Preservation frequency of tissue-like structures in vertebrate remains from the upper Campanian of Alberta: Dinosaur Park Formation

Journal Pre-proof Preservation rates of tissue-like structures in vertebrate remains from the upper Campanian of Alberta: Dinosaur Park Formation Aaron J. van der Reest, Philip J. Currie PII:

S0195-6671(19)30087-4

DOI:

https://doi.org/10.1016/j.cretres.2019.104370

Reference:

YCRES 104370

To appear in:

Cretaceous Research

Received Date: 4 March 2019 Revised Date:

29 November 2019

Accepted Date: 30 December 2019

Please cite this article as: van der Reest, A.J., Currie, P.J., Preservation rates of tissue-like structures in vertebrate remains from the upper Campanian of Alberta: Dinosaur Park Formation, Cretaceous Research, https://doi.org/10.1016/j.cretres.2019.104370. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Ltd.

1

Preservation rates of tissue-like structures in vertebrate remains from the upper Campanian of Alberta:

2

Dinosaur Park Formation.

3 4

Aaron J. van der Reest* and Philip J. Currie

5 6

Department of Biological Sciences, University of Alberta, Edmonton, Alberta, Canada, T6G 2E9.

7 8 9

*To whom correspondence should be addressed.

10 11

ABSTRACT In recent years, several papers have claimed that soft tissue can preserve within bone matrix of

12

extinct vertebrates, some dating back over 100 million years. Work conducted on specimens from

13

Montana suggested sediment type may influence preservation of original tissues and proteins. An

14

alternative hypothesis is that soft tissue preservation may be linked to the time that a specimen is

15

exposed to the environment prior to burial. The time of exposure can be estimated by the degree of

16

disarticulation of a skeleton. A study was conducted to determine if these factors truly contribute to the

17

preservation of soft tissue-like structures in the geological record. This study is not intended to verify the

18

presence of proteins but is simply to determine how common are macrostructures that look like soft

19

tissue preservation. Samples were placed into a 0.5 M solution of (ethylenedinitrilo)tetraacetic acid,

20

disodium salt, dihydrate (EDTA) for two months to dissolve mineral components. All specimens studied

21

were collected from the Dinosaur Park Formation (upper Campanian) to minimize stratigraphic variation

22

that may influence preservation. Dissolution of vertebrate remains sampled indicate an unexpectedly

23

high rate of structural preservation. Fifteen dinosaur, two crocodilian, one fish, and one turtle were

24

sampled for a total of nineteen specimens. Specimens were chosen based on sediment type and degree

25

of articulation. Approximately half of the samples were recovered from sandstones, and the other half

26

originated in mudstones. Additionally, approximately half of the samples were collected from

27

articulated or closely associated skeletons, and the other half were taken from isolated bones or

28

specimens from micro vertebrate sites. Of the nineteen specimens tested, eighteen specimens produced

29

“vessel” structures, eighteen had extracellular organic-like matter, and seven revealed “osteocyte”

30

structures. Although “vessel” and extracellular organic-like structures are not associated to a specific

31

matrix type or degree of association, “osteocyte” structures appear to be associated more often with

32

articulated/associated specimens, especially if they are preserved in sandstones.

33

34 35

INTRODUCTION Methods in palaeontology changed little over the first 150 or so years since the first dinosaur

36

fossils were described (Buckland, 1824), with most of the research focussing on basic anatomical

37

description of bones and the analysis of relationships. Description of anatomy and classification of

38

different species are fundamental steps towards more complex, derived studies, such as rates of

39

evolution or speciation. Studies of extant taxa through the use of molecular data from species genomes

40

have started to dominate phylogenetic research for modern clades (Jarvis et al., 2014). Although

41

modern molecular data has been questioned in the past, recent work has improved the resolution of

42

evolutionary rates and divergence dates (Morlon et al., 2011). Palaeontology, which was slow to adopt

43

new methodologies until the last few decades, is now adopting more sophisticated methods of study

44

from other scientific fields. Many of these developments have benefited from statistical analyses using

45

computer programs such as “R” (Benson et al., 2014; Campione et al., 2014).

46

In 1999, a theropod dinosaur from Mongolia, Shuuvuia, was shown to have small white

47

filamentous fragments preserved in the matrix surrounding the specimen. Resulting

48

immunohistochemistry (IHC) trials indicated the fibres preserved around Shuuvuia were in fact

49

fragments of unaltered feathers with protein still intact (Schweitzer et al., 1999). Following the

50

discoveries with Shuuvuia, pliable tissues were reported from dissolved bone of a Tyrannosaurus rex

51

(Schweitzer et al., 2005). These tissues consisted of fibrous, flexible, collagen structures, as well as

52

anastomosing “vessel” structures containing red, spherical objects with darker cores. It was proposed

53

these structures represented unaltered soft tissues – including collagen-I, blood vessels, and red blood

54

cells – preserved within the bone matrix. Further investigation into the preservation of soft tissues in

55

vertebrate fossil and sub-fossil elements from around the world (Schweitzer et al., 2007b) indicates

56

these structures are not uncommon. Taxa with soft tissue structures published by Schweitzer et al.

57

(2007b) included Brachylophosaurus canadensis, Mammuthus, Megoedon, Triceratops horridus,

58

Trichechidae sp, Tyrannosaurus rex, a dicynodont and two unidentified theropods (one from

59

Madagascar and the other from the Santana Formation of Brazil). Although, in this previous study, not

60

all specimens sampled contained preserved soft tissue, many specimens were discovered to possess

61

some form of “vessel” structures, collagen, or osteocytes (Schweitzer et al., 2007b). Furthermore, a

62

Triceratops horridus revealed exceptionally well preserved tissues, including vessels and osteocytes, of

63

which the latter contained structures resembling nuclei (Armitage and Anderson, 2013). Subsequent to

64

the release of Schweitzer et al. (2007b), the analysis of unaltered preserved soft tissue was taken to the

65

next level. Immunohistochemistry was again employed to investigate collagen-like bundles that were

66

observed within the bone matrix of Tyrannosaurus rex (Schweitzer et al., 2007a). By using several anti-

67

body treatments, it was shown that collagen-I persisted, although degraded, within the bone matrix.

68

This was the first instance where collagen-I was found to be present within bone matrix of a dinosaur

69

and was the first bone older than 66 Mya. Unsurprisingly, the thought that protein fragments could

70

survive through deep time was contentious. It was claimed that the more likely interpretation of these

71

structures would be biofilms formed by bacterial colonies during decomposition (Kaye et al., 2008). On

72

the other hand, it has also been suggested that if the presence of original protein is in fact accurate,

73

then bacterial biofilms must play a role in preserving it (Peterson et al., 2010). Following the

74

identification and confirmation of the presence of unaltered proteins in dinosaur bone using

75

immunohistochemistry, sequences of collagen-I were sequenced using mass spectrometry (Asara et al.,

76

2007; Schweitzer et al., 2009a). This work was duplicated several years later using more precise

77

measurements and sequencing technology, confirming that the results were repeatable, and showing

78

that original results were accurate (Schroeter et al., 2017). In addition to collagen I fragments,

79

polypeptides have been reported in fossil osteocytes (Schweitzer et al., 2013) and vessels (Cleland et al.,

80

2015). Recent work has suggested that proteins are altered into N-heterocyclic polymers during

81

fossilization (Wiemann et al., 2018).

