Naturphilosophie received considerable pushback with the rise of classical embryology. On the continent, scientists like Christian Pander, Karl Ernst von Baer, and Heinrich Rathke, all helped to define and expand the field of embryology and reduced the a priori, transcendental approach to morphology that Naturphilosophen such as Carus embraced (Churchill, 1991 1991). Most of these embryologists worked on accessible specimens, such as sea urchins, chickens, and mice, to understand laws and mechanisms driving the formation of new morphologies. And, although tortoises were difficult to obtain and to keep alive, some early embryologists and zoologists used them to refine and test their ideas about development as well as evolution. Most notably Martin Rathke and Richard Owen with their embryological investigations looked to the Chelonians for a fresh perspective on their work.
Rathke and Owen shared an epistemic commitment to understanding the development and diversity of organismal form through reference to embryological processes; however, their goals and conclusions about the turtle's shell were vastly different. Owen, unlike Rathke, was tied, like Cuvier, Geoffroy, and Carus, to a theory that could unite organismal form. For Owen, his archetype theory was the explanatory basis of all vertebrate diversity (although the mechanisms of producing diversity changed over time, as we shall see). This framework both drove his research, and colored his explanations of form—as we will see with his interpretation of the turtle's shell. Meanwhile, Rathke signifies a break with the tradition of grand theories of nature—his embryological investigations were in search of laws of development and thus mechanisms underlying common morphologies (Rathke, 1825). In this respect, Rathke is a precursor of important traditions in the history of embryology such as the later Entwicklungsmechanik (Churchill, 1991 1991).
Rathke's Comparative Embryology of Turtles
While Karl Ernst von Baer may be the most well known of the proto-embryologists on the continent, Martin Rathke (1793–1860), a professor of zoology and anatomy at the University of Königsberg, is more rightly considered the founder of comparative embryology (Churchill, 1991 1991). Von Baer still embraced the transcendental principles of Naturphilosophie, and was concerned with generalizations in embryology and taxonomy (Churchill, 1991 1991). On the contrary, Rathke was a “humble observer” rather than a bold theorizer (Menz, 2000; 217)—his attention was on the details of individual development. According to Rathke, explaining a morphological trait meant following it throughout development and reconstructing the embryological processes that produced it (Menz, 2000). The impressive 1848 monograph “Über die Entwicklung der Schildkröten” dealing with the development of the turtle shell was borne out of the epistemic commitments of the new embryology: attention to the details and focus on embryological mechanisms underlying the emergence of any morphological trait. Rathke's 1848 text shifted the attention of turtle studies from anatomy to embryology. Following the publication of this work, it became clear that explaining the puzzling morphology of Chelonians could not rely on the anatomy of adult specimens, but had to address the underlying developmental mechanisms. As we will see, this approach drastically changed the field of Chelonian studies.
Like many scientists on the continent at the time, Rathke started his career under the influence of Naturphilosophie. However, only brief glimpses of this approach can be found in his works. In his early studies, such as “Beiträge zur Geschichte der Tierwelt” (1825), Rathke admitted that most ideas coming from natural philosophy were just dreams (Traumerei) that could inspire thinking and not metaphysical commitments about the real nature of the biological world. From the start, Rathke's works were largely focused on tracing through empirical observations, and in meticulous detail, the development of organisms. By tracing individual development, Rathke hoped to discover the laws, according to which different forms in nature emerge and take shape. Of the scientists covered thus far, Rathke is the first to truly embrace a mechanistic understanding of the constitution of form. His meticulous approach to chronicling the details of development led to immensely dense treatises, and also helped to establish the importance of looking to embryos and development to understand morphologies (Churchill, 1991 1991).
Rathke started his research on turtle embryonic development when he moved to Königsberg, by the Baltic Sea. He found that some lakes in southeastern Prussia supported thriving turtle populations. In a letter to his publisher, Friedrich Vieweg, Rathke described why he cared about turtles as well as why anybody interested in morphology should care about them (Menz, 2000). He explained that, although past masters of biology had acknowledged the importance of understanding turtles both because of their unusual morphology and because of their interesting place in the history of life on earth, scientists did not actually know much about turtle morphology. Rathke found a new way to fill this gap. He wrote “… research about the development of these animals will lead to more satisfactory explanations of their wondrous and puzzling body structure, than the ones we have produced so far by dissecting adult specimens” (Rathke, , 756). In these words lies the shift from the anatomy to the embryology of turtle morphology. Rathke showed that, if scientists, wanted to explain how such a puzzling morphology comes about, they had to pay attention to its embryonic development. In an important passage referring to Carus's work on turtles, Rathke wrote: “Carus set up the above mentioned view [the mixed hypothesis] on the basis of studies of the skeletons of adult turtles: but such can not be held for sufficient, instead we need here further confirmation or correction by the developmental history (Rathke, 1848; 115).” Here Rathke clearly criticized Carus's ideas not just because they are incorrect. Rather he argues that Carus's approach, similar to the approach of all the other scientists who had investigated turtle morphology, was not adequate, as it was not informed by the mechanistic understanding of developmental processes.
