Speciation by physiological selection of environmentally acquired traits

A chance mutation affecting a single or extremely few individuals in a continuous population will be quickly diluted through interbreeding. Charles Darwin fully appreciated this difficulty with relying on natural selection alone, and suggested an enabling role for geographical isolation in the origin of species. However, Darwin also believed in evolution by the inheritance of acquired traits and in populations of interbreeding animals, both of which would need a different isolating mechanism to overcome dilution and play a role in animal evolution. Historically disputed, the inheritance of acquired characters is now increasingly accepted as a phenomenon, and Charles Darwin himself is acknowledged as closely pre‐empting the type of physiology necessary to mediate it in his hypothesis of ‘pangenesis’. In this article, we question how the inheritance of acquired traits might overcome the problem of dilution by interbreeding and contribute to evolution. Specifically, we describe how Darwin's young protégé, George Romanes, developed ideas he discussed with Darwin and extended pangenesis to include a conceivable solution published after Darwin's death: physiological selection of fertility. In light of the ‘rediscovery’ of pangenesis, here we recount physiological selection as a testable hypothesis to explain how environmentally acquired characteristics could become coupled to the generation of species.


Introduction to the question: is natural selection the only origin of species?
The problem with Darwin and Wallace's (1858) theory of natural selection as a complete explanation for the Origin of Species is easily understood.A chance mutation affecting a single or extremely few individuals in a continuous population will be quickly diluted through interbreeding.This fact is also clear from the outcome of artificial selection.Breeders need to prevent interbreeding when they select new variants.Charles Darwin fully appreciated this difficulty with relying on natural selection alone, and suggested an enabling role for geographical isolation in the origin of species.However, Darwin also believed in evolution by the inheritance of acquired traits and in populations of interbreeding animals, both of which would require a different isolating mechanism if they were to overcome interbreeding and help to explain animal evolution.The inheritance of acquired characters is now increasingly accepted as a widespread phenomenon, and Darwin himself described and experimented on the type of physiology necessary to mediate it, namely his hypothesis of 'pangenesis' , which is slowly being acknowledged as presenting a cutting-edge biology of soma-germline communication.In this article, we question if and how the emerging reality of the inheritance of acquired traits can overcome the problem of interbreeding and contribute to evolutionary change.We describe how Darwin's protégé, George Romanes, developed Darwin's ideas after his death and extended them to include a conceivable solution: physiological selection of fertility.We suggest physiological selection as a promising and testable hypothesis to explain how acquired characteristics become coupled to the generation of new species.

Historical background
Immediately following his ground-breaking publication of the theory of natural selection with Alfred Russel Wallace in 1858 (Darwin & Wallace, 1858), Darwin concluded that it could not be the sole process by which evolution occurs.Even as early as 1859, in The Origin of Species (Darwin, 1859), Darwin acknowledged the effects of the use and disuse of the parts in the evolution of organisms, which implied acceptance of the possibility of the inheritance of acquired characteristics.This would eventually become one of the bases for the subsequent rift between Darwin and Wallace.Wallace went on to become the effective founder of the narrow theory of neo-Darwinism.He would later be joined by August Weismann in 1883, shortly after Darwin's death in 1882.Weismann sealed the narrowing down of Darwin's ideas by formulating the theory of the Weismann Barrier (Weismann, 1883(Weismann, , 1892)), the declared aim of which was to exclude the inheritance of acquired characteristics (Weismann, 1883).The subsequent incorporation of Mendel's discrete inheritance ideas (Bateson, 1902;van Dijk & Ellis, 2016; see review by Auffray & Noble, 2022) formed the basis of the neo-Darwinist Modern Synthesis, defined by Julian Huxley's book Evolution: The Modern Synthesis in 1942 (Huxley, 1942).
Whilst the theory of evolution has continued to evolve since the conception of the Modern Synthesis (see historical details in Noble & Noble, 2023), we question whether Darwin would wish his name to be attached to a neo-Darwinian model still widely interpreted as including the Weismann Barrier as a central premise.In contrast to neo-Darwinism, Darwinism proposed a more varied but now largely forgotten vision of how evolution occurs, which not only implied but eventually emphasised a generative role for the environment and the inheritance of acquired traits.Darwin himself invested considerable effort in explaining the broad type of physiology needed for the inheritance of acquired traits to be possible in a theory he privately described as his 'beloved child' (Darwin, 1887, p. 75): the theory of 'pangenesis' .Pangenesis was not simply a new proposal for the mechanics of inheritance, acquired or otherwise, but was instead intended to be an ambitious unifying theory explaining a swathe of different (largely animal and plant developmental) phenomena as variations on a wholly new conception of organic growth, reliant on a purely hypothetical mode of intercellular communication and signalling molecule(s).In his 1868 book, The Variation of Animals and Plants under Domestication, Darwin wrote: I assume that cells … throw off minute granules or atoms, which circulate freely through the system … these granules for the sake of distinctness may be called cell-gemmules, or, as the cellular theory is not fully established, simply gemmules.
