The Phanerozoic diversity of the agglutinated foraminifera

The oldest known fossil representative of the agglutinated foraminifera is the genus Titanotheka, which was found in sedimentary rocks of the latest Pre-Cambrian age (the Vendian or Ediacarian stage) in Argentina (Gaucher and Sprechmann 1999). This supposed ancestral form was a singlechambered sac-shaped genus, with later two-chambered and multichambered morphologies derived from this primitive ancestor. Early Cambrian forms such as Platysolenites from the lower Cambrian of Estonia have been discovered to possess a primitive proloculus (McIlroy et al. 2001) and were transferred to the Hippocrepinacea. This adds to the evidence that the agglutinated foraminifera are derived from a simple sac-shaped ancestor.

The curves showing the number of genera in each stage and mean standing diversity are given in Figs. 1 and 2. These curves suggest the evolutionary history of the agglutinated foraminifera can be divided into five main stages:

Fig. 2.

Phanerozoic mean Standing diversity of agglutinated foraminifera and mean sea level. The mean sea level curve was generated using Time Scale Creator, version 4.0.2 [ http://www.tscreator.com]

The timing and magnitude of diversification and decline vary widely among the different suborders of agglutinated foraminifera (Fig. 3). All four suborders that have representatives in the Cambrian (Astrorhizina, Saccamminina, Hippocrepinina, and Ammodiscina) are still extant today. The Ammodiscina is the dominant group in the Cambrian, while the Saccamminina show a rapid radiation until the Late Silurian, and become the dominant group among the four suborders at that time. The Hormosinina first appear in the Ordovician and diversify during the Devonian. The rest of suborders appear by the Oxfordian. The Lituolina, Spiroplectamminina, Trochamminina, Verneuilinina, Loftusiina, and Ataxophragmiina increase in diversity very rapidly during the Mesozoic, and most of these groups show a decline at the end-Cretaceous. Another peculiarity in the diversity patterns of the suborders is that the Hippocrepinina, Verneuilinina, Biokovinina, and Ataxophragmiina appear to be Lazarus taxa. The Ataxophragmiina, for example, are absent from the Hettangian to Callovian, an interval of about 35 Myr. The Cainozoic is characterised by the rapid radiation of the Textulariina. The stratigraphic records of the Astrorhizina, Hormosinina, and Trochamminina need to be further resolved, as these groups possess many genera that are only recorded from the Holocene.

Fig. 3.

Stratigraphic ranges of agglutinated foraminiferal suborders (Kaminski 2004) over the Phanerozoic.

For the diversity falls identified in the estimated mean standing diversity at the Changhsingian/Induan, Norian/ Rhaetian, Cenomanian/Turonian, Campanian/Maastrichtian, Maastrichtian/Danian, Bartonian/Priabonian and Messinian/Zanclean boundaries, simultaneous decreases in the diversity of several suborders are recognised. The Spiroplectamminina, Ataxophragmiina, and the suborders of larger agglutinated foraminifera (Loftusiina and Orbitolinina) in particular show prominent evolutionary bottlenecks. Three groups that first appeared in the Mesozoic (Cyclolinina, Nezzazatina, and Orbitolinina) suffered extinction in the Campanian, Maastrichtian, and Chattian, respectively.

Extinction and origination rates

Taxonomic metrics performed on the curve include percent originations and extinctions in a given geological stage, and these rates were normalised to differing stage durations, providing origination and extinction rates per million years (Figs. 4, 5). The origination and extinction rates were then ranked to graphically illustrate the events of highest amplitude, and to determine whether their distribution is continuous (Fig. 6).

The extinction record shows that these marine protozoans were adversely affected by the five major Phanerozoic mass extinction events recognised by Sepkoski and co-workers. In Sepkoski's data (Sepkoski 2002), the highest extinction rate is observed at the end-Permian, with other major peaks in extinction observed at the end-Ordovician, end-Devonian, end-Triassic, and end-Cretaceous. In comparison to Sepkoski's (2002) record of marine macrofaunal biodiversity, the diversity decrease in agglutinated foraminifera that we observe associated with these mass extinction events is much lower, which highlights the resilience of these more primitive marine protozoans to environmental change. However, unlike Sepkoski's (2002) biodiversity curve, we also observe extinction rates that are significantly above background levels in the Gorstian, Callovian, end-Cenomanian and end-Miocene. In our record, the end-Cenomanian extinction event rivals the Cretaceous/Palaeogene event in terms of magnitude, with extinction of about 16% of genera. The end-Miocene extinction appears to be an especially significant one, when about 12% of the genera went extinct, while Sepkoski's (2002) data show an extinction rate of less than 10%. The time-corrected extinction rate (per-capita rate), however, suggests it is the most rapid extinction event in the Phanerozoic record for agglutinated foraminifera. The background extinction rate shows an increasing trend towards the Recent, which is in disagreement with the findings of Raup and Sepkoski (1982). This increasing trend is likely due in part to the two devastating extinction events observed mostly among larger agglutinated genera at the end-Cenomanian and end-Cretaceous, as well as the pull of the end-Miocene extinction event, which may be caused in part by a sample/observer bias (see discussion below).

