COMPARING STRATIGRAPHIC OCCURRENCE WITH CLADISTIC HYPOTHESES: A CASE STUDY USING ORDOVICIAN GRAPTOLITES

Stratigraphic occurrences and cladistic branching order are often compared to provide the most accurate representation of evolutionary order in fossil taxa. A stratigraphic character was added a posteri to choose among competing cladistics hypotheses (most parsimonious trees) in a 33 taxon data set. These taxa represent planktic graptoloids that range in age from lower Ordovician (Tremadocian) to Upper Ordovician (Katian). Stratigraphic data are binned by biozone, based on the succession in the Cow Head Group, Western Newfoundland. We were able to decrease the number of retained most parsimonious trees from 80 to 24 based on which were least modified by the addition of the stratigraphic character. The degree to which the most parsimonious trees were modified is based on the change in total number of steps after the inclusion of the first appearance datum (FAD) as an ordered character. The branching sequence in the cladogram was then compared to the FAD's. Any species that had FAD's that were stratigraphically higher than those species that occurred next in the cladistic branching sequence had ghost ranges extended downward. The resulting range chart then represents a stratigraphic hypothesis that best matches the cladistics analysis, while our consensus tree represents a cladistical hypothesis that best matches the stratigraphy. Incomplete sampling of the fossil record often invalidates temporal occurrence, and incorporating parsimony data into graptolite ranges has provided an alternate tool for stratigraphic interpretation. A stratigraphic consistency index (SCI) was calculated to determine the proportion of nodes that are consistent with the tree and provide a means of estimating congruence. Many nodes provide a good fit to the cladogram, but there is still some incongruence in other nodes. Planktic graptolites have a reasonably complete fossil record, but there still remains some disagreement with morphology based parsimony.