Taxonomy and distribution of recent species of the subfamily Nodosariinae (Foraminifera) in Icelandic waters

Taxonomy of fourteen very little known species of Nodosariinae Ehrenberg, 1838 in Icelandic waters is revised. Knowledge of these species in the North Atlantic relies mainly on studies in the late 19th and early 20th centuries, using large volume samplers. Later studies have emphasized quantitative samples of a few cm3 where the Nodosariinae are very rare. This study analysed 879 dredging samples where Nodosariinae occurred in 492 samples, comprising 7598 specimens of about 415 000 of all picked foraminifera. Ordination analysis of species distributions refl ects prominent temperature and salinity differences that exist in the sampling area (753 000 km2) north and south of the Greenland-Scotland Ridge (GSR). Eight species are restricted to southern temperate waters (> 2°C): Dentalina mutabilis (Costa, 1855), Dentalina antarctica Parr, 1950, Dentalina antennula d’Orbigny, 1846, Dentalina fi liformis (d’Orbigny, 1826), Grigelis pyrula (d’Orbigny, 1826), Grigelis guttifera (d’Orbigny, 1846) comb. nov., Grigelis semirugosus ? (d’Orbigny, 1846) and Nodosaria subsoluta Cushman, 1923. Four species (Nodosaria haliensis Eiland & Guðmundsson, 2004, Nodosaria incerta Neugeboren, 1856, Dentalina elegans d’Orbigny, 1846 and Dentalina frobisherensis Loeblich & Tappan, 1953) occur mainly north of Iceland. Two species, Dentalina obliqua (Linnaeus,1758) and Pseudonodosaria subannulata (Cushman, 1923), have wide tolerance ranges for physical variables.


Introduction
Throughout most of the Cenozoic, species of the subfamily Nodosariinae Ehrenberg, 1838 (Foraminifera) as well as other calcareous, elongate, and uniserial foraminifers, were among the most significant components of foraminiferal assemblages in the deep seas of the world oceans, but with a progressive and severe decline in abundance to recent times (Hayward 2002). Today, these forms are moderately diverse and constitute a minor component of modern faunas (Hayward et al. 2012). The world diversity of Nodosariinae comprises seventeen genera and 1580 accepted names of recent and fossil species (Hayward et al. 2022).
The first comprehensive account of Foraminifera in Icelandic waters reported three species of Nodosariinae (Nørvang 1945). Later studies in this area have emphasized quantitative studies of a few cm 3 , where Nodosariinae and most other large foraminifera are either absent or very rare Mackensen 1987;Jennings et al. 2004). The study of Eiland & Guðmundsson (2004) is an exception, since they analysed about one third of the dredging samples from the BIOICE project (Steingrímsson et al. 2020) that were then available, and reported twelve nodosariine species in Icelandic waters, but only with a preliminary note on their distributions, because of the limited number of samples available at the time. Since then, all samples of the BIOICE project have been processed, revealing fourteen species that are extant in Icelandic waters.
The objective of this study is to use the new material to review morphologic variation of species of Nodosariinae in Icelandic waters, update the taxonomy, redescribe species with high resolution photographs of local specimens, and map the geographic distributions of the species in relation to different water masses, as characterized by bottom water temperature, salinity, and water depth.

Material and methods
Samples were collected during the BIOICE program (Benthic Invertebrates of Icelandic Waters), with a general objective to investigate faunal composition of benthic invertebrates in Icelandic waters and map their distributions. The study area is the 200-mile economic zone of Iceland (758 000 km 2 ). Sampling proceeded during the years 1991-2004 and resulted in 1031 zoological samples, taken at 579 stations. Selection of 62% of the stations followed stratified random sampling (Palsson et al. 1989), and 38% were selected to sample interesting habitat types and topographic bottom features discovered during the cruises, and some were the same stations that are part of the annual ground fish survey program of the Marine and Freshwater Research Institute in Iceland (BIOICE 2005;Guðmundsson et al. 2014;Steingrímsson et al. 2020). At each station, several devices were used to sample different components of the benthos. A modified RP-sledge (Rothlisberg & Pearcy 1976;Brattegard & Fosså 1991) sampled epifauna on relatively smooth and soft bottom types; a detritus Sneli-sledge (Sneli 1998), sampled infauna of soft and mixed sediment substrates; an Agassiz trawl (McIntyre 1956) sampled megafauna on soft and even bottom; and an 80 × 80 cm triangular dredge (KCDenmark 2022) was the only device that could be applied on rough rocky bottoms. These dredging samplers were routinely towed at speed of 1 knot for 20 minutes, which amounts to about 600 m of towing length. At each station, near bottom water temperature and salinity were measured with CTD meter (Conductivity-Temperature-Depth), except for a few stations where it malfunctioned. The near bottom water temperature at the BIOICE stations ranges from -0.9°C to 9.64°C; water depth from 23 m to 3006 m, and salinity varies between 34.17‰ and 35.20‰. Detailed information on the BIOICE samples and location data is available in open access (Steingrímsson et al. 2020).

