In 1856, the Danish naturalist Johannes Japetus Smith Steenstrup, published, in the Memoires of the Royal Academy of Sciences and Letters of Denmark, a detailed study of an “essential deviation from the symmetrical structure” in the octopods Argonauta argo and Tremoctopus violaceus (Steenstrup, 1856a). The identified structure went by the arcane name of “Hectocotylus” (hundred-fold tube) and, since its first description by Stefano Delle Chiaje (who had christened it Trichocephalus acetabularis, “hair-sized head with suckers”, Delle Chiaje, 1825: p. 225ff), it had undergone several changes of identity. Delle Chiaje (1825) and Cuvier (1829) described it as a parasitic worm, endowed with great liveliness and motility, as well as a staggering resemblance to an octopod arm. To the zoologist Rudolf Kölliker and his colleague, the comparative anatomist Carl von Siebold, the hectocotylus was instead the (never observed before) male form of three octopus species (Argonauta, Tremoctopus and Eledone rugosa), on account of its complex internal structure (Kölliker, 1846; Kölliker, 1849);. In advancing this hypothesis, they both relied on personal examinations on Tremoctopus specimens and on some earlier observations by the French zoologist Jeannette Villepreux-Power on argonauts (Villepreux-Powers, 1837). Although Siebold was confident enough to include this explanation in his influential manual of comparative anatomy (Siebold, 1848: p. 363ff), his optimism was to prove hasty: soon after, the Würzburg anatomist Heinrich Müller, and the Italian amateur naturalist Jean-Baptiste Vérany, through a series of well-aimed (and lucky) observations, put the matter to rest. Vérany had indeed engaged in a census of the marine species of the Mediterranean coast since the early 1830s. Between 1847 and 1851, he condensed the results in the first part of his Mollusques Méditerraneens, devoted to cephalopods (Vérany, 1851). The last entry of this census was on the Hectocotylus (p. 126), and contained an abridged history of the controversy, followed by his suggested solution: the Hectocotylus octopodis, proposed by Cuvier, was nothing else than the deciduous, regenerating sexual arm of the octopus, while this was not the case for Argonauta and Tremoctopus (Figure 1). However, in 1851, a short note by Müller announced the identification of male Argonauta, described as much smaller in size than female specimens, and the recognition of Hectocotylus argonautae as part of the animal (Müller, 1851). Thus, the nature of the hectocotylus as detachable sexual arm was confirmed for argonauts and, by inference, for T. violaceus, the male of which was still unknown.

The following year, Vérany and the Swiss zoologist Carl Vogt, published a lengthy account of the anatomy and behaviour of the hectocotylus (Vérany and Vogt, 1852), based on the observation of living animals and of the fresh specimens obtained from anglers in Nice and Genoa. Most of these studies were on O. carenae Vérany, 1839 (accepted name Ocythoe tuberculata Rafinesque, 1814), which lent itself especially well to in vivo anatomical examination due to the transparency of its tissues, allowing observation of the structure of the hectocotylus as part of the living animal (Vérany and Vogt, 1852: p. 176). Through regular visits to the fish markets the two were also able to secure a few specimens of argonauts and T. violaceus, for comparison, but could not find any males and thus relied on personal communication by Müller on the argonaut, and analogical reasoning for the other species. In 1853, also Müller completed his study, which found place in Kölliker and Siebold’s Zeitschrift für Wissenschaftliche Zoologie (Müller, 1853).

The three scholars finally put order in the puzzling series of observations and interpretations of the previous decades providing a thorough description of the main features of the organ, including the constancy and species-specificity of its position, the greater ease with which it could be removed from its basis, as opposed to the other arms (Vérany and Vogt, 1852: p. 155) and the persisting liveliness of the separated hectocotyli. These last two characteristics found an explanation (at least in the case of the argonauts) in the special challenges posed by copulation in species with such a remarkable sexual dimorphism. Both works also proposed speculations about the regeneration potential of the hectocotyli, by implication from the “well known fact” (Vérany, 1851) that it was difficult to find any living octopod without at least one regenerating arm.

These conclusions had immediate diffusion among naturalists, through translations (Henfrey and Huxley, 1853) and textbook summaries (Owen, 1855: p. 630-632).

Steenstrup’s 1856 contribution (Steenstrup, 1856a; Steenstrup, 1856b; Steenstrup, 1857) added an argument for the taxonomic relevance of this “essential deviation from the symmetrical structure” (Steenstrup, 1857: p. 79), due to its species- and sex-specific location, and its (now undeniable) role in the reproduction of the animal.

As to the phenomenon of regeneration, Steenstrup emphasised its specificity to octopoda, which “possess this power in the highest degree”. “All the Decapoda”, on the contrary, “[appeared] to be incapable of replacing accidental injuries of the arms, or the loss of parts of them, by a new growth” (Steenstrup, 1857: p. 107). This was to prove a long-standing myth in the field of cephalopod regeneration studies, despite the numerous testimonies to the contrary (for a review see Imperadore and Fiorito, 2018).

In the natural-historical debate sketched here, regeneration of cephalopod appendages emerges as a peripheral, but important element in the characterisation of taxa, structures and modes of life, in close relation with sexual dimorphism, as well as sexual and “defensive” autotomy (on the categorisation of autotomy, cf. Stasek, 1967). Steenstrup, like Vérany, Vogt and Müller, put on the scientific record the fact of arm regeneration in cephalopods, which before was a matter of common experience.

