Over 200 exploration and delineation wells have been drilled on the Scotian Margin since 1966. Many of these have been the subject of micropaleontological investigation, including over 20 with data and interpretations concerning the Cenozoic interval (Ascoli 1976; Thomas 2002). The Cenozoic benthic foraminiferal assemblages in most of these wells, as described from cuttings samples plus occasional sidewall core samples, are typical for neritic or upper or lower bathyal conditions. They usually reflect a net regression from bathyal conditions in the early Paleogene to a shallower, neritic setting by the late Neogene. An extensive bibliography of published and unpublished data on the micropaleontology of these wells is given in Thomas (2002).

Cenozoic benthic assemblages on the Scotian Margin are generally rich and diverse, commonly with 20 or more foraminiferal species per sample (Ascoli 1976; Gradstein and Agterberg 1982; Thomas 2002 and unpublished data). However, at some levels in some wells, diagenesis, rapid sedimentation, unfavourable bottom conditions or other factors limit assemblages to low numbers of a few species: nevertheless, a moderate diversity of species is retained in all but the poorest samples. In some cases calcareous forms are nearly or entirely absent, replaced by agglutinated assemblages, which may be rich and diverse. These trends are evident from examination of cuttings samples, even though the nature of such samples precludes detailed statistical counts (see Materials and Methods). In only a few cases have discrete, unmixed benthic assemblages of Cenozoic age been recovered from east coast wells, providing rare opportunities for paleoecological study (Thomas 1991).

In one well however, Petro-Canada-Shell Wenonah J-75, examination of sidewall cores from the Oligocene interval has revealed unusual, oligospecific assemblages of benthic foraminifera. Significantly, seismic records indicate that this well was drilled in a site that was, during much of the mid Cenozoic, the location of a deep canyon (informally termed "Wenonah Canyon" by R.A. MacRae, personal communication 2001) that debouched onto the slope. Wenonah Canyon is approximately 11 km across at its widest point, and its maximum depth, as inferred from seismic data, is approximately 475 m (uncorrected for compaction). At its greatest development, in the mid Oligocene, its actual depth may have been some 500– 600 m (J. Shimeld, personal communication).

The purpose of this paper is to compare these unusual assemblages with coeval, "typical" benthic foraminifera from a nearby well, Canterra et al. Whycocomagh N-90, situated in a shallower setting, closer to a normal shelf environment (Thomas 2001b; Fig. 1).

All cuttings samples used in this study were prepared by the Canada-Nova Scotia Offshore Petroleum Board (CNSOPB), and picked by Atlantic Paleo Services of Dartmouth, Nova Scotia, in the late 1970s using methods described in Thomas (2002). The sidewall core samples (SWC's) were prepared and picked by or for the well owner, and processing details are not available. Selected specimens were imaged by the author using an ElectroScan E3 Environmental Scanning Electron Microscope, housed at the Geological Survey of Canada (Atlantic).

Note that, unlike Whycocomagh N-90, Wenonah J-75 was drilled prior to metrification, and depths on all samples and slides stored both at CNSOPB and at Geological Survey of Canada (Atlantic) were recorded in feet. All depths have been converted to metres for the purposes of this paper. Location and engineering data for both sites are provided in Table 1. Depths of all samples from both wells used in this study are given in Appendix 1.

The oligospecific assemblages in the Wenonah sidewall cores are distinct from any others so far encountered in Cenozoic Canadian Atlantic offshore sites and it seems reasonable to ascribe this uniqueness to unusual sedimentological/chemical conditions in their original environments in the Wenonah Canyon. Some of the more common species in the sidewall core samples are illustrated in Fig. 6. The nearby Whycocomagh N-90 well, situated in a less incised area between canyon tributaries, contains no unusual assemblages. Figure 7 shows the Cenozoic biostratigraphic correlation between the two wells.

Studies reported in the literature clearly show that submarine canyons can contain benthic environments that are different from those in similar depths in normal slope environments. In a hydrologic study of currents in modern submarine canyons, Shepard et al. (1979) described alternating vertical currents of 25– 35 cm/sec, measured 3 m above the sediment surface. Because of their frequency and nature they were believed to be tide-related, and varied with depth along the axis of the canyons. They also reported short-lived turbidity currents flowing downhill at speeds of roughly 80 cm/sec.

Whether or not these special conditions in canyons always result in unusual benthic foraminiferal assemblages is a complex question, and the existing literature seems to contain a variety of answers. Weston (1985), in a study of modern canyon assemblages in piston cores in the eastern North Atlantic, found that canyons tend to support so-called "cosmopolitan" species (e.g. Brizalina dilatata andBrizalina subaenariensis); and furthermore, that normal depth stratifications of benthic species do not apply as well in canyons as in normal slope environments. Weston cited frequently changing water masses and variability of sediment supply as probable causes of these phenomena. In an examination of core samples from submarine canyons off the New Jersey coast, Jorissen et al. (1994) found that canyon samples tend to have relatively low benthic foraminiferal densities, and the infaunal suite of species is poorly developed, presumably due to sediment instability. These authors postulated that periodic small-scale mass wasting and bottom scour by currents may prevent benthic foraminiferal faunas from reaching a high-diversity equilibrium state. Furthermore, periodic shifts of the oxic-anoxic interface may tend to favour species with wider tolerance to low dissolved-oxygen levels. Vénec-Peyré (1990) found that, in canyons in the western Mediterranean, smaller foraminiferal taxa (both benthic and planktic) tend to be concentrated by hydrological factors near the canyon axes, and that diversity increased toward these axes.

