22 In the summer of 1956, Romer organized a field season to search for early tetrapods in Nova Scotia with financial support from the National Science Foundation. The target was the province’s Carboniferous strata, which, with the exception of the world-famous Joggins section (Dawson 1882; Calder et al. 2006), had produced only two tetrapod specimens from an underground coal mine and vertebrate trackways and occasional scraps of bone from its scenic coastal exposures (Owen 1854; Watson 1926; Beaumont 1977; Sarjeant and Mossman 1978) (Fig. 2). Don would later observe that “[t]his Acadian reconnaissance shaped my field activities for decades to come” (Baird 1997, p. 24).

23 Romer divided the MCZ field party into three groups to work the known fossil vertebrate localities and prospect for new deposits. In addition to Romer and his wife Ruth, participants comprised Don and Lucy with their son Andy, Romer’s research assistant Nelda E. Wright, and paleontological preparators Richard Van Frank, Arnie Lewis and James A. Jensen.

24 The party first worked at Joggins before going separate ways. During July, Don and Lucy prospected outcrops at Parrsboro and Horton Bluff between rain showers and visits to mostly non-productive coastal exposures, then returned to Parrsboro to make latex molds of trackways as the weather improved. Late in the month, Lucy discovered an upright lycopsid stump at Parrsboro that contained a small temnospondyl skull, the first tetrapod fossil ever found at a locality known for its trackways. Milner (1982) considered this specimen (MCZ 2774) possibly referable to Dendrerpeton, which had first been discovered by Dawson and Lyell at Joggins in 1852 (Lyell and Dawson 1853).

25 In early August, the MCZ field party reunited briefly at Crystal Cliffs near Antigonish before moving on to Cape Breton Island (Fig. 3). The Bairds worked along the northwest coast of the island, finding and molding trackways and collecting fragmentary remains of Carboniferous vertebrates. The most significant locality they discovered was an exposure of freshwater limestone in the Pomquet Formation (Mabou Group; Mississippian: Serpukhovian: Arnsbergian) at Grand Étang that would yield fish and fragmentary tetrapod remains (Baird 1978a; Holmes et al. 1995). By mid-August, Don and Lucy turned east to meet up with the others, who had just discovered tetrapod remains in upright lycopsid tree stumps near Florence, on the northeastern side of the island (Fig. 2).

26 The Florence stump assemblage, discovered by Arnie Lewis, would prove to be the most important find of the 1956 fieldtrip to Nova Scotia, even though it was too young to sate Romer’s desire for fossils of genuinely ancient tetrapods. The fossil locality is in an abandoned surface mine (Dominion Coal Company No. 7) in the Sydney Mines Formation (Cumberland Group; Pennsylvanian: Moscovian), about 2 km north of Florence. Younger in age than the famous assemblage from the lycopsid stumps at Joggins (Pennsylvanian: Bashkirian: Langsettian), vertebrate fossils from Florence generally are better preserved than those from Joggins and more or less coeval with the Linton cannel coal, so that meaningful comparisons between the Florence and Linton assemblages could be made. Although collecting at the strip mine proved challenging — the bone-bearing tree stumps were exposed in a high wall of a strip cut, the floor of which is flooded (Fig. 4). Don and others would repeatedly revisit the locality in later years to recover additional fossils. Carroll (1967, 1969), Reisz (1972), Rieppel (1980), Klembara (1985), and Godfrey and Holmes (1995) have published detailed studies of individual tetrapod taxa from Florence.

27 On the return trip from the Sydney area, the Bairds continued to find remains of Carboniferous freshwater fishes and abundant tetrapod trackways. Don carefully recorded stratigraphic observations and locality data that he would use for future fieldtrips. He realized that the geology of Nova Scotia was much more complex than anything he had studied previously, and although he had assembled the appropriate literature and maps for the trip, he found the generalized nature of earlier geological reports insufficient for his research.

