Christopher R.C. Paul
Department of Earth Sciences, University of Liverpool, L69 3GP, UK. email: crcp@liv.ac.uk
Although the fossil record is the only evidence for the history of life on earth, it suffers from serious biases. Soft-bodied organisms are rarely preserved. Even those with skeletons have lower preservation potentials if they live in areas of active erosion rather than areas of sediment accumulation. Even if preserved, skeletons have differing survival rates, as with calcitic versus aragonitic carbonate. And, of course, there is the legendary incompleteness of the fossil record. So perhaps the most fundamental question in palaeontology is “Can the fossil record be used to interpret patterns of biotic change, or is it so distorted as never to reveal the true course of events?” A related question is “Does increased precision of stratigraphic sampling improve our understanding of the real pattern of events?”
The Cretaceous-Tertiary boundary (KTB) is the best global time plane in the entire rock record. Many sections reveal a boundary clay with all or most of the following characteristics: a high abundance of iridium and other platinum group elements, shocked quartz, microspherules, finely divided carbon often interpreted as soot, a ‘fern spike’ and a 2 ‰ fall in d 13C values. And, of course, the boundary clay corresponds to a major biotic turnover. It is inconceivable that the boundary clay is not the same deposit wherever it is encountered. The boundary clay occurs in terrestrial, shallow and deep marine sections world-wide. Furthermore, it exhibits characteristics of a fall-out deposit. Historical volcanic events (e.g. Krakatao, Mt. St. Helens) which were minor events compared to the proposed impact at the KTB, produced atmospheric/climatic effects globally within one year of the eruptions. It is distinctly possible that the base of the boundary clay is of the same age globally to within a year. Even if it is the same age to within a millenium, this represents unrivalled stratigraphic precision compared to any other event or dating technique. Thus the KTB is the best place in the fossil record to test the reliability of patterns of biotic turnover.
The record of planktonic foraminifera is used as an example of biotic turnover across the KTB because it has proved controversial as regards the timing (sudden versus gradual), severity (proportion of species which became extinct) and the relative completeness of KTB sections. The timing is affected by the Signor-Lipps Effect (and its mirror image, the Jaanusson Effect). Signor and Lipps (1982) suggested that a combination of varying sample sizes and range truncation could cause a sudden mass extinction to appear gradual. This has since become known as the Signor-Lipps Effect. Similarly, Jaanusson (1976) argued that is was extremely difficult to date the time of immigration of ostracods into the Baltic area during the middle Ordovician because relative abundance and range truncation combined to make the first record occur at some unknown level above the actual immigration event. I refer to this latter as the Jaanusson Effect (also known as the Sppil-Rongis Effect). Both result from the patchiness of the fossil record (which in turn reflects the real local distributions of living organisms), and hence they are both universal in the fossil record. They do not just affect mass extinctions. It follows that before interpreting any pattern of biotic turnover in the fossil record, these effects must be quantified and compensated for. I present a new technique which allows direct comparison of observed patterns of diversity decline approaching a mass extinction with the mean value for the Signor-Lipps Effect. Applied to the record of benthonic foraminifera across the Cenomanian-Turonian Boundary, what appears to be a classic example of gradual decline can be shown to have been very sudden. Alternatively, the decline in diversity of planktonic foraminifera across the KTB at El Kef undoubtedly started well before the boundary event, but even here the Signor-Lipps Effect accounts for half the observed decline in diversity at the boundary.
The severity of the extinction relates to whether or not Cretaceous species found in the Tertiary are genuine survivors or reworked specimens. Arz, Arenillas and Molino (1999) have shown that examples of both occur in the Zumaya section. The best technique to test for survivorship is probabilistic stratigraphy (Hay, 1982). I have confined this to the most basic question: “Is it more probable that the extinction of a given species occurred before or after the boundary?” Results suggest that perhaps 22% of Cretaceous planktonic foraminifera survived the boundary. This is almost certainly a generous estimate because I have accepted any record above the boundary as genuine. One may reasonably ask, does a single occurrence 1 cm above the boundary constitute survival? Filtering data by level (e.g. only accepting records at least 10 or 20 cm above the boundary), by frequency (e.g. rejecting all single occurrences above the boundary), etc. would reduce this proportion of surviving species. Since it is known that stratigraphic ranges can be extended up or down by bioturbation by > 1 m, and up by even more by physical reworking of sediments, applying such filters is not unwarranted. Indeed it is this aspect that raises the question of whether refined biostratigraphic sampling provides a better picture of the true course of events. This is a particular problem with events such as the KTB mass extinction, because so many range terminations occur within such a narrow stratigraphic interval.
Finally, it has been argued that only KTB sections with all zones present are complete and that sections with a thick zone P0 (the first zone of the Tertiary) are the best to investigate patterns of biotic turnover across the KTB because sedimentation rates were high. However, zone P0 is an interval zone defined by a last occurrence below and a first occurrence above. If the fossil record is degraded across the KTB, the Signor-Lipps Effect will tend to move any last occurrence down section and the Jaanusson Effect will tend to move any first occurrence up section. Thus it is possible to widen zone P0 artificially by degrading the quality of the fossil record, for example by randomly removing records of fossil occurrences. It follows that sections with a wide zone P0 may result from the poor quality of the local fossil record rather than from high sedimentation rates. Equally, it is possible that sections in which zone P1(a) lies directly above the boundary clay may, in fact, be the most complete.
The inevitable conclusion from the above is that first and last occurrences of fossils are not necessarily good proxies for their times of origin/ immigration or extinction/ emigration. The Signor-Lipps and Jaanusson effects are universal in the fossil record and it is essential to quantify them before interpreting any pattern of biotic turnover. Probabilistic stratigraphy is the best technique to test for genuine survival across the KTB or any other mass extinction. Interval zones are best avoided, but where unavoidable care is needed in interpreting the significance of their apparent durations.
References
Arz, J. A., Arenillas, I. and Molina, Y. E. 1999. Extinción de foraminíferos planctónicos en el tránsito Cretácico-Terciario de Zumaya (Guipúzcoa): supervivencia o reelaboración? Revista Española de Micropaleontología, 31, 297-304.
Hay, W.W. 1982. Probabilistic stratigraphy. Eclogae Geologicae Helvetiae, 65, 255-266.
Jaanusson, V. 1976. Faunal dynamics in the Middle Ordovician (Viruan) of Balto-Scandia, in The Ordovician System: proceedings of a Palaeontological Association symposium, Birmingham, September 1974: Cardiff, Bassett, M.G. (ed.), 301-326. University of Wales Press & National Museum of Wales.
Signor, P.W. and Lipps, J.H. 1982. Sampling bias, gradual extinction patterns and catastrophes in the fossil record. In Geological implications of impacts of large asteroids and comets on the earth, Silver, H. T. and Schultz, P. H. (eds). Geological Society of America Special Paper, 190, 291-296.