Geographic and temporal morphological stasis in the latest Cretaceous ammonoid Discoscaphites iris from the U.S. Gulf and Atlantic Coastal Plains

Abstract We examine temporal and spatial variation in morphology of the ammonoid cephalopod Discoscaphites iris using a large dataset from multiple localities in the Late Cretaceous (Maastrichtian) of the U.S. Gulf and Atlantic Coastal Plains, spanning a distance of 2000 km along the paleoshoreline. Our results suggest that the fossil record of D. iris is consistent with no within-species net accumulation of phyletic evolutionary change across morphological traits or the lifetime of this species. Correlations between some traits and paleoenvironmental conditions as well as changes in the coefficient of variation may support limited population-scale ecophenotypic plasticity; however, where stratigraphic data are available, no directional changes in morphology occur before the Cretaceous/Paleogene (K/Pg) boundary. This is consistent with models of “dynamic” evolutionary stasis. Combined with knowledge of life-history traits and paleoecology of scaphitid ammonoids, specifically a short planktonic phase after hatching followed by transition to a nektobenthic adult stage, these data suggest that scaphitids had significant potential for rapid morphological change in conjunction with limited dispersal capacity. It is therefore likely that evolutionary mode in the Scaphitidae (and potentially across the broader ammonoid clade) follows a model of cladogenesis wherein a dynamic morphological stasis is periodically interrupted by more substantial evolutionary change at speciation events. Finally, the lack of temporal changes in our data suggest that global environmental changes had a limited effect on the morphology of ammonoid faunas during the latest Cretaceous.


Introduction
Ammonoid cephalopod mollusks (commonly referred to as "ammonites" by Cretaceous ammonoid workers; Wright et al. 1996) are classic organisms for evolutionary studies thanks to their excellent fossil record, high rates of speciation and extinction, and preservation of morphological changes through ontogeny (see references in reviews by De Baets et al. 2015a; Monnet et al. 2015;Yacobucci 2015). It is often argued that they show high levels of morphological/ecophenotypic plasticity similar to modern cephalopods (Doubleday et al. 2016). Intraspecific variability is rarely quantified in ammonoids, particularly for "heteromorph" or uncoiled ammonoids (Kakabadze 2004), but their fossil record provides the opportunity to examine large sample sizes geographically and temporally-perfect for testing hypotheses of macroevolutionary change versus stasis in the fossil record.
In this study, we examine the morphology of the ammonoid species Discoscaphites iris (family Scaphitidae) across the Atlantic Coastal Plain (ACP) and Gulf Coastal Plain (GCP) of the United States ( Fig. 1) using collections from multiple sites. Scaphitid ammonoids are FIGURE 1. Locality map illustrating sites where Discoscaphites iris is found across the Atlantic Coastal Plain (ACP) and Gulf Coastal Plain (GCP) (black and white circles). Labeled black circles are sites that yielded specimens for this study. Black crosses are important drill cores used for biostratigraphic correlation of microfossil taxa (see Landman et al. 2004a;Larina et al. 2016). SRC, Santee Reserve Core; BRC, Bass River Core. Lower portion of figure is a cross section (SW-NE) illustrating general lithologic character/facies distribution and formations that contain the D. iris Zone across the ACP and GCP. well preserved, widely distributed, and locally abundant in Late Cretaceous shallowmarine sediments across the Northern Hemisphere (e.g., Landman and Waage 1993;Machalski 2005;Landman et al. 2010Landman et al. , 2014. Discoscaphites in particular is one of the last common ammonoid genera found at the end of the Cretaceous (see review by Landman et al. 2015), and in North America, numerous species occur in the Maastrichtian of the ACP, GCP, and the Western Interior Seaway (WIS) (Fig. 2). Discoscaphites iris is abundant in upper Maastrichtian sediments across the ACP and GCP and defines the D. iris Assemblage Zone-the highest ammonoid biostratigraphic zone in North America (Fig. 2) (Landman et al. 2004a;Larina et al. 2016;Witts et al. 2021). A single fragmentary specimen has also been reported from the upper Maastrichtian of Libya (Machalski et al. 2009), but this occurrence requires further study, and the species is otherwise unknown outside North America. Our sample from the ACP and GCP therefore represents a well-known and well-delineated "time slice" (see Koch 1996) and provides an opportunity to examine morphological variation in an ammonoid species across virtually its entire geographic and temporal range. These data provide quantitative information on the range, limits, and potential drivers of intraspecific variation and adaptational change (or lack thereof) that characterize an ammonoid species. They also contribute to broader issues of evolutionary tempo during the latest Cretaceous (e.g., the species stasis debate; see Lidgard and Hopkins 2015), and the drivers and predominate mode of ammonoid evolution and speciation FIGURE 2. Biostratigraphic framework for the Gulf and Atlantic Coastal Plains, USA, based on (1) calcareous nannofossils and (2) ammonite zonation (modified from Larina et al. 2016), with tentative correlation to the Western Interior Seaway (WIS) zonation (based on data in Landman et al. 2004aLandman et al. , 2007Cobban et al. 2006). Also plotted is the phylogenetic analysis of relationships among North American Discoscaphites species from the consensus tree of Landman et al. (2007), with the genus Trachyscaphites as the outgroup. Vertical black lines indicate observed stratigraphic ranges. MS, magnetostratigraphy; B. clinolob., Baculites clinolobatus; H. birkelu., Hoploscaphites birkelundae; H. nebras., Hoploscaphites nebrascensis; Trachyscaph., Trachyscaphites; D. sphaero., Discoscaphites sphaeroidalis; D. jerseyen., Discoscaphites jerseyensis. (Wani 2011;Yacobucci 2015). Finally, a better understanding of the intraspecific evolutionary changes of species immediately preceding the end-Cretaceous (K/Pg) mass extinction event will shed light on the abundance and "health" of ammonoid populations immediately before their terminal extinction.