82

Research into the primary mechanism for preservation of unaltered tissues has investigated

83

several potential factors. Originally, it was proposed that the higher porosity within a sandstone matrix

84

would facilitate the drainage of water and decay enzymes away from remaining tissues within bone

85

(Schweitzer et al., 2007a). Further research into chemical influences on preservation of unaltered

86

proteins suggested that molecular cross-linking in collagen fibrils stabilizes the molecule (San Antonio et

87

al., 2011). McGowan (1983) theorized that ironstone surrounding and within fossil bone likely derived

88

from the iron within blood. Schweitzer et al. (2013b) gave an explanation to how this may occur,

89

suggesting that the iron released from haemoglobin, myoglobin, and cytochromes bind to proteins and

90

lipids, leading to their crosslinking. Investigations into the preservation of collagen and osteocytes

91

recovered from Lufengosaurus (Early Jurassic, 190-197 Ma) proposed that hematite cementation may

92

play a role in preventing the degradation of soft tissue within bone (Lee et al., 2017). The authors

93

suggest that the iron used to form the hematite may come from erythrocytes (red blood cells). However,

94

they also point out that it would require approximately 20,000 cells to form a single 10 µm hematite

95

sphere. Ullmann et al. (2019) investigated tissue preservation in a South Dakotan, Hell Creek Formation,

96

Edmontosaurus annectens bonebed. Results of that study indicate that all skeletal elements potentially

97

preserve tissue, and that preservation is not associated with the thickness of overburden over an

98

element.

99

Extraction of soft tissues from fossilised vertebrate bone has raised the potential for studies

100

once thought to be impossible for taxa from deep time. Studies could potentially produce molecular

101

phylogenies, evolutionary rate estimates, divergence events, molecular evidence of cladogenesis and/or

102

anagenesis, and genetic diversity within populations to name a few. To facilitate these studies however,

103

bone dissolution is the primary, and most successful procedure currently known to extract preserved

104

soft tissues from fossil bone.

105

Select bones were dissolved to extract tissue-like structures from a multitude of vertebrate taxa

106

from the Dinosaur Park Formation (upper Campanian, ca. 76.3-74.4 Ma) primarily within Dinosaur

107

Provincial Park in southern Alberta (Fig. 1) to determine the potential for a large-scale study involving a

108

multitude of taxa. The primary goal of this project was to obtain a better understanding of the required

109

environmental and sedimentological conditions during burial of vertebrate remains to preserve possible

110

soft tissue-like structures. Two potential preservational criteria were tested: 1) Sedimentological

111

association (sandstone or mudstone) was identified to determine if soft tissue-like structures are in fact

112

limited to sandstones. 2) Degree of articulation within a specimen (articulated/associated skeletons or

113

isolated elements/microvertebrate sites) to test if the length of time exposed to environmental

114

conditions prior to final burial decreases the likelihood of soft tissue-like structure preservation.

115

Structures that were most commonly recovered and identified in this study are “vessels” (Fig. 2, Fig. S1),

116

“extracellular organic matrix” (EOM)(Fig, 3), or “osteocytes” (Fig. 4, Fig. S2). Because this study is not

117

intended to determine if these structures are unaltered original tissues, structures are referred to as

118

“vessel(s)”, “osteocyte(s)”, “EOM”, etc. for simplicity.

119

ABBREVIATIONS

120

Ethanol (ETOH); (ethylenedinitrilo)tetraacetic acid, disodium salt, dihydrate (EDTA); extracellular

121

organic matrix (EOM); Mega-annum (Ma – million year); sodium hydroxide (NaOH); University of Alberta

122

Laboratory for Vertebrate Palaeontology (UALVP)

123

MATERIALS AND METHODS

124

Specimen Selection

125

Specimens used within the study were selected based on three main criteria. First, specimens

126

were chosen to cover as many different taxa as possible. This provides future opportunities to expand

127

on potential protein sequencing and thus molecular phylogenetic studies. Of the nineteen specimens

128

sampled, there are fifteen dinosaurs, two crocodilians, one fish, and one turtle. All specimens were

129

collected from the Dinosaur Park Formation to reduce stratigraphic variation that may influence

130

preservation. Secondly, specimens were chosen on the basis of the depositional matrix composition –

131

sandstone or mudstone – that they were found in. This was investigated to determine if preservation

132

potential differs between depositional environments as proposed by Schweitzer at al. (2007a). Thirdly,

133

specimens were chosen based on whether they were articulated or associated skeletons, or were

134

isolated or recovered from bonebeds (including microvertebrate sites). This was done to determine

135

whether the length of time a specimen was exposed to environmental conditions prior to burial

136

influences soft tissue-like preservation. It is assumed that in most cases articulated specimens were

137

buried more rapidly than disarticulated and isolated bones. However, it is recognized that this approach

138

is simplistic as processes and timing of disarticulation between the death of an animal and its final burial

139

are highly variable. Nevertheless, it was used as a rough proxy to see if differences exist.

140

All samples were collected from specimens from the collections of the University of Alberta

141

Laboratory for Vertebrate Palaeontology (UALVP). Specimens from microsites were generally small

142

enough to be placed directly into dissolution dishes without needing to be broken up. Samples from

143

larger articulated or isolated specimens were typically fractured off by using a hammer and awl, or wire

144

cutters. The work surface was clean, and fragments were usually taken from the insides of the

145

specimens, rather than the exposed surfaces. This method was employed to reduce the number of

146

potential contaminants introduced into the bone matrix by cutting pieces off with saw blades. All

147

samples were collected while wearing nitrile gloves to prevent the contamination of proteins from

148

contact with human skin. Once removed from the original specimen, samples were then placed into

149

clean plastic baggies during transport to prevent contamination.

150

Preparation of samples

151

Because several specimens that were used had been collected, prepared, and stored in non-sterile

152

environments for up to a century, the following protocols were used to remove as much previous

153

contamination as possible. To reduce the potential of personnel contamination within the laboratory

154

environment, all lab work was performed while wearing a lab coat, hair net, medical face mask, safety

155

glasses, and nitrile gloves to prevent contamination from persons involved.

156

To clean specimens of any potential contaminating debris, the following steps were taken:

157

1. All samples were sonicated using de-ionized water to remove surface contaminants.

158

2. All samples were then sonicated in 50% ethanol (ETOH) solution to remove any unknown

159

consolidants, to sterilize bacteria, and to further remove any debris from the surfaces.

160

3. All samples were then sonicated one further time in de-ionized water to flush ETOH.

161

Once cleaned, samples were taken from the final sonication directly into separate wells in 6-well,

162

non-tissue-culture treated sterile culture plates.