He also added a remark, as an agenda for future investigators:

Such a plea came from an authority in natural history, with three museums at his disposal (Steenstrup, 1857, note *: p. 83). Similarly, Vérany’s research critically depended on his institutional position, with all the connections it entailed, and the informal knowledge they contributed. Finally, the meaning of Müller’s decisive input (the observation of sex-specific traits in the argonaut male) had emerged against the backdrop of a close interaction between the locally-connected Vérany and the German colleagues.

Only 3 years after Steenstrup’s plea, the first European marine station was founded, at Concarneau (in 1859), on the Atlantic coast of France (Caullery, 1950). By 1900, there were more than sixty stations throughout the world (Dayrat, 2016), arguably an ideal infrastructure for pursuing Steenstrup’s programme. By the end of the XIX century, however, Steenstrup’s comparative-morphological approach had been superseded by a decidedly experimental one, with marine stations such as those of Naples (Italy) and Woods Hole (United States) playing a central role in the shift (Allen, 1975). On the one hand, regeneration became ever more firmly entrenched in a developmental framework, which entailed a focus on general “molecular” mechanisms (in animals, plants, and even crystals. Cf. Morgan, 1901) and a preference for simpler models, like the sea urchin embryo or the starfish, in addition to the traditional ones (e.g. salamanders and hydras; cf. Churchill, 1991). On the other hand, seashore laboratories contributed to the growing popularity of cephalopods mostly as physiological models, thanks to a level of organisation comparable to that of vertebrates (especially the closed circulatory system, unique among invertebrates, the complex nervous system, etc. Cf. Steiner, 1898), their tolerance to surgery and the remarkable viability of the explanted organs (Grimpe, 1928). In 1909, Bauer announced that “inkfish, and especially octopodes [were] about to rival frogs and rabbits” as physiological models (Bauer, 1909: p. 150). Just 2 years before, in his review of regeneration in the animal kingdom, Hans Przibram had remarked that knowledge of regeneration in cephalopods was limited to observational evidence, mentioning only Riggenbach’s, 1901 work on autotomy in O. defilippii (accepted name Macrotritopus defilippi Vérany, 1851) (Riggenbach, 1901) as the only experience with a bearing on the problem, under controlled conditions (Przibram, 1909: p. 130). It would indeed take the best part of a decade for a young scholar, by the name of Mathilde Margarethe Lange to devise the first systematic investigation of cephalopod regeneration in “standardised” conditions.

Lange was especially qualified for the task. Since 1910, she had read Zoology at Leipzig, Freiburg i. B, and Jena, attending the courses of the teuthologists Carl Chun (her first doctoral advisor), and Georg Grimpe. At Zurich, where she moved in 1917, she was supervised by Karl Hescheler, and attended the lectures of Adolf Naef, the authority in cephalopod systematics.

Lange experimented on live O. vulgaris, Eledone moschata and Sepia officinalis, at the Naples Zoological Station (in 1914), and the Musée Océanographique of Monaco (in 1915), providing macro- and microscopical description of all the stages of the process (cicatrisation, de- and regeneration), drawing comparisons between regeneration and embryonic development in cephalopods, and with the current results in invertebrates and vertebrates.

Cytological investigation yielded challenging results, especially as regarded the crucial mechanism of blastema formation. Since the 1880s, several competing theories of blastema formation had been proposed (Liversage, 1991). The prevailing one, named “epimorphosis” by Morgan (1901), had it derive from the dedifferentiation of neurones and muscle cells. These de-differentiated cells constituted the initial mass of the blastema, divided mitotically and re-differentiated returning to their original identity. What Lange observed in the octopus was instead a “double blastema”, as she named it. The “primary blastema” appeared to derive from the leucocytes carried by the blood vessels to the site of injury, where they phagocytised the cellular debris and formed the protective scar by agglutination (as cephalopod blood does not contain fibrin). After the regenerating skin had covered the site, the leucocytes appeared to transform, perhaps directly, into fibrocytes, the units of connective tissue. Lange’s “secondary blastema” (what we would today regard as the proper one) only began to appear after two or more days, displacing the primary without mixing with it.

Hescheler, Lange’s supervisor, was especially critical of her hypothesis of a direct transformation of one cellular type into another, as he made clear in his assessment of the dissertation ¹ . The US zoologist, however, was unshaken by this opposition and concluded that the dermal connective constituted an exception to the accepted view that like tissues derive from like precursors.

As the differentiation process was concerned, Lange remarked how it was directly dependent on the contact with the regrowing tip of central axons, thus confirming the regulative role of nerves in regeneration, another hotly debated topic at the time (cf. Reiß, 2022).

By the time her dissertation appeared on the Journal of Experimental Zoology, Lange had moved back for good to the United States, where she made a career as Professor of Biology at Wheaton College (a women’s college in Massachusetts. McCoy, 2016, 139ff). She returned to Naples only once (November 1927-May 1928, at the American Women’s Table), to pursue further research on cephalopods, but no information is available either at the Zoological Station Archive, nor at Wheaton College about the activities she conducted during this visit ² .

