Historical and Philosophical Perspective Darwin’s Theory of Descent with Modification, versus the Biblical Tree of

David Penny* Institute for Molecular BioSciences, Massey University, Palmerston North, New Zealand

Thereisastrongurbanmyththat to explain events in the past—an ap- Darwin’s Alternative View of introduced and/or ad- proach that I interpret as being learned ‘‘Descent with Modification’’ vocated a ‘‘’’ for the classi- from his geological background (Figure 1) fication of living organisms. This has [6,7]. Having established that Darwin was not recently been highlighted by Lawton [1], the originator of the term, my second but dates back to at least Doolittle [2]. It How Old Is the Tree of Life main point is that, instead of using the tree is often combined with the idea that we Concept? of life concept, Darwin referred to his require a ‘‘paradigm shift’’ to take into theory as ‘‘descent with modification’’. For account extensive lateral gene transfer, The Tree of Life is a biblical phrase that example, in the first edition of the Origin of especially in prokaryotes. Yes, extensive predates the 19th century by several (Darwin 1859, see http://darwin- lateral gene transfer occurs widely in millennia. In all, the phrase occurs 11 online.org.uk/) Darwin uses the phrases prokaryotes (see [3–5]), and this enriches times in Genesis, Proverbs, and the Book ‘‘theory of descent’’ and/or ‘‘descent with our understanding of by em- of Revelations (http://www.biblegateway. modification’’ 21 times (on pages 12, 14, phasizing a more gene-centered view. In com/quicksearch/). In Genesis, it is found 189, 206, 320, 349, 351, 354, 358, 361, contrast, though acknowledging the prior three times (2:9 and 3:22 and 3:24); in the 399, 429, 444, 477, 479, 484, 493, 494, availability of the tree of life simile, middle of the Garden of Eden were both 496, and 497). As such, it includes the idea Darwin continually referred to his ‘‘the- the Tree of Life and the ‘‘Tree of the of an evolutionary tree but also includes ory of descent with modification’’, an knowledge of good and evil’’ (from which hybrids (long a standard technique for expression that encompasses a wide Eve supposedly ate the fruit). Eating of the plant breeders), of which Darwin was fully variety of fundamental processes includ- fruit of the Tree of Life would apparently aware, as well as ideas from genetics that ing both vertical descent and lateral gene have given eternal life, and that eternal life would have been unfamiliar to him. These transfer, thus removing any need for theme appears again in Revelations 2:7, include mechanisms of evolution such as paradigm shifts. and 22:2, 14, and 19. The other references the transfer of genes between bacteria, or For Western Europe, the Tree of Life are in Proverbs (3:18, 11:30, 13:12, 15:4), from bacteria to eukaryotes by endosym- phrase is biblical in origin and, although in where it appears more metaphorical in biosis (by mitochondria and chloroplasts), Darwin’s words, it is ‘‘a useful simile’’, it is intent. and so on. No doubt there are other best not interpreted literally. The basic Thus, my first point is that Darwin mechanisms, but they all fit well with the confusion may have arisen because it is not certainly did not coin the phrase. Indeed, concept of descent with modification. recognised that Darwin was mainly inter- the phrase has widespread use in many Basically, descent with modification allows ested in the mechanisms of evolution. cultures (see Wikipedia) and is clearly an ‘‘cycles’’ in ‘‘graphs’’ (technically, a tree is These include both whether mechanisms ancient idea with strong mystical over- a ‘‘connected graph without cycles’’ [8]). that could be studied in the present could tones. This ancient origin of the term the In real life, cycles could result from explain past events, and establishing the Tree of Life does not preclude its use in hybridisation, endosymbiotic gene trans- continuity between all living creatures. He science, but it certainly means that we fer, lateral gene transfer, recombination, was not primarily interested in the de- need to be careful how we use it—we lineage sorting, the complexities of gene- scription of the patterns of evolution. always use it in lower case letters (tree of alogical relationships, etc. Importantly, We need to reappraise the different life) to help avoid unintended implications. Darwin’s more general phrases, including concepts, what Darwin meant by them, It perhaps should also be pointed out that descent with modification, emphasise the and how he used the terms. The Tree of searching the same biblical Web source continuity between populations, subspe- Life example is not the only case where does not find any occurrences in either cies, sibling species, etc., which was Darwin’s work has potentially been mis- the Old or New Testament of either the perhaps the more fundamental issue in understood (and I will mention some of ‘‘theory of descent’’, nor of ‘‘descent with the mid-19th century (rather than that of these later). In contrast to being interested modification’’! the precise form of relationships). in describing patterns of relationships, I argue that Darwin appears to have been Citation: Penny D (2011) Darwin’s Theory of Descent with Modification, versus the Biblical Tree of Life. PLoS more interested in the extent to which Biol 9(7): e1001096. doi:10.1371/journal.pbio.1001096 mechanisms that can be studied in the Published July 5, 2011 present are both necessary, and sufficient, Copyright: ß 2011 David Penny. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. The Perspective section provides experts with a Funding: No specific funding was received for this work. forum to comment on topical or controversial issues of broad interest. Competing Interests: The author has declared that no competing interests exist. * E-mail: [email protected]

PLoS | www.plosbiology.org 1 July 2011 | Volume 9 | Issue 7 | e1001096 Figure 1. In 1830, (effectively Charles Darwin’s mentor) published his Principles of Geology as an attempt to explain past geological changes by mechanisms that can be studied in the present. Lyell’s frontispiece was the remains of the ancient Greek Temple of Serapus at Puzzuoli, near Naples in Italy. Known forces of land subsidence and earthquakes could account for its unusual appearance today. It had been built on land, and then subsided over a millennium, which allowed marine molluscs to leave the high water mark. The remains of the temple had then been raised up again in an earthquake in September 1438 to its present position. It was an example of how known forces and mechanisms could be used to explain past events. Continuity was essential, but the known forces could vary in their intensity over time. doi:10.1371/journal.pbio.1001096.g001

This modification with descent phrase is ogies. As such, descent with modification, analogy of a branch of the tree overgrow- used in several contexts, including main especially in relation to continuity over ing, and supplanting a ‘‘feebler branch’’ headings in the Origin. It occurs in the short- and long-term time scales, is central (see Darwin [9], p. 148). The full quote is discussion around difficulties for the ‘‘the- to Darwin’s view of evolution. All this is ‘‘buds give rise by growth to fresh buds, and these, ory of continuity’’—including the absence fully acceptable to modern evolutionists, if vigorous, branch out and overtop on all sides or rarity of transitional varieties (leading to even though it is important to acknowl- many a feebler branch, so by generation I believe it a discussion on the relative completeness edge that Mendel’s particulate view of has been with the great Tree of Life, which fills of the fossil record, and why the fossil inheritance was not at all well known to with its dead and broken branches the crust of the record should not be read literally—still Darwin, and (as I note later) authors in the earth, and covers the surface with its ever excellent reading for paleontologists). It is mid 19th century were still puzzled by the branching and beautiful ramifications’’. As such, used as an explanation for the affinities of principles of inheritance. it is not a description of the relationship the extinct forms of life both to each other between taxa, but rather a suggestion that and to living forms. Another usage is that How Darwin Used the Concept a living tree analogy can be applied to species of different classes do not neces- of the Tree of Life lineages of species competing (and sup- sarily change together, or at the same rate, planting) other lineages or groups of or to the same degree; yet, in the long run, In contrast to the 21 occurrences of the species—whether fungi, plants, or animals. they all undergo some modification. It is theory of descent, in the Origin of Species In other words, microevolutionary pro- also an explanation of the naturalness of there is only a single usage of the Tree of cesses are similar in principle to macro- the classification of eukaryotic groups, and Life. However—and this is my third evolutionary processes, a central part of also an explanation for biogeographic point—it is not used as a description of his thinking [7]. He also comments that distributions. Again, the phrase is used relationships, but rather as an analogy for ‘‘the affinities of all the beings of the same class with reference to variability under domes- competition between species (and groups have sometimes been represented by a great tree’’, tication, and as an explanation for homol- of species) during evolution. It is the and immediately follows with the sentence

