Morphometric changes in two Late Cretaceous calcareous nannofossil lineages support diversiﬁcation fueled by long-term climate change

Morphometric changes have been investigated in the two groups of calcareous nanno-fossils, Cribrosphaerella ehrenbergii and Microrhabdulus undosus across the Campanian to Maastrichtian of the Zagros Basin of Iran. Results reveal a common episode of size increase at c. 76 Ma, with a sudden increase in the size of C. ehrenbergii and with the emergence of a newly deﬁned, larger species Microrhabdulus zagrosensis n.sp. An even larger species emerges at c. 69 Ma within the Microrhabdulus lineage, Microrhabdulus sinuosus n.sp. The timing of these size changes and origination events matches global changes in nannoplankton diversity and/or in diversity of other planktonic organisms and marine invertebrates. Comparison with long-term


Introduction
Extant and fossil calcareous microplankton commonly comprise numerous cryptic and/or pseudocryptic species (Sáez et al., 2003;de Vargas et al., 2004). Morphometric measurements such as size variation are regularly conducted on calcareous nannofossils in order to assess changes in morphology that often lead to the erection of new species and lineages (Bollmann, 1997;Geisen et al., 2004;Shamrock & Watkins, 2009). Among planktonic microfossils, the fossil record of calcareous nannoplankton is particularly important as it is intimately linked to the carbon cycle, global changes in climate, and oceanography. Morphological changes in calcareous nannofossils have been the subject of numerous studies across their evolutionary history in the Meso-Cenozoic. Perhaps due to better preserved records, Cenozoic biometric studies are abundant, while Mesozoic studies have focused mostly on specific intervals such as the Early to Middle Jurassic Suan et al., 2010;Peti & Thibault, 2017;Peti et al., 2021;Menini et al., 2021;Faucher et al., 2022;López-Otálvaro et al., 2012;Ferreira et al., 2017) or Early to mid-Cretaceous (Bornemann & Mutterlose, 2006;Barbarin et al., 2012;Lübke & Mutterlose, 2016;Bottini & Faucher, 2020;Wulff et al., 2020). In comparison to the latter periods, and considering that the Santonian to Maastrichtian interval bears the highest diversity within the overall evolutionary history of calcareous nannoplankton (Bown et al. 2004), this significant part of the Late Cretaceous has been the subject of relatively few biometric studies mostly focused on Arkhangelskiellaceae and the Eiffelithus lineages (Girgis, 1987;Faris 1995;Thibault et al., 2004;Linnert & Mutterlose, 2009;Shamrock & Watkins, 2009;Thibault, 2010;Linnert et al., 2014). Such studies are of fundamental importance in taxonomy, often revealing the presence of pseudo-cryptic taxa (Shamrock & Watkins, 2009;Suchéras-Marx et al., 2010;Thibault, 2010;Peti & Thibault, 2017;Peti et al., 2021), but also revealing trends that coincide with paleoenvironmental change Ferreira et al., 2017;Lübke & Mutterlose, 2016;Bottini & Faucher, 2020;Wulff et al., 2020). During the investigation of samples from the Shahneshin section (Zagros basin, Iran, Razmjooei et al., 2018; Fig. 1), we noticed interesting patterns in size and morphology of Cribrosphaerella ehrenbergii and Microrhabdulus undosus, two species that are frequent to common in Santonian to Maastrichtian low-latitude assemblages. The paleoecology of M. undosus is unclear, and it is essentially a cosmopolitan taxon (Lees, 2002). Although C. ehrenbergii is also a cosmopolitan taxon (Thierstein, 1981;Henriksson & Malmgren, 1997;Lees, 2002), several authors have considered this species as having a greater affinity toward cool sea-surface temperatures (Wise, 1983;Pospichal & Wise, 1990;Watkins, 1992;Ovechkina & Alekseev, 2002, 2005Razmjooei et al., 2020b), possibly enhanced in more proximal conditions (Razmjooei et al., 2020b), while others postulated a controversial affinity to nutrient availability. For instance, Erba et al. (1995) suggested that blooms of C. ehrenbergii could indicate increased surface water productivity, while Linnert et al. (2011) inferred a lower nutrient affinity for this species. The purpose of this study is to better constrain Cribrosphaerella and Microrhabdulus taxonomies, investigate biometric changes in these two lineages, and evaluate the timing of these changes relative to global changes in temperature and the carbon cycle.

