The science of paleontology has long been concerned with searching out the origins of modern groups of animals, but at the turn of the 20th century there were frustratingly few transitional fossils. That evolution had occurred was generally agreed upon, but where the transitional forms might be found, what they would look like, and what mechanisms drove their evolution remained disputed. Among the murkiest of these subjects was the origin of mammals.
In an 1898 letter published in the journal Science the American paleontologist O.C. Marsh reviewed the problems involved in studying the origin of mammals. Chief among the factors that aggravated paleontologists was a lack of transitional fossils. The origin of fishes, amphibians, reptiles, birds, and mammals (the five great vertebrate classes everyone learns about in elementary school) could not be determined without more evidence, and “The answer to each [group's origins],” Marsh wrote, “belongs to the future, and how far it may profitably be sought in the present limited state of our knowledge is a fair question in itself.”
Marsh could not simply set the question aside, however. The origin of mammals was a mystery but this did not mean that he could not review a few clues that might eventually be of benefit to ongoing discussions. In order to answer the question, though, naturalists would have to determine whether mammals constituted a single, natural group or if different groups of mammals had separate origins. The group as a whole seemed disparate enough in form that Marsh could not say with certainty whether all mammals had a single common ancestor or not.
Whether they had a single origin or several, however, mammals were most popularly considered to have evolved from reptiles. This was in spite of the fact the fossil record seemed to only contain creatures that could be called “mammals” or “reptiles.” If someone discovered a mammal with multiple bones in its lower jaw (a tell-tale reptilian feature) rather than a single one, that would be something, but Marsh claimed to have seen every extinct mammal ever discovered and found no such evidence. If mammals had evolved from reptiles then paleontologists had not yet found the intermediate forms to confirm the hypothesis.
Perhaps paleontologists had been looking in the wrong place. Marsh suggested that the resemblances between reptiles and mammals were only “adaptive likeness[es]“, otherwise known as evolutionary convergence. Maybe they did not share a close evolutionary relationship after all. There was another group that showed some resemblances to both reptiles and mammals, though. These were the earliest tetrapods (what Marsh called the “earliest amphibians”), and they seemed to be better candidates for the progenitors of mammals. Marsh concluded;
In the amphibians, especially in the oldest forms, there are hints of a true relationship with both reptiles and mammals. It seems to me, therefore, that in some of the minute primitive forms, as old as the Devonian, if not still more ancient, we may yet find the key to the great mystery of the origin of mammals.
This implied a ghost lineage of mammal-like amphibians (entirely bypassing any reptile-like form) that had not yet been found, but it seemed like a potentially fruitful hypothesis if extinct reptiles failed to bridge the gap to more mammalian types. This would have to be confirmed or refuted by further research, but it at least offered a fresh start to a somewhat beleaguered debate.
Today much of what we know about the origin of mammals has changed, and Marsh’s hypothesis has not held up. There were no “mammal-like amphibians.” Instead mammals are synapsids, part of a larger group including stem-mammals like the sail-backed Dimetrodon. Indeed, during the Permian and Early Triassic there were many synapsids that looked a lot more like “reptiles” than living mammals, yet they were more closely related to us than lizards or crocodiles. Some of these creatures were just the sort Marsh was looking for and even known during the time he wrote his letter, but for one reason or another they were considered to be more “reptile-like” than “mammal-like.” Today we known better, but even now there are many questions about the evolution of “true” mammals from within the synapsids that survived the Permian mass extinction. Details of this transformation will rely on future fossil evidence.