PLoS Biology | www.plosbiology.org 2 July 2011 | Volume 9 | Issue 7 | e1001096 ‘‘I believe this simile largely speaks the truth’’. aside, it is interesting that the tree is a non- advocated an increase in the use of And that is the usage that is best used binary tree, not a bifurcating tree, and this networks in . Trees (formally, today: the tree of life is a useful simile. But must be of concern to cladists (a small ‘‘connected acyclic graphs’’) are too simple the context in the Origin shows that readers group who attempted to maintain a strict to reflect all the signals in sequence data— were already expected to know about the relationship between a binary tree and the even though a tree is generally an excellent tree simile; it was neither a novel nor a formal classification). summary of the main effects. For example, central part of his theory. I and others have emphasised the multiple Thus, it seems clear that Darwin What Were Darwin’s Intentions? (and conflicting) signals in sequences [20], deliberately chose not to use the phrase have published a program (SpectroNet) for Tree of Life; he was certainly aware of it Is there any reason that Darwin basi- reconstructing networks [21], and shown and used the phrase in his Notebooks [10] cally rejected the Tree of Life phrase in that assuming a strict tree leads to biases in that were started over 20 years before the favor of the theory of descent with maximum likelihood when additional sig- Origin was published. However, even here modification? Before Darwin there were nals are present [22]. This should make it his usage was very different, and he many ideas about the relationships of clear that researchers like myself are not suggested in his notes that the Tree of species [13]. Perhaps the most favoured rejecting the tree per se but enriching the Life ‘‘should perhaps be called the of life, was the [11], but tree concept into a network. Indeed, the base of branches dead; so that passages cannot be there were many others that are referred more mathematically focussed side of seen’’. Nevertheless, the continuity between to in Stevens [14] and Winsor [15], and a phylogenetics has long used networks forms of life is still there. The 21 usages of diagram summarising many of them is in (e.g., [23]) to show more complex rela- the theory of descent in the Origin, and the [16]. tionships. Huson and Bryant [24] have one reference to the Tree of Life (which It certainly follows that species both updated the SplitsTree program, and was not as a description of relationships) changed with time and could split into Holland [25] discusses ways of forming demonstrates that Darwin rejected the more than one species. Put another way, consensus networks. concept of a Tree of Life to describe his there was no ‘‘unchangeable essence’’ to a views of evolution in favour of his theory species—this was an ‘‘idea’’ from Plato Lessons from Darwin of descent with modification. This does that was added into the concept of species not deny that he was well aware of the that developed in the late 17th century So far I have concentrated on what was concept, just that he declined to use it for [17]. Given Darwin’s ideas above, a tree- written by Darwin. My aim has been to his theory; instead, he focussed on mech- like form of relationship could have been a illustrate why it is important for both anisms. natural prediction. But, as I mentioned itself, and for our previously, scientists of the time already communication of the subject more wide- A Continuous Process knew about hybrids, and their importance ly, that Darwin’s work and his concept of in plant breeding. Even a century earlier, descent with modification is not misun- This leads to a very interesting fourth hybridisation was known to Linnaeus derstood. It is unfortunate, however, that main point emphasising further the con- [18]—indeed, Linnaeus suspected ‘‘gen- there are some other examples where cept of continuity that was mentioned era’’ were the true ‘‘fixed’’ units, and biologists seem to have insufficient knowl- earlier. Clearly, an evolutionary tree is a species themselves could change and vary edge of what Darwin wrote and of his useful concept and Darwin was certainly over time. So Darwin considered a tree a reasoning. This probably affects us all aware of it, and has the well-known tree useful analogy, but he was not very some of the time, and it is unfortunate that diagram in the Origin. However, an interested in the tree itself. He was more there is not more communication between important function of that diagram (and interested in asserting the continuity evolutionary biological researchers, and one less appreciated today) was that it between populations and species and historians of evolution who regularly study explicitly asserted the continuity of popu- genera, and on whether the mechanisms the older manuscripts [26,27]. lations, subspecies, species, sibling species, that could lead to change were sufficient. One example that demonstrates such a genera, etc. At a time when separate Why Darwin did not really like the tree of need for more communication is in regard creation of each species was the dominant life concept is worth a major study in its to the neutral theory of molecular evolu- theme, it was important to establish the own right, perhaps equivalent to Lovejoy’s tion, in which King and Jukes [28] idea that there were