The blind sample test was designed to determine whether the observed species extinction pattern across the K/T boundary supports Smit’s (1982, 1990) scenario of all but one species suddenly extinct at the K/T boundary, or Keller’s (1988b) scenario of gradual extinctions with some species disappearing below and the majority at the K/T boundary with l/3 ranging into the Danian. The blind test can only resolve the controversy regarding the observed pattern of extinction, and not the controversy regarding the interpretation of this pattern.

The Smit and Keller extinction models are shown in Figs. 11 and 12. Since Smit has not published a complete census list of Cretaceous taxa, his model is illustrated in Fig. 11 without species names. Keller’s (1988b) data are shown in Fig. 12 for six stratigraphic levels that are equivalent to the six blind test samples.

The four testers were asked to collect species census data from the >63 pm size fraction, provide relative abundances of each species and the benthic/ planktic ratio based on population counts of 300 to 400 individuals in a random sample split. Unfortunately, not all testers used the same data gathering methods and as a result the relative species abundances and benthic/planktic ratio data differ by more than one magnitude. These data are excluded from this analysis. The blind sample test therefore rests upon taxic census data only. Taxic census data, however, have their own problems. They essentially vary from tester to tester based on taxonomic concepts. This may result in different species names used for the same morphotypes among the four testers, or in several different species names given to morphotypes that some testers consider to be morphologic variants of the same species.

The degree to which different taxonomic concepts influenced the taxic census data is seen by the number of species identified and by the common species names used. In Maastrichtian samples, the number of species identified by each of the blind testers are: Canudo 47, Olsson 45, Masters 52, and Orueetxebarria 59; in similar samples, Keller (1988b) identified 51 species. All four testers used the same species names for 14 to 16 species and three testers used the same species names for 10 to 16 species. Taxonomic agreement for the remaining species is low. This illustrates the fact that taxic census data of the four testers cannot be compared on a species by species basis. Even if the same species names are used, there is no guarantee that all testers applied that species name to the same morphotype. But, we can be reasonably sure that each tester applied each species name to a distinct and different morphotype. For these reasons, comparison of the patterns of extinction of all taxa is more instructive than comparing extinction of species by species using the same species names.

There is no way taxonomic differences between testers can be identified and isolated in the data sets without getting them all together to sort out their taxonomic differences. But, doing so would defeat the purpose of the blind test which was to get a spectrum of outside views to compare and contrast with &nit’s (1982, 1990) and Keller’s (1988b) studies. The degree of variation between the testers is in itself an important and real variable of differences between workers and therefore must be preserved, if the test is to approximate the spectrum of opinions. The real test will be whether a common extinction pattern emerges from these analyses. If so, then individual variations in taxonomic concepts are less relevant. If no common extinction pattern emerges, then taxonomic differences override the actual data. In this section, patterns of extinction, irrespective of species names, are discussed first, followed by analysis of the statistical similarity between the stratigraphic position of species identified by each tester.   PDF