The Panda's Thumb: More Reflections in Natural History

by Stephen Jay Gould

  D’Arcy Wentworth Thompson, classical scholar, Victorian prose stylist, and glorious anachronism of twentieth-century biology, dealt with this dilemma in his classic treatise On Growth and Form.

  An algebraic curve has its fundamental formula, which defines the family to which it belongs…. We never think of “transforming” a helicoid into an ellipsoid, or a circle into a frequency curve. So it is with the forms of animals. We cannot transform an invertebrate into a vertebrate, nor a coelenterate into a worm, by any simple and legitimate deformation…. Nature proceeds from one type to another…. To seek for steppingstones across the gaps between is to seek in vain, forever.

  D’Arcy Thompson’s solution was the same as Goldschmidt’s: the transition may occur in simpler and more similar embryos of these highly divergent adults. No one would think of transforming a starfish into a mouse, but the embryos of some echinoderms and protovertebrates are nearly identical.

  1984 will mark the 125th anniversary of Darwin’s Origin, the first major excuse for a celebration since the centenary of 1959. I hope that our “new speaking” these few years hence will be neither dogma nor vacuous nonsense. If our entrenched, a priori preferences for gradualism begin to fade by then, we may finally be able to welcome the plurality of results that nature’s complexity provides.

  19 | The Great Scablands Debate

  THE INTRODUCTORY PARAGRAPHS of popular guidebooks usually tout prevailing orthodoxy in its purest form—dogma unadulterated by the “howevers” of professional writing. Consider the following from our National Park Service’s auto tour of Arches National Park:

  The world and all it contains is in a continuous process of change. Most of the changes in our world are very tiny and so escape our notice. They are real, however, and over an immense span of time their combined effect is to bring about great change. If you stand at the base of a canyon wall and rub your hand on the sandstone, hundreds of grains of sand are dislodged. It seems like an insignificant change, but that’s how the canyon was formed. Various forces have dislodged and carried away grains of sand. Sometimes the process is “very fast” (as when you rub the sandstone) but most of the time it is much slower. If you allow sufficient time, you can tear down a mountain or create a canyon—a few grains at a time.

  As the primary lesson of geology, this pamphlet proclaims that big results arise as the accumulated effect of tiny changes. My hand rubbing the canyon wall is an adequate (if anything, overeffective) illustration of rates that carved the canyon itself. Time, geology’s inexhaustible resource, performs all the miracles.

  The channeled scablands of eastern Washington.

  Yet, when the pamphlet turns to details, we encounter a different scenario for erosion in Arches. We learn that a balanced rock known as “Chip Off the Old Block” fell during the winter of 1975–76. Before and after photographs of the magnificent Skyline Arch receive the following commentary: “It remained thus for as long as man knew the arch, until, late in 1940, the block of stone fell, and Skyline was suddenly twice its former size.” The arches form by sudden, intermittent collapse and toppling, not by imperceptible removal of sand grains. Yet gradualist orthodoxy is so entrenched that the authors of this pamphlet failed to note the inconsistency between their own factual account and the stated theory of their introduction. In other essays of this section, I argue that gradualism is a culturally conditioned prejudice, not a fact of nature, and I make a plea for pluralism in concepts of rate. Punctuational change is at least as important as imperceptible accumulation. In this essay I tell a local, geologic story. But it conveys the same message—that dogmas play their worst role when they lead scientists to reject beforehand a counterclaim that could be tested in nature.

  Flow basalts of volcanic origin blanket most of eastern Washington. These basalts are often covered by a thick layer of loess, a fine-grained, loosely packed sediment blown in by winds during the ice ages. In the area between Spokane and the Snake and Columbia rivers to the south and west, many spectacular, elongate, subparallel channelways are gouged through the loess and deeply into the hard basalt itself. These coulees, to use the local name, must have been conduits for glacial meltwaters, for they run down gradient from an area near the southern extent of the last glacier into the two major rivers of eastern Washington. The channeled scablands—as geologists designate the entire area—are puzzling as well as awesome, and for several reasons:

  1. The channels connect across tall divides that once separated them. Since the channels are hundreds of feet deep, this extensive anastomosis indicates that a prodigious amount of water must once have flowed over the divide.