Geological setting and a summary of earlier works
The Upper Cretaceous pelagic deposits of the Gurpi Fm. in the Shahneshin section consist of regular carbonate cycles of dark-to pale-gray marl and dark-gray to light-yellow argillaceous limestones with a number of resistant carbonate beds forming ridges in the landscape that constitute excellent markers across the field area. The section possesses the advantage of bearing a detailed and precise stratigraphic framework of calcareous nannofossils, planktonic foraminifera, marine palynomorphs, and bulk carbonate carbon isotopes Fig. 2). The Shahneshin section is situated in the central part of the Zagros folded zone, west of Fars province, on the northeast of Kazerun city and the northwest of the Shahneshin anticline, with geological coordinates of N29° 44´ 47´´; E51° 46´ 31´´ for the base of the section, and N29° 44´ 40.69´´; E51° 46´ 26.87´´, for the top of the Cretaceous sequence (Fig. 1).  (Barrier et al., 2018). B) Recent geological map of the Zagros Basin and the location of the studied area.
The first stratigraphic framework on Shahneshin was carried out in 2014 based on integrated calcareous nannofossil biostratigraphy and carbon isotope stratigraphy (Razmjooei et al., 2014). In 2018, the detailed integrated bio-(calcareous nannofossils, planktonic foraminifera and marine dinoflagellate cysts) and stable carbon isotope stratigraphy, with a subsequent graphic correlation to Tethyan reference sections (Gubbio, Italy) led to proposing an age model for the Gurpi succession and for numerous Late Cretaceous planktonic foraminifera and calcareous nannofossil bioevents   (Fig. 2). According to these two earlier works, two significant hiatuses were recorded in the latest Campanian and/or across the Campanian-Maastrichtian boundary and across the topmost Maastrichtian to early Danian. Later on, combined with facies change, carbonate content and changes in absolute abundance of calcareous nannofossils, a sequence stratigraphic interpretation was proposed for the Gurpi Fm. at Shahneshin (Razmjooei et al., 2020a). Moreover, detailed abundance counts and statistical analysis were investigated on the calcareous nannofossil assemblage to explore its response to Late Cretaceous climatic changes. According to this study, a Late Cretaceous intensification of cooling across the late Campanianearly Maastrichtian, a mid-Maastrichtian warming episode, and a late Maastrichtian cooling episode could be delineated in Zagros along with accompanying proximal/distal trends possibly influencing the nannofossil assemblage (Razmjooei et al., 2020b).

Material and methods
The calcareous nannofossil biometric analysis has been carried out on smear slides prepared by Razmjooei et al. (2018) based on the standard preparation method of Bown & Young (1998). The calcareous nannofossil zonations are based on the CC biozonation of Sissingh (1977) modified by Perch-Nielsen (1985) and the UCTP (Tethyan Province) of Burnett (1998), using the taxonomic concepts of Perch-Nielsen (1985), Young & Bown (1997) as well as Shamrock & Watkins (2009).
Due to the moderate to poor preservation of calcareous nannofossils (Razmjooei et al., 2020b), only the best-preserved samples with frequent to common abundance of the two lineages C. ehrenbergii and M undosus in the upper Campanian-Maastrichtian were chosen for biometric measurements (Fig. 2). The length of C. ehrenbergii has been investigated in standard smear-slides for a total of 29 samples spanning the entire Campanian and Maastrichtian intervals of the Shahneshin section (Table 1; Supplementary Appendix 1). Because of the lower abundance, changes in the size and morphology of M. undosus group have been investigated in 17 samples from 170 to 342 m (upper lower Campanian to Maastrichtian), where the abundance of the species was high enough to allow for the measurement of several specimens (Supplementary Appendix 1). The sample spacing varies mostly between 7 to 9 m with time intervals varying between 250 and 900 kyr according to the age-model presented in Razmjooei et al. (2018) (Supplementary Appendices 3 and 4). When possible, we measured the length and width of up to 50 specimens of C. ehrenbergii. Three samples at 106.2, 123.1, and 186 m had a low abundance of this species so only 19, 25 and 40 specimens were measured in them, respectively (Table 1).

Figure 2.
Integrated biostratigraphy and bulk carbonate carbon isotope records of the Shahneshin section based on Razmjooei et al. (2018), along with the relative abundances of Microrhabdulus spp. and C. ehrenbergii according to Razmjooei et al. (2020b). The two green areas with dashed lines indicate the two stratigraphic levels with major shifts observed in the morphology of the two taxa. Stratigraphic and isotopic data from Razmjooei et al. (2018). The given absolute ages are based on the age-model presented in Razmjooei et al. (2018).
For M. undosus group, we could measure the length and width of up to 30 specimens in only 8 samples. The total number of specimens measured in other samples is provided in Table 2. The biometric measurements have been performed using a light microscope (Leica DM750P) at a magnification of 1250× with a Leica DFC320 digital camera and the ImageJ 1.50i software. In total, 1385 pictures of C. ehrenbergii and 336 pictures of M. undosus were captured throughout the Campanian-Maastrichtian interval. All the images were taken under the cross-polarised light (XPL). Despite the heterogeneity of the M. undosus group dataset, our investigations of size changes in this group were paralleled with systematic observations on the patterns of individual laths that characterize this taxon (Supplementary Appendix 2). As described further, significant changes in the size of the M. undosus group occur at precise stratigraphic intervals and coincide with significant changes in lath patterns (Figs. 3 and 4). Accordingly, biometric measurements on individual samples are used to follow the progression of these patterns across the stratigraphy of the section and the statistical significance in the difference between potential morphotypes is tested via density plots using the Matlab® script of Thibault et al. (2018) and via the comparison of distribution histograms produced in PAST® (Hammer et al., 2001) for three distinct stratigraphic intervals that bear enough specimens for reliable statistics (Fig. 5).
A settling technique was adapted from Bown & Young (1998) to obtain suitable slides for Scanning Electron Microscope (SEM) analyses and eliminate particles bigger than 30 µm and smaller than 1 µm. Moreover, to have a cleaner surface for observing coccoliths and taking pictures of coccolith individuals under SEM, the suspensions from gravity settling were subjected to the filtration technique using membrane filters with 0.8 µm pore size. Finally, to observe and capture images of the same nannofossil individuals under the light and SEM microscopes, we applied the slide preparation technique proposed by Gallagher (1988) andPirini Radrizzani et al. (1990) using a copper grid with the coordinate system. Table 1. Number of measured specimens, mean length (μm), mean width (μm), standard deviation and 95% confidence intervals for each sample investigated for the size of Cribrosphaerella ehrenbergii.
latest Campanian Zone CC22, a significant increase is observed with values reaching up to 15%. An acme of the species with relative abundances higher than 10% and peaking up to maximum values of 20% is observed throughout the uppermost Campanian CC22 Zone to the lower Maastrichtian CC24 Zone. The relative abundance of C. ehrenbergii decreases again down to 5% in the mid-Maastrichtian and remains below 10% within CC25. The upper Maastrichtian subzones CC26a and UC20b-c TP are characterized by a second acme of the species reaching values as high as 17%. A slight decrease of the species is observed in the uppermost Maastrichtian samples from CC26a and UC20c TP to the Cretaceous-Paleogene (K-Pg) boundary (Fig. 2). From our biometric measurements, it appears that the size of C. ehrenbergii remains essentially stable around a mean length of 5.5 µm throughout the lower Campanian to upper Campanian Zone CC21. A significant increase in the average size of C. ehrenbergii occurs in coincidence with the onset of the first acme of the species within upper Campanian Zone CC22. Thereafter, the size of C. ehrenbergii remains stable around a mean of 6.5 µm in the remaining upper Campanian to Maastrichtian (Fig. 3). Comparison of two distribution histograms for stratigraphic intervals below and above 240 m highlights the near complete disappearance of C. ehrenbergii representatives smaller than 5 µm and appearance of giant specimens > 8 µm in upper Campanian to Maastrichtian zones CC22 to CC26 (Fig. 5, distribution histograms A and B).