Surely, her first, ground-breaking stint had left open fronts. She had not followed the regeneration of suckers, passingly mentioned that of the eye lens, and only just raised the possibility of a different regeneration mechanism for the tip of the arm, where she had observed a permanent reservoir of undifferentiated embryonic cells. Finally, she had not really pursued a comparison between octopods and decapods, despite the general title of her dissertation (referring to the “arms of cephalopods”). Cuttlefish, Lange admitted, had proven too difficult to keep long enough. Nevertheless, she reported two intriguing cases of “compensatory regulation”, shown to her by Adolf Naef at Naples.

By the time Lange visited Naples for the first time, the Swiss zoologist Adolf Naef (1883-1949) was already well-known in the Station’s community, where he had been working since 1910 on a monograph on cephalopods for the series Fauna und Flora des Golfes von Neapel.

To Naef, the comparative study of the anatomy and embryology of a whole class afforded the possibility of an epistemological and methodological reassessment of morphological science, against two extremes: Haeckel’s phylogenetic morphology, with its emphasis on the recapitulation of developmental stages, and the excessive centrality of the phenotype and proximal causes preached by developmental mechanics (cf. Breidbach, 2003; Rieppel et al., 2013).

The debate between these two opposite positions had developed around the proper method for identifying homologies among organisms, and the very use of homology as a criterion in classification (cf. Laubichler, 2000). Since the early 1900s, the phenomenon of regeneration had taken centre stage in this debate. In a 1902 experiment, Hans Spemann and Hilde Mangold had extirpated the eye lens of a salamander embryo, and watched it regenerate completely, but from a different layer of tissue than its original precursor. This result disproved Haeckel’s theory of the Gastraea, a gastrula-like common progenitor of all animal forms (cf. Hoßfeld and Olsson, 2005). To Spemann, it also had wider consequences. If, as he argued in a later theoretical paper (Spemann, 1915), the regenerated lens had to be considered homologous to the extirpated one, then the very concept of homology had to be revised, and risked to lose most of its meaning. The problem was not only of explanatory frameworks, but also of methodology and approach: once accepted that ectopic regeneration was not an aberration, but true regeneration, then the proper way of elucidating the links between phylogeny and development was the study of the local conditions and mechanical processes that determined the phenomenon. This represented a complete reversal of Haeckel’s view on the relations between phylogeny and ontogeny, in which the latter became the basis for explaining the former. In methodological terms, this entailed the superiority of the experimental analysis of the mechanisms of development, over the systematic comparison of developmental stages.

Naef took an intermediate stand. On the one hand, he acknowledged the importance of Entwicklungsmechanik to morphology, and the criticism of Haeckel’s dogmatism. On the other hand, he found Spemann’s devaluation of homology too rush a conclusion to be drawn from a single experiment. To Naef, only a critical combination of all three approaches (comparative anatomy, plus descriptive and experimental embryology) could conclusively tell if regenerates of the kind observed by Spemann and Mangold were actually aberrations, or true homologies. The class Cephalopoda was of the right size for such an endeavour: large enough to allow empirical definition of homologies, but also small enough to be worked out by a single researcher, on the basis of a well-defined epistemic strategy. Comparative study of cephalopods held promise of yielding general concept of “type” and “typical stages” of development, based on the comparison between adult forms, to which he devoted the first volume of his work (Naef, 1972 [1921-1923]), and of developmental series of the greatest possible number of species (object of the second volume. Naef, 2000 [1928]: p. 342). Naef (1972) [1921-1923] Naef framed the phenomenon of regeneration as one element of a complex epistemological edifice, with the purpose of assessing the proper hierarchy of the different perspectives on morphology. To him, a science of form could only be founded on a comparative outlook, and the generalisation of results from single experiments, was misleading (Naef, 2000 [1928]: p. 342). Far from having consigned the problem of homology to the dustbin of history, experimental embryologists had to accept that an appropriate grasp on developmental mechanisms rested on a proper assessment of the relation between local, mechanical forces and typical, inherited developmental mechanisms (Naef, 2000 [1928]: p. 343). The brief experimental coda, attached to his great systematic effort, was meant to show just how this could be done.

In the succinct section two of the second volume (On Disturbed and Abnormal Morphogenesis and Its Relation to Normal Development), Naef built the case for cephalopods as a unifying model for morphology, by providing some hints on their proper use in the laboratory. The section opened with regeneration of the outer organs, followed by two parts on abnormal development (naturally occurring and experimentally induced). Naef noted the ubiquity of regeneration within the class (including, most clearly, arms and tentacles of decapods), the relative ease of obtaining it experimentally (Naef, 2000 [1928]: p. 343), and the possibility of contrasting several species-specific patterns of regeneration. He mentioned autotomy in O. defilippii (M. defilippi), as well as the interesting case of the loss of one dorsal arm in the argonaut, in which the remaining arm takes over the function of generation and repair of the shell. As Steenstrup had done before, Naef also warned of the possible misleading effect of arm regeneration on the identification of freshly caught specimens (p. 344).