163

Chemical Laboratory Protocols

164

All solutions used within the study were mixed by hand in the Earth and Atmospheric Sciences

165

radiocarbon dating wet lab in the Centennial Centre for Interdisciplinary Sciences at the University of

166

Alberta, Edmonton. Fifty percent ETOH solution for the sterilization sonic bath was made by mixing

167

equal parts of 100% ETOH and deionized water. Dissolution of bone used a 0.5 mol solution of

168

(ethylenedinitrilo)tetraacetic acid, disodium salt, and dihydrate (EDTA) (molecular weight = 372.240

169

g/mol). To create a 2 litre solution, approximately 372.240 g of solid EDTA was weighed into a 4 litre

170

beaker and then deionized water was added until a volume of 2 litres was reached. Due to the low

171

solubility at low pH of EDTA, sodium hydroxide (NaOH) was slowly added until both EDTA and NaOH was

172

dissolved. The final pH was tested using pH indicator strips with a target of 8.0. This solution was then

173

filtered through a 2 µm filter to remove all undissolved crystals and other debris that may have entered

174

the solution by air.

175

Dissolution of samples

176

Samples (three samples from each specimen) of bone were each placed into a single well in a

177

sterile 6-well non-TC treated microculture plate (two specimens per each 6-well plate). Approximately 9

178

mL of 0.5 M EDTA was then added to each sample well by sterile transfer pipette. Each well was

179

inspected daily for evidence of tissue-like structures as the bone fragment samples dissolved. Samples

180

that contained identifiable soft tissue-like structures had organics removed and were placed into

181

microfuge tubes containing 0.1 mol/L EDTA to prevent further dissolution and damage. Transfers of soft

182

tissue-like structures were performed using sterile glass Pasteur pipettes with manual bulbs.

183

Mounting and imaging of tissues

184

Selected tissue samples were transferred to single depression slides using sterile glass Pasteur

185

pipettes and manual bulbs. Depression slides were chosen to prevent three dimensional structures such

186

as vascular networks from being crushed and/or distorted during imaging. Cover slips for microscope

187

slides were held in place using nail-polish.

188

Imaging was conducted in the Department of Biological Sciences Microscopy Laboratory using a

189

Leica compound microscope with attached camera. Scale bars were calibrated using a stage

190

micrometre.

191

Microcrystals of gypsum within “vessels” were identified by using a petrographic microscope

192

and morphology. Hematite and pyrite are identified by morphology because petrography could not be

193

used due to their high opacity (Fig. S3).

194

RESULTS

195

Although all specimens had at least one type of tissue recovered post dissolution, the amount

196

recovered varied dramatically. Several specimens produced such high quantities that some samples

197

were removed from the EDTA for other analyses. Results herein are reported phylogenetically, followed

198

by specimen numbers if multiple specimens of the same genus and species were sampled.

199

Anchiacipenser acanthaspis (Actinopterygii, UALVP 56596 from Q266, articulated, river sandstone, Fig

200

1-“1”)

201

UALVP 56596, the holotype of Anchiacipenser acanthaspis (Sato et al., 2018), had three samples

202

taken from a dorsal osteoderm. In all thee samples, small fragments of “vessels” (Fig 2f) were recovered

203

as well as an unidentified white mesh-like structure. In one of the three samples, a tightly bound fibrous

204

mat was found. This mat originated from the bottom-most layer of the osteoderm and was oriented

205

anteroposteriorly along the midline. Currently, the morphology of the mat most closely resembles that

206

of dermal collagen, which is not unexpected because osteoderms derive from dermal tissues during

207

ontogeny. Small clumps of “EOM” material were recovered.

208

Basilemys sp. (Testudines, UALVP 59613, disarticulated ungual from a microsite in mudstone Fig 1-“2”)

209

UALVP 59613, a Basilemys ungual, was broken into three portions. From each sample, large

210

quantities of degraded “vessels” were recovered. Visually, they appear to be of crystalline structure,

211

suggesting that there has been some form of alteration to the original cellular matrix. Until further work

212

can be done, the true nature of these organics cannot be positively determined. However, they are

213

considered as tissue-like structures in this study. Small clusters of “EOM” material were visible, although

214

rare.

215

Crocodylia indet. (Archosauria, UALVP 59615, isolated osteoderm within a mudstone microsite, Fig 1-

216

“3”)

217

UALVP 59615, a single, unfragmented, small crocodylian osteoderm was placed into one dish

218

well. This specimen produced a large number of “vessel” structures and “EOM” fibrous matrix. Although

219

“vessels” are abundant, most appear to have undergone some form of alteration and now appear as

220

crystalline-like structures. Some “vessel” structures, however, do appear to contain structures

221

resembling endothelial nuclei, suggesting there is the possibility of some residual organics remaining.

222

Leidyosuchus canadensis (Alligatoroidea, UALVP 59616, articulated skull, mudstone, Fig 1-“4”)

223

UALVP 59616, an articulated skull and mandibles of Leidyosuchus, had fragments from the left

224

parasphenoid and left dentary removed because of fractures within the specimen. Vinac consolidant

225

had been applied to the specimen when it was in the field, but was removed by soaking samples in

226

acetone for 10 minutes with agitation. Although few “vessels” were recovered, those that were (Fig 2e),

227

contained large numbers of hematite spheres. Large bundles of “EOM” fibres were also recovered (Fig

228

3a), and several “osteocyte” structures were observed (fig 4d). “Osteocyte” structures are poorly

229

preserved and have badly damaged ‘filipodia’.

230

cf. Euoplocephalus tutus (Ankylosauridae, UALVP 56607, isolated partial skull found in a well cemented

231

sandstone, Fig 1-“5”)

232

UALVP 56607, a partial Euoplocephalus skull from Q265, had a small fragment and a larger

233

fragment (subsequently broken into two) of the posterior part of the skull sampled for dissolution. The

234

initial small fragment was recovered from a piece that had been weathered off the main specimen in the

235

field. Although “vessels” are common, most of these are heavily infilled with solid hematite or crystalline

236

hematite spheres (Fig 2d). In areas where the original pliable “vessel” tissues are visible, structures

237

resembling endothelial nuclei can be discerned (Fig. 2d). Interestingly, regions of “vessel” structures are

238

consistently capped on either end by large amounts of hematite. Although “EOM” was not uncommon,

239

no “osteocytes” were observed.

240

Panoplosaurus mirus (Nodosauridae, UALVP 12, partial cranium with isolated osteoderms, no other

241

skeletal elements, sandstone concretion with high iron, no locality data so not figured in Fig. 1)

242

UALVP 12 is a partial cranium of Panoplosaurus associated with osteoderms collected in 1920 by

243

G. Sternberg. A fragment of dermal armour with original ironstone matrix still attached was sampled.

244

After nearly three months of dissolution in EDTA, virtually no possible soft anatomical material was

245

recovered. A few small, and heavily stained “EOM” clumps were visible, and only a couple of

246

“osteocyte” structures were observed (Fig 4c). Similar to other “osteocyte” structures, these are poorly

247

preserved with limited extensions of ‘filipodia’.

248

Mercuriceratops gemini (Ceratopsidae, UALVP 54559, squamosal with associated cranial fragments

249

(Ryan et al., 2014), sandstone, Fig 1-“6”)

250

UALVP 54559 had a fragment of associated cranial bone of Mercuriceratops divided into two

251

smaller samples and a third small fragment of surface collected bone from the main specimen was used

252

as the third sample. Two of the three fragments released high amounts of iron staining upon dissolution.

