The Great Fossil Enigma (15 page)

After years of economic and political turmoil in Germany, 1933 saw Hitler rise to become chancellor and the beginnings of anti-Semitism and book burning. Soon anti-German sentiments began to affect the reception of German science in the United States just as Schmidt's discovery was beginning to reach conodont workers there. In 1935, Stauffer was the first to record the arrival of Schmidt's innovations. He was convinced that the associations of conodont elements were valid and that “when good specimens or even good fragments of all these are found in close proximity, or as an assemblage of the remains of one individual, a long step towards the reconstruction of the animal can be made.” Branson and Mehl felt the “long step” had not been taken. They looked at Scott's assemblages and dismissed them as chance inclusions in “excretionary matter.” Their proof was that samples of isolated conodonts “fail to show proportional numbers of kinds supposedly found in one individual.” They added that the different types of conodont making up these assemblages had different stratigraphic ranges – how could these different kinds belong to a single animal? Their objections were to become fair tests of the theory. If refuted, then assemblages would be proved anatomical rather than mere coincidences. Until these doubts were answered, Branson and Mehl would continue to class conodonts as the remains of fish and deny the validity of assemblages. As evangelists for conodont stratigraphy their interests lay elsewhere and they were actively attracting new recruits to their way of thinking.
²²
As the most published and therefore the most authoritative workers in this field, their words had impact, particularly on those making practical use of microfossils. Increasingly, stratigraphers began to doubt the truth of Scott's claims.

In Continental Europe things were different. Eichenberg and Schmidt's fish was still being developed and enhanced. At the Royal Belgian Institute of Natural Sciences in Brussels, paleontologist and stratigrapher Félix Demanet had recently found his first conodonts in the Belgian Carboniferous. When, in 1937, Girty's successor at the
USGS
Jim Steele Williams, called on Demanet on his way home from a Moscow trip with his former mentor and colleague, Ted Branson, he found the Belgian at a loss as to how to deal with his new fossils. Williams asked Branson to send Demanet some offprints. Demanet, however, soon became attached to the German way of thinking. The huge number of fossils he had collected confirmed the truth of Schmidt's repeated association of different types.
²³
Schmidt's fish entered his mind. Conodonts were the remains of filtering “appendices” or “processes.” He agreed with Schmidt that there remained just one final question: To which group of fishes do conodonts belong?

Demanet knew that, in 1937, D. M. S. Watson of University College, London, had illustrated the gill rakers of the fossil fish,
Acanthodes.
He read over Watson's description and then went back to T. H. Huxley's much older account of another fossil fish where a similar arrangement was described. Demanet noted Huxley's comment: “Minute horny or osseous filaments seem to have been set at right angles to the branchial [gill] arches along their edges.” This awakened a thought: Did this specimen preserve conodonts in situ? Demanet had Huxley's fish sent over from the Geological Survey Museum in the UK. It arrived as part and counterpart, one of which had formed the basis of Huxley's illustration. But to Demanet's initial disappointment he found that Huxley's illustration wasn't an entirely true reflection of the fossil but had been restored “at least with regard to the normal presence of “filaments” on both sides of the branchial arches.” On the specimen itself there were mere scatterings of these filaments – the very thing Demanet wanted to study. Nevertheless, there was sufficient information in the fossil to confirm Huxley's interpretation and to relate it to Stadtmüller's more recent work. And then he found it, there, on the internal border of a gill arch, a 1.3-millimeter fragment of what he thought was a conodont: “It resembles a small toothed straight comb.” This could be matched with a 4-millimeter-long impression in the counterpart that showed “the Conodont
in situ,”
in contact with the gill arch. Although rather large for a conodont, he thought it proved Schmidt's interpretation was correct.

The Europeans' linear research program had in three important steps turned an idea into a material reality. Schmidt's combination of conodonts was also found in Scotland at this time, further confirming the truth of the European fish.
²⁴

Rather different evidence for conodont fish was found at this time by James S. Cullison of the Missouri School of Mines and Metallurgy in Rolla. He possessed two examples of “an outstanding specimen of a jaw” collected from two different localities. “These are of a bone-like substance on which cones like conodont teeth are set in a single row. These two specimens support the theory that at least some conodonts are the teeth of ancient fishes rather than the jaws of annelids.”
²⁵
Cullison imagined that these tiny and easily detached fish teeth might be taken for primitive conodonts. His strange jaws then became something of an enigma within the enigma. Some, like Ellison, were willing to let them stand as conodont jaws – at least until disproven. They remained objects of controversy.

