Biogeography and Species Durations of Selected Cenozoic Shallow and Deep-Water Smaller Calcareous Benthic Foraminifera – A Review

This review has been prepared to pay tribute to the enormous contribution of Marty Buzas to foraminiferal research over more than six decades. BWH was his first post-doctoral fellow at the Smithsonian Institution, and they continued spasmodic collaboration ever since (Hayward & Buzas, 1979; Hayward et al., 2003, 2010, 2014; Buzas et al., 2007, 2013). Marty, along with his second post-doctoral fellow, Steve Culver, has been interested in analysing the biogeography, originations, and species durations of modern smaller benthic foraminifera for several decades (e.g., Buzas & Culver, 1980, 1984, 1986, 1990, 1998, 2001; Culver & Buzas, 1981a, b, 1982a, 1999). These studies were based on an enormous database, assembled by Culver, of modern records of benthic foraminifera from thousands of samples right around North America which had been studied by dozens of previous workers (e.g., Culver & Buzas, 1980, 1981c, 1982b, 1985, 1986, 1987). For the dataset to be useful, all species identifications were reassessed and standardised (e.g., Culver et al., 1987)—an enormous task only possible if undertaken by one or several people using consistent criteria and taxonomy.

Marty's role was to recognise the possibilities that this standardised database could be used to answer some major questions about species distributions, evolution, and dispersal and to use his statistical analytical skills to help address them. These authors used cluster analysis on presence/absence data to identify modern benthic foraminiferal provinces (clusters) around the coasts of North and Central America (e.g., Buzas & Culver, 1980, 1990). Many workers have used a criterion of >25–50% endemic species to identify separate macrofaunal biogeographic provinces (e.g., Kaufmann & Scott, 1976), but Buzas and Culver used the more flexible definition of provinces proposed by Valentine (1968) as “collections of communities associated in space and time.” A total of 11 shallow-water (<200 m) provinces were recognised right around the North American continent as well as a number of more-widespread bathyal and abyssal provinces (>200 m depth; e.g., Culver & Buzas, 1999). They concluded that temperature was the major driver of these biogeographic provinces but that their boundaries often coincided with those of major water masses (e.g., Buzas & Culver, 1980, 1990).

Buzas and Culver extended the standardised taxonomy into the fossil record of their extant eastern North American species and used this to investigate species durations and possible centres of origin for their benthic species. After analysis of their data, Buzas & Culver (1989, 2009) concluded that there was no strong evidence to support the hypothesis of benthic foraminifera having one more likely centre of origin and noted that many species appeared to disperse rapidly after their origination (Buzas & Culver, 1989). They also used these data to conclude that commonly occurring extant species have a far higher percentage with a fossil record than those that are less frequently encountered (Buzas & Culver, 1991) and that both common and rare extant species on the Atlantic coast of North America had the same average partial species duration (Buzas & Culver, 1989). These authors also found that the partial species durations of shelf-restricted (<200 m) species was slightly shorter than for those living in deep-water (>200 m; Buzas & Culver, 1984).

The prime requirement for the documentation and understanding of the truly global biogeographic and paleobiogeographic distribution patterns of organisms is that their taxonomy is standardised and consistent throughout (Culver et al., 1987; Gooday & Jorissen, 2011). The existence of numerous regional taxonomies of Cenozoic smaller benthic foraminifera has always been a hindrance to global biogeographic and paleobiogeographical understanding. Until recent decades this was because of the lack of easily available and widespread literature, often published in many different languages. A second reason for the taxonomic confusion was the existence of a wide variety of morphospecies approaches. At one end of the spectrum, the “lumpers” considered that most of the morphological variability was a result of environmental drivers (e.g., Parker, Jones, Williamson, Carpenter), whereas the “splitters” considered that even the smallest differences between test characters was genetically-driven with many distinct species (e.g., Cushman, McCulloch; Cifelli, 1990). For example, in the 1980s–2003 many workers around the world identified only one extant species of the genus Ammonia (as A. beccarii) or three forms or species of it (f. beccarii, f. parkinsoniana, f. tepida), yet 110 extant Ammonia species had been described and named, some of which were being used in different regions (e.g., Hayward et al., 2021).

With the growth of the internet, establishment of taxonomic databases such as Forams in WoRMS (Hayward et al., 2023a), and the availability of most foraminifera literature in digital format online, the first cause of the taxonomic confusion has now largely disappeared. The second reason has also been resolved, at least in part, through the results of molecular taxonomy, which now shows that in large part the “splitters” were correct and minor variations in morphology usually reflect genetically separate species, with occasional examples of cryptic molecular species having little or no morphological differences in their tests (e.g., Darling et al., 2016; Hayward et al., 2021; Holzmann et al., 2022). Obtaining genetic material from live species of imaged smaller benthic foraminifera is an enormous task. Thus, only the most abundant and easily collected near-shore, smaller benthic foraminiferal groups (e.g., Ammoniidae, Elphidiidae) have anywhere near a significant number of sequenced specimens to make them useful in biogeographic studies, and even now these databanks are far from complete (fewer than 30% of species sequenced; e.g., Hayward et al., 2021, 2023b).

Until now, only a few families or genera of Cenozoic smaller benthic foraminifera have had modern (last 50 years) global reviews of the taxonomy and biogeography of their morphospecies or significant documentation of their molecular species. This review is by choice restricted to these groups, with examples from sheltered near-shore, open shelf and deep-sea environments from both the extant and fossil record. The bathymetric terminology used in this review is shown in Figure 1 .