253

These two samples also produced high numbers of flexible “vessels” and “EOM” fibres after dissolution.

254

“EOM” matrix is abundant around vessels and throughout the bone. Interestingly no “osteocyte”

255

structures were recovered. Nearly all “vessels” observed are three-dimensional anastomosing

256

structures. Within the “vessel” walls, endothelial nuclei-like structures can be discerned, although large

257

pyrite crystals (identified by their cubic shape and complete lack of light penetration) are also present

258

within the lumina of “vessels”. The third sample produced organic-like structures, although in

259

significantly fewer amounts.

260

Chasmosaurus belli (Ceratopsidae, UALVP 52613, articulated skeleton of juvenile in sandstone Fig 1-“7”)

261 262

UALVP 52613, an articulated juvenile Chasmosaurus belli from Q255, had fragments of surface collected cranial bone sampled from a complete skull (Currie et al., 2016). Of all the specimens sampled,

263

UALVP 52613 produced the highest number of “vessels” and “EOM”. During dissolution of the

264

specimens, three-dimensional anastomosing “vessels” could be seen protruding from every surface.

265

“Vessels” were surrounded by a fibrous matrix of “EOM” after the dissolution of bone matrix. “EOM”

266

fibers are more common than expected and densely intertwined but can easily be pulled apart.

267

Although “vessels” remain intact throughout all three fragments, they appear to have been altered

268

significantly. Large amounts of microscopic gypsum crystals are found throughout most of the “vessels”.

269

Fractured, orange organic-like structures are visible on the inner surfaces of the “vessel” walls. One

270

“osteocyte” structure was observed (Fig. 4b), however, was poorly preserved.

271

Chasmosaurus sp. (Ceratopsidae, UALVP 55926, partially articulated skeleton from Q263, mud filled

272

channel mudstone, Fig 1-“8”)

273

Three fragments of the tibia of a Chasmosaurus (UALVP 55926) were removed as samples. Each

274

sample contained significant numbers of “vessels” and “EOM” fibrous filaments. The “vessels” are well-

275

preserved and appear to frequently contain endothelia nuclei-like structures. Most “vessels” appear

276

slightly pitted (Fig 2c).. Vessels are three-dimensional anastomosing structures and several are observed

277

with “EOM” matrix surrounding them. Minerals morphologically similar to be hematite and gypsum are

278

observed throughout most “vessels”.

279

Centrosaurus apertus (Ceratopsidae, UALVP 55794, isolated eroded skull in sandstone, and because

280

there is no locality data, the specimen is not included in Fig. 1)

281

Samples were removed from a fragment of cranial bone (UALVP 55794) recovered as an eroded

282

portion of a Centrosaurus skull. Although only three fragments of “vessels” were recovered from one

283

sample and no “osteocyte” structures were observed, several loosely woven “EOM” fragments were

284

observed in this specimen.

285

Styracosaurus albertensis (Ceratopsidae, UALVP 52612, partially articulated skeleton from Q256 from

286

very fine sand, Fig 1-“9”)

287

Samples were removed from a fragment of unidentified cranial bone found eroding out along

288

with the skull of a Styracosaurus (UALVP 52612). Few tissue structures were recovered from any of the

289

three samples. Only a few small fragments of “vessels” were recovered, and only small portions of

290

”EOM” fibres were observed. No “osteocyte” structures were observed.

291

Styracosaurus albertensis (Ceratopsidae, UALVP 55900, articulated skeleton from a channel lag

292

sandstone in Q262, Fig 1-“10”)

293

Two skeletal elements from UALVP 55900 (Styracosaurus) were sampled. Three samples were

294

removed from the left scapula and another three from the left humerus. Interestingly few “vessels”

295

were recovered from the scapula. However, all three samples from the humerus produced high

296

numbers of “vessels” (Fig 2a) and “EOM” structures. “Vessels” are fully three-dimensional structures

297

connecting through multiple levels of the cortex. Regardless of the number of these structures, no

298

“osteocytes” were observed. Small, red, spherical objects are present within a few “vessels”, although

299

many precipitated minerals are also present.

300

Hadrosauridae indet. (UALVP 59614, isolated caudal vertebra centrum, mudstone, Fig 1-“11”)

301

UALVP 59614 represents a young hadrosaur that had an unfused neural arch. Three samples,

302

two from the lower half (mainly internal spongy bone) and one from the top half (mainly cortical bone

303

from the attachment point for neural arch), were removed. Of the three samples, the only one to

304

produce structures of interest was one that came from the attachment point of the neural arch. This

305

one sample produced a high number of “vessels” and “EOM” material. “Vessels” are highly flexible with

306

few minerals within the lumen. “Osteocyte” structures were not observed.

307

Prosaurolophus maximus (UALVP 55880, articulated specimen from a fluvial sandstone in Q264, Fig 1-

308

”12”)

309

Six samples in total were removed from UALVP 55880 – three from the right radius and the rest

310

from an ossified tendon. The tendon was sampled to determine if tendons produce structures with soft

311

tissue morphology, because if they do there may be less need of sampling from important skeletal

312

elements in the future. After dissolution of the tendon, small fragments of “vessel” were observed (Fig

313

2b), as well as “EOM” bundles; however, no “osteocytes” were recovered. The amount of material of

314

interest in the tendon fragments was not significant. Unlike the tendons, fragments from the radius

315

produced significant amounts of soft tissue-like material. “Vessels” are three dimensionally

316

anastomosing flexible structures like those seen in other specimens. “EOM” material is fibrous and is

317

observed in some cases surrounding “vessel” structures. Several “osteocytes” were recovered from the

318

radius, although these were poorly preserved with little of the “filipodia” remaining.

319

Gorgosaurus libratus (UALVP 10, closely associated skeleton found in sandstone with highly cemented

320

nodules, Fig 1-“13”)

321

UALVP 10 was collected from Q048 by G. Sternberg in 1921, and as a result, it was not expected

322

to produce much soft tissue-like structures. Samples were removed from the pubic shaft where the

323

bone was still encased within the sandstone. Of the three samples, all produced “EOM” structures and

324

two produced “vessels”. Several poorly preserved “osteocyte” structures were also recovered.

325

Gorgosaurus libratus (UALVP 49500, partially articulated, mud with traces of sand and concretions, Fig

326

1-“14”)

327

Three samples were taken from a larger unidentified long bone fragment collected in the quarry

328

(Q253) of UALVP 49500, a sub-adult Gorgosaurus. All three samples had significant quantities of

329

anastomosing “vessels” surrounded by “EOM” matrix. Most “vessels” contain large amounts of

330

precipitated minerals resembling hematite spheres. However, several well-preserved “vessels” contain

331

spherical red structures that resemble red blood cells (Fig 5). No “osteocyte” structures were recovered.

332

Ornithomimus edmontonicus (UALVP 52531, articulated skeleton from Q254, sample removed from

333

mud with traces of sand, Fig 1-“15”)

334

Samples were removed from the right femur of the Ornithomimus edmontonicus, UALVP 52531.

