Notes on Ceratopsians and Ankylosaurs at the Royal Ontario Museum

Transcription

1 Notes on Ceratopsians and Ankylosaurs at the Royal Ontario Museum Andrew A. Farke, Ph.D. Raymond M. Alf Museum of Paleontology 1175 West Baseline Road Claremont, CA Introduction and Overview As part of my dissertation research and general research program focused on horned dinosaurs, I traveled to the Royal Ontario Museum (ROM, Toronto, Ontario, Canada) in the summer of 2006 for a few days of research. The notes were typed in Microsoft Word, and are presented in unedited format on the following pages. Taxa discussed in these notes include specimens referred to Centrosaurinae, Centrosaurus, Chasmosaurus, Triceratops, Anchiceratops, and the ankylosaur Panoplosaurus. Note that the taxonomic identifications in the notes may be tentative or outdated in light of new work. In most cases, I simply used the identification in the museum databases (which may or may not match with identifications published in Farke, 2010). The notes on Panoplosaurus are not connected to any published project, and my opinion on ornithischian cheeks has changed subsequently. I now am convinced that they did have cheeks. Disclaimer The sketches, measurements, and notes are all my personal work and interpretations. These data are thus presented as is, and users should be appropriately cautious when relying upon my notes for their own research. Although I strive to be accurate, inadvertent mistakes or inaccuracies are possible. In the end, there is no substitute for personal examination of a specimen in order to answer some questions. Acknowledgments Thank you to Kevin Seymour and David Evans (ROM) for facilitating access to the specimens. Portions of the research were funded by the Jurassic Foundation, Paleontological Society, and a National Science Foundation Graduate Research Fellowship. Key Citations Farke AA (2010) Evolution, homology, and function of the supracranial sinuses in ceratopsian dinosaurs. J Vertebr Paleontol 30: How to Cite This Some aspects of these notes were incorporated into a previous publication (Farke, 2010), and you may wish to cite those if the data were included there. Otherwise, I do request that you cite these notes through reference to the appropriate page on figshare.com, or another appropriate link if figshare.com is not available at some point in the future. License: This work is licensed under the Creative Commons Attribution 3.0 Unported License. To view a copy of this license, visit or send a letter to Creative Commons, 444 Castro Street, Suite 900, Mountain View, California, 94041, USA. This version of the document was assembled and uploaded on May 14, 2012

2 Royal Ontario Museum - Toronto Tuesday, August 22, 2006 Friday, August 25, 2006 Arrived on Monday, 21 August, and stayed night at Victoria University. Arrived at museum on Tuesday morning, at 10 a.m. Met Kevin Seymour, collections manager, and was given collections tour. Spent morning taking pathology data in main collections room. Finished out morning with beginning of sinus work on some selected specimens. ROM 3521 Centrosaurinae, anterior parietal portion Oldman Formation, Upper Cretaceous. Little Sand Creek, Badlands, Alberta Accession number A137. Specimen consists of apparently surface-collected fragment of anterior parietal. Little or no reconstruction is present, aside from some plaster patches in cracks. As an aside, a small circular resorption pit, measuring 5.31 x 5.47 mm in breadth and 3.16 mm deep is just to the left of the midline, on the textured portion of the dorsal surface of the parietal. Moderately vascularized texture on dorsal surface of parietal indicates that the individual was probably an adult. The very posterior portion of the frontal sinus is preserved. Unfortunately, there is breakage in the area that would indicate whether or not the frontal sinus is continuous with the supratemporal fenestrae in this specimen. The preserved portion of the frontal sinus is divided into two portions; an upper portion on the dorsal surface of the parietal and a lower portion recessed into the base of the parietal. I cannot tell if the upper portion connected with the supratemporal fenestrae, due to the breakage in the necessary area. There is a right and left upper sinus, separated by a small tab of bone from the vascularized dorsal parietal. The right side is slightly better preserved, so I will base my description on that. The right upper sinus is recessed from the main dorsal surface of the parietal by about 2.63 mm. The bone in this area is perfectly smooth. It measures at least mm wide (medio-laterally) by mm long (antero-posteriorly). The upper sinus sits on top of a thick projection of bone which overhangs the supratemporal fenestra (the fenestra undercuts this bone projection). Anteriorly, a smoothly rounded lip of bone marks the descent into the lower frontal sinus. The lip is a little more sharply defined on the right than on the left side, but this may be due to preservation (a small crack filled with plaster for a part of the right side). The lower frontal sinus is at least mm tall and mm wide, as preserved. It descends directly ventrally; no portion of the floor of the sinus is preserved, however. The lateral, entire anterior, and ventral margins of the sinus are not preserved. The bone lining the lower frontal sinus is mostly smooth, without vascularization grooves. Laterally, it is separated from the supratemporal fenestrae by a [broken] very thin piece of bone.