continuous sets of classic study on The Great Chain of Being suggested that neutral amino acid change intermediates (generations and popula- [11]. I should emphasise again that at the was ‘‘non-Darwinian’’ evolution—leading tions) between all these levels. This aspect time the Origin was published, virtually no to a decade of fruitless debate over of continuity was not unique to Darwin; it biologist knew the basis of Mendelian naming. However, Darwin stated ‘‘Varia- would be favoured by some versions of genetics. Even as late as 1870 ([19] pp. tions neither useful nor injurious would not be , in which evolution was 76–77), some researchers were studying affected by , and would be left thought to be driven continuously by some whether there was any inherited (‘‘genet- either a fluctuating element, as perhaps we see in external force, and some evolutionary ic’’) effect on a second foal, from a horse certain polymorphic species, or would ultimately interpretations of the Great Chain of that fathered the first foal. Given the lack become fixed,…’’. (This version is from the Being [11], which posits a linear hierarchy of formal genetic information at the time, 6th edition of the Origin, and is expanded with humans (especially European males) descent with modification was certainly from earlier editions.) Thus, Darwin was at the top! This continuity aspect of the keeping options open and focussed just on well aware of neutral changes, ‘‘neither diagram is not now generally recognised what was known, and on the continuity useful nor injurious’’, and commented on the because Darwin’s continuity between spe- between generations and populations. two aspects that eventually (when made cies (‘‘evolution’’) was basically accepted A final aside is that, although many quantitative) led to the neutral theory of among scientists within about a decade of research groups have studied trees exten- (namely, the high level the publication of the Origin [12]. As an sively, several research groups have long of in natural populations,

PLoS Biology | www.plosbiology.org 3 July 2011 | Volume 9 | Issue 7 | e1001096 and the apparent ‘‘’’ for the These three examples, together with the history of our subject if we wish to rate of fixation). A second example was the that of the Tree of Life, reflect situations claim something as novel. My own suggestion that Darwin favoured phyletic where some excellent evolutionary biolo- interpretation of why Darwin focused on gradualism (over punctuated equilibria), gists were not fully aware of Darwin’s a mechanistic view of evolution is that he but this claim has already been shown to thinking and reasoning. However, there is started his professional career as a geolo- be incorrect [29,30]. He carefully differen- almost certainly a more general issue here gist [6,7], effectively with Charles Lyell as tiated between ecological and geological in that biologists studying evolution, and his mentor [33], and Lyell sought to time scales when using terms such as ‘‘fast’’ historians studying the history of evolu- explain past geological events by mecha- or ‘‘slow’’. tion, do not meet regularly. Most of the nisms that could be studied in the present Another example is Mayr’s interpreta- issues I have discussed with respect to the (Figure 1). Whether this interpretation of tion of species as the fundamental unit of Tree of Life are well known to historians Lyell’s influence stands future tests re- evolution. But, as mentioned earlier, of biology; it is the lack of opportunity to mains to be seen, but certainly Darwin was Darwin assumed a continuum of interme- interact between the groups that has much more interested in mechanisms that diate stages between individuals, popula- hampered communication. could explain evolution than in describing tions, varieties, subspecies, sibling species, So where does this leave us? With patterns of relationships. As mentioned species, subgenera, genera, and so on; respect to the Tree of Life, it is unambig- earlier, given Darwin’s theory of descent there was nothing really special about the uous that Darwin neither invented this and his interest in explaining the past by level of species per se. This was highlighted ancient phrase, nor used it to describe the known mechanisms, we should welcome by Mallet’s [31,32] analysis that recent fundamentals of his evolutionary under- lateral gene transfer as another mecha- work with molecular data supports Dar- standing. In contrast, he routinely used the nism that can help explain past biology. win’s interpretation. Nevertheless, many term theory of descent with modification, (most?) biologists still appear to consider and focussed more on the mechanisms of Acknowledgments species as a quite fundamental unit, even evolution and the continuity of life. I though Mallet (as well as other authors) would therefore argue that we evolution- I am pleased to thank Mike Hendy for have argued otherwise. ists need to have a better understanding of stimulating and encouraging this work.