335

This specimen is important because of the presence of preserved feathers across nearly the entire body

336

(van der Reest et al., 2015). UALVP 52531 was chosen to be sampled because it was believed that since

337

feathers were preserved it would have the best chance of preserving tissue-like structures. All three

338

samples produced “vessels” and “EOM” structures. However, “osteocyte” structures were not

339

recovered.

340

Saurornitholestes langstoni (UALVP 55700, articulated skeleton from Q260, iron rich sandstone, Fig 1-

341

“16”)

342

Samples were taken from small portions of a caudal vertebrae of the Saurornitholestes, UALVP

343

55700. Despite well preserved keratin sheaths on manual and pedal claws, very little soft tissue-like

344

structures were recovered. Three small fragments of “vessels” and small portions of “EOM” structures

345

were identified. Interestingly, although there were few “vessels” and little “EOM” structures present,

346

“osteocyte” material was recovered (Fig 4a). The potential for soft tissue-like structures in the bone is

347

still high, and a long bone should be sampled for further “vessel” structures and collagen-like material.

348

Latenivenatrix mcmasterae (UALVP 55804, isolated articulated pelvis, channel sandstone, Fig 1-“17”)

349

Samples were taken from fragments of the left ilium of Latenivenatrix (UALVP 55804)(van der

350

Reest and Currie, 2017). Each sample produced small numbers of “vessels” but higher amounts of

351

“EOM” fibrous material (Fig 3b). One “vessel” in particular is preserved exceptionally well (Fig 6). The

352

fragment possesses several endothelial nuclei-like structures bulging into the lumen. Most of these

353

possess what appears to be a secondary internal membrane-like structure resembling the morphology

354

of a nucleus (Fig 6b,d). The surface texture of each nuclear membrane-like structure is textured with

355

small bumps. Also, within a branching portion of the “vessel” a single red blood cell-like object is

356

present. Within this structure, a central darker spherical region, resembling a nucleus, is observed. A

357

second organic-like structure is also observed within the lumen of the “vessel”. This structure is orange

358

in colouration and appears to be bilobed (Fig 6d). Surrounding this structure are several granules. All the

359

granules and the main lobed structures are contained within thin membrane like structures. If this

360

structure is organic in origin, the closest morphological comparison is that of a basophil, which possess a

361

two lobed internal structure with many small granules within a cellular membrane (Vellis, 2011). This

362

“vessel” is the best-preserved structure with the highest definition of all samples within the study. This

363

quality of preservation suggests that if this is original organic material, proteins may persist.

364

DISCUSSION

365

The number of specimens examined that produced soft tissue-like structures in this study was

366

significantly higher than anticipated. All samples produced soft tissue-like structures. A previous study

367

performed at Imperial Collage London (London, U.K), using specimens from the Natural History Museum

368

(London, UK) performed on specimens collected by Cutler and Sternberg also recovered similar organic-

369

like structures (Bertazzo et al., 2015). Their study of material from the Dinosaur Park Formation

370

produced similar results to this study, although they did not associate any specimens with sediment

371

type. In this current study, only UALVP 12, Panoplosaurus (Nodosauridae) did not produce any sign of

372

“vessel” structures. Despite the fact that this is an osteoderm, these elements are vascularized and so

373

should produce vessels (Burns and Currie, 2014). Therefore, there are two possible reasons that the

374

specimen did not preserve soft tissue-like structures. First, UALVP 12, was one of the first specimens

375

ever collected for the University of Alberta vertebrate paleontology collections (1920). If length of time

376

exposed to modern atmospheric conditions effects preserved soft tissues, nearly a century of exposure

377

could have destroyed any recognizable tissue structures. However, one would expect to see similar

378

effects in UALVP 10 (Gorgosaurus), which was collected only a year later. Secondly, the major difference

379

between UALVP 12 and other specimens examined is the significant amount of ironstone that encased

380

the specimen. The sample was taken from an osteoderm enclosed in a hard, dense layer of concreted

381

ironstone several millimetres thick. It is possible that as the ironstone precipitated out in the vascular

382

and Haversian canals of the cortical bone, tissue-structures were destroyed. Although nearly all other

383

specimens that possessed “vessel” structures contained some form of precipitated crystalline mineral,

384

the most common precipitates are hematite spheres. This has been suggested in previous studies as

385

having originated from hemoglobin within red blood cells (McGowan, 1983; Lee et al., 2017). The

386

amount of iron in concretions around the bone of UALVP 12, however, likely originates from another

387

source. Further investigation into the fractionation of stable isotopes may reveal the origin of this iron.

388

The second most common crystalline mineral precipitate found in “vessel” lumina is gypsum. The third

389

most common mineral observed within “vessels” is pyrite. Pyrite is identified by its cubic crystalline

390

structure that is impermeable to light. Only three specimens were observed to possess pyrite, two of

391

which had hard sandstone nodules associated with them.

392

Although the original hypothesis proposed by Schweitzer et al. (2007a) that sedimentology

393

affects the preservation of tissue structures in general does not extend to the Dinosaur Park Formation

394

of southern Alberta (Table 1), it does appear to be related to “osteocyte” structures from

395

articulated/associated specimens from sandstone deposits (Table 2). All specimens that were sampled

396

from mudstone matrix produced some form of tissue-like structures, suggesting that high porosity in the

397

surrounding sediment is not required to facilitate the drainage of the enzymes responsible for decay.

398

The hypothesis proposing that degree of articulation may play a role in tissue-like structure preservation

399

was tested by sampling specimens that were either articulated/associated or isolated in origin. For

400

tissues in general, there was no difference in either category that could be determined for preservation

401

of tissue. “Osteocytes” specifically, however, do appear to be more readily preserved in specimens that

402

were recovered as articulated/associated skeletons. It must be noted, however, that all “osteocyte”

403

structures that were recovered were poorly preserved with none possessing complete, or even partial,

404

‘filipodia’. “Osteocyte” structures are identified in these specimens by size, morphology, and the

405

presence of numerous ‘filipodia’ bases that are visible as raised protrusions from the body of the

406

structure. Most tissue-like material recovered was brown-orange in colour like other tissues reported

407

previously (Schweitzer et al., 2014, 2009b, 2007b, 2005). Wiemann et al. (2018) proposed brown-orange

408

coloured tissue is caused by oxidative crosslinking of peptides into N-heterocyclic polymers, and as a

409

result, these coloured tissues should not produce sequenceable polypeptides. However, a

410

Brachylophosaurus canadensis specimen preserving tissues with this colouring (Schroeter et al., 2017;

411

Schweitzer, 2011; Schweitzer et al., 2009b, 2007b) was discovered to preserve polypeptide fragments

412

within the walls of vessels and collagen (Cleland et al., 2015). This colouring suggests that the structures

413

that were recovered in this study are potentially altered original organics, although further chemical

414

analyses are required for confirmation.

415

CONCLUSIONS

416

This study suggests that neither sediment type nor degree of articulation have any influence on

417

the preservation of tissue-like structures in vertebrate remains from the Dinosaur Park Formation.