28 During the 1959 field season Don undertook fieldwork with William F. Take, then Curator of Geology at the Nova Scotia Museum of Science in Halifax. They prospected Carboniferous exposures where, in the Port Hood Formation (Pennsylvanian: Langsettian) near Port Hood in Inverness County, they found the anterior portion of a poorly preserved tetrapod skeleton (YPM VPPU 016982). Don and Robert L. Carroll (McGill University) would later describe this specimen as Romeriscus periallus (Baird and Carroll 1967). Interpreting it as a member of the Limnoscelidae, which were then regarded as stem reptiles, they introduced the specimen as the oldest-known reptile. A subsequent re-examination of this fossil by Laurin and Reisz (1992) failed to identify any diagnostic features in support of this assignment, and the affinities of Romeriscus remain uncertain.

29 Another “discovery” of the 1959 Nova Scotia trip was Eldon George, proprietor of the Parrsboro Rock and Mineral Shop. Eldon has been an exceptional collector and, during Don’s many trips to Parrsboro, they usually would prospect the tremendous coastal exposures of the area together. Don and Eldon collected skull roofs of haplolepidid fishes and a fin-spine of a small acanthodian fish (Gyracanthus) that Don later described in detail (Baird 1962a, 1978a). Don introduced Eldon to collecting and identifying Carboniferous tetrapod footprints and provided latex for him to mold select trackways.

30 During the Nova Scotia trip of 1960, Don continued to work closely with Bill and Jane Take. On their tour of Carboniferous exposures, Don found a skull fragment near Point Edward on Cape Breton Island that he described as Spathicephalus pereger (YPM VPPU 017182; Baird 1962b). It was the first North American record of this distinctive stem tetrapod, the type species of which, Spathicephalus mirus, is known from the Rumbles Ironstone (Mississippian: Serpukhovian: Pendleian) of the Loanhead District near Edinburgh (Scotland). Spathicephalus has a flat, greatly expanded skull with numerous small, chisel-shaped teeth (Beaumont and Smithson 1998). At Point Edward, vertebrates occur in a freshwater limestone and overlying mudstone of the Mabou Group (Mississippian: Serpukhovian: Chokierian). Although fragmentary fish remains had first been discovered there in 1913 (Carroll et al. 1972), the first tetrapod from Point Edward was collected in 1956 by Romer, who described a partial lower jaw (MCZ 2772) as ?Pholiderpeton bretonense (Romer 1958). Fifteen years later, Don would find, and recognize instantly, the counterpart of Romer’s specimen, which is now regarded as an indeterminate proterogyrinid anthracosaur (Holmes et al. 1995). In 1968, Don collected an additional tetrapod specimen at Point Edward (YPM VPPU 020100), which is represented by an articulated pectoral girdle and trunk segment of a colosteid stem tetrapod similar to Greererpeton burkemorani, known mainly from the more or less coeval Mauch Chunk Group of West Virginia (Smithson 1982; Godfrey 1989).

31 Don first visited Horton Bluff (Fig. 2) in 1956 and made numerous subsequent visits to this Mississippian locality, the source of the earliest-known Carboniferous tetrapod remains in North America. The Horton Bluff Formation has been palynologically dated as Tournaisian (Ivorian) (Utting et al. 1989; Rygel et al. 2006). Despite the great potential of this site, tetrapod fossils are very rare and fragmentary, as well as difficult to prepare. Moreover, bones of tetrapodomorph fishes occur in the same deposit, appear to be more common, and often are difficult to distinguish from those of basal tetrapods. In their discussion of the vexing nature of the tetrapod remains from Horton Bluff, Clack and Carroll (2000, p. 1042) aptly remarked, “Rarely have so many people worked so hard and so long for so few fossils.”

32 The exposures of the Horton Bluff Formation at Horton Bluff (Blue Beach) had long been known for the occurrence of Mississippian tetrapod footprints, some of them attaining a print width of 20 cm (Sarjeant and Mossman 1978). Although Logan’s original 1841 material (Hylopus logani Dawson 1882) represents Schwimmspuren (swimming traces), later collecting by Don and others documented an array of tetrapod tracks (Carroll et al. 1972). This material includes well preserved pentadactyl and tetradactyl trackways of at least six different forms of both small and rather large tetrapods that were capable of locomotion on land (Smithson et al. 2012).

