Chen Xu¹, Michael J. Melchin², David Sheets³, Charles E. Mitchell⁴, and Fan Jun-Xuan¹

¹ Nanjing Institute of Geology and Paleontology, Academia Sinica, Nanjing, China, <xuchen@jlonline.com>,
² Geology Department, St. Francis Xavier University, P.O. Box 5000, Antigonish, Nova Scotia B2G 2W5, Canada <mmelchin@stfx.ca> and
³ Dept. of Physics, Canisius College, 2001 Main St, Buffalo, NY 14208
⁴ The State University of New York at Buffalo, Buffalo, 14260-3050, <cem@geology.buffalo.edu>

We have studied the pattern of graptolite species turnover during the latest Ordovician mass extinction based on four continuous Ashgillian to earliest Llandovery sections together with data from more than 30 other published sections in South China. The studied sections represent relatively shallow water and deeper water belts in the Yangtze platform region. Graphic correlation among these sections reveals that the mass extinction was gradual or stepwise and began with a major extinction event that spanned an interval from near the top of Tangyagraptus typicus Subzone to the middle of the Normalograptus extraordinarius-N. ojsuensis Zone. Graptolite extinctions swept across the Yangtze basin from shallow-water belt to the central Yangtze deeper water belt during this interval. Species diversity fell most rapidly and reached the lowest diversity in the shallower environment. A secondary, more minor pulse of extinction took place late in the interval of the upper Normalograptus persculptus Zone. Overall, fifty-six percent of genera and seventy-seven percent of species of the Ashgillian graptolites expired during the course of the first and more dramatic extinction event. The last five surviving species of the dominant Ordovician clades (the Diplograptidae, Dicranograptidae, and Orthograptidae) and some normalograptid species expired during the succeeding minor extinction. The principal graptolite faunal replacement was completed prior to the end of Ordovician although high rates of extinction also occurred within the Parakidograptus acuminatus Zone of the Lower Llandovery.

Graptolite species range data have been compiled into a composite standard section using graphic correlation. This composite standard section has been calibrated to a temporal scale with a resolution of at least 300 Ky. Using these temporally scaled range data, species diversities, extinction and origination rates and probabilities have been calculated using a variety of analytical methods, including estimated per-capita extinction and origination rates, cohort survivorship and renascence curves, and a capture-mark-recapture method. In addition, several methods, including Monte Carlo simulations and contingency analysis, have been used to test the statistical significance of the observed extinction and origination peaks, the changes in mean rates, and the apparent taxonomic selectivity of the extinctions and originations.

Comparison of the duration and magnitude of the major and minor extinction events among graptolites with the patterns of mass extinction exhibited in other fossil groups reveals several common elements that may reflect the process of species turnover during mass extinction. The graptolite major extinction was preceded by a substantial evolutionary radiation. Speciation among graptolites continued, albeit at a diminished level, during the mass extinction interval. In particular, we recognize a survival-recovery interregnum characterized by continued extinction and the roots of recovery (continued speciation) in the interval between the major and minor extinction events. All the new species that appeared during the survival-recovery interregnum and minor extinction episode belong to what appears to be ecologically generalized species. None of the post-extinction graptolite species appear to be Lazarus taxa, but rather are all newly evolved from a few normalograptid ancestors, the majority of which had themselves evolved during the mass extinction events. The discernible biologic factors associated with the risk of succumbing during graptolite extinction are mainly population size, specialized colony structure and probably a factor related to the pattern of juvenile colony development (astogeny).

Mid-Ashgillian graptolite biogeographical distribution suhhests a latitudinal diversity gradient, which may be controlled mainly by water temperature. This climate gradient becomes much less evident by late Hirnantian time in which most parts of the world have a relatively low diversity fauna totally dominated by normalograpid species, many of which appear to have been eurytopic. Throughout the Ashgillian, however, the Yangtze platform shows a high diversity and long persistence of DDO taxa the mass extinction interval. This may be a consequence of the semi-restricted nature of the basin in which conditions relatively favorable to graptolite survival and speciation existed throughout all or most of the Hirnantian.