In contrast, other studies of benthic foraminiferal assemblages in canyons, both modern and ancient, have found no particular anomalies. In an examination of material from Oligocene-Miocene canyon fill in west Africa, Petters (1982, 1984) reports only more or less "normal" benthic foraminiferal assemblages, in this case essentially the same mix of calcareous and benthic species found in contemporaneous deposits in nearby wells drilled on ancient slope environments. Similarly, Pícha (1979) reported no special benthic foraminiferal assemblages in Eocene-Oligocene canyon fill in central European sites, but did note the common presence of shallower-water taxa presumably carried down the canyon channel, and older forms supposedly eroding out of the canyon walls. It may be significant, however, that both of these studies relied almost exclusively on cuttings samples, which usually comprise material from a stratigraphic interval rather than a discrete horizon. In such samples, the characteristics of discrete samples from one instant in time can be totally masked by the addition of assemblages from adjacent levels.

There have been several studies of modern submarine canyons using samples from piston or box cores, or other relatively precise means, that have found no unusual benthic foraminiferal assemblages. Stanley et al. (1986) collected samples from the lower end of Wilmington Canyon, offshore Delaware, that contained "normal" benthic suites, although with a mix of shallower-water taxa ostensibly due to down-canyon transport. Petrological evidence suggests that deposition in the area was intermittent, and included a component of slumped sediment from eroding canyon walls. In a study of modern sediments from canyons off the Portuguese coast, Young et al. (1995) studied the accumulation rates of benthic foraminiferal tests. They found that these rates vary widely from canyon to canyon, and, though no especially high concentrations of these species were seen, the taxa Brizalina, Bulimina and the Cassidulininae had the highest overall accumulation rates. These taxa also figure prominently in the Wenonah samples.

In a later study of Wilmington and South Heyes canyons, Lundquist et al. (1997) found only more or less "normal" benthic assemblages, although several samples contained taxa indicative of shallower water as a result of downslope transport. They found no evidence, however, of high-energy downslope transport events in the last 200– 400 years, further suggestive of a stable benthic regime. An additional finding of this study was that the linearity of the canyon may affect the impact of bottom currents, with relatively straight stretches of canyon less affected than crooked ones.

Finally, Schmiedl et al. (2000) found that benthic foraminiferal faunas of canyons in the western Mediterranean do not generally contain opportunistic species, and are adapted to rather stable environments. In their study area there was a relatively large amount of phytodetritus and other organic matter as food sources, high standing benthic foraminiferal stock and diversity, a well developed deep infauna, and relatively high oxygen consumption rates. The open shelf areas at the bottom of the canyon actually showed lower standing stocks than the canyon samples. Possibly these assemblage characteristics, so different from the Wenonah material, may be attributable to the relative quiescence of the modern highstand, in contrast to the activity of the sea-level fall associated with the erosion of the Wenonah canyon during the mid Cenozoic.

While little or no relevant ecological data exists for most of the foraminiferal species in the Wenonah canyon assemblages, there are a few with close physical resemblances to modern forms. In such cases, while a direct lineal relationship with the Cenozoic form is conjectural, it is at least possible that the older form had similar environmental preferences to its modern counterpart.

For example, Trifarina wilcoxensis, which appears in relatively large numbers in two sidewall cores (1030.2 m and 1158.2 m) from the upper section of the canyon interval, may be an ancient analogue to modern Trifarina angulosa. The latter species has been listed as indicative of high-energy environments in areas off Antarctica (Violanti 1996), and one of two species (the other being Cibicides lobatulus) associated with strong bottom currents at depths of 200– 600 m on the Norwegian continental margin (Mackensen et al. 1985).

The globobuliminids, buliminids and bolivinids so common in some of the sidewall cores from the lower part of the canyon section may reflect reduced-oxygen conditions, as in the case described by Miller and Lohmann (1982) for the northeast United States continental slope. Furthermore, one bolivinid, Bolivina subaenariensis, has been identified with muddy, high-carbon sediments on the northern Florida continental slope by Sen Gupta et al. (1981). In this case, the sediments are from an area known for periodic local upwelling.

Finally, Cassidulinoides parkerianus, also common in the lower canyon interval, is one form which is still extant. While not widely distributed, it does appear to be more common at sites near continental margins, (Brady 1884; Matoba 1967; Ishman and Webb 1988) and so may have some affinity for areas of upwelling.

Possibly this succession of assemblages may reflect the evolution of the canyon, from a deeper-water environment subject to incursions of lower-oxygen bottom waters favouring the buliminids and bolivinids to somewhat shallower conditions with better oxygenation but more local dynamism in the benthic environment.

Significantly, the Oligocene interval of Whycocomagh N-90, some 10 km to the north, contains no unusual assemblages of benthic foraminifera (see Appendix 2); in fact none of the species occurring as important components of the sidewall core sample assemblages in Wenonah J-75 occur at all at Whycocomagh. Figure 7 illustrates the paucity of micropaleontological similarities between the two sites.

Seismic records indicate that the Wenonah J-75 well on the Scotian Shelf was the location, during the mid Cenozoic, of a tributary of a major submarine canyon incised into the shelf, debouching onto the lower slope.

Sidewall cores in the Oligocene section of Wenonah J-75 provide a unique opportunity to observe discrete benthic foraminiferal assemblages of Cenozoic age on the Scotian Margin. These samples contain unusual, oligospecific associations apparently related to the physiographic setting in the canyon. Unusual hydrographic conditions, especially at times during the Oligocene, may have led to the development of these associations, often dominated by relatively high percentages of just one, two or three species – taxa apparently capable of withstanding the unstable conditions in the canyon. As the canyon became more infilled and the depositional environment shallowed, these faunas gave way to more mixed benthic assemblages, often exhibiting breakage and other effects of a more dynamic environment. The coeval interval of nearby Whycocomagh N-90, a well in a non-canyon setting, contains no such assemblages.