Gould, Dinosaurs in the Haystack

Darwin himself had to admit the apparent exceptions:

In some cases, however, the extermination of whole groups of beings, as of the ammonites towards the close of the secondary period, has been wonderfully sudden.

We now come to the central irony that inspired this essay. So long as Darwin's gradualistic view of mass extinction prevailed, paleontological data, read literally, could not refute the basic premise of gradualism--the "spreading out" of extinctions over a good stretch of time before the boundary, rather than a sharp concentration of disappearances right at the boundary itself. For the geological record is highly imperfect and only a tiny fraction of living creatures ever become fossils. As a consequence of this imperfection, even a truly sudden and simultaneous extinction of numerous species will be recorded as a more gradual decline in the fossil record. This claim may sound paradoxical, but consider the following argument and circumstance.

Some species are very common and easily preserved as fossils; we may, on average, find specimens in every inch of strata. But other species will be rare and poorly preserved, and we might encounter their fossils only once every 100 feet or so. Now suppose that all these species died suddenly at the same time, after 400 feet of sediment had been deposited in an ocean basin. Would we expect to find the most direct evidence for mass extinction--that is, fossils of all species through all 400 feet of strata right up to the very top of the sequence? Of course not.

Common species would pervade the strata to the top, for we expect to find their fossils in every inch of sediment. But even if rare species live right to the end, they only contribute a fossil every 100 feet or so. In other words, a rare species may have lived through 400 feet, but its last fossil may be entombed 100 feet below the upper boundary. We might then falsely assume that this rare species died out after three-fourths of the total time had elapsed.

Generalizing this argument, we may assert that the rarer the species, the more likely that its last fossil appears in older sediments even if the species actually lived to the upper boundary. If all species died at once, we will still find a graded and apparently gradualistic sequence of disappearances, the rare species going first and the common forms persisting as fossils right to the upper boundary. This phenomenon--a classic example of the old principle that things are seldom what they seem and that literal appearances often obscure reality--even has a name: the Signor-Lipps effect, to honor two of my paleontological buddies, Phil Signor and Jere Lipps, who first worked out the mathematical details of how a literal petering-out might represent a truly sudden and simultaneous disappearance.

We can now sense the power of Darwin's argument about needing theories to guide observations. We say, in our mythology, that old theories die when new observations derail them. But too often, indeed I would say usually, theories act as straitjackets to channel observations toward their support and to forestall data that might refute them. Such theories cannot be rejected from within, for we will not conceptualize the potentially refuting observations. If we accept Darwinian gradualism in mass extinctions, and therefore never realize that a graded series of fossil disappearances might, by the Signor-Lipps effect, actually represent a sudden wipeout, how will we ever come to consider the catastrophic alternative? For we will be smugly satisfied that we have "hard" data to prove gradualistic decline in species numbers.

New theories are to this conceptual block what Harry Houdini was to straitjackets. We escape by importing a new theory and by making the different kinds of observations that any novel outlook must suggest. For "all observation must be for or against some view," and a new view can therefore engender original observations. I am not making an abstract point or waving arms for my favorite Darwinian motto. Recently, two lovely examples with the same message have been published by a pair of my closest colleagues: studies of ammonites and dinosaurs through the last great extinction.

Anyone who keeps up with press reports on hot items in science knows that a new catastrophic theory of mass extinction has illuminated the paleontological world (and graced the cover of Time magazine) during the past decade. In 1980, the father-son (and physicist-geologist) team of Luis and Walter Alvarez published, with colleagues Frank Asaro and Helen Michel, their argument and supporting data for extraterrestrial impact of an asteroid or comet as the cause of the Cretaceous-Tertiary extinction, most recent of the great mass dyings and the time of extinction for dinosaurs along with some 50 percent of marine invertebrate species.

This proposal unleashed a furious debate that cannot be summarized in a page, much less an entire essay, or even a book. Yet I think it fair to say that the idea of extraterrestrial impact has weathered this storm splendidly and continually increased in strength and supporting evidence. At this point, very few scientists deny that an impact occurred, and debate has largely shifted to whether the impact caused the extinction in toto (or only acted as a coup de grâce for a process already in the works), and whether other mass ex-

395