  2. As another item favoring channels filled to the brim with water, the sides of the coulees contain many hanging valleys where tributaries enter the main channels. (A hanging valley is a tributary channel that enters a main channel high above the main channel’s modern stream bed.)

  3. The hard basalt of the coulees is deeply gouged and scoured. This pattern of erosion does not look like the work of gentle rivers in the gradualist mode.

  4. The coulees often contain a number of high-standing hills composed of loess that has not been stripped away. These are arranged as if they were once islands in a gigantic braided stream.

  5. The coulees contain discontinuous deposits of basaltic stream gravel, often composed of rock foreign to the local area.

  Just after World War I, Chicago geologist J Harlen Bretz advanced an unorthodox hypothesis to account for this unusual topography (yes, that’s J without a period, and don’t ever let one slip in, for his wrath can be terrible). He argued that the channeled scablands had been formed all at once by a single, gigantic flood of glacial meltwater. This local catastrophe filled the coulees, cut through hundreds of feet of loess and basalt, and then receded in a matter of days. He ended his major work of 1923 with these words:

  Fully 3,000 square miles of the Columbia Plateau were swept by the glacial flood, and the loess and silt cover removed. More than 2,000 square miles of this area were left as bare, eroded rock-cut channel floors, now the scablands, and nearly 1,000 square miles carry gravel deposits derived from the eroded basalt. It was a debacle which swept the Columbia Plateau.

  Bretz’s hypothesis became a minor cause célèbre within geological circles. Bretz’s stout and lonely defense of his catastrophic hypothesis won some grudging admiration, but virtually no support at first. The “establishment,” as represented by the United States Geological Survey, closed ranks in opposition. They had nothing better to propose, and they did admit the peculiar character of scabland topography. But they held firm to the dogma that catastrophic causes must never be invoked so long as any gradualist alternative existed. Instead of testing Bretz’s flood on its own merits, they rejected it on general principles.

  On January 12, 1927, Bretz bearded the lion in its lair and presented his views at the Cosmos Club, in Washington, D.C., before an assembled group of scientists, many from the Geological Survey. The published discussion clearly indicates that a priori gradualism formed the basis for Bretz’s glacial reception. I include typical comments from all detractors.

  W. C. Alden admitted “it is not easy for one, like myself, who has never examined this plateau to supply offhand an alternative explanation of the phenomena.” Nonetheless, undaunted, he continued: “The main difficulties seem to be: (1) The idea that all the channels must have been developed simultaneously in a very short time; and (2) the tremendous amount of water that he postulates…. The problem would be easier if less water was required and if longer time and repeated floods could be allotted to do the work.”

  James Gilluly, this century’s chief apostle of geological gradualism, ended a long comment by noting “that the actual floods involved at any given time were of the order of magnitude of the present Columbia’s or at most a few times as large, seems by no means excluded by any evidence as yet presented.”

  E. T. McKnight offered a gradualist alternative for the gravels: “This writer believes them to
be the normal channel deposits of the Columbia during its eastward shift over the area in preglacial, glacial, and postglacial times.”

  G. R. Mansfield doubted that “so much work could be done on basalt in so short a time.” He also proposed a calmer explanation: “The scablands seem to me better explained as the effects of persistent ponding and overflow of marginal glacial waters, which changed their position or their places of outlet from time to time through a somewhat protracted period.”

  Finally, O. E. Meinzer admitted that “the erosion features of the region are so large and bizarre that they defy description.” They did not, however, defy gradualist explanation: “I believe the existing features can be explained by assuming normal stream work of the ancient Columbia River.” Then, more baldly than most of his colleagues, he proclaimed his faith: “Before a theory that requires a seemingly impossible quantity of water is fully accepted, every effort should be made to account for the existing features without employing so violent an assumption.”