If Naef’s coverage of regeneration in octopods was an orderly summary of the state of knowledge, the part on decapods offered new, first-hand observations, which he thought had potential for opening a few fronts of research. He noted that, apart from arms and tentacles, also small parts of the fin, arm membranes, eyelids and mantle regenerated easily, and that the phenomenon was easily controllable in the laboratory. Abnormal regenerates (heteromorphoses) were also often encountered in decapods, and in this connection Naef provided a lengthy description of the two extraordinary specimens mentioned by Lange in 1920. Probably because of the special position of the injuries, very close to the base of the arm, and to the buccal lappets, both specimens presented some mechanism of compensation (the “compensatory regulation” mentioned by Lange): the injured arms had not regrown, but in their stead, the corresponding buccal lappets had grown, slightly changed their position, fused with the injured stumps and started to develop suckers. The result was an intermediate condition between prehensile and buccal arms, confirmed by histological examination of their muscular connections. To Naef, the value of these exceptional instances was epistemic, in the first place. Sound knowledge of “the animal studied or developmental stage in all its details and […] multiple relationships with other members of the greater framework of order” (Naef, 2000 [1928]: p. 343), of the kind his monumental work had provided, allowed to determine whether these were cases of atavistic regeneration, or the expression of a “normally existing tendency” (p. 346). A firm experimental science of the mechanisms of adaptation, therefore, was critically dependent on the distinction between typical and atypical phenomena, which could only be rooted in comparison.

Naef intended to publish a more detailed study on the two cuttlefish specimens, but this promise, to the best of our knowledge, remained unfulfilled. He was never to see the Naples Station again, after his last 10-month visit in 1926 to complete the volume, and never to return to cephalopods (cf. Boletzky, 1999; Rieppel et al., 2013).

In their diversity, Lange and Naef’s takes on cephalopods as models for regeneration studies nicely complement each other. The former broke the ground for an experimental study and mechanistic interpretation of appendage regeneration in a so-far neglected animal class. The latter tried to reconcile two apparently opposing epistemic stands, by fashioning cephalopods as research models allowing the convergence of the comparative-anatomical and experimental-physiological approaches to morphology. Yet, both conspicuously failed to make any impact on contemporary regeneration research.

Lange’s dissertation was published in 1920, in the Journal of Experimental Zoology, which counted among its editors the US authorities on regeneration: Ross G. Harrison, Jacques Loeb, and Thomas H. Morgan. None of them seemed to take notice, however, for their way of framing regeneration was different. Although all of them researched on a variety of organisms, they did so mostly on account of the experimental advantages these offered towards a general physico-chemical, or at least mechanistic interpretation, rather than in a traditional comparative spirit. As Loeb put it in 1924, “We are already in possession of a number of enigmatic though often interesting observations on regeneration”, relic of a “stage of blind empiricism”, which made it difficult to discern whether one was getting lost in “a jungle of futile experiments”. What was needed, instead, were models amenable to precise quantitative work (Loeb, 1924: vi-vii), or well-chosen examples of generalizable mechanisms (Cf. Maienschein, 1991; Maienschein, 2010, on Harrison; Sunderland, 2010 on Morgan). The comparative approach loosely informing Lange’s study, and the interesting peculiarities she highlighted, were not what the US-American masters of the field cherished most. Nor did their European counterparts, reared in the same experimental-embryological tradition (cf. Barfurth, 1923; Przibram, 1926).

The fate of Naef’s synthesis is more nuanced. His Fauna und Flora monograph was saluted upon appearance as “the Bible of Theutologists” (Boletzky, 1999), and his epistemological stance was taken seriously and developed by a number of German-speaking scholars (from Adolf Portmann to Willi Henning), eventually constituting one pillar of the cladistics approach in the 1950s (Williams and Ebach, 2008). Yet, his ecumenical program for comparative and experimental embryology, centred on cephalopod regeneration, went completely unnoticed, as it fell in-between different audiences. On the side of systematics, the rise of the Evolutionary Synthesis, between the 1930s and the 1950s (Huxley, 1942), marked a disciplinary shift, consolidating around a nexus between the genetic, palaeontological and populational approaches, at the expense of the developmental. Despite occasional attempts of “translation” and introduction to Anglophone audiences (e.g. Zangerl, 1948), systematic morphology was actively side lined by the leaders of the Synthesis as a rear-guard approach (cf. Williams and Ebach, 2008: p. 62-63): Naef’s works were only translated into English from the 1970s (Naef, 1972 [1921-1923]).

As for the morphological disciplines of comparative anatomy and developmental mechanics, Naef’s call to collaboration, and his idea of cephalopod regeneration as a common field, also fell on sterile ground, because of the diverging paths of regeneration research, on the one side, and the perception of cephalopods as models, on the other side. On both shores of the Atlantic, regeneration was more than ever entrenched in an embryological framework, encompassing explanatory paradigms, methodology and the whole organisation of experimental systems, including animal models. Already before Mangold and Spemann’s spectacular demonstration of the “organiser effect” (Churchill, 1991: p. 116), and even more so after it (and Spemann’s 1935 Nobel Prize), the experimental object of choice for regeneration research were amphibians, especially urodeles. Apart from their very long association with regeneration since Spallanzani, salamanders and other germane species represented the perfect point of encounter between many different takes on regeneration. They afforded observation of normal and disturbed development at three different stages (embryo, larva, adult), and comparison among different species, which were not overly difficult to rear in captivity. Finally, and crucially, it was on such models that the practices of homo- and heteroplastic transplantation had been developed and perfected (what Reiß, 2022 calls “the practices of the cut and paste”). Cephalopods, on the contrary, raised many difficulties of management and interpretation. They were much harder to breed in captivity; their developmental stages were not as uniform and well understood as those of amphibians (Young and Harman, 1988), their taxonomy was constantly under revision, and even their age was extremely difficult to assess. Finally, such extreme experimental procedures were not possible, either because the animals were not resistant enough (this is the case for cuttlefish), or because those that were, like the octopus, presented peculiar problems: their arms could reach any part of the body, and boycott the recovery process (Boycott et al., 1965). A basic approach like Lange’s, or even the more refined one, only sketched by Naef, could not compete at the same level with Spemann’s experimental system. Moreover, the times of intense discussion of the evolutionary origin of the regeneration capacity (c.f. Goss, 1992) were long gone. Proximal causes and environmental influences were the name of the new game, and wide comparison across classes was a luxury that, perhaps, only a few, well equipped marine stations (like those of Naples or Woods Hole) could offer. Even there, knowledge of the material and methods for long term, comparative studies of regeneration were limited to a narrow circle of connoisseurs.