Rathke published his first preliminary results on studies of embryonic turtles nearly 15 years after Carus developed the mixed hypothesis, in his 1846 short essay, “Vorläufingen Bemerkungen betreffend die Entwicklung der Schildkröten.” Years of experience impressed upon Rathke the difficulty of amassing sufficient evidence about embryonic development, and so, his efforts to find and grow turtle embryos were quite impressive. Von Baer, also an inhabitant of Königsberg at the time and a close friend of Rathke had already addressed the difficulties of obtaining specimens of tortoises (Von Baer, 1834; 546). Rathke first decided to get 100 eggs of Emys europea sent from other regions of Germany, as it was extremely difficult to obtain viable specimens in the Königsberg region (Menz, 2000). Such eggs, unfortunately, did not give rise to any embryos and all his attempts to have the eggs develop into embryos did not work, likely because of “stress” during transportation. Rathke then decided to have specimens sent from other countries—eventually yielding two viable embryos of Emys europaea. Eventually, in June 1847, Rathke finally obtained three recently laid eggs from different species: Testudo gracca; Chelonia midas, and E. europea. Using these three embryos, Rathke was finally able to observe development from the beginning to the middle time of their embryonic development, but no further, so more specimens were ordered to extend his investigation to more advanced stages.
In his embryological observations, Rathke found that no osseous plates developed independently in the corium that would coalesce afterward with the neural spines and ribs. According to him, the neural plates were expansions of the spinous processes of the vertebrae and the costal plates were extensions of the greater tubercle of the ribs (Rathke, 1848). Therefore, ribs and vertebrae build the carapace, and the carapace in combination with the scutes forms the external shield. In other words, according to Rathke, there are no osseous plates that develop independently in the dermis and afterward coalesce with the neural spines and ribs (Rathke, 1848). This is why he concluded that the carapace of Chelonians is composed exclusively of endoskeletal elements: “The spinous processes are already developed from the second to the eighth dorsal vertebrae before the exclusion of the embryo, […] they remain pretty short, but contrary to the general laws of development of the vertebrate animals, they grow so much in breadth, that they form, after their ossification, a series of horizontal plates of moderate size” (1848). In his 1848 work, Rathke also held that the plastron has exoskeletal origins. He refers to the various parts of the plastron as “supplementary plates” and takes them for equivalent to the Nuchal, the Marginalia, and Pygalia in the Carapace.
In order to support his hypothesis about the endoskeletal nature of the carapace, in his 1848 monograph, Rathke made an unprecedented use of comparative analyses and is therefore also considered one of the founders of comparative embryology. In “Über die Entwicklung der Schildkröten,” he wrote first about the “Characteristics of the Eggs” and first stages of development of the tortoise E. europea (Part 1). Then, he addressed developmental formation of the carapace in several kinds of turtles both from his observations and from previous studies, from E. europea to T. gracca, Chelonia virgata, and Tryonix aegiptiacus (Part 2). At the end, Rathke focused again on E. europea (Part 3). He showed that similarity in adulthood was not conclusive evidence of shared origins (Rathke, 1848; Agassiz, 1857). Among the firsts, Rathke emphasized the importance of understanding comparative embryology. The process of development for understanding evolutionary trajectories would become a major factor in the future of biology as well as in the study of turtle morphology (Goette, 1899).
Owen's Archetypes and His Turtles
While Rathke focused on detailing individual development over the creation of elaborate theories of nature's unity, Richard Owen (1804–1892), the man whose intensive works would establish the Natural History Museum in London, took a different stance. Throughout the 1840s, Owen established within his anatomical treatises an understanding of the unity of type that Geoffroy had embraced, lending his vast knowledge of the natural world to the creation of the archetype theory through inductive reasoning (Fig. 3). Owen's archetype, outlined in two major texts, On the Archetype and Homologies of the Vertebrate Skeleton in 1848 (Owen, 1848) (Owen had discussed the archetype in several earlier contexts, but did not fully explicate his theory in print until 1848), and On the Nature of Limbs in 1849, was highly influential to both the scholarly community and his own peripheral works, as we will see with the case of Owen's work on turtles. And, it is Owen's work on the archetype theory that ties him to the construction of grand theories of natural systems that his predecessors, like Cuvier, Geoffroy, and Carus, had attempted, but Rathke had rejected.