3 The entire parietal portion measures mm long by mm wide, as preserved. The specimen probably includes the portion heading down to the exoccipital. I am guessing that only parietal is preserved here, because if any other bones are present, the sutures are entirely obscured. ROM 839 Chasmosaurus brevirostris type Skull is extremely well-preserved, except the supratemporal fenestra region and parts of the area leading up to the frontal fontanelle are not all that well prepared out. The left supratemporal fenestra is prepared better and is in better condition, so I will rely primarily upon that. Frontal fontanelle The frontal fontanelle is oval in shape, compressed slightly mediolaterally. The fontanelle itself measures mm long by mm wide, as preserved. It was at least mm deep, but the bottom is obscured by matrix and plaster. The left side is pushed up relative to the right side. The border between the dorsal surface of the skull and the margin of the frontal fontanelle is quite abrupt. Anteriorly, it is obscured by plaster. Laterally, it drops off quite sharply, and even overhangs the sinus somewhat. Due to crushing, I unfortunately can t tell a whole lot about the posterior aspect of the sinus. Two possibilities exist as to its morphology: 1) The posterior border of the frontal fontanelle is a sheet of bone that overhangs the entire sinus. In this interpretation, it is just crushed down into the posterior portion of the sinus. I think this is most likely, given the rough bone texture on this sheet of bone (probably vascularized). Unfortunately, the tunnel that should exist is covered by plaster. 2) There is no posterior border of the frontal fontanelle, but a shallow channel of bone that connects the anteriormost part of the frontal sinus to the anterior chamber of the posterior part of the frontal sinus (terminology following that used above). I don t think this as likely, because of the apparently vascularized texture on the floor of this midline channel. I favor interpretation one outlined above, for the given reasons. So, this implies that the frontal fontanelle, if originally open, was partially roofed over in this specimen, at least at the posterior portion. A tunnel in the bone would connect the anterior frontal sinus (open to the frontal fontanelle) to the posterior frontal sinus. Posterior frontal sinus Again, based primarily on the left side. The right side has a decent bit of plaster, so I don t trust the morphology. The upper portion of the posterior sinus is quite well-preserved. It is clearly continuous with the supratemporal fenestra. On the midline, a tab of the parietal projects over the frontal sinus. It extends approximately 9.59 mm over the sinus. The left channel is recessed 6.94 mm below this central tab. Laterally, it is recessed much more (17.27 mm), but at least some of this may be accentuated by crushing. The floor of the channel is smooth and avascular. The channel was at least mm wide, but reconstruction laterally means I can t say much more about the morphology. Plaster obscures the lateral portion of the sinus. The upper part of the sinus

4 continues for at least mm, and then a smoothly arced lip dives toward the lower sinus. The lower sinus is at least mm below the tab on the dorsal surface of the parietal, but the lower portion of the sinus is still obscured by matrix. Argh! The bone flooring the upper part of the sinus is undercut by the supratemporal fenestra. Thus, adductor muscles would have coursed underneath this bone during life. That s about all I have to write about the frontal sinus! It s a shame it s not prepared out more. Measurements of ROM 839 Bilateral measurements given as right / left 1: Minimum width of nasals above external naris : Greatest width across orbits : Minimum medio-lateral width of postorbital horn at base, immediately above orbital rim / : Minimum rostro-caudal length of postorbital horn at base, immediately above orbital rim / : Greatest width across jugals, including epijugal : Least distance from rostral border of orbit to tip of rostral / : Distance from tip of rostral to caudal end of maxillary tooth row / : Minimum dorso-ventral depth from dorsal surface of nasals to ventral surface of maxilla / : Distance from tip of rostral to end of occipital condyle (basal skull length) 746 mm 10: Minimum distance from caudal end of external naris to rostral end of orbit / : Length of postorbital horncore from tip of horn to dorsal margin of orbit / : Length of jugal from ventral margin of orbit to ventral limit of jugal / : Depth of skull, from rostral limit of supratemporal fenestra to distal tip of epijugal / Note: The rostral bone is reconstructed (probably pretty realistically, based on its articulation with the tip of the predentary). So, measurements involving the rostral are estimated, but are off at most by mm. Anterior depression on premaxilla is not perforate, on left side at least (which is better preserved than right). Only a small bit of plaster (~5 mm across) bridging damaged area, so I think it s safe to say it s solid in life. On Thursday, I went to take another look at the skull. I wanted to see if there is any clear pathway where paratympanic pneumaticity could have gone up into the supratemporal region. Taking a look at the otic region, the stapes would have to do something sorta funny and long. It would start out posterior to the quadrate, and then pass between the pterygoid and exoccipital in order to make it to the oval window on the side of the braincase (which is anterior to the exoccipital). The sketch on the Lull 1903 Triceratops in this document shows the possible path. This would imply that the stapes would pass right next to the jaw adductor musculature. Assuming I ve pegged this correctly, this means that it would be possible for paratympanic pneumaticity to make it up to the supratemporal region and run with the adductor musculature.