Most of the previous work on the biogeography and species duration of smaller calcareous benthic foraminifera has been undertaken by Buzas and Culver, as outlined above. Prior to their studies, Cushman (1948) had produced a map of global benthic foraminiferal provinces based purely on his extensive knowledge of modern foraminiferal distribution. He provided no faunal criteria for how he recognised and separated these provinces. He recognised cosmopolitan cold-water faunas in high latitudes in the northern and southern hemispheres, and four warm-water provinces, each largely restricted to one ocean, although the Atlantic had a western and eastern province. Boltovskoy & Wright (1976) refined Cushman's map and recognised two cold-water, eight temperate-water, and seven warm-water provinces. Like Cushman, Boltovskoy & Wright's (1976) provinces were recognised qualitatively based on improved knowledge of the distributions of modern benthic foraminifera. Boltovskoy (1976) recognised four benthic foraminiferal provinces around the South American continent and Boltovskoy et al. (1980) further subdivided his Argentinian Province into two subprovinces.

There have been a number of studies based on literature reviews which accept the identifications of many workers and their species’ records from all around the world. Sen Gupta & Smith (2010) and Hayward et al. (2010) assessed the relative abundance of endemic and cosmopolitan species in their total extant foraminiferal faunas from the Gulf of Mexico (987 spp.) and New Zealand (580 spp.). Both estimated that a high proportion of their species had a cosmopolitan distribution (35% and 64%, respectively) and both estimated a relatively low proportion that were restricted to their respective study regions (7 and 9%, respectively. Both recorded a greater proportion of endemic species at shelf depths (<200 m) and more cosmopolitan in the deep sea. Both studies found that the majority of endemic species belonged to two orders: Miliolida and Rotaliida. Recent molecular studies of just one of the numerous families present in these studies, Ammoniidae (Hayward et al., 2021), shows the unreliability of identifications in the unstandardised international literature, and that these studies probably underestimate the proportion of endemic species by 50% and overestimate the cosmopolitan. If each study region was doubled in size to include the Caribbean or the rest of the Southwest Pacific, again the number of endemic species would likely be more than doubled.

Gooday & Jorissen (2011) reviewed the biogeography of extant foraminifera in the deep sea based on the international literature and the established distributions of a relatively small number of well-known morphospecies. They concluded that the majority of coastal and sublittoral species have restricted distributions with an increasing proportion of widespread, cosmopolitan species with increasing depth down the continental slope to the abyssal depths (>2000 m).

In contrast to the smaller benthic foraminifera, two groups of Cenozoic foraminifera extensively studied for their application in biostratigraphy have relatively consistent global taxonomy and well-documented biogeographic and paleobiogeographic distribution patterns. These are the mostly extinct, larger benthic foraminifera of warm waters (e.g., Pignatti, 1994; Langer & Hottinger, 2000; Boudagher-Fadel & Price, 2013, 2014, 2022; Förderer et al., 2018) and the planktic foraminifera. In this latter group, the majority of the extinct Cenozoic morphospecies have been rigorously reviewed by teams of global experts over the past three decades (e.g., Olsson et al., 1999; Pearson et al., 2006; Wade et al., 2018), and the extant species have mostly been sequenced with the recognition of a number of cryptic, often biogeographically separate, molecular species (e.g., Brummer & Kučera, 2022).

The modern deep abyssal (sub-calcium carbonate compensation depth) agglutinated foraminifera have a largely cosmopolitan distribution and mostly originated in the Cretaceous and Paleogene (Kaminski et al., 1999). Species in this group adapted to more eutrophic conditions have changed little since that time but oligotrophic faunas have evolved and diversified during intervals of more oligotrophic conditions (Kaminski et al., 1999).

Hayward et al. (2021) presented the results of more than 20 years of collecting and sequencing specimens of the common intertidal and inner shelf (mostly shallower than 50 m) genus Ammonia combined with a review of the morphology of all described and well-illustrated (with SEM images) extant specimens and species. Twenty-seven species/infraspecies were recognised by DNA sequencing and morphologically distinguishable. An additional 33 morphospecies/infraspecies were recognisably distinct, and Hayward et al. (2021) speculated that the total number of extant Ammonia species globally was probably nearer 90.

Of the 60 recognised species of Ammonia, there are two widely distributed and commonly encountered species (A. veneta, A. batava) that have been recorded in >19 of their 52 global regions (Fig. 2) with 65% (39) rare species encountered in only a few places (three or fewer regions; Hayward et al., 2021, supplementary appendix 3). This is similar to the pattern observed by Buzas et al. (1982) in all smaller benthic foraminifera along the east coast of North America. Only one species is (A. veneta, T1) is truly cosmopolitan (Fig. 3 A) having been recorded from 72% of their regions, many as DNA-sequenced records.

Of the other species, molecular studies have shown that two (A. aberdoveyensis, T2 and A. batava, T3) occur on both sides of the North Atlantic (Figs. 3A, B); one (A. kitazatoi, T10) on both sides of the North Pacific (Fig. 3B); and one (A. aoteana, T5) on both sides of the South Pacific Ocean (Fig. 3B). The morphologically distinct species A. convexa (T13) has the second most widespread distribution with records stretching from the Red Sea, across the northern Indian Ocean, through East Indies and equatorial West Pacific to the central South Pacific islands of French Polynesia (Fig. 3B). Another morphospecies with a widely recorded distribution is A. ariakensis with occurrences stretching from the northeast Indian Ocean, through the East Indies, right around Australia, north to Japan, and west to French Polynesia (Fig. 3C). The distributions of the morphospecies A. convexa and A. ariakensis have not been tested by molecular studies.

Based on presence/absence data of all the Ammonia species, ten “biogeographic provinces” were recognised (Hayward et al., 2021, text-figure 17). The provinces with the largest diversity of Ammonia species are all tropical–subtropical regions – northwest Pacific (16 spp.), Australian (15 spp.), Mediterranean (15 spp.), central Americas (11 spp.), Middle East (11 spp.), Southeast Asia (10 spp.), and Indo-Pacific (9 spp.). The more temperate provinces in the North Atlantic, South Atlantic, northeast Pacific and South Pacific each have 2–8 Ammonia species recorded (Fig. 4).