Gemmules from across the entire body (hence the 'pan') would then 'reunite' in the germ cells to generate an organisation (hence 'genesis') complete with information about the parental body-plan.To explain use-and-disuse, pangenesis simply relied on these particles to communicate the physiological state of their parent cell to the egg and sperm cells.Summarising this schema, Darwin continues: In variations caused by the direct action of changed conditions, whether of a definite or indefinite nature, as with the fleeces of sheep in hot countries, with maize grown in cold countries, with inherited gout, &c., the tissues of the body, according to the doctrine of pangenesis, are directly affected by the new conditions, and consequently throw off modified gemmules, which are transmitted with their newly acquired peculiarities to the offspring.
Interacting in the germline, these modified gemmules would develop into an offspring recapitulating the modified physiology that was their origin.However, it is important to note that Darwin was well aware that this would not generally include bodily mutilations: With respect to mutilations, it is certain that a part may be removed or injured during many generations, and no inherited result follow.
The reason was that gemmules produced with 'acquired peculiarities' were only those generated by the originating tissues in actively adapting to changed environmental conditions.Darwin therefore fully anticipated the negative results of Weismann's 20th century experiments (Wiesmann & Poulton, 1891) showing that mice whose tails had been cut off at birth did not produce tailless progeny.Moreover, the non-inheritance of spontaneous mutilations would not even have been a merely uninteresting negative result to Darwin, because under his new topsy-turvy view of heredity transmission 'it is not inheritance, but non-inheritance, which is the anomaly' (cited by Olby, 1963).Under this unified view, the 'multitude of facts' Darwin had originally sought to explain could now conceivably be made soluble by his radical pangenic view of organismal growth and plasticity (see Liu, 2018).What was missing was experimental support.

Collaboration with George Romanes and the rediscovery of pangenesis
George Romanes was only 10 years old when The Origin of Species was first published.It was later, as a student of physiology at Cambridge, that he first met Charles Darwin.They became life-long friends and collaborators.Romanes became a frequent visitor to Down House, and worked on attempts to prove the existence of gemmules through experiments on grafts of different vegetable species or varieties.They were looking for the buds that could 'combine the characters of the two united forms' .Sadly, not much came of these experiments (but see Liu, 2018 for a modern reappraisal).Nineteenth-century microscopy could not help them either.Darwin knew this too, believing that although 'many thousand gemmules must be thrown off from the various parts of the body at each stage of development' they are 'inconceivably minute and numerous as the stars in heaven') so as to be virtually undetectable.Without physical evidence, gemmules, like fictional particles evoked by physicists, were, as later put by August Weismann (1891, p. 319), 'a pure invention' lacking any evidence.Francis Galton also severely criticised Darwin's ideas (Bulmer 2003), and was responsible for conducting a blood 'intertransfusion' experiment largely taken to be the final nail in the coffin for pangenesis (Galton, 1871).Nonetheless, Darwin remained hopeful that with time this 'airy nothing' D. Noble and D. Phillips J Physiol 602.11 would evolve into a 'substantial theory' (Schafer, 1896, pp. 481-483).
Yet, it would be a mistake to think that nothing came from the long-lasting Darwin-Romanes collaboration in search of the additional physiological processes necessary to complete their evolutionary vision.What emerged was not new experimental results, but new thinking about the physiology of a 'Lamarckian' origin of species.This now appears fortuitous, because modern evidence for the old idea of pangenesis is beginning to emerge, the details of which are dealt with in an accompanying article (Phillips & Noble, 2023) and centre around the communication of extracellular RNAs (exRNAs) from the body to its gametes wherein they reprogram offspring development.The reality of the inheritance of acquired traits and underlying gemmule-like mediator is now acknowledged by even the most mainstream voices and genetic societies and goes to 'illustrate that inheritance can be entirely non-genetic' (Wolf et al., 2022, p. 2).Indeed, in a thoughtful rebuttal of claims by proponents for an extended evolutionary synthesis (EES), prominent evolutionary biologist Douglas Futuyma makes an exception for the increasingly robust evidence for 'epigenetic inheritance' , stating that 'a largely unrecognized, fairly ubiquitous phenomena has come to light, methods to study it have matured, and it irresistibly cries out for study and for integration into our ever-growing knowledge of evolution' (Futuyma, 2023;p. 532).We share this tempered enthusiasm, and argue that in light of Darwin's and Romanes's prescience pertaining to pangenesis and soma-germline communication, by returning to study their mostly forgotten developments of these ideas we can gain further insight into what we deem to be perhaps the most pressing question crying out for study and integration: if and how the inheritance of acquired traits can overcome the problem of interbreeding and contribute to the origination of species.