This is not to say that cephalopods had not consolidated their position as laboratory animals, on the contrary. A curious work, published in 1928 by Georg Grimpe, testifies to the growing demand of cephalopods as physiological and zoological models. A chapter of Emil Abderhalden’s encyclopaedic “Handbook of biological work-methods” (Handbuch der biologischen Arbeitsmethoden 1911-1939. Cf. Grote, 2018; De Sio et al., 2020, suppl. mat.) bore the title Pflege, Behandlung und Zucht der Cephalopoden für Zoologische und Physiologische Zwecke (“Care, Treatment and Rearing of Cephalopods for Zoological and Physiological Purposes”). What is revealing of this highly technical precis on methods and techniques is its focus on the demands of inland research aquaria—a sign of the growing fame of these “marine Guinea-pigs”, as he called them (Grimpe, 1928). Pupil and successor of Chun, Grimpe was a frequent guest of the Naples station and of many others, and could rely on the wisdom of the greatest teuthologists of the time. In fact, a great share of the technical information conveyed by Grimpe came from personal experience, or personal communication, but the overall picture he painted was one of great progress, especially in prolonging the survival of both captive octopods and decapods. Significantly, the concluding section (Grimpe, 1928: pp. 388-402), was devoted to the rearing of animals from the egg, a feat that had been tried with varying success since the 1880s (c.f. Joubin, 1888; Gravely, 1908; Drew, 1911) and to which Naef (1928) had attached a great importance as a means for turning cephalopods into the connecting link between systematic and experimental approaches. Although the rearing techniques for cephalopods (especially octopods) were nowhere near the level of development necessary for competing with amphibians or echinoderms in embryological studies, Grimpe’s summary conveyed the hope that, with a wider, planned effort, the difficulties could be overcome. In this voluminous chapter (mostly focussed on the common Mediterranean species), however, regeneration appears only marginally, and mostly in connection with the care of the animals. Lange’s procedures are duly described, and there is mention of regeneration of the eye lens, as well as of autotomy in O. defilippii (M. defilippi), but no treatment of regeneration experiments is provided, comparable to the much-better developed descriptions of physiological and psychological experimental systems. Moreover, Grimpe fell victim to the same misinterpretation of decapod regeneration as Lange. Although he gratefully listed Naef among his confidential sources, Grimpe (1928) bluntly stated that “no reliable proof of a natural regeneration has yet been adduced”, and, therefore “that Sepia, and even the more so the other decapods, are not suitable for experiments of this kind”.

The public Grimpe addressed had mostly other uses for cephalopods in its mind. Throughout the first half of the XX Century (Ponte et al., 2013) the greatest use of such animal models was in the field of neurophysiology, especially by means of chemical and electrical stimulation. Indeed, Grimpe reproduced almost the same list of experimental advantages as that proposed almost 20 years before by Bauer (see above). From the late 1920s, a new, productive front of investigation was opened, on the physiology, pharmacology and biochemistry of hormones and neurotransmitters (Bacq and Mazza, 1935; Erspamer and Boretti, 1951; Axelrod and Saavedra, 1977). Moreover, two pioneering experiences, by the Dutch animal psychologists Johannes A. Bierens de Haan and Frederik J. J. Buytendijk (academically “born” a physiologist) inaugurated the experimental study of octopus behaviour and of its neural underpinnings, a field that was to witness a great expansion after World War II (Bierens de Haan, 1926; Buytendijk, 1933).

Young only returned to the problem of octopus regeneration upon his retirement from academic life, in 1974, once again with an intriguing but solitary stint. He and Geoff Sanders (Sanders and Young, 1974) returned to the dynamics of pallial nerve regeneration on O. vulgaris, in a preliminary attempt at exploring the underlying physiological mechanisms. The landscape of regeneration research had changed dramatically in the four decades since Young’s last contribution to the field: new evidence (c.f. Gaze, 1970) had revealed the full complexity of neural development and regeneration, and the undeniable role of chemical signalling and an increasing number of growth factors in it. This evidence derived from studies on a variety of models: chick and frog embryos, in vitro cultures and fish. To Sanders and Young, the pallial nerve-system in octopods, once developed, could outclass all existing experimental systems: it allowed observation in vivo until completion of the regenerative process and each single animal afforded comparison of the operated vs. intact side of the mantle. Crucially, the pallial system combined a relative simplicity of access and intervention on the nerve, with a very refined “tool” for the quantitative assessment of regeneration: the rate of recovery of texture and colour-patterning in the skin. The variety and highly stereotyped character of both colour- and skin-patterns of octopus (cf. Borrelli et al., 2006) offered reliable external marks of the progress of regeneration.