418

Nevertheless, the preservation of “osteocyte” structures appears to be correlated with

419

articulated/associated skeletons from sandstone deposits. These findings support, for “osteocytes” but

420

not any other tissue-like structures, the hypothesis proposed by Schweitzer et al. (2007a). This suggests

421

the depositional factors and/or taphonomic processes between different geological formations play a

422

role in tissue preservation. This is further supported by the association of “osteocytes” with articulated

423

specimens. The underlying cause behind the lower preservational rates of “osteocyte” structures is

424

currently unknown. However, the smaller filipodia found to have been destroyed on virtually all those

425

observed suggests that if these are indeed original cells, the phospholipid bilayer has been compromised

426

and thus the cells are destroyed. Endothelial cells and extracellular organic matrix that contain collagen

427

and elastin may be more resilient and thus survive longer in the geological record. Further investigations

428

into different paleoenvironmental conditions are required to shed more light on favourable conditions

429

for tissue-like structure preservation.

430 431 432

ACKNOWLEDGMENTS

433

Dr. Felix Sperling (University of Alberta) provided access to a laminar flow hood in which the dissolution

434

work and slide mounting was conducted, and Alberto Reyes made his laboratory available for the

435

dissolution of the samples. The Dinosaur Research Institute gave a student research grant to cover lab

436

supplies for this study. Two anonymous reviewers provided insightful comments that improved the

437

manuscript.

438

LITERATURE CITED

439

Armitage, M.H., Anderson, K.L., 2013. Soft sheets of fibrillar bone from a fossil of the supraorbital horn

440

of the dinosaur Triceratops horridus. Acta Histochem. 115, 603–608.

441

https://doi.org/10.1016/j.acthis.2013.01.001

442 443 444

Asara, J.M., Schweitzer, M.H., Freimark, L.M., Phillips, M., Cantley, L.C., 2007. Protein Sequences from Mastodon and Tyrannosaurus Rex Revealed by Mass Spectrometry. Science (80-. ). 316, 280–285. Benson, R.B.J., Campione, N.E., Carrano, M.T., Mannion, P.D., Sullivan, C., Upchurch, P., Evans, D.C.,

445

2014. Rates of Dinosaur Body Mass Evolution Indicate 170 Million Years of Sustained Ecological

446

Innovation on the Avian Stem Lineage. PLoS Biol. 12. https://doi.org/10.1371/journal.pbio.1001853

447

Bertazzo, S., Maidment, S.C.R., Kallepitis, C., Fearn, S., Stevens, M.M., Xie, H.N., 2015. Fibres and cellular

448

structures preserved in 75-million-year-old dinosaur specimens. Nat. Commun. 6, 1–8.

449

https://doi.org/10.1038/ncomms8352

450 451 452

Buckland, W., 1824. Notice of the Megalosaurus or great fossil lizard of Stonesfield. Trans. Geol. Scociety 1, 390–396. Burns, M.E., Currie, P.J., 2014. External and internal structure of ankylosaur (Dinosauria, Ornithischia)

453

osteoderms and their systematic relevance. J. Vertebr. Paleontol. 34, 835–851.

454

https://doi.org/10.1080/02724634.2014.840309

455

Campione, N.E., Evans, D.C., Brown, C.M., Carrano, M.T., 2014. Body mass estimation in non-avian

456

bipeds using a theoretical conversion to quadruped stylopodial proportions. Methods Ecol. Evol.

457

n/a-n/a. https://doi.org/10.1111/2041-210X.12226

458

Cleland, T.P., Schroeter, E.R., Zamdborg, L., Zheng, W., Lee, J.E., Tran, J.C., Bern, M., Duncan, M.B.,

459

Lebleu, V.S., Ahlf, D.R., Thomas, P.M., Kalluri, R., Kelleher, N.L., Schweitzer, M.H., 2015. Mass

460

Spectrometry and Antibody-Based Characterization of Blood Vessels from Brachylophosaurus

461

canadensis. J. Proteome Res. 14, 5252–5262. https://doi.org/10.1021/acs.jproteome.5b00675

462

Currie, P.J., Holmes, R.B., Ryan, M.J., Coy, C., 2016. A juvenile chasmosaurine ceratopsid (Dinosauria,

463

Ornithischia) from the Dinosaur Park Formation, Alberta, Canada. J. Vertebr. Paleontol. 36.

464

https://doi.org/10.1080/02724634.2015.1048348

465

Jarvis, E.D., Mirabab, S., Aberer, A.J., Zhang, G., Li, C., Li, Q., Li, B., Larkin, D.M., Lee, C., Storz, J.F.,

466

Antunes, A., Greenwold, M.J., Meredith, R.W., Odeen, A., Cui, J., Zhou, Q., Xu, L., Pan, H., Wang, Z.,

467

Jin, L., Zhang, P., Hu, H., Yang, W., Hu, J., Xiao, J., Yang, Z., Liu, Y., Xie, Q., Yu, H., Lian, J., Wen, P.,

468

Zhang, F., Li, H., Zeng, Y., Xiong, Z., Liu, S., Zhou, L., Huang, Z., An, N., Wang, J., Zheng, Q., Xiong, Y.,

469

Wang, G., Wang, B., Wang, J., Fan, Y., da Fonseca, R.R., Alfaro-Nunez, A., Schubert, M., Orlando, L.,

470

Mourier, T., Howard, J.T., Ganapathy, G., Pfenning, A., Whitney, O., Rivas, M. V., Hara, E., Smith, J.,

471

Farre, M., Narayan, J., Slavov, G., Romanov, M.N., Borges, R., Machado, J.P., Khan, I., Springer,

472

M.S., Gatesy, J., Hoffmann, F.G., Opazo, J.C., Hastad, O., Sawyer, R.H., Kim, H., Kim, K.-W., Kim, H.J.,

473

Cho, S., Li, N., Huang, Y., Bruford, M.W., Zhan, X., Dixon, A., Bertelsen, M.F., Derryberry, E., Warren,

474

W., Wilson, R.K., Li, S., Ray, D.A., Green, R.E., O’Brien, S.J., Griffin, D., Johnson, W.E., Haussler, D.,

475

Ryder, O.A., Willerslev, E., Graves, G.R., Alstrom, P., Fjeldsa, J., Mindell, D.P., Edwards, S. V., Braun,

476

E.L., Rahbek, C., Burt, D.W., Houde, P., Zhang, Y., Yang, H., Wang, J., Gilbert, M.T.P., Wang, J., Ye,

477

C., Liang, S., Yan, Z., Zepeda, M.L., Campos, P.F., Velazquez, A.M. V., Samaniego, J.A., Avila-Arcos,

478

M., Martin, M.D., Barnett, R., Ribeiro, A.M., Mello, C. V, Lovell, P. V, Almeida, D., Maldonado, E.,

479

Pereira, J., Sunagar, K., Philip, S., Dominguez-Bello, M.G., Bunce, M., Lambert, D., Brumfield, R.T.,

480

Sheldon, F.H., Holmes, E.C., Gardner, P.P., Steeves, T.E., Stadler, P.F., Burge, S.W., Lyons, E., Smith,

481

J., McCarthy, F., Pitel, F., Rhoads, D., Froman, D.P., 2014. Comparative genomics reveals insights

482

into avian genome evolution and adaptation. Science (80-. ). 346, 1311–1320.

