparella et al., 2020). The sacrum in Neodiapsida shows some contrasting patterns in the number of vertebrae. In Archelosauria, there is a tendency to increase the number of sacral vertebrae (Bona & Alcalde, 2009; Nesbitt, 2011), while most lepidosauromorph reptiles retain two sacral vertebrae, and there is a trend to reduce the sacral numbers in limb-attenuated forms (Borsuk-Białynicka, 2008; Hoffstetter & Gasc, 1969; Tschopp, 2016). We have represented this variation by mapping sacral vertebrae counts from the literature onto a metatree of neodiapsid reptiles, which combines phylogenies of Lepidosauromorpha (Pyron et al., 2013), Archosauria (Nesbitt, 2011), and turtles, according to the broader base tree by Crawford et al. (2015; Figure 1). Turtles generally possess just two sacral vertebrae, although some pleurodire species incorporate one or two additional vertebrae into the sacrum (Bona & Alcalde, 2009; Olukole et al., 2014; Sheil, 2003; Valente et al., 2007; Zehtabvar et al., 2022). Within Pseudosuchia, modern crocodylians retain two sacral vertebrae, while others (e.g., Sillosuchus) possess as many as five (Nesbitt, 2011; Stocker, 2019). Pterosaurs are notable in that they possess a minimum of three sacral vertebrae, but the exact number is highly variable across taxa, with some confusion arising due to the overwhelming prevalence of juvenile specimens which may not represent the full extent of vertebral fusion that exists in adults (Hyder et al., 2014; Vidovic & Martill, 2014). Within dinosaurs, an even greater degree of variability exists. The ancestral condition in Dinosauromorpha appears to be two sacral vertebrae (Moro et al., 2020). Incorporation of additional sacral vertebrae has been considered a characteristic of true dinosaurs; however, a paucity of fossils from the Triassic makes it difficult to determine if this is the plesiomorphic condition (Moro et al., 2020). Early dinosaurs (e.g., Herrerasauridae, Silesauridae; Norman et al., 2022) have been found to possess three sacral vertebrae with varying degrees of fusion (Bittencourt & Kellner, 2009; Dzik & Sulej, 2007; Piechowski & Dzik, 2010). Ornithischians possess an increased number of sacral vertebrae, ranging from 5 to 11 in some ceratopsians (Carrier & Farmer, 2000; Erickson, 1966). Fossils of Guaibasaurus, typically considered an early sauropodomorph, have been found with two intact sacral vertebrae, although their morphology suggests that there may be a missing third sacral (Langer & Benton, 2006). Basal sauropods possess four sacral vertebrae, with an increase to five in more derived taxa, and the addition of a sixth in titanosaurs (Wilson, 2011). Theropods show a similar increase in sacral vertebrae counts throughout their evolution. Basal neotheropods, such as Liliensternus lilensterni, may possess as few as three sacral vertebrae, but the typical condition in later, non-avian theropods is a synsacrum of five vertebrae, with an increase to six or seven in ceratosaurians (Bittencourt & Kellner, 2004; Moro et al., 2020). A substantial increase is seen in the modern avian pelvis where at least nine, and regularly as many as 12 or 14, fused vertebrae are present (Berger, 1956; Christ et al., 2000; Rashid et al., 2014). In modern birds, it is hard to determine which of the many vertebrae in the synsacrum are the sacrals, although in embryos, two or more vertebrae near the acetabulum appear to be the original sacrals (Jollie, 1962). From this general review, it appears that the formation of a synsacrum has occurred independently in multiple lineages of archosauromorphs. Lepidosauromorph reptiles are far more conservative, almost invariably retaining the basal condition of two vertebrae (Borsuk-Białynicka, 2008), including early diverging fossil taxa like Cryptovaranoides, Huehuecuetzpalli, and possibly in Parviraptoridae (Benson et al., 2025; Reynoso, 1998; Whiteside et al., 2022). Among living lepidosaurs, Sphenodon and nearly all non-ophidian squamates possess just the basal two sacral vertebrae (Hoffstetter & Gasc, 1969; Holder, 1960; Seligmann et al., 2008; Werner, 1971). This includes, to our knowledge, all dibamids, nearly all