Background.-The Discoscaphites iris Zone is characterized in the ACP and GCP by the co-occurrence of the index ammonoid with (1) the calcareous nannofossil Micula prinsii-the marker of the uppermost Maastrichtian Subzone CC26b in the scheme of Perch-Nielsen (1985) or UC20d TP in the Tethyan scheme of Burnett (1998)-and (2) the dinoflagellates Palynodinium grallator, Disphaerogena carposphaeropsis, and Thalassiphora pelagica (Landman et al. 2004a,b;Larina et al. 2016). It also coincides with planktonic foraminiferal zones CF1 (Plummerita hantkeninoides), CF2, and possibly the upper part of CF3 (Pseudoguembelina hariaensis) (Abramovich et al. 2011;Witts et al. 2021). These biostratigraphic markers all overlap with the upper part of magnetochron C30N and overlying chron C29R (Landman et al. 2004a;Abramovich et al. 2011;Larina et al. 2016;Witts et al. 2021), the base of which is currently dated at ∼340-400 ka before the K/Pg boundary (66.04 Ma) (Fig. 2) (Batenburg et al. 2012;Thibault 2018;Ogg 2020). Although its precise duration is unknown, these combined data constrain the D. iris Zone to the last 1 Myr and more likely the last ∼400 kyr of the Maastrichtian. As a minimum estimate for its duration, recent cyclostratigraphic analyses in the GCP suggest the D. iris Zone could represent as little as 185 kyr (Naujokaitytė et al. 2021).
As previously mentioned, this time interval is of interest due to its proximity to the K/Pg mass extinction event. Although substantial evidence suggests the primary cause of this event is a bolide impact at Chicxulub in the Gulf of Mexico (e.g., Schulte et al. 2010;Hull et al. 2020), the final ∼400 kyr of the Maastrichtian are characterized by dynamic environmental changes. These coincide with emplacement of the "main phase" of the Deccan Traps Large Igneous Province (LIP) in India (Schoene et al. 2019;Sprain et al. 2019), linked to a period of rapid global climate warming and cooling (Barnet et al. 2018;Hull et al. 2020; also see Dzombak et al. 2020). Shifts in the distribution, richness, and abundance of various groups of marine plankton occur alongside these climate changes before the K/ Pg boundary (Olsson 2001;Thibault 2016;Vellekoop et al. 2019), while morphological changes (dwarfing) and increased fragmentation in some planktonic foraminifera hint at stressful conditions in the oceans (MacLeod et al. 2000;Keller and Abramovich 2009;Henehan et al. 2016;Gilabert et al. 2021). The influence of these environmental changes on ammonoid species has been debated, with some workers suggesting widespread global decline of ammonoids leading into the K/Pg (e.g., Stinnesbeck et al. 2012) and others supporting species' health and even increased abundance during the latest Cretaceous (Landman et al. 2014Witts et al. 2015Witts et al. , 2018Witts et al. , 2021.

Methods
To date, Discoscaphites iris has been recorded from >15 localities across the ACP and GCP (see summary in Witts et al. 2021). The abundance of specimens at each locality varies due to local differences in preservation and differing sampling intensities. Our final morphometric dataset consists of 328 specimens collected from nine localities in Texas, Missouri, Mississippi, and New Jersey, representing a ∼2000 km transect from SW to NE and encompassing the full geographic spread of these occurrences ( Fig. 1). Some examples of D. iris are illustrated in Supplementary Figures 1 and 2. Specimens are reposited in the collections of the American Museum of Natural History (Division of Paleontology) (AMNH), Department of Earth and Planetary Sciences, University of New Mexico (UNM), U.S. Geological Survey (USGS), Monmouth County Amateur Paleontological Society (MAPS), and the Museum at the Black Hills Institute of Geological Research (MBHI) (full details for each specimen are available in Supplementary Dataset 1). The stratigraphy of most of the localities these specimens are derived from is detailed in the literature (e.g., Landman et al. 2004bLandman et al. , 2007Larina et al. 2016;Witts et al. 2018Witts et al. , 2021 and briefly reviewed in the following sections; a summary of locality information is provided in Table 1.
New Jersey.-The D. iris Zone is present at the top of the Tinton and New Egypt Formations, which outcrop in Monmouth County, New Jersey. Both units are glauconite-rich mudstones. At AMNH locality 3345, D. iris specimens are derived from the upper 20 cm of the New Egypt Formation below the K/Pg boundary, which at this locality is an unconformable lag deposit (also known as the "Main Fossiliferous Layer") containing reworked macro-and microfossils (Landman et al. 2004b). In the Manasquan River basin (e.g., AMNH localities 3335 and 3372), material comes from the upper 1 m of the Tinton Formation, primarily from the top 20 cm, which is a richly fossiliferous unit known as the "Pinna Layer" (Landman et al. 2007). An iridium anomaly occurs at the base of the "Pinna Layer," suggesting a conformable K/Pg sequence, but the boundary itself is probably best placed at the top of this unit (see discussion in Landman et al. 2012b). Measurements from AMNH localities 3335, 3345, and 3372 were mostly taken from the published literature (Landman et al. 2004b(Landman et al. , 2007 and are combined due to geographic proximity. Mississippi.-The D. iris Zone occurs in both the mixed carbonate-clastic Prairie Bluff Chalk and the laterally equivalent siliciclastic Owl Creek Formation (Fig. 1). A 1.55-m-thick section of the Owl Creek Formation crops out at AMNH locality 3481 in Union County (Witts et al. 2018), comprising three distinct units: a 52-cm-thick lower, brown-colored muddy sandstone; a 35-cm-thick gray-colored micaceous muddy siltstone; and a 68-cm-thick yellow-colored bioturbated and well-bedded silty sandstone. Discoscaphites iris specimens come from both the brown and gray units, which are abundantly fossiliferous. Only fragments were found in the upper, yellow-bedded unit. The K/Pg boundary occurs at the base of the overlying Clayton Formation, which is a 30-cm-thick poorly sorted muddy quartz sand containing rip-up clasts of Owl Creek Formation, bioclasts, and altered Chicxulub impact spherules (Witts et al. 2018). Lithologically similar sediments are also exposed in Union County at AMNH locality 3461, with specimens of D. iris abundant in the upper Owl Creek Formation. At this site, the Owl Creek Formation is a ∼1-m-thick yellow-colored, bioturbated, and fossiliferous silty sandstone. The K/Pg boundary occurs at the base of the Clayton Formation, which is a 15-cm-thick crossbedded sandstone that again contains altered Chicxulub impact spherules overlain by a 10-cm-thick bioclastic limestone.
