DISCUSSION

Discussions of the behavior of trace fossils have generally hypothesized that movement patterns are "hard-wired," e.g., they result from some innate set of behaviors. For example, Raup and Seilacher (1969) hypothesized that spiral and meandering trails resulted from a combination of phobotaxis (avoidance of crossing a previous trail), thigmotaxis (staying close to older parts of the trail) and strophotaxis (periodic sharp turns). These movement patterns may instead result from simple chemotaxis interacting with heterogeneities of resource distribution and encounters with patch boundaries (Kitchell 1979; Koy and Plotnick 2007). If correct, phobotaxis, strophotaxis, and thigmotaxis are thus consequences, and not the causes, of the observed movements.

The results for inter-patch behavior discussed here also support the importance of resource distributions as a major control of movement and thus of trace morphology. For example, Figure 9 illustrates a portion of the movement trail of a simulated organism approaching an isolated patch within which the distribution of resources is fractally distributed (Plotnick and Gardner 2003; Koy and Plotnick 200 7). The relatively simple trace geometry outside of the patch becomes far more complicated within the patch.

The simulations also suggest that when resources are continuously distributed, only short-range detection (contact chemoreception) is necessary for efficient foraging in heterogeneous environments. As resource distributions become increasingly patchy, the advantage of possessing the ability to detect these resources at a distance (distant chemoreception) becomes greater. As a result, there may be morphological and anatomical differences between organisms living in patchy vs. non-patchy settings.

These inferences have potential implications for our understanding of some of the biotic changes occurring during the Ediacaran-early Phanerozoic. During this interval there was a major increase in trace fossil diversity and complexity (Crimes and Fedonkin 1994; Jensen 2003; Droser et al. 2005). Re-analyses of taxa used in these analyses have indicated that many suggested Ediacaran trace fossils, in particular ones purportedly showing complex movement traces, are probably not traces (Droser et al. 2005). Nearly all traces currently accepted as valid are small and simple meandering horizontal forms and are presumed to have formed at or near the sediment water interface within or just below microbial mats (Jensen 2003).

Gehling et al. (2005) discussed a conceptual model for feeding on biomats by the larger possible bilaterian Dickinsonia, based on preserved traces, which suggest movement to immediately adjacent areas (Fedonkin 2003). In their scenario, Dickinsonia absorbed nutrients over its ventral surface. When an area was fully exploited, the organism shifted to exploit the neighboring area. This behavior would have required contact chemoreception only.

In addition, conceptual models and some data also suggest a general increase in mobile bilaterian body size over this interval (Valentine 2002; Novack-Gottshall 2005). Typical body widths suggested by Ediacaran trace fossils are on the order of millimeters (Jensen 2003); Cambrian bilaterians are clearly much larger. An increase in body size, as well as shift from grazing under biomats to foraging over the substrate, would have changed the relevant fluid mechanical environment from one dominated by diffusion to one dominated by convection.

Another change is in the nature of sense organs. As most recently pointed by Marshall (2006), Ediacaran body fossils show a noticeable lack of macroscopic sensory organs of any kind. This is in contrast with the presence of organs such as eyes and antennae in Cambrian animals, in particular the early arthropods.

Finally, there are possible changes in substrate heterogeneity associated with the "agronomic revolution" or "substrate revolution" concept originally proposed by Seilacher and Pflüger (1994), Seilacher (1999), and expanded by Bottjer et al. (2000). In this concept, matground environments offered relatively low spatial patchiness at scales relevant to the earliest benthic bilaterians. With resources more-or-less continuously distributed, contact chemoreception and accompanying relatively simple movement patterns were appropriate for effective resource exploitation.

The advent of burrowing greatly increased the spatial complexity of the sea floor (Meysman et al. 2006). The previously extensive subtidal mats became disrupted by bioturbation, scarcer, and perhaps much patchier. Modern supratidal mats described by Hagadorn and Bottjer (1999) are patchy. Resource patchiness was further increased by the advent of macroscopic bilaterians. These would produce spatially discrete carcasses and fecal pellets (McIlroy and Logan 1999). The advent of higher trophic levels likely enhanced spatial resource heterogeneities by producing patches of higher quality food (Bengston 2002). Taken together, this suggests that the spatial heterogeneity of biomass distribution in the benthic environment rapidly increased in the early Paleozoic. Resource distributions became increasingly patchy with an increase in the range of resource concentrations.

As spatial heterogeneity increased, there should have been concomitant changes in the foraging responses of the animals. These changes would be a direct consequence of the increased patchiness and, most probably, of increased rates of predation (Bengston 2002; Dzik 2005). With the increase in patchiness and heterogeneity, the complexity of movement patterns between and within resource patches increased, and thus trace fossil diversity increased. The Cambrian growth of trace fossil diversity, therefore, may be part of a larger cascade within the evolving biosphere (McIlroy and Logan 1999; Bengston 2002; Marshall 2006).

In sum, the critical innovations that would have led to optimal foraging by early mobile marine organisms in an environment of increasing patchiness are directly related to an organism's abilities to obtain, process, and retain information about the spatial properties of its environment, similar to a conclusion reached by Hammer (1998). Selection would have favored the evolution of mechanisms for obtaining the location and richness of resources both close by and at a distance and for the assessment of predation risk. Such mechanisms include chemosensory organs such as the antennae of trilobites and other arthropods and the osphradia of molluscs, and the apparently polyphyletic evolution of complex eyes (Fernald 2004; Nilsson 2005). These ideas are admittedly speculative and await further analysis and testing.