In late 1848, Owen submitted his manuscript, “On the Development and Homologies of the Carapace and Plastron of the Chelonian Reptiles” to the Royal Society of London. The text was read to the society in January of 1849. Owen's work on turtles, based off of careful dissections of at least six species of turtle, contrasts heavily with Rathke's both in terms of what he perceived to be the developmental origins of the carapace and plastron, and how he arrived at these conclusions.
By way of introduction to the issue of how the carapace and plastron develop, Owen offered his readers a review of previous works on the subject. From Geoffroy to Cuvier and Carus, Owen considered the theories of the turtle's shell and found the choice between any of them to be difficult to accept based upon the merits of their evidence, because they addressed a question of development with evidence gleaned from adult specimens. On this note, Owen tells us, “The guide to our choice of either of these [theories of carapace/plastron development], or of any other view that has been offered of the nature and signification of the thoracic-abdominal case of the Chelonia, must be the light afforded by a true perception and explanation of the phenomena of its development” (Owen, 1849a; 158). Thus, Owen sought his own evidence and explanation of shell development through embryology.
Through dissection and comparison of the embryos (and adults) of several species of turtles, Owen reached the conclusion that the carapace is a mixed structure—a finding reminiscent of Carus, but radically different in its inception. In his work, Owen looked at the development of the neural, costal, and marginal plates as separate phenomena. He identified eleven neural plates (this is in keeping with previous authors), and lumped the nuchal plate in with this classification. Neural plates 1–8 Owen saw as serial homologues, each beginning as what Owen referred to as “dermal cartilage” (this term most likely refers to dermal connective tissue), but only becoming ossified once the neural spine of the subadjacent vertebra extended into the dermis. Thus, neural plates 1–8, according to Owen, were endoskeletal, because their ossification was an extension of the underlying vertebrae. Meanwhile, neural plates 9–11 and the nuchal bone, also began as dermal cartilage, but continued their development as independent centers of ossification. Because, Owen reckoned, these plates do not ankylose with the vertebral spines, they must be considered dermal bone, like that of a crocodile, and are thus exoskeletal.
While Owen thought the neural plates to be mixed in their composition, he saw the costal plates as purely endoskeletal. These plates began as dermal cartilage, and became ossified as the process spread from near the head of the underlying rib into the costal plate. While the diagnosis of endoskeletal costal plates is in agreement with the views of Rathke, the mechanism through which the ossification occurs was different enough for Owen to spend a great deal of time to explain. To Rathke, the costal plates could be considered as outgrowths of the greater tubercle of the ribs, just as the neural plates could be considered as expansions of the spinous processes of the vertebrae. To Owen, however, the plates could not develop simply as outgrowths of ribs for the following reasons: (1) the placement of the costal plates over the ribs is variable (in fact, he found that the placement of the plates was more closely aligned with the development of the scutes), and (2) within the vertebrates, the greater tubercle of the ribs is not variably placed (i.e., it always develops from the true neck of the rib). It follows that if the placement of the tubercle is invariable, but the placement of the plate over the tubercle is variable, then the plate is not a simple outgrowth of the tubercle. Thus, Owen and Rathke agreed on the developmental origins of the costal plates, but not on the mechanism of their development.
When it came to the issue of the plastron, Owen's interpretation, based on dissections of embryonic tortoises and adult Trionyx specimens led him to the conclusion that the plastron is an endoskeletal structure (Fig. 4)—a finding that directly contradicted Rathke's exoskeletal determination. Additionally, Owen saw evidence in the development of the tortoise to hypothesize that the nine pieces of the plastron are modifications of the sternum (entosternum [s] and episternals [es]) and the haemapophyses (hyosternals [hs], hyposternals [ps], and xiphisternals [xs]) found within the larger vertebrate clade. Owen's finding of homology between the plastron and the sternum resonates with the determination of Cuvier (1799, 1819) and Geoffroy-St-Hilaire (1809), but Cuvier and Geoffroy had not made the comparison of some of the plastral plates to the haemapophyses (sternal ribs/costal cartilage). Burke (1989) has claimed that “Owen's interpretation of the plastron was influenced by his own theory of the archetypal vertebral segment, and he homologized the plastral bones with thoracic haemapophyses … (Burke, 1989; 364)” which is a point that bears further discussion here.
Figure 4. Owen's diagram of plastron of Chelone caouanne (modern day Caretta caretta). From Owen (1849a); Figure 3, pg. 153. es = episternal, s = sternal, hs = hyosternal, ps = hyposternal, xs = xiphisternal.