5 This wouldn t be hugely likely though, due to the pressure changes in the middle ear every time the animal chewed! I would then predict you might see pneumaticity on lateral surface of pterygoid or on braincase in that region. But...no real evidence that I can see! Red arrow indicates path of stapes on its way to oval window.

6 Shortest distance from posterior surface of quadrate to opening for CN XIII is about 162 mm (left estimate, some reconstruction) or 157 mm (right side, actual measurement). The minimum volume of the frontal sinus, as measured with sand, is 120 ml. It may have been up to twice this. Sacriceratops Triceratops from BHI, sold to ROM recently On examination, I realized that this is the specimen I have cast in my office. Yay! This means I can now use the casts for research. Coded the specimen for pathology Also, I measured the specimen, for potential use in the morphometric analysis. The frontal fontanelle is reconstructed over, so there s not much I can see there. Also, there is a bit of reconstruction is portions of the frontal sinus preserved. So, the casts I have at home in SD probably correspond pretty well with the actual specimen as it is reconstructed/mounted at present. Not probably worth trying to see the sinuses in their current state, as I have them on the casts at home and because it is very tough to get to. As a caveat, the specimen was originally disarticulated. It looks to be a good reconstruction, but it means that some medio-lateral measurements may be off a touch. 1: Minimum width of nasals above external naris : Greatest width across orbits : Minimum medio-lateral width of postorbital horn at base, immediately above orbital rim / : Minimum rostro-caudal length of postorbital horn at base, immediately above orbital rim / : Greatest width across jugals, including epijugal : Least distance from rostral border of orbit to tip of rostral 731 / 739 7: Distance from tip of rostral to caudal end of maxillary tooth row 759 / 744 8: Minimum dorso-ventral depth from dorsal surface of nasals to ventral surface of maxilla / : Distance from tip of rostral to end of occipital condyle (basal skull length) 1042 as reconstructed - I think it was probably about 800 or 900 in actuality, as braincase as reconstructed seems to be set too far back (should be flush with end of nasals at front end, more or less) 10: Minimum distance from caudal end of external naris to rostral end of orbit / : Length of postorbital horncore from tip of horn to dorsal margin of orbit / : Length of jugal from ventral margin of orbit to ventral limit of jugal / : Depth of skull, from rostral limit of supratemporal fenestra to distal tip of epijugal / Antorbital fenestra should open just anterior to the ridge running down from the orbit to the maxilla (anterior to the the trough of Witmer and Sampson). It is much higher than the other opening into the maxilla. ROM Chasmosaurus sp. postorbital

7 Red Deer River, Alberta Isolated?right postorbital with horncore, probably collected as float. Much of postorbital is eroded away. Sinus is partially exposed. Laterosphenoid cotylus not preserved. Contrary to Farke 2006, the frontal sinus does extend partly under the postorbital horncore about halfway under medial half of horncore. But, there is no cornual sinus extending up into the horn itself. As preserved, sinus is mm long (antero-posterior length), mm tall (dorsoventrally) and mm wide (medio laterally). Thus, the total sinus width would have been at least 72 mm (probably more). No sutures are visible anywhere on this specimen, but given banged-up state of specimen I don t know if I d see them anyhow. Bone texture and size of horn, however, indicate probable maturity. Horncore dimensions: mm long (height above orbit), basal length, basal width. Two gentle folds or ridges are present in the wall of the sinus. Both extend vertically, from roof to floor of sinus. Middle of each ridge separated by about 28 mm. At middle portion of sinus, the bony roof is relatively thin about 7.32 mm.. It thickens posteriorly, however. Neurovascular channels are present on the wall of the sinus. They are relatively numerous anteriorly, but disappear completely posterior to the posterior-most fold in the lateral wall of the sinus. The grooves in the space between the two folds are visibly connected to neurovascular foramina in the medial wall of the orbit. The grooves terminate in a canal that passes superiorly into the postorbital horncore and some go into the bone on the dorsal roof of the skull. Thus, blood could have flowed from the orbit into the wall of the frontal sinus and then into the horncore or dorsal roof of the skull (or vice versa). The sinus extends at least through the posterior 2/3 of the orbit. ROM Centrosaurinae Left postorbital (partial) Oldman Formation, Red Deer River, Alberta. Museum Expedition Two big, adjacent resorption pits on postorbital. A possible third on anterior border of palpebral buttress at rostral margin of orbit. A portion of the frontal sinus is preserved. It extends at least for the posterior 2/3 of the orbit; can t tell if it goes much further, because the anterior margin is broken. The preserved portion of the sinus is mm long and at least 20 mm deep (medio-laterally). The sinus extends only up to the medial base of the postorbital horncore, but does not actually extend underneath of it. Only four or five neurovascular grooves trend dorso-ventrally in the sinus. One is broken open to show that it opens through a foramen into the orbit. The proximal end of this groove looks to continue onto the dorsum of the skull.

