Dispatch R513 Dispatches Palaeontology: Turtles in Transition One of the major remaining gaps in the vertebrate fossil record concerns the origin of turtles. The enigmatic little reptile Eunotosaurus could represent an turtles [13,14], while analyses important transitional form, as it has a rudimentary shell that resembles the focused on identifying turtle turtle carapace. relatives within diapsids excluded poorly known anapsids (such as Michael S.Y. Lee of the amniote radiation. One anapsid Eunotosaurus) [11,12]. lineage, the ‘parareptiles’, includes This Gordian knot was recently cut Turtles (tortoises, terrapins and sea three historical contenders for turtle when the striking similarities between turtles) have a very bizarre and highly relatives: the procolophonids — Eunotosaurus and turtles were modified anatomy that has long lizard-shaped reptiles with often reiterated [15,16], and the two taxa hindered attempts to decipher their spinose skulls [5]; the pareiasaurs — were finally simultaneously included in evolutionary origins and relationships. large, stout animals varyingly covered rigorous phylogenetic analyses [15]. The most notable feature of their highly with armour plates [6]; and, The results were intriguing (Figure 1A). aberrant body plan is the external shell, Eunotosaurus — an odd little creature When turtles were added to analyses which incorporates vertebrae, ribs, with a short rigid body encased in wide of anapsids, they fell next to shoulder and sometimes the pelvis. leaf-shaped ribs [7]. However, all Eunotosaurus (and thus within This highly derived anatomy means putative anapsid relatives fell from parareptiles in general). When that morphological traits are often not favour when genomic data robustly Eunotosaurus was added to analyses readily comparable between turtles placed turtles within diapsid reptiles, focusing on diapsids, it again fell with and their putative relatives, leading to usually as sister-group to archosaurs turtles, this pairing again nesting within numerous disputed homologies. (birds and crocodiles) [8–10]. This parareptiles. There was a consistent Therefore, turtles have been arrangement implied that turtles could morphological signal uniting turtles particularly difficult to place within the not be related to any primitively anapsid with Eunotosaurus in particular, and reptile evolutionary tree [1,2]. Now, reptiles: rather, their anapsid-like skulls with parareptiles generally. writing in this issue of Current Biology, must be secondary (atavistic) rather Lyson and colleagues [3] undertake a than representative of the primitive Eunotosaurus: No Longer a re-evaluation of the neglected Permian reptilian condition, and their nearest Pariah-Saur (w260 million year old) fossil relatives should be sought amongst The potential importance of Eunotosaurus. Their analysis reveals diapsid reptiles, notably Eunotosaurus as a transitional taxon that this small, stiff-bodied terrestrial sauropterygians, an extinct clade has spurred a detailed reassessment reptile possessed an expanded ribcage that includes marine reptiles such of the carapace-like structure of this that shares many detailed similarities as plesiosaurs, placodonts and neglected reptile, known from a with the turtle carapace [3]. The overall ichthyosaurs [11]. The sauropterygian handful of good specimens from South morphology of Eunotosaurus is also hypothesis raised the possibility that Africa. In this issue of Current Biology, consistent with that of a turtle ancestor turtles evolved in the ocean, boosted by Lyson and colleagues [3] now predicted by recent ontogenetic the recent discovery of the most document additional turtle-like features studies. These discoveries should shed primitive known turtle, the small aquatic in Eunotosaurus, which encompass light on the broader phylogenetic Odontochelys [12]. However, this gross anatomy as well as fine structural relationships of turtles, and the hypothesis has some inconsistencies: detail. As in turtles, the trunk region evolutionary origins of their highly for instance, while genomic data place of Eunotosaurus is wide and stiff, distinctive body plan. turtles with archosaurs [8–10], consisting of only 9 elongate vertebrae sauropterygians are generally each with a pair of broadened Evolutionary Relationships Turning considered related to the other major leaf-shaped ribs (Figure 1a). Other Turtle living branch of diapsids, the reptiles typically have over twenty short Despite their uniquely specialised lepidosaurs (lizards, snakes and vertebrae, each with narrow cylindrical bodies, turtles have rather primitive tuataras) [11,12]: if both relationships ribs. The similarities also extend to the ‘anapsid’ skulls, characterised by a are true, then turtles and underside. Most