The palimpsest theory

by Greg Mayer

One of the things Jerry mentioned in introducing me was that I had coauthored, with my late friend and mentor John A.W. Kirsch, a paper entitled “The platypus is not a rodent”.  While there’s a certain pure amusement value in such a title (which alludes to a series of papers concerning the relationships of guinea pigs with titles such as “Is the guinea-pig a rodent?”,  “The guinea-pig is not a rodent”, and “Are guinea pigs rodents?”; btw, the guinea pig is a rodent, and the platypus is an egg-laying monotreme, nowhere close to rodents), Jerry might have mentioned the paper because of its subtitle: “DNA hybridization, amniote phylogeny and the palimpsest theory”. In WEIT, Jerry likens the bodies and genomes of organisms to palimpsests. In ancient and medieval times, parchment to write on was expensive, but writing was cheap. To save parchment, the writing would be scraped off a book (for in those days, all books were written by hand), and a new book written over the old. Such reused parchments are called palimpsests. The original writing, however, can often be seen or retrieved, and thus the history of the parchment’s uses can be inferred from the existing parchment. (Here’s an example mentioned by Jerry.)

A palimpsest possesses both recently acquired features (the new writing) and remnants of old features (the old writing).  So do organisms. They possess immediately adaptive characters, as well as characters from earlier in their history.  This has long been recognized, and the analogy with palimpsests has been explicit. In 1910 W.K Gregory of the American Museum of Natural History made a distinction between ‘caenotelic’ and ‘paleotelic’ characters; he later called these ‘habitus and heritage’ (Gregory was the major professor of my major professor, E.E. Williams, and thus I am Gregory’s academic ‘grandson’). In 1947, Gregory christened his ideas the ‘palimpsest theory’.  As John and I explained:

Habitus characters (equivalent to caenotelic features) become in time transformed into, or at least included among, those of heritage as the collected adaptive wisdom of the lineage at more general levels, by a process of sequential adaptation …. Habitus and heritage are thus ‘correlative terms’, so that ‘the remainders of the successive habitus of the remote ancestors become incorporated into the heritage of later times’ (Gregory 1947, p. 8). Heritage features are therefore of utmost importance in determining the broad affnities of a higher-category taxon, because they may be ones shared with a similarly inclusive but different group.

Thus the habitus characters are the new writing, the heritage the old. The palimpsest analogy had been published earlier by the great South African paleontologist Robert Broom (best known for his later work on australopithecines, the predecessors of our own genus, Homo) who in 1924 wrote about turtles that

Unfortunately members of the order are all extremely specialized and in some respects degenerate, so that the picking out of the ancestral [=heritage or paleotelic] characters amid the more recent specializations [=habitus or caenotelic] is somewhat like the reading of a difficult palimpsest.

Broom and Gregory were well-acquainted with one another, and Broom visited the American Museum in 1913-1914, so the use of the term palimpsest by the two of them is probably not independent.

Not all historical processes leave clear traces of their paths: if a ball rolls downhill to a resting place, we cannot infer from where on the surrounding heights it began; and one molecule of water is just like another (of the same isotopes), no matter whence it came.  We are fortunate that descent with modification is a history-conserving process: bodies and genomes of organisms are documents of evolutionary history. As the many examples in WEIT show, even when a fossil record is lacking, we can learn much about an organism’s evolution.