and Early Jurassic sediments, and patterns of the Triassic-Jurassic PAUL E. OLSEN AND tetrapod transition HANS-DIETER SUES Introduction parent answer was that the supposed mass extinc- The Late Triassic-Early Jurassic boundary is fre- tions in the tetrapod record were largely an artifact quently cited as one of the thirteen or so episodes of incorrect or questionable biostratigraphic corre- of major extinctions that punctuate Phanerozoic his- lations. On reexamining the problem, we have come tory (Colbert 1958; Newell 1967; Hallam 1981; Raup to realize that the kinds of patterns revealed by look- and Sepkoski 1982, 1984). These times of apparent ing at the change in taxonomic composition through decimation stand out as one class of the great events time also profoundly depend on the taxonomic levels in the history of life. and the sampling intervals examined. We address Renewed interest in the pattern of mass ex- those problems in this chapter. We have now found tinctions through time has stimulated novel and com- that there does indeed appear to be some sort of prehensive attempts to relate these patterns to other extinction event, but it cannot be examined at the terrestrial and extraterrestrial phenomena (see usual coarse levels of resolution. It requires new fine- Chapter 24). The Triassic-Jurassic boundary takes scaled documentation of specific faunal and floral on special significance in this light. First, the faunal transitions. transitions have been cited as even greater in mag- Stratigraphic correlation of geographically dis- nitude than those of the Cretaceous or the Permian junct rocks and assemblages predetermines our per- (Colbert 1958; Hallam 1981; see also Chapter 24). ception of patterns of diversity, extinctions, and Second, like the Cretaceous-Tertiary boundary, the originations. This poses an especially difficult prob- Triassic-Jurassic boundary heralded a new, long- lem for the Early Mesozoic because there are vir- lasting regime of dominant animals, the dinosaurs. tually no unquestioned Early Jurassic continental Third, but unlike the Cretaceous, a definite bolide vertebrate assemblages. Correlations are of such impact structure is known in the Late Triassic. The paramount importance to any study of change during 70 km Manicouagan crater in Quebec, Canada has the Early Mesozoic that we devote the first part of been dated at 210 Â4 MY (Grieve 1982), which is this chapter to a summary description of the ratio- within the margin of error of the currently accepted nale for correlating various continental sequences dates for the Triassic-Jurassic boundary. Despite with those of the Late Triassic and Early Jurassic considerable uncertainty, scenarios of asteroid im- type areas of Europe. pact have already been proposed to explain the The second part of this chapter details the tax- Triassic-Jurassic extinctions, much as they have onomic changes through the Early Mesozoic based been for the Cretaceous-Tertiary extinctions (Raup on these correlations. We examine the skeletal rec- and Sepkoski 1984; Rampino and Stothers 1984); ord of continental tetrapods by looking at the global but as such attempts at explanation proceed, the record at the family and stage levels. Unfortunately. pattern itself must be continually examined. Experts the family and stage levels are too coarse a level of on the patterns of each of the supposed mass ex- analysis for these kinds of questions. We try to cir- tinction events must define the terms and taxa in- cumvent this problem by examining two subsets of volved, and ask whether each "event" is real or the world data, data from the Newark Supergroup artificial. (Froelich and Olsen 1984) and data from the Eu- Olsen and Galton (1977) previously asked this ropean Early Mesozoic. question of the Triassic-Jurassic extinctions. The ap- The unique periodic lacustrine cycles of the From: K. Padian (ed.), The Be inning of the Age of Dinosaurs, Faunal Change Across the Triassi~JurassicBoundary, Cambridge University Press. few ~ork,p. 321-351. Paul E. Olsen and Hans-Dieter Sues Newark allow us to look at chronometric sampling Lower Jurassic (Gall, Durand, and Muller 1977; intervals of two million years as well as the individual Chapter 1). Terrestrial tetrapod remains are fairly stage lengths. Although the data are reliable at the common in the Keuper, but are all but absent in the generic level, we use these data at the level of the Lias. As a consequence, it is impossible to compare family rather than the genus or species because too Triassic and Jurassic tetrapod assemblages directly many species are from single localities for the com- in the type area of the Early Mesozoic. The main pilations of generic distributions to be meaningful. goal of this chapter is to make just such a compar- In addition, it can be argued that when looking at ison, and thus it is necessary to correlate, by what- one geographic area, sampling can exert a severe ever means