The Great Fossil Enigma (31 page)

Thus far Seddon and Druce had been working in parallel, and broadly with the same geographic, stratigraphic, and intellectual goals. By 1970, conodont ecology was becoming a hot field, not least because it tested the idea of the universal conodont. Increasing numbers of workers began to develop similar and overlapping topics of study. Competition and debate increased. This parallel production of data and knowledge meant that individuals tended to know different things, and this did not depend solely on personal networks and research groups. Some would gather new data or ideas by attending meetings while others would not receive this information until the paper was published – which might be years later or not at all. Some important studies were published merely as short abstracts of orally presented papers. The abstracts were hardly scientific arguments, but they acted as important markers and were often referred to. In 1970, both Druce and Seddon heard each other speak and both published short abstracts of their papers.
²⁰
Druce made a commitment to write up his paper for the book arising from the conference, but this book would have such a catastrophically delayed gestation that the paper would not appear for three years. And when it did, it was not the paper Druce presented at the meeting but one that reflected upon that meeting and Seddon's presentation there. We shall come to that paper shortly.

Seddon, who produced a rather different interpretation of the data for the 1970 meeting, chose not to publish in the conference volume. Instead, he teamed up with Sweet to explain his filtering mechanism in more detail. In order to do so the two men looked for a “likely ecologic analogue” and chose the chaetognaths, or arrow worms. These are typically carnivores, three centimeters long, “that spend their entire existence floating or swimming in the water without relation to the bottom.”
²¹
Although Seddon and Sweet suggested no direct relationship between conodonts and these animals, they could clearly talk of the analogue darting forward to capture prey and visualize their own animal doing the same. There were other similarities too, as one authority considered the chaetognaths something of an enigma: “They may be the most isolated group in the animal kingdom.” Doubtless their “paired batteries of anterior and posterior teeth, and grasping spines” were sufficiently unique to support Seddon and Sweet's contention that they had before them a “suggestive analogy.” In doing so, they drew upon a number of books to act as their authorities on these animals, many of which came from the 1950s. These sources told them that the animal's wide distribution was controlled by temperature, salinity, and available food, and that species were vertically stratified: Most species occurred in water depths of less than two hundred meters, but more specialized forms could be found between two hundred and one thousand meters, and a few even beyond that depth (figure 9.2b). This suggested a possible mechanism for the operation of the biological filter that affected conodont distribution. It had long been suspected that
Icriodus
was a shallow-water form but now this genus could be visualized as occupying surface waters, overlying the deeper waters where
Palmatolepis
swam or floated. On death, conodonts elements would sink to the seafloor. Those that accumulated in deeper waters would as a result contain both shallow- and deep-water forms.

Drawing upon the contemporary literature, Seddon and Sweet could generalize further and suggest that the shallow zone contained just a few unspecialized conodont species, while at depth, where the environment was more stable, there was greater diversity. Sweet's Ordovician then became the testing ground for this new depth-stratification model, but here Seddon and Sweet could not call upon the relative simplicity of the reef model to infer water depth or relationships between communities. They were, of course, dealing with fundamentally different genera, but they still felt that within the different provincial faunas it was possible to detect this two-way, depth-controlled division of genera. Indeed, they thought the ratio between two particular genera might provide a crude index of water depth, which could be confirmed by lithological data. As Seddon and Sweet were developing this idea, others reported Seddon's distinctive
Icriodus
fauna in various parts of the United States, demonstrating its wide distribution.

When Druce's paper finally did appear, two years after Seddon and Sweet's, it showed the influence of Seddon's work. It further distinguished a third (
Belodella
) fauna, which Seddon had included with
Icriodus
but Druce thought occurred still closer to the reef. Druce sought to test and extrapolate his model across the whole of the Upper Paleozoic and into the Triassic, aware that very similar apparatus architectures and element morphologies had repeatedly evolved across long periods of time. Homeomorphy, the repeated appearance of identical forms as a result of environmentally determined convergent evolution, suggested this close link between the form of an animal and the nature of the environment.
²²
This kind of morphological convergence had been seen across the natural world in everything from mollusk shells to teeth. It was reasonable to expect it in conodonts even if the animal remained unknown and unimaginable. How wonderfully informative these tiny fossils would be if this was the case; find a microscopic fossil, and there, in your hand, you have the key to understanding the world in which it lived. Druce was not alone in thinking in this way; Ronald Austin compared Walliser's Silurian species to his Carboniferous forms and, surprised by how similar they were, had similar thoughts.

Ecological models reflected this desire for generalization: They emerged almost subliminally from the data and would then perform as spectacles through which others would look at fossils and the rocks that held them. These workers might accept this new way of seeing, cast these spectacles aside, or make improvements to them. Druce chose to do the latter. He was sure that Seddon's model was better than his own but still not perfect. He knew the
Palmatolepis
fauna was more widespread than that of
Icriodus
, but Seddon's model predicted the opposite. So Druce refined it, concentrating populations toward the shore and effectively superimposing his earlier model on Seddon's (
figure 9.1c
).