Hayward et al. (1997) presented a taxonomic review of the extant morphospecies of the genus Elphidium in the Southwest Pacific (33 spp.). They also reviewed the taxonomy and well-documented time ranges from New Zealand of the 18 extinct and ten extant species (Fig. 5 ). Cenozoic foraminiferal time ranges in New Zealand are among the most complete for any region in the world. They are based on 70 years of species identifications from inner shelf-abyssal faunas throughout the Cretaceous–Recent documented in the New Zealand Fossil Record File and summarised in Hayward et al. (2012b). Today Elphidium species are largely restricted to inner–mid shelf depths (0–100 m) with near monospecific Elphidium faunas sometimes found in sheltered shallow subtidal and intertidal environments (Hayward, 2014). The main exception are two species of mid-Eocene Elphidium (E. hampdenense, E. saginatum) that are restricted to bathyal faunas.

The New Zealand fossil record is interpreted to consist of parts of seven branches of the global Elphidium family tree (Hayward et al., 1997). Each branch is inferred to have started in the New Zealand fossil record with a successful immigration event from across the oceans (Fig. 5). Thereafter the branch evolved within New Zealand waters with species originations and local extinctions. Species from four branches are still present around New Zealand today with one probable recent ship's ballast introduction (E. vellai) from Japan (Hayward et al., 1999).

This review is based on the morphological and numerical taxonomic study of Hayward (1990) as no extant specimens have been sequenced so far. The Bolivinellidae are a group of rather rare, tropical and subtropical biserial foraminifera having cribrate apertures with rows of fine beads radiating out from the apertural area—an adaptation for their unusual plastogammic style of reproduction in the gamont generation (Hayward & Brazier, 1980, plate 1). They live in normal marine salinity at inner shelf depths (0–60 m), possibly as deep as 100 m in the shelter of coral reef lagoons, on the seaward slopes of coral reefs and around the open mouths of bays and harbours (Hayward & Brazier, 1980).

Nine extant morphospecies are restricted to the Indian and Pacific Oceans (Fig. 6 ). The most common and widespread species (Rugobolivinella elegans) occurs throughout the range of the whole family, except for the southern half of Australia. It is the only species recorded from the Indian Ocean. The second-most widespread species (Punctobolivinella philippinensis) has many records from the equatorial West Pacific, from southern Japan in the north to Lord Howe Island in the south and Samoa in the east (Fig. 6). The third most widespread species (Bolivinella folium) occurs along the entire southern coast of Australia from West Australia to New South Wales (Fig. 6). The remaining six species all appear to be restricted to small areas or island groups and are local endemics (Fig. 6).

The global fossil Cenozoic record is more extensive than in the recent with seven genera and 52 morphospecies/infraspecies recognised (Hayward, 1990; Popescu & Crihan, 2005). The origins of the family are unknown with the two oldest records from the mid-Eocene of the Tethys (India, Nodobolivinella jhingrani) and Central America (Bolivinella n.sp. C; Fig. 7 ). By the late Eocene, there were seven endemic species (from four genera) in the Caribbean region, five species (four endemic) in the European part of the Tethys, one species in the equatorial Pacific (Bolivinella cf. cubensis), and one endemic species (Quasibolivinella taylori) in Australia (Fig. 7).

The time of greatest geographic spread was in the Oligocene with six species in each of the European part of the Tethys, central America and New Zealand, four in Australia, and one in the equatorial Pacific (Fig. 7). In the Miocene, eight species have been recorded from Australia (five endemic), seven each from New Zealand (five endemic) and Europe (all endemic), and five from the equatorial Pacific (two endemic; Fig. 7). Thus, the fossil Cenozoic record of the family shows changing biogeographic patterns with dispersal around the world in the late Eocene, diversification in the Oligocene and Miocene, and a progressive withdrawal from the Americas by the end of the Oligocene, and from Europe and New Zealand near the start of the late Miocene (Fig. 7). By the Pliocene, the family had attained its present biogeographic record, restricted to the Indopacific and Australia. Its disappearance from Europe and New Zealand at the end of the mid Miocene is inferred to be due to cooling temperatures in these regions. Also noteworthy is the observation that this family never appears to have colonised the east Pacific, the South Atlantic, or the western North Atlantic. Records from the Indian Ocean south of the equator are limited to the Holocene.

As with the extant species of this family, the extinct fossil species include a few (three) common and geographically-widespread morphospecies that occur in several oceanic regions with differing time ranges that may reflect dispersal events (Nodobolivinella nodosa, Rugobolivinella rugosa, Inflatobolivinella subrugosa; Figs. 7, 8 ). The geographically closer regions of Australia and New Zealand share two species (Q. taylori, N. nodosa) and jointly share with the equatorial Pacific a further two species (R. rugosa, Bolivinella profolium). Thirty-six (65%) of 55 bolivinellid morphospecies/infraspecies are rather rare and endemic to just one of the global regions of central America, Europe, Australia, New Zealand, and the equatorial Pacific (Hayward, 1990).

At the generic level, three genera (Rugobolivinella, Inflatobolivinella, Nodobolivinella) occur in all the regions colonised by the family, at least during the Oligocene or early Miocene (Figs. 7, 8), although the genus Nodobolivinella is only present (as two species) in the European Tethys during the late Eocene and Oligocene. The other four genera are more restricted in their geographic range throughout the Cenozoic: Rhombobolivinella occurring mostly in Europe and two species in the late Miocene of the Pacific; Quasibolivinella mostly occurring in Australasia but with its earliest species record in the late Eocene of the Caribbean; Bolivinella is primarily an Australian genus with several sporadic Miocene records in nearby New Zealand and the equatorial Pacific; and Punctobolivinella mainly occurring in the tropical and subtropical West Pacific including occurrences off the coast of northern Australia (Hayward, 1990).

This is another inner shelf, warm-water genus that is inferred to have lived in similar environments to the bolivinellids (Wendler et al., 2011). Unlike the bolivinellids, Tubulogenerina is extinct, and this review is based on the morphospecies identifications and inferred phylogeny of Gibson (1987, 1989) and Gibson et al. (1991). Tubulogenerina differs from the externally similar Nodosarella by the presence of internal toothplates and frequently more than one aperture (Gibson, 1987). The oldest, possibly ancestral, species of Tubulogenerina comes from the mid-Cretaceous of East Africa (Wendler et al., 2011), but there is a gap in the record until the early Eocene. Here, we only consider the paleobiogeography of the Cenozoic species of this genus (Gibson, 1989; Gibson et al., 1991).