Figure 3.
Mean length (μm) of C. ehrenbergii and mean width (μm) of the Microrhabdulus undosus group along with their standard deviation and 95% confidence interval envelopes. Density plots for both groups provide the percentage of data for each size bin. The two green areas indicate the stratigraphic levels where major shifts in morphology are observed. Relevant nannofossil and foraminifera biohorizons provide an age estimate for these two stratigraphic levels. The white dashed line on the density plot of C. ehrenbergii is a simplified interpreted trajectory of the highest population density intended to show the significant difference before and after 76 Ma.

M. undosus relative abundance and size variations
Microrhabdulus spp. represent only a minor component of the calcareous nannofossil assemblage for much of the Coniacian to lower Campanian, where they are generally below 2.5% (Fig. 2). The relative abundance of the genus starts to increase up to 5% within the upper Campanian Zone CC21 and shows a first significant peak abundance in the top of CC21 and within Zone CC22 with values up to 9%. This significant increase in the relative abundance of the genus is actually related to the first common occurrence of M. undosus which is the dominant species within this lineage. Microrhabdulus spp. relative abundance fluctuates gently around 4% in the remaining of the upper Campanian. Above the Campanian/Maastrichtian boundary, Microrhabdulus spp. increase and show a lower Maastrichtian acme with two double peaks of abundance reaching values as high as 12%. The top of this acme is in the upper part of zones CC25/UC19 and coincides with minimal values in C. ehrenbergii. Values then fluctuate between 1 and 7% in the remaining Maastrichtian (Fig. 2). The genus is essentially represented by two groups, M. decoratus and the M. undosus group. M. decoratus is responsible for a peak abundance of the genus up to 6% across the Coniacian/Santonian boundary. Otherwise, M. decoratus is generally below 1.5%, so the abundance fluctuations described above essentially correspond to that of the M. undosus group (Fig. 2). Variations in the length of the M. undosus group are generally biased by the frequent fragmentation observed for this species. Microrhabdulus nannoliths are elongated rods generally >10 µm in length for 1 to 3 µm in width, hence the rods are commonly fragmented and edges of the rod often lack the typical tapering observed in complete specimens. Despite this bias, the maximum length observed in each sample can still represent a valuable index as this parameter is more likely to represent the length of complete, non-fragmented specimens, but evidently, the average length is meaningless and hence, not considered further in this study. A major shift in the morphology of the M. undosus group occurs in the upper Campanian Zone CC22 in coincidence with the shift in the size of C. ehrenbergii and with a peak abundance of Microrhabdulus spp. (Fig. 3). This change is expressed by a significant increase in the width of the rod (Figs 3 and 4), together with the occurrence of specimens with wavy patterns in the extinction line of the central axis of the rod and a slight change in morphology of the laths from triangular to rosethorn shaped (Fig. 4). Moreover, we also observe an increase in maximum length of our specimens that parallels the increase in width of the rod (Fig. 6). It must be noted that the observed range of maximum length observed here varies between 14 and 27 µm which matches the given rod size range of 15 to 30 µm in illustrated specimens of the Nannotax3 database: (https://www.mikrotax.org/Nannotax3/index.php?taxon=Microrhabdulus%20undosus&module=ntax _mesozoic).
A second episode with a significant change in morphology is observed at around 315 m, at the top of subzone CC25a/Zone UC19. The maximum length of the M. undosus group increases significantly to reach up to 27 µm towards the top Maastrichtian. In parallel, a progressive, significant increase in the width of the rod is also observed in the same interval (Figs 3 and 4), together with the occurrence of numerous forms that bear a sinuous central axis and additional, overlapping rows of rose-thorn shape laths visible within the central axis (Fig. 4). These observations lead us to define two new species Microrhabdulus zagrosensis and Microrhabdulus sinuosus and append the definition of M. undosus sensu stricto (see section 4.1., taxonomy). The significant difference in distribution histograms of the M. undosus group below and above 240 m strongly supports the mixing of two or three distinct morphospecies in samples above 240 m (Fig. 5, distribution histograms C and D). We also produced distribution histograms for the width of the three distinguished morphospecies following the taxonomy given in the next chapter (Fig. 6, distribution histograms A, B, and C).