Sanders and Young compared photographs of ca. 30 specimens after acclimatisation, and then at different stages of recovery after crushing, resection, or complete transection of the pallial nerve. Their conclusions were as intriguing as they were tentative, and raised baffling questions. In particular, five specimens showed “practically complete” recovery of chromatophore control, i.e., a “fully normal” pattern of response, as shown by comparison between the operated and un-operated side, and between pre-operation and post-recovery photographs (cf. Sanders and Young, 1974). How to account for such precise functional restitution in terms of the physiology of regeneration, however, remained mysterious. How could the colour- and skin-pattern changes be so faithfully restored? Having excluded the unlikely extremes of random re-innervation, and a total rewiring of the nervous system, Sanders and Young were left with the hypothesis that the regenerating axons reconnected “with their original end-organs” (p. 10), a mechanism about which, by their own admission, they remained “totally ignorant”. A personal communication by Andrew Packard was reported at the end of the article, pointing at some “degree of functional control of patterning within the skin”. This hypothesis helped at least to reduce the complexity of the phenomenon: instead of a one-to-one relation between nerve fibres and chromatophores, it posited that innervation may occur through a single axon reconnecting with all the chromatophores involved in a patterning component. Just how, exactly, the regenerating axon was supposed to find its way (either by guidance from the muscle fibres, or by reconnection to their “own labelled tubes”), was a matter for further research, exploiting the favourable experimental conditions offered by the system.

Neither Young, nor anyone of his entourage did follow up on this effort. In a number of studies published in the 1990s, Packard resumed experimentation on de- and regeneration of the pallial nerve, in the framework of his whole-animal investigations on the central control of chromatophores. There again, although regeneration as a phenomenon resurfaced in intriguing ways (see, e.g. Packard, 1991; 1995), regeneration as a problem, to be mechanistically accounted for, did not.

Recently, cephalopod arm wound healing was subjected to closer investigation (Shaw et al., 2016): wound closure was followed for the first 24 h after amputation in O. vulgaris, using classical histological staining, immunohistochemistry (IHC), high-resolution ultrasound imaging and electron microscopy. Despite the diverse experimental settings (water temperature, animals’ age, species, sex, surgical method, site of lesion) the newer studies (Tressler et al., 2014; Shaw et al., 2016; Imperadore et al., 2022) confirm the earlier macro- and microscopic accounts, and especially the key role of amoebocytes/hemocytes (Polglase et al., 1983; Imperadore, et al., 2022). Shaw et al. (2016) also suggested i. a role for muscles cells in plug development, ii. the involvement of apoptotic skin, muscle and nerve cells (assessed through the use of TUNEL Assay, for the identification of fragmented DNA) and iii. the hypothetical formation of a belt-like structure below the wound apparently functioning as an actin cable involved in a purse-string contraction.

The pallial nerve model has also been resumed after a 40 years-long eclipse (Imperadore et al., 2017; Imperadore et al., 2018; Imperadore et al., 2019a), with a wider scope. In connection with the recent, rising concern for cephalopod welfare, the more recent studies have expanded their focus to the behavioural responses to injury, as an index of the severity of the procedure. Soon after lesion and at recovery from anesthesia, a few animals exhibit intense grooming behaviour around the denervated mantle area, an action tending to last for a few hours after surgery; no other signs of pain or distress were ever observed (Imperadore, et al., 2017; Imperadore, et al., 2019a). In addition, evaluation of the animal welfare, assessed through the measurement of their predatory performance in terms of readiness to attack and type of attack (e.g., Amodio et al., 2014; following Borrelli, 2007), highlighted no significant differences between the lesioned and control groups or among individuals of the lesioned groups before and after surgery (Imperadore et al., 2019a).

Another element neglected by previous studies is breathing resumption, a conspicuous and easily measurable index of regeneration (Imperadore et al., 2019a). Finally, the great difference in the time of recovery, as well as in the number of successfully regenerating individuals, between the old and new experiments catches the eye, although it is difficult to account for on the basis of the published accounts.

In the first account (Sereni and Young, 1932) 65 days were needed to identify earliest signs of chromatic functional recovery and muscle contraction (both incomplete), highlighted through electrical stimulation of the skin and of the stellate ganglion post mortem, with functional regeneration only observed during summer. In the experiment by Sanders and Young (1974), some animals never or only partially recovered the chromatic function; those described as ‘fully regenerated’, required a minimum time of 50–59 days in summer (23°C) after crush and 60–69 days in autumn (18°C) after nerve cut. Additionally, authors highlighted a mismatch between nerve regeneration and functional recovery.

The most recent work on the subject (Imperadore et al., 2019a), however, reports a remarkable 100% structural and functional regeneration, both during spring and autumn (water temperature between 18°C and 22°C). Particularly, while the timing for skin pattern recovery varied (fastest complete recovery at 45 days), the time required to observe regained control over papillae raising and breathing muscles contraction was set at around 1 month after lesion (30–37 days following surgery), independently of the temperature in all specimens. Possible causes of such a stunning difference may be the type of lesion performed, its localization on the nerve, or the different anaesthetics employed.