Fig. 4.

Per-capita (A) and percent (B) extinction rates of agglutinated foraminiferal genera over the Phanerozoic, compared with the “Big-5” mass extinctions of Sepkoski (1984).

The origination rate appears to vary greatly through geological time (Fig. 5), and the ranked dataset of origination rates (Fig. 6) shows a discontinuous distribution. The per-capita origination rate (diversification) is particularly high in the Early Ordovician, early Silurian, late Carboniferous, Early Jurassic, mid-Cretaceous, early Eocene, and early Miocene. Curiously, the background extinction rates are higher for the Jurassic to Holocene part of the record than for the pre-Jurassic.

In the Palaeozoic record of per-capita origination rate the peaks are generally higher from the Early Cambrian to the late Silurian (Gorstian) than those in the remainder of the Palaeozoic, and the maximum is reached in the Dapingian. The “universal depression” of origination and extinction rates displayed by marine invertebrates during the Late Carboniferous to Early Permian ice ages (Stanley and Powell 2003) is not apparent in our record. If anything, origination rates during the period between 275–325 Ma are characterised by their high variablity. As is the case with the fusulinid taxa (Groves and Lee 2008), average per-capita origination rates within the Late Palaeozoic Ice Age (LPIA) period are higher than in the intervals immediately before or after the LPIA. Instead of displaying elevated rates in the late Permian as is the case with marine invertebrates, with the exception of the basal Permian Asselian stage, the origination rate of agglutinated foraminifera remains at zero for the rest of the Permian. After the Permian period of no origination, there is a massive radiation in the beginning of the Triassic, the Induan, followed by an interval of relatively low rates. The origination rate reaches its Phanerozoic maximum in the Hettangian and then drops rapidly in the next stage. During the Middle and Late Jurassic, the rate remains relatively high in comparison with rates in the rest of the Mesozoic and the Cenozoic. There is a fluctuation of the origination rate during the Cretaceous, and the rate reaches its highest value in the Santonian. The Cenozoic appears to be a period of low origination, except for the peaks in the Aquitanian (Early Miocene), and the Pleistocene. The rate remains at zero in the Late Miocene, and the Pliocene. The background origination rate indicated by a liner regression line decreases towards the Holocene (Fig. 5). This is in agreement with the findings of Raup and Sepkoski (1982) and Gilinski and Bambach (1987).

Fig. 5.

Per-capita (A) and percent (B) origination rates of agglutinated foraminiferal genera over the Phanerozoic.

Origination and extinction events

Global climate (and associated species-area effects) undoubtedly played a role in the diversification of agglutinated foraminifera. For example, nearly all the peaks in origination rate correspond to well-known marine transgressions or long-term sea level rises, which are linked to global climatic trends (Miller et al. 2005). Twenty years of subsequent work on the stratigraphic records of the agglutinated foraminifera after the publication of the Loeblich and Tappan (1987) volume (especially some recent key studies on the Palaeozoic and Mesozoic) now give us much improved resolution of the early diversification of the group, bringing the early diversification rates more in line with general trends in the Sepkoski (2002) curve. Several origination and extinction events stand out above the background rate. Extinction events, such as ones at the Cenomanian/Turonian and Cretaceous/Palaeogene boundaries, are relatively well studied, but the record of agglutinated foraminfera points to some lesser-known extinction and origination events:

