Fossil footprints of vertebrates are known in remarkable quantity and diversity from many stratigraphic levels in the Mississippian and Pennsylvanian of the Maritimes Basin in Nova Scotia (Dawson 1855 and supplements; Matthew 1903, 1905; Carroll et al. 1972; Sarjeant and Mossman 1978; Calder et al. 1995). In marked contrast, none had been described from the neighbouring province of Prince Edward Island nor from the Permian of the Maritimes Basin previous to the discovery in 1984 of the footprints reported here. Since the discovery of trackways described here, a subsequent discovery from Prince Edward Island was published by Mossman and Place (1989). The small but prolific trackway-bearing slab here described was discovered at Beech Point, Malpeque Bay by EBU in August of 1984. When the significance of the find had been confirmed by Dr. John Ostrom of Yale University (written communication, 25 March 1987), he revisited the site in 1987 and in subsequent years collected additional, isolated footprints of the same types with Mr. Charles Corkum of Prince Edward Island.

The fossil locality is situated on the eastern shore of Malpeque Bay at Beech Point, north of the end of Beech Point Road and due west of the town of Hamilton, in GSC Map Area 11L5 at latitude 46° 29' 40" north, longitude 63°45' west (Fig. 1). The tracks occur as natural casts within redbeds of the Hillsborough River Formation (van de Poll 1989). The trackway-bearing horizon occurs as a locally reduced, pale grey sandstone horizon within a sequence of tabular, reddish brown sandstone beds with internal large-scale, shallow cross-sets and less commonly, narrow troughs (Fig. 2). Greenish-grey mottling occurs on bedding surfaces and vertically. The shore there is covered with a shingle of flat, somewhat water-worn, sandstone cobbles and boulders derived from the typical red bluffs along the beach.

Small slab (Figs. 3, 5, 6) with natural casts of three or more partial trackways; three isolated pedal imprints (Fig. 4). Originals in the possession of EBU; casts reside in the National Museum of Natural Sciences, Ottawa, the Nova Scotia Museum of Natural History (NSM 988GF72.1), and the Peabody Museum of Yale University (YPM-PU 23973).

Trackway A comprises a left manus-pes set and the following right pes print. (The isolated pes is identical to A in size and form.) Trackway B was made by an individual about three-fourths the linear size of A. It begins with a right manus-pes set; the following left set is partly obscured; and the second right pes, truncated by the slab edge, overprints the third and fourth toes of the right pes in Trackway A. No tail-trace is evident. Measurements are given in Table 1.

The assignment of trackway A to Gilmoreichnus and track-way B to Notalacerta is consistent with the summary descriptions of the ichnogenera in the treatise of Haubold (1971, cf. his figs. 13.14 and 20.1, 2). Amongst the attributes that differentiate these two ichnogenera, Notalacerta differs from Gilmoreichnus in its lacertoid morphology and smaller size. A possible third trackway ‘ C' with poorly preserved footprints and medial, straight tail drag is of uncertain affinity.

The number of trackways on one small slab indicates concentrated activity at this horizon, which bodes well for future discoveries. The occurrence of such a concentrated footprint record within a reduced redbed horizon is suggestive of the presence of organic material, but whether organic matter in the presence of the trackmakers bears any paleoecological significance is uncertain. A trackway site in coeval redbeds of the Pictou Group at Brule, Nova Scotia (Calder et al. 1995; unpublished material) occurs within a walchian conifer forest interpreted to be a dryland waterhole.

In comparing Paleozoic trackways with taxa based on skeletal material, we are hampered by the fact that, in most of the latter, the feet are unknown. Whole groups of animals must therefore be omitted from comparison for lack of evidence. Even in those taxa for which foot material is known, it is seldom complete enough to permit a strictly factual reconstruction of both manus and pes. Our attempts at correlation will therefore be, at best, approximations.

The foot structure represented in the Malpeque Bay footprints is of a primitive reptilian type. Eliminating, therefore, the various Permo-Carboniferous amphibian groups and those reptiles with more specialized feet, our attention logically focuses on the small, generalized "stem-reptiles" of the Suborder Captorhinomorpha, the group that appears to have given rise to all the more advanced reptiles and thus, indirectly, to birds and mammals. Following Heaton (1979), the more primitive Pennsylvanian and early Permian captorhinomorphs constitute the Protorothyrididae ("Romeriidae" of other authors), while the more derived genera from the early Permian constitute the Captorhinidae, with the captorhinid Romeria providing a morphological transition between the two families. What is known of their foot structure is described for Carboniferous forms by Carroll and Baird (1972) and for Permian ones by Clark and Carroll (1973) and Holmes (1977). The body proportions of these reptiles are similar to those deduced for the Notalacerta and Gilmoreichnus trackmakers, and a close correspondence in the skeletal structure and proportions of the manus and pes is evident. Trackmaker ‘ B' evidently was a moderately short-bodied animal that walked with a sprawling gait. Trackmaker ‘ A', however, appears to have had a more elevated posture and less sprawling gait.