A detailed microscopical analysis of the events occurring after axotomy of the pallial nerve, confirmed the occurrence of Wallerian degeneration, where intense axon swelling, fragmentation and death is observed as a consequence of the separation from the soma (Imperadore et al., 2017, 2018). A few days after the trauma, fibres of the central stump start regenerating toward the scar tissue to penetrate it, while it requires much longer (around 2 weeks) to observe the same effect in the opposite stump (regenerating sensory neurones, Imperadore et al., 2017). Despite the disorganised appearance of the regenerating fibres, the two stumps direct regenerating fibres toward their end targets, eventually crossing the lesion site (Imperadore et al., 2017).

As regards the process of correct orientation of the regenerating fibres, the observations of (Imperadore et al., 2017; Imperadore et al., 2018; Imperadore et al., 2019a) lend support to Fèral’s hypothesis of the leading role of connective tissue, against Sanders and Young’s (1974) proposal of the “orientated strand of muscle” beneath the nerve as the means used by the fibres.

Finally, backfilling experiments on the regrowing nerve up to 5 months after lesion, are in agreement with Young and Sanders’ hypothesis of functional recovery though end-target reinnervation, with some fibres reconnecting to motoneurons in the stellate ganglion, and other crossing it to reach chromatophores at the periphery. However, although physiological and functional regeneration is achieved, the pallial nerve never restores its original structure, showing fibers aberration, swelling and branching even several months post lesion (Imperadore et al., 2019a). Unlike arm injury, pallial nerve regeneration remains a laboratory model, as no published account on the injury frequency in the wild for this structure is available.

Also Fèral’s open questions about the role of neurotransmitters in nerve regeneration have been resumed with the aid of more advanced techniques, with an indirect approach. Fossati et al. (2013); Fossati et al. (2015) measured the metabolism of acetylcholinesterase (AChE), the enzyme responsible for the breakdown of acetylcholine (ACh) in the regenerating arm of O. vulgaris. AChE activity, inversely correlated to ACh abundance, was found to drop during wound healing (3–17 days after damage) and reversed the trend only at the onset of regeneration (ca 18 days post lesion). In this instance, the active enzyme is restricted to the sole axial nerve cord. Return to activity basal level corresponds to complete morphological restoration, exactly like in Triturus (Singer et al., 1960).

A non-classical and non-cholinergic function was also suggested for AChE during arm morphogenesis in embryo development and in adult regeneration, again in O. vulgaris (Fossati et al., 2015). The enzyme was found to be expressed in non-neural regions, i.e., in the blastemal differentiating mesenchymal cells of the newly developing limb and in the blastemal structure that forms just after wound healing in the adult damaged arm, mainly composed of undifferentiated cells. These phases are characterized by intense cell proliferation. Mitotic cells appear diffuse in the whole early arm rudiment, later restricting to the most distal part of the tip as differentiation progresses (Nödl et al., 2015). Cell proliferation during adult arm regeneration appears to follow a similar pattern (Fossati et al., 2013), although Lange’s (1920) speculation on the “permanent reservoir of undifferentiated embryonic cells” at the tip of the adult arm (see above) remains to date unsubstantiated.

The characterization of HOX and Wnt genes in the regulation of development and regeneration (established for several metazoans: Ruddle et al., 1994; Petersen and Reddien, 2009; Holstein, 2012), is still at an incipient phase in cephalopods, as is that of the molecular fingerprint guiding and controlling arm growth and regeneration. The few available data are intriguing and encourage a specific attempt to pursue this research further.

Indeed, the expression patterns identified for HOX genes during embryo development in Euprymna scolopes, design a precise temporal and spatial distribution in some structures, suggesting a correlation between the localized gene expression and cephalopod morphological innovations (for instance, the funnel tube, the buccal crown or the light organs; Lee et al., 2003). Wnt proteins (Wnt1, Wnt2, Wnt5 and Wnt7), together with other molecules responsible for the regulation of proximodistal, anteroposterior, and dorsoventral axes, were instead proven active during limb development in cuttlefish embryos (S. officinalis and S. bandensis) and showed molecular regionalization, consistently with what observed in arthropods and vertebrates’ limb development (Tarazona et al., 2019).

Baldascino (2019) has explored the expression profile of about 30 genes in uninjured and regenerating octopus arms. Results reveal differential expression in the proximal arm areas, as compared to the tip. Moreover, some genes appeared up- and/or downregulated during different phases of arm regeneration (e.g., Wnt proteins, Hox-B7, Antennapedia; Baldascino, 2019).

In recent years, epigenetic regulation of gene expression during regenerative phenomena has become of great interest (for a review see Katsuyama and Paro, 2011), fuelled by the hope of finding ways to induce structural recovery in poor regenerators, such as humans (Barrero and Izpisua Belmonte, 2011). The questions of how these regulatory pathways work, and how they evolved, are still begging an answer, and investigation of cephalopods’ epigenome, among other regeneration competent organisms, may help filling some of these lacunas in a comparative perspective.

