L17 - Turtles: Building a Shell From Ribs

where do turtles fit on the amniote tree? old view

the solid skull roof (“anapsid”) allied turtles with parareptiles - especially small pariasaurs

where do turtles fit on the amniote tree? molecular view

>1000 ultra conserved elements and turtle genomes place turtles as sister to archosaurs

where do turtles fit on the amniote tree? genomic split

turtle-archosaur divergence estimated at about 57 million years ago

where do turtles fit on the amniote tree? caveat

microRNAs once grouped turtles with lizards; a reanalysis supported archosaurs

what makes a turtle a turtle?

• bony shell

• keratin scutes

• toothless beak

• ribs outside the girdles

• flexible neck

• hinged shells (some)

what makes a turtle a turtle? bony shell

carapace (dorsal, usually domed) + plastron (ventral, usually flat), joined by a bridge

what makes a turtle a turtle? keratin scutes

horny shields cover the bony shell (lost in soft shells and leatherbacks)

what makes a turtle a turtle? toothless beak

jaws covered by a keratin ehampthotecha; no teeth in crown turtles

what makes a turtle a turtle? ribs outside the griddles

shoulder and hip girdles lie beneath the carapace

what makes a turtle a turtle? flexible neck

cervical joints allow the head to be tucked or withdrawn

what makes a turtle a turtle? hinged shells (some)

plastron hinges evolved at least 8 times

what makes a turtle a turtle? Alabama examples

box and mid turtles shut their shells with plastral hinges; mud and mush turtles have one or two hinges

what makes a turtle a turtle? convergence

hinges arose independently in box turtles, asian box turtles, African mud turtles, and kinodternids; hinge-back tortoises hinge the carapace instead

the “anapsid” turtle skull is

secondary

the “anapsid” turtle skull - pappochelys

two clearly defined temporal openings

the “anapsid” turtle skull - eorhynchochelys

upper opening closed; lower opening open

the “anapsid” turtle skull - crown turtle

temporal region roofed over; openings lost

the “anapsid” turtle skull - reading the sequence

openings closed step by step along the stem, upper first

the “anapsid” turtle skull - one hypothesis

an unfenestrated temporal region spreads neck-muscle forces more evenly

eunotosaurus (Bever) - what they did

• scanned skulls of eunotosaurus Africans with CT scans

• scored them in a new character-taxon matrix

• tested whether skull characters alone support a turtle link

eunotosaurus (Bever) - ontogeny evidence

• juvenile skull: “clear expression of both” loser and upper temporal openings

• opening shrink late in postnatal growth

• an elongating supratemporal bone covers the upper opening

eunotosaurus (Bever) result

“strong cranial support” for Eunotosaurus as a stem turtle

eunotosaurus (Bever) - timing

extended the turtle stem by 40 million years

eunotosaurus (Bever) setting

moved the ecological context of turtle origins “back onto land”

eunotosaurus (Bever) - teaching point

counting skull openings is a misleading guide to relationships

June 2026: Is Eunotosaurus a Turtle at All? YES

• turtle stem reaches the Middle Permian

• broad ribs, lost intercostals, and skull closure began early

• supports a burrowing (land) origin

June 2026: Is Eunotosaurus a Turtle at All? IF MILLERETTID

• turtle-like ribs and skull closure are convergent

• stem sequence starts with Pappochelyx

• turtle origin 255 Ma leaves a Permian ghost lineage

June 2026: Is Eunotosaurus a Turtle at All? what survives

the embryology, the rib-derived carapace, and how living turtles breathe

June 2026: Is Eunotosaurus a Turtle at All? what falls

the permian date for the breathing change and the 40 million year stem extension

Eorhynchochelys - beak and teeth

beak; teeth only at back of jaw

Eorhynchochelys (lower bed) - plastron

no

Eorhynchochelys - ossified carapace

no

Eorhynchochelys - supper temporal opening

closed

Odontochelys - beak and teeth

no beak; teeth along jaw

Odontochelys - plastron

complete

Odontochelys - ossified carapace

no

Odontochelys - upper temporal opening

no reported

crown turtles - beak and teeth

beak; no teeth

crown turtles - plastron

yes

crown turtles - ossified carapace

yes

crown turtles - upper temporal opening

closed

same rock, conflicting traits - the puzzle

Eorhynchochelys branches off earlier in the tree yet has a beak; Odontochelys branches off later yet keeps teeth and lacks a beak

same rock, conflicting traits - either way

at least one trait evolved more than once or was lost and regained

two shells in one - carapace core

• costal plates grow from rib periosteum

• neural plates grow from vertebral neural arches

• “the major part of the carapace is derived purely from endoskeletal ribs”

two shells in one - carapace rim

• nuchal, peripheral, suprapygal, and petal plates: dermal bone

• nuchal = cleithrumi; never lost, even in leatherbacks

two shells in one - plastron

• dermal bone

• epiplastra + endoplastron = clavicles + inter clavicle

• rear elements probably = gastralia (belly ribs)

• alternative: sternal-cartilage homology

Did the Shell Evolve on Land or in Water? - Evidence for water

• stem turtles from Guizhou, China, occur in marine rocks

• 2026 press summary: shell “may have evolved in water from burrowing ancestors”

Did the Shell Evolve on Land or in Water? = Evidence for land

• bone histology of Pappochelys and Norman turtles: non-aquatic

• Pappochelys limb histology read as burrowing (via secondary source)

Did the Shell Evolve on Land or in Water? - Hand length as a habitat proxy

• short hands = land

  • long hands = water

Did the Shell Evolve on Land or in Water? - reanalysis

83% classification accuracy; Proganochelys on land (99.9), palaeochersis mostly on land (95.5)