We are not prepared to nominate any particular genus as the closest correlative of the Prince Edward Island trackmakers. Haubold (1971) ascribed Notalacerta to the temnospondyl amphibians; however, we feel that the attributes described above are inconsistent with that group. The strong affinity of Gilmoreichnus with pelycosaurs such as Ophiacodon (Haubold 1971), the osteological remains of which have been reported from the Pictou Group redbeds of Prince Edward Island (Langston 1963), we consider more tenable. Prince Edward Island redbeds have provided a fossil record of large vertebrate taxa (Fig. 7) including eryopoid and brachyopoid temnospondyl amphibians, the problematic reptiliomorphs Seymouria and Diadectes, and reptiles that include ophiacodont, sphenacodont and edaphosaur pelycosaurs (Langston 1963). Up to now, however, no bones of primitive captorhinomorphs have been identified. This is hardly surprising, for small reptiles are relatively uncommon, even in the more abundantly fossiliferous beds of the Ohio Valley and the New Mexico-Texas-Oklahoma area. In Atlantic Canada, despite energetic and extensive prospecting by Langston and his colleagues, the Permian redbeds of Prince Edward Island have so far yielded only half-a-dozen fragmentary specimens of reptiles. Thus, the footprints from Malpeque Bay serve as still another example of how the ichnological evidence can supplement the skeletal record to provide a fuller understanding of an ancient fauna.

The footprints of Malpeque Bay are broadly Permo-Carboniferous in morphology but include ichnogenera that if discovered in isolation from one another would be taken to indicate either a Pennsylvanian (Notalacerta) or early Permian (Gilmoreichnus) age (Lucas 2002a, 2003). The youngest occur-rence known previously of Notalacerta is from the Stephanian of the western United States, and the oldest occurrence known of Gilmoreichnus dates from the late Stephanian of western Europe (Lucas 2003). The ichnological record therefore coincides most parsimoniously with the late Stephanian Dromopus biochron of Lucas (2003). Although the ichnogenus Notalacerta is represented in this biochron, the Malpeque Bay locality represents its youngest known occurrence, which conservatively is late Stephanian: the palynological and macrofloral records of the redbeds of Malpeque Bay place them in the late Stephanian to Lower Permian, whereas the osteological record of vertebrates favours a Lower Permian age.

Frankel and Crowl (1970) and Prest (1972) assigned outcrops in the region of Spring Valley (Fig. 7, between Malpeque Bay and New London Bay) to the informally named stratigraphic Subunit A2, which they assigned to the Lower Permian on paleontological evidence. Subsequent lithostratigraphic study by van de Poll (1983) assigned the strata, in which the footprints of Malpeque Bay and Prim Point (Fig. 7) both occur, to his informal Megacyclic Sequence III of the Pictou Group. More recently, van de Poll (1989) introduced formal lithostratigraphic nomenclature for the redbed outcrops on PEI, and in so doing he renamed the unit in which the track-bearing sequence occurs, the Hillsborough River Formation. The correlation of the formations of van de Poll with those in informal use for the sub-surface geology of the Gulf of St. Lawrence (Rehill et al. 1995; Giles and Utting 1999) is somewhat problematic. Petroleum wells in the area have been interpreted as being collared in the Naufrage Formation (Giles and Utting 1999), although a direct correlation of the two formational schemes awaits. The age of the Hillsborough River Formation was given by van de Poll (1989) as early Permian (Artinskian).

The southeastern district of Prince Edward Island, in the area around Hillsborough Bay (Fig. 7), yields a spore flora identified by Barss et al. (1963) that agrees with the amphibians and reptiles described by Langston (1963) in indicating an early Permian age for the fossiliferous beds. In the north-central district around Malpeque Bay, however, the two disciplines appear to yield discordant ages. The uppermost strata of borehole Imperial MacDougall No. 1, situated west of Malpeque Bay were dated by Barss et al. (1963) as Stephanian (late Pennsylvanian) on the basis of palynology. (In boreholes north and south of it, however, Stephanian spore assemblages were not encountered above the 524m level.) The most recent interpretation of the age of these strata, which have been assigned to the Naufrage Formation (Giles and Utting 1999), is Stephanian to early Permian based on a Stephanian palynoflora from well samples and inferences of a younger age for overlying, barren redbeds, in part drawn from the vertebrate record discussed below (P.S. Giles, personal communication, 2001). Recently, Ziegler et al. (2002) have undertaken an assessment of the macrofloral record of the Prince Edward Island redbeds in an effort to resolve local phytogeographic effects from global climate change accompanying the fall in global sea level during the early Permian. Their assessment of the combined paleontological data for Prince Edward Island supports a Stephanian-Lower Permian age range for the redbeds, with a general younging to the northeast, as purported by van de Poll (1989).

East of Malpeque Bay are two bone-bearing sites, Spring Valley (P-6108) and Irishtown (P-6111), from which vertebrate fossils were described by Langston (1963). A review of osteo-logical finds and their biostratigraphic significance follows.

Vertebrate trackway and skeletal discoveries from the red-beds of the Maritimes Basin exposed on Prince Edward Island, together with macrofloral evidence framed by subsurface palynology, increasingly agree in a late Stephanian to early Permian age for the redbeds of Prince Edward Island. The combined biostratigraphic data indicate an age of late Stephanian at Malpeque Bay, younging northward through early Permian. Reflected in the fossil record is the global rise to dominance of terrestrially adapted tetrapods during the Permian as continental conditions and moisture deficits became widespread across the paleoequatorial regions of Pangea. The first records of Notalacerta and Gilmoreichnus from Prince Edward Island, combined with the subsequent discovery of Mossman and Place (1989), reflects the high probability that these redbeds are the repository of a significantly more extensive vertebrate trackway record than previously recognized. The youngest known occurrence of Notalacerta emphasizes the significance of this record.