By now Branson and Mehl's students were regularly writing master's dissertations that used conodonts in stratigraphic study. Almost every one of these introduced some new species, and one, by Gertrude Burnley, even reported finding conodont associations. She did not, however, take the next step and suggest these were reflections of the animal in life. Dan Jones continued his studies at the University of Chicago, writing a doctoral thesis on the Pennsylvanian Seminole formation under the supervision of Croneis. Using his fellow micropaleontology students of the class of 1937 to help with picking and mounting his finds, Jones was then coming around to Scott's way of thinking. He told him, “The evidence furnished by my assemblages tallies very closely with your conclusions that conodonts can be hardly anything else but the masticatory apparatus of annelid worms.”
²⁶
His finds, he said, repeated earlier ones and put the biological validity of natural assemblages beyond doubt; there were “definite types of associations, each containing its characteristic genera and species of conodonts.” He did not give up the ideas of Denham and remained of the view that conodonts could possibly be grasping appendages in a buccal cavity of some sort, probably belonging to an annelid worm. Under Croneis's influence, Chicago was becoming another center for conodont studies. His micropaleontology course was an important draw, and before long pioneer conodont worker Chalmer Cooper, then with the Illinois State Survey, would arrive to register for a PhD.

By the end of the 1930s there were a number of conodont animals in the minds of European and American workers. In Europe the animal was a fish because the Americans had said so, but it had become a peculiarly European fish with a complex apparatus, which gave each conodont element a different functional role. It had been the result of a simple linear research project that revealed and extended the truth in logical steps. In the United States there were some, from Bryant onward, who had been building the fish from isolated parts who welcomed Schmidt's work. Yet another, far simpler fish swam in the minds of those who were devotees of Ulrich and Bassler, or who, like Branson and Mehl, wanted above all else a simple basis for their stratigraphic work. Another group, centered in the state of Illinois, particularly Chicago, had rejected the fish entirely: Here the animal was a worm. What precisely the animal was, it seemed, depended rather more on where one was than on the fossils themselves. But surely the science could not stay like this? More than anything science appeals to universal understanding, not intellectual segregation. It was inevitable that there would be a coming together of the various sides and when they then looked at what they had done, surely they would see that they had created a horribly complicated mess?

 

THERE WERE THREE WAYS TO SOLVE THE RIDDLE OF THE
conodont. The first was to think differently about things known, but if anything too many people were thinking differently. The second was to find better material but this seemed only to deepen the problem. The third – taking advantage of the kind of technological change Zittel and Rohon thought empowering – was to journey into the object itself, and no one had attempted that since the late nineteenth century. An unexplored trail down which progress might be found, in the late 1930s it called to a number of those who had recently become fascinated by the fossil. Among them was Clinton Stauffer, who was perplexed by that simple paradox that now seemed to be at the heart of the problem: an animal with a wormlike arrangement of teeth composed of material indicative of a vertebrate. It prompted him to ask, was the phosphate truly part of the tooth or mere contamination? If simply contamination, then the mystery was solved: The animal was a worm. He asked his technically minded Minneapolis colleague, Duncan McConnell, to resolve the matter. McConnell examined the fossil's chemistry and crystallography and reported that the conodont was indeed composed of material structurally and chemically similar to that making up vertebrate teeth. Stauffer could only conclude, “It becomes evident that the only way to relate conodonts to the worms is to postulate an entirely new group of extinct forms with vertebrate-like teeth.” He continued, “Which might be equivalent to suggesting that they are primitive vertebrates.” Stauffer had been on this vertebrate track for some time, but just as McConnell seemed to give him the confirmation he needed, Bill Furnish, supported by Branson and Mehl, debunked his earlier suggestion that the bar-like conodonts were jaws with teeth inserted in them.
¹

Furnish, who was looking very closely at his fossils too, also believed he had examples of conodont fossils in his own collections showing breakage and repair but not abrasion or wear. This suggested to him that conodonts were perhaps used for grasping rather than mastication.

Stauffer's interpretation of the bar conodonts as jaws indicates how little the Americans knew of the interior structure of conodont fossils even in 1940. Up until then, every internal investigation had been undertaken in German-speaking Europe. This, however, was about to change. The young Wilbert Hass was convinced that too much faith had been put in the simple analogies that might be drawn from an examination of the fossils' external morphology. He was not the first of this new generation who aspired to look into the interior structure of fossil, but those before him had been prevented from doing so by the poor quality of the fossils they found. Situated in Washington, at the heart of the U.S. Geological Survey, Hass was undoubtedly in a privileged position as he could call upon some of the best conodont fossils in the country, and among these he found translucent specimens collected from the Mississippian by Roundy and Chalmer Cooper.
²
By grinding these down, he could produce thin sections for the microscope.

