Of Embryos and Transmutation IV – there and back again

By Robbert Striekwold

Throughout the 19^th century, ideas concerning embryonic and species development were joined together in a unification that many biologists took to be self-evident. Near the end of that century they split up, however, and embryology was largely left out of the Modern Synthesis of Darwin and Mendel. That is, until several spectacular advances in developmental biology allowed for a re-integration of the two fields a few decades ago. This remarriage has led to such enthusiasm that some practitioners now speak of a ‘More Modern Synthesis’, unifying all of biology under the banner of developmental evolution.[1]

Entwicklungsmechanik

The field of ‘evolutionary embryology’, which had sprung from the work of Darwin and Haeckel, was immensely popular during the mid-1880s. This was soon to change, however, with the rise of Entwicklungsmechanik (developmental mechanics) in Germany. Led by embryologists such as Wilhelm Roux (1850-1924), one of Haeckel’s students, this movement sought mechanical (rather than historical) explanations for developmental phenomena, and quickly garnered much following.[2]

Using frogs as model organisms, Roux performed a series of experiments with which he hoped to show how hereditary material was distributed in early development. In his most famous experiment, he took a frog embryo in its two-cell stage and killed one of the cells, leading to the development of only half a tadpole. This led to a heated debate with his colleague Hans Driesch (1867-1941), who had separated the cells of sea-urchin embryos at the two-cell stage, ending up with somewhat smaller but nonetheless fully formed larvae.[3]

Hans Spemann (1869-1941) would improve upon such experiments by tying a baby’s hair around the cells instead of separating or killing them, creating two compartments between which some exchange of material was possible. Depending on the orientation of the hair, this would lead to two complete tadpoles, one tadpole and a mass of undifferentiated tissue, or a two-headed tadpole (figure 1). This approach made it possible to find out how the planes and axes (up-down, left-right, front-back) of the embryo are determined early in development, without ever needing to turn to evolution.[4] If, for example, one ends up with a two-headed tadpole, the hair must have constricted exchange between left and right, leading to the growth of a head on both sides.

The Modern Synthesis

The school of Entwicklungsmechanik became increasingly popular during the late 19^th and early 20^th centuries, largely to the detriment of the Haeckelian school of evolutionary embryology. Meanwhile, the early 20^th century also witnessed the birth of Mendelian genetics. One of the chief distinguishing features of this approach was its particulate view of organisms and their underlying hereditary factors, seeing the organism as a bag of loosely connected characteristics rather than an integrated whole. The British geneticist William Bateson (1881-1926) neatly summed it up when he wrote: “The organism is a collection of traits. We can pull out yellowness and plug in greenness, pull out tallness and plug in dwarfness.”[6] Experiments were done with fruit flies to investigate the behaviour of genes and chromosomes, assuming a nice one-to-one mapping of traits onto genes (figure 2).[7]

With the Modern Synthesis of the 1930s and 1940s, Mendelian genetics was wedded with Darwinian natural selection to become a supposedly unified theory of evolution. Developmental biology was left out almost entirely, for two main reasons. First, Mendelian genetics basically ignored embryos as very complicated stages between the easily quantifiable genes and their conspicuous effects in adults. Second, most embryologists, while accepting evolution, rejected natural selection, and were more interested in the mechanical causes of development than in the historical explanations offered by evolution.[9]

Heterochrony

This is not to say that, during the first decades of the 20^th century, no one tried to open the embryonic black box and integrate evolution and development; just that none of the attempts really caught on. Historians often (somewhat artificially) date the resurgence of research combining evolution and development to the publication of Ontogeny and Phylogeny by the American palaeontologist Stephen Jay Gould in 1977. In this book, Gould argued that concepts and processes from embryology could be hugely important for the study of evolution.

Gould focused on the concept of heterochrony, which involves variations in the timing and rate of developmental processes between ancestor and descendant, thus leading to evolutionary change. One example is the salamander Axolotl, which retains certain juvenile features (such as gills and a large, larval tail) into adulthood, which other salamanders lose as they mature. A very intriguing possibility (to many) was the idea that this could be applied to humans. Chimp and human infants, it was observed, look very much alike, the adults much less so (figure 3). Maturing chimpanzees experience a significant warping of the skull, with the jaw protruding outwards and the skull dome pressing down; humans do not. The idea, then, was that human skull development had slowed down dramatically since it split from the lineage leading to chimpanzees.[10]

Evo-Devo

But the real ‘revolution’ came with the increased involvement of geneticists. When molecular genetics arrived on the scene in the 1960s and 70s, developmental biologists suddenly had at their disposal a new set of concepts and techniques that allowed them to study processes of embryonic development with a level of detail never seen before. Development could now be studied by inducing targeted genetic changes in organisms and studying the effects in adults. One spectacular experiment involved the expression of the mouse gene Pax6, which functions in eye development, in the antenna region of fruit fly embryos. This led to the development of adult flies with an extra set of faceted eyes, in the place where their antennae would normally have been (figure 4).[12]

The black box of the embryo had thus been opened, and during the 1980s and 90s developmental biologists discovered an entire toolkit of genes controlling the architecture of the embryo. Now, the new field of evolutionary developmental biology (or “evo-devo”) could contribute its own mechanisms to the theory of evolution, and it was quickly argued that the Mendelian genetics of the Modern Synthesis was too simplistic. Moreover, based on the observation that the genetic toolkit genes were remarkably similar in organisms as distantly related as mice and fruit flies, developmental biologists argued that most evolution occurred in big jumps resulting from subtle changes in these toolkit genes, with Darwinian natural selection being a marginal force in evolution.[14]

A More Modern Synthesis?

The volume of this rhetoric has waned a bit over the past decade, and commentators on the study of evo-devo now call for a “more modern synthesis” rather than an entirely new one. Where this will go in the future is anybody’s guess, and at any rate not what I want to comment on here. The point I want to make is that few biologists now will want to argue that a mature theory of evolution can be constructed without major input from the embryonic development. And even though few (if any) will assent to the belief that development and evolution are essentially the same thing (see part 2 of this series), or that they are parallel processes (see part 3) – it cannot be gainsaid that the strong connection between them has once again become undeniable.[15]

o-o-o

Robbert J Striekwold is a M.Sc. student in the History and Philosophy of Science programme at Utrecht University, specializing in the history and philosophy of evolutionary theory. He is writing his thesis on conceptual issues in modern evolutionary developmental biology.