33 Don first discovered tetrapod bones at Horton Bluff in 1966. Although he had some ideas concerning the affinities of his tantalizing finds (Carroll et al. 1972), the absence of more complete, articulated remains from Horton Bluff, as well as the dearth of comparative material from coeval strata elsewhere, prevented definitive study. Only in the last two decades has the situation begun to improve with the discovery of well-preserved, articulated Tournaisian-age tetrapod fossils in Scotland and the recovery of additional (if still only dissociated) skeletal remains at Horton Bluff. Clack (2012) most recently reviewed the tetrapod fossils from Horton Bluff. She recognized a previously unexpected diversity of taxa, including a possible reptiliomorph, a possible colosteid stem tetrapod, and a form resembling the famous stem tetrapod Ichthyostega from Late Devonian (Famennian) strata of East Greenland.

34 Although Don’s final published contribution to the paleontology of the Maritimes is based on material from Prince Edward Island (Calder et al. 2004), the findings relate broadly to Nova Scotia and illustrate the serendipitous and often prolonged nature of Don’s endeavors. While visiting Malpeque Bay, Prince Edward Island, in 1984, Elliott B. Urdang found tetrapod footprints and brought them to the attention of John H. Ostrom at Yale University, who, in turn, referred the inquiry to Don. Although he had never visited the island, Don recognized the trackways as the first footprint record for the province and encouraged Urdang to obtain further documentation, which he did in 1987. Once Don had received adequate photographs and peels of the specimens, he wrote an analysis of the material and added it to his thick file of unpublished data on Permo-Carboniferous tetrapod trackways. More than a decade would pass before John H. Calder (Nova Scotia Department of Natural Resources, Halifax) became interested in the project after consulting with Don on Nova Scotian trackways. John combined his assessment of the age of the Pictou Group in the Malpeque Bay area with Don’s information to conclude that the tracks represented a latest Pennsylvanian to earliest Permian occurrence of the ichnogenera Notalacerta and Gilmoreichnus, most likely made by a basal reptile and a basal synapsid, respectively.

35 Otherwise, aside from comments in Carroll et al. (1972), the information that Don amassed on Carboniferous footprints from Nova Scotia has largely gone unpublished. He freely shared his considerable insights with other workers and often was acknowledged for his assistance. After Don’s death, his large collection of trackway latex peels, including material from both the Paleozoic and Mesozoic and from numerous North American localities, was donated to the Peabody Museum of Natural History, along with his detailed catalogue.

36 After Don had assumed his curatorial duties at Princeton he undertook a three-day visit to the Nova Scotia Museum of Science in Halifax in 1958. On a snowy day in late November, Bill Take introduced Don to fossil localities that he had recently discovered in the Late Triassic (Carnian) Wolfville Formation (“Evangeline Member”) along the shores of the Minas Basin in the Bay of Fundy, including Evangeline Beach (Fig. 2). Don found a small incomplete skull that 40 years later became the holotype of a new taxon of procolophonid parareptile, Acadiella psalidodon (Sues and Baird 1998; NSM 996GF69.1). The informal stratigraphic terms “Economy Member” and “Evangeline Member” for units within the Wolfville Formation are used here in lieu of “lower Wolfville” and “middle Wolfville” (Olsen et al. 1989), and thus are cited in quotation marks. The need for distinguishing these two members was initially driven by Don’s discovery of different Triassic faunal assemblages in different areas around the Minas Basin. The early Mesozoic succession of the Fundy Group of the Newark Supergroup comprises (in ascending order) the Wolfville Formation, the Blomidon Formation, the North Mountain Basalt, and the McCoy Brook Formation (Olsen et al. 1989).