At the Owl Creek type locality in Tippah County (AMNH loc. 3460), D. iris occurs throughout a ∼9-m-thick succession of bioturbated micaceous siltstones and sandstones with occasional laminated intervals (Larina et al. 2016). Macrofossils at this locality are often found in "pods" that may represent relict burrows and are beautifully preserved with original aragonitic shell material permitting isotopic analyses (e.g., Sessa et al. 2015;Ferguson et al. 2019). The Clayton Formation again overlies the Owl Creek Formation and is composed of 5 m of poorly fossiliferous orange fine-grained quartz sand and intervals of bryozoan-rich bioclastic limestone. The K/Pg boundary is coincident with the unconformable contact at the top of the Owl Creek Formation.
Missouri.-The Owl Creek Formation outcrops along Crowley's Ridge in Stoddard and Scott Counties (Stephenson 1955). The D. iris biozone occurs at AMNH locality 3458, which consists of a 1.5-m-thick outcrop of micaceous sandy siltstone overlain by 0.5-m-thick glauconite-rich muddy sandstones of the Danian Clayton Formation. The unconformable contact between these units marks the K/Pg boundary (Dastas et al. 2014). Fossils are abundant in the Owl Creek Formation, and like the Owl Creek type section in Mississippi, often occur in pods that may represent in-filled burrows.
Texas.-Material comes from outcrops along the Brazos River and its tributaries in Falls County that preserve a continuous and expanded record across the K/Pg boundary . Discoscaphites iris were collected from AMNH localities 3620 (Darting Minnow Creek) and 3621 (Cottonmouth Creek), which both contain 1.5to 2-m-thick outcrops of the upper Corsicana Formation, a fossiliferous dark mudstone. The lower part of the overlying Kincaid Formation contains complex clastic "event deposits" related to the Chicxulub impact event, the base of which marks the K/Pg boundary (Hansen et al. 1987;Hart et al. 2012;Yancey and Liu 2013). Most D. iris come from the Corsicana Formation, but one specimen is derived from the basal unit of the K/Pg event deposit.
To assess paleoenvironmental conditions at each locality, grain-size analyses were conducted on bulk sediment samples collected from stratigraphic units or intervals containing D. iris specimens (Table 1). Data for localities in New Jersey were taken from the published literature (Landman et al. 2004b(Landman et al. , 2007. For other localities, sediment samples were physically disaggregated and mixed with a 6% hexametaphosphate solution for 10 minutes. They were then wet sieved in a 63 μm sieve to separate sand-and mud-size (<63 μm) fractions, and percentages of these grain-size fractions were then calculated based on weight loss.
Morphometric Measurements.-Morphometric parameters were measured on wellpreserved adult specimens and are the same as those described in Witts et al. (2020) for the closely related species Hoploscaphites nicolletii (Fig. 3). All measurements were made using electronic calipers (accuracy of 0.01 mm) on actual specimens. The adult shell of D. iris consists of two parts, a closely coiled phragmocone and a slightly to strongly uncoiled body chamber. The adult phragmocone is the part of the phragmocone that is exposed in the adult shell. The body chamber consists of the shaft, beginning near the last septum, and a hook terminating at the aperture. The point at which the hook curves backward is called the point of recurvature. Dimorphism is present in D. iris, as in all scaphitid ammonoids and many modern cephalopods, and is interpreted as sexual in nature (Landman and Waage 1993;Davis et al. 1996;Landman et al. 2010). The dimorphs are referred to as the macroconch, presumably the female, and the microconch, presumably the male. Dimorphism is generally expressed by differences in size, robustness, and degree of uncoiling. Results for each dimorph are described separately. A full taxonomic description is available in Supplementary Material.
Maximum length (LMAX) was measured from the venter of the phragmocone to the venter of the hook parallel to the umbilical margin; whorl height was measured at three points on the shell: the end of the phragmocone (H P ), midshaft (H S ), and the point of recurvature (H H ); whorl width was also measured at these same points (W P , W S , W H ). The ratios of whorl width to whorl height were calculated at each of these three points (e.g., W P /H P ) and provide a measure of the degree of whorl compression. Two additional shape ratios were also calculated; the ratio of maximum length to whorl height of the phragmocone along the line of maximum length (LMAX/ H P ) is a measure of the degree of uncoiling. The ratio of maximum length to whorl height at midshaft (LMAX/H S ) is a measure of the degree of curvature of the body chamber in lateral view. If the outline of the body chamber in lateral view is a semicircle, the ratio equals 2. The ratio applies only to macroconchs, because the umbilical seam of the body chamber usually coincides with the line of maximum length in these forms, and the whorl height is the distance from the line of maximum length to the venter of the body chamber (equivalent to the radius in the case of a semicircle).
Shape measurements were made at each locality, but whorl compression ratios were not measured on specimens from the Brazos River localities, as these specimens were still embedded in matrix. It is important to note that these data represent variation among mature adults rather than ontogenetic variation (Landman et al. 2008). This is because the mature stage of scaphitid ammonoids is well defined by the uncoiling of the body chamber (e.g., Landman and Waage 1993;Landman et al. 2010); whorl compression ratios taken at different points on the shell provide some information on ontogenetic changes.