Did the Shell Evolve on Land or in Water? - mixed signals

odontochelys leans aquatic; proterochersis may have shifted with size

Did the Shell Evolve on Land or in Water? implication

tortoise life on land is a secondary return if the crown ancestor was aquatic

the body plan

1. ribs redirected

2. shoulder enclosed

3. breathing moved

4. blood rerouted

5. skin patterned

6. limbs retuned

body plan core idea

turtles did not add armor to a lizard; they rebuilt the trunk in the embryo

the body plan modular

bone, skin, and muscle parts can change or be lost separately

the body plan - physiology follows

a rigid trunk changed how turtles breathe and how blood reaches the lungs

the carapacial ridge steers the ribs - remove the ridge

ribs lose their path or stop fanning

the carapacial ridge steers the ribs - block FGF

the FGF-blocking drug SU5402 degrades the ridge; ribs deflect into th body wall

the carapacial ridge steers the ribs - limb-bud logic

the FGF loop resembles limb meschyme and apical ectodermal ridge

the carapacial ridge steers the ribs - co-option

genes used elsewhere in development are reused here

the carapacial ridge steers the ribs

carapacial ridge → FGF loop → ribs fan outward

why is the shoulder inside the rib cage? folding model

• ribs undergo “axial arrest”: they stop and grow outward

• late in development the body was folds inward, carrying the scapula

  • some muscle connections kept, some new

why is the shoulder inside the rib cage? ribs-over-scapula

• a scapula in front of the rib cage is the ancestral amniote condition

• the scapula never moves; ribs grow over it

• odontochelys: scapula lies in front of the ribs

why is the shoulder inside the rib cage? both agree

axial arrest the key transition, and rigid carapace came before the shoulder was enclosed

why is the shoulder inside the rib cage? the real dispute

the reference frame

• was the ancestral scapula besides the ribs or in front of them

transversus abdominis action and breathing phase

sling presses viscera against the lungs

• exhalation

obliques abominés action and breathing phase

dome flattens when it contracts, enlarging the body cavity

• inhalation

breathing without moving the ribs - the problem

lose their intercostal muscles in the embryo, so rib-based breathing is impossible

breathing without moving the ribs - solution

an antagonistic pair of addominal muscles

breathing without moving the ribs - not the limbs

breathing is not locked to the stride; the girdles are not th pump

breathing without moving the ribs - beyond the lungs

some breathe through its cloaca on dives of up to 3 days; soft shells exchange gases and excrete urea through the mouth lining

when did the breathing change happen? Lyson 2014

• eunotosaurus ribs lack intercostal attachment fibers on their front edges

• fibers match transverses abdomens origin

• “division of labour”: broad ribs support; abdomen breathes; ribs fuse 50 millions years later

when did the breathing change happen? June 2026 reassignment

• if eunotosaurus is not a turtle, the permian date and the 50 million year figure fall

• its rib evidence becomes an independent case of the same change

when did the breathing change happen? June 2026 preprint

• not peer reviewed

• turtle embryos lose intercostals by apoptosis before carapace forms

• proposes muscle loss comes first and enable axial arrest

turtle heart - structure

two atria and one ventricle divided into three connected compartments (cava)

turtle heart - partial separation

a muscular ridge keep O2-poor and O2-rich streams largely apart

turtle heart - adjustable

how blood splits between lungs and body shifts with vascular resistance

turtle heart - development

the transcription-factor gene Tbx5 becomes restricted to the left side of the ventricle - a partial step toward sepatation

shunting =

steering blood around the lungs

shunting: breathing holding

right-to-left

• vagal nerves constrict the pulmonary circuit

• O2-poor blood bypasses the lungs to the body

• lung blood flow falls; heart slows

shunting: breathing or swimming

left-to-right

• vagal tone is withdrawn

• long blood blow rises

• atropine (blocks vagal effects) raises lung flow

shunting: established

turtles shunt, the direction tracks breathing state, and vagal control of lung resistance drives it

shunting: debated

why - removing the shunt in alligators did not change dive time or metabolism

shunting: link

the three-compartment ventricle makes is possible

what scutes are

keratin shields on the skin, patterned separately from the bone beneath

scutes signaling spots -

bone morphogenetic proteins (Bmp2, Bmp4), Sonic hedgehog (Shh), and Msx2 mark placode-like spots; the BMP blocker Gremlin marks boundaries.

scutes model

a Turing reaction-diffusion system places spots; colliding growth waves set the seams

scutes soft shells

same genes only as a line around the ridge - no placodes, no scutes

scutes temperature

hot, dry nests raise anomalies, mostly in midline

softshells

• lost = scutes (no placodes)

• kept = rib0derived costal plates

• leathery skin covers the bony disc

leatherback shells

• lost = most typical carapace bones

• new = mosaic of small skin bones (osteoderms)

• evolved independently of hard-shelled sea turtles

pancake tortoise shell

• kept = scutes

• reduced = thin, flat, flexible shell with openings in the bone

• edges into rock crevices

evidence that the shell is modular

the skin, dermal-bone, and rib-bone parts of the shell can each be lost separately

lizard trunk

bends

turtle trunk

cannot bend

• shell lacks the backbone; no trunk wave to swing the shoulders

trunk - most sprawling tetrapods

bend the trunk side to side as they walk; each bend swings a shoulder forward and lengthens the step

turtles trunk

the shell fuses the ribs and trunk vertebrae, so the trunk cannot bend at all

trunks old prediction

the shell should also hold the girdles still, leaving the limbs to do all the work

shoulder girdle - three prongs

scapula, acromion, and coracoid meet at the shoulder joint

shoulder girdle - scapula

runs up to the underside of the carapace, so from below it looks short