  The story has a happy ending, at least from my point of view, for Bretz was delivered from the lion’s lair by later evidence. Bretz’s hypothesis has prevailed, and virtually all geologists now believe that catastrophic floods cut the channeled scablands. Bretz had found no adequate source for his floodwaters. He knew that the glaciers had advanced as far as Spokane, but neither he nor anyone else could imagine a reasonable way to melt so much water so rapidly. Indeed, we still have no mechanism for such an episodic melting.

  The solution came from another direction. Geologists found evidence for an enormous, ice-dammed glacial lake in western Montana. This lake emptied catastrophically when the glacier retreated and the dam broke. The spillway for its waters leads right into the channeled scablands.

  Bretz had presented no really direct evidence for deep, surging water. Gouging might have proceeded sequentially, rather than all at once; anastomosis and hanging valleys might reflect filled coulees with gentle, rather than raging, flow. But when the first good aerial photographs of the scablands were taken, geologists noticed that several areas on the coulee floors are covered with giant stream bed ripples, up to 22 feet high and 425 feet long. Bretz, like an ant on a Yale bladderball, had been working on the wrong scale. He had been walking over the ripples for decades but had been too close to see them. They are, he wrote quite correctly, “difficult to identify at ground level under a cover of sagebrush.” Observations can only be made at appropriate scales.

  Hydraulic engineers can infer the character of flow from the size and shape of ripples on a stream bed. V. R. Baker estimates a maximum discharge of 752,000 cubic feet per second in the scabland flow channels. Such a flood could have moved 36-foot boulders.

  I could end here with a cardboard version of the story much to my liking: Perceptive hero suppressed by blinded dogmatists stands firm, expresses his allegiance to fact over received opinion, and eventually prevails by patient persuasion and overwhelming documentation. The outline of this tale is surely valid: gradualist bias did lead to a rejection of Bretz’s catastrophic hypothesis out of hand, and Bretz (apparently) was right. But, as I read through the original papers, I realized that this good guy-bad guy scenario must yield to a more complex version. Bretz’s opponents were not benighted dogmatists. They did have a priori preferences, but they also had good reasons to doubt catastrophic flooding based on Bretz’s original arguments. Moreover, Bretz’s style of scientific inquiry virtually guaranteed that he would not triumph with his initial data.

  Bretz proceeded in the classic tradition of strict empiricism. He felt that adventurous hypotheses could only be established by long and patient collecting of information in the field. He eschewed theoretical discussion and worried little about the valid conceptual problem that so bothered his adversaries: where could so much water come from so suddenly?

  Bretz tried to establish his hypothesis by toting up evidence of erosion in the field, piece by patient piece. He seemed singularly uninterested in finding the missing item that would render his story coherent—a source for the water. For this attempt might involve speculation without direct evidence, and Bretz relied only upon fact. When Gilluly challenged him on the absence of a source for the water, Bretz simply replied: “I believe that my interpretation of channeled scabland should stand or fall on the scabland phenomena themselves.”

  But why should an opponent be converted by such an incomplete theory? Bretz believed that the southern end of the glacier had melted precipitously, but no scientist could imagine a way to melt ice so quickly. (Bretz tentatively suggested volcanic activity under the ice, but quickly abandoned the theory when Gilluly attacked.) Bretz stayed in the scablands, while the answer sat in western Montana. Glacial Lake Missoula had been in the literature since the 1880s, but Bretz did not make the connection—he was working in other ways. His opponents were right. We still do not know a way to melt so much ice so quickly. But the premise shared by all participants was wrong: the source of the water was water.