Statistical Analyses.-Morphological changes were evaluated using box-and-whisker plots to graphically examine the entire distribution of morphological parameters across the D. iris Zone (sensu Monnet et al. 2012). We also calculated the coefficient of variation (CV) for size and shape ratios as well as whorl compression ratios at each site. CV is the standard deviation divided by the mean value of each morphological trait and is often used as a measure of intraspecific variation in ammonoids (e.g., De Baets et al. 2013Baets et al. , 2015aKlein and Landman 2019). Nonparametric Mann-Whitney U-tests were used to evaluate the statistical significance of changes in mean morphological trait values between localities. To correct for multiple comparisons we adjusted p-values generated by the Mann-Whitney tests for each morphological trait using three methods. We applied a Bonferroni correction and also controlled for the false discovery rate using the methods of Benjamini and Hochberg (1995) and Benjamini and Yekutieli (2001). Only results considered below are those that were consistently statistically significant across all four of these methods, or at least three out of four methods if the conservative Bonferroni correction did not show significance.
We also explored relationships between morphological traits and several environmental variables available at each locality using linear modeling via the LM function in R (R Core Team 2021). Variables considered for each locality were paleolatitude (available using the Paleolatitude Calculator: www.paleolatitude.org; van Hinsbergen et al. 2015) and present-day longitude, which act as reliable proxies for original geographic location and distribution of the samples along the paleoshoreline of the ACP and GCP, and sediment grain size defined as percent of sandsized material derived from our analyses (Table 1). Grain size varies according to lithology, which itself varies according to numerous factors such as energy level in the environment of deposition, relative water depth, or position relative to shore (Jacobs et al. 1994;Klein and Landman 2019). These physical changes occur along with changes in various biological factors (see "Discussion"). Models were run for each morphological trait and separately for macroconchs and microconchs (Table 2). Separate linear models were also run excluding specimens from the Brazos River localities, which lack shell compression data. All statistical and linear modeling analyses were conducted in R (R Core Team 2021).

Results
The total distribution of size, shape, and shell compression measurements for specimens of Discoscaphites iris are illustrated in Figures 4 and 5 as box-and-whisker plots. Plots are arranged geographically from Texas to New Jersey along a general SW-NE transect. For shape traits like maximum length (LMAX) and degree of uncoiling (LMAX/H P ), a clear difference is visible between macroconchs and microconchs at each locality (Fig. 4A,B). Macroconchs are larger than microconchs, while microconchs are more rounded, reflecting the inflated body chamber and straight umbilical margin in macroconchs. All shape traits (also including degree of curvature [LMAX/H S ] for macroconchs; Fig. 4C) show overlap between localities. The ratio of the largest to smallest specimen ranges from 1.14 (AMNH loc. 3461, Ellis Pit) to 1.83 (AMNH loc. 3458, Crowley's Ridge) for macroconchs, and from 1.34 (AMNH loc. 3481, "4 th St") to 2.38 (AMNH loc. 3458) for microconchs. Whorl compression ratios also show overlap at all sites between macro-and microconchs (Fig. 5), and between localities across the entire extent of the D. iris Zone with no consistent pattern. Full morphological and environmental datasets are available in Supplementary Datasets 2 and 4.
The range of intraspecific variation is similar to estimates from other studies of morphological traits in scaphitids (Landman 1987;Klein and Landman 2019;Witts et al. 2020). Values of the CV for shape traits range from 5% to 21% for both dimorphs and show few clear patterns (Fig. 6, Table 3). Most shape traits show CV values of <10%. LMAX generally shows the largest variation in both dimorphs (Fig. 6A,B), while LMAX/H P and LMAX/H S (for macroconchs) are less variable. Whorl compression CV values in macroconchs range from 7% to 23%, with W P /H P (whorl compression at the base of the phragmocone) consistently the most variable (Fig. 6C,D). Microconchs show a similar pattern, with values of CV ranging from 8% to 22%; W P /H P again consistently shows the largest variation and W H /H H (compression of the hook) the least. Interestingly, CV values for all three whorl compression traits from the Owl Creek type section (AMNH loc. 3460)-the locality with the longest stratigraphic time series-show the smallest range, tightly clustered around 14%. No correlation is observed between CV values and either sample size or stratigraphic thickness of any locality ( Supplementary Figs. 3, 4).
The results of pairwise Mann Whitney U-tests show that significant differences in morphology of both D. iris macro-and microconchs are present between several localities. However, after adjusting p-values for multiple comparisons using three different methods (a Bonferroni correction and attempting to control for the false discovery rate using the methodologies of both Benjamini and Hochberg [1995] and Benjamini and Yekutieli [2001]), only five results were consistently statistically significantly different (Table 4). Maximum length (LMAX)-essentially a proxy for size at maturity-of D. iris macroconchs between AMNH localities 3460 and (1) AMNH localities 3620/3621 (Brazos River), (2) AMNH locality 3481, and (3) New Jersey localities. In two cases (AMNH loc. 3460 and AMNH loc. 3481 vs. New Jersey localities), macroconchs also show a consistently significant difference in whorl compression at midshaft (W S /H S ). Five comparisons were statistically significant in three out of four analyses; macroconch LMAX Linear modeling indicates that grain size shows a statistically significant relationship with macroconch LMAX in all models (models A and C) ( Table 5). Microconch LMAX also shows a statistically significant relationship with both present-day longitude and paleolatitude when data for all localities are included (model B), but these relationships are not maintained for longitude when data from the Brazos River localities in Texas (AMNH locs. 3620 and 3621) are excluded (model D). Macroconch LMAX shows a significant relationship with paleolatitude, but in this case only when data from Texas are excluded (model C). LMAX/ H P and LMAX/H S show no significant relationships with environmental variables. Whorl compression at the base of the phragmocone (W P /H P ) in microconchs shows a statistically significant relationship with paleolatitude and grain size (models B and D). Whorl compression at the hook (W H /H H ) in both dimorphs shows a significant relationship with paleolatitude (models A-D). Despite these correlations, however, all linear models, including those producing statistically significant relationships, have very low r 2 values (<0.2/20%) ( Table 6). The relationship between macroconch LMAX and grain size shows the highest r 2 values, suggesting that it explains 12%-18% variance in models A and C. Overall, these analyses indicate at best a weak relationship with environmental parameters in some traits, with no single environmental parameter influencing all traits. Model results are illustrated in Supplementary  Figures 5-10.