37 Don returned to Nova Scotia in June 1959 and collected additional Triassic vertebrate fossils with Bill Take. With the help of Lloyd Duncanson, they reached by boat the northwest side of Boot Island, near the mouth of the Avon River east of Wolfville. From beach exposures of the “Evangeline Member” accessible only at low tide (Fig. 5), Don collected a partial procolophonid skull that would later become the holotype of Scoloparia glyphanodon (Sues and Baird 1998; NSM 996GF83.1) and an impression in sandstone of an incomplete dentary of a rhynchosaur — the first record of these presumably herbivorous archosauromorph reptiles from North America (Baird 1963). Following the 1959 field season, Don and Bill published an abstract with a preliminary list of the Wolfville tetrapod assemblage and a provisional correlation scheme for the Triassic of Nova Scotia and parts of New Jersey and the Connecticut Valley (Baird and Take 1959). In 1960, they spent considerable time prospecting outcrops of the Wolfville Formation at Burntcoat, near Noel, Hants County (Fig. 2), and found a maxilla that would later become the holotype of a third taxon of procolophonid parareptile, Haligonia bolodon (Sues and Baird 1998; NSM 996GF74.1), along with other fragmentary tetrapod remains.

38 Unfortunately the collaboration between Bill Take and Don did not continue after 1960. An underlying reason was a disagreement over the final disposition of the vertebrate fossils: Don proposed to deposit all the Triassic vertebrate fossils at the Nova Scotia Museum of Science in Halifax and take the Carboniferous material to Princeton. Take wanted all of the specimens returned to the care of his museum. When Don returned to Halifax in 1963, he was dismayed to find many Triassic specimens still unwrapped, few labeled, and no cataloguing system in place. Mindful of the scientific value of this material, he arranged with the director of the museum a long-term loan of all fossils to Princeton for proper curation and study. Take, who was away at the time, never forgave Don for this action.

39 In July 1965, Bob Carroll, while collecting with Don, found a fragment of a very large dentary (YPM VPPU 019190) at Burntcoat, which proved to be the first record of a traversodont cynodont from North America. Hopson (1984) tentatively referred it to the Late Triassic genus Scalenodontoides, previously known only from Lesotho, but recognized it as a distinct species ?Scalenodontoides plemmyridon. Later, the recovery of a large upper molariform tooth by a field party led by one of us (H-DS) prompted a reassessment of this taxon and the naming of the new genus Arctotraversodon (Sues et al. 1992). Don continued to collect vertebrate fossils from the “Evangeline Member” of the Wolfville Formation until his last visit to Nova Scotia in 1985. Later two of us (H-DS and PEO) successfully continued this work, ably assisted for a number of years by the local fossil collector George P. Hrynewich.

40 Don also discovered an assemblage of fragmentary remains of tetrapods from the “Economy Member” of the Wolfville Formation near Lower Economy in Colchester County in 1966 (Baird 1986). This occurrence is distinctly different in faunal composition from the better-known tetrapod assemblage from the overlying “Evangeline Member” and may be significantly older, possibly as old as early Middle Triassic. This is based on the fact that in North America and Morocco tetrapod assemblages with only nonmetoposaurid temnospondyls are found in strata of Middle Triassic age and are overlain by Late Triassic strata with metoposaurids (Sues and Olsen, personal observations). However, in comparison with the well-dated tetrapod assemblages from the German Triassic, there are no taxa that actually constrain the age more precisely than Middle to early Late Triassic.

41 In October 1973, Don and one of us (PEO) were prospecting exposures in the Wolfville area when Don spotted a sandstone slab covered with natural casts of several three-toed, rather bird-like foot impressions associated with small hand imprints. Don realized immediately that this find belonged to a new ichnogenus that he had identified previously in the Newark and Gettysburg basins of New Jersey and Pennsylvania, based in part upon what he had described separately as Anchisauripus milfordensis and Grallator sulcatus in his doctoral dissertation (Baird 1957) and in part on material he had collected at Milford, New Jersey in the 1960s. Olsen and Baird (1986) later named this new ichnogenus Atreipus, and the Nova Scotian slab (YPM VPPU 021713 B) became the holotype of the new ichnospecies Atreipus acadianus. Don had coined the generic name in honor of Atreus Wanner, a school teacher from York, Pennsylvania, who had discovered a beautifully preserved set of these distinctive tracks in the Gettysburg basin nearly a century earlier (Wanner 1889). Olsen and Baird (1986) argued that either an early ornithischian dinosaur or a non-dinosaurian archosaur converging on ornithischians in the structure of its feet left the distinctive tracks of Atreipus. The track makers now are thought to have been close relatives of dinosaurs such as Silesauridae (Olsen et al. 2005), the first skeletal remains of which were only discovered quite recently (Irmis et al. 2007).