Description
Test shape elongate, cylindrical, slightly curved to straight, moderately nodular; initial end rounded in megalospheres and only slightly tapering; the microsphere is gradually tapering to a pointed initial end. Length of test 4-5 mm, the largest 7 mm; test width 0.3-0.5 mm. Chambers initially globular to subglobular, becoming slightly more elongated as added; rectilinearly arranged in both generations; number of chambers 5-9 in megalospheres, 15-21 in microspheres. Chamber embracement minimal, sutures horizontal. Aperture radial, nearly central, slightly raised, sometimes protruding, with up to 12 symmetric tines that are fused in the center (Fig. 2C-D). Short costae are confined to base of chambers; initial end usually with one or more basal knobs or a short spine. Secondary surface laminations absent (atelo-lamellar). Wall medium thick, finely perforated.

Remarks
Of the 63 examined specimens, 22 were microspheres. Diameter of proloculus in the megalospheres is about 0.5 mm, but less than 0.03 mm in the microsphere. The illustration of the lectotype of D.
antennula shows an open, rounded and crenulate aperture, but they also questioned the validity of this species, stating "that the independence of this species remains doubtful" (Papp & Schmid 1985). Judging from the illustrations it seems that the tip is worn away and it is assumed here that it originally had fused apertural tines as the present specimens. Several names of fossil forms have been erected that seem very similar to D. antennula, like Dentalina capitata (Boll, 1846) described from the Oligocene by Boll (1846: 177, pl. 2 fig. 13a-b); Dentalina buchi from the Eocene described by Reuss (1851: 60, pl. 3 fig. 6); and Dentalina philippii also from the Eocene and erected by Reuss (1851: 60, pl. 3, fig. 5

Description
Test shape elongate, cylindrical, somewhat tapering, slightly nodular, more so in the latest chambers, initial end rounded, distinctly curved, rarely almost straight. Length of test 3-4 mm, the largest 5 mm; test width 0.4-0.6 mm. Chambers subglobular, sometimes slightly cylindrical in later chambers, rectilinearly arranged, embracement minimal, distinctly to slightly curved; number of chambers usually 4-7, at most 9. In microspheres, the initial 3-5 chambers are wound in about half of an evolute planispiral (Fig. 3A). Sutures horizontal. About 90% of the specimens have a basal knob or short spine, exterior otherwise smooth. Secondary surface laminations cover the whole test (ortho-lamellar) or only part of the test (plesio-lamellar). Aperture radial, nearly central, slightly raised or protruding, with up to 12 symmetrically arranged tines, that are fused in the center. Wall thick and finely perforated.

Remarks
Of the 1364 examined specimens of D. elegans, 6 were microspheres. Diameter of proloculus in the megalospheres is 0.2-0.4 mm, but < 0.05 mm in the microsphere. Dentalina elegans was tacitly considered to be a megalospheric form of D. frobisherensis (Grønlund & Hansen 1976), since the morphology of the adult chambers is very similar. The species Dentalina baggi was distinguished from Dentalina elegans (Eiland & Guðmundsson 2004), stating that the test of the former had less inflated chambers, and that it lacks a prolocular spine, but noted though that about 10% of D. elegans also had no prolocular spine. The added material from the BIOICE renders this distinction arbitrary for the material at hand, as variation in test width and chamber inflation overlaps in specimens with or without a prolocular spine. It seems that forms identified here as D. elegans are relatively common in the North Atlantic and may have been reported under several names. The species Nodosaria advena, first described by Cushman (1923) off the northeast coast of USA seems to have a nearly identical morphology as the present D. elegans but differs only in being much larger viz up to 7 mm in length. Specimens under the name Nodosaria advena have been reported from the North Sea (Gabel 1971), off Scotland and the Faroe Islands (Cushman 1923)

Description
Test shape elongate cylindrical and slender; distinctly curved, slightly nodular, and evenly tapering in a rounded but narrow end. Length of test 2-3 mm, the longest is 4.5 mm; test width 0.15-0.25 mm. Chambers initially cylindrical, becoming slightly inflated and more elongated as added. Chamber arrangement rectilinear, minimal embracement; number of chambers commonly 6-11, or up to 14. Initial chambers in microsphere are rectilinear. Sutures oblique. Aperture radial, nearly central and usually protruding, with up to 16 symmetrically arranged tines that are fused in the center. Surface smooth without ornamentation; secondary surface laminations absent (atelo-lamellar). Wall finely perforated, thin, and transparent.