In the Cenozoic, Tubulogenerina first appears in the Eocene of Europe, and the fossil record suggests that it dispersed to North America in the late Eocene, the equatorial Pacific in the Oligocene and to Australasia and India in the early to mid-Miocene (Fig. 9 ). As with the bolivinellids, this genus disappeared from different regions at different times. They became extinct in Europe and North America at the end of the Oligocene, from New Zealand in the early Miocene, Australia, and India in the mid Miocene, and survived longest in the equatorial Pacific until the early Pliocene (Gibson, 1989; Gibson et al., 1991).

No morphospecies of Tubulogenerina is shared between these regions. The greatest diversity in a single region at any one time was four species in the late Eocene of Europe and four species in the mid-Miocene of Australia (Fig. 9).

Molecular studies on specimens from Rotaloidea have shown the presence of a number of distinct clades, one of which correlates with the family Notorotaliidae and is a sister to the Elphidiidae (Holzmann & Pawlowski, 2017; Hayward et al., 2023b). Sequenced specimens from the Notorotaliidae clade are assigned to the morphogenera Notorotalia, Buccella, and Porosorotalia. Two unsequenced extant morphogenera are also allocated to this family – Parrellina and Cristatavultus (Hayward et al., 2023b). This biogeographic summary is largely based on the combined molecular and morphological study of extant species of the Notorotaliidae by Hayward et al. (2023b).

The two common extant genera, Buccella and Notorotalia are largely restricted to mid–high latitudes, with Notorotalia confined to the Southern Hemisphere and Buccella predominantly in the Northern Hemisphere but also extending down the coasts of the Americas to southern South America where it is also common (Fig. 10 ). The three species of Parrellina occur around the subtropical to temperate east coast of Australia with records of one of these species (P. papillosa) from the Mediterranean where it has probably been introduced by human activities (Hayward et al., 2023b). The genus Porosorotalia has one extant southern species (P. meridionalis), around South America and one extant northern species (P. makiyamai) around Japan (Fig. 10). The monospecific genus Cristatavultus is the exception in this family with its tropical distribution in the western Pacific (Fig. 10). This latter taxon has not been sequenced and so it is possible that, although morphologically aligned, it belongs in another family. The coasts of Africa, the Mediterranean, and Indian Ocean are essentially devoid of any notorotaliids (Fig. 10).

The 11 extant morphospecies of the genus Notorotalia occur in a globe-encircling belt in the Southern Hemisphere temperate and subantarctic climate zones, north of the Subantarctic Front (Figs. 10–11 A). The centre of diversity is around New Zealand (eight morphospecies) with just one endemic species, each in southern Australia (N. clathrata), southern South America (N. patagonica), and the Kerguelen Islands (N. kerguelenensis; Fig. 11A). Around New Zealand, one species (N. profunda) occurs in deeper water (outer shelf to mid-bathyal, 100–1000 m) whereas the other seven are shelf-restricted and most abundant at inner–mid shelf depths (0–100 m). Of this latter group, four species occur right around New Zealand from North Cape to the subantarctic Auckland and Campbell Islands (Hayward et al., 1999): one (N. olsoni) is restricted to the warm waters of northeastern North Island, one (N. aucklandica) to the subantarctic islands, one (N. inornata) occurs between the southern North Island and the subantarctic islands, and the other (N. zealandica) occurs around both main islands but does not extend into the subantarctic zone (Hayward et al., 1999; Fig. 12 ).

Sixteen extant species of Buccella are recognised by a combination of molecular and morphological studies (Hayward et al., 2023b). They live in a wide range of water depths with their greatest abundance usually at shelf depths (0–200 m). The greatest diversity (six endemic species) occurs around the southern two-thirds of South America, with four rare species and two more common and widespread. Of the latter two, B. peruviana occurs on the Pacific coast whereas B. alvarezii occurs on the South Atlantic coast and around Cape Horn (Fig. 11B). Five rather rare Buccella morphospecies occur along the Pacific coast of central America, with one of these (B. hannai) also occurring sporadically in the Caribbean and up the east coast of the USA as far as Virginia (Fig. 11B). Five species are restricted to the temperate–arctic zones of the Northern Hemisphere. Two are rare and are largely restricted to the northeast (B. troitzkyi) and northwest (B. angulata) Pacific (Fig. 10B). The other three (B. floriformis, B. frigida, B. tenerrima) are widespread in the Arctic Ocean and extend down into the subpolar and temperate parts of the North Atlantic and North Pacific (Fig. 11B).

The fossil record of Buccella is not well documented in contrast to Notorotalia. As with the extant species, the centre of diversity throughout the mid-Eocene–Recent history of Notorotalia has been in New Zealand (43 morphospecies, 35 extinct) compared with Australia (six) and South America (two; Hayward et al., 2012b; Fig. 13 ). Interestingly, New Zealand was also the centre of diversity for the genus Porosorotalia (as Cribrorotalia) in the late Eocene–Miocene (nine species) with only five extinct species described from the northwest Pacific and six from South America (Bertels, 1977; Hayward et al., 2023b). The oldest recorded (possibly ancestral) specimens of Porosorotalia occur in Patagonia. For reasons unknown Porosorotalia became extinct near the start of the late Miocene in Australasia but survived through to the present in the other two regions.

