Madtsoia

Madtsoia is a genus of large extinct snakes known mainly from vertebrae and ribs. Its traditional fossil range links Late Cretaceous Madagascar and India with Paleogene South America, but named species are not always recovered together in modern phylogenetic analyses. The type species, M. bai, is unusually informative because it includes long associated vertebral series; other species are known from different, often isolated elements. No skull is known for the genus, so its head shape and jaw mechanics remain uncertain. The fossils and comparisons are presented in the ancient lizard and snake catalogue with other Gondwanan snakes such as Najash.

Quick facts

The type species and its long vertebral series

George Gaylord Simpson named Madtsoia bai in 1933 from the Notostylops beds of Chubut Province, Argentina. The name-bearing material, AMNH 3154, includes two associated series containing roughly 45 trunk vertebrae with ribs, plus several additional vertebrae. Modern stratigraphic interpretation places the locality in the lower Eocene of the San Jorge Basin.

A long sequence is especially useful for fossil snakes. Vertebrae change shape from the front of the trunk toward the tail, so a series from one individual reveals that within-body variation. If isolated bones are compared without knowing their position, a front-to-rear difference can be mistaken for a difference between species. The type material therefore provides a reference for interpreting less complete samples.

Simpson derived the genus name from Tehuelche words associated with valley and cow, referring to the locality rather than the snake’s appearance. The species name bai also reflects a local geographic name. Etymology does not indicate body form, diet or behaviour.

The Maastrichtian Madagascar sample

Jean Piveteau reported a large snake vertebra from the Marovoay region in 1933 but did not name a species. Robert Hoffstetter established Madtsoia madagascariensis in 1961 from holotype MNHN MAJ 5, a posterior trunk vertebra from Gîte du Guide near Berivotra. Later work by the Mahajanga Basin Project added dozens of bones from the Maevarano Formation.

The expanded sample includes anterior, middle and posterior trunk vertebrae, cloacal elements, tail vertebrae and rib fragments. All are from Maastrichtian deposits of the Late Cretaceous. This range along the spine makes it possible to describe regional anatomical differences instead of treating every isolated vertebra as if it came from the same position.

The vertebrae carry a combination of features used to identify madtsoiids: parazygantral foramina beside the zygantral facets, paracotylar foramina, broad rib articulations, hypapophyses limited to the anterior trunk, and a haemal keel on middle and posterior vertebrae. Some bones also have short paired projections at the rear of the keel. The diagnosis depends on the combination, not one feature alone.

More species, a wider map and an unsettled genus

The name Madtsoia camposi was applied to Paleocene Brazilian material that includes numerous vertebrae, ribs and some jaw elements. The 2010 revision of the Madagascar snake assemblage retained M. bai, M. camposi and M. madagascariensis within the genus. This historical arrangement is useful to describe the literature, but is not the same as demonstrating that every species forms one natural branch.

In 2011, Dhananjay Mohabey, Jason Head and Jeffrey Wilson named M. pisdurensis from large vertebrae in the Maastrichtian infratrappean beds of central India. The deposits lie beneath the Deccan lava sequence. A low, flattened triangular projection on the haemal keel was among the features described for that species. The authors used vertebrae reaching about 1.83 centimetres long and 4.35 centimetres high to estimate body length at approximately five metres.

Later phylogenetic analyses that sampled madtsoiids more broadly do not always recover all traditional Madtsoia species together. In some trees, the Indian form is closer to the African Gigantophis than to the Argentine type species. The named species and their map therefore represent a history of classification, not a settled family tree. Similar large vertebrae can preserve genuine relationship, but incompleteness can also blur the diagnosis.

How paleontologists read the vertebrae

Madtsoia vertebrae are robust and relatively short, with the paired zygosphene and zygantrum joints that help stabilise the snake’s axial column. In many specimens, one parazygantral opening lies in a fossa on each side of the zygantral facet. Wide rib articulations and the haemal keel add further characters for comparison.

Researchers assess the neural spine’s height and angle, the shape of the cotyle and condyle, the depth of lateral fossae, and the size of the rib processes. These traits vary along one individual’s spine and may also change as an animal grows. A diagnosis must compare homologous vertebral positions and account for age, breakage and distortion.

This is why the associated series of M. bai is valuable and why claims based on a single vertebra can remain tentative. The delicate margins that carry a diagnostic ridge may be broken away. A fossil can still be attributable with a combination of preserved characters, but the confidence depends on how much anatomy remains available.

Size estimates and missing anatomy

Madtsoia is commonly illustrated as a giant snake, yet a reliable total length is not available for most named species. Large trunk vertebrae show that the animals were substantial, but converting one bone into a body length requires a model derived from the proportions of living snakes. That model may not fit an extinct lineage with a different number or shape of vertebrae.

The approximately five-metre estimate for M. pisdurensis is the result of such a comparative calculation using measured vertebrae. It is not the measured length of a complete skeleton. Much larger estimates circulated for other species, but those depend on less secure assumptions about body proportions and should not be presented as established measurements.

No Madtsoia skull is known. Consequently, researchers cannot directly assess the shape of the head, the width of the gape or the degree of jaw mobility. Vertebrae and ribs support a long, limbless body; they do not yield a detailed portrait of the face or a precise body outline.

Gondwanan geography and ecological limits

Late Cretaceous records traditionally assigned to Madtsoia occur in Madagascar and India, while South American species are Paleocene or Eocene. These regions had separated as Gondwana fragmented. Mohabey and colleagues proposed either an older dispersal history before final separation or later exchange routes that remain poorly documented. The fossils do not decide between those scenarios on their own.

The Malagasy species lived within the seasonal river systems represented by the Maevarano Formation. The same broad deposits preserve dinosaurs, crocodyliforms, mammals and the snake Sanajeh. This faunal association establishes that those animals occur in the same geological setting; it does not document a Madtsoia attack on a particular dinosaur or mammal.

Madtsoia was almost certainly predatory, as expected for a snake, but no stomach contents, coprolite or prey fossil with a diagnostic bite has been tied to the genus. Powerful vertebrae are compatible with strong body bends and restraint of prey, yet they do not directly demonstrate constriction. A fossorial or aquatic lifestyle also cannot be established from vertebrae alone.

What a reconstruction can responsibly show

The fossils directly show robust vertebrae, associated ribs in some species, and in a few cases long connected portions of the spinal column. A long, limbless body with lateral bending is a well-supported general reconstruction. Rib dimensions and joint surfaces can help estimate trunk thickness, but the result remains dependent on comparison with other snakes.

The skull, eye placement, scale pattern and colour are not preserved. A full-body illustration therefore combines known axial anatomy with assumptions drawn from living snakes. The animal should not be shown with a particular head shape or constricting prey as if those features had been found in the fossil.

The most useful interpretation keeps three scales of certainty separate: vertebral anatomy is directly observed; body length is modelled for some species; and the genus-wide range and evolutionary unity remain debated. Future skulls or additional associated skeletons could substantially change that picture.