Geoscience Reference
In-Depth Information
compilation of the extinction records of genera, and the concept of
waiting time, to build what he called a "kill curve" (Figure 23).
17,1
8
It shows how much time passes on the average between extinction
events of various sizes: the 1-million-year extinction, the 10-million-
year extinction, the 100-million-year extinction, and so on.
Extinctions that destroy 5 percent of species occur about every
million years. Interestingly, this is the approximate length of the
"biostratigraphic zone," the minimum unit of geologic time that
paleontologists can detect using specific assemblages of fossils. An
extinction the size of the K-T has a waiting time of about 100 mil-
lion years, whereas one the magnitude of the more lethal Permian-
Triassic occurs at intervals of 1,000 million or even 10,000 million
years. Since life has never been completely exterminated, Raup
assumes that the curve must level off to form an "S," never reaching
100 percent killed, no matter how long the waiting time.
Raup next turns to Shoemaker's observations of comets and
asteroids, which allow an estimate of how frequently craters of dif-
ferent sizes form (Table 5).
2
0
Figure 23 and Table 5 use two com-
pletely independent sets of data, one obtained from the record in the
rocks, the other derived from searching the heavens. The first relates
waiting time to percent species killed; the second relates waiting time
to crater size. We could write an equation that would plot out each
graph, and from high school algebra we know that we could elimi-
nate the common variable from the two equations, waiting time, and
relate percent species killed directly to crater size. In Figure 24, Raup
has done so. (The curve is dashed above the 150-km-crater diameter
because that is as far as Shoemaker's estimate went. Since these esti-
mates inevitably have large associated errors, the dashed upper and
FIGURE 23
Raup's kill curve,
showing the average time spacing
between extinction events of
different intensity. The Big Five
extinctions (with the possible
exception of the Permian-Triassic
at 96 percent), are
100
-million-
year events. [After Raup.
19
]