gekkotans and lateratans, all scinciformatans, and most toxicoferans (Di-Poi et al., 2010; Malashichev, 2001; Moch & Senter, 2011; Reynoso & Callison, 2000). In general, little attention has been paid to the variation of sacral vertebrae numbers which is present in Lepidosauromorpha. In this paper, we present observations from x-rays and high-resolution computed tomography (HRCT) in combination with accounts from the literature of the condition of the sacrum in a variety of squamates. We focused on knob-tailed geckos (Nephrurus with 10 species), which have been previously reported as having additional sacral vertebrae (Bauer, 1990; Couper et al., 1993; Hoffstetter & Gasc, 1969; Holder, 1960). We considered additional vertebrae from the lumbar or pygal series to be part of the synsacrum if they were in contact with the ilium (via a cartilaginous joint) and/or functionally linked to the sacrum by being at least partially fused with the other sacrals. Our examination of x-rays of specimens representing 12 lizard families verifies the typical condition of two sacral vertebrae in squamates (Hoffstetter & Gasc, 1969), as well as some notable exceptions. Within Toxicofera, these exceptions include modern snakes, which possess no sacral vertebrae, although the fossil species Pachyrhachis problematicus and Najash rionegrina possess one and two sacrals, respectively (Apesteguía & Zaher, 2006; Lee & Caldwell, 1998). Species in Iguania and Anguimorpha have the usual two sacrals (Akita, 1992; Čerňanský et al., 2019; Rosa et al., 2025). In limb-attenuated squamates (dibamids, pygopods, skinks, fossorial gymnophthalmids, amphisbaenians, anguids, and snakes), it is common to find only vestiges of the pelvic girdle (Urben et al., 2014), and the iliosacral joint shifted dorsally (Borsuk-Białynicka, 2008). Members of Dibamia have been reported to have one sacral (Borsuk-Białynicka, 2008; Hoffstetter & Gasc, 1969) although in recent papers using HRCT the number seems to be two (Kliukin et al., 2023; Koppetsch et al., 2019). In some specimens, there have been reports of anomalous sacra composed of three vertebrae, but with asymmetric fusion of pleurapophyses (S1 and S2 on the left and S2 and S3 on the right; see illustration of Gerrhosaurus falvigularis in Hoffstetter & Gasc, 1969). The same anomalous fusion is seen in Cordylus namakuiyus (Stanley et al., 2016; https://sketchfab.com/3d-models/cordylus-namakuiyus-skeleton-6aa5d09c910341d59e7fcb013cd33677). We found similar asymmetrical sacra in two specimens of the gecko Diplodactylus pulcher (WAM R65775, WAM R67075). In modern Amphisbaenia, there is no connection between the pelvic girdle and the vertebral column; therefore, there are no vertebrae which can be assigned to the sacrum (Hoffstetter & Gasc, 1969), although in Amphisbaena caeca, we observed two vertebrae between the last thoracic and first pygal vertebrae. The fossil Slavoia darevskii does possess two robust sacrals connected to the pelvis, although its phylogenetic position within Amphisbaenia has been called into question (Simões et al., 2018; Tałanda, 2017). Previous studies using skeletonized, cleared, and stained specimens, and x-rays of squamates (Bauer, 1990; Hoffstetter & Gasc, 1969; Holder, 1960; Kluge, 1962; Wellborn, 1933) have only identified a few taxa with three symmetrical (i.e., non-anomalous) sacral vertebrae. The majority of gekkotans have two sacrals, but pygopods seem to have one sacral (Hoffstetter & Gasc, 1969; this study). Although retained in other genera, in Aprasia the connection between the pelvic girdle and the sacrum is completely lost (Shea, 1993). Eublepharid geckos, Uroplatus, some Phyllurus platurus, and all species of Nephrurus have been reported to possess an extra vertebra from the pygal series to form a synsacrum (Bauer, 1990; Hoffstetter & Gasc, 1969; Holder, 1960; Kluge, 1962; Wellborn, 1933). In Uroplatus, we found that the lateral processes of the first caudal pygal vertebra are directed toward the sacrum, but we did not observe any bony contact (Uroplatus fimbriatus AMNH-herp-173394, Uroplatus henkeli JFBM 15833), while in eublepharids this seems variable (Kluge, 1962). In