Remarks
Of the 24 examined specimens, 14 had an intact initial end, with prolocular size that varied nearly continuously from less than 0.05 to about 0.12 mm, presumably including both generations with rectilinear chamber arrangement of the initial part. Loeblich & Tappan, 1953 Fig. 5 Dentalina frobisherensis Loeblich & Tappan, 1953: 55, pl  minimally embracing; number of chambers 9-14, sometimes up to 21. Initial chambers in microsphere are rectilinear. Sutures horizontal. Aperture radial, nearly central, slightly raised, sometimes protruding, comprising up to 12 tines, symmetrically arranged, usually fused in the center, sometimes only partly fused. Surface smooth, except for a very short basal knob or spine. Secondary surface laminations cover the whole test (ortho-lamellar) or partially (plesio-lamellar), wall finely perforated and thick.

Remarks
Prolocular size is commonly around 0.1 mm, but ranges continuously from about 0.03 to 0.14 mm. The secondary laminations result in successive thickening of the whole test as new chambers are added (Eiland & Guðmundsson 2004), which partially levels out the originally nodular character of the initial part of the test as new chambers and laminations are added.

Description
Test shape elongate, cylindrical, commonly distinctly curved, sometimes nearly straight, slightly nodular in latest chambers, somewhat tapering into a pointed end. Length of test 3-5 mm, the largest specimens nearly 7 mm; test width usually 0.6-0.8 mm. Chambers subglobular, rectilinearly arranged; number of chambers 8-11 or up to 13. Initial 6-9 chambers in microsphere are arranged in an evolute planispiral, nearly completing one whorl (Fig. 6C). Sutures horizontal, sometimes slightly oblique. Aperture radial, eccentric, protruding, made of up to 12 rather broad symmetrically arranged tines, that are fused in the center. Initial end with a short spine. Conspicuous longitudinal costae run the whole length of the test along slightly twisted lines; the costae divide or fuse as the diameter of the test changes; the apertural face is free of costae (Fig. 6A, D) and in occasional specimens the last one or two chambers are smooth. Secondary surface laminations seem to cover whole test (ortho-lamellar). Wall thick and finely perforated, except for nearly imperforate narrow segment close to the aperture.

Remarks
Seven microspheres were found among the 1082 examined specimens. Diameter of the proloculus in microspheres is about 0.02 mm, and > 0.2 mm in the megalospheres. Secondary laminations of successive chambers levels out the slightly nodular character of the earlier part of the test, resulting in an evenly tapering cylindrical shape.

Description
Test shape elongate, cylindrical, usually straight, sometimes only slightly curved, somewhat nodular near the apertural end; tapering to a pointed initial end. Length of test commonly 4-5 mm, the largest 7 mm; test width 0.3-0.4 mm. Chambers sub-globular, rectilinearly arranged, number of chambers 8-14, or up to 16. Sutures horizontal. Aperture radial, central and protruding, with up to 14 tines that are symmetrically arranged and fused in the center. Initial end with relatively long spine; conspicuous longitudinal costae run in nearly straight lines along the whole length of the test and are fused with the base of the apertural tines ( Fig. 7D-E); the costae divide or fuse as the diameter of the test changes. Secondary surface laminations cover the whole test (ortho-lamellar). Wall thick and finely perforated.

Remarks
All the 322 examined specimens of this species had a rectilinear chamber arrangement in the initial part. Diameter of proloculus falls usually within the range of 0.10-0.15 mm, but in few specimens the diameter reaches 0.40 mm. Robust specimens of D. mutabilis bear some superficial resemblance to exceptionally slender individuals of D. obliqua. Nevertheless, D. mutabilis differs from D. obliqua in having a central aperture and the surface costae run in direct continuation of the apertural tines (Fig. 7E), and nearly parallel to the axis of the test, whereas the costae are slightly twisted in D. obliqua (Fig. 6A-B).
Genus Grigelis Mikhalevich, 1981 Diagnosis "Test elongate, uniserial; ovoid to fusiform proloculus may be up to twice the length of the first chamber; later chambers usually, but not always separated by long narrow necks; aperture terminal, at end of a long, narrow neck, radiate bars join at center giving petalloid effect." (Hayward et al. 2012: 139)

Description
Test shape elongate, nearly straight to slightly curved, strongly nodular, barely tapering, or successive chambers are of equal diameter, except the proloculus which is often largest. Length of test fragments 2-4 mm, the largest 8 mm; test width 0.15-0.25 mm. Chambers pear-shaped, rectilinearly arranged, embracement minimal; chambers attached at the very end of a long apertural neck; number of chambers unknown, the largest fragment has 11 chambers. Proloculus often larger than later chambers. Sutures horizontal. Aperture radial, central and protruding, at the end of a long neck, with up to 12 symmetrical tines, that are fused in center. Surface smooth, except for short longitudinal costae extending for about one third to half of the upper part of the apertural neck; initial end pointed or with long spine. Secondary surface laminations absent. Wall transparent of medium thickness, finely perforated.

Remarks
The 29 examined specimens were all fragments, of which four had a proloculus. The top of the aperture is commonly damaged, exposing a circular, crenulated opening.