Download figure to PowerPoint
Taking a closer look at Owen's archetype concept, recall the previous discussion of Carus' Urtyp from 1828—Owen borrowed heavily from Carus to develop his own archetype (although he was wont to admit it). Owen, much like Carus, considered the archetype to be the most elementary form of vertebrate. The data, or evidence, upon which Owen built his archetype was similar to the epistemology proposed by Geoffroy—that of homologies (what Geoffroy had referred to as “analogies”). The principle of homology (a term that Owen coined, and which helped to standardize the anatomical lexicon) meant that the significance of an organ or structure was derived from its place in the make-up of the whole organism (recall Geoffroy's “Principle of Connections”). This structuralist understanding of anatomy stood at odds with the functionalist view that dominated some of the scientific and social circles in which Owen was embedded in England. These circles, most notably scholars at Oxbridge, like Adam Sedgwick and William Whewell, grew increasingly wary of Owen's homological program, even though his inductive approach to anatomy gave transcendental morphology a systematic basis that it had thus far lacked (Rupke, 1994).
The transcendental turn of Owen's comparative anatomy, coupled with his embracing the archetype as an idealized generalization (and not an actual entity) strained his relations with Oxbridge scholars during the late 1840s, such that, by the time his On the Nature of Limbs came out in 1849 (1 year after the archetype text), Owen had reversed his stance on the archetype. Whereas previously, and in keeping with the Naturphilosophen, Owen viewed the archetype as the result of a material force that propelled nature to differentiate, by 1849, he expounded the archetype as a metaphysical entity that was the blueprint for the formation of animal life (Rupke, 1993). This new vision of the archetype is evidenced by his reference to it as a “predetermined pattern” in On the Nature of Limbs (Owen, 1849b; 2). This new instantiation made Owen's archetype concept teleological, and although Owen continued his research in a mostly structuralist vein, his reincarnation of the archetype put that concept in line with both the Oxbridge functionalist's and Cuvier's highly teleological worldview.
This volte-face has interesting implications for the context of Owen's work on turtles. To begin, Owen's turtle treatise is sandwiched in time between his two texts on the archetype, and contains evidence of his pull from both sides (structuralism and functionalism). At first, we see Owen giving evidence against a functional interpretation of the carapace, “… their presence [the costal plates] seems to be determined rather by the angle of the union of the superincumbent vertebral scutella with the lateral or costal scutella, than by the necessity for additional strength in the articulation of the ribs with the spine (Owen, 1849a; 162).” Here Owen refers to the previous discussion of the placement of the costal plates—which he sees as being more closely aligned with the development of the overlying scutes than with the underlying ribs. Despite this, Owen also employs functionalist interpretations: “The main purpose of the augmentation of the ordinary vertebral elements in the thoracic-abdominal region of the Chelonia, by the extension of ossification from them into the corium, and the consequent connation with those elements of dermal bony plates, being the formation of a strong defensive abode …” (Owen, 1849a; 165) [Emphasis is ours]. This passage shows that at the time of writing his turtle treatise, Owen's interpretations were caught-up in the transition that his own archetype concept was making—from structuralist and transcendental, where the cause of structural differentiation from the archetype was the presence of a material force, towards teleological explanations, where structures are explained by the purpose they serve.
Owen's commitment to the archetype concept may have colored his interpretation of his evidence regarding the issue of the plastron. Although Owen adopted a different methodology and forms of evidence than Cuvier, Geoffroy, and Carus, his works were driven, much like theirs, by an underlying theory of natural unity and the requirement that all nature somehow fit this unity. While his turtle treatise is not an overt overture to the archetype theory, the treatise is firmly entrenched within that theory. Owen bases the interpretation of his evidence on an understanding that the turtle's morphology must be a modification of a more basal plan, as evidenced in the following quote:
The most obvious, and, I believe, the most natural explanation of this first complete segment of the thoracic-abdominal region of the young Tortoise, according to the typical vertebra, and the composition of the corresponding segment in the nearest allied Vertebrata, is—that the centrum, the neural arch, and the pleurapophysis, are parts so indicated by the initial letters; that the hyosternals are the haemapophyses (sternal ribs or costal cartilages), and the entosternum is the ‘haemal spine' or sternum proper (Owen, 1849a; 163).
In this quote, Owen is telling his reader that the only way to account for the turtle's peculiar morphology is to envision that some parts—namely those of the plastron—are extensively modified homologues of extant vertebrate parts. In this theoretical scheme, the entire turtle body plan can be thought of as a revised form of the archetypal vertebrate, and it is this systematic comparison and understanding of modification from a simple, shared vertebrate structure, that drives Owen's understanding of the natural system and interpretation of evidence.