Abstract
Trace fossils are abundant and diverse in many clastic sequences spanning the Precambrian-Cambrian boundary and may prove to be the most useful palaeontological method for global correlation in this stratigraphic interval. The ichnofaunas of the latest Precambrian (Vendian) rocks include some forms whose range does not extend into the Cambrian (e.g. Bilinichnus, Intrites, Palaeopascichnus, Vendichnus, Vimenites) and others which continue throughout most or all of the Phanerozoic (e.g. Arenicolites, Aulichnites, Cochlichnus, Didymaulichnus, Gordia, Neonereites, Planolites, Skolithos). At least 50 ichnogenera make their first appearance below the lowest trilobites in sections with broad geographic spread. A few of these appear to have a short time range, extending to about the incoming of the trilobites (e.g. Astropolichnus, Didymaulichnus miettensis, Plagiogmus, Taphrhelminthopsis circularis), but the majority continue through most or all of the Phanerozoic.

For correlation of Precambrian-Cambrian boundary sequences it is therefore possible to use both the occurrence of those ichnogenera with a short time range and the incoming of those with an extended range. Three stratigraphical zones can be recognized with respect to the incoming of trace fossils. Zone I is of Upper Vendian age and includes Arenicolites, Bilinichnus, Cochlichnus, Didymaulichnus, Gordia, Harlaniella, Intrites, Nenoxites, Neonereites, Palaeopascichnus, Skolithos, Vendichnus and Vimenites. In Zone II, of Lower Tommotian age, the earliest examples of Bergaueria, Phycodes, Teichichnus and Treptichnus are encountered. Many trace fossils appear in Zone III, which extends from Upper Tommotian to Lower Atdabanian, but the most important are: Astropolichnus, Cruziana, Diplichnites, Diplocraterion, Dimorphichnus, Plagiogmus, Rusophycus and Taphrhelminthopsis circularis.

This vertical zonation of trace fossils allows an attempt at world-wide correlation, from which the most significant conclusions are that the Vendian/Tommotian boundary can probably be placed: (i) near the middle of the McNaughton Formation in the Rocky Mountains, Canada; (ii) at the base of the Deep Spring Formation or in the underlying Reed Dolomite in the White Inyo Mountains, California, U.S.A.; (iii) low in the Chapel Island Formation in the Burin Peninsula, Newfoundland, Canada; (iv) at or close to the base of the Candana Quartzite in North Spain; (v) at or below the base of the Breivik Member in Finnmark, Norway; and (vi) near or below the base of the Zhongyicun Member at Meischucun, China.

The sections in the Burin Peninsula, Newfoundland and Meischucun, China are favoured candidates for the global stratotype for the Precambrian-Cambrian boundary. In the Burin Peninsula, the trace fossils suggest that the Tommotian/Atdabanian boundary may be within or at the base of the Random Formation, thereby implying that the Tommotian may include a thickness of 500 m of sediment comprising at least most of the Chapel Island Formation. At Meishucun, the ichnofaunal evidence implies that the Tommotian/Atdabanian boundary is probably no higher than the top of the Zhongyicun Member. The thickness of the Tommotian is therefore possibly only about 20 m here, implying a very condensed sequence, a conclusion consistent with an abundance of phosphorites. Two stratotype reference points for the Precambrian-Cambrian boundary have been suggested in this section. The lower point (0.8 m above the base of the Xiawaitoushan Member) may be near the Vendian/Tommotian boundary or younger, while the higher point (base of Unit 7 of the Zhongyicun Member) is probably Upper Tommotian or even Lower Atdabanian. The higher point would place the boundary above the world-wide dramatic increase in trace fossil abundance and diversity but probably before the first trilobites. This would almost certainly have advantages for correlation. The inference that the Meishucun section is younger than most Chinese work suggests should not therefore, by itself, prejudice its adoption as global stratotype.

In general, where comparative data are available, the trace fossil correlations agree well with pre-existing proposals based on small shelly fossils. The degree of resolution of the two methods would appear at present to be similar but trace fossils, being found mainly in clastic facies, may benefit from more frequent occurrence.