References 1. Lawton G (21 January 2009) Why Darwin was 13. Voss J (2003) Darwins Diagramme – Bilder von 22. Penny D, White WT, Hendy MD, Phillips MJ wrong about the tree of life. New Scientist. Issue der Entdeckung der Unordnung. Preprint 249. (2008) A bias in ML estimates of branch lengths number 2692. pp 34–39. Berlin: Max Planck Institute for the History of in the presence of multiple signals. Mol Biol Evol 2. Doolittle WF (1999) Phylogenetic classification Science, Available: http://www.mpiwg-berlin. 25: 239–242. and the universal tree. Science 284: 2124–2128. mpg.de/de/ressourcen/preprints.html. Accessed 23. Bandelt HJ, Dress AWM (199) Split decomposi- 3. Beiko RG, Doolittle WF, Charlebois RL (2008) 27 May 2011. tion: a new and useful approach to phylogenetic The impact of reticulate evolution on genome 14. Stevens PF (1994) The development of biological analysis of distance data. Mol Phylog Evol 1: phylogeny. Syst Biol 57: 844–856. classification: Antonie-Laurent de Jussieu, nature, 242–252. 4. Dagan T, Artzy-Randrup Y, Martin W (2008) and the natural system. New York: Columbia 24. Huson DH, Bryant D (2006) Applications of Modular networks and cumulative impact of University Press. phylogenetic networks in evolutionary studies. lateral transfer in prokaryote genome evolution. 15. Winsor MP (1976) Starfish, jellyfish, and the Mol Biol Evol 23: 254–267. Proc Natl Acad Sci U S A 105: 10039–10044. order of life: issues in nineteenth-century science. 25. Holland BR, Huber KT, Moulton V, Lockhart PJ 5. Kunin V, Goldovsky L, Darzentas N, New Haven (Connecticut): Yale University Press. (2004) Using consensus networks to visualize Ouzounis CS (2005) The net of life: reconstruct- 16. Penny D, Hendy MD, Poole AM (2003) Testing contradictory evidence for species phylogeny. ing the microbial phylogenetic network. Genome fundamental evolutionary hypotheses. J Theor Mol Biol Evol 21: 1459–1461. Res 15: 954–959. Biol 233: 377–385. 26. Bowler PJ (2009) Darwin’s originality. Science 6. Herbert S (2005) Charles Darwin, geologist. 17. Raven CE (1986) John Ray, naturalist: his life and 323: 223–226. Ithaca: Cornell University Press. works. Second edition. Cambridge: Cambridge 27. Bowler PJ (1990) Charles Darwin: the man and 7. Penny D (2009) Charles Darwin as a theoretical University Press. his influence. Oxford: Blackwells. biologist in the mechanistic tradition. Trends in 18. Reiseberg LH (2001) . In: 28. King JL, Jukes TH (1969) Non-darwinian evolu- Evolutionary Biology 1: e1. Encyclopedia of life sciences John Wiley and tion. Science 164: 788–798. 8. Semple C, Steel M (2003) Phylogenetics. Oxford: Sons, Available [requires subscription]: http:// 29. Penny D (1983) Charles Darwin, gradualism and Oxford University Press. onlinelibrary.wiley.com/doi/10.1038/npg.els. . Syst Zool 32: 72–74. 9. Darwin C (1859) . 0001753/pdf. Accessed 27 May 2011. 30. Rhodes FHT (1983) Gradualism, punctuated London: Murray. 19. Burkhardt F, Secord JA, edsThe Editors of the equilibrium and the origin of species. Nature 10. de Beer G, ed (1960) Darwin’s notebooks on Darwin Correspondence Project, editors (2010) 305: 269–272. transmutation of species. Bulletin of the British The correspondence of Charles Darwin. Volume 31. Mallet J (2008) Mayr’s view of Darwin: was Museum (Natural History). Historical Series Vol. 2, 18. Cambridge: Cambridge University Press. Darwin wrong about speciation? Biol J Linn Soc No. 2. Available: http://darwin-online.org.uk/ 20. Lento GM, Hickson RE, Chambers GK, Penny D 95: 3–16. content/frameset?itemID=F1574a&viewtype=image (1995) Use of spectral analysis to test hypotheses 32. Mallet J (2010) Why was Darwin’s view of species &pageseq = 1. Accessed 27 May 2011. on the evolutionary origin of the pinnipeds. Mol rejected by twentieth century biologists? Biol 11. Lovejoy AO (1936) The great chain of being; a Biol Evol 12: 28–52. Philos 25: 497–527. study of the history of an idea. Cambridge 21. Huber KT, Langton M, Penny D, Moulton V, 33. Lyell C (1830) Principles of geology: being an (Massachusetts): Harvard University Press. Hendy MD (2002) SpectroNet: a package for attempt to explain the former changes of the 12. Hull DL, Tessner PD, Diamond AM (1978) computing spectra and median networks. Appl earth’s surface by reference to causes now in Plancks principle. Science 202: 717–723. Bio Inf 1: 159–161. operation. Volume I. London: John Murray.

PLoS Biology | www.plosbiology.org 4 July 2011 | Volume 9 | Issue 7 | e1001096