Interpretations of the functional morphology of fossil cephalopods have tended to concentrate on those that resemble extant cephalopods (e.g., belemnites compared to squid, Monks et al. 1996; regularly coiled ammonites to Nautilus, Chamberlain 1976, 1980). Fossil cephalopods that do not have such clear analogues, such as the Lower Cretaceous heteromorph ammonites, are more problematical. Moreover, it is unclear how far Nautilus is a valid analogue for ammonites.

Nautilus has paired retractor muscles that can pull the head and arms into the shell when the animal is threatened. The aperture is also sealed by a tough hood, functionally similar to the gastropod operculum (Wells et al. 1992). However, the Nautilus body chamber is short, approximately one third of a whorl in length, and the animal cannot withdraw far. In contrast, the living chambers of ammonites are very long and often narrow. Moreover, ammonites lacked a hood or operculum. [Note: the aptychi are now generally accepted to have been part of the jaws (Lehmann 1981)]. Consequently, the assumption that the soft parts of the ammonite filled the entire living chamber of the shell is not neccessarily correct, and it is possible that without a way of sealing the aperture, ammonites may have been capable of withdrawing further into the shell than has heretofore been acknowledged.

Body position in heteromorph ammonites is important because the association of the soft body (e.g., head, arms, viscera) with the shell is a key part of understanding how these animals lived and functioned. The discussion by Trueman (1941) on this topic is widely quoted (e.g., in the Treatise on Invertebrate Paleontology, Arkell et al. 1957). Trueman assumed that the entire living chamber was filled, and used this to deduce the centres of mass and buoyancy of the ammonite. Ammonite functional interpretations based on this model are characterised by a single stable orientation, with the centre of mass vertically below the centre of buoyancy. For heteromorph ammonites this model typically predicts upturned apertures (Anim. 1). Many workers have cited these upturned apertures as evidence that heteromorphs were pelagic plankton feeders (e.g., Klinger 1980, Westermann 1996).

More recently, Kakabadzé and Sharikadzé (1993) proposed that the heteromorph shell had two stable orientations, one with the aperture pointing upwards (as in previous interpretations) and a second with the aperture tilted towards the sea floor. This alternative orientation would make it possible for heteromorph ammonites to feed off the sea floor. Kakabadzé and Sharikadzé (1993) also proposed that these ammonites might have been able to adjust their bouyancy by shifting cameral fluid within the shell. This would change the position of the centre of mass, and consequently the orientation of the shell (Anim. 2).

The idea that ammonites may have had more than one stable orientation has not been widely discussed by ammonite workers, but if true could force a re-appraisal of ammonite paleoecology. The Kakabadzé and Sharikadzé (1993) model relies on ammonites being able to rapidly change the distribution or amount of fluid within the shell. As far as is known, movement of fluid within the shell of ectocochleates is slow. In Nautilus, changes in the amount of fluid within the chambers is too slow even to assist diurnal migration by adjusting overall buoyancy (Chamberlain 1991). Cuttlefish have a highly modified and characteristic shell, divided into numerous small chambers, subdivided by conchiolin walls, and exhibiting a broad siphuncular region rather than a narrow siphuncle (Denton & Gilpin-Brown 1961a). This appears to allow the cuttlefish to make changes to the buoyancy of the shell more quickly than Nautilus (Denton & Gilpin-Brown 1961b). Even so, buoyancy regulation in cuttlefish seems not to be used as the primary device for rising or sinking in the water column, but rather for acquiring neutral buoyancy once a preferred depth has been reached (Denton 1973). Ammonite shells are quite unlike the shells of cuttlefish and so it is more likely that Nautilus is an appropriate analogue in this respect. As a result, the problem remains that while changes in orientation may have been possible, no entirely convincing mechanism has been proposed for how this might have been acomplished. We propose a third possible anatomy: That the heteromorph ammonite animal was a relatively small, mobile creature that was able to move within its floating shell rather like a modern gastropod. As with the previous model this allows significant changes in orientation, due in this case to changes in the distribution of the mass as the body moves within the shell (Anim. 3).

The only direct evidence for soft part morphology in ammonites is provided by the muscle scars. These are consistently located near the rear of the living chamber, close to the final septum (Crick 1898). Such a muscular arragement is compatible with our model of a small retractable body although clearly it does not prove it. Heteromorph ammonite shells are characterized by a departure from the regular planispiral shape of most other ammonites. One of the most common themes in heteromorph shell design is that of the planispiral phragmocone and hook-shaped living chamber. The simplest form is that of the aspinocone, where the living chamber consists of a shallow U-shaped hook held slightly away from the planispiral phragmocone (Fig. 1). By drawing the hook deeper into two parallel shafts, the ancyclocone is produced (Fig. 2). Finally, straightening all but the earliest stages of the phragmocone and tightening the bends further results in the hamiticone (Fig. 3).