Nephrurus, this trait has been previously described and found to be variable for the genus, with individuals possessing between two and four sacrals (Bauer, 1990; Holder, 1960), but it has never been illustrated in the literature, and little has been said about the nature or evolutionary reason for this unique morphological modification among the Lepidosauria. We confirmed that the typical condition in all examined species of Nephrurus is three sacral vertebrae (Holder, 1960; Bauer, 1990; Figure 2), although there was variation in sacral numbers in some taxa (Table 1). In addition to those specimens possessing the typical three, we found individuals with just two sacrals in Nephrurus levis, Nephrurus stellatus, Nephrurus asper, Nephrurus vertebralis, and Nephrurus laevissimus. In Nephrurus wheeleri, we observed three sacral vertebrae in four specimens and one individual (WAM R116901) possessed a third sacral vertebra from the pygal series and a fourth vertebra from the lumbar series with an asymmetrical sacral rib on the right side. We found no individuals with normal, symmetrical fourth sacrals, although Holder (1960) reported this condition in N. asper s.l. We extended our observations to other members of the family Carphodactylidae. We found two sacral vertebrae in all specimens of Carphodactylus laevis (n = 6), Orraya occultus (n = 1), Saltuarius cornutus (n = 4), Saltuarius salebrosus (n = 4), P. platurus (n = 7), Phyllurus caudiannulatus (n = 3), and Uvidicolus sphyrurus (n = 3). In addition to our findings, Holder (1960) reports that P. platurus may occasionally have three sacral vertebrae. In Underwoodisaurus milii, we found five individuals with the typical two sacral vertebrae, and one individual (NHMW 17426) with an additional asymmetrical rib on the right side. Both U. milii and U. sphyrurus were previously placed in Nephrurus (Oliver & Bauer, 2011), but typically possess just two sacral vertebrae, compared to three in Nephrurus (Bauer, 1990). Despite this minor variation, the predisposition to develop a synsacrum is a putative synapomorphy for the genus Nephrurus and may therefore be phylogenetically informative. It is difficult to establish a functional explanation for the prevalence of synsacra in Nephrurus geckos. Considering the dichotomy of Neodiapsida, between Archelosauria and Lepidosauromorpha, one might correlate the presence of a synsacrum with their predominant locomotor gait—parasagittal in Archelosauria and sprawling in lepidosauromorphs (Dingus & Rowe, 1997). In synapsids and archosaurs, the transition from sprawling to parasagittal locomotion required some reorganization of the musculoskeletal system (Brocklehurst et al., 2022; Gallina & Otero, 2008; Lamas et al., 2014; Romer, 1923). In reptiles, two powerful ventral limb muscles—the caudofemoralis longus and caudofemoralis brevis—arise from the underside of the sacral and caudal vertebrae (Romer, 1923; Romer & Parsons, 1977). In Iguana iguana, the caudofemoralis brevis inserts on pygal vertebrae (postsacrals lacking haemal arches) while the caudofemoralis longus inserts on true caudal vertebrae (Russell & Bauer, 2008). In geckos, it has been reported that the caudofemoralis brevis originates from the transverse processes of the sacral vertebrae or the posterior margin of the ilium, and that the longus originates from the first four or five caudal vertebrae (Zaaf et al., 1999). We investigated this topic further using diffusible iodine-based contrast-enhanced computed tomography (diceCT) of two carphodactylid taxa. We looked at the muscles that interact with the sacrum and tail in geckos—P. platurus with two sacrals and a broad tail, and N. asper with three sacrals and a stumpy tail. There is a posterior shifting of the caudofemoralis muscles in N. asper, when compared with P. platurus. In P. platurus, the caudofemoralis brevis originates from the ventral surface of the transverse processes of the second sacral and first pygal and the caudofemoralis longus originates from the vertebral bodies of the second–fourth pygals. In N. asper, the caudofemoralis brevis does not interact with the first two sacrals but originates from