In 1970, the appearance of
Icriodus
and other Devonian conodonts in various parts of the world raised a number of questions about the faunas Ziegler was finding. His collections inexplicably lacked these fossils. Whichever of the models was correct, there was good reason to believe these conodonts should also occur in Germany, and not simply because the conodont was celebrated there as the universal fossil. Druce wondered if some kinds of conodont had not been recorded because they were less exotic, often long-ranging, and therefore less useful to stratigraphy. Austin and Rhodes had similar suspicions and began to doubt Ziegler's data. They had found that British, Irish, and Belgian conodont faunas were similar to one another but different from those Ziegler had published. (Later these faunas would turn up in Pakistan, Russia and China.) Ziegler found himself increasingly under assault on this point, and attention now focused on the methods he used to process his fossils. Doubts had been raised about Beckmann's method, which Ziegler used, as early as 1964. At the time, Beckmann, Ziegler, and all the other British, American, and German heavyweights in the field closed ranks and rejected these claims. But with the identification of statistical assemblages a few years later it became increasingly obvious that Ziegler's faunas were incomplete. The arrival of this new shallow-water
Icriodus
fauna simply accentuated these doubts.

Ziegler remained resistant to change until Lindström moved to Marburg and used Beckmann's method on his Ordovician conodont samples. To his great surprise, Lindström found the conodonts seriously affected. Indeed, the results were quite spectacular. From two hundred grams of limestone, using Beckmann's method and leaving the sample bubbling away for two weeks, Lindström obtained just six badly corroded cone fragments. Repeating the exercise but recovering the residue every two days, he obtained eight hundred identifiable specimens, a few of which were corroded. This dissolution of conodonts was selective, bar-like fossils being particularly vulnerable. This seemed to explain why these elements were missing from Ziegler's collections. In contrast, the platforms, which so interested Ziegler, were much more resilient. But surprisingly, this discovery did not greatly alter Ziegler's use of acids; he simply removed his fossils from the acid solution more regularly. He believed this was fine for stratigraphic work in which only the platforms were required but that where more detailed paleobiological studies were to be carried out, acetic acid should be used.
²³
Ziegler was perhaps typical of most conodont workers: If they could find sufficient specimens, they tended to believe the method was fine.

Others were more inclined to frown at Ziegler's disregard for producing accurate samples. One in particular, Lennart Jeppsson, would take the opposite approach and go to great pains to improve his methods and reduce the risks of loss or distortion. Jeppsson needed to process huge quantities of limestone to get just a few specimens. Sometimes he was literally looking for needles in haystacks.

The ecological models that soon became well known in the conodont research community did not go unchallenged. Glen Merrill, for example, had argued from the early 1960s that conodonts were environmentally controlled. For him, the universal animal was an illusion resulting from inadequate sampling. Having submitted a paper on the subject to the same volume that would eventually publish Druce's, he felt compelled to recall it so he could criticize Seddon and Sweet's model. Merrill's contribution was to point to a phenomenon he had seen in Pennsylvanian faunas and that Chalmer Cooper had briefly mentioned back in 1947. (Indeed, it was not unlike that distribution of fossils Rexroad had spotted in the late 1950s.) Merrill could show that alternative abundances in two platform genera depended upon lithology.
²⁴
One was dominant in shales, the other in limestones. The relationship was both consistent and remarkable.

Merrill knew that the shales were produced near the shore while the limestones represented fully marine conditions farther out to sea. The rocks representing these two different environments were stacked one upon the other, which Merrill suggested represented fluctuating sea levels. As the sea level rose, the shoreline would move inland and these two environments and their conodonts would follow.

Merrill then asked himself how rapidly these changes took place. He could answer that question by using an old idea that required a little lateral thinking. On the presumption that animals were dying at a constant rate and falling to the seafloor to form fossils, high concentrations suggest that the rate of sedimentation has been low and a thin layer of rock represents a long period of time. Finding these fossils heavily diluted by sediment could then be interpreted as evidence of rapid sedimentation. Thus in a rock section, time may be both condensed and stretched. A rough measure it may be, but it permitted Merrill to consider how rapidly the sea had transgressed the land. He found this often occurred so rapidly that it prevented the shale fauna from developing locally. The shallowing (regression) of the sea caused by the growth of a delta was, however, slower, and in these circumstances the expected fauna could develop. Merrill thought he was looking at shallow and very shallow water communities that were not greatly distinguished in terms of water depth. Instead, he believed salinity a controlling factor as the shales contained brackish-water fossils while the limestones were fully marine. He deduced that the animals actively maintained a link to the environment and did not, like chaetognaths, simply waft around in the currents: “The conodont governed its own occurrences in a much more direct manner…. It was nektonic; an active organism fully capable of exerting important controls upon its own distribution, depth and destiny.” Seddon and Sweet's model simply did not fit the bill.