How Did Snakes Lose Their Legs? Fossils, Embryos and a Single Switch

Several extinct snakes preserve hind limbs, and they come mostly from marine deposits of the late Cretaceous around the eastern Mediterranean. Pachyrhachis, described from Israel, has a clear pelvis and small but complete hind legs with a femur, tibia, fibula and foot bones, attached to a body that is otherwise a snake with more than a hundred and fifty vertebrae. Haasiophis and Eupodophis are similar. The most significant find came from Brazil in 2015: Tetrapodophis, a fossil from the Crato Formation described as having four limbs, though the interpretation has been disputed and some researchers argue it is a limbed lizard rather than a snake. What all of these have in common is that the forelimbs are already gone or vestigial while the hind limbs persist, which establishes the order: front legs first, back legs later, and the hind limb girdle lingering in some living families to this day.

The argument about where the body plan evolved has run for more than a century. The marine hypothesis points to the fossils, since Pachyrhachis and its relatives are found in marine sediments alongside mosasaurs, and proposes that the elongated limbless body was an adaptation for swimming, in the way that eels and sea snakes are shaped. The burrowing hypothesis points to anatomy, since the earliest-branching living snakes are small burrowing forms, and since the snake eye is rebuilt in a way that looks like the recovery of vision by an animal whose ancestors had lost it underground: the lens, the retina and the focusing mechanism all differ from those of lizards in ways best explained by degeneration followed by redevelopment. Evidence has moved toward burrowing. A 2015 reconstruction of the ancestral snake, based on a large anatomical dataset, described a nocturnal, ground-dwelling forest hunter with hind limbs still present, and the marine fossils are now generally read as a specialised offshoot rather than the main line.

The developmental story turned out to be unexpectedly simple, and it rests on two mechanisms:

• •Hox genes, which set out the body's regions along its length, are expressed differently in snakes, extending the thoracic identity along almost the entire trunk so that rib-bearing vertebrae are produced by the hundred instead of stopping at a shoulder and a hip
• •A regulatory sequence called ZRS, which controls where and how long the gene Sonic hedgehog is active in a developing limb bud, carries deletions in snakes that cripple it
• •In 2016 two teams tested this directly: the snake version of the sequence was put into mice, and the mice grew severely truncated limbs, while repairing the deleted portion in a snake sequence partly restored function
• •Python embryos do begin to form hind limb buds, which start developing and then stall and regress, so the machinery is still there and switched on briefly
• •The forelimb loss has a different cause, connected to the Hox changes that eliminate the neck and shoulder region entirely, which is why no snake has any trace of a front limb

Losing limbs is not obviously an improvement, and the group's success suggests it was. Snakes move by at least four distinct methods, including the lateral undulation everyone pictures, rectilinear creeping in which the belly scales pull the animal forward in a straight line, concertina movement for narrow spaces, and sidewinding on loose sand, which collectively give access to burrows, crevices, water, leaf litter and tree canopy. The elongated body permitted a rearrangement of the organs, with one lung reduced or lost and the others strung out in series, and with a highly flexible skull whose bones are loosely connected, allowing prey far wider than the head to be swallowed whole; this is done by walking the jaws forward alternately rather than by unhinging anything, which is a persistent myth. Venom evolved in this context as a way for an animal with no limbs to subdue prey it cannot hold. There are around four thousand species today on every continent except Antarctica, which is not the record of a compromised design.

Limb loss is not rare and not confined to snakes. It has happened independently at least twenty-five times among squamate reptiles alone, producing slow worms, legless skinks, amphisbaenians and several others that are lizards rather than snakes and are frequently mistaken for them; the test is usually eyelids and external ear openings, which snakes lack and most legless lizards retain. It has happened in amphibians, in the burrowing caecilians, and it is approached in whales, which retain internal pelvic bones with no external hind limbs at all. The repetition is the useful part, because a change that occurs twenty-five times independently in one group is not a freak event but a predictable response to a particular way of life, and the fact that it has arisen so often through the same developmental switches suggests the number of ways to build a long body without legs is small.

Pythons and boas retain a vestigial pelvis with external spurs, and Cretaceous fossils including Pachyrhachis preserve complete hind limbs on otherwise snake-like bodies, showing forelimbs were lost first. The ancestor is now generally reconstructed as a burrowing forest animal rather than a swimmer, partly because the snake eye looks rebuilt after degeneration underground. Hox genes extend rib-bearing vertebrae along the whole trunk, and deletions in a regulatory sequence controlling limb development stall the hind limb buds that python embryos still begin to grow.