Evolutionarily conserved elements involved in DNA and histone methylation/acetylation were identified and found active in different tissues of O. bimaculoides, E. berryi and Doryteuthis pealeii (Macchi et al., 2022). Moreover, transcriptional analysis of control and regenerating structures in O. vulgaris highlighted dynamic gene expression profiles for some epigenetic regulators (Imperadore, 2017; Baldascino, 2019). In particular, the limb of adult individuals showed differential expression along its length, as for the case of the polycomb group (PcG) proteins of the PRC1 and PCR2 complexes, usually involved in the methylation of Histone H2 and H3, generally marking gene repression. These were found to be upregulated in the uninjured arm tip compared to medial and proximal arm areas, data also corroborated by gene expression analysis in another octopus species, E. moscata (Baldascino, 2019). A few genes of the same complexes, e.g., EZH2 and SUZ12 were also found upregulated during arm regeneration, particularly during blastema formation (Baldascino, 2019). It is worth noting that both, the adult arm tip and the regenerating blastema, are characterized by cells actively proliferating. As it happens for other species, PcG repressive marks may serve to induce or promote proliferation and allow for arm continuous growth and stump regeneration (Barrero and Izpisua Belmonte, 2011), although this remains speculation.

Although the summarised data appear to robustly support the commonality of developmental and regenerative pathways in cephalopods and other metazoans, we have to consider that a biased approach has been utilized so far. Unsurprisingly, research on cephalopods has until now replicated the pattern established on model organisms, relying on the available information from other species, and adapting the technology developed on them. Recently (Ritschard et al., 2019; Schmidbaur et al., 2022), class- and species-specific orphan genes have been identified and tentatively linked to the evolution of cephalopod morphological novelties. It is at least plausible that such novel, still uncharacterized genes could also be involved in regenerative processes, although this again remains conjectural. Alternatively, it is also possible that known conserved molecules have pleiotropic functions (Sánchez Alvarado, 2004) as was observed for Hox genes in E. scolopes (Lee et al., 2003) where these well-known conserved genes are expressed in novel structures, specific to the class Cephalopoda.

Interest in deciphering and characterizing the regeneration toolkit of competent organisms has recently been boosted by the emergence of the relatively new interdisciplinary field of tissue engineering and regenerative medicine (Polykandriotis et al., 2010; Berthiaume et al., 2011; Mehta and Singh, 2019). Crucial features, such as accessible genomic and molecular resources and tools, amenability to genetic manipulation, fast generation time and ease of maintenance in laboratory conditions, mainly restricted regenerative studies to a few well-established animal models (Mokalled and Poss, 2018; Mehta and Singh, 2019), leaving a variety of species unexplored. In some cases, these epistemic advantages prevailed over the most crucial aspect of high regenerative capacity, allowing poor regenerators, such as Drosophila melanogaster, Caenorhabditis elegans and Mus musculus, to take the lead in translational studies.

The accelerating methodological and technological spillover, together with the release of publicly available Omic datasets, and, not least, the cost-optimization of cutting-edge technologies, revamped the interest for still largely overlooked, proficient regenerators, determining the possibility to elucidate common pathways as well as novel genes involved in the process (Sánchez Alvarado, 2004; Smith et al., 2011; Franco et al., 2013; Brockes and Gates, 2014; Casco-Robles et al., 2018). The release of the first cephalopod genome (O. bimaculoides, Albertin et al., 2015) set the ground for a new era: in less than a decade, the genome and transcriptomes of more than ten species have been published, together with chromatin profiling and mass-spectrometry datasets, for some. The enormous flow of new data highlighted some unique features of this class: extensive RNA editing, gene duplication, gene family expansion (e.g., GPCRs, Protocadherins, C2H2 ZNFs), large scale genome reorganization and emergence of novel genes. All these elements have been tentatively correlated with the organismal novelties identified in cuttlefish, squid and octopus (e.g., Liscovitch-Brauer et al., 2017; Ritschard et al., 2019; Schmidbaur et al., 2022).

The interest raised by these findings inspired deeper examination of cephalopod nervous system, the largest, most complex, and most cell-dense among invertebrates (Young, 1963; Grimaldi et al., 2007). A brain atlas (Deryckere et al., 2020; Deryckere et al., 2021), massive single-cell and single-nuclei datasets (Styfhals et al., 2022) were produced for O. vulgaris paralarval stages, allowing for novel insights into the characterization of molecular signatures of brain cells at early stage of development for the first time in a mollusc. A measure of the effort required, however, is given by the consideration that only for 9% of the 200,000 brain cells estimated in an octopus brain at hatching (compared to the 200 million in the adult), a single-cell expression profile could be obtained.

The possibility of altering gene expression in vivo, through loss and gain-of-function experiments, is a new standard in the study of regeneration. A range of genetic tools have been developed upon, and are currently employed in model organisms: RNA interference, transgenesis, chemical- and UV-induced mutagenesis, and, not least, CRISPR-CAS9 technology (Mehta and Singh, 2019), which, since its development in the 2010s, has held promise of connecting basic life science with biomedical and biotechnological applications. Cephalopod models have long been kept at the margin of this tumultuous development, due to the absence of these genetic tools. Very recently, however, Crawford et al. (2020) successfully applied CRISPR-CAS9 to squid embryos (D. pealeii) obtaining completely disrupted skin pigmentation: the first ever gene knockout in cephalopods.

Imaging regeneration has also proved advantageous, in several species, to investigate regeneration. Despite limited access to commercial markers or techniques for real time imaging, some tools have recently been developed: label-free multiphoton microscopy (Imperadore et al., 2018; Imperadore et al., 2022), 18F-FDG PET (Zullo et al., 2018), optimized CUBIC clearing protocol (Deryckere et al., 2020) and neural tracing (Imperadore et al., 2019a; Imperadore et al., 2019b).