Taxonomy
Family AXOPODORHABDACEAE Wind and Wise in Wise & Wind (1977) Genus Cribrosphaerella Deflandre in Piveteau (1952) Cribrosphaerella ehrenbergii, (Arkhangelsky, 1912) Deflandre in Piveteau (1952) Figure 7 Remarks. Reinhardt (1964) first made a distinction between distinct forms of Maastrichtian Cribrosphaerella species and this distinction was confirmed later on by Perch-Nielsen (1968) in the Maastrichtian of Denmark. In particular, they noticed a distinction between C. ehrenbergii and C. hilli based on the number of elements composing the shield, C. hilli having a clear tendency towards bearing more elements on the shield than C. ehrenbergii. In her SEM illustrations, C. hilli appeared more elliptical than subrectangular, contrary to C. ehrenbergii, which was clearly subrectangular. Reinhardt (1964) and Perch-Nielsen (1968) did not mention any significant difference in size between C. ehrenbergii and C. hilli, but Perch-Nielsen (1968) showed that the minimum size of her measured specimens of C. hilli was 7 µm long whereas her smaller specimen of C. ehrenbergii is 5 µm long. The definition of C. hilli by Reinhardt (1964) provides a range of large sizes between 8 and 10 µm, thus pointing towards systematically large specimens of Cribrosphaerella as compared to a total size range of 5-10 µm for C. ehrenbergii. Therefore, we cannot completely exclude that our observation of a significant rise in the mean length of C. ehrenbergii recorded in the late Campanian of Shahneshin is possibly due to the first occurrence of C. hilli, this species is probably slightly bigger than C. ehrenbergii. We were not able to make any clear consistent distinction here in the outline (subrectangular versus elliptical) between our measured specimens of C. ehrenbergii in our investigations under XPL (Fig.  7).
Remarks. All specimens measured in the present study were previously assigned to the species M. undosus. Morphological observations and measurements led us to redefine this group as a lineage that comprises three distinct species. From the detailed observations and biometric results obtained here, we redefined M. undosus sensu stricto (s.s.) as comprising narrow and large specimens of Microrhabdulus with undulating, slightly triangular laths and a neat, straight extinction line in the center of the rod when observed under XPL (Fig. 4). The sister species M. decoratus exhibits rows of strictly parallel-sided and butting, well-aligned laths of the same dimension under the SEM with no empty space in between distinct rows, and laths appearing as strictly rectangular under XPL with a straight extinction line in the center (Fig.  10A-B). Under the SEM, M. undosus s.s. exhibits less regular, non-strictly butting laths and visible open space in between the rows that allow observation of a preceding inner layer of laths beneath the outer layer ( Fig. 10C-D). The resulting appearance of this morphology under XPL appears to be a characteristic slightly triangular aspect of the individual laths but we note that, in contrast to the two new species defined below, the central extinction line remains straight in M. undosus s.s. (Fig. 10C-D). In the Late Cretaceous of Shahneshin (Zagros, Iran), narrow and smaller specimens with a width of the rod measuring between 1.2 and 1.9 µm (mean: 1.4 µm) and a maximum length of 20 µm dominate in the lower Campanian and to the top of upper Campanian Zone CC21 (UC15c TP ) (Figs. 3, 5 and 6). The uppermost Campanian specimens from Zone CC22 and Maastrichtian specimens are a bit wider, and vary in width between 1.4 and 2.1 µm (mean 1.8 µm) but have a maximum length of 21 µm, which is almost similar to narrower specimens (Figs 5, 6 and 8).