The Late Palaeozoic Ice Age period.—The “universal depression” of origination and extinction rates displayed by marine invertebrates during the Late Carboniferous to Early Permian ice ages between 275–325 Ma (Stanley and Powell 2003) is not apparent in our record. Stanley and Powell (2003) claimed that non-fusulinid foraminifera conform to the general pattern of suppressed origination and extinction rates during the LPIA, but this is clearly not the case with agglutinated foraminifera. As is the case with the fusulinid taxa (Groves and Lee 2008), average per-capita origination rates within the LPIA are variable, but are distinctly higher than in the intervals immediately before or after the LPIA. The three peaks in origination in our record correspond to the Serpukhovian, Kasimovian, and Asselian stages, the former and latter of which represent the cooler intervals (glacials II and III) of the LPIA. In the case of the fusulinids, the invention of algal symbiosis and fragmentation of their geographical ranges are widely regarded as key factors that promoted their diversification, but among the agglutinated foraminifera no evolutionary innovations or geographical provincialism are observed. The agglutinated genera appearing during the LPIA are mainly simple smaller ammodiscids, hippocrepinids, and astrorhizids, as well as some multichambered genera that today inhabit both shallow and deep environments taxa such as Trochammina, Spiroplectammina, and Haplophragmoides. Expansion and contraction of epicontinental habitat size would have had the same effect on foraminifera as on other shallow-marine invertebrate taxa, but does not explain why marine invertebrates had subdued evolution while some groups of foraminifera display enhanced evolution during the LPIA. However, one feature relating to the habitat of agglutinated foraminifera is clearly apparent in our data set: eight out of the 15 agglutinated genera that suffered extinction during the LPIA or were confined to the interval were sessile, whereas most of the genera that either survived the LPIA or originated during the period (45 out of 49 genera) were motile. Perhaps sessile forms were at a disadvantage during a period of rapidly fluctuating sea levels? In any case, the classic studies of late Palaeozoic agglutinated foraminifera were largely based on acid residues of shallow-water carbonates, mostly in the North American and European coal measures. Therefore a sampling bias is likely present in our data set. We need more studies on agglutinated foraminifera from a range of environments across the LPIA to resolve the question of elevated origination rates.

Fig. 6.

Ranked per-capita extinction (A) and origination (B) rates, showing events of highest amplitude.

The Hettangian origination.—This event represents the largest peak in the per-capita origination rate, and is characterised by the first appearance of some “larger” agglutinated foraminiferal genera, derived from Lituola, such as Cym- briaella, Haurania, Orbitopsella, Everticyclammina, Lituosepta, Biokovina, etc. This origination event actually covers the interval from the Hettangian to Pliensbachian, and the background origination rate remains above the pre-Jurassic rate well into the Cretaceous. The appearance of larger foraminifera with complex inner structures is associated with the Hettangian transgression and the development of widespread carbonate platform environments in the Peri-Tethyan region, with the establishment of a stable, oligotrophic, shallow-water environment. These conditions led to the development of larger foraminifera with chamberlets that may have housed algal symbionts, and such forms become well established by the late Sinemurian (Septfontaine 1988; Fugagnoli 2004). At first glance, from our diversity record may appear unusual for a major origination event to appear without a precursor extinction event, as the two phenomena in the marine biodiversity record are causally linked (Stanley 2007). However, the larger agglutinated foraminifera of the early Jurassic have taken over the ecological niche vacated by the calcareous microperforate fusulinids. This foraminiferal group suffered a double extinction event in the late Permian (Stanley and Yang 1994), with the last morphologically primitive survivors becoming extinct in the late Triassic. By the Pliensbachian, the Tethyan carbonate platforms had been colonised by new lineages of large, morphologically complex loftusiids and textulariiids, including the pfenderinids, hauraniids, mesoendothyrids, biokovinids, and chrysalidinids (see review by BouDagher-Fadel 2008).

The Cenomanian turnover.—The Cenomanian stage represents the Phanerozoic maximum in the diversity of agglutinated foraminifera, with a total of 204 genera present (Fig. 1). It also contains the largest number of genera that possess stratigraphic ranges restricted to a single stage (16 genera). We note the first occurrences of 35 genera within the Cenomanian stage, most of which belonging to the loftusiids, including Orbitolinella, Dohaia, Cyclolina, Lituonella, Quataria, Spirocyclina, Coxites, and others. In addition, some notable deep-water forms such as Praecystammina and Bolivinopsis also make their appearance in the stratigraphic record. The Cenomanian witnessed the Phanerozoic peak in global sea levels (Miller et al. 2005), and most of Cenomanian larger agglutinated foraminifera had their origin in the subtropical lagoons and shallow carbonate shelf seas of the Tethyan region, with only one genus restricted in the Caribbean (BouDagher-Fadel 2008). The Cenomanian extinction rate (ca. 16%) ranks just below the Cretaceous/Paleogene boundary extinction event (Fig. 6). The extinction of many of the Cretaceous larger agglutinated foraminifera, as well as Jurassic survivors such as Everticyclammina and Siphovalvulina is causally linked to the drowning and collapse of the tropical reef ecosystems in the late Cenomanian (see review by Skelton 2002). A number of smaller agglutinated genera did not survive Oceanic Anoxic Event 2 at the Cenomanian/Turonian boundary (Schlanger and Jenkins 1976), including Spiroplectinata, Quasispiroplectammina, Kaminskia, Aaptotoichus, Buccicrenata, and Pfenderina. Carbon-rich sediments and associated taxonomic turnover in agglutinated foraminifera have been documented from numerous DSDP and ODP sites and onshore sections worldwide (e.g., Kuhnt 1992; Kaiho and Hasegawa 1994; Coccioni et al. 1995), and reflect a major perturbation of the global carbon cycle that lead to widespread oceanic anoxia.

