The hook-like shape of heteromorph ammonites has puzzled scientists for years. In such ammonites, including those in the family Scaphitidae (called scaphites for short), the body chamber consists of a shaft, beginning near the last septum and a hook terminating at the aperture (Fig. 1). The point at which the hook curves backward is called the point of recurvature. In primitive members of the Scaphitidae, such as Scaphites, the hook is well developed with a large gap between the phragmocone and body chamber (Fig. 2A, B). In more derived members of this group, for example, Hoploscaphites, the hook is reduced, but a small gap is still present between the phragmocone and body chamber (Landman et al., 2010: fig. 49). Recently, Arkhipkin (2014) published a study arguing that the hook-like shape of such ammonites was an adaptation for stationary attachment to the branches of kelp forests on the sea floor. He also included in his argument members of the Ancyloceratidae, including large specimens of Ancyloceras almost 1 m in length.
Although this ‘stationary hooked-adult’ hypothesis is provocative, it is highly speculative. We discuss five main pieces of evidence in its support, none of which bears scrutiny. We expand on each of these points below, introducing new data based on additional specimens. These specimens are deposited in the American Museum of Natural History, New York (AMNH), Black Hills Institute, Hill City, South Dakota (BHI), and the US National Museum of Natural History, Washington, D.C. (USNM).
Arkhipkin (2014: 355, fig. 2) reported flattening of ribs, “suggesting wear” on the inside (dorsal) surface of the hook in several species of heteromorph ammonites. He argued that worn ornament was produced as a result of the shell rubbing against the kelp branch. However, most of the specimens he examined are steinkerns with little or no original shell wall still attached, making it impossible to detect the existence of wear. In contrast, in specimens in which the original shell wall is still preserved, the inside surface of the hook is pristine, without any scratch marks. The outer wall in this area of the shell consists of an outer prismatic and inner nacreous layer. An examination of the dorsal surface of a well-preserved specimen of S. carlilensis reveals a smooth surface marked with growth lines (Fig. 2C). The surface is not worn or abraded in any way. Irregularities are present on the surface and consist of kinks in the growth lines. However, these are present everywhere on the shell, including the ventral surface of the body chamber, and have been interpreted as growth irregularities that formed during the normal course of shell secretion (Landman & Lane, 1997). Similarly, in the ancyloceratid Didymoceras cheyennense, which also possesses a hook-like body chamber (Fig. 2D, E), the dorsal surface of the body chamber reveals no traces of wear (Fig. 2F).
Arkhipkin (2014: fig. 1) reported that the hook-like body chamber in scaphites is asymmetric so that the gap is larger on one side than the other, for example, the left side vs the right side. He argued that if the animal were attached to a branch of kelp that was angled upward, the formation of the hook would be similarly affected, so that it would mirror the orientation of the branch. He determined the degree of asymmetry by comparing published photos of the left and right sides of the same specimen and tracing the outline of the gap (what he called the foramen). However, the appearance of the gap in such photos is very dependent on the exact position of the shell when it was photographed and the shadows that were created.
A better sense of the shape of the gap is provided by dorsoventral cross-sections through the shell at the point of recurvature. Such cross-sections in H. nodosus and H. brevis from the Upper Cretaceous Western Interior of North America reveal that the gap is perfectly symmetrical from one side of the shell to the other (Fig. 3). To provide more quantitative data, we measured the whorl height between the dorsum and venter of the hook at the point of recurvature on the left and right sides of 20 specimens of these species. The data reveal hardly any disparity between the left and right sides, not nearly enough to require a special explanation, other than the asymmetry due to normal growth irregularities (Fig. 4).
Fossils of kelp-like brown macroalgae are absent in any of the areas in which heteromorph ammonites lived during the Late Cretaceous. Even though the preservation potential of such noncalcified material is admittedly low, one might expect to find at least some evidence of epizoans that encrusted these branches. Such evidence does not exist. Arkhipkin (2014) noted that molecular studies of kelp suggest that the radiation of this group occurred in the Early Cretaceous, but the absence of any kelp-like fossils associated with heteromorph ammonites is a serious handicap to the stationary hooked-adult hypothesis.