8 At its thinnest, the sinus is separated from the orbit by only about 9 mm, and from the skull roof by about mm. Thus, the sinus is sandwiched between two thin layers of bone, top and bottom. Height of horncore: mm. Basal length, mm. Basal width, approximately mm. ROM Centrosaurinae Right postorbital (mostly complete), lacrimal, prefrontal, and palpebral (fused) No locality data This specimen was probably from a youngish animal, as the nasal was very nicely and neatly separated from the specimen after depth. Interestingly, the preorbital elements overlapped the nasal for quite a ways about 9 cm or so, anteriorly! A midline suture is present, probably where the prefrontals would have attached. A lightly textured area is present below this (with some vertically-trending neurovascular grooves), which probably represents the top of the nasal cavity (or, potentially the sinus...but, it seems pretty far forward for this). The texture on the external surface of the elements is that of an adult. So, I infer that this was likely a young adult animal. Only a small portion of the frontal sinus is preserved, measuring about 58 mm long. Anteriorly and posteriorly, its borders are obscured by plaster. Thus, I cannot tell what its relationship to the nasal cavity would have been, if any. The sinus extends past at least the posterior 2/3 of the orbit. The sinus was at least mm deep, but possibly more than that. It does not extend anywhere near the base of the postorbital horncore, and was probably separated from it by at least mm. The sinus definitely extends anteriorly to the laterosphenoid cotylus in fact, the entire preserved section is anterior to the laterosphenoid. No sutures are visible between the frontal and postorbital, so I can t tell if the sinus extended into the postorbital itself. A sharp ridge of bone divides the preserved sinus into an anterior and posterior part. The anterior part is longer, measuring at least mm (as preserved). Although the roof of the sinus has been augmented by plaster, where genuine bone is preserved on both sides it appears that the roof of the sinus was quite thin. At the bony portion, the roof is only mm thick. No vascular traces are preserved in the sinus, although preparation isn t great in that area. Part of the margin of the frontal fontanelle may be preserved. It is in the same plane as the middle of the postorbital horncore. However, due to erosion, I am not 100 percent confident that it represents fontanelle margin. Postorbital horncore length = mm, antero-posterior length = mm, and medio-lateral width = mm.

9 23 August 2006 Arrived at museum at 8:30 or so; spent morning taking pathology survey of tyrannosaurid phalanges (spreadsheet in Madagascar pathology folder), for comparison with Majungatholus material. Paper idea: Cheeks in ornithischians: the evidence from bone texture in ankylosaurs Specimen examined: ROM 1215 (Panoplosaurus mirus) (described by L. S. Russell as Edmontonia rugosidens) Complete skull, with left lower jaw. Lower jaw In lateral view, lower 2/3 has a raised surface relative to the upper 1/3. Bone texture in this lower portion displays numerous neurovascular foramina In upper 1/3, bone is relatively smooth. Anteriorly (at location of predentary), a few round neurovascular canals occur. However, even the predentary is inset from the plates fused to the lower jaw. At deepest inset (mid-toothrow), it is mm inset. Posteriorly, below coronoid process, it is 9.04 mm inset. Anteriorly, below predentary, it is 7.86 mm inset. Predentary is well-fused with dentary. Area below dentary teeth is relatively smooth, with only one broad neurovascular impression anteriorly, leading into a largish foramen laterally. In medial view, only the lower ¼ or so of jaw has the raised neurovascular area. Bottom line: If I remember correctly from ontogenetic studies on ankylosaurs, the raised area consists of osteoderms fused onto the lower part of the jaw. Thus, the osteoderms are not fused onto the predentary area (which almost certainly was covered with keratin). Thus, inset area in lower jaw of ankylosaurs does not necessarily indicate presence of cheeks (i.e., inset area below that raised osteoderm area). Upper jaw: areas where you wouldn t expect osteoderms, etc. don t have them (e.g., quadrates, etc.). Maxilla does have big inset emargination. Horizontal part ( buccal emargination) and vertical part, leading down to the toothrow. The horizontal part does have some neurovascular impressions on it (photographed). This would strongly suggest to me that no cheek muscle is attaching here, and no fleshy patch of tissue is attaching here, either. The vertical portion of the maxilla doesn t have these impressions. But, there isn t really an area for the muscles to insert here, either. So, I don t think that ankylosaurs had fleshy cheeks. Or, even non-fleshy cheeks. There is no real evidence for it that I can see based on this. Now if those supposed loose cheek plates that Jaime Headden had once mentioned really existed, I might change my mind.. Anchiceratops ornatus - ROM 802