483

https://doi.org/10.1126/science.1251385

484 485 486

Kaye, T.G., Gaugler, G., Sawlowicz, Z., 2008. Dinosaurian soft tissues interpreted as bacterial biofilms. PLoS One 3, 1–7. https://doi.org/10.1371/journal.pone.0002808 Lee, Y.-C., Chiang, C.-C., Huang, P.-Y., Chung, C.-Y., Huang, T.D., Wang, C.-C., Chen, C.-I., Chang, R.-S.,

487

Liao, C.-H., Reisz, R.R., 2017. Evidence of preserved collagen in an Early Jurassic sauropodomorph

488

dinosaur revealed by synchrotron FTIR microspectroscopy. Nat. Commun. 8, 14220.

489

https://doi.org/10.1038/ncomms14220

490 491 492 493 494

McGowan, C., 1983. The Successful Dragons: A Natural History of Extinct Reptiles. Samuel Stevens and Company, Toronto, Ontario, Canada. Morlon, H., Parsons, T.L., Plotkin, J.B., 2011. Reconciling molecular phylogenies with the fossil record. Proc. Natl. Acad. Sci. U. S. A. 108, 16327–32. https://doi.org/10.1073/pnas.1102543108 Peterson, J.E., Lenczewski, M.E., Scherer, R.P., 2010. Influence of microbial biofilms on the preservation

495

of primary soft tissue in fossil and extant archosaurs. PLoS One 5.

496

https://doi.org/10.1371/journal.pone.0013334

497

Ryan, M.J., Evans, D.C., Currie, P.J., Loewen, M.A., 2014. A new chasmosaurine from northern Laramidia

498

expands frill disparity in ceratopsid dinosaurs. Naturwissenschaften 101, 505–512.

499

https://doi.org/10.1007/s00114-014-1183-1

500

San Antonio, J.D., Schweitzer, M.H., Jensen, S.T., Kalluri, R., Buckley, M., Orgel, J.P.R.O., 2011. Dinosaur

501

peptides suggest mechanisms of protein survival. PLoS One 6.

502

https://doi.org/10.1371/journal.pone.0020381

503

Sato, H., Murray, A.M., Vernygora, O., Currie, P.J., 2018. A rare, articulated sturgeon (Chondrostei:

504

Acipenseriformes) from the Upper Cretaceous of Dinosaur Provincial Park, Alberta, Canada. J.

505

Vertebr. Paleontol. 0, 1–15. https://doi.org/10.1080/02724634.2018.1488137

506

Schroeter, E.R., Dehart, C.J., Cleland, T.P., Zheng, W., Thomas, P.M., Kelleher, N.L., Bern, M., Schweitzer,

507

M.H., 2017. Expansion for the Brachylophosaurus canadensis Collagen I Sequence and Additional

508

Evidence of the Preservation of Cretaceous Protein. J. Proteome Res. 16, 920–932.

509

https://doi.org/10.1021/acs.jproteome.6b00873

510 511

Schweitzer, M.H., 2011. Soft Tissue Preservation in Terrestrial Mesozoic Vertebrates. Annu. Rev. Earth Planet. Sci. Vol 39 39, 187–216. https://doi.org/10.1146/annurev-earth-040610-133502

512

Schweitzer, M.H., Suo, Z., Avci, R., Asara, J.M., Allen, M. a, Arce, F.T., Horner, J.R., 2007a. Analyses of soft

513

tissue from Tyrannosaurus rex suggest the presence of protein. Science (80-. ). 316, 277–280.

514

https://doi.org/10.1126/science.1138709

515

Schweitzer, M.H., Watt, J.A., Avci, R., Knapp, L., Chiappe, L., Norell, M., Marshall, M., 1999. Beta-keratin

516

specific immunological reactivity in feather-like structures of the cretaceous alvarezsaurid,

517

Shuvuuia deserti. J. Exp. Zool. 285, 146–157. https://doi.org/10.1002/(SICI)1097-

518

010X(19990815)285:2<146::AID-JEZ7>3.0.CO;2-A

519

Schweitzer, M.H., Wittmeyer, J.L., Horner, J.R., 2007b. Soft tissue and cellular preservation in vertebrate

520

skeletal elements from the Cretaceous to the present. Proc. Biol. Sci. 274, 183–97.

521

https://doi.org/10.1098/rspb.2006.3705

522

Schweitzer, M.H., Wittmeyer, J.L., Horner, J.R., Jan, K., Science, S., Series, N., Gut, T., Tube, I., Niiio-like,

523

E., 2005. Soft-Tissue Vessels and Cellular Preservation in Tyrannosaurus rex. Science (80-. ). 307,

524

1952–1955.

525

Schweitzer, M.H., Zheng, W., Cleland, T.P., Bern, M., 2013. Molecular analyses of dinosaur osteocytes

526

support the presence of endogenous molecules. Bone 52, 414–423.

527

https://doi.org/10.1016/j.bone.2012.10.010

528

Schweitzer, M.H., Zheng, W., Cleland, T.P., Goodwin, M.B., Boatman, E., Theil, E., Marcus, M.A., Fakra,

529

S.C., 2014. A role for iron and oxygen chemistry in preserving soft tissues, cells and molecules from

530

deep time. Proc. R. Soc. B 281, 20132741. https://doi.org/10.1098/rspb.2013.2741

531

Schweitzer, M.H., Zheng, W., Organ, C.L., Avci, R., Suo, Z., Freimark, L.M., Lebleu, V.S., Duncan, M.B.,

532

Vander Heiden, M.G., Neveu, J.M., Lane, W.S., Cottrell, J.S., Horner, J.R., Cantley, L.C., Kalluri, R.,

533

Asara, J.M., 2009a. Biomolecular characterization and protein sequences of the Campanian

534

hadrosaur Brachylophosaurus canadensis. Science (80-. ). 324, 626–629.

535

Schweitzer, M.H., Zheng, W., Organ, C.L., Avci, R., Suo, Z., Freimark, L.M., Lebleu, V.S., Duncan, M.B.,

536

Vander Heiden, M.G., Neveu, J.M., Lane, W.S., Cottrell, J.S., Horner, J.R., Cantley, L.C., Kalluri, R.,

537

Asara, J.M., 2009b. Biomolecular characterization and protein sequences of the Campanian

538

hadrosaur B. canadensis. Science (80-. ). 324, 626–31. https://doi.org/10.1126/science.1165069

539

Ullmann, P. V., Pandya, S.H., Nellermoe, R., 2019. Patterns of soft tissue and cellular preservation in

540

relation to fossil bone tissue structure and overburden depth at the Standing Rock Hadrosaur Site,

541

Maastrichtian Hell Creek Formation, South Dakota, USA. Cretac. Res. 99, 1–13.

542

https://doi.org/10.1016/j.cretres.2019.02.012

543

Van Der Reest, A.J., Currie, P.J., 2017. Troodontids ( Theropoda ) from the Dinosaur Park Formation ,

544

Alberta , with a description of a unique new taxon : implications for deinonychosaur diversity in

545

North America. Can. J. Earth Sci. 935, 919–935. https://doi.org/10.1139/cjes-2017-0031

546

Vellis, P.K., 2011. Preface, in: Vellis, P.K. (Ed.), Cell Biology Research Progress Series: Basophil

547

Granulocytes. Nova Science Publishers, Inc., New York, New York, USA, pp. ix–xiii.