Description
Test shape elongate, nearly straight to slightly curved, strongly nodular, barely tapering, or successive chambers are of equal diameter. Length of test fragments 2-3 mm, comprising 3-7 chambers; test width 0.2-0.3 mm. Chambers pear-shaped with an elongated long neck, rectilinearly arranged, with minimal embracement; chambers attached at the very end of the apertural neck; number of chambers unknown, the largest fragment has 7 chambers. Sutures horizontal. Aperture radial, protruding and central, at the end of a long neck; radial tines up to 12, symmetrically arranged and fused in the center. Longitudinal striae cover the apertural neck and the upper and lower part of the chambers, leaving the central part smooth. Secondary surface laminations absent, wall finely perforated thin and transparent.

Remarks
Typical G. semirugosus? has distinct costae or furrows that are mostly confined to the base of the chambers. However, the six specimen fragments at hand differ in having rather faint striations or furrows, covering the upper half and the lower part of the chambers, leaving the central part smooth (Fig. 9A).
In addition to being striated, the putative P. semirugosus? differs from P. pyrula in having a relatively shorter and more conical neck. The initial end in available material is missing.  (Fig. 11B), rarely smooth. Secondary surface laminations absent. Wall thick and relatively coarsely perforated, except for an imperforate segment next to the aperture.

Remarks
Of the 521 examined specimens, 17 were microspheres. Diameter of proloculus is < 0.04 mm in microspheres and 0.24-0.46 mm in the megalospheres. Some of the larger specimens of G. guttifera resemble N. subsoluta, but the former differs in being distinctively much more coarsely perforated (Fig. 10A, C) with less embracing chambers and a longer neck. This species is placed in Grigelis since the top of the fused apertural tines is distinctively more protruding than in Nodosaria and the chambers are less embracing. It seems likely that the rare reports of Nodosaria simplex Silvestri, 1872 in the North Atlantic (Flint 1899;Cushman 1923) are of megalospheric juveniles of G. guttifera; illustrated specimens have two chambers, with an elongated neck and radiate aperture. However, the types of N. simplex from Pliocene, Sicily, have only two chambers but are described to have a rounded aperture (Silvestri 1872). Flint (1899) and Cushman (1923) reported G. guttifera as rare off Ireland and perhaps also in the Faroe Channel, but most reports are from deep waters in the Gulf of Mexico, Caribbean, and off Carolina. Revision of the BIOICE material, previously identified as N. subsoluta (Eiland & Guðmundsson 2004), was found to include several specimens of G. guttifera.

Description
Test shape elongate, slightly curved, sometimes straight, usually moderately nodular, rarely strongly nodular, megalospheres slightly tapering, with round initial end; evenly tapering to a pointed end in microspheres. Length of test usually 3-6 mm, the largest nearly 10 mm; test width 0.6-1 mm. Chambers subglobular, rarely pear-shaped. Chamber arrangement rectilinear, chamber embracement usually moderate; number of chambers 4-8 in megalospheres, up to 16 in microspheres. Sutures horizontal. Aperture radial, central, slightly raised or protruding; radial tines up to 15, symmetrically arranged, usually fused in center (Fig. 12D), sometimes partly fused (Fig. 12E). Initial end with a short spine or thick knob. Secondary surface laminations absent, wall finely perforated and thick.

Remarks
Resembles N. subsoluta but differs in being less nodular and in having short secondary surface laminations that cover only the upper most parts of preceding chambers. The edges of the laminations are clearly visible in stereo microscope as faint irregular lines. Available material comprised 923 specimens, of which 21 were microspheres. Diameter of proloculus in megalospheres is 0.3-0.6 mm, and < 0.02 mm in microspheres.

Description
Test shape elongate, cylindrical, straight, barely to moderately nodular and slightly tapering; initial end rounded. Length of test 0.7-1.5 mm, the largest 2.2 mm; test width 0.4-0.5 mm. Chambers subglobular to cylindrical. Chamber arrangement rectilinear, embracement moderate to minimal; number of chambers 4-6 in microsphere, up to 12 in microsphere. Initial 3-5 chambers in microsphere are wound in about half of an evolutive planispiral (Fig. 14A). Sutures horizontal. Aperture radial, central and slightly raised; radial tines up to 12, symmetrically arranged and fused in the center. Wall smooth, finely perforated of medium thickness.

Remarks
Available material of this species comprised 371 specimens, of which 9 were microspheres. Diameter of proloculus in megalospheres is 0.15-0.30 mm, and < 0.05 mm in microspheres.

Diagnosis
Test ovoid, uniserial, early chambers strongly embracing and increasing rapidly in diameter; sutures horizontal and flush; aperture terminal with radiate slits (Hayward et al. 2012).

Remarks
Variation within this species has been noted as varieties A and B (Eiland & Guðmundsson 2004): variety A has embracing rectilinear chambers, where the height of the last chamber may constitute more than half of the test length; whereas the B variety has less embracing chambers, with irregularly slanted sutures. Frequent intermediate forms render distinction between these varieties rather arbitrary. Available material of this species comprised 717 specimens. Diameter of proloculus varies between 0.25-0.40 mm.

Species distributions and environmental parameters
Foraminifera occurred in 879 of the of the 1031 BIOICE zoological samples; consistently in the RP-and detritus sledges (362 and 335 samples, respectively), but only sporadically in the Agassiz trawl and the triangle dredge (95 and 87 samples, respectively), because of the large mesh size of the tail bags. Over 415 000 foraminifera specimens were picked from all samples, with an average of about 470 specimens per sample. The larger species of Nodosariinae were present in 56% (492) of the samples that contained foraminifera, comprising 7598 Nodosariinae specimens, about 1.5% of all picked foraminifera of the BIOICE samples. Most specimens of Nodosariinae were collected with RP-and detritus sledges (271 and 176 samples, respectively), but intermittently in the triangle dredge and the Agassiz trawl (25 and 20 samples, respectively).
The Nodosariinae occur ubiquitously within the sampling area, albeit appearing rather sparsely in the strait between Iceland and Greenland (the Denmark Strait) and along the shallow shelf (< 400 m) between Iceland and the Faroe Islands (Fig. 16A). Most samples of Nodosariinae are from shallow waters (< 500 m) where sampling intensity is highest and have relatively higher occupancy in samples below 1200 m depth (Fig. 16B).
Species richness of Nodosariinae in samples from the RP-and detritus sledges varies from 1-7 species per sample (Fig. 17A) and increased sample occupancy of a species (Fig. 17B) tends to result in higher number of picked specimens for that species (Fig. 17C). All the 14 species occur in more than one sample (no singletons), and only one species (G. semirugosus?) is present in two samples (one doubleton). That satisfies the condition of a heuristic stopping rule, i.e., a satisfactory biodiversity sampling is achieved when every species is represented by at least two individuals, or occurs in two samples (no singletons), after which further sampling is superfluous, as estimated species number equals observed species number (Chao et al. 2009). Nonparametric estimation of species richness for incidence data (Chao2) indicates negligible probability that further sampling with present methods will reveal more undetected species. Each of the species occurs within different and overlapping ranges of water depth, temperature, and salinity (Fig. 18). Distributions of the species are variously restricted to only a part of the sampling area , which seems to coincide with dispersion of near bottom water temperature and salinity (Fig. 16C). The 2-dimensional NMDS ordination converged with acceptable stress (0.094). The gradient axes were well correlated (Kendall's t = 0.92 and 0.32 for first and second axes, respectively) with an independent analysis by detrended correspondence analysis (DCA; vegan function decorana). Gradient fits between NMDS scores and latitude, longitude, and depth were highly significant (Fig. 22); we could not include temperature and salinity in this analysis due to missing values. The close alignment between latitude and longitude vectors reflect that the main geographical gradient was running diagonally from SE to NW, probably reflecting the distinct water masses on opposite sides of the Greenland-Scotland Ridge (Fig. 16C).
The species of Nodosariine show different distribution patterns. Two species, G. guttifera and D. mutabilis, are mostly confined to the medium salinity and cooler deep waters south off Iceland, albeit the former occurs also in shallower waters (Figs 18,. Two species, N. haliensis and N. incerta occur mainly in shallow and temperate to cool waters northeast and east of Iceland, the latter with some scattered occurrences south and west off Iceland (Figs 18,. The species D. frobisherensis and D. elegans occur mainly in the deep and shallow waters north and east of Iceland, with scattered occurrences in the more temperate and shallower to deep waters south of Iceland (Figs 18,. Three species, N. subsoluta, D. antennula and D. antarctica occur mainly in shallow waters south and west off Iceland ( Fig. 20A-C), each occurring within somewhat different ranges of physical variables (Fig. 18). The species Dentalina obliqua and Pseudonodosaria subannulata are widely dispersed in Icelandic waters (Fig. 20E-F), albeit P. subannulata seems more ubiquitous and spans a wider range of physical variables (Fig. 18). The main distribution of Dentalina filiformis and Grigelis pyrula occur in temperate and shallow waters south and west of Iceland (Figs 18, 21A-B). The distribution of Grigelis semirugosus? is known from two locations only (Fig. 20D), leaving little room for further inferences on wider distribution.