The Aquitanian origination.—In the latest Oligocene to earliest Miocene a number of trochospiral agglutinated genera with alveolar or canaliculate walls first appeared in the fossil record. Genera, such as Alveovalvulina, Guppyella, Goesella, and Alveovalvulinella, are typical of assemblages found in subtropical oxygen minimum zones, especially in West Africa and the Caribbean. It is likely that the global warming of the latest Oligocene to Early Miocene contributed to intensification of dysoxic conditions in upwelling regions, creating an expanded niche for these organisms that flourished in low-oxygen conditions. In a recent study of the lower bathyal Miocene foraminiferal assemblages of the Congo fan, offshore Angola, Kender et al. (2008) noted steadily increasing diversity and proportions of infaunal morphotypes over the lower Miocene interval. The proportion of infaunal types increased dramatically in the lower mid-Miocene, indicating expansion of the Oxygen Minimum Zone into deeper waters. In addition to the trochospiral alveolar genera, species of Reticulophragmium and Cyclammina display rapid diversification into numerous separate lineages that are at present not reflected in our generic diversity record owing to their poorly established taxonomy. It is likely that a more detailed phylogenetic approach to these genera would result in the description of new genera for individual lineages.

The Messinian extinction.—This event appears to be the largest extinction event in the diversity record of agglutinated foraminifera (Fig. 6), and it is arguably the most poorly understood event. The extinction is partly due to the disappearance of some of the alveolar or canaliculate genera that evolved in the Early Miocene. These include Matanzia, Jarvisella, Pseudotriplasia, Alveovalvulinella, and Colominella. The fact that so many of these genera belong to the infaunal morphogroup suggests that the end-Miocene extinction may be a precursor of the “Stilostomella Extinction Event” noted among calcareous deep sea benthic foraminifera (Schönfeld 1996) However, the actual “Messinian extinction event” among agglutinated genera may be influenced by an observer bias (see below). The lack of published early Pliocene records means that the stratigraphic ranges of some genera are simply reported as “Miocene”. This interval in particular requires further study to improve stratigraphic resolution, in particular in offshore upwelling regions where the alveolar genera are found.

Possible bias in the diversity curve

We cannot exclude the possibility that at least some features of the diversity curve may be due to observer or stratigraphie bias or uneven sampling intensity. The Phanerozoic history of the group has probably been influenced by various biases, such as the rock-volume, sampling intensity, and taxonomical effort biases. The effects of possible biases on diversity records have been reviewed by Smith (2003) with reference to marine invertebrates. It is perhaps no coincidence that the large rise in our mean standing diversity observed in the Jurassic also corresponds to the largest increase in the curve of marine rock outcrop area in Western Europe compiled by Smith and McGowan (2007). The influence of these possible biases on the diversity record of agglutinated foraminifera needs to be evaluated.

In the case of agglutinated foraminifera we see additional biases. For example our oldest record of the deep-sea environment (the largest biotope on our planet) sampled by the Ocean Drilling Program hails from the Late Jurassic. Whole environments where agglutinated foraminifera are common (such as salt marshes, mangrove swamps, the Central Arctic, and the Antarctic seas) have very patchy stratigraphie records. Single studies of agglutinated assemblages from deep-sea cores can still produce numerous changes to the known ranges of agglutinated genera. The end-Miocene diversity fall in particular may be a precursor of the “Stilostomella Extinction Event” or may be partly due to the fact that we have so few good published records of agglutinated foraminifera from Pliocene and Pleistocene strata. We list a total of 293 living genera of agglutinated foraminifera, whereas our Pliocene record shows only 132 fossil genera. Part of this apparent discrepancy may be due to the fact that not all modern forms have shells that are readily preserved, as already noted by Tappan and Loeblich (1988), but part of the discrepancy is no doubt due to the “incompleteness of the sampled fossil record”. If we accept that all the genera listed as existing only in the modern ocean might some day be found preserved as fossils in Mesozoic to Cenozoic sediments, then the diversification rate of the group would appear to continue undiminished since the Early Jurassic. It is possible that in the future, a more phylogenetic approach (rather than a simple taxic approach) to compiling the diversity record may improve the robustness of our curve and give us higher diversification rates over the post-Triassic interval. Nevertheless, the current shape of our diversity curve does indeed mimic the records of calcareous nannoplankton (Bown et al. 2004) and planktonic foraminifera (Tappan and Loeblich 1988)—groups of marine protists that have much more complete and well-established fossil records. The effect of global climate on the diversity record of these primitive single-celled organisms is undeniable, but to exactly what extent remains to be quantified.