the ventral surface of the transverse processes of the third sacral (first pygal) and second pygal, while the caudofemoralis longus originates from the vertebrae bodies of the third–fourth caudal pygals, and fifth caudal (i.e., the first non-pygal caudal; Figure 3). Other caudal synapomorphies of these geckos include the distinctive enlarged “knob” at the end of the tail, a reduction in tail length and number of caudal vertebrae, and lack of autotomy planes (N. asper group) or their limitation to a single vertebra (all others), which has been proposed to be a byproduct of heterochrony (Bauer, 1990; Russell & Bauer, 1987; Oliver & Bauer, 2011). It is not possible to determine the order of appearance of these caudal characters outside of the context of a phylogenetically informed analysis. The caudofemoralis muscle inserts on the femur and typically functions to retract the hindlimb, but when the hindlimb is fixed it has the action of bending the tail (Russell & Bauer, 1992). During their defensive display, Nephrurus geckos fully extend their legs and raise their hindquarters. The posterior extension of the sacrum in Nephrurus shifts the origin of the caudofemoralis muscles posteriorly. In some knob tailed geckos (e.g., N. asper group), the tail has lost its function as defense (i.e., distracting predators with the autotomized tail) and fat storage (Oliver & Bauer, 2011) and is used in caudal behavioral display in conjunction with the bizarre terminal knob, and their facultative parasagittal gait. In eublepharid geckos with an additional sacral vertebra (Kluge, 1962; Wellborn, 1933), the tail is not reduced and has not lost the fat storage capability. In summary, the synsacrum of knob-tailed geckos indeed produces a reorganization of the caudofemoralis muscles, which might be linked to other morphological and behavioral adaptations involving the tails of this group. More information is needed about pelvic muscles in other geckos and reptiles with a synsacrum to confirm whether they also reorganize the caudofemoralis muscles. Additional observations in other squamates can help us to understand the rarity of synsacra in Lepidosauromorpha. Peter J. Babcock: Conceptualization; methodology; software; data curation; investigation; validation; visualization; writing – original draft; writing – review and editing. Aaron M. Bauer: Conceptualization; investigation; writing – review and editing; project administration; writing – original draft. Juan D. Daza: Conceptualization; methodology; software; data curation; investigation; validation; visualization; supervision; project administration; writing – original draft; writing – review and editing. Daniel J. Paluh: Visualization; writing – review and editing; writing – original draft; investigation. Edward L. Stanley: Methodology; software; data curation; investigation; formal analysis; funding acquisition; visualization; writing – original draft; writing – review and editing. We thank the curators and collection managers of the following museums for access to x-ray, high-resolution computed tomography, and diffusible iodine-based contrast-enhanced computed tomography data: California Academy of Sciences (CAS); Museum of Vertebrate Zoology at the University of California, Berkeley (MVZ); University of Florida, Gainesville (UF); American Museum of Natural History (AMNH); Natural History Museum of Los Angeles County (LACM); Field Museum of Natural History, Chicago (FMNH); Museum of Comparative Zoology at Harvard University (MCZ); Centrum für Naturkunde, Universität Hamburg (ZMH); South Australian Museum (SAMA); Naturhistorisches Museum, Vienna (NHMW); Australian Museum, Sydney (AMS); and Western Australia Museum, Perth (WAM). We also thank the James Ford Bell Museum of Natural History at the University of Minnesota for access to specimens in their collection. United States National Science Foundation, oVert: Open Exploration of Vertebrate Diversity in 3D, DBI 1701714 (Edward L. Stanley). All computed tomography and diffusible iodine-based contrast-enhanced computed tomography data are available from Morphosource. x-ray data will be available from the authors upon request.