Aspects of sampling
Taxonomic knowledge of Recent Nodosariinae in the North Atlantic and the Greenland-Iceland-Norwegian seas (GIN) is mainly based on classical publications in the late 19 th and the early 20 th centuries (Brady 1884;Goës 1894;Cushman 1923;Loeblich & Tappan 1953). These early studies relied mostly on dredging gears that cover wide areas of the sea floor as well as other large-volume samplers, which yielded rich material of Nodosariinae and other large foraminifera. Later and more recent studies on foraminifera in Icelandic waters and nearby seas have focused on quantitative core samplers or smaller grabs (< 0.05 m 2 ), often processing less than a few hundred cm 3 of sediment through finely gridded (> 63 µm) sieve series (Adams & Frampton 1962;Belanger & Streeter 1980;Sejrup et al. 1981;Mackensen et al. 1985;Mackensen 1987;Wagener 1988;Rytter et al. 2002;Murray 2003;Jennings et al. 2004;Bouchet et al. 2012). Some of these studies used a standard Shipek grab (Flannagan 1970) with a sampling surface of 0.04 m 2 , or smaller subsamples of 0.004 m 2 from large box corers, and indicated that the larger species of Nodosariinae are either absent or very rare (Mackensen 1985(Mackensen , 1987Rytter et al. 2002;Jennings et al. 2004). This is expected as the larger foraminifera (> 0.5 mm) rarely appear in a small sample volume of a few cm 3 (Bouchet et al. 2012). In addition, studies of macrofauna (> 0.125-0.5 mm) in the North Atlantic and the GIN seas, have often omitted larger foraminifera as a part of the studies, since these are not metazoans (Nørrevang et al. 1994;Bluhm et al. 2011;Buhl-Mortensen et al. 2015).
The rich material of Nodosariinae in the present study was obtained after processing large sample volumes, sometimes several liters. That a large sample volume is needed to detect Nodosariinae may perhaps stem from different dispersion patterns compared to smaller foraminifera. These relatively large foraminifera might be more widely dispersed, possibly occurring in denser aggregation, which in turn may have patchy distribution. Species with such heterogenous dispersion patterns are more effectively collected with samplers that cover large surface areas. During the MAREANO project in Norwegian waters, larger infaunal species (> 0.5 mm) were collected with a large Van Veen grab, that cover a sampling area of 0.25 m 2 (Buhl-Mortensen et al. 2015) and with a volume capacity of 80 liters. Foraminifera were picked from a few of these samples, along with other macrofaunal taxa which resulted in rich material of Nodosariinae, using the same sorting procedure as for the BIOICE samples. The RPsledge is designed to sample waters that lie slightly above the bottom (26-59 cm). Nevertheless it also samples the bottom dwelling organisms, as surface sediments are stirred up because a turbulence wave is Fig. 18. Kernel density smoothed distribution (violin plots) of species along physical gradients. A. Depth (m). B. Temperature (°C). C. Salinity (ppt); omitted are two outlier salinity records for N. haliensis Eiland & Guðmundsson, 2004 (33.98 and 34.34 ppt) and one for P. subannulata (Cushman, 1923) (Linnaeus, 1758). F. Pseudonodosaria subannulata (Cushman, 1923). created in front of the sledge, as it is towed along the sea floor (Rothlisberg & Pearcy 1976;Buhl-Jensen 1986;Brattegard & Fosså 1991). In the BIOICE program the RP-and the detritus sledges were routinely towed for about 20 minutes (bottom time) at a speed of one knot (0.5144 m/s), resulting in a potential sampling surface of about 600 m 2 of the sea floor. A closing mechanism of the RP-sledge ensures that it is only open while in contact with the sea floor, with a potential filtering capacity of about 250 m 3 of water, after 20 min towing at 1 knot (Rothlisberg & Pearcy 1976).

Bottom waters and ocean currents off Iceland
The study area encompasses a large part of the GSR which forms a barrier that affects major ocean currents and shapes the physical conditions of the near bottom waters of the North Atlantic. The deep basins north of the GSR in the Nordic Seas, at water depths greater than 600-1200 m, the near bottom water temperature ranges from -0.5°C to -1.5°C with a of low salinity of < 35 ppt. In contrast, the deepsea basins south of the GSR, near bottom water is more saline (> 35 ppt) with an average temperature around 1°C in the Irminger Basin, and about 2°C in the Iceland Basin (Fig. 16C). As a result, the nearbottom water masses in the deep basins north and south of GSR hold major temperature differences (Hansen & Østerhus 2000;Jochumsen et al. 2016).
The cold and dense deep waters north of the GSR flow southward along the sea floor through the deepest sills of the GSR, i.e., in the Faroe-Shetland Trough, the strait between Iceland and the Faroe Islands, and in the Denmark Strait (Fig. 23). These different overflows jointly form a cold overflow bottom water current system, that is directed westward along the southern slopes of the GSR, at about the 1000 m depth contour, and then turn southward as the current reaches the continental slopes east off Greenland and join the overflow branch that falls through the Denmark Strait (Dickson et al. 1990;Hansen & Østerhus 2000;Hansen et al. 2001). The recently discovered North Icelandic Jet bottom water current (NIJ), centered at the 650 m isobath at the continental slope north off Iceland, is a significant source of dense water to the overflow plume passing through Denmark Strait, and similarly the Iceland-Faroe Slope Jet (IFSJ) adds to the overflow between Iceland and Faroe Islands and to the overflow in the Faroe-Shetland Trough (Jochumsen et al. 2017;Semper et al. 2019).
The near bottom water temperature and salinity on the Icelandic shelf (< 500 m), and to some extent on the upper parts of the GSR slopes, is largely controlled by near surface currents, i.e., the temperate Irminger Current and the cold East Iceland Current, which is a branch off the East Greenland Current (Fig. 23) (Valdimarsson et al. 2012;Logemann et al. 2013). The annual mean bottom water temperature of the Iceland shelve is highest southwest and south off Iceland (> 8°C), where the Irminger Current reaches the Icelandic shelf (Figs 16C, 23). This current gradually cools as it flows westward in a clockwise direction around Iceland until it mixes with the cool East Iceland current (2°C to 3°C) that bathes the northeast and eastern shelve off Iceland. The location where the Irminger current meets the East Iceland and East Greenland currents, known as the Polar Front (Fig. 23), has fluctuated for the last 6000 years (Eiríksson et al. 2000;Símonarson et al. 2021), and for the last decades the effects of the cold East Iceland Current have diminished, especially at the northeastern part of the shelf (Astthorsson et al. 2007). Nevertheless, the East Iceland Current still directs relatively cold waters to the eastern parts of the Icelandic shelf (Stefansson 1962;Stefansson & Jónsdóttir 1974;Hansen 1985;Malmberg 1985;Malmberg & Désert 1999;Eiríksson et al. 2000Eiríksson et al. , 2011Hansen & Østerhus 2000;Jochumsen et al. 2016).

Relation of species distributions to water masses and ocean currents
Differences in near bottom water properties associated with the GSR shape the well-known biogeographic boundaries of the benthic faunas of the Boreal and Arctic regions (Einarsson1948 ;Ekman 1953;Briggs 1970Briggs , 1974G. Guðmundsson 1998;Weisshappel & Svavarsson 1998;Brix & Svavarsson 2010). The overall distributions of the Nodosariinae species seem to follow the steep gradients in hydrographic properties, sea floor elevation, and major sea currents within the study area (Figs 16C,(22)(23).
Distributions of eight species seem restricted to the cool and temperate waters at the southern side of the GSR, and without any traces of occurring in the deep cold waters north of the GSR or at northern and eastern parts of the Icelandic shelf, where colder bottom waters are prevalent (Figs 19A 21A-B). Two of these species (G. guttifera and D. mutabilis) occur mostly in deep waters, whereas six are mainly found in shallower waters (D. antarctica, D. antennula, D. filiformis, G. pyrula, N. subsoluta, G. semirugosus?), although the depth distribution of these species overlap to varying degree (Fig. 18A).
Two species, N. haliensis and N. incerta, occur mainly on the shelf area north and east of Iceland ( Fig. 19C-D), where the warmer Irminger current meets the colder East Iceland Current, also known as the Polar Front (Símonarson et al. 2021). Both species occur in shallower waters than 1300 m but occur nevertheless over a wide range of temperature and salinity ( Fig. 18B-C), covering nearly the whole temperature and salinity spectrum of the greater shelf area around Iceland. That the distribution of these two species is mostly limited to the northern and eastern part of the Icelandic shelf might partly be their tolerance to highly fluctuating oceanographic conditions in this area, which is influenced by interannual variations in influx of Atlantic, Arctic and even Polar water masses (Astthorsson et al. 2007). Annual primary production is generally much lower and more variable north and east off Iceland compared to the southern and western shelf areas, or 4.3 and 9.2 mg C -1 m -3 h -1 , respectively (K. . The lower primary production is partly because of a more stratified water column that prevents adequate transfer of nutrients to the surface layers. This may, to some degree, shape the benthic habitat type of the north-eastern shelf regions to be more suitable to N. haliensis and N. incerta. Two species, D. obliqua and P. subannulata, are frequent and widely dispersed within the sampling area ( Fig. 20E-F), especially P. subannulata which occurs within the whole range of depth, temperature and salinity that is spanned by the BIOICE samples (Fig. 18). The apparent high morphologic variation within P. subannulata, and wide tolerance limits to physical variables, might indicate that it is a complex of two or more species. D. obliqua occurs mainly in shallower waters than 1500 m but with a wide tolerance range for temperature and salinity (Fig. 18), that are characteristic of the Icelandic shelf (Fig. 16C).
The relatively strong overflow current of cold bottom waters from the north and to the southern sides of the GSR may facilitate southward dispersal and sporadic distribution of deep water arctic species, especially for those that have a floating larval stage (Bouchet & Warén 1979). The annual mean core speed of the bottom water current in the Denmark Strait is 25-30 cm s -1 , with maximum speed of 100 cm s -1 off East Greenland (Dickson et al. 1990;Jochumsen et al. 2017). The distribution of D. elegans and D. frobisherensis might be shaped by this overflow current (Fig. 23). The main occurrences of both species are in the subzero waters north and east off Iceland, but with scattered occurrences to the south and west off Iceland ( Fig. 19E-F), which might be facilitated by a winnowing and cooling effects of the cold-water overflow currents, and perhaps also providing slightly more favorable environmental conditions at the southern sides of the GSR for cold affinity species like D. elegans and D. frobisherensis (Fig. 18B). A few specimens of D. elegans and D. frobisherensis, collected at the southern side of the GSR, may contain traces cytoplasm, indicating that perhaps these northern species can maintain viable breeding populations.
In this study, specimen counts include both living and dead tests, but worn and damaged tests were excluded. That practice is prone to overestimate distributional ranges of living populations, especially for species residing on the shelf and close to steep continental slopes. It has been demonstrated that dead tests of smaller foraminifers are to some degree translocated from the north to the southward side of the GSR, by the winnowing action of the relatively strong overflow currents (Mackensen 1987) and this would also apply to the larger tests of Nodosariinae. Clear patterns of a variety of surface bedforms and related depositional features match the direction and intensity of the cold overflow currents on the southern flanks of the GSR; like obstacle marks, sand ribbons, barchan-like features, dune fields, and asymmetrical channel fillings (Dorn & Werner 1993). The potential sources of these sands lie upstream, including the IFSJ and the NIJ (Fig. 23) which might be translocating sediment runoffs from the several glacial rivers in the north and east off Iceland, and also foraminifera tests, both dead and living.

Other variables affecting species distributions
Apart from temperature, salinity and depth, other environmental parameters may affect the local distribution of the species of Nodosariinae at hand, like sediment characteristics, local influx of particulate organic matter, predation, and competition, as has been suggested for other foraminifers (Svavarsson et al. 1993;Guðmundsson et al. 2000Guðmundsson et al. , 2003Rytter et al. 2002;Jennings et al. 2004). Little is known about life habits of the Nodosariinae, nor if the relatively large test has some adaptive significance, perhaps serving as protective case around accumulated reserves of nutrients in the cytoplasm. The function of the peculiar radiate aperture is largely unknown: it might serve as a defence mechanism to keep out predators or parasites from entering the test, similar to what has been suggested for the apertural toothplate in Globobulimina turgida (Bailey, 1851) (Glock et al. 2019); or perhaps radiate apertures serve some purpose in a feeding strategy of burrowing species. A box core sample from 1075 m depth, off Nova Scotia contained living Pseudonodosaria sp. with a maximum frequency around 2 cm depth in the sediment (Corliss 1991). The morphology of Nodosariinae shares some of the characteristics that have been associated with a burrowing habit, like circular circumference, rectilinear chamber arrangement and evenly dispersed pores over the test (Corliss 1985). If the species of Nodosariinae have a burrowing habit, then the physicochemical properties of the sediment may significantly affect their distribution, as has been suggested for other infaunal foraminifera (Corliss 1985;Jorissen et al. 1995). Nodosariinae are often absent in samples taken in shallower waters, especially in the Denmark Strait, between Iceland and Greenland, and in the sound between Iceland and the Faroe Islands (Fig. 16A), perhaps because the bottom types in these areas consist mainly of coarse sediments and sand (ICES 2019), where strong currents of COBW are prevailing (Fig. 23).
Mean annual primary production is higher in the Atlantic waters off the southern and western shores, than in the more variable waters north and east off Iceland, and it is higher closer to land than farther offshore (Astthorsson et al. 2007). This may also contribute to variable occurrence and abundance of the species of Nodosariinae within the sampling area (Fig. 16A), along with effects of predation and competition, although the extent of such relations remains unclear.
Each of the fourteen species of Nodosariinae is present in two or more samples, where G. semirugosus? has the lowest incidence, occurring in only two samples and the Chao2 estimator of species richness indicates that further sampling using the present procedure will hardly unveil any undetected species. However, several shallow water species of Nodosariinae that occur in the east and the west Atlantic were not encountered in the present study, although the environmental conditions in these areas resemble those of the Icelandic waters. The species Dentalina melvillensis and Dentalina ittai (now Botuloides ittai) described by Loeblich & Tappan (1953) from shallow waters (< 100 m) off Baffin Island, in Frobisher Bay, and off east Greenland, were not found in the present material. The species Nodosaria emaciata (Reuss, 1851) has been recorded in samples from the Barents Sea and in the eastern and the western parts of the North Atlantic, and Pseudonodosaria aequalis (Reuss, 1863) has been reported from the Arctic and off Norway, but not from the western side of the Atlantic (Cushman 1923). It seems that these species are absent from Icelandic waters, although the possibility remains that further sampling with different gears and sample processing might unveil these and some other little known Nodosariinae in Icelandic waters.