The microhabitat of epifaunal taxa foraminifera can only satisfactorily be determined from direct observation of living individuals. Strictly speaking, epifaunal refers to living either on the sediment surface or else on a firm substrate, such as a shell or other structure, on or above the sediment surface. If the sediment is soft, the distinction between epifaunal and shallow infaunal may be negligible. Examination of stained samples from depth slices below the surface provides data on infaunal taxa but the resolution is dependent on the thickness of the sediment slices. However, as the sediment-water interface forms the upper surface of the topmost sediment sample it is impossible to separate those epifaunal on the sedi

ment surface from those having a shallow infaunal modes of life. In this study, the samples from the open shelf were ~1 cm thick while the topmost samples from the deeps were 0.5 cm thick. Thus, the Textularia sagittula group which is known to be epifaunal (see Murray 1991, plate 3, fig. b) occurs in the 0.0-0.5 cm sample of Muck Deep (core MD7b). Because of this sampling problem Corliss (1991) extended the use of the term epifaunal to include those living in the surface 1 cm of sediment (the most common thickness of a sediment sample). This method has been followed by other authors (Barmawidjaja et al. 1992), and they extended its use down to 2 cm. However, although it may be convenient to do this it may lead to confusion or erroneous conclusions. For instance, live Ammonia beccarii, observed in freshly collected sediment from near Southampton, are rarely seen at the sediment surface as they live just below the sediment-water interface (personal observation) and are therefore considered to be infaunal. However, Barmawidjaja, et al. (1992) describe this species as 'apparently epifaunal'. Even infaunal taxa are not consistently present at the same depth in different areas (Corliss and Van Weering 1993; Jorissen 1999) because the controls on their position in the sediment, which are mainly food availability/type and sediment porewater geochemistry, vary both spatially and temporally. Nevertheless, it is important to gather data on microhabitat in order to define the limits of variability so that it might be applied to the interpretation of the fossil record.

With the exception of Muck Deep core MD7a, live forms are abundant only in the top 1 cm of sediment. The upper layers of replicate core MD7a were disturbed by a burrow of the polychaete worm Pectinaria and, as a consequence of this, live forms extended more abundantly to a greater depth (Murray, in press). Beyond the observation that most living forms are in the topmost 0.5 cm layer, there are no clear depth preferences evident. Living forms do not extend down deeper than the redox boundary; in the case of SN6, they do not extend down even close to that boundary (Table 1). In Stanton Deep core SN6b there is an anomalously abundant occurrence of Cornuspira involvens which must be due to a local bloom.

As described above, there are no major differences of temperature or salinity across the shelf. The bottom waters must be well-oxygenated as the redox boundary in the surface sediments is located at 1 cm or greater. However, there are differences in sediment grain size and sorting with muddy sediments being confined to the enclosed deeps. This is in turn related to the effects of storm waves and tidal currents. There may also be differences in the availability of food. Based on macroscale hydrodynamic modelling, Delhez (1998) calculated that annual mean primary production for 1989 for the areas of Muck and Stanton deeps was 150 and 75 g orgC m^-2 year^-1, respectively. However, there is almost certainly advection of organic detritus into the deeps as this is likely to be transported with the fine-grained sediment. The precise controls on the niches of individual species remain unknown.

Although Heron-Allen and Earland (1916) had some samples from deeper water to the west of Scotland, most were from the inner shelf, and their faunas include diverse miliolids not seen in this study. They recorded 324 benthic species and varieties, 27 of which were new records for British seas. In the present study, several additional species have been found which have not previously been recorded from the Scottish shelf. These species include Cuneata arctica, Eggerella europea, Eggerelloides medius, Morulaeplecta bulbosa, Portatrochammina murrayi, Recurvoides trochamminiformis, Cornuloculina balkwilli, Ammonia falsobeccarii, Nonionella iridea, Robertina subcylindrica, and Rosalina anomala.

The relationship between shape of agglutinated taxa (morphogroups) and environment was explored by Jones and Charnock (1985). The ATAs of both Muck Deep and Stanton Deep (Table 3) are dominated by morphogroup C1 (elongate tests) with subsidiary B3 (flattened or lenticular, essentially planispiral and globular trochoid tests). According to Jones and Charnock this composition is correct for Muck Deep (geographically inner shelf but with outer shelf water depth) but not for Stanton Deep (outer shelf position and water depth). In comparison with the diversity for ODAs, the ATAs fit well with the pattern from other studies. For the Fisher alpha index, the three deeps samples fall in the field previously defined for shelf basins (Murray and Alve 2000); for the information function, H(S), they fall in the area of overlap of shelf, shelf basin, fjords, and bathyal/abyssal (Figure 11). Thus, again it is confirmed that although the ATAs are drawn from only a tiny proportion of the ODAs (0.07-2.35%, Table 3), they still preserve useful ecological information.