Forty-six Cenozoic morphospecies in this family were recognised in a recent review (Hayward et al., 2012a) based on the records in 38 deep-sea drill holes and the world literature (Hayward et al., 2023a). The earliest species (Staffia tosta) appears in the Paleocene, followed by rapid diversification of the family, especially in the late Eocene, with all but one species (Mucronina resigae) now extinct, with 17 species (37%) disappearing in the last 5 myrs (Pliocene–Pleistocene; Fig. 14 ). Records are most complete for the Pacific and Atlantic Oceans and show that 50% (23) of species occur in both oceans with 11 restricted to the Atlantic and 12 to the Pacific (Fig. 14). It is impossible to determine how complete the stratigraphic records are, even in the Atlantic and Pacific Oceans and therefore how reliable the following observations might be.

Of the species that lived in both oceans, only one (M. dumontana) has the same recorded time range in each. Eight species appeared at approximately the same time in both oceans, suggesting rapid dispersal after origination. Even among the most common species many appear to have evolved in one region and later dispersed to another with more species recorded as first appearing in the Pacific (11 spp.) than first appearing in the Atlantic (five spp.). Of the 17 species that became extinct in the Pliocene–Pleistocene, seven became extinct in both oceans at the same time but not the other ten (Fig. 14).

A qualitative conclusion would be that approximately 50% of the species of Plectofrondiculariidae were widespread, possibly cosmopolitan and present in all oceans at bathyal–abyssal depths with the remainder being restricted to one ocean or part of an ocean.

A global morphotaxonomic review of five families of deep-sea calcareous benthic foraminifera has recently been completed (Hayward et al., 2012a). The review was based on species records from over 1000 Cretaceous–Quaternary faunas in 38 deep-sea cores from all of the world's oceans, including the Mediterranean Sea. Records of well-illustrated specimens from the global literature were also included. These families are characteristically elongate with specialised apertures and are often uniserial. The 203 recognised morphospecies lived at mid-bathyal–abyssal depths, mostly >700 m deep (Hayward et al., 2012a).

The deep-sea core sites span the four major oceans (Atlantic, Pacific, Indian, and Southern; Hayward et al., 2012a). These deep-sea families have a high proportion of widespread species with 40% (81) of the species in all five families recorded in all four oceans, 11% in three oceans, 21% in two, and 28% in just one ocean (Fig. 15). Eighty-six percent of the species are recorded from the Atlantic and 79% from the Pacific—the oceans with the most deep-sea core study sites. The lower percentage of species in the Indian (63%) and Southern (53%) oceans reflects, at least in part, the lower number of study sites in these oceans (Fig. 15).

At the family level, the Glandulonodosariidae and Pleurostomellidae have the greatest proportion of widespread species with 60% of each in all four oceans and 20% or less recorded from just one or two oceans combined (Hayward et al., 2012a). Approximately 35% of the species in each of the other three families are recorded from all four oceans with the Ellipsoidinidae having the greatest proportion of rarer species occurring in just one (43%) or two oceans (21%). The Chrysalogoniidae and Stilostomellidae also have about 50% of their species recorded from only one or two oceans (Fig. 15).

Although 60% of the species in the Pleurostomellidae are recorded in all four oceans, their stratigraphic ranges are often not consistent globally (Fig. 16 ). Some of this may be a result of insufficient study, especially of rarer species; it probably also reflects origination in one region (mostly in the Atlantic or Southern Oceans) and progressive dispersal as well as progressive withdrawal prior to final extinction (most commonly in the Pacific) of many species (Hayward et al., 2012a; Fig. 16). In the present data set, 43% of species that occur in two or more oceans (143 spp.) originated in the Atlantic, 28% in the Southern Ocean and 23% in the Pacific. The earliest withdrawals (local extinctions) are predominantly from the Southern Ocean. Throughout the Paleocene–Miocene, these five deep-water families maintained a cosmopolitan distribution and only in the Pliocene and Pleistocene did their range contract away from the higher southern latitudes (Fig. 16).

The most studied interval for these five deep-water families is the Pliocene–Pleistocene (Hayward et al., 2012a). The geographic distribution of records of some of the more common species are shown in Figure 17 and illustrate their cosmopolitan distributions (e.g., Orthomorphina perversa, Fingerina weaveri, Cribronodosaria stimulea, Ellipsoidina abbreviata, Strictocostella matanzana, S. hyugensis). Some common species appear to have been absent from part of the world oceans (e.g., Carchariostomoides dentaliniformis from the Atlantic; Scallopostoma karreri and Siphonodosaria lohmanni from the North Atlantic; Anastomosa gomphiformis and Epelistoma crassitesta from the central and north Pacific; Siphonodosaria campana from the Indian and South Pacific). The Mediterranean Sea lacked many species in the Pliocene–Pleistocene (Fig. 17)—inferred to be a result of the end Miocene Messinian crisis local extinctions and limited re-invasion from the Atlantic (Hayward et al., 2009).

In two of the few molecular studies on deep-water (>600 m) calcareous benthic foraminifera, Pawlowski et al. (2007) and Lecroq et al. (2009) showed that Arctic- and Antarctic-sourced specimens of three common morphospecies (Cibicides wuellerstorfi, Epistominella exigua, Oridorsalis umbonatus) are molecularly indistinguishable and the molecular species in the deep sea likely have a cosmopolitan distribution and, unlike shallow-water species, have no intrinsic dispersal limitations.

In this review, morphospecies durations have been documented in families and genera that have well-documented time ranges and rigorous taxonomic reviews (Table 1 ). In the inner shelf, warm-water family Bolivinellidae, numerous morphospecies have short known species ranges and a few (mostly widespread) species have longer durations (Fig. 18 ). Twenty-six species/infraspecies (50%) have documented time ranges of <1 myrs and a further ten (19%) have species durations of <3 myrs (Table 1). The longest species durations belong to Rugobolivinella rugosa (26 myrs) and Nodobolivinella jhingrani (20 myrs). The mean species duration of all the morphospecies/infraspecies of Bolivinellidae is 4 myrs. These results are similar to those for the extinct, inner shelf, warm-water genus Tubulogenerina, where the mean species duration is 5 myrs. Nine (50%) of the 18 Cenozoic Tubulogenerina species have short species durations (<3 myrs). The two longest species durations are of T. toddae (23 myrs) in the equatorial Pacific and T. conica (15 myrs) in the Eocene–Oligocene of Europe (Gibson et al., 1991). The average New Zealand total species duration for the 28 species of Elphidiidae is 19 myrs, with an average 12 myrs for the 18 extinct species and an average 34 myrs (double partial species durations) for the ten extant species (Hayward et al., 1997).

