Saurolophus (/sɔːˈrɒləfəs/; meaning "lizard crest") is a genus of large hadrosaurid dinosaur from the Late Cretaceous period of Asia and North America, that lived in what is now the Horseshoe Canyon and Nemegt formations about 70 million to 66 million years ago. It is one of the few genera of dinosaurs known from multiple continents. The type species, S. osborni, was described by Barnum Brown in 1912 from Canadian fossils. A second valid species, S. angustirostris, is represented by numerous specimens from Mongolia, and was described by Anatoly Konstantinovich Rozhdestvensky.

Saurolophus is distinguished by a spike-like crest which projects up and back from the skull. It was a herbivorous dinosaur which could move about either bipedally or quadrupedally.

Barnum Brown recovered the first described remains of Saurolophus in 1911, including a nearly complete skeleton (AMNH 5220). Now on display in the American Museum of Natural History, this skeleton was the first nearly complete dinosaur skeleton from Canada. It was found in rocks of early Maastrichtian age, in the Upper Cretaceous Horseshoe Canyon Formation (then known as the Edmonton Formation) near Tolman Ferry on the Red Deer River in Alberta. Brown wasted little time in describing his material, giving it its own subfamily. Saurolophus was an important early reference for other hadrosaurs, as seen in the names of Prosaurolophus ("before Saurolophus") and Parasaurolophus ("near Saurolophus"). However, little additional material has been recovered and described.

Instead, more abundant remains from Asia have provided more data. Initial remains were not promising; a partial fragmentary ischium from Heilongjiang, China, that Riabinin named S. kryschtofovici. Much better remains were soon recovered, though, but from Mongolia's early Maastrichtian-age Nemegt Formation. The 1946–1949 Russian-Mongolian paleontological expeditions recovered the large skeleton that became S. angustirostris as described by Anatoly Rozhdestvensky. Other skeletons from a variety of growth stages have also been discovered, and S. angustirostris is now the most abundant Asian hadrosaurid.

Two species are regarded as valid today: the type species S. osborni, and S. angustirostris. S. osborni (Brown, 1912) is known from a skull and skeleton, two other complete skulls, and skull fragments. S. angustirostris (Rozhdestvensky, 1952) is known from at least 15 specimens. It differs from S. osborni by some details of the skull, as well as in the pattern of scales found in skin impressions. The Mongolian species had a longer skull (by 20%) and the front of the snout (the premaxillary bones) were more upwardly directed. S. angustirostris also had a distinctive row of rectangular scales along the midline of the back and tail, known as 'midline feature-scales'; these are not currently preserved in S. osborni. In S. angustirostris, the scales on the tail flank were arranged in vertical patterns, which may have corresponded to striped coloration in life. This area was covered in radial scale patterns in S. osborni, possibly indicating a more mottled or spotted coloration.

S. kryschtofovici (Riabinin, 1930) is not considered valid; either it is regarded as a dubious name, or as a synonym of S. angustirostris (although the name antedates S. angustirostris).

Until a 2011 reevaluation of the species by Phil R. Bell, S. angustirostris was not well-described. No autapomorphies, unique derived traits, had been established distinguishing it from S. osborni. Bell found in a publication earlier in the year that the two previous studies of S. angustirostris, by Rozhdestvensky in 1952, and Maryanska and Osmolska in 1981, do not provide a comprehensive enough description to compare the species with S. osborni.

In 1939–40, two partial skeletons were found in the late Maastrichtian age Moreno Formation of California. These specimens were referred to cf. Saurolophus sp. In 2010, one of the skulls was instead assigned to Edmontosaurus. A 2013 study placed the two specimens in a new species, S. morrisi. In 2014, the species was reassigned to a new genus, Augustynolophus.

Saurolophus is known from material including nearly complete skeletons, giving researchers a clear picture of its bony anatomy. S. osborni, the rarer Albertan species, was around 8.2–8.5 m (27–28 ft) long, with its skull 1.0 m (3.3 ft) long. It has been estimated to have weighed around 3 tonnes (3.0 long tons; 3.3 short tons). S. angustirostris, the Mongolian species, was larger; it got as large as 13 m (43 ft) in length, and larger remains are reported. It has been estimated to have potentially weighed up to 11 tonnes (11 long tons; 12 short tons). The largest known skull of S.angustirostris measures 1.22 m (4.0 ft) in length. Aside from size, the two species are virtually identical, with differentiation hindered by lack of study.