42 In August 1976, a field party comprising Don, PEO, and Princeton technician John R. “Jack” Horner (who would go on to become one of the world’s leading experts on Cretaceous dinosaurs) made history when PEO discovered what would be the first unequivocal skeletal remains of dinosaurs known from the Maritimes. On a rainy August day, while prospecting exposures of the uppermost Triassic and lowermost Jurassic McCoy Brook Formation at Wasson Bluff near Parrsboro (Fig. 2), PEO collected numerous white, unusually large, angular bits of what appeared to be bone originating from a basalt-clast-bearing sandstone bed between eolian sandstones. There was initially some doubt concerning the identity of these pieces because zeolite minerals often superficially resemble fossil bone and are abundant in the basalts, and bone had not been previously reported from this area. A few evenings later, however, they assembled the fragments into a three-dimensional jigsaw puzzle that Don identified as a cervical vertebra of an early sauropodomorph (“prosauropod”) dinosaur (YPM VPPU 022196). During two subsequent visits to the site, the field party collected a partial femur, some caudal vertebrae, and a phalanx from the foot, all of which belong to sauropodomorph dinosaurs. The Wasson Bluff material is currently the oldest known skeletal material of dinosaurs from eastern North America. As Don would observe, however, Late Triassic tracks of dinosaurs from Nova Scotia predate these finds in geological age (Olsen et al. 1989, 2005). Don and PEO collected tetrapod footprints and fish from other latest Triassic localities in Nova Scotia, with Baird being the first to discover the important locality for semionotid fishes at Five Islands Provincial Park in 1976. Eldon George, in addition to his discoveries of Pennsylvanian fossils, also collected many excellently preserved footprints of dinosaurs and other reptiles from the McCoy Brook Formation, mostly at McKay Head and Blue Sac, both between Wasson Bluff and Five Islands.

43 Work in the 1980s by teams from Columbia and Harvard, with financial support from the National Geographic Society, made a concerted collecting effort at Wasson Bluff and discovered a diverse assemblage of latest Triassic to earliest Jurassic vertebrates from various facies representing fully terrestrial to aquatic depositional settings (Olsen et al. 1987, 1989, 2005; Shubin et al. 1991, 1994; Sues et al. 1994, 1996). One of the most common taxa is the small sphenodontian Clevosaurus bairdi, which Sues et al. (1994) named in Don’s honor. During the next two decades, researchers from the Nova Scotia Museum of Natural History in Halifax and the Fundy Geological Museum in Parrsboro recovered additional sauropodomorph remains, including articulated partial skeletons. This material formed the subject of an unpublished Ph.D. dissertation by Timothy Fedak at Dalhousie University in Halifax. Unfortunately, tectonic deformation of these specimens renders their anatomical interpretation very difficult.

44 Don maintained an interest in the Triassic-Jurassic of Nova Scotia through the 1980s, although he did little additional collecting himself. During this time, PEO continued to bring footprint material to Princeton where Don worked on identifying it. One particular slab had an especially colorful history. Mark Anders, then a graduate student at Berkeley (now a professor at Columbia University), told PEO in the spring of 1984 that he had seen a slab bearing interesting tracks of Batrachopus at McKay Head. Together with Joseph Smoot (United States Geological Survey), Anders and PEO collected the block (now YPM PU 23631) in July 1984. This proved to be a difficult undertaking because the slab was very heavy, covered with mud, and far from the vehicles. There was still quite a bit of mud on the rock when it was hauled into Don’s preparation room in Guyot Hall at Princeton. When Don cleaned off the slab with the tracks of Batrachopus he discovered a minute dinosaur trackway and photographed the best of its footprints. Eldon George subsequently told PEO that he had seen that slab, and had turned it over but not collected it. Thus he gets the credit for finding the slab, but Don deserves the credit for finding the dinosaur track. A photograph of a footprint appeared in Time in 1986 as part of the publicity generated by a press release from the National Geographic Society on the discoveries at Wasson Bluff (even though the block was collected at McKay Head.) It was labeled, “World’s smallest dinosaur footprint”.