Derivation of name.
Referring to the Zagros basin from which the species is hereby described. Diagnosis. A long, wide species of Microrhabdulus with rose-thorn-shaped laths with pronounced wavy patterns of the laths and a central extinction line that starts to deviate from a straight line, when observed in XPL (Fig. 4 and Fig. 9A-E). Under SEM, we noticed open spaces in-between rows of laths as in M. undosus s.s., but more variations in the thickness of the individual laths with rows composed of thicker laths surrounding rows composed of thinner laths (Fig. 10F). Also, the width of the space between distinct rows can vary (Fig. 10E-F).
Differentiation. This species greatly resembles M. undosus s.s., from which it originates but can be distinguished from the latter by the rose-thorn shape of the laths under XPL, M. undosus sensu stricto bearing triangular laths; and by the wavy pattern of the extinction line of the central axis ( Fig. 9A-E), M. undosus sensu stricto bearing essentially straight central extinction lines. Microrhabdulus zagrosensis differs from the newly defined Microrhabdulus sinuosus by the lack of any additional rows of laths in the central axis and by a narrower diameter of the rod. Under the SEM, the space left between two distinct upper/outer rows of laths is wider in Microrhabdulus zagrosensis compared to M. undosus s.s, but is narrower compared to Microrhabdulus sinuosus (Fig. 10).
Paratypes. Figure 9B,   Derivation of name. Referring to the pronounced sinuosity of the central axis. Diagnosis. A very long, thick species of Microrhabdulus with rose-thorn-shaped laths on the sides, much pronounced wavy patterns of the laths and bearing additional rows of laths that are visible in the central axis in XPL ( Fig. 4 and Fig. 9G-L). As interpreted from our SEM observations, the strong wavy pattern of the laths under XPL is likely caused by the more complex undulating arrangement of the laths, with individual laths that vary slightly in width, hence they are less strictly rectangular in shape. There are also large open spaces between distinct rows of laths that allow observation of an inner layer and we notice some laths showing distinct kinks and/or trapezoidal shapes (Fig. 10H).
Differentiation. This species greatly resembles M. undosus s.s., but can be distinguished from the latter by the rose-thorn shape of the laths under XPL (M. undosus s.s. bearing triangular laths), by the much pronounced wavy pattern of the central axis (M. undosus s.s. bearing essentially straight extinction lines), and by the presence of additional rows of laths visible in the central axis (M. undosus s.s. showing a simple extinction line in this axis). M. sinuosus differs from the newly defined M. zagrosensis by a slightly wider diameter of the rod and the systematic presence of an additional row of laths in the central axis ( Fig. 9G-L), feature that is absent in M. zagrosensis (Fig. 9A-E). Under the SEM, the spaces between two distinct rows of laths in M. sinuosus appear wider than those in M. undosus and M. zagrosensis, and the laths composing the outer rows vary in width along the row, creating strong undulations that are likely responsible for the pronounced wavy patterns observed in XPL. The maximum length of M. sinuosus is greater than that of M. zagrosensis and M. undosus s.s., suggesting that the complete, non-fragmented specimens of the species essentially represent a giant species of Microrhabdulus. In the Late Cretaceous of Shahneshin (Zagros, Iran), Microrhabdulus sinuosus has a maximum length of 27 µm and a width that varies between 2 and 3.2 µm (mean of 2.5 µm) (Fig. 6).
Remarks. Note that there is likely a continuum in the evolution from M. zagrosensis to M. sinuosus as expressed by the progressive Maastrichtian increase in the diameter of the rod delineated in our data as well as by the presence of rare transitional forms that show very slightly undulating median extinction lines as in M. zagrosensis together with a faint additional row of laths in the center visible only in the upper third of the specimen (Fig. 9F).
Paratypes. Figure 9H Occurrence. Shahneshin; First occurrence is recorded in the upper part of early Maastrichtian CC25a subzone (top of UC19) and last occurrence at the Cretaceous-Paleogene boundary.

Lessons from the Meso-Cenozoic
Important lessons for understanding coccolith size changes and their potential causes have been brought by the study of the three main Cenozoic genera of coccolithophores (Reticulofenestra, Cyclicargolithus, and Coccolithus), namely that coccolith size is strongly linearly correlated to coccosphere and cell-diameter, and hence reflect changes in calcareous phytoplankton cell size (Henderiks, 2008). Out of these three genera, the Paleogene trend towards smaller reticulofenestrid cells across the Eocene to Miocene has been primarily interpreted as reflecting an adaptive response to increase aqueous CO2 limitation caused by a decrease in atmospheric pCO2 (Henderiks & Pagani, 2008;Hannisdal et al., 2012). As such, these cell size changes appear to be linked to coccolithophore calcification and cell growth. However, other factors have been postulated as potential controls on coccolith sizes and varying responses have been observed in distinct taxa. For instance, while the genus Toweius sees a decrease in coccolith length interpreted as a response to a rise in temperature and pCO2 across the Paleocene/Eocene Thermal Maximum (PETM), Coccolithus pelagicus increases in size across that interval, which has been interpreted as a response to slowed cell division (O'Dea et al., 2014). Cyclic, rapid Pleistocene changes in the size of Noelaerhabdaceae have been related to species radiation and pulses of extinction (Bendif et al., 2019), but also to orbital changes in insolation for the past 400 kyr, with enhanced seasonality at insolation highs favoring speciation, expressed in biometric data by a larger range of coccolith sizes with distinct small and large populations (Beaufort et al., 2021). Lessons from the Cenozoic hence tend to show that prominent, rapid coccolithophore size changes primarily reflect episodes of speciation favored by climate change, but also possible long-term responses to changes in pCO2 via calcification or cell growth rate.
In contrast, biometric studies in the Mesozoic tend to show a wider variety of possible controls on calcareous phytoplankton size variations. Cyclic changes in the size of Pliensbachian Crepidolithus crassus have been related to an orbital control, but the primary causes evoked to explain these features are changes in water turbidity and nutrient recycling in the photic zone via orbitally-controlled storm intensity . Large Pliensbachian nannolith Schizosphaerella size changes have been interpreted as variations in abundance of three distinct morphotypes as a response to variations in temperature and proximity to the coastline (Peti & Thibault, 2017;Peti et al., 2021), whereas the drop in size of this same taxon across the Toarcian Oceanic Anoxic Event has been interpreted as a response to a calcification crisis Clémence et al., 2015;Faucher et al., 2022). Changes in the size of Early to Middle Jurassic coccolith Lotharingius species might reflect the opposition between stable environmental conditions favoring large specimens versus periods dominated by ecological stress favoring small specimens (Ferreira et al., 2017). Mid-Cretaceous size changes in Biscutum constans have been interpreted as a response to both temperature variations and environmental stability versus altered ocean chemistry from widespread volcanism (Bottini & Faucher, 2020). In contrast, size changes in B. constans, Rhagodiscus asper and W. barnesiae across a mid-Barremian episode of black shale deposition in the Boreal Realm have been primarily interpreted as a response to changes in nutrients (Wulff et al., 2020). Despite the large variety of causes invoked in Mesozoic studies to explain the latter morphological changes, putative distinct morphotypes and hence pseudo-cryptic paleontological species have often been inferred and such variations in size appear correlated to pronounced climatic and environmental changes.    Aubry et al. (2005) demonstrated that the average size of coccoliths has increased from the Early Jurassic through the Santonian, stabilized until the Campanian and then decreased during the Maastrichtian. Remarkably, they noticed that this size history parallels the diversity (species richness) history of the Mesozoic Coccolithophorids. Consequently, these authors claimed that these observations constitute an illustration of Cope's rule. The same idea was later reiterated for a number of Jurassic-Cretaceous coccolith taxa (López-Otálvaro et al., 2012;Ferreira et al., 2017;Gollain et al., 2019).