10 Specimen is top half of a skull. Maxilla, premaxillae, palate, rostral, most of jugals are missing. Right supratemporal fenestra has lots of plaster, so I will base most of my description on the left side. The dorsal midline of the parietal also has lots of plaster. Fortunately, however...the frontal fontanelle and associated region is reasonably well-preserved. There is a little bit of plaster in the lateral portion of the left supratemporal fenestra, but I think everything else should be ok. Channels leading from supratemporal fenestra to frontal fontanelle are not particularly wellpreserved. The bone texture on the dorsum of the skull is somewhat fragmented (generally true throughout for this specimen). Channel is better preserved antero-laterally, where it is definitely present. Although I can infer its presence (based on other specimens), I can t say much more about it. Posteriorly (at supratemporal fenestra border), I would say that the channel is about mm wide. Anteriorly, I would say it is around 60 mm wide. At posterior border of frontal fontanelle, it looks like the two channels run together before diving into the superior portion of the frontal sinus. In this animal, the channel runs for mm before it hits the posterior portion of the frontal sinus. It communicates posteriorly with the upper chamber, and the anterior portion of the channel communicates with the lower chamber. These channels are smooth-floored as near as I can tell, without any major vascular traces. So, channels communicate with both upper and lower sinus. Morphology of the frontal fontanelle The frontal fontanelle is rostro-caudally elongated, expanding slightly rostrally. Thus, the fontanelle is broader anteriorly than posteriorly. The fontanelle measures mm long, mm wide posteriorly, mm wide at midpoint, and mm wide at widest point anteriorly. The posterior 62 mm of the fontanelle is bordered by the channels from the supratemporal fenestrae. Anteriorly, there is a pretty sharp edge that overhangs the frontal sinus itself. The anterior margin of the fontanelle is approximately coincident with the plane passing vertically through the middle of the orbits. Superior portion of frontal sinus (posterior chamber). The posterior chamber of the frontal sinus is completely preserved in this specimen. It measures around 44 mm wide at widest, and passes back approximately mm before a rounded lip drops down into the lower portion of the frontal sinus. A smoothly rounded portion of the parietal overhangs the posterior portion of the upper chamber, by about 17 mm. The upper chamber communicates directly with the posterior portion of the supratemporal channels, separated by a smoothly rounded lip that doesn t really overhang the posterior chamber. The lateral borders of the upper chamber are more or less vertical, and appear to be smooth (although there is somewhat fragmented bone quality in areas). Anterior part of frontal sinus (inferior portion) The anterior part of the frontal sinus (i.e., that portion anterior to the superior chamber) can be divided into two parts, a rostral (anterior, part II) and caudal (posterior, part I) division.

11 Unfortunately, the floor (ventral wall) of the sinus is obscured by plaster in this part. So, I can only really see the lateral portions. The caudal portion is distinctive in that its lateral walls are pretty much vertical. They are completely continuous with the vertical walls of the superior chamber. Anteriorly, they run up to a slight lip of bone from the dorsal surface of the skull. This condition continues for approximately 55 mm, before the sinus expands laterally quite rapidly and drastically, to underlie the postero-medial bases of the horns. A rounded lip of bone separates the portion where the wall is pretty much vertical from the portion where the wall heads medially. The entire lateral portion of the sinus underlying the horncore is preserved yay! I would say that this is a rudimentary cornual sinus. From the medial lip of the frontal fontanelle (approximately coincident with the medial edge of the postorbital horncore), the sinus extends approximately 50 mm on the right and 50 mm on the left. From the midline, this would be approximately 90 mm wide. This part of the sinus is hemispherical and somewhat domed superiorly against the lower border of the very base of the horncore. Anteriorly, I can t tell what the end of the sinus looks like. It appears to just end abruptly. By pouring sand into the sinus, I estimate its minimum volume as 720 ml. Measurements 1: Minimum width of nasals above external naris : Greatest width across orbits : Minimum medio-lateral width of postorbital horn at base, immediately above orbital rim / : Minimum rostro-caudal length of postorbital horn at base, immediately above orbital rim 133 / 149 5: Greatest width across jugals, including epijugal not preserved 6: Least distance from rostral border of orbit to tip of rostral not preserved 7: Distance from tip of rostral to caudal end of maxillary tooth row not preserved 8: Minimum dorso-ventral depth from dorsal surface of nasals to ventral surface of maxilla not preserved 9: Distance from tip of rostral to end of occipital condyle (basal skull length) not preserved 10: Minimum distance from caudal end of external naris to rostral end of orbit / : Length of postorbital horncore from tip of horn to dorsal margin of orbit 632 / : Length of jugal from ventral margin of orbit to ventral limit of jugal not preserved 13: Depth of skull, from rostral limit of supratemporal fenestra to distal tip of epijugal not preserved ROM Centrosaurus apertus Identification based on procurved nasal horn and form of supraorbital horn (visible and partly preserved on right side) Measurements

