Did Dinosaurs Swim? Scratch Marks on an Ancient River Bed

The traces differ from ordinary footprints in specific ways. They consist of scratches or grooves rather than impressions of a whole foot, since only the tips of the claws contacted the sediment. They are frequently arranged in sets of two or three parallel lines, matching the digits of a theropod foot. They may be unevenly spaced, longer than a normal stride and irregular in direction, which is what would be expected if the animal was being carried by a current while intermittently touching the bottom. Some sequences show the marks becoming shallower and then disappearing, which has been read as the animal lifting clear as the water deepened. They occur in sediments deposited in water, which is consistent with the interpretation and does not establish it.

Several alternative explanations produce comparable marks:

• •An animal walking on a very soft substrate where the foot sinks and only the claws register on a firmer layer below
• •An animal walking through shallow water without floating, where buoyancy reduces the load without supporting the animal
• •Marks produced by the feet of an animal swimming and by different behaviour on the same surface being conflated
• •Scratches made by objects dragged by current rather than by any animal
• •Undertracks, where a mark registers on a layer beneath the surface the animal walked on and loses its outline
• •Sediment deformation after the mark was made, which alters shape and spacing

Independently of these particular traces, the general question has a reasonable answer. Almost all vertebrates can swim, including animals with no aquatic adaptation, since a limbed body with lungs floats and paddles adequately, and living birds, crocodilians and large mammals with no swimming specialisation cross water routinely. There is no reason to think dinosaurs could not, and the interesting question is not whether they were physically capable but how often they did it and whether any group made a living in water. Distributional evidence bears on that, since dinosaur remains occur on islands that were separated by water at the time, which implies crossings happened. The debate about the traces is therefore about what a particular set of marks records rather than about whether swimming occurred at all.

Fossil surfaces record several other kinds of contact with water that are worth distinguishing. Wading traces show full footprints with distortion from water moving around the foot, which differs from marks made in air. Impressions made in sediment beneath standing water have characteristically softer margins, since the material was saturated. Rippled surfaces record current direction and water depth independently of any animal and supply the setting a trace has to be interpreted within. Raindrop impressions and desiccation cracks establish that a surface was exposed rather than submerged, which excludes a swimming interpretation for marks on the same bed. Reading a trace therefore begins with reading the rock it sits in, and a claim about behaviour that ignores the sedimentology is weaker than one built on it, which is the main methodological point in this whole area.

A stronger claim concerns whether any dinosaur was genuinely adapted to water, and that has focused on one large predatory group with a long crocodile-like snout, conical teeth, dense bones and, in one much-publicised reconstruction, a paddle-like tail. Papers arguing for an aquatic or semi-aquatic lifestyle have been followed by papers disputing the biomechanical modelling, the bone density comparison and the tail reconstruction, with the exchange continuing across several rounds in prominent journals. The current position is that a fish-based diet is well supported by tooth chemistry and by preserved gut contents, that shoreline foraging is widely accepted, and that whether the animal pursued prey underwater is not settled. The episode is a good illustration of a disputed question being argued with quantitative evidence rather than assertion, which does not guarantee it resolves quickly.

Parallel claw scratches rather than full footprints have been read as a floating animal pushing off the bottom, and soft substrate, shallow wading and undertracks produce comparable marks. Almost any limbed vertebrate can swim, so the open question is how often rather than whether. Whether one large predatory group pursued prey underwater is argued with quantitative evidence and remains unsettled.