Are Cope-Bergmann's rules applicable to phytoplankton?
Cope's rule postulates that "evolution proceeds in the direction of increasing body size" (Ghiselin, 1972, p. 141). In parallel, another rule of evolution in organisms named Bergmann's rule stipulates that organisms evolve larger sizes under cold temperatures (Timofeev, 2001;Meiri & Dayan, 2003). This correlation between body size and temperature variations led to the idea of an integrated Cope-Bergmann hypothesis, suggesting that Cope's rule may simply be an evolutionary manifestation of Bergmann's rule (Hunt & Roy, 2006). Cope-Bergmann hypothesis predicts that the size of organisms increases in relation to climatic cooling, and it was applied to micro-organisms such as verified in Cenozoic deep sea ostracods (Hunt & Roy, 2006) and even Cenozoic planktonic foraminifera (Schmidt et al., 2004(Schmidt et al., , 2006. However, Cope's rule implies a number of causal factors to the increase of body size through time (such as endothermy and prey-predator relationships) that only applies to multicellular organisms with sexual reproduction (Hone & Benton, 2005). Moreover, a common character of Cope's rule is the observation that the general increase in body size operates at the level of higher clades such as classes and orders but not particularly at the level of lower clades down to the family, genera, and species (Novack-Gottshall & Lanier, 2008). This is therefore in strong contrast to unicellular calcareous phytoplankton for which we see a transient increase in cell size within lower clade levels as exemplified in many examples of the literature given above. Similarly, Bergmann's rule has been essentially evoked in animal evolution, and causal factors of this rule involve temperature regulation as the primary adaptive mechanism for this rule is a decrease in the surface area to volume ratio, reducing heat loss in colder conditions (Timofeev, 2001;Meiri & Dayan, 2003). These implications make it difficult to apply this rule blindly to unicellular phytoplanktonic organisms that are essentially eurythermal.
Finally, although observations by Aubry et al. (2005) on the size of Mesozoic coccolithophorids are undeniable, a counter-example by the same prime author was given for the Neogene where most lineages of coccolithophorids underwent a general decrease in size in parallel to a long-term cooling trend, in strong contrast with the Cope-Bergmann hypothesis (Aubry, 2009). The applicability of Cope-Bergmann's hypothesis to calcareous nannoplankton thus appears questionable, both due to the counter-examples occurring in various lineages of this group, and to the causal factors that this rule implies which operate at a higher level than the cell and higher taxonomic ranks.
However, some analogy might be drawn in our understanding of how climate change affects biodiversity because, as recently demonstrated in the Quaternary, rapid phenotypic size changes in calcareous nannoplankton almost always imply speciation events, which appear to be related to significant climate changes (Bendif et al., 2019;Beaufort et al., 2022). Moreover, direct causal factors that can be invoked for size variation in plankton may be indirectly related to temperature and, in fact, driven by other factors, like fertility and light availability. Since the density of cold water is higher than warm water, a probable explanation for the size variation in plankton through geological time could be changing in buoyancy (Eppley et al., 1967;Walsby & Reynolds, 1980). In cold periods, calcareous nannoplankton may be able to compensate for some of the effects of increasing buoyancy by increasing their size and consequently their weight. This is especially the case for those phytoplanktons that only can tolerate a specific range of light intensity, so the increasing size would allow them to sink to levels where light intensity is suitable (see Morgan & Kalf, 1979;Meeson & Sweeney, 1982;Atkinson 1994). On this basis, phytoplankton species that cannot adjust their size with buoyancy variation may become extinct, while those with wider light intensity tolerance may show no change in size.

Are morphological changes in C. ehrenbergii and M. undosus lineages related to climaticallycontrolled episodes of speciation?
If rapid phenotypic size changes observed in calcareous nannoplankton lineages often result from diversification, then short-and long-term climate and environmental changes likely operate as causal factors via biogeographical partitioning. A compilation of calcareous nannoplankton diversity through time (Bown et al., 2004) allowed Bown (2005) to suggest that Cretaceous nannoplankton diversification occurred during cold intervals, supported by increased paleobiogeographical partitioning of oceanic photic-zone environments and the establishment of high-latitude provinces.

