Ceratopsid dinosaurs changed considerably during growth, which caused early palaeontologists to name different growth stages of the same animal as different species, or even genera. The Maastrichtian (Latest Cretaceous, 70.6-65.5 million years ago) Ceratopsids of North America were at one point divided into 22 species in 10 genera, but as palaeontologists have uncovered intermediate forms this has been consolidated into three species in two genera, Torosaurus and Triceratops.
In 2010 John Scannella and Jack Horner of Montana State University published a paper in the Journal of Vertebrate Paleontology in which they propose that the small-frilled Triceratops is an immature form of the large-frilled Torosaurus, so that there should only be a single genus, Triceratops (when species or genera are combined the older name takes preference, in this case Triceratops which was first named in 1889, two years ahead of Torosaurus in 1891).
In a paper published in the journal PLoS ONE on 29 February 2012, Nicholas Longrich and Daniel Field of the Department of Geology and Geophysics at Yale University discuss the results of a study of both the physical and chronological distribution of Torosaurus and Triceratops, and a study of a large number of individual specimens to look for intermediate characters.
The skulls of (A) Triceratops and (B) Torosaurus. (B) has a larger crest than (A), with two promanent openings, but this might be a later developmental stage. From Longrich and Field (2012).
Longrich and Field found that Triceratops and Torosaurus did overlap closely in both chronological and physical distribution, with the majority of locations that produced fossils of one also producing fossils of the other, and those sights which produced only one or the other being otherwise fossil-poor. This leaves open the possibility that the two dinosaurs might be members of the same species.
The distribution of Triceratops and Torosaurus. (1) Scollard Formation, Alberta; (2) Frenchman Formation, Saskatchewan, (3) Hell Creek Formation, Montana; (4) Hell Creek Formation, North Dakota; (5) Hell Creek Formation, South Dakota; (6) Lance Formation, Wyoming; (7) Denver Formation, Colorado; (8) North Horn Formation, Utah; (9) Javelina Formation, Texas. From Longrich and Field (2012).
After this Longrich and Field looked at the skulls of a large number of Triceratops and Torosaurus specimens, examining both the morphological development and the age based upon bone tissue development. They found that a full range of ages were present in both dinosaurs, strongly implying that they were not the same species.
Scanella and Horner had reported depressions in the crests of some Triceratops, which they argued were analogous to the gaps in the crest of Torosaurus, but Longrich and Field argue that these do not form on the same bones, and therefore cannot develop from one to another.
The frills of (A) Triceratops and (B) Torosaurus; fen, fenestra; fos, parietal-squamosal fossa; par, parietal; sq, squamosal. P0, midline epiparietal; p1–p5 epiparietals 1–5; eps, epoccipital crossing the parietal-squamosal suture. From Longrich and Field (2012).
They also found that two forms of Torosaurus previously thought to be different developmental stages did not appear to be sequential when assessed by tissue development, the larger, apparently more developed, form often (but not always) being younger than the smaller, less developed form. Longrich and Field examined several possibilities for this. Firstly our understanding of bone tissue development in Ceratopsid Dinosaurs might be inaccurate. This would have implications for the wider study of the group, so Longrich and Field suggest this needs to be investigated further. Then there is the possibility that development in Torosaurus was variable, either between populations or in response to climatic variations, so that individuals of the same age might be at different developmental stages. Finally they examined the possibility that Torosaurus might be sexually dimorphic, so that the different forms could represent the two sexes of a single species.
The two morphotypes of Torosaurus. B was formerly thought to grow into A, but tissue development suggests A is younger than B. From Longrich and Field (2012).
For some reason Longrich and Field did not discuss the possibility that the two morphotypes of Torosaurus might tie in with the two 'species' of Triceratops, to give two sexually dimorphic species, each with a large and small crested form, though it seems unlikely that this did not occur to them.
See also New dinosaurs from old draws, Nesting behavior and parental care in an Ornithischian Dinosaur? and Dinosaurs on Sciency Thoughts YouTube.