Recognizing that he must compare his findings with Pander's original descriptions, Hass had Pander's book translated. Armed with those descriptions in one hand and “beautiful sections” in the other, he began the most detailed examination of the structure of the conodont in nearly a century and in doing so called upon technologies far better than those available to Pander. Hass also had the advantage of large systematic collections, which even from a casual survey seemed to suggest that the animal's evolution had led to an increase in the surface areas of the elements, culminating in plate-like forms. He also concluded that conodonts grew cone
on
cone by the addition of material to the surface and not, as Pander had suggested, by internal secretion.

Hass's biggest contribution, however, concerned the more complex compound conodonts, which grew like the simple cones but did so from a number of points and in various directions (
figure 4.1
). Some of these conodonts became progressively simpler in form as the animal matured and as some growth points were suppressed. In other words, some of Pander's teeth – unlike vertebrate teeth – actually changed shape markedly as they grew. If these really were teeth, then this suggested that their function changed over time. This discovery itself seemed remarkable, yet it raised a rather more pressing issue: how to distinguish a mature specimen from a growth stage. There was simply no way to tell from the exterior morphology, as two quite different fossils might simply be growth stages of the same species. For Hass this too was a revelation because it meant one could not identify species on the basis of an additional bump or ridge, as had commonly been the practice. The conodont fossil had now acquired even greater ambiguity; it really was a master of illusion. Hass believed that given time it might be possible to document these growth stages, but in the meantime he advised workers to turn their attentions to the “pulp cavity.” As this was the first formed part of the conodont and not altered, it was the best indication of the fossil's true identity and its evolutionary relationships. He then revisited all the new species he had invented in his earlier study of some Montana conodonts and deleted everyone.
³
The conodont population was nowhere near as diverse as he had previously thought.

4.1.
Hass's journey into the anatomy of the fossil. Hass's thin sections through conodont elements viewed in transmitted light.
Left
, a blade showing lamella growth.
Right
, an element showing ‘aberrant effects due to suppression and rejuvenation of parts. From W. H. Hass,
Journal of Paleontology
15 (1941).
SEPM
(Society for Sedimentary Geology).

Hass's evidence looked disappointingly conclusive; conodonts were not teeth and the animal was probably not a vertebrate after all. He concluded, “Conodonts functioned as internal supports for tissues within or on the body of some marine organism at places subject to stresses.” He could add, confirming Furnish's observation, that conodonts did not show tooth-like wear but that, in some cases, breakages had been followed by renewed growth. Some would dispute his “internal supports” theory, but there was widespread recognition that Hass had made the first major step in understanding the internal morphology of the conodont since Pander. Nevertheless, Hass remained uncertain about what the animal might be and refused to speculate. He was a believer in Scott's assemblages but also shared Scott's desire to maintain a utilitarian classification for the sake of the stratigraphic advances that had been made. His was a solid rather than flamboyant contribution.

The curatorially minded Sam Ellison was already making massive strides in introducing a simple yet comprehensive logic to conodont stratigraphy. He was clearly riding a wave of optimism, for he too believed that the answer to the biological conundrum was within reach. In 1944, he began to gather and curate facts. For him the riddle was merely a jigsaw to be solved by finding, sorting, and placing all the pieces.
⁴
With the aid of his colleagues, he compiled a comparative summary of the physical and chemical properties of conodonts, fossil and recent bones and teeth, and various naturally occurring phosphate minerals. Like many papers published on conodonts, it was short, consisting of just four pages of text, one figure, and three tables. His conclusion from this data was equally succinct: “The composition of conodonts is the same as the mineral matter in fossil and modern vertebrate hard parts.” With his first critical piece in place, he now began to assemble the jigsaw around it in hopes that a picture would take shape. To do so, he needed to carefully select the parts: Cullison's jaws, Scott's assemblages, and the new structural and chemical data. Anyone assembling a jigsaw puzzle knows that reference to the box lid simplifies matters, and it seems that the box lid in Ellison's mind carried the image of a vertebrate. He had no more data than the cautious Hass, but he drew his conclusions with a conviction that sometimes comes with youth: “It is evident that conodonts may be considered as vertebrates on the basis of composition, size, shape, associated bone material, and assemblage associations. They are further restricted to the fish or lower vertebrates on the basis of internal structure and stratigraphic occurrence. To consider conodonts as belonging to any other group is to disagree with the evidence afforded by the composition.” More seasoned workers might have seen in Ellison's assertiveness something of the young Scott of a decade before. Each was convinced of his rightness, though both were convinced of quite different things.