A possible relationship with the peak in global nannoplankton diversity at 76 Ma
We cannot exclude that the observed shift in the size of C. ehrenbergii reflects the origination of a new, larger species that became common in the late Campanian-Maastrichtian (possibly C. hilli, see taxonomy in section 4.1). We notice that this size change is relatively fast in the term of geologic time (less than 500 kyr, Supplementary Appendices 3 and 4, and Fig. 3), and the length of C. ehrenbergii remains almost constant around the same average after the shift (Fig. 3). Such a pattern is quite typical of Gould & Eldredge (1977) punctuated equilibrium model characterized by long periods of stasis punctuated by rapid shifts in morphology interpreted as evolutionary pulses. Such patterns are common in coccolithophores (Knappertsbusch, 2000;Geisen et al., 2004). Remarkably, the timing of the shift in size from a mean of around 5.5 to 6.5 µm in C. ehrenbergii coincides with the first occurrence (FO) of Microrhabdulus zagrosensis which also corresponds to a transient, rapid shift in morphology of the M. undosus group. Mircrorhabdulus rods resemble in many ways central processes observed in many coccolith lineages and bear various morphologies. However, it has never been found attached so far to a Cretaceous coccolith and the origin of this nannolith lineage remains obscure. Hence, size changes related to M. undosus cannot be attributed with confidence to any putative changes in the cell size of a fossil phytoplanktonic algal group. Rapid changes observed in this taxon can at best represent episodes of diversification within the lineage. The coincidence in the timing of the rapid increase in C. ehrenbergii and the emergence of M. zagrosensis can be dated at around 76 Ma thanks to the first occurrences of R. cf. R. calcarata (planktonic foraminifera) and Uniplanarius trifidus at around 250 m (Fig. 3). Moreover, the timing of these events also coincides with a significant increase in the abundance of benthic foraminifera in the Shahneshin section that was interpreted as a relative sea-level low  and correlates precisely with a global sealevel low delineated in the Miller et al. (2005) sea-level estimates of New Jersey (details in Kominz et al., 2008, figure 10). We cannot exclude that the apparent rapidity of what we interpret as an evolutionary pulse in both C. ehrenbergii and M. undosus group is due to a hiatus in nannofossil Zone CC22 associated with this sea-level low. Nevertheless, the coincidence of these morphological shifts in both lineages suggests a possible common external forcing on nannoplankton evolution. Figure 11 summarizes the results obtained in our study, updates the phylogeny of Microrhabdulus through the Late Cretaceous and compares these evolutionary trends to (1) the recent TEX86 compilation of O'Brien et al. (2017), (2) the compilation of bulk carbonate carbon isotopes drawn from the Late Cretaceous English chalk standard of Jarvis (2006) and the late Campanian-Maastrichtian Danish chalk (Thibault et al., 2012;2016) and (3) Bown et al. (2004) global diversity of calcareous nannofossils. This figure supports a strong link between rises in the global diversity of nannoplankton and cooling intervals in the lower Cenomanian as well as in the Santonian to Maastrichtian. In particular, the peak of global diversity reached at 76 Ma corresponds broadly to an episode of acceleration of surface-water cooling in the late Campanian and to the so-called Late Campanian Event, a ca. 1 per mil negative carbon isotope excursion (Fig. 11). This timing corresponds exactly to that of the positive shift in the sizes of C. ehrenbergii and M. undosus and to the first occurrence of Microrhabdulus zagrosensis (Figs 5 and 6). We infer here that our observations illustrate an intimate link between climatic cooling, the Campanian carbon cycle and speciation, and strongly support that the Late Cretaceous nannoplankton peak in diversity was essentially fueled by cooling. The rise in diversity that appears to precede this interval centered around 76 Ma is most likely due to the wide binned average used to draw nannofossil global diversity. Both C. ehrenbergii and M. undosus groups show a significant increase in relative abundance through the late Campanian to Maastrichtian interval and the onset of this increase coincides with the shift in morphology observed in both taxa at 76 Ma.
The late Campanian acceleration of cooling and associated Late Campanian carbon isotope event around 76 Ma have been linked to tectonic uplifts and formation of reliefs around the Tethys, enhancing atmospheric CO2 consumption by increasing continental weathering and erosion (Chenot et al., 2016(Chenot et al., , 2018Corentin et al., 2022). Such a supra-regional tectonic uplift associated with a global sea-level low not only could have a prominent role in reducing global temperature but could also have a significant effect on seasurface water fertility and oceanic circulation in Tethys. In such conditions, a bigger cell size could have been advantageous in environments with high nutrient availability since as cell size increases, the surfaceto-volume ratio decreases, and the ability to obtain nutrients and light decreases (e.g., Marañón, 2015). Hence, considering the fact that nutrient supply and light availability are regarded as two crucial factors in controlling modern phytoplankton size structure (e.g., Marañón, 2015), the record of size changes in our study may not be explained by temperature alone. The transient increase in abundance of C. ehrenbergii, together with the emergence of larger C. ehrenbergii and Microrhabdulus zagrosensis across the late Campanian to early Maastrichtian matches well the cool Late Campanian Event (Fig. 11), and this episode might have triggered further modifications in the ocean circulation and structure, affecting the buoyancy of plankton, and fertility levels. This interaction could have increased the persistent selection pressure on calcareous nannoplankton toward phenotypic divergence and provincialism. The results of this work suggest that the peak in global nannoplankton diversity reached at 76 Ma could even be underestimated as very few of the numerous calcareous nannofossil lineages of the Campanian interval have been so far the object of consistent biometric studies.