In many of these ammonites, notably the Scaphitidae, two morphs (dimorphs) are present, known as macroconchs and microconchs. They are interpreted as sexual in nature, the macroconch being the female and the microconch being the male, following the traditional view as outlined by Lehmann (1981) and Davis et al. (1996). Macroconchs and microconchs of the same species, e.g. S. whitfieldi, are equally uncoiled, with the implication according to Arkhipkin (2014: 8) that both morphs would have been permanently attached to kelp at maturity. All living cephalopods engage in reproductive behaviour including copulation and spawning, so extinct species are likely to have done so also. If these two scaphite morphs represent sexual dimorphs, such permanent attachment to kelp branches would have posed considerable challenges to copulation.
Arkhipkin (2014: 9) speculated that the female ammonite brooded her eggs in the shell while she was attached to the kelp. Indeed, recent finds of embryonic shells in the body chambers of Early Cretaceous ammonites suggest that brooding may have occurred in some forms (Mironenko & Rogov, 2015). Thus, unlike the hypothesis of kelp attachment, which we completely reject, the hypothesis of brooding behaviour in heteromorph ammonites is plausible, although not much evidence is available to test it (DeBaets, Landman & Tanabe, 2015). The most widely cited piece of evidence (cited, e.g. by Lewy, 2002) is a study by Landman (1985) documenting the presence of embryonic and young postembryonic ammonite shells in the body chambers of 15 mature specimens of S. ferronensis from the Upper Cretaceous Mancos Shale of Colorado. However, these embryonic shells also occur in the surrounding matrix and may have been even more abundant on the outcrop. They are frequently broken and are intermixed with clams, fish scales and larger juveniles. A CT-scan of one of the adult body chambers reveals the presence of small clams inside it, in addition to ammonites, suggesting that this fossil assemblage resulted from post-mortem transport and does not constitute evidence of brooding (Fig. 5). Similar associations of embryonic ammonite shells with other fossil debris have also been reported elsewhere in the stratigraphic record (Wetzel, 1959; Landman, 1982; Mapes & Nutzel, 2008).
In summary, the stationary hooked-adult hypothesis is not well supported by observations. In marine animals that occasionally attach to objects on the sea floor, the structure of the animals is usually flexible enough to permit bending and twisting while grasping. For example, the seahorse attaches to mangrove roots and sea grasses on the sea floor. However, the design of its prehensile tail is completely different from the rigid shell of a scaphite. It is composed of a series of articulated plates with specialized joints that allow the animal to grasp and hold onto an object (Porter et al., 2015).
The hook-like body chamber of scaphites and similar ammonites undoubtedly imposed constraints on locomotion. It probably limited these animals to only backward or downward swimming. In many heteromorphs in which the hook is well developed, the body chamber culminates in an upturned aperture. This may have been an adaptation to feeding on small organisms in the water column, which is consistent with the shape of their jaws and radula (Kruta, Landman & Tanabe, 2015; Tanabe, Kruta & Landman, 2015). These ammonites may have lived near the bottom, occasionally swimming to remain in place or evade predators, with limited vertical and lateral migration (Landman et al., 2010; Landman, cobban &Larson, 2012). Indeed, recent studies of the oxygen isotopic compositions of scaphite shells reveals that they are very similar to those of bivalves, suggesting that these ammonites lived close to the bottom (Sessa et al., 2015). As such, they may have acted as floating filter feeders and occupied a different niche from that of their fast-swimming coleoid cousins. But, to sum up, we find no evidence in support of the stationary hooked-adult hypothesis, as proposed by Arkhipkin (2014).
We thank Neal L. Larson (Larson Paleontology Unlimited, Keystone, South Dakota) and Kazushige Tanabe (University of Tokyo) for stimulating discussions over the years about the mode of life of heteromorph ammonites, Jamie Brezina (Rapid City, South Dakota) for the generous donation of the specimen of Didymoceras cheyennense, Steve Thurston (AMNH) for help in preparing the figures, Henry Towbin (AMNH) for assistance with scanning electron microscopy and K.C. McKinney (USGS, Denver, Colorado) for facilitating the loan of USGS material in his care. We also thank the three reviewers of this manuscript for their insightful comments, and Editor David Reid for his encouragement and help.