The most distinctive feature of Saurolophus is its cranial crest, which is present in young individuals, but is smaller. It is long and spike-like and projects upward and backward at about a 45° angle, starting from over the eyes. This crest is often described as solid, but appears to be solid only at the point, with internal chambers that may have had a respiratory and/or heat-regulation function. The unique crest of Saurolophus is made up almost completely by the nasal bones, and in S. angustirostris it is solid. In adult specimens the crests are a rounded triangular shape in cross section. The crest protrudes past the edge of the skull backwards. Thin processes from the frontals and prefrontals extend along the underside of the crest, probably to strengthen it. At the end of the crest is a swelling of the nasal, which is often termed differently.

The holotype of S. angustirostris is a skull and postcrania, so the cranium of the species is well-described. Bell et al. re-evaluated the entire species in a 2011 publication with Acta Palaeontologica Polonica. Their description found the skull to be generalized among hadrosaurines, and are much larger than any skulls of S. osborni. The most unusual feature for a hadrosaurine is the long, protruding, solid crest that extends upwards diagonally from the back of the skull roof. Unlike lambeosaurines, the crests are made up completely of the nasal bone. The premaxilla bones make up almost 50% of the entire skull length, and both sides are filled with small holes. Only in adult individuals has the front of the premaxillary contact been fused. Longer than the premaxilla, the nasal bones are the longest in the skull. They make up the entire length of the crest, and are never preserved as fused.

Barnum Brown, who described the first specimens, put it in its own subfamily in "Trachodontidae" (=Hadrosauridae), the Saurolophinae. At the time, this also included Corythosaurus and Hypacrosaurus, the only well-known examples of what would become the Lambeosaurinae. Brown thought that Saurolophus had an expanded tip to the ischium bone in the hip, as dinosaurs now recognized as lambeosaurines had, but this appears to have been based on a mistakenly associated lambeosaurine ischium. Additionally, he misinterpreted the crests of Saurolophus and lambeosaurines as being made of the same bones.

Most publications before 2010 classified Saurolophus as a member of Hadrosaurinae, often known colloquially as the "flat-headed hadrosaurs". In 2010, the subfamily Saurolophinae was brought back into use because Hadrosaurus appears to have branched off prior to the "hadrosaurine"–lambeosaurine split. As a result, Hadrosaurinae by definition cannot include the traditional "hadrosaurines". Saurolophinae is the oldest available name for the former "hadrosaurine" clade. Saurolophus, as the name suggests, is a saurolophine, as it has a saurolophine pelvis and a (largely) solid crest.

The following cladogram of hadrosaurid relationships was published in 2013 by Alberto Prieto-Márquez et al. in Acta Palaeontologica Polonica:

As a hadrosaurid, Saurolophus would have been a bipedal/quadrupedal herbivore, eating a variety of plants. Its skull permitted a grinding motion analogous to chewing, and its teeth were continually replacing and packed into dental batteries that contained hundreds of teeth, only a relative handful of which were in use at any time. Plant material would have been cropped by its broad beak, and held in the jaws by a cheek-like organ. Its feeding range would have extended from the ground to about 4 m (13 ft) above.

The distinctive spike-like crest of Saurolophus has been interpreted in multiple ways, and could have had multiple functions. Brown compared it to the crest of a chameleon, and suggested it could provide an area for muscle attachment and a connection point for a nonbody back frill like that seen in the basilisk lizard. Peter Dodson interpreted similar features in other duckbills as having use in sexual identification. Maryańska and Osmólska, noting the hollow base, suggested that the crest increased the surface area of the respiratory cavity, and helped in thermoregulation. James Hopson supported a function as a visual signal, and further mentioned the possibility that the inflatable skin flaps over the nostrils could have acted as resonators and additional visual signals. This idea has been picked up by authors of popular dinosaur works, such as David B. Norman, who discussed hadrosaurid display at length and included a life restoration of such an adaptation in action.