12 Some are difficult because specimen is in cradle on its side, too fragile to bring upright. The tip of the rostrum is gone, so I have to estimate some of those measurements. However, the estimate is probably within 20 to 30 mm. 1: Minimum width of nasals above external naris : Greatest width across orbits : Minimum medio-lateral width of postorbital horn at base, immediately above orbital rim can t measure 4: Minimum rostro-caudal length of postorbital horn at base, immediately above orbital rim can t measure 5: Greatest width across jugals, including epijugal not preserved. By measuring up from lateral surface of jugal to midline of braincase, I would estimate around 260 mm. 6: Least distance from rostral border of orbit to tip of rostral (not on right) / 560 mm. 7: Distance from tip of rostral to caudal end of maxillary tooth row (not on right) / mm 8: Minimum dorso-ventral depth from dorsal surface of nasals to ventral surface of maxilla mm (left side only) 9: Distance from tip of rostral to end of occipital condyle (basal skull length) 735 mm 10: Minimum distance from caudal end of external naris to rostral end of orbit (no right) / 275 mm 11: Length of postorbital horncore from tip of horn to dorsal margin of orbit can t measure 12: Length of jugal from ventral margin of orbit to ventral limit of jugal can t measure 13: Depth of skull, from rostral limit of supratemporal fenestra to distal tip of epijugal can t measure Description of specimen First: there is no connection between the nasal cavity in the front (i.e., the roof of the nasal) and the frontal sinus. This area is beautifully preserved. The anterior part of the frontal sinus is in line with the plane passing vertically through the middle of the orbit. A portion of the margin of the frontal foramen is preserved on the left side. Describing from front to back... The anterior portion of the sinus still has matrix attached. Too bad. I have to guess the rough position of the front of the sinus based on where the wall of bone is at the rear of the nasal. Just posterior to the anterior portion of the sinus is a ridge of bone that has a slight posterior tilt when running from ventral to dorsal. This ridge of bone is just anterior (by about 10 mm or so) to the position of the plane that has the laterosphenoid. This is probably equivalent to what s called the transverse buttress in Triceratops. This buttress is about 13 mm thick. Maximum width of sinus is just posterior to this; here it measures mm across. At this point, it is below part of the frontal fontanelle opening. Length of sinus, as measured from caudal-most portion of parietal to anterior-most possible part on matrix at front of sinus, is mm long. It had a maximum depth of approximately mm.

13 The frontal fontanelle has only the vascularized portion preserved; none of the channels are preserved here. The textured, vascularized portion on the dorsal roof of the skull turn abruptly to smooth, avascular bone when you reach the ventral surface of the fontanelle (i.e., the part inside the sinus). Crushing has appeared to have skewed the medial margin of the fontanelle to the left (from the right side). So, the border of the fontanelle is moved past the midline. Thus, I can t estimate width of fontanelle. In coronal cross-section, the sinus appears to have the outline of a rounded triangle (with the point ventral, as in V). So, it broadens dorsally. The sinus in this specimen has a minimum volume of 200 ml (as measured by pouring sand into the sinus). It probably had a maximum volume of about 300 ml (estimated by how much of sinus is gone on right side). Description of the nasal cavity and related areas First, a note on the braincase. Leading up to the oval window (fenestra ovalis), there is a shallow groove trending supero-medially (on the anterior surface of paroccipital process of the exoccipital). Lehman identifies this as the groove for the tympanic cavity. This makes sense. No pneumatic recesses are visible coming off of it, however. The groove measures about mm wide and is about mm long before it runs into the oval window. The oval window is mm tall by 9.22 mm wide, approximately. A second groove runs antero-inferiorly from the oval window. It is about 4.73 mm deep (for stapedial artery, perhaps?). Question: what s in the space between the basisphenoids? Ventral surface of the nasal bone: A midline ridge confluent with the ridge on the front of the braincase occurs. This is probably just matrix left in this specimen. There is some dark-colored organic matter along a midline layer. One of two possibilities: 1) just two different burial events; 2) sediment sorting due to soft tissue (nasal septum) left over before final decay. Interesting... On either side of the midline are two ridges that start level with the top of the orbits and trend anteriorly on the underside of the nasals. They are roughly triangular in cross section, raised about mm above the surrounding bone. About 5 cm along the ridge (at least visible on the right side), a second smaller, sharper ridge breaks off and heads ventrally (paralleling the orbit), towards the internal maxillary fossa. It stops about midway though the palatine. A shallow trough posterior to the second ridge is equivalent to the internal maxillary fossa of Witmer s ceratopsid nose manuscript. This trough starts at the very top of the anterior surface of the orbit, and is separated from the orbit by a second sharp ridge of bone posteriorly (as you move inferiorly, this is probably formed in part by the jugal). The trough has a constant width, around mm, along its entire length. It stops just superiorly to the palatomaxillary foramen, from which it is separated by a sharp ridge of bone. A little pocket of bone is at the very end of the trough. I wonder if this trough isn t the path of the nasolacrimal duct? So, I wouldn t say (as in the MS) that the internal maxillary fossa continues up onto the lacrimal.

