Last updated on 21 Jun 2018
Notes on Novelty series:
2. Historical considerations – before and after evolution
3: The meaning of evolutionary novelty
4: Examples – the beetle’s horns and the turtle’s shell
5: Evolutionary radiations and individuation
6: Levels of description
8: Conclusion – Post evo-devo
The Beetle’s horns
Beetles often have projections on their carapace for various functions, but one group in particular, the scarab beetles, which includes the rhinoceros beetles, have exaggerated and complex structures that are often thought of as novelties. Darwin discussed them in the Descent and explained them in terms of sexual selection (a view that is still accepted), but the mechanisms by which these weird structures occurred were unclear until recently.
These horns are used in mating combat, and occur on the males. The can account for as much as 30% of the body weight of the beetles, and interfere with their ability to fly; the beetles that occur in this group in Australia, with which I am familiar, sound like helicopters as they settle on your hair at night. They can grow very large indeed. Darwin focused on a particular species, the Atlas Beetle Chalcosoma atlas, from South Asia, which can grow to 130 mm or over 5 inches.
The Atlas beetle.
Various beetle horns, from Darwin’s Descent of Man.
Now from one perspective, a beetle horn is a novel structure, unquestionably. It is a structure that is not found in relatives of this group, and which clearly has novel functions in mating behaviours. From these two facts it can be inferred (but, let it be noted, not asserted as known by observation) that the common ancestor of all beetles did not have them, and so they evolved. They are, as Calcott suggested, exemplars of evolutionary novelties.
But what is most interesting in this case is that the underlying developmental and genetic systems are not novel, but modified in minor ways (mostly to do with gene expression regulation and timing). Shubin et al. show the following figure:
The top image is the limb of a drosophilid fly, and the gene expression in each segment. The lower image is the horn of an Onthophagus dung beetle, part of the group we are discussing, and the expression of the very same genes. Drosophilids and dung beetles have an evolutionary separation of some 300–355 million years, and yet these genes are not only conserved, but employed for wildly different functions and structures even in the same broad group, insects.
What counts as a novel structure at one level has homologous underlying mechanisms. Of court, we still need to account for why these novel structures evolved, which is another way of asking why all beetles, and indeed all organisms that share these four genes, do not have these structures.
The turtle’s shell
For some time the turtle’s shell has been a problem for evolutionary accounts. The shell is formed from skeletal structures that in what we think are the closest “reptile” relatives of turtles and tortoises (Testudinae), are inverted. For example, the scapulae (“shoulder bones”) of ordinary tetrapods like us and lizards are external to the rib cage and anchor the limbs with muscle attachments to ribs and the vertebrae from the outside. In turtles, the scapulae are internal to the rib cage, and the muscle attachments are also internal, to the ribs and the plastron (the lower part of the shell). It is a puzzle debated since the days of Cuvier (1800) and Carus (1834, refs in Nagashima et al.), and which has played a substantial role in recent debates about development, evolution and novelty.
Here is the oddity of turtle anatomy, from the Nagashima et al. paper (figure 2A):
The scapula is sc, the other features are osteoderms (os), neural plates (net), costal plates (cos), carapacial dermis (cd), ribs (r), and the vertebra (v).
There are several hypotheses how these evolved (again, not why, as this is assumed; carapaces are for protection against predators). The two main ones are Rucke’s 1929 hypothesis where the shell evolves to fold over the scapula (ax = axial domain, lbw = lateral body wall, cor = coracoids):
and Burke’s 1989 hypothesis where precursor cells for these structures migrate early in development (cr = carapacial ridge, fl = forelimb, rpc = rib precursor cells, spc = pectoral girdle primordium, nt = neural tube, n = notochord):
The evidence is in favour of Rucke’s hypothesis: changes to muscle attachments from the vertebrae to the ribs are a side effect of changes in the expression of the carapacial ridge, blocking the formation of vertebral attachments. In effect, the turtle morphology and shell is the topological shift, a folding during development, as shown by comparing early and late formation in a bird (chicken) and the turtle:
What looks to be an evolutionary novelty is in reality a shift in expressions of genes and one novelty, the development of the carapacial ridge, which subsequent selection for ectoskeletal plates. The authors say:
These anatomical and embryological analyses have changed the former perspective on turtle evolution, from an inconceivable morphology made by a simple developmental change to a probable morphology achieved through many turtle-specific ontogenetic processes. The latter idea does not always have to presume saltatory evolution; instead, a stepwise evolutionary path is probable.
What is interesting here is that developmental systems play a crucial role in uncovering the evolution of the turtle’s shell, but the work is not based on the Muller-Wagner notion of unique developmental modules; in fact the work has uncovered quite standard genetic expression and developmental effects.
Deep homology and subjectivity
Now this might seem to be a reductionist argument, but in fact it is about what level the description of characters must be for something to count as a homology. At the level of gross morphological or phenotypic description, the scarabid horn is not homologous to the general insect limb (one supposes the same genes are expressed in scarab limbs also). But at a deeper level, they are; as Shubin and colleagues term it, these are deep homologies. They are developmental systems or modules that are co-options of prior systems. The novelty here lies in the identification of the structures and functions; in an observer independent manner, these are just Arthurian phylogenetic apomorphies. The novelty lies in the salience of these apomorphies to observers, where other apomorphies are not so salient.
Evo-devo accounts like Muller’s and Wagner’s is no less vulnerable to the subjectivity issue than the evolutionary: since all aspects of development have underlying mechanisms, the choice of which mechanisms are to be regarded as “non-homologous” depends critically on what descriptions are given in order to make the traits or characters novel. This will be the subject of the next post in this series.
Note: I am not a biologist, but a philosopher, so I may very well have got the details wrong. Whether or not I have, the real issue is whether this materially affects the argument being developed.