Harold Scott joined the University of Illinois in Urbana in 1937, where he would stay for the next thirty years. He had remained silent on conodonts since the publication of his first, potentially groundbreaking, paper in 1934. Fueled by Branson and Mehl's skepticism, that paper had received a muted reception – many simply did not believe it – but Scott seemed in no hurry to answer their tests. He eventually did in May 1942.
⁵
Now a highflyer with a number of noteworthy discoveries under his belt, this was no longer the Scott who had once decided not to rock Branson and Mehl's boat. He had 180 assemblages from Montana, each composed of two or more conodonts in close association, as well as a further 3,000 isolated conodonts showing relative abundances in the proportions expected if complex apparatuses really did exist. All this evidence proved Branson and Mehl were wrong. Strengthened by the doubt that had greeted Branson and Mehl's methods, Scott asked why their lists of conodonts did not match the distribution he was able to demonstrate. Plainly the weakness lay not in the fossil apparatus but in their increasingly criticized lists of names. Scott further suggested that older rocks might contain rather different kinds of apparatus, meaning that one should not expect to find the same relative abundances of the different types anyway. But by then even Branson and Mehl understood that they had contributed to a proliferation of names, many of which were now considered invalid or unnecessary.

Scott had seen Schmidt's paper and he was not unconvinced by his arguments for the fish, though he found Demanet's discoveries very doubtful. Indeed, he was delighted by Schmidt's excellent specimens and the proof they provided that assemblages were real. Conodonts had by then been found in similar arrangements in Illinois, Kentucky, Oklahoma, and Montana, which to Scott made the objections from Missouri ludicrous: “It would be strange indeed to find a group of animals that had such a perfectly balanced diet that the excretal material would consist time after time of one pair of prioniods, one pair of spathodgnaths, one pair of prioniodells, and approximately four pairs of hindeodells.” Branson and Mehl's test, which had so effectively undermined Scott's discovery, was now turned into a crushing blow against them.

Having emphatically destroyed the opposition, Scott set about reconstructing his assemblages, though he remained uncertain about the position of all the components: “They probably operated as rights and lefts, or possibly they were placed in a circular position around an esophageal tract.” He included about twice as many elements in his reconstructions as had Schmidt. Indeed, Scott's reconstruction looked nothing like Schmidt's. In Scott's animal, each conodont element was set in soft tissue: “Such an apparatus would not only form an excellent screen to prevent undesirable objects from entering, but would also present a formidable barrier for the escape of desirable food once it had passed beyond the battery of teeth…. It could operate with equal ease either as the jaw apparatus of an annelid or as gill rakers of a fish.”

Scott had sufficient material to define two genera. One –
Lochriea
– was composed of two species, and he remarked that it was possible to distinguish these on the basis of an individual “tooth,” as he still called them. He also described another genus, which he named
Lewistownella.
These were new names, and as such they broke the zoological rules, for his named assemblages were composed of conodonts that themselves already possessed names. He had decided not to take – for the name of the animal – the oldest established name, as Schmidt and Eichenberg had done. Rather, he argued that he could not be certain that that name belonged to precisely the same element and thus animal. Perhaps that name was associated with a slowly evolving “tooth,” while the animal itself, as understood from its apparatus, showed much more rapid change. How could that unadventurous tooth really help distinguish the different animals? It was clear to Scott that to follow the rules would be to throw conodont studies into “utter confusion.” Instead, he introduced a dual system. The names of isolated conodont fossils were now to be understood as artificial “form genera” and “form species.” It was these that the stratigraphic community would continue to use, but he could not use them as names for the elements within his assemblages. Instead, he took a form name, such as
Hindeodella
, and turned it into a noun (hindeodell) or adjective (hindeodellid) to describe the component elements. Although this approach was a radical step, it seemed logical to Scott, who had established a similar system for the spicules (the tiny skeletal components) of sponges back in 1936. It was a compromise that permitted both camps – the utilitarians and the biologists – to “have their cake and eat it.” It also showed that Scott possessed a very sophisticated understanding of the problem.