Is the emergence of M. sinuosus related to the mid-Maastrichtian event?
One of the striking features of the Maastrichtian nannofossil zonal and subzonal schemes is a sudden shift from the neat dominance of biohorizons represented by last occurrences in the early Maastrichtian to the sole presence of first occurrence biohorizons in the late Maastrichtian interval. Thibault (2016) discussed this feature and showed that at least for the South Atlantic, the mid-Maastrichtian, and to a lesser extent the late Maastrichtian, are characterized by several discrete origination events. Several nannofossil species, including Micula premolisilvae, Micula praemurus, Micula murus, Lithraphidites quadratus, Ceratolithoides amplector and Ceratolithoides kamptneri have been shown to first originate worldwide in the mid-Maastrichtian. This interval is also characterized by a significant increase in the size of the common genus Arkhangelskiella due to the massive occurrence of a large representative of this group (A. maastrichtiensis Burnett, 1997, syn. A. cymbiformis var. W, Varol, 1989Thibault, 2010). The 2 myr-long bins of the estimated nannoplankton global diversity curve somewhat prevent the distinction of this event in the global compilation, but a clear increase in origination rate and diversity of nannofossils has been reported at the same time interval in the southern high latitudes and in the South Atlantic (Huber & Watkins, 1992;Thibault, 2016). This trend is also observed in planktonic foraminifera in the areas mentioned above as well as at low and middle latitudes (Boersma & Schackleton, 1981;Boersma, 1984;Caron, 1985;Huber, 1990Huber, , 1992Li & Keller, 1998). Interestingly, the rise in the diversity of planktonic organisms appears to be paralleled by potential diversity increases in cephalopods (southwestern Francenorthern Spain, (Ward et al., 1991); northwest Pacific, (Jagt-Yazykova, 2011, 2012; Mexico, (Ifrim et al., 2004(Ifrim et al., , 2010; Seymour Island, Antarctica, (Witts et al., 2015)) and bivalves (Seymour Island, Antarctica, (Macellari, 1988)), though non-regulated inoceramids and rudists suffered a significant decline during the same interval (Ward, 1990;MacLeod & Ward, 1990;MacLeod, 1994;Johnson & Kauffman, 1996). Such changes in diversity among independent lineages and clades reduce the plausibility of a taphonomical bias. Overall, this event is spread across the early/late Maastrichtian boundary interval over ca. 2 myr and spans the two transitions from the early Maastrichtian cooling event to the mid-Maastrichtian warming, and the mid-Maastrichtian warming to the late Maastrichtian cooling (Thibault, 2016). The first occurrence of Microrhabdulus sinuosus at around 69.25 Ma is inscribed into this general mid-Maastrichtian diversification event.
The Campanian/Maastrichtian transition and lower Maastrichtian are marked by a pronounced longterm global cooling (close to 7ºC) paralleled by the decrease in global nannoplankton species richness following the late Campanian peak in diversity. While the Late Campanian cooling might have favored higher diversity via biogeographic partitioning, the following long-term trend of early Maastrichtian cooling could also be responsible for the extinction of a number of species that were better adapted to higher CO2 levels and/or higher temperatures and fertility levels, such as the biostratigraphic markers Reinhardtites levis, Tranolithus orionatus and Broinsonia parca constricta (Thibault, 2016;Razmjooei et al., 2020b). In contrast, the widespread episode of speciation in mid-Maastrichtian planktonic organisms does not coincide with climatic cooling but rather with higher temperature instability when global climate rapidly shifted from cool to warmer temperatures to cool again, from 69 to 67 Ma (Li & Keller, 1998;Thibault et al., 2016). Thus, the processes at play here differ from the case of the late Campanian and may be better explained by the conjunction of globally cooler temperatures with temperature instability (Fig. 11). The cooler temperatures of the Late Cretaceous could have fostered a rapid evolution in many lineages by quickly changing the direction of selection. In contrast, the temperature instability of the mid-Maastrichtian may have enhanced climatic heterogeneity, and hence geographical heterogeneities and provincialism triggered phenotypic divergence in nannoplankton and many other marine groups. Figure 11. Schematic synthesis of data obtained in this study with a reconstructed phylogeny of the Microrhabdulus lineage, across the late Albian to Maastrichtian, and comparison with global changes in sea-surface temperatures from TEX86 (O'Brien et al., 2017), global calcareous nannoplankton diversity (Bown et al., 2004), and a standard bulk carbonate carbon isotope curve compiled from the English chalk standard of Jarvis (2006) and the late Campanian-Maastrichtian curve of the Danish Basin from Thibault et al. (2012Thibault et al. ( , 2016. The phylogeny of Microrhabdulus is based on Nannotax website (Young et al., 2017). Two distinct SST estimates derived from Tex86 are provided: (a) Tex86 H -SST calibration and (b) linear Tex86-SST calibration. See O'Brien et al. (2017) for details on these calibrations as well as for symbols used in the figure to distinguish the different deep-sea sites and sections used in this compilation.

Conclusions
Changes in the morphology of two Late Cretaceous nannofossil lineages have been investigated here in deposits from the Zagros Basin (Iran). A common event occurs at c. 76 Ma in the late Campanian with a sudden significant increase in the mean length of C. ehrenbergii and in the mean width, as well as a change in the shape of M. undosus that led us to define the emergence of Microrhabdulus zagrosensis n.sp. at that time. Later on, at c. 69 Ma in the mid-Maastrichtian, the second increase in mean width and maximum size of the M. undosus group lineage is observed, along with an increased complexity in the arrangement of the laths of the rod, leading us to define the emergence of Microrhabdulus sinuosus n.sp. Comparison of the timing of these morphological changes and emergence of new species with nannoplankton global diversity, global climate change and global patterns in planktonic organisms allowed us to suggest that these two origination events might illustrate so far hidden but significant episodes of morphological change linked to episodes of diversification within Late Cretaceous nannoplankton. Indeed, the first speciation event coincides with a global peak in nannoplankton diversity and a Campanian episode of acceleration of cooling associated with the Late Campanian carbon isotope event, while the second episode coincides with the mid-Maastrichtian Event, a ~2 myr-long interval of climatic instability and ocean reorganization characterized by diversification in planktic organisms and several marine invertebrate groups, as well as by the extinction of non-regulate inoceramids. Our results call for a closer examination of potential morphological change within the numerous Santonian to Maastrichtian calcareous nannofossil lineages.