without a thorough knowledge of changes in classification for all major taxonomic groups
over the last forty years or so.
An alternative method is to accept the classification in the later volume (Benton 1993)
and record when historically the earliest record for a species within a family of fossils was
first published. If stratigraphic ranges are known very well and genuine ghost ranges are
rare, then it is probable that the earliest known stratigraphic record of a given family was
discovered in the 19th century. Alternatively, if the fossil record is full of ghost ranges we
should still be regularly discovering ever older stratigraphic occurrences within families.
Thus the historical distribution of publication dates for the earliest stratigraphic
occurrence of any species within a family should give us a measure of how fossil ranges of
families have been extended over time. Fortunately Benton (1993) is a good source for
the relevant information because most chapters include dates of publication of the earliest
species or of the earliest stratigraphic record of species within families. It soon became
apparent, however, that actual curves tend to be concave up, rather than convex up as the
theory suggests. They have shallow initial portions and steep later portions (see
Figure 5.7). The most likely reason for this is that the theoretical curve assumes the same
search effort throughout, whereas there are far more palaeontologists actively publishing
stratigraphic data now than there were, for example, in the 19th century.
Standard curveA measure of the level of palaeontological activity is needed against which to compare real
discovery curves. In this case the rate of description of new taxa is not an adequate
measure, unlike the arguments used to estimate the completeness of the cystoid fossil
record (Paul 1980, 1982). To extend a stratigraphic range requires at least two
publications; the first to establish the stratigraphic occurrence and the second to extend the
known range. Furthermore, this is an open-ended experiment. The fact that a second
publication did not extend a stratigraphic range does not mean that subsequent
Figure 5.5 Hypothetical collector curve. As more effort is expended it becomes progressively more
difficult to discover stratigraphically earlier occurrences.
GHOST RANGES 99