  Events that “cannot happen” according to received wisdom rarely gain respectability by a simple accumulation of evidence for their occurrence; they require a mechanism to explain how they can happen. Early supporters of continental drift ran into the same difficulty that Bretz encountered. Their evidence of faunal and lithological similarities between continents now widely separated strikes us today as overwhelming, but it failed in their time because no reasonable force had been proposed for moving continents. The theory of plate tectonics has since provided a mechanism and established the idea of continental drift.

  Moreover, Bretz’s opponents did not rest their case entirely on the unorthodox character of Bretz’s hypothesis. They also marshaled some specific facts on their side, and they were partly right. Bretz originally insisted upon a single flood, while his opponents cited much evidence to show that the scablands had not formed all at once. We now know that Lake Missoula formed and re-formed several times as the glacial margin fluctuated. In his latest work, Bretz called for eight separate episodes of catastrophic flooding. Bretz’s opponents were wrong in inferring gradual change from the evidence of temporal spread: catastrophic episodes can be separated by long periods of quiescence. But Bretz was also wrong in attributing the formation of the scablands to a single flood.

  I prefer heroes of flesh, blood, and fallibility, not of tinseled cardboard. Bretz is inscribed on my ledger because he stood against a firm, highly restrictive dogma that never had made any sense: the emperor had been naked for a century. Charles Lyell, the godfather of geological gradualism, had pulled a fast one in establishing the doctrine of imperceptible change. He had argued, quite rightly, that geologists must invoke the invariance (uniformity) of natural law through time in order to study the past scientifically. He then applied the same term—uniformity—to an empirical claim about rates of processes, arguing that change must be slow, steady, and gradual, and that big results can only arise as the accumulation of small changes.

  But the uniformity of law does not preclude natural catastrophes, particularly on a local scale. Perhaps some invariant laws operate to produce infrequent episodes of sudden, profound change. Bretz may not have cared for this brand of philosophical waffling. He probably would brand it as vacuous nonsense preached by an urban desk man. But he had the independence and gumption to live by a grand old slogan from Horace, often espoused by science but not often followed: Nullius addictus jurare in verba magistri, “I am not bound to swear allegiance to the words of any master.”

  My tale ends with two happy postscripts. First, Bretz’s hypothesis that channeled scabland reflects the action of catastrophic flooding has been fruitful far beyond Bretz’s local area. Scablands have been found in association with other western lakes, most notably Lake Bonneville, the large ancestor of a little puddle in comparison—Great Salt Lake, Utah. Other applications have ranged about as far as they can go. Bretz has become the darling of planetary geologists who find in the channelways of Mars a set of features best interpreted by B
retz’s style of catastrophic flooding.

  Second, Bretz did not share the fate of Alfred Wegener, dead on the Greenland ice while his theory of continental drift lay in limbo. J Harlen Bretz presented his hypothesis sixty years ago, but he has lived to enjoy his vindication. He is now well into his nineties, feisty as ever and justly pleased with himself. In 1969, he published a forty-page paper summarizing a half century of controversy about the channeled scablands of eastern Washington. He closed with this statement:

  The International Association for Quaternary Research held its 1965 meeting in the United States. Among the many field excursions it organized was one in the northern Rockies and the Columbia Plateau in Washington…. The party…traversed the full length of the Grand Coulee, part of the Quincy basin and much of the Palouse Snake scabland divide, and the great flood gravel deposits in the Snake Canyon. The writer, unable to attend, received the next day a telegram of “greetings and salutations” which closed with the sentence, “We are now all catastrophists.”

  Postscript

  I sent a copy of this article to Bretz after its publication in Natural History. He replied on October 14, 1978:

  Dear Mr. Gould,

  Your recent letter is most gratifying. Thank you for understanding.

  I have been surprised by the way my pioneer Scabland work has been applauded and further developed. I knew all along that I was right but the decades of doubt and challenge had produced an emotional lethargy, I think. Then the surprise following Victor Baker’s field trip in June woke me up again. What! Had I become a semi-authority on extra-terrestrial processes and events?