Temporal trends in morphometric data were assessed at AMNH locality 3460 (the Owl Creek type section, Mississippi), which provides the longest continuous stratigraphic record among our localities (Fig. 7) (Sessa et al. 2015;Larina et al. 2016). Data from AMNH localities 3620 and 3621 (Brazos River, Texas) were also plotted against stratigraphy (Fig. 8). Here, data from the Corsicana Formation are presented in a composite section relative to the position of the K/Pg boundary at the base of clastic event deposits in the overlying Kincaid Formation (for details, see Witts et al. 2021). These data generally support the geographic box-and-whisker plots; in both cases, no clear directional trend is apparent in any morphological characters up-section, but a range of variation is present at any given stratigraphic horizon. Likewise, comparison with published paleotemperature estimates at AMNH locality 3460 (Sessa et al. 2015;Ferguson et al. 2019) reveal no trend (Fig. 7D).

Discussion
Ecophenotypy and Intraspecific Variation.-The range of morphometric and CV values at each locality represent varying degrees of intraspecific variation, and together with the occasional significance of paleolatitude and longitude in our linear modeling analysis, suggest that local populations of Discoscaphites iris with slightly differing morphologies developed along the shallow margins of the ACP and GCP during the latest Maastrichtian. Despite low statistical support, other trends in our data may support limited local ecophenotypic (plastic) variation in distinct paleoenvironmental settings. For example, the relationship between grain size and maximum length (LMAX) of macroconchs in linear models (Table 5) suggests that size at maturity in this dimorph of D. iris varied predictably with this parameter, with smaller specimens more common in more sand-rich environments. There are many factors affecting size at maturity in modern cephalopods that were likely also important in ammonoids (see De Baets et al. 2015a). Grain size is generally considered a proxy for the energy level in the environment of deposition and a rough proxy for relative water depth or position relative to shore (Jacobs et al. 1994;Klein and Landman 2019), which coincides with changes in turbidity, food/ nutrient availability, light levels, and even predation and competition. Because LMAX could also signify the age at maturity, it is possible that individuals living in quieter, offshore environments represented by finer-grained sediments grew larger and reached maturity later than those in more nearshore, sandy settings. Flexibility in developmental timing has been hypothesized as a major driver of rapid evolution in ammonoids (Yacobucci 2016), although it is challenging to assess the evolutionary significance of ecophenotypic variations in mollusks (e.g., Grey et al. 2008). LMAX is the trait most likely to show statistically significant differences between localities in our study (Table 4). This is consistent with the findings of Hopkins and Lidgard (2012), who demonstrated that organismal body size and related traits are particularly labile, with this result consistent across multiple clades. Previous studies of morphological change in both juvenile and adult scaphitids have noted a consistent pattern of increasing whorl compression in specimens and species from nearshore versus offshore environments-also using grain-size as a proxy (e.g., Jacobs et al. 1994;Klein and Landman 2019;Landman et al. 2020;Witts et al. 2020). Hydrodynamics suggest that a streamlined morphology is more suited to higher-energy conditions in shallow, nearshore settings and that whorl compression is a particularly plastic trait in ammonoids. Statistically significant differences in compression of the body   4. Morphological traits that show consistent statistically significant differences between localities for macroconchs (M) and microconchs (m). Based on four analyses: Mann-Whitney U-test, analysis of p-values using a Bonferroni correction for multiple comparisons, and assessment of the false discovery rate using the methods of Benjamini and Yekutili (2001) and Benjamini and Hochberg (1995 (Table 4), which are consistent with this hypothesis. The whorl compression trait that shows the largest intraspecific variability in both D. iris dimorphs is typically W P /H P (Fig. 6). CV values for whorl compression traits from stratigraphically expanded AMNH locality 3460 (Owl Creek type locality) are tightly constrained around 14% in both macro-and microconchs (Fig. 6C,D). This suggests lower levels of ecophenotypic variation in a stable environment, supported by the stratigraphically expanded nature and invariant sedimentology of this section, as well as paleotemperature data (Fig. 7). However, whorl compression values do not vary with grain size in our linear models, indicating this general relationship between water depth/shell compression is complex and potentially not fully captured using average grain-size data. This calls into question suggestions of a 1:1 relationship between environment and morphology/ecology, which is sometimes assumed in the literature for ammonoid taxa (Westermann 1996;Ritterbush and Bottjer 2012). It has also been suggested that offshore environments may be more permissible to morphological variations in scaphitids, as populations in these areas did not face the same selective pressures for compressed shells as those in more inshore, higher-energy FIGURE 7. Stratigraphic changes in selected Discoscaphites iris morphometric traits and paleotemperature at AMNH locality 3460 (Owl Creek type locality, Mississippi). Section modified from Larina et al. (2016). A, LMAX; B, LMAX/H P ; C, W S / H S . Filled black circles in A-C are macroconchs, open gray circles are microconchs. D, Paleotemperature data derived from oxygen isotope analysis of various groups of macrofossils (Sessa et al. 2015;Ferguson et al. 2019), converted to temperature using the equation of Grossman and Ku (1986) and assuming a δ 18 O seawater value of −1‰). Black circles, scaphitids; gray circles, other ammonoids (baculitids and sphenodiscids); +, infaunal bivalves; ×, epifaunal gastropods and ostreid bivalves. environments where such a shape is hydrodynamically favorable Peterman et al. 2020). Water-depth estimates for ACP and GCP localities in this study suggest most sites ranged from 20 to 50 m (Landman et al. 2004b(Landman et al. , 2007Sessa et al. 2015;Larina et al. 2016;Witts et al. 2018). It is possible that the higher CV values in size and shape traits for specimens of both dimorphs of D. iris from Brazos River are an example of this pattern; estimated water depths for these localities are greater than those for other GCP localities (Hart et al. 2012;Witts et al. 2021). Unfortunately, whorl compression values from this site are not available for comparison.