Fortunately, clues to Darwin's and Romanes's thinking during that fertile period can be pieced together from an extensive articles that Romanes published in both Nature and the Zoological Journal of the Linnean Society in 1886, 4 years after Darwin's death.Significantly, the articles are entitled 'Physiological selection: an additional suggestion on the origin of species' (Romanes, 1886a,b).They contain many references to Darwin's own views, so we can almost listen in to what must have been their frequent conversations at Down House a few years earlier.

Requirements for speciation by selection or direct influence of the environment
Romanes first asks the important question we have already highlighted: does natural selection suffice to explain the origin of species, as the title of Darwin's (1859) book suggests?The problem can be stated quite simply.Artificial selection has been responsible for the development of innumerable varieties of various species of cats, dogs, fish, birds and plants.But this form of selection has not resulted in new species.The variants can almost all interbreed.Romanes begins his article by pointing out that precisely the reverse is true of natural speciation, and thus concludes natural selection to be insufficient.This seemingly marked a first important break from Darwin's position: Mr Darwin himself acknowledges, and therefore suggests a wholly independent hypothesis by which to explain the fact.This hypothesis is, that varieties occurring under nature 'will have been exposed during long periods of time to more uniform conditions than have domesticated varieties, and this may well make a wide difference in the result.(Romanes, 1886a, p. 314) Romanes reacts: For my own part, I deem the hypothesis wholly insufficient to meet the facts.When we remember the incalculable number of species, living and extinct, we immediately feel the necessity for some much more general explanation of their existence than is furnished by supposing that their mutual sterility, which constitutes their most general or constant distinction as species, was in every case due to some incidental effect produced on the generative system by uniform conditions of life … yet this [just before long-term exposure to the same conditions] is just the time we must suppose the infertility with the parent form arose. Because, if not, the incipient variety would have been reabsorbed into its parent form by intercrossing.(Romanes, 1886a, p. 314) Romanes's logic is correct.Relying on natural selection alone cannot work since random mutations would initially affect only one or very few individual organisms so that cross-breeding in a large population would quickly dilute the frequency of the mutation.Romanes was acknowledging the now indisputable need for a reproductive barrier in any adequate theory of speciation.Darwin himself had already commented on this in the form of geographical barriers in The Origin of Species and elsewhere, for example: The most general conclusion which the geographical distribution of all organic beings appears to me to indicate is, that isolation is the chief concomitant or cause of the appearance of new forms.(Life and Letters,vol. II.,p. 28.) Darwin also referred to isolation as 'the corner stone of the breeders' art' (1868, p. 85) and in a letter to stated: With respect to original creation or production of new forms, I have said that Isolation appears the chief element.(Life and Letters,vol. II.,p. 28.)So what of the apparent disagreement here between Darwin and Romanes on the requirements of speciation?It seems to us that Darwin's answer lingered largely around the evolutionary scenarios attended to in the origin of species, of classical ecological isolation followed by adaptive radiation, whereas Romanes's mind had committed also to those 'post-Darwinian' questions at the heart of his physiological collaborations with Darwin and his three volume magnum opus, Darwin, and After Darwin (1892-97), namely, how the physiological system of pangenesis could permit speciation also through the inheritance of acquired traits and by the direct impact of the environment in sympatry.Romanes was recognising that for a theory of evolution to be consistent with all the facts of heredity he and Darwin had sought to explain with pangenesis, something more than ample time and uniform conditions was required.Put simply, Romanes was stating that 'Where [Darwin] went wrong was' in the same vein as Moritz Wagner: 'in not perceiving that geographical isolation is not the only form of isolation' (Romanes, 1897 p. 4, emphasis added).

