Geological duration of ammonoids controlled their geographical range of fossil distribution

Relationship between latitudinal range and ammonitella diameter

The scatter diagrams for the aggregate Devonian–Cretaceous interval and each constituent period are shown in Fig. 2, which are derived from data summarized in Table 1. No relationship between latitudinal ranges and ammonitella diameters was found for the aggregate Devonian–Cretaceous interval nor for any constituent period (ANOVA, F = 0.309 and P = 0.58 Devonian–Cretaceous interval, F = 2.015 and P = 0.17 for the Devonian, F = 0.055 and P = 0.82 for the Carboniferous, F = 1.065 and P = 0.31 for the Permian, F = 0.354 and P = 0.55 for the Triassic, F = 0.691 and P = 0.41 for the Jurassic, F = 0.267 and P = 0.61 for the Cretaceous).

Relationship between latitudinal range and geological duration

The scatter diagrams for the aggregate Devonian–Cretaceous interval and each constituent period are shown in Fig. 3. Statistically significant relationships between latitudinal ranges and geological durations were found (ANOVA, F = 94.340 and P < 0.001 for the Devonian–Cretaceous interval, F = 7.739 and P < 0.01 for the Carboniferous, F = 8.148 and P < 0.01 for the Permian, F = 7.785 and P < 0.05 for the Jurassic, F = 36.976 and P < 0.001 for the Cretaceous). No significant correlation was found in the Devonian (F = 0.988 and P = 0.33) and Triassic (F = 1.624 and P = 0.21).

Relationship between ammonitella diameter and index of latitudinal range and geological duration

The scatter diagrams for the aggregate Devonian–Cretaceous interval and each constituent period are shown in Fig. 4. Statistically significant relationships between ammonitella diameter and the index of latitudinal ranges and geological durations were not found (P = 0.06, 0.54, 0.14, 0.76, 0.79, 0.98, and 0.37, for the aggregate Devonian–Cretaceous interval, Devonian, Carboniferous, Permian, Triassic, Jurassic, and Cretaceous, respectively). However, the larger indices are observed only in the smaller ammonitella diameters. In the Carboniferous, Triassic, and Jurassic, the larger indices tend to shift to the middle of the observed ranges of ammonitella diameters (Figs. 4C, 4D and 4F).

Controlling factors for latitudinal ranges of ammonoids

The scatter diagrams between latitudinal ranges and ammonitella diameters (Fig. 2) revealed no relationship between the chosen parameters. In contrast, the hatchling sizes of modern cephalopods together with developmental modes (planktic or benthic) influence the geographical range (Villanueva et al., 2016). The evolutionary history of ammonoids is long (Devonian–Cretaceous), thus the geographical ranges of ammonoids can be considered as an accumulation of multiple geological time slices. Latitudinal ranges of ammonoids had a statistically significant positive correlation with the geological durations, except for the Devonian and Triassic (Fig. 3). These relationships show that the longer the geological duration of taxa, the broader the latitudinal ranges through ammonoid evolutionary history. Such a positive relationship is similar to those in other fossil marine organisms (e.g., gastropods, bivalves, arthropods, foraminifers, radiolarians, crinoids; Jablonski, 1987; Miller, 1997; Liow, 2007; Foote et al., 2008). Evolutionary longevity of ammonoid taxa influenced their geographical ranges of fossil distributions (Fig. 3).

Figure 4 illustrates that the larger indices of the latitudinal ranges and the geological durations are observed only in the smaller ammonitella diameters. The exceptions are those in the Carboniferous, Triassic, and Jurassic, in which such relationships are blurred. These tendencies in the analyses that reduce the influence by the geological durations imply that the hatchling size was one of the main controlling factors for geographical ranges in ammonoid habitats at a given geological point in time: ammonoids with smaller hatchling sizes tended to have broader geographical ranges of ammonoid habitats.

Most ammonoid hatchlings are thought to have had a planktic mode of life (Kulicki, 1974; Kulicki, 1979; Kulicki, 1996; Drushchits, Doguzhayeva & Mikhaylova, 1977; Tanabe, Fukuda & Obata, 1980; Tanabe et al., 2001; Tanabe, Landman & Yoshioka, 2003; Tanabe & Ohtsuka, 1985; Landman, Tanabe & Shigeta, 1996; Westermann, 1996; Rouget & Neige, 2001; De Baets, Landman & Tanabe, 2015). In modern cephalopods, hatchlings smaller than approximately 3 mm are considered to have a planktic habit in the post-embryonic stages (Wani, 2011; De Baets et al., 2012; De Baets, Landman & Tanabe, 2015; Villanueva et al., 2016), which also suggests a planktic mode of life in most ammonoids. A planktic mode of life at the post-embryonic stage is thought to have had a significant impact on their geographical ranges (Landman, Tanabe & Shigeta, 1996; Westermann, 1996; Tajika & Wani, 2011; Ikeda & Wani, 2012; Brayard & Escarguel, 2013; Yahada & Wani, 2013; De Baets, Landman & Tanabe, 2015; Zacaï et al., 2016). However, there has been little discussion of how the hatchling sizes are related to their geographical ranges. The analyses that reduce the influence of geological duration in this study imply that the hatchling sizes were of importance to the geographical ranges of ammonoid habitats at each geological point in time, similar to those in modern cephalopods. This similarity to modern cephalopods shows a common reproductive strategy in ancient and modern cephalopods: the planktic mode of life at the post-embryonic stages was important in order to achieve wide geographic ranges.

Taxonomic treatment of ammonoid genera could also affect the geographical range and geological duration of the examined genus depending on how the examined species are attributed to a given genus. For example, Maeda (1993) examined “four” species of “two” genera of Late Cretaceous ammonoids and suggested that the examined nominal species was one species (i.e., one genus), based on the stratigraphic occurrence, morphological variation, and shell ontogeny. Such integration and separation of a genus would influence the geographical range and geological duration, the averages of ammonitella diameters, which impact analyses such as this study.

Comparison between different periods

The examination of the scatter diagrams for each period (Figs. 2–4) illustrates the characteristics of the relationship between the hatchling sizes and geographical ranges for each period. Scatter diagrams for the Permian and Cretaceous data demonstrate the same tendency as data for the aggregate Devonian–Cretaceous interval.

But scatter diagrams of the Devonian, Carboniferous, Triassic, and Jurassic (Figs. 2–4) demonstrate different tendencies. The ammonitella diameters in the Devonian tend to be larger than found in other periods and the maximum ammonitella diameter within the Devonian attained at least 6.0 mm (Table 1; De Baets, Landman & Tanabe, 2015). Several genera had ammonitella diameters larger than 3 mm, which is thought to be the approximated critical size for the planktic habit of the post-embryonic stage in modern cephalopods (Wani, 2011; De Baets et al., 2012; De Baets, Landman & Tanabe, 2015). De Baets, Landman & Tanabe (2015) mentioned a trend towards smaller ammonitella diameters from the Early to Late Devonian: the maximum ammonitella diameters were more than 3 mm in the Early Devonian, and less than 3 mm in the Late Devonian. This implies that ammonoids developed a reproductive strategy with smaller hatchling sizes during the Devonian, which possibly blurs the relationship between the latitudinal ranges, ammonitella diameters, and geological durations in the Devonian, at the examined temporal resolution (Fig. 3). Laptikhovsky, Nikolaeva & Rogov (2017) also concluded that the evolution of reproductive strategies in cephalopods in the geological past was marked by an increasing abundance of small-egged taxa, which agrees with the findings in this study.

The scatter diagram between the latitudinal ranges and geological durations in the Jurassic (Fig. 3C) shows a significant relationship. If only one long-ranged genus (Phylloceras of Phylloceratina) is excluded from the analyses, the scatter diagram showed no statistically significant correlation (F = 3.253 and P = 0.09). Such long-ranged genera are few in the examined data set and might indicate a possible problem with species determination; most ammonoid genera in the Triassic and Jurassic have shorter geological durations (Table 1; Table S1) (see Korn & De Baets, 2015, for a similar bias on the paleogeography of Devonian ammonoids). The larger indices of the geographical ranges and geological durations are observed not only in the smaller ammonitella diameters but also in the middle of the observed ranges of ammonitella diameters (Figs. 4C and 4D). Based on these values, the relationships between the hatchling diameters, geological duration, and geographical ranges in the Triassic and Jurassic are regarded as similar.

There were mass extinction events at the end of the Permian and Triassic when drastic taxonomic turnovers of ammonoid fauna occurred (e.g., House, 1988; Monnet, Brayard & Brosse, 2015; Brayard & Bucher, 2015; Longridge & Smith, 2015). In the Early Triassic and Early Jurassic, origination rates of ammonoids were high (Ceratitida originated in the Triassic and Lytoceratina and Ammonitina originated in the Jurassic; Neige & Rouget, 2015; Yacobucci, 2015; Longridge & Smith, 2015), which probably indicate radiations into open niche followed mass extinction events. At these times, ammonoids appear to have originated and resulted in an increased diversity rather than an increased geographical spread of each taxa (with the exception of some long-ranged taxa). This could blur the relationship between latitudinal ranges, geological durations, and ammonitella diameters in the Triassic and Jurassic (Figs. 2–4).

There was also a mass extinction event at the end of the Devonian, after which ammonoid diversity increased during the Carboniferous (Kullmann, 1981; Becker & Kullmann, 1996; Korn, Klug & Walton, 2015). As with the Triassic and Jurassic, originations following mass extinction events, those during the Carboniferous might influence the relationship between latitudinal ranges, geological durations, and ammonitella diameters. This is possibly seen in Fig. 4F where the larger indices tend to shift to the middle of the observed range of ammonitella diameters.

A more complete and detailed dataset of geographical range and geological duration with larger data size, together with the paleobiogeography of each period and morphological changes at the embryonic and post-embryonic stage, would allow us to analyze with more elaborate approaches and therefore to better understand the geological transition of ammonoid reproductive strategies during the Devonian–Cretaceous with finer geological resolution (i.e., at the stage level) and on a lower taxonomic level (i.e., at the species level).