The stratigraphic ranges of the endemic morphospecies of the shelf to upper bathyal genus Notorotalia and shelf genus Porosorotalia (as Cribrorotalia) are well documented from New Zealand (Hornibrook, 1996; Hayward et al., 2012b; Fig. 13), but more sporadic and imprecise from Australia and South America. In New Zealand, the mean species duration for extinct Notorotalia species is 7 myr and for Porosorotalia is 10 myrs. The longest species durations for each genus are 11 myrs (N. spinosa) and 28 myrs (P. dorreeni). Another group of mid-shelf–bathyal foraminifera having relatively well-documented stratigraphic ranges globally are the extinct morphospecies of the Plectofrondiculariidae (Fig. 14). Approximately 50% have species durations shorter than 10 myrs (Table 1), but the mean species duration for the 45 species is 20 myrs with the longest ranging species (Staffia tosta) having a 65-myr duration. Many species of Plectofrondiculariidae originated in the late Eocene (15 spp., 33%; Fig. 14) whereas in Notorotalia and Porosorotalia there was a more evenly progressive origination and extinction of species (Fig. 13).

Also available are the documented global time ranges of species in the five extinct deep-water families Chrysalogoniidae, Ellipsoidinidae, Glandulonodosariidae, Pleurostomellidae, and Stilostomellidae all of which lived at mid-bathyal–abyssal depths (>700 m; Hayward et al., 2012a). All had extremely low numbers (1%) of short-lived species (<3 myrs) and together had nearly 50% of their morphospecies with species durations of >50 myrs (Table 1). One species (Nodosarella rotundata) had a species duration of 150 myrs and three (Anastomosa nuttalli, Pleurostomella subnodosa, Strictocostella matanzana) of 120 myrs (Hayward et al., 2012a). The mean species durations in these five families were 41, 44, 46, 47, and 50 myrs with an overall mean of 45 myrs for 207 species (Table 1). Some intervals of enhanced diversification (species origination) are apparent in these deep-water families, such as in the Late Cretaceous in the Ellipsoidinidae (30 spp., 58% of total; figs. 61a–b in Hayward et al., 2012a).

Molecular studies on species of the shallow-water genus Ammonia support the general biogeographic pattern, although not always the detailed species-specific distributions, obtained from morphospecies review (Hayward et al., 2021). This gives us confidence in the recognition of similar patterns in fossil shallow-water and shelf-dwelling smaller calcareous benthic foraminifera only able to be studied through the morphology of their tests.

We acknowledge that the records and time ranges of many of the rarer morphospecies are likely to be incomplete. The fossil-based paleobiogeographies and species durations reviewed above are based on only those studies with the most comprehensive fossil documentation using standardised taxonomy and are the best estimates currently available.

It is clear that widespread, at times trans-oceanic, dispersal of shallow-marine (<200 m) benthic foraminifera has occurred, but many species have remained with restricted geographic distributions both in the fossil record and today. This dispersal most probably occurred by current transport of floating 1–3 chambered microscopic juvenile shells (Alve & Goldstein, 2002, 2003, 2010). Exactly why some shallow-water species have been successful at dispersal (e.g., Ammonia veneta), and many others not, is unclear. It appears that the more abundant species with longer time ranges are the ones that have dispersed most widely, suggesting that successful dispersal and colonisation events are a result of greater numbers of individuals giving a greater chance of dispersal. The timing of successful trans-oceanic dispersal in the Cenozoic, via the Tethys or central American seaway, was at times influenced by plate movement and tectonics.

Compared with the shallow-water-restricted benthic foraminifera, those in the deep sea (mid-bathyal and deeper, >600 m) are predominantly cosmopolitan and have been widely dispersed. The most likely transport agents in the deep sea are deep-water bottom currents carrying along small, suspended, juvenile tests (propagules; Gooday & Jorissen, 2011). Hayward et al. (2012a) showed that after some elongate deep-sea taxa disappeared from the record in higher latitudes during colder periods in the Pliocene to mid-Pleistocene climate cycles, they did not recolonise the sea floor in an up-current direction during the warmer interglacial periods. This observation supports the hypothesised role of bottom currents in rapid deep-sea benthic foraminiferal dispersal.

Our studies suggest that the biogeographic distribution of each individual species is largely independent of any others and that almost every possible pattern of contiguous geographic range exists, confirming the findings of Buzas & Culver (1998). We agree with previous hypotheses that the main drivers of shallow-water benthic foraminiferal biogeography are sea temperature, oceanic barriers, and possibly surface current directions (e.g., Buzas & Culver, 1980; Langer & Hottinger, 2000). Majewski et al. (2021) used molecular and morphotaxonomy to show that the evolution and biogeography of coastal cassidulinid foraminifera in the Southern Ocean was driven by the Neogene climatic history and movement of oceanic fronts in this region. These drivers do not influence the biogeography of the more widespread deep-sea species where we hypothesise that the main drivers may be bottom currents and carbon flux (e.g., Gooday & Jorrisen, 2011; Hayward et al., 2012a).

In addition to the natural modern biogeographic distribution of species, we recognise several disjunct distributions that we attribute to human-assisted introductions, mostly in ship's ballast water but possibly also with macrofauna that has been translocated for aquacultural purposes (e.g., oysters, crabs).

Within the shallow-water genus Ammonia, one commonly occurring molecular species (A. confertitesta, T6) has numerous records around Japan and northeast China and from around northern Europe (Fig. 3C). The European records appear to show that this species was introduced to the Baltic Sea from the northwest Pacific more than 30 years ago and has rapidly spread westwards from there (Hayward et al., 2004; Pawlowski & Holzmann, 2008; Bird et al., 2020; Richirt et al., 2021).

We also infer that records from New Caledonia and New Zealand of the otherwise Atlantic–Mediterranean species Ammonia batava are best explained as recent introductions (Hayward et al., 2021). The New Zealand occurrences of the sheltered, shallow-water Elphidium vellai is inferred to have resulted from a shipping-assisted introduction from Japan (Hayward et al., 1999). The Mediterranean records of the otherwise Australia-restricted Parrellina verriculata are also inferred to be a result of shipping-related introductions (Hayward et al., 2023b).

Previous estimates of species durations have been determined for all extant benthic taxa with fossil records from two regions: eastern North America and New Zealand (Table 1; Buzas & Culver, 1984, 1989; Hayward et al., 2010). These are effectively partial species durations as all are of species still living. An estimate of their full species durations can be made by doubling the mean values using the generalised assumption that the average living species is halfway through its species duration. Collectively, the 267 benthic foraminiferal species from all depths along the northwest margin of the Atlantic Ocean had a mean partial species duration of 21 myrs or an estimated full species duration of 42 myrs. This is comparable to the mean partial species duration of 20 myrs (40 myrs full duration) for the 249 species from all depths around New Zealand (Table 1). Around New Zealand, the estimated full species duration for the 108 dominantly inner–mid shelf (<100 m) was 38 myrs and for the 138 deeper water species was 42 myrs. On the northwest Atlantic margin, the estimated mean full species duration for 75 morphospecies living on the shelf (<200 m) was 32 myrs and for 33 living deeper was 52 myrs (Table 1). It has been suggested by Gooday & Jorissen (2011) that longer species durations and the greater proportion of cosmopolitan foraminiferal species in the deep sea may be due to an inferred dominance of asexual reproduction in the deep sea, which would favour much slower evolution.

In these earlier studies the full species durations for the deeper-water morphospecies are comparable to those determined for our five extinct deep-water families (40–50 myrs) but are considerably longer for the shallow-water (32–38 myrs) than in our family- and genus-based studies (4–20 myrs). The reasons for this are unclear, but maybe the shallow warmer water benthic foraminifera (e.g., Bolivinellidae, Tubulogenerina) evolve faster. However, this does not explain the cooler-water notorotaliids and cosmopolitan plectofrondiculariids. The two studies using extant species in a region did not include species (mostly rare) that did not have a well-documented time range, and this exclusion possibly explains a good deal of the discrepancy as all species (even the rarest with <3 myr documented time ranges) were included in the estimates based on family/genus reviews.

The above discussions on species durations are based on the premise that originations and extinctions have occurred at a steady rate throughout geological time. This is not true, however, as we know in the deep sea, for example, that there have been two intervals of enhanced extinction in the Cenozoic during the Paleoecene–Eocene Thermal Maximum (e.g., Thomas, 1998) and the Mid-Pleistocene Climate Transition (e.g., Hayward et al., 2012a). Our present data are insufficient to prove periods of enhanced origination for benthic foraminifera. Some datasets hint that there may have been certain times of enhanced diversification, such as the late Eocene for the Plectofrondiculariidae (Fig. 14) or late Cretaceous for the Ellipsoidinidae (Hayward et al., 2012a, figs. 61a–b), or enhanced trans-oceanic dispersal of shallow water-species, such as the Oligocene for the Bolivinellidae (Fig. 7).

• There is one known truly cosmopolitan species (Ammonia veneta) in tropical to temperate waters.

• There are several known widely dispersed, but not cosmopolitan, warm-water species (e.g., Rugobolivinella elegans, Rugobolivinella rugosa). The latter extinct species had the longest species duration (29 myrs) in the Bolivinellidae.

• One species (Ammonia batava) occurs along the eastern coast of the Atlantic Ocean in both hemispheres (temperate–tropical) and in the Mediterranean but has not managed to disperse to the other side of the Atlantic.

• Several Northern- or Southern-Hemisphere-restricted subtropical and temperate species are widespread but restricted to one of the major oceans (e.g., Ammonia aberdoveyensis, A. aoteana, A. kitazatoi).

• Several warm-water species (e.g., Ammonia ariakensis, A. convexa, Elphidium crispum, Punctobolivinella philippinensis) are widespread in the Indian and Pacific Oceans but not fully globe encircling.

• The majority of both warm- and cool-water species in these groups are regionally or locally-restricted endemics (92% of Bolivinellidae, 100% of Tubulogenerina, 73% of Ammoniidae).

• The temperate–subtropical notorotaliid genera Notorotalia and Parrellina are restricted to the Southern Hemisphere.

• The temperate notorotaliid genus Porosorotalia has one regionally-restricted extant species in each hemisphere.

• The tropical notorotaliid genus Cristatavultus occurs on both sides of the equator in the West Pacific.

• The dominantly cool-cold water notorotaliid genus Buccella occurs in about half of the world's oceans with its greatest abundance and diversity in the Arctic Ocean and around South America. The genus does not occur in the east Atlantic, Indian Ocean, Mediterranean Sea nor in the tropical and southern West and central Pacific.

• One species (Buccella tenerrima) occurs throughout the Arctic Ocean and on both sides of the North Pacific and another (B. floriformis) is also widespread in the Arctic.

• Another species (Buccella hannai) has sporadic occurrence on the warmer Atlantic and Pacific coasts of central America.

• The majority of notorotaliid species are regionally or locally restricted to one part of the coast of one ocean.

• Approximately 50% of the species of the extinct Plectofrondiculariidae had widespread, near cosmopolitan, distributions in all of the world's oceans, with the remainder being more restricted to one ocean or part thereof.

• At least 50% of species (102) in these five extinct families had a cosmopolitan distribution, occurring in at least three and mostly all four of the major world oceans (Pacific, Atlantic, Indian, Southern).

• 60% of the species in the Glandulonodosariidae and Pleurostomellidae are recorded from all four oceans.

• Approximately 50% of the species (mostly rare) in the Chrysalogoniidae, Ellipsoidinidae, and Stilostomellidae had more restricted recorded distributions, occurring in no more than two of the world's oceans.

• A few species that were common in some parts of the world's oceans appear to have been absent from other large parts (e.g., Anastomosa gomphiformis, Carchariostomoides dentaliniformis, Epelistoma crassitesta, Scallopostoma karreri, Siphonodosaria campana, S. lohmanni).

• The widespread, cosmopolitan distribution of many common deep-sea benthic foraminifera has been confirmed by limited molecular studies that show the same three molecular species occurring in both the Arctic and Antarctic realms.

• The paleogeography of the two groups of warm shallow-water foraminifera (Bolivinellidae, Tubulogenerina) changed through time. The Bolivinellidae originated in the Caribbean and Asian Tethys and dispersed to and diversified in Europe, the Pacific and Australasia in the late Eocene and Oligocene before becoming extinct in Europe and New Zealand in the early late Miocene. The family's modern distribution in the Indo-Pacific is completely different from its Eocene–Oligocene distribution primarily in the Caribbean and Europe regions.

• The extinct Tubulogenerina appears to have originated in Europe in mid Eocene and progressively dispersed to the Caribbean in the late Eocene, tropical Pacific in the Oligocene and to Australasia and Indian Ocean in the Miocene. The genus's time of regional extinction also occurred earliest (late Oligocene) in Europe and the Caribbean and lasted longest (till the end of the Miocene) in the tropical Pacific.

• The genus Notorotalia has maintained its Southern Hemisphere distribution around the coasts of Australasia and South America throughout its Eocene to Recent history with the centre of greatest diversity consistently around New Zealand.

• The genus Parrellina has been restricted also to the south and east coasts of Australia throughout its Cenozoic history from Eocene–Recent.

• The genus Porosorotalia had its greatest Eocene–Oligocene diversity in southern Australia and New Zealand but is now extinct there and only survives around southern South America and in the northwest Pacific.

• All but one of the species (Staffia tosta) of Plectofrondiculariidae originated in the Eocene or later.

• There are very few species in our dataset from the Indian Ocean and only one from the Southern Ocean, but this may reflect the fact that all but one of our deep-sea cores in these oceans were from abyssal depths (>2000 m) and not from the more favoured outer shelf–bathyal depths.

• The only subtle patterns in the paleobiogeographic history of the extinct Plectofrondiculariidae is that more species appear to have originated in the Pacific Ocean and dispersed subsequently to the Atlantic than the reverse and that there was an interval of enhanced origination in the late Eocene.

• The majority of species in these families originated in the Cretaceous–mid-Eocene.

• The available data set suggests that more species originated in the Atlantic Ocean, followed by the Southern Ocean and Pacific Oceans but many species rapidly dispersed to other oceans.

• All families and the majority of genera had a cosmopolitan distribution throughout the Paleocene to early Miocene, with progressive disappearance and withdrawal from higher latitudes and greater depths in the Pliocene–mid-Pleistocene, during which time they became extinct.

• The Ellipsodiinidae had an interval of enhanced origination and diversity in the Late Cretaceous (31 spp., 70% of total) and declined in diversity thereafter. They are not recorded from the Southern Ocean after the early Miocene.

• In two warm-water inner-shelf groups (Bolivinellidae, Tubulogenerina) the mean full species durations of all species are unusually short (4–5 myrs). Of these species, 50–70% have species durations of <3 myrs each and all are locally or regionally-restricted in their distributions. The longest species durations in these two groups are 26 and 23 myrs respectively.

• The mean full species duration for the 18 temperate, dominantly inner–mid-shelf Elphidiidae in New Zealand is 19 myrs with a mix of longer- and shorter-ranged species. The mean partial species duration for just the 10 extant species is 17 myrs, which we double to estimate full durations.

• The temperate shelf to upper bathyal notorotaliid genera Notorotalia and Porosorotalia in New Zealand have only endemic species where their mean species durations are 7 myrs (43 spp.) and 11 myrs (9 spp.), respectively. The longest species durations are 11 and 28 myrs, respectively. Near 50% of the Notorotalia species have species durations <3 myrs, whereas only 10% of Porosorotalia species.

• The deeper mid-shelf to bathyal, extinct Plectofrondiculariidae have a mean species duration of 20 myrs with the longest duration of 65 myrs.

• The five extinct, deep-water (mid-bathyal–abyssal) families (207 spp.) have the longest mean species durations of 41–50 myrs, with ∼50% of species with global durations of >50 myrs and almost none (1%) with durations of <3 myrs.

• Previous estimates of partial species durations based on the fossil records of extant species off the coasts of eastern North America and New Zealand are 16 and 19 myrs, respectively, for shelf-restricted species (<100–200 m) and 26 and 21 myrs for deeper-water (>100–200 m) species. If doubled to estimate full species durations the means are 35 myrs for shelf-restricted and 47 myrs for bathyal–abyssal species.

• Based on the above results, we conclude that deep-water benthics have the longest mean full species durations (40–50 myrs) and inner-shelf-restricted species have the shortest (4–11 myrs) with outer shelf–mid-bathyal species longevities in between (7–20 myrs).

• These differences in species duration may reflect more widespread and constant conditions in the deep sea compared with the rapidly fluctuating and stressed environmental conditions at shelf depths, particularly intertidally and shallow subtidally.

We wish to acknowledge the extensive studies of Marty Buzas, Steve Culver, Tom Gibson, Hugh Grenfell, Chris Hollis, Norcott Hornibrook, Shungo Kawagata, Jan Pawlowski, and Ashwaq Sabaa that underpin significant parts of this review article. The New Zealand-based parts of this review would not have been possible without the New Zealand-wide system, established by the late Norcott Hornibrook, that records the stratigraphic ranges of all foraminifera in New Zealand's extensive Cretaceous and Cenozoic marine sediments. We thank Stephen Culver, Martin Langer, and Mimi Katz for reviews and comments that have greatly improved the original manuscript.