Effect of Differential Sampling, Time Averaging, and Facies.-It is important to consider the effects of time averaging on these data. Variations in sedimentation/accumulation rates can affect the range of morphological variability within a sample. Hunt (2004) suggested that greatly inflated variance is expected in samples from condensed intervals that accumulated over several tens of thousands of years or that are "lumped" from multiple horizons.
Although all our samples are temporally constrained to within the ∼200-500 kyr D. iris Zone and can be taken as a single time slice, it is unlikely that all localities are precisely temporally equivalent. Discoscaphites iris specimens were collected from restricted stratigraphic intervals (Table 1) <2 m thick, and data are binned at this scale (Figs. 4,5). The exact duration of these intervals is unknown, and sedimentation rates are clearly variable as reflected in lithological changes (Larina et al. 2016). Landman et al. (2007Landman et al. ( , 2012b and Witts et al. (2018) argued that the mode of occurrence of fossils in the "Pinna Layer" in New Jersey (AMNH locs. 3335 and 3372) and the "gray unit" at AMNH locality 3481 (which furnished most material in our study from these localities) indicate autochthonous accumulations that formed over short intervals of time and thus provide robust snapshots of in situ marine communities.
Despite temporal uncertainty, CV values for shape traits from the stratigraphically expanded and lithologically homogenous AMNH locality 3460 align closely with those of other, presumably more condensed, localities and would therefore suggest a similar degree of intraspecific variation despite time averaging (Fig. 4, Fig. 6A,B). Whorl compression traits may differ for other (ecophenotypic) reasons outlined in the previous section. At several localities (e.g., AMNH loc. 3460, AMNH loc. 3458) specimens were commonly found concentrated in fossiliferous pods that may represent hydrodynamic accumulations in seafloor burrow systems (Larina et al. 2016). This, combined with lowered sedimentation rates, could explain the greater range of morphological variability present in samples from AMNH locality 3458.
It should also be noted that our study only contains data from siliciclastic successions across the ACP and GCP. While these sites generally capture the range of environments that D. iris occurs in across its geographic range (Fig. 1), the species also occurs in the more carbonate-rich facies of the Prairie Bluff Chalk in Mississippi and Alabama (Larina et al. 2016;Naujokaitytė et al. 2021;Witts et al. 2021). Unfortunately, there are currently too few complete specimens from these localities to usefully compare with siliciclastic successions. This is partly a sampling artifact/bias linked to increased preservation potential for complete specimens in clastic successions, but it may also reflect a true environmental preference whereby D. iris is simply more common in siliciclastic environmental settings that tend to represent more nearshore environments. Studies of molluscan occurrence data from the ACP and GCP margins during a comparable latest Maastrichtian time slice (e.g., Sohl and Koch 1983, 1984, 1987 demonstrate lateral changes in the relative abundance of benthic taxa due to shifting paleoenvironments ( Fig. 1) (for summaries, see Koch 1996). Further work on the occurrence in different facies and preferred habitat of D. iris is needed to test this hypothesis.
"Dynamic Stasis" in Scaphitid Ammonoids: Clues from Modern Cephalopods, Paleoecology, and Phylogeny.-Modern cephalopods are well known for exhibiting morphological plasticity (Boyle and Rodhouse 2005). This is most often related to differing environmental conditions (ecophenotypy) and/or the development of complicated population structures depending on factors such as dispersal or migration ability regulating gene flow (e.g., Boyle and Boletzky 1996;Pérez-Losada et al. 2007;Zaleski et al. 2012;Braga et al. 2017). These features also figure prominently in the debate surrounding plausible mechanisms for speciation and morphological stasis in the fossil record (e.g., Eldredge et al. 2005;Yacobucci 2016). Scaphitids have small embryonic shells suggesting high adult fecundity (Landman 1987;De Baets et al. 2015b), and values of oxygen isotopes in earliest ontogeny are consistent with a surface-water habitat immediately following hatching, presumably as part of the plankton (Linzmeier et al. 2018). During this phase, passive dispersal via ocean currents is plausible, which in the absence of geographic or oceanographic barriers could lead to development of genetic cohesion among well-connected populations. However, this is strongly dependent on the amount of time juvenile ammonoids spent in the plankton Villanueva et al. 2016;Wani 2017). There are currently no reliable estimates for growth rates in ammonoids, but rapid growth (reaching maturity in <5 years) is likely , which might suggest limited dispersal potential in the plankton. Later-stage juvenile and adult scaphitids were nektobenthic and lived close to the seafloor based on overlap of shell oxygen isotopic data with benthic organisms (Fig. 7) (Sessa et al. 2015;Ferguson et al. 2019);Linzmeier et al. (2018) demonstrated the change from a planktonic to nektobenthic mode of life in scaphitids probably occurred relatively early in ontogeny, at approximately one whorl of postembryonic growth.
Analysis of the functional morphology of shells and well-preserved jaws (aptychi) suggest adult scaphitids exploited a low-energy planktivorous lifestyle (Landman et al. 2010(Landman et al. , 2012a. Despite their uncoiled heteromorph shells, scaphitids were capable of active swimming (Peterman et al. 2020), but evidence suggests adults did not undergo long-distance migration. For example, Cochran et al. (2003) found differences in the strontium (Sr) isotopic composition of scaphitid shells from different but laterally equivalent shallow-marine paleoenvironments in the Maastrichtian Fox Hills Formation of the WIS. Yahada and Wani (2013) examined morphology through ontogeny in samples of scaphitids from laterally equivalent but geographically separate localities in the Turonian of Japan. They found no statistically significant differences in the size of embryonic shells from the two localities but did observe significant differences in the thickness ratio (shell breadth/width) later in ontogeny. They argued that planktonic hatchlings were transported in surface currents between the two localities, but nektobenthic juveniles and adults did not migrate over long distances and formed distinct populations subject to local (ecophenotypic) selective pressures. Taken together, these studies suggest that scaphitids likely had significant potential for rapid morphological change in conjunction with limited dispersal capacity. This combination could fuel rapid evolutionary change (either punctuated, cladogenetic, or phyletic) in scaphitids responding morphologically to unique local environments and limited gene flow between isolated/semi-isolated populations; if paired with speciation, this is an interpretation that fits expectations of the broader ammonoid clade that demonstrates both high diversity and disparity (Yacobucci 2005;Wani 2011).