reptiles have multiple solid cheek region, an arrangement sauropterygians cannot be close kin. longitudinal rows of rod-like bones resembling that of early reptiles. In The emerging consensus that turtles along their belly (gastralia), whereas contrast, all other living reptiles have were aberrant diapsid reptiles stymied Eunotosaurus has only two rows, more advanced ‘diapsid’ skulls with further consideration of anapsid-grade perhaps a precursor to the turtle two large openings (fenestrae) in the relatives, including all parareptiles. plastron which similarly consists of two cheek region [4] (Figure 1). Based on Eunotosaurus was thus overlooked in rows of fused bony plates. skull morphology, the search for turtle recent debates on turtle origins. Eunotosaurus also appears to have ancestors historically focused on Analyses focused on relationships lost intercostal muscles (which extinct anapsid-grade reptiles, the among Eunotosaurus and other normally extend between the ribs earliest and most primitive members anapsids explicitly excluded and are involved in breathing and Current Biology Vol 23 No 12 R514 A Turtles Archosaurs Lepidosaurs fans out to overhang the shoulder 0 Genomic ‘scaffold’ girdle. Furthermore, development B from single rather than multiple enozoic primordia suggests that costal Cenozoic C SauropterygiansSauropterygians es s elements evolved via broadening of rtl 66 aurs an Turtles Tu idosaurs p ribs alone, rather than via fusion of e terygi Lepidosaurs L Archos Archosaurs s rib and overlying armour plating saurus Saurop Sauropterygians o s oto [17,18]. All this ontogenetic evidence tile retaceou p Euno Eunotosaurus C suggests that the earliest stages in Parare Parareptiles the evolution of the turtle shell should 145 c be represented by an animal with only expanded ribs (costals), no neurals, a Jurassi ParareptilesParareptilees 201 C shoulder girdle anterior to a wide trunk c i region, and no dermal s ass es ri s armour — precisely the Gestalt T rtl aurs u osaur ian T Turtles g 252 id EunotosaurusEunotosauruus exhibited by Eunotosaurus [3]. ep L Lepidosaurs Archos Archosaurs n While the new studies consistently Sauroptery Sauropterygians ermia PermianP Triassic Jurassic Cretaceous unite Eunotosaurus with turtles, saurus s 299 DiapsidDiapsid skullskull noto s many questions remain. First, the ptiles Euno E Eunotosaurus rare Anapsid skull a striking similarities between Parareptiles Eunotosaurus and turtles are currently arboniferou 359 Carboniferous C all associated with a single adaptive Current Biology complex (the shell), raising the possibility of convergent evolution. Figure 1. Phylogenetic relationships and scenarios for turtle evolution. Further study of this reptile is required, Affinities of Eunotosaurus and turtles, suggested by morphology alone (left) and morphology in to identify turtle-like features the context of genomic evidence (right). Anapsid-skulled taxa in blue, diapsid-skulled taxa in (synapomorphies) in other anatomical red; extinct taxa in light shading, living taxa in darker shading. (A) Morphology alone suggests areas, especially in the poorly-known that Eunotosaurus is a stem-turtle, and places both taxa within anapsid-skulled reptiles called skull region [16]. Such traits would parareptiles, outside the clade of all diapsid reptiles (archosaurs, lepidosaurs and sauroptery- more robustly corroborate the link gians) [3,15]. However, this position conflicts with robust genomic evidence placing turtles between Eunotosaurus and turtles. within diapsids, next to archosaurs [8–10]. If one accepts the latter evidence, two possible res- olutions to this dilemma are: (B) Eunotosaurus is a stem-turtle with a primitive carapace, Second, the position of the and thus falls within the turtle-archosaur clade, perhaps along with sauropterygians; Eunotosaurus turtle clade within (C) alternatively, Eunotosaurus is an anapsid-grade parareptile, distant from the turtle-archo- parareptiles is acknowledged to saur clade, and has convergently evolved numerous carapace-like traits. be unstable, alternating between near the lizard-shaped millerettids [15] locomotion), again a novel condition ancestral turtle lineage just before and the large stout pareiasaurs [3]. characteristic of turtles. The histology Odontochelys (most primitive ‘true’ These positions have implications of the ribs of Eunotosaurus reveals turtle) elucidates the order in which for wider homologies of some major Sharpey’s fibres (indicating muscle turtles evolved their novel adaptations turtle novelties, e.g. some attachments) are only present on [3]. Notably, the short trunk region, millerettids have broadened ribs and the ventral (i.e. internal) surface, expanded ribs (i.e. costal bones in the possible precursors of the turtle