Physiological isolation and selection of fertility
Perhaps Romanes's main contribution to evolutionary theory was to argue for the evolutionary primacy of reproductive isolation in speciation and clearly demarcating its varied forms and consequences.Principally, Romanes recognised two main categories of isolation: discriminate and indiscriminate, which referred to isolation by like-with-like or otherwise, respectively.Natural selection too was considered to be a mechanism for discriminate isolation, by (at least partially) removing the unfit from the population, leaving only the fit to reproduce, and was thus 'but a particular form of a [more] general principle' (Romanes, 1897, p. 10).Put differently, natural selection 'is a sub-species of the species [discriminate isolation], which in its turn is but a constituent part of the genus Isolation' (Romanes, 1897, p. 10).Indeed, there was yet another sub-species of discriminate isolation which, due to its 'ubiquitous operation' (Romanes 1897, p. 9), was deemed to be more potent than, or at least logically antecedent to, natural selection: physiological isolation.The deepest premise of physiological isolation is that the reproductive system is intrinsically variable in function between individuals and taxa, with sufficiently large differences between pairings resulting in sterility.Sterility between members of two taxa is both now and then (approximately) robust enough to indicate their belonging to separate species.However, 'some degree of infertility is not unusual as between different varieties of [even] the same species' , and, furthermore, due to being of the discriminate modality (reproductive like-with-like), 'wherever this is the case, it must clearly aid the further differentiation of those varieties [into new species]' (Romanes, 1897, p. 8).In other words, natural populations conceal cryptic 'fault lines' of differential fertility caused by inherent reproductive variability, which by isolating (absolutely or statistically) collectives of the population from others, furnishes that population with an inherent physiological capacity to overcome interbreeding and diverge into new species: natural selection certainly makes populations adapted to their environment, but it is the action of physiological selection that separates them into different species.
Physiological selection of fertility was a most ubiquitous mechanism of speciation to Romanes because its operation was logically guaranteed to apply in continuously breeding populations as in those isolated by geography.Moreover, it was the default mode of natural populations, because, in anticipation of the concept of genetic drift, it would generally be a stochastic process, driven by the underlying stochastic ('independent') variability of the populations' underlying reproductive biology: At first the difference would only be in respect of the reproductive systems; but eventually, on account of independent variation, other differences would supervene, and the variety would take rank as a true species.(Romanes, 1897, p. 42) Crucially, however, Romanes recognised this inherent stochasticity could in theory be moulded by the physiology of the organism in its interactions with the environment, whereby a phenotypic change positively influences discriminate fertility and thus additionally drives its own inheritance and gradual fixation in the isolating populations: I do not deny that in some of these cases changes of structure, size, colour, &c., may be the causes of the physiological [fertility] change by reacting on the sexual system in the required way.But in such cases free intercrossing will have prevented the perpetuation of any morphological changes, save those which have the power of so reacting on the reproductive system as to produce the physiological [fertility] change, and thus to protect themselves against the full and adverse power of free intercrossing.(ibid p. 55) Therefore, if affecting a large share of a population at once, such 'phenotypic drive' of reproductive divergence could, in theory, couple the direct effect of the environment and the inheritance of acquired traits to the generation of species.It is to the feasibility of these special cases of a physiological origin of species that we dedicate the remaining section of this article.

Physiological selection of inherited acquired traits
It should first be noted that one way for the inheritance of acquired traits to contribute to evolution is to directly D. Noble and D. Phillips J Physiol 602.11 change the phenotype of a population, such as if a sustained change in climate, pathogen or other challenge is first accommodated by phenotypic plasticity before being genetically assimilated in the manner described by the likes of Baldwin (1896), Waddington (1956) and West-Eberhard (2003).Such cases are conceivable by mechanisms including the mutagenic (and thus genetic 'hardening') effect of epigenetic markings like DNA methylation (Gonzalgo & Jones, 1997) or environmentally stimulated action of mobile genetic elements, and was accepted by each of Romanes, Darwin and even Weismann himself (for bacteria; e.g.see 1892, pp. 190-191).However, in these cases one species population transmutates into another, which has already been discussed at length (e.g.Corning et al., 2023), whereas the emphasis of physiological selection is as a mechanism of one species fracturing into multiple in sympatry.The novel pathway proposed by Romanes features similarities to Dawkins' 'green beard effect' (1976) reformulation of Hamilton's model of inclusive altruism (1964a; 1964b), and bears a similar logical structure.It is that: P1: there must exist a phenotype induced by the environment capable of being transmitted to offspring.P2: the phenotype is not only inherited but influences reproductive behaviour or fertility.P3: this influence is discriminate: there is differential reproduction between affected/unaffected.
Assuming the phenotype is first acquired by no small proportion of the population, the logical structure of this formulation essentially dictates the environmentally acquired trait be 'hoist by its own petard' to overcome the swamping effects of interbreeding and concurrently enable speciation.A further similarity with the green-beard effect is its immediate criticism, namely that a phenotype satisfying all criteria seems unlikely.