Our results provide empirical support for morphological stasis in the fossil record of D. iris throughout its entire geographic and stratigraphic range, which would suggest a punctuated or cladogenetic mode of evolution in this clade. Preliminary reconstruction of phylogenetic relationships among North American examples of Discoscaphites (Landman et al. 2007) and comparison with stratigraphic ranges support this hypothesis (Fig. 2). These data suggest that the geologically older and younger species comprise two clades; D. iris belongs to the geologically younger "Clade B" (Fig. 2) and co-occurs with Discoscaphites minardi, Discoscaphites sphaeroidalis, and Discoscaphites jerseyensis in the GCP and ACP. The recently described species Discoscaphites mullinaxorum from the upper Maastrichtian of Texas probably also belongs to this clade . Discoscaphites minardi first appears in the GCP and ACP in the middle part of the late Maastrichtian but persists alongside D. iris to the K/Pg boundary in the ACP.
"Clade A" contains Discoscaphites conradi, Discoscaphites rossi (which together form a separate clade), and Discoscaphites gulosus. These taxa occur in the ACP, GCP, and WIS in the early late Maastrichtian (Landman and Waage 1993;Landman et al. 2004a;Larina et al. 2016), and D. gulosus ranges to the K/Pg boundary (Landman et al. 2007). Morphological stasis in D. iris, alongside the co-occurrence of ancestor and/or sister species, lends support for a punctuated or cladogenetic pattern and suggests that anagenesis might not be an important contributor to evolutionary trends in this clade.
It is important to note that potentially ecophenotypic or intraspecific changes outlined above, especially with regard to shifting CV values at each locality, reflect population-scale variability that is not at odds with stasis of the species overall (see discussion in Lieberman et al. 1995;Lieberman 2009). Despite some statistically significant differences between localities, most morphological traits remain invariant across the geographic extent of the D. iris Zone and do not exhibit any kind of directional trend in terms of size or shape. Some statistically significant differences in trait values do occur, but there is generally no consistent pattern that might support morphological adaption across time or geography at the species level. In addition, morphological traits show only weak relationships with environmental parameters included in the linear models. Although a limited test of the hypothesis of ecophenotypy, these data do not support strong ecophenotypic variation at the species level. Size, shape, and compression traits measured at the population level also appear to exhibit low levels of intraspecific variation. These results suggest that the fossil record of D. iris is consistent with shifting trait values within local populations due to either some limited plastic response to local paleoenvironment or random intrapopulational variation (see discussion De Baets et al. [2015a] for other examples from the ammonoid fossil record), which may simply be compounded by the time-averaging and sampling constraints mentioned earlier.
Despite these caveats, our results clearly support no within-species net accumulation of phyletic evolutionary change across traits or, where temporal data are available, the lifetime of this species. Notably, data from the Owl Creek type locality (AMNH loc. 3460), the most expanded succession in our dataset, reveal a range of trait values comparable to more restricted sections, with no consistent temporal trends (Fig. 7). Any given stratigraphic horizon at this locality captures a range of variation similar to that of the entire dataset. In addition, paleotemperature estimates based on oxygen isotope analysis of macrofossils (Sessa et al. 2015;Ferguson et al. 2019) show a similar pattern. Both morphology of D. iris and the overall paleoenvironment at Owl Creek are consistent with fluctuating regimes around a relatively stable mean but no long-term directional trend. Whereas the precise duration of the Owl Creek stratigraphic section is unknown, the calcareous nannofossil Micula prinsii appears at the base of the succession alongside D. iris (Larina et al. 2016); as noted earlier, this taxon is restricted to the latest Maastrichtian. In many successions with wellconstrained age models, it first occurs within the ∼350 kyr interval corresponding to magnetochron C29R (Fig. 2) (Thibault 2018;Gale et al. 2020). Witts et al. (2020) demonstrated similar statistical support for evolutionary stasis across the temporal and geographic range of another closely related scaphitid ammonoid species, Hoploscaphites nicolletii, in the Maastrichtian WIS of North America. More specifically, this study suggested that the fossil record of H. nicolletii conforms to the expectations for "dynamic stasis" (e.g., Eldredge et al. 2005; Lieberman 2009); although there were statistically significant morphological changes both temporally and spatially, these were reversible, with a wide range of variation and geographic differences among local populations at any given horizon, some of which could be related to specific environmental conditions (i.e., ecophenotypy) (see also Landman et al. 2008). The exact mechanisms driving patterns of dynamic stasis are still unclear, with both differential selection across paleoenvironments acting in aggregate (Eldredge et al. 2005;Estes and Arnold 2007) or stabilizing selection with a fluctuating optimum associated with shifting environments (Hunt 2007;Hunt and Rabosky 2014) considered plausible by Witts et al. (2020); our results are in good accord with this interpretation. Thus, the study by Witts et al. (2020) also supports a model of punctuated/cladogenetic evolutionary change in scaphitid ammonoids.