14 In this specimen, the palatine forms the entire medial border of the palatomaxillary foramen, and the maxilla forms the entire lateral border. No internal antorbital fenestra is visible in this specimen. Measuring from the top of the internal choana, the nasal cavity was at least 235 mm tall and mm wide. The internal choana has an estimated length of at least mm. Maximum length, from end of vestibule to olfactory bulbs, is mm. ROM 767 Centrosaurus apertus skull on display 1: Minimum width of nasals above external naris : Greatest width across orbits : Minimum medio-lateral width of postorbital horn at base, immediately above orbital rim / : Minimum rostro-caudal length of postorbital horn at base, immediately above orbital rim / : Greatest width across jugals, including epijugal : Least distance from rostral border of orbit to tip of rostral / : Distance from tip of rostral to caudal end of maxillary tooth row / : Minimum dorso-ventral depth from dorsal surface of nasals to ventral surface of maxilla / : Distance from tip of rostral to end of occipital condyle (basal skull length) : Minimum distance from caudal end of external naris to rostral end of orbit / : Length of postorbital horncore from tip of horn to dorsal margin of orbit / : Length of jugal from ventral margin of orbit to ventral limit of jugal / : Depth of skull, from rostral limit of supratemporal fenestra to distal tip of epijugal / This specimen is beautifully preserved!!! The animal is an adult, but probably not an old adult. Bone texture indicates adult, but sutures are still open for the most part in fact, almost entirely open on face and skull roof. Note: the jugal actually extends anteriorly to form part of the antorbital fenestra in this specimen. The skull is probably very slightly medio-laterally compressed. Frontal fontanelle is mm long, mm wide. Stays pretty constant in width along its entire length. The texture around the fontanelle is pretty standard for the dorsal skull roof; it s rather rough and vascularized. It has a prominent lip that overhangs most of the sinus mm in length until you get back to the posterior portion of the sinus. The anterior-most portion of the sinus is in line with the plane marking the division between the anterior 1/3 and the posterior 2/3 of the orbit. The fontanelle is, at least anteriorly, bordered entirely by the frontals; laterally, by the frontals, too. Unfortunately, postorbital / parietal division is a little hazy on this specimen, so I can t tell what s bordering it posteriorly.

15 The entire floor of the sinus is obscured by matrix; they didn t prep it out, darn it! However, it is at least mm deep anteriorly, mm deep in the middle of the sinus, and mm deep posteriorly (towards the posterior / upper portion of the sinus). A bit of transverse buttress is present, about mm posterior from the anterior margin of the frontal fontanelle. This area was bashed up a little bit during prep, so I can t tell much more on morphology, other than that it was present. It probably won t photograph well either, unfortunately. However, the floor of the sinus is plaster posterior to this. In the posterior chamber of the sinus (or the superior chamber, as I ve variably called it), the walls of the sinus run pretty much vertically. At the anterior portion of this chamber, the walls curve laterally to expand into the main body of the sinus. The maximum depth here (with the bit of matrix at the bottom) is mm below the supratemporal channels. The walls of this superior chamber are pretty much vertical, and smooth. Posteriorly, there isn t really an overhang of parietal. This chamber is about mm long. The supratemporal channels are extremely well-preserved, definitively showing that they exist in Centrosaurus. The left supratemporal channel measures mm in length; the right one measures mm from the supratemporal fenestra to the dive into the frontal sinus. The right channel is mm wide, and the left one is mm wide at its narrowest (posteriorly), expanding to mm anteriorly. The left channel has a depth of approximately 6.47 mm below the dorsal surface of the parietal; the right around 3.57 mm. The channel is smooth, without any vascular impressions or foramina. A small expansion of the sinuses, on their posterior border, cuts slightly into the parietal bar. This creates an expanded \ \M/ / appearance to the channels and dorsal (textured) parietal bar. The \ \ and / / represent the channels, and the bottom gaps on the M represent the little diverticula of the channels that project posteriorly. The supratemporal channels overhang the supratemporal fenestrae, and the floor of these channels is apparently quite then (<3 mm). A thin slice of bone (the V on the M separates the right and left channels on the dorsal surface, although not by a lot!). Pathologies: only one (assume all other bones of skull are complete and present), on the left parietal. The bone is imbricated, with the lateral portion of the parietal shoved on top of the medial part of the parietal. No real fracture calluses or anything on dorsal or ventral surface. Definitely real this area is not particularly plastered. Thicknesses: on left side, squamosal Thursday, August 24, 2006 Arrived at the museum at 8:30 a.m. Spent a little time discussing nasal cavities with Dave. For part of the morning, I m going to spend more time with ROM My notes will be up in the section with that skull, in order to keep everything together. The focus will be on nasal cavity features. ROM 843 Chasmosaurus belli Skull is very nicely preserved. However, the rostrum is unfortunately absent. Measurements