Besides mentioning that both cases of the green-beard effect and pangenesis were initially considered unlikely and later proven to be true, we further suggest, similarly to George Gaylord Simpson (1953) when proposing the Baldwin effect, that each of the underlying processes 'are known to occur, there is no reason to doubt that they could co-occur, [and] that there is even some probability that they must have produced that effect sometimes' (Futuyma, 2023;citing Simpson, 1953).Without making claims of their frequency of occurrence, and first recognising that any such occurrence seems to us entirely compatible as an additional mode with existing models of evolution, we now turn to if and how each premise of the argument might indeed (co-)occur.
That cases of inheritance of acquired traits (P1) exist we believe is now undeniable.One such case is reported by Dias and Ressler (2014), whereby rats trained to fear a specific chemical (classical conditioning) produced naive offspring that also feared the chemical, but not other control chemicals, even when fertilisation was in vitro (IVF).This example is particularly interesting because the inheritance of associative learning (see also Deshe et al., 2023), at least in higher organisms, provides an intuitive mechanism through which an acquired phenotype transmitted to offspring can be coupled to a change in reproduction (P2), such as a sexual preference.Indeed, some recent cases of inheritance of acquired traits (P1) have been shown to directly impact mating behaviour.One example is Toker et al. (2022), who found that C. elegans hermaphrodites exposed to higher temperature produced a pheromone making them more attractive to male conspecifics and that this phenotype was transmitted to offspring via a small-RNA based mechanism.As a result, the worms engaged in significantly increased out-crossing and their genes were more readily transmitted throughout the population, satisfying P1 and P2 of the argument stated above.Whilst positive discriminate reproductive isolation (P3) was not satisfied in this case, it is interesting to note that negative discriminate reproduction was instead likely to occur, as mutants unable to transmit acquired phenotypes via such a small-RNA based mechanism are known to become sterile if continuously self-crossed (see Rechavi & Lev, 2017).Similarly, Hoffmann et al. (2023) reported that in a model of recurrent stress (preconception glucocorticoid injection) male mice exhibited increased anxiety-like behaviours and transmitted this phenotype to their male offspring, along with an unexpected increased attractiveness to female conspecifics.Such cases of an epigenetically inherited trait being associated with a associated change in reproduction are only now beginning to emerge.We hypothesise that as more examples begin to surface, some will be discovered to feature the additional requirement of having a discriminate nature, with organisms that inherit the traits being more likely to reproduce with others also inheriting the trait, or vice versa.The likelihood of this being true is increased by the fact that this schema is unconcerned with the precise mechanism used to enact differential reproduction, with a change in sexual preference being only one of many possible routes forward.Any mechanism leading to discriminate reproduction will satisfy the condition, with other modalities including a change in reproductive morphology (i.e.pre-zygotic isolation), alternative pollinator or seasonal (epi-)genetic incompatibility (i.e.post-zygotic isolation), or even simple behaviour (e.g. by influences upon migration or mating displays).Any combination of such occurrences described above will complete the logical formula for a Romanesian pathway towards the evolution of species by the physiological selection of an environmentally acquired trait.

Conclusions
In this article we have shown that Charles Darwin was deeply concerned in his later years with whether his theory of natural selection was a sufficient explanation of the origin of species, and he collaborated with the physiologist George Romanes in experiments and theoretical developments attempting to complete his theory of evolution to better explain the title of his 1859 book, The Origin of Species.Following Darwin's death in 1882, Romanes (1886aRomanes ( , 1886b) ) outlined the ideas he had developed in collaboration with Darwin, which included his theory of physiological selection.
Modern experimental studies on epigenetically led evolution already partially vindicate Romanes's theory.Through clarifying the basis of reproductive isolation in populations, further physiological studies, including emerging epigenetic methodologies, could now lead the way to a more complete understanding of the origin of species.
That this hope is not restricted to those who are critical of the Modern Synthesis is illustrated by this quote from a strong supporter of the Modern Synthesis, Douglas Futuyma, in his recent (2023) review, where he asks many pertinent questions: Are there any generalizations about which epigenetic marks are inherited, and which not, and if so, do these bear the stamp of adaptation?How do epigenetic states change in response to selection for or against expression of particular traits or genes?How is the inherited basis of adaptive phenotypes reflected in changes in DNA sequence?Is the origin of epigenetically inherited adaptive modifications ever not grounded in sequence change?Does epigenetic state, and its inheritance, affect certain classes of traits more than others?
These are questions we ask ourselves too.The study of Evolutionary Biology is clearly in a state of flux.It is almost as though the more open spirit of Huxley (1942) is returning.The days when the physiological sciences were viewed as irrelevant to evolution have clearly gone.