Implications of Temporal Stasis for Pre-K/Pg Paleoenvironmental Change.-Size and shape changes documented by other authors in planktonic foraminifera during the latest Maastrichtian, temporally equivalent to the D. iris Zone (Larina et al. 2016;Naujokaitytė et al. 2021;Witts et al. 2021), have been related to environmental stresses associated with the emplacement of the Deccan Traps LIP, before the Chicxulub impact event and K/Pg mass extinction (Keller and Abramovich 2009;Henehan et al. 2016;Gilabert et al. 2021). Morphological changes have been detected in ammonoids before other mass extinction events linked to episodes of environmental change driven by LIP volcanism: for example, Kiessling et al. (2018) documented a reduction in size and morphological complexity in ammonoid assemblages from deep-water limestones correlated to the last 700 kyr of the Permian in Iran, coincident with the onset of Siberian Trap LIP volcanism and disruption to the global carbon cycle, but preceding the main phase of the end-Permian mass extinction. However, morphological variation in response to environmental stressors more broadly is a well-studied feature of the ammonoid fossil record. For example, spectacular morphological changes and evolutionary "jumps" within lineages occur during recoveries from carbon cycle perturbations and extinctions (e.g., Monnet et al. 2013), and morphological changes are commonly seen coincident with sea-level fluctuations and faunal invasions in epeiric seaway basins, particularly within endemic lineages (Yacobucci 2003;Klug et al. 2005;Witts et al. 2020).
Although it is difficult to assign a precise age to many samples of D. iris below the K/Pg boundary, the stratigraphic data from the Owl Creek type locality outlined earlier suggest no significant directional size or shape changes in this species of ammonoid during the latest Maastrichtian before the K/Pg boundary and Chicxulub impact event. Limited stratigraphic data from the Brazos River localities (AMNH locs. 3620 and 3621) are also consistent with temporal patterns from Owl Creek in showing no directional trend through time (Fig. 8).
Here, D. iris co-occurs in the uppermost Corsicana Formation below the K/Pg boundary with the index planktonic foraminifera Plummerita hantkeninoides, representing as little as 140 kyr of the latest Maastrichtian (Husson et al. 2014;Gale et al. 2020). In the Manasquan River basin, New Jersey (AMNH locs. 3335 and 3372), D. iris and eight other species of ammonite occur in the 20-cm-thick unit known as the "Pinna Layer" at the top of the Tinton Formation, associated with a prominent iridium anomaly (Landman et al. 2007). As mentioned earlier, fossils in the "Pinna Layer" likely represent a terminal Cretaceous community immediately before the time of the Chicxulub impact (see discussion in Landman et al. 2007Landman et al. , 2012bMiller et al. 2010). Morphology of D. iris from these localities falls within the range of variation seen at other, potentially older, sites (Figs. Fig. 4, 5), again suggesting no differences in size or shape through the duration of the D. iris Zone besides the limited population-scale ecophenotypic variations outlined earlier.
These results are inconsistent with the LIP-induced ecophenotypic changes proposed for other groups and suggest that any environmental changes related to Deccan volcanism had a limited effect on the morphology of North American ammonoid faunas before the Chicxulub impact and K/Pg mass extinction event.

Conclusions
Research on modern and ancient cephalopods supports the potential for rapid, ecophenotypic morphological change. Moreover, inferred poor larval and adult dispersal ability of scaphitids (Landman et al. 2012a;Yahada and Wani 2013;Linzmeier et al. 2018) increases the likelihood of isolation between populations, which, if combined with morphological change, could promote speciation resulting in high levels of diversity and disparity within the clade. The current paradigm of speciation by some combination of sympatry and micro-allopatry (Wani 2011;Yacobucci 2016), is consistent with this model and the overall pattern of high diversity and disparity in ammonoids more generally. What has remained less clear is how this combination of traits interacts with evolutionary mode to produce the observed patterns (i.e., evolutionary change primarily through phyletic adaptation vs. more punctuated cladogenesis). The results of this analysis indicate that the latest Cretaceous scaphitid Discoscaphites iris (like its relative Hoploscaphites nicolletii; Witts et al. 2020) demonstrates morphological stasis across its lifetime at the species level and even limited intraspecific variability at the population level. Moreover, no directional changes in species morphology are observed that could be related to contemporaneous Earth system changes during the latest Cretaceous (e.g., environmental stress driven by emplacement of the Deccan Traps LIP) before the K/Pg boundary and Chicxulub bolide impact. It is possible that evolutionary mode in the Scaphitidae follows the punctuated model of cladogenesis, wherein a dynamic morphological stasis is periodically interrupted by more substantial evolutionary change at speciation events. Preliminary phylogenetic analysis of Discoscaphites supports this claim. Additional studies that focus specifically on abundance and morphological change across the entire geographic and temporal duration of fossil species (especially those that hypothetically represent ancestors and descendants) could further demonstrate empirically whether the cladogenetic evolutionary mode is indeed pervasive in the highly speciose ammonoid clade or a phenomenon unique to scaphitids. Such work is certainly possible given the nature of the ammonoid fossil record and promises to shed further light on the broader mechanisms of speciation within this diverse clade.
R. Rovelli, A. Rowe, D. Ryan, and the late S. Klofak. We acknowledge and thank numerous property owners for providing access to field sites: A. Carrol and B. Carrol, S. Coffey (Owl Creek Type Locality, MS), W. C. Smallwood ("4 th St" MS), B. Stinchcomb (Crowley's Ridge, MO), and J. Mullinax and R. Mullinax (Brazos, TX). Thanks to S. Thurston, M. Slovacek, B. Husseini, and M. Conway (AMNH) for assistance with figures, specimen preparation, and facilitating access to AMNH collections. We are grateful to the staff at the Black Hills Institute of Geological Research (Hill City, SD, USA) for access to specimens in their care. Finally, we thank the associate editor (J. Crampton) and two reviewers (W. Allmon and K. De Baets) for their thoughtful comments, which greatly improved the paper. This research was funded by a Lerner-Gray Postdoctoral Research Fellowship at the AMNH and Richard Gilder Graduate School and a Postdoctoral Fellowship at UNM, both awarded to J.D.W. C.E.M. and N.H.L. acknowledge support from U.S. National Science Foundation grant no. 1924807. Additional funding was provided by the N.D. Newell Fund (AMNH).