available, other continental beds with bias, especially in the Newark, where osseous re- those of the European Early Mesozoic. Unfortu- mains are not common and there is a bias toward nately, correlation of principally continental beds lacustrine taxa. Therefore, we also look at vertebrate with the marine Jurassic beds has proved very dif- ichnotaxa because they are sampled in the same kind ficult. Obviously, whatever sections we choose as of depositional environment through the Newark, correlative with those of the European Lias deter- and they are extremely abundant and therefore not mine our view of the transition in tetrapods. This as subject to the problems of small sample size that view is necessarily indirect, and therefore our con- plague bony remains. Pollen and spore taxa are ex- clusions based on cumulative faunal lists of the world amined as an independent check on the diversity are somewhat uncertain (as stressed by Colbert in patterns. They are reliable at a much finer taxonomic Chapter 1). level than bones or ichnotaxa, and they do not suffer from small sample size. Correlation problems within the We examine the European Early Mesozoic type areas in Europe only at a stage level, because there is as yet no way Colbert (Chapter 1) reviews the origins of the to calibrate the section independently at a finer main divisions of the Triassic and Jurassic within stratigraphic level. Relative dates are not reliable, Europe. For our purposes, it is necessary to say a and chronostratigraphic measures are elusive. For little more about these divisions and outline the cru- the most part, absolute dating has not been done. cial problems of correlating the type areas of the As we do for the global data, we also restrict our Triassic and Jurassic systems with the type areas of analysis of the European fossil vertebrates to the the standard marine stages. family level because we are unsure of potential syn- The type area of the Triassic is the Germanic onymies at lower taxonomic levels within the assem- Basin of Central Europe. The earliest Mesozoic in blages themselves. Finally, we compare taxonomic the Germanic Basin consists of three vertically seg- rates from the global record of tetrapods to those of regated facies: (1) a lower continental and paralic the marine invertebrate record of the Early Meso- sequence, the Buntsandstein; (2) a middle marine zoic, at the stage and family level. Of interest here sequence, the Muschelkalk; and (3) an upper con- is the comparison among the different patterns. We tinental and paralic sequence, the Keuper. These are treat all of these data, including the global data, with lithological divisions. consistent methods of calculating the average num- The wholly marine sequences of the Alps pro- ber of taxa, normalized origination and extinction vide the type areas for the stages of the Late Triassic; rates per million years, and probabilities of extinc- their history is reviewed by J. T. Gregory (in prep.). tion and origination. The stages in the Alpine Triassic are recognized prin- In the third section of this chapter, we identify cipally by marine invertebrate zones, especially am- which taxa in particular are responsible for the ob- monites. These stages are time-stratigraphic, not served patterns. We go on to examine physical and lithological, units, and there are problems in cor- biological changes through the Early Mesozoic and relating the type areas of the stages of the Late Trias- comment on events that might be synchronous with sic. Specifically, the youngest of the stages of the (and therefore perhaps related to) faunal changes Triassic, the Rhaetian, contains only one ammonite that stand above background levels. zone and is now generally included as the uppermost division of the Norian (Tozer 1974, 1979; Hallam The rationale for correlation of 1981; Pearson 1970). Accordingly, we do not rec- continental Early Mesozoic ognize the Rhaetian as a separate stage. On the other tetrapod assemblages hand, correlation of the Alpine zones with the Mus- A dramatic change in facies marks the tran- chelkalk of the Germanic Basin has been fairly sition between Triassic and Jurassic systems in the straightforward, with the Anisian and Ladinian Germanic Basin of Central Europe. The continental stages mostly represented by the Muschelkalk. How- and paralic Germanic facies of the Upper Triassic, ever, the upper and lower boundaries of the Mus- the Keuper, give way to the fully marine Lias of the chelkalk have proved diachronous, as might be The Triassic-Jurassic tetrapod transition 323 expected. The upper part of the Middle and the thickness of around 500 m (Rutte 1957; Brinkmann Upper Buntsandstein and the Lower Keuper are in- 1960). When correlated with the Alpine section, the cluded in the Early Anisian and Late Ladinian, re- sequence covers the interval from late Ladinian to spectively. The rest of the Buntsandstein is the Triassic-Jurassic boundary.