16 Note: the rostrum and portion of premaxilla are absent. I think the reconstruction is probably pretty good, but note that basal skull length, etc., are estimated (probably within 5 to 10 cm). 1: Minimum width of nasals above external naris : Greatest width across orbits : Minimum medio-lateral width of postorbital horn at base, immediately above orbital rim can t measure, because horns are so low and knobby 4: Minimum rostro-caudal length of postorbital horn at base, immediately above orbital rim can t measure, because horns are so low and knobby 5: Greatest width across jugals, including epijugal 704 6: Least distance from rostral border of orbit to tip of rostral 635 / 649 7: Distance from tip of rostral to caudal end of maxillary tooth row left side only right side busted off at end of tooth row 797 8: Minimum dorso-ventral depth from dorsal surface of nasals to ventral surface of maxilla / : Distance from tip of rostral to end of occipital condyle (basal skull length) : Minimum distance from caudal end of external naris to rostral end of orbit / : Length of postorbital horncore from tip of horn to dorsal margin of orbit / : Length of jugal from ventral margin of orbit to ventral limit of jugal / : Depth of skull, from rostral limit of supratemporal fenestra to distal tip of epijugal / The frontal sinus is open and exposed, but the entire floor is covered in plaster, and/or covered in matrix (anteriorly). This specimen could stand a little repreparation in this region. The lateral walls are exposed in places. Frontal fontanelle The frontal fontanelle measures mm long by mm wide (anteriorly), mm wide (middle) and mm wide (posteriorly). The rough, vascularized texture of the dorsal surface of the skull surrounds the entire fontanelle, except for the very posterior portion (which is slightly distorted, but does show the supratemporal channels running into it). The lateral borders are mostly irregular in shape. The anterior margin of the frontal fontanelle is coincident with the plane passing through the anterior 1/3 of the orbits. Frontal sinus description As mentioned above, the floor of the sinus is entirely covered by matrix and plaster. Anteriorly, the sinus undercuts the frontal for at least 6 mm. Can t tell beyond this due to plaster. At its anterior-most point, the sinus is at least mm deep. The inferior/anterior chamber undercuts the frontals laterally, for a distance of at least 20 mm. At the middle of the sinus, it has a maximal depth below the surface of the skull of at least mm. The anterior chamber is the widest part of the sinus, and it has a length of at least 97 mm. It ends at the nearly vertical wall of the posterior chamber (which is best seen on the left side; too much plaster on the right). The posterior chamber has a maximum depth of at least mm

17 (although plaster floors it, of course), and continues back for at least 43 mm. The wall of the chamber is completely smooth. The posterior chamber connects with the supratemporal channels. They are preserved on both sides, although the left is slightly better. The floor of the supratemporal channels is completely smooth. On both sides, the rounded posterior margin leading down into the supratemporal fenestrae is preserved. The left and right channels connect anteriorly, and dive over a quite rounded (almost half-circular) lip into the posterior frontal sinus. I see the stretched out M arrangement again in this specimen for the tab of parietal that separates the two channels. The channels do come together quite smoothly at the middle, though. On the left side, the supratemporal channels are mm long, at least mm wide, and 5.57 mm deep (below the dorsal surface of the parietal in that area). On the right side, the channel is mm long, I can t measure width, and it is mm deep right next to the supratemporal fenestra, and 2.62 mm deep before it dives into the posterior chamber of the sinus. On the left side, it s pretty clearly lipped over the supratemporal fenestra by a thin (<4 mm) piece of bone. As measured by pouring sand in, the minimum volume of the sinus is 300 ml. Friday, August 25, 2006 Arrived at museum by 8:45. I will spend my last partial day here wrapping up odds and ends in the collection. First, I will spend a bit of time with ROM Notes are above with the notes for that specimen from the other day. Conclusions about sinuses: Blood flow There was no major bloodflow between the arteries of the supratemporal fenestral area (i.e., the complex leading out onto the frill) and the frontal sinus. There was bloodflow between the frontal sinus and the vessels of the orbit. No major blood vessels were carried in the supratemporal channels. Major blood vessels were carried out of the supratemporal fenestrae onto the frill. Blood vessels from the horncores passed into the orbital region. Implications: probably not related to cooling blood for the brain (at least not directly). Hmm...I wonder if this bloodflow pattern could say something important about the development of the sinus? Would it be worth CT scanning a Chasmosaurus postorbital in order to look at the connections (assuming resolution was adequate?)? Origin Possibilities: paratympanic Paranasal: can t rule out, based on examination of Made some sketches of this.

18 Possible figures Cross-section of Centrosaurus skull (based on specimen here), showing potential pneumatic pathways.