548

Wiemann, J., Fabbri, M., Yang, T.-R., Stein, K., Sander, P.M., Norell, M.A., Briggs, D.E.G., 2018.

549

Fossilization transforms vertebrate hard tissue proteins into N-heterocyclic polymers. Nat.

550 551 552

Commun. 9, 4741. https://doi.org/10.1038/s41467-018-07013-3

553

FIGURE CAPTIONS

554

Figure 1. Map of Alberta and Dinosaur Provincial Park indicating where specimens were recovered. See

555

specimen descriptions for corresponding numbers. Orange numbers indicate specimens recovered from

556

sandstone and blue from mudstone.

557 558

Figure 2. Select blood “vessel” structures. a) UALVP 55900, Styracosaurus. b) UALVP 55880,

559

Prosaurolophus. c) UALVP 55926, cf. Chasmosaurus sp. d) UALVP 56607, cf. Euoplocephalus. e) UALVP

560

59616, Leidyosuchus. f) UALVP 56596, Anchiacipenser acanthaspis.

561 562

Figure 3. Extracellular organic matrix-like structures. a) UALVP 59616, Leidyosuchus. b) UALVP 55804,

563

Latenivenatrix. Arrows identify “EOM” structures, grey fibrous material is undissolved bioapatite.

564 565

Figure 4. Select “osteocyte” structures. a) UALVP 55700, Saurornitholestes. b) UALVP 52613,

566

Chasmosaurus belli (juvenile). c) UALVP 10, Panoplosaurus. Note that two “osteocyte” structures remain

567

in situ within the bioapatite bone matrix. d) UALVP 59616, Leidyosuchus.

568 569

Figure 5. Anastomosing “vessel” network in UALVP 49500, Gorgosaurus subadult. Grey fibrous material

570

is remaining bioapatite bone matrix. a) Overview of structure indicating enlarged images in b and d. b)

571

enlarged area indicated in a. c) Enlarged portion outlined in b, black arrows point to red blood cell-like

572

structures within the lumen of the “vessel” structure. d) Enlarged area as indicated in a, black arrow

573

points to red blood cell-like structure.

574

575

Figure 6. Portion of “vessel” structure from the ilium of Latenivenatrix. a) Overview of “vessel” indicating

576

where enlarged areas in a-c are from. b) Portion of main length of the “vessel” structure. Black arrows

577

indicate the nucleus-like structures bulging into the lumen of the “vessel”. c) Red blood cell-like

578

structure remaining within the “vessel”. Note dark centre, suggesting possible nucleus-like structure. d)

579

Large bilobed object possessing similar morphology to a basophil. Black arrows indicate endothelial

580

nuclei-like structures.

581

582 583 584 585 586

Table 1. List of samples derived from specimens and types of soft-tissue preservations noted for each. Sediment: M, Mudstone; S, Sandstone. Association: AR, Articulated skeleton; AS, Associated skeleton; I, Isolated bones; M, Micro vertebrate site. Paired grey rows indicate specimens that originated from the same UALVP specimen. Abundance of tissue-like structures, X, rare; XX, common; XXX, abundant. Precise numbers are exact numbers found in 40 µL of dissolution fluid.

Taxon Actinopterygii

Element

Specimen number

Year Collected

Vessels

EOM

Anchiacipenser

osteoderm

56592

2017

X

ungual

59613

2016

osteoderm parasphenoid, dentary

59615

Osteocytes

Sediment

Association

XX

S

AR

XXX

XX

M

M

2016

XXX

XXX

M

M

59616

2016

XXX

XX

M

AS

cranium

56607

2018

XXX

X

S

I

osteoderm

12

1920

S

I

Mercuriceratops

squamosal

54559

2012

S

AS

Chasmosaurus

cranium

52613

2010

XXX

XX

S

AR

cf. Chasmosaurus

tibia

55926

2014

XXX

XXX

M

AR

Centrosaurus

cranium

55794

2008

XX

XX

S

I

Styracosaurus

cranium

52612

2010

XX

XX

S

AR

Styracosaurus

humerus

55900

2015

XXX

1

S

AR

scapula

55900

2015

XX

3

S

AR

Hadrosauridae indet.

caudal centrum

59614

2016

XX

XX

M

I

Prosaurolophus

tendon

55880

2015

X

X

S

AR

radius

55880

2015

XXX

XX

1

S

AR

Gorgosaurus

pubis

10

1921

X

X

1

S

AS

Gorgosaurus

femur

49500

2008

XX

XX

M

AS

femur, pedal phalanx

52531

2009

XX

XX

M

AR

caudal centrum

55700

2014

X

X

S

AR

ilium

55804

2014

XX

XX

S

I

Testudines Basilemys sp. Archosauria Crocodylia indet. Leidyosuchus

2

Ankylosauridae cf. Euoplocephalus Nodosauridae Panoplosaurus

X

2

Ceratopsidae XXX

XX 1

Hadrosauridae

Tyrannosauridae

Ornithomimidae Ornithomimus Dromaeosauridae Saurornitholestes

1

Troodontidae Latenivenatrix

587 588

Table 2. Relationship between sediment type and degree of articulation for preservation rates of “osteocyte” structures. Sediment type Sandstone Mudstone Articulated 62.5% 25% Degree of articulation Isolated 33% 0%

589

Figure S1. Representative “vessels” recovered from bone samples from the Dinosaur Park Formation. a) Anchiacipencer, UALVP 56592. b) Basilemys, UALVP 59613. c) Crocodylia indet., UALVP 59615. d) Leidyosuchus, UALVP 59616. e) cf. Euoplocephalus UALVP 56607. f) Mercuriceratops, UALVP 54559. g) Chasmosaurus, UALVP 52613. h) cf. Chasmosaurus, UALVP 55926. i) Centrosaurus, UALVP 55794. j) Styracosaurus, UALVP 52612. k) Styracosaurus, UALVP 55900. l) Hadrosauridae indet, UALVP 59614. m) Prosaurolophus, UALVP 55880. n) Gorgosaurus, UALVP 10. o) Gorgosaurus, UALVP 49500. p) Ornithomimus, UALVP 52531. q) Saurornitholestes. UALVP 55700. r) Latenivenatrix, UALVP 55804.

Figure S2. Representative “osteocytes” recovered from bone samples from the Dinosaur Park Formation. a) Leidyosuchus, UALVP 59616. b) Panoplosaurus, UALVP 12. c) Chasmosaurus, UALVP 52613. d) Styracosaurus, UALVP 55900. e) Prosaurolophus, UALVP 55880. f) Gorgosaurus, UALVP 10. g) Saurornitholestes. UALVP 55700.

Figure S3. Representative mineral precipitates observed in vessels. a) gypsum in regular light. b) gypsum in cross polarized light. c) spheroid crystals inside “vessel” matching hematite morphology. d) cubic crystals inside “vessel” matching pyrite morphology.

Authors contributions Aaron van der Reest: Conceptualization, Methodology, Investigation, Writing, Visualization, Funding acquisition. Philip Currie: Review & Editing, Supervision, Funding acquisition.

Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: