Discussion

These specimens confirm previous findings on the nature of the ammonitella of early Devonian ammonoids (Klofak et al., 1999). The end of the ammonitella is marked by a reduction in both the spacing and size of the ornament, until a relatively smooth surface appears. A constriction is present with no accompanying

-i—i—i—i—i—i—i—i—i—i—i—r-10 20 30 40 50 60

Lirae Space

Fig. 2.11 (4, C) Agoniatites vanuxemi (NYSM3545, Devonian, New York State, USA) (A) Lirae spacing on embryonic shell for ventral, midflank, and dorsal positions. Symbols given in graph. X axis is the number of lirae space. Y axis is the measured distance between two lirae given in microns (mm). (C) Ratio ofthe ventral lirae space/dorsal lirae space (V/D) on the embryonic shell. Data are given in Appendix 10. (B, D) Fidelites fidelis (AMNH 50417, Devonian, Morocco). (B) Lirae spacing on embryonic shell for ventral, midflank, and dorsal positions. Symbols given in graph. X axis is the number of lirae space. Y axis is the measured distance between two lirae given in microns (mm). (D) Ratio of the ventral lirae space/dorsal lirae space (V/D) on the embryonic shell. Data are given in Appendix 11.

-i—i—i—i—i—i—i—i—i—i—i—r-10 20 30 40 50 60

Lirae Space

210 190 170 150

O 110

VW'

0 10 20 30 40 Lirae Space

-i—i—i—i—i—i—i—i— 10 20 30 40 Lirae Space

Fig. 2.11 (4, C) Agoniatites vanuxemi (NYSM3545, Devonian, New York State, USA) (A) Lirae spacing on embryonic shell for ventral, midflank, and dorsal positions. Symbols given in graph. X axis is the number of lirae space. Y axis is the measured distance between two lirae given in microns (mm). (C) Ratio ofthe ventral lirae space/dorsal lirae space (V/D) on the embryonic shell. Data are given in Appendix 10. (B, D) Fidelites fidelis (AMNH 50417, Devonian, Morocco). (B) Lirae spacing on embryonic shell for ventral, midflank, and dorsal positions. Symbols given in graph. X axis is the number of lirae space. Y axis is the measured distance between two lirae given in microns (mm). (D) Ratio of the ventral lirae space/dorsal lirae space (V/D) on the embryonic shell. Data are given in Appendix 11.

varix. Previously, the absence of a primary varix was documented for the genus Archanarcestes in the family Mimagoniatitidae (see Klofak et al., 1999). A second genus in the Mimagoniatitidae demonstrates the same type of ammonitella aperture, that is, a constriction with no varix. Finally, we have also extended this apertural type to a second family, the Anarcestidae. It is known that a varix is present in the Tornoceratina (House, 1965), a phylogenetically more advanced taxon (Becker and Kullmann, 1996; Klug and Korn, 2002). To date, no varix has been found in phylo-genetically more primitive ammonoids, increasing the likelihood that this may be the ancestral state for ammonoids. Bactritids also do not possess a varix, but the microstructure of their embryonic shells is different from that of these ammonitellas and may represent a different mode of formation. Bactritids are different enough that some workers have separated them from the Ammonoidea (Ruzhentshev, 1960, 1974; Erben, 1964a; Mapes, 1979; Doguzhaeva, 2002; Korn and Klug, 2002).

The postembryonic shell can be seen emerging from beneath the ammonitella edge of several specimens (AMNH 46646, Fig. 2.3f; AMNH 46645, Fig. 2.2c, d). In the specimen of Latanarcestes, the area just anterior to the ammonitella edge shows several breaks, with new shell emerging from below the previous surface (Fig. 2.3f). This is not dissimilar to the end of the embryonic shell in modern Nautilus where the thin apertural edge is prone to breakage at hatching (Arnold et al., 1987). It also supports an idea proposed by Kulicki (1974) that the primary varix in post-Devonian ammonoids evolved to prevent breakage during hatching.

It is also clear from the data collected, that while marking the end of a phase in embryonic development, the end of the initial chamber does not exhibit any discernable breaks in shell production (Fig. 2.3c). There is an obvious change in the shape of the shell (see "1. Wachstums-Anderung" of Erben, 1962, 1964b) and a reduction in the distance between lirae in all three taxa at the end of the intial chamber. In addition, Erben (1964b) noted that the ventral sinus appears at this point, perhaps marking the first appearance of the hyponome. This was part of his reason for interpreting this point as the end of embryonic development, though this interpretation has been disproved. Studies on Nautilus have shown that a functional hyponome is present early in embryonic development and its appearance is not correlated with hatching (Tanabe et al., 1991).

The original intent of this project was to discover if there were taxonomic differences in the distance between lirae in different taxonomic groups. It became apparent that there was enough variation across an individual ammonitella that any individual measurement or a mean measurement was not informative (Table 2.1). However, the pattern itself proved to be useful.

Based on the shell geometry of a coiled tube with radial ornament, one would expect the distance between any two lirae to be widest on the venter, narrowest on the dorsum, and somewhere in between on the midflank. Projecting a three dimensional object onto two dimensions may distort some measurements. However, the venter and dorsum are not affected because they are measured in the same plane. Distortion could affect the midflank position on the initial chamber but this is minimized because the midflank measurements always lie between the venter and dorsum. Therefore, the patterns in our data are real and not artifacts of the methodology.

There are differences in the distances between lirae on the initial chamber among the different ammonoid families. In the Mimagoniatitidae, the ventral and dorsal parts of the ammonitella show the same pattern (Figs. 2.6a, 2.7a, 2.8a). This is not the case for the Anarcestidae where, as the ventral distance between lirae increases across the initial chamber, the dorsal distance between lirae decreases (Figs. 2.9a, 2.10a). This is also true for the Agoniatitidae (Figs. 2.11a, b). The distance between lirae on the initial chamber of the Agoniatitidae also shows a rapid increase before the decrease marking the end of the initial chamber. These changes reflect subtle changes in the shape and symmetry of the nearly spherical initial chamber. In the Mimagoniatitidae, the arc of both the dorsum and venter are symmetrical, whereas in the Anarcestidae, the arc of the venter is circular but the dorsum is flattened.

With these data, it might be possible to use the change in the distance between lirae as an indicator of the end of the initial chamber. Currently, the initial chamber is defined based on the appearance of the first septum (proseptum) (Owen, 1878; Branco, 1879-1880; Hyatt, 1883; Schindewolf, 1933; Erben, 1960; Bandel, 1982; Landman et al., 1996). It makes better biological sense to define a developmental stage based on characters that formed or are present at that stage (change in distance between lirae) rather than ones added at a later stage (first septum). In other words, the end of the initial chamber would be the point where the distance between lirae shows a marked decrease.

The actual distance between lirae seems to vary widely across the ammonitella. The only constraint on lirae formation on the ammonitella is that the prescribed number of lirae is present by the end of embryonic development. This number is distinctive and taxonomically significant at a family level: 70-77 are present in the Mimagoniatitidae, 95 in the Anarcestidae, and an estimated 65 for the Agoniatitidae. In the Mimagoniatitidae and the Anarcestidae, the difference in number can be directly attributed to the different number of lirae formed on the initial chamber. There are 25-30 in the Mimagoniatitidae and 35-40 in the Anarcestidae, but approximately the same number occur on the ammonitella coil. For the Agoniatitidae, there seems to be a reduction in the number of lirae on the ammonitella coil. There are 30-35 lirae on the initial chamber, more than that in the Mimagoniatitidae, but only a total of 65 for the ammonitella, less than that in the Mimagoniatitidae. Both the Mimagoniatitidae and Anarcestidae have umbilical perforations; the Agoniatitidae do not. It is likely that the tighter coil of the agoniatitid ammonitellas led to a reduction in the size of the ammonitella coil and, hence, a reduction in the number of lirae. The initial chamber, however, remains large. Not only are the agoniatid ammonitellas more tightly coiled, but also the nature of the coiling appears different. The end of the initial chamber can only be defined based on the fold of the dorsal part of the initial chamber against the dorsal part of the ammo-nitella coil. The venter shows only a slight flattening, similar to the description given for some ammonitellas of Mesozoic ammonoids (Bandel, 1986; Landman et al., 1996). This is visible in the specimen of Fidelites fidelis (Fig. 2.1e).

A comparison of the V/D ratio of the embryonic shell and the beginning of the postembryonic shell in the Mimagoniatitidae and Anarcestidae shows the relationship between the dorsal and ventral parts of the shell. The V/D ratios for the postembry-onic shells show that there is a closer correlation between the placement of the lirae on the dorsal and ventral sides of the postembryonic shell, than in the embryonic stage. This pattern is especially clear in Fig. 2.6c, d.

It is assumed that the embryo inside the egg case would have produced a shell in a constant environment. Upon hatching, the growth of the animal and production of shell would have been affected by many variables in the environment, such as temperature, salinity, food supply, predators, and water currents. For example, in Fig. 2.2f, less than a whorl after hatching, a small repaired break can be observed on the shell. Rather than be constrained by a more constant environment, the variability of the embryonic shell formation speaks of a high degree of developmental plasticity. This variability has already been documented by Erben (1950) where he observed several coiling shapes in Mimagoniatites fecundus. Such developmental plasticity likely played a role in the rapid radiation of the early ammonoids.

Several models for the development of the embryonic shell of ammonoids have been proposed (see Landman et al., 1996). These models feature opposing views. Either the ammonitella was calcified in its entirety (Bandel, 1982; Tanabe, 1989) or it was accreted as small increments at the aperture as in the adult shell (Erben, 1964b, 1966; Erben et al., 1968; Tanabe, 1989). The data presented here do not suggest a good fit with either of these views. The presence of transverse lirae has been used as evidence for an accretionary mode of formation (Erben, 1964b, 1966; Tanabe, 1989), with growth occurring at the apertural edge in small increments followed by hiatuses, producing growth lines (Bucher et al., 1996). The transverse lirae are not growth lines (Bucher et al., 2003) and no growth lines appear between lirae on the embryonic shell. The marginal accretionary mode of formation would likely result in a regular spacing as the mantle edge moved forward in small increments. The distance between any two consecutive lirae on the venter should be proportional to the distance between the same lirae on the dorsum. This is not what is observed for these primitive ammonoids. The V/D ratio shows wide fluctuations between consecutive lirae (Fig. 2.6c, d).

Other models for ammonitella formation suggest a nonaccretionary mode of formation (Bandel, 1982; Tanabe, 1989). The presence of the "wrinkle-like" creases between the transverse lirae suggest an organic component. The production of an organic nonmineralized shell eliminates the need for the lirae to be produced in an even fashion and the V/D ratio could then vary. This might also explain the differences in shape observed by Erben (1950) in the coiling of the ammonitella of Mimagoniatites fecundus. There is also no need for the entire nonmineralized shell to calcify simultaneously. Stepwise growth at the aperture has been observed in molluscs during the formation of large scale ornamental features (Vermeij, 1993). And similar models have been proposed for ammonoids (Checa, 1995; Bucher et al., 2003). While neither of these models produces ornament like that on these embryonic shells, it does suggest that ammonoids have the ability to produce the shell wall in a variety of ways, including growing short segments of organic material, and then calcifying it. This would also account for the fluctuations in the spacing observed if the segments produced were from lira to lira.

The models proposed for ammonitella formation are based, at least in part, on the embryonic development of extant taxonomic groups. Nonaccretionary models rely heavily on archaeogastropod developmental models (Bandel, 1982) and accretionary models on Nautilus development (Arnold et al., 1987). The embryonic shell of Nautilus grows by forward accretion at the aperture (Arnold et al., 1987). It produces a reticulate ornament with both longitudinal and transverse elements strongly developed. A similar ornament has been described in some Jurassic nautilids (Chirat and von Boletzky, 2003). This ornament is different from that described for Devonian ammonoids. The transverse elements in the nautilids appear as if imbricated and the longitudinal elements are not the "wrinkle-like" creases described here, but rather strongly developed longitudinal ridges that intersect the transverse lines. While examining related modern taxa is useful, the phylogenetic distance of these taxa must also be considered (Jacobs and Landman, 1993). The Nautiloidea lie some distance away from the Ammonoidea (Berthold and Engeser, 1987; Engeser, 1990, 1996). Furthermore, between the nautiloids and the three families in this study lie several taxonomic groups (Orthocerida, Bactritida, and several more basal ammonoid families, see Becker and Kullmann, 1996; Korn and Klug, 2002) several of which have distinctly different embryonic shells [see Doguzhaeva (1996a, b, 2002) for the bactritids and Ristedt (1968) for the orthocerids]. While data do exist for more basal ammonoid taxa (Sandberger, 1851; Schindewolf, 1933; Erben, 1950, 1960, 1962, 1964b, 1965; Chlupác and Turek, 1983), they are scarce and the material is not as well preserved. Due to the open nature of their coiling, the inner whorls are often lost, and when they are preserved, they are often steinkerns. The external morphology of the shell is necessary for comparison. What is known is that members of one clade (the Agoniatitida) show remarkable variation at a very early developmental stage. Such developmental plasticity very early in ontogeny may have aided the rapid radiation of the ammonoids in the Devonian.

Was this article helpful?

0 0

Post a comment