th the exception of some very simple vertical burrows the only Ediacaran burrows are horizontal, lying on or just below the surface of the seafloor. Such burrows have been taken to imply the presence of motile organisms with heads, which would probably have had a bilateral symmetry. This could place them in the bilateral clade of animals but they could also have been made by simpler organisms feeding as they slowly rolled along the sea floor. Putative "burrows" dating as far back as 1,100 million years may have been made by animals that fed on the undersides of microbial mats, which would have shielded them from a chemically unpleasant ocean; however their uneven width and tapering ends make a biological origin so difficult to defend that even the original proponent no longer believes they are authentic. The burrows observed imply simple behaviour, and the complex efficient feeding traces common from the start of the Cambrian are abs ent. Some Ediacaran fossils, especially discs, have been interpreted tentatively as trace fossils but this hypothesis has not gained widespread acceptance. As well as burrows, some trace fossils have been found directly associated with an Ediacaran fossil. Yorgia and Dickinsonia are often found at the end of long pathways of trace fossils matching their shape; these fossils are thought to be associated with ciliary feeding but the precise method of formation of these disconnected and overlapping fossils largely remains a mystery. The potential mollusc Kimberella is associated with scratch marks, perhaps formed by a radula. Classification and interpretation Classification of the Ediacarans is difficult, and hence a variety of theories exist as to their placement on the tree of life. Martin Glaessner proposed in The Dawn of Animal Life (1984) that the Ediacara
