Larva, nymph and naiad – for accuracy's sake

Abstract

Having noticed the lack of ‘consensus as to their use in modern entomological papers’ (p. 669), Bybee et al. (2015) intended to eliminate a supposed ‘confusion’ recognised by them in the usage of the terms ‘larva’, ‘nymph’ and ‘naiad’. We agree with the authors – indeed it is noteworthy that a significant proportion of zoologists, particularly experimental biologists (Wigglesworth, 1954, 1972; Siewing, 1969; Nijhout, 1999; Klowden, 2013; Shi, 2013), refer to all preimaginal stages (with the exception of eggs) of all insects invariably as ‘larva’. Albeit the term ‘nymph’ seems to be still in relatively frequent usage first of all by taxonomists, ecologists and other ‘classical’ entomologists, but also by many textbook authors (e.g. Gullan & Cranston, 2015), the nearly complete avoidance of the term ‘naiad’ (sometimes spelled as ‘najad’) in modern works on Ephemeroptera (Domínguez, 2001; Kluge, 2004; Hauer et al., 2008), Odonata (Corbet, 2002; Garrison et al., 2006; Heckman, 2006; Córdoba-Aguilar, 2008) and Plecoptera (Domínguez, 2001; Zwick, 2004; Hauer et al., 2008) is striking. Before reverting to the supposedly ‘correct’ usage of these terms as proposed by Bybee et al. (2015), it is worth raising the question: what is the reason that a significant portion of contemporary entomologists do not adopt them? The authors' answer is simple: entomologists of our time are not as well-educated as those who lived a century ago and they ‘struggle with the appropriate use’ (p. 667) of these terms. Harsh words; could they be true? It is conspicuous that the single argument used by Bybee et al. (2015) for proposing a consistent usage of their preferred terminology is authority: they provide a review of the history of the terms, they claim that present-day entomologists should follow the terminology as it was introduced by Comstock (1918, 1920) and they point out that their proposal is supported by ‘the authoritative modern scientific literature’ (p. 669) – here simply citing the well-known general entomological glossary of Torre-Bueno (1989). Languages as well as individual words constantly evolve, and the meaning of a given word is always decided by the actual speech community. Scientific terminology is not any different. There is no ‘original meaning’ (p. 668) of any term which could be a priori (without considering its scientific relevance) considered more correct than a subsequent usage, particularly if the latter is overwhelmingly accepted by the community. Bybee et al. (2015) believe that ‘a clear history is required to restore a concrete lexical entry for each of these words [larva, nymph, naiad] within entomology’ (p. 667). Accordingly, about one third of their article is devoted to a survey of the history of the terms, with much attention to ancient and medieval usages of them, illustrating the usages of these words (not the scientific terms!) by Pliny the Elder or in the Greek mythology. Such trivia might certainly be entertaining, but they have no relevance to science. Although there is a certain ‘rule of priority’ in order to maintain stability, terminology ideally follows progress of the scientific concepts. Besides, it is of course also influenced by authoritative and influential publications (textbooks, monographs), or frequently simply by the personal preference of the members of the scientific community. Because of this, claiming that the usage of Comstock (1918, 1920) is the single ‘exact […] use of these terms’ (p. 669), without examining the biological relevance of the terms in concern, is absurd. It is equally irrelevant to simply refer to Torre-Bueno's (1989) glossary as ‘authoritative modern scientific literature’ (p. 669). Similarly to all glossaries, this work is merely a compilation intending to document usages of terms taken from the whole entomological literature but without any critical evaluation of their scientific accuracy. Besides ‘larva’, ‘nymph’ and ‘naiad’, also the now outdated and rather unfamiliar terms ‘gaead’ and ‘neanide’ are listed in the referred work. All of these are equivalent either with ‘larva’, ‘nymph’ or ‘naiad’; does their occurrence in Torre-Bueno's (1989) glossary mean that they are ‘correct’ terms and we should use them? Two major hypotheses have been proposed to explain the evolution of the hemimetaboly–holometaboly transition and the acquisition of the pupal stage (hemimetaboly is understood here broadly, comprising the development of all nonholometabolous ectognathous hexapods, thus including paurometaboly). Most of the other alternatives reviewed by Tichomirova (1991) and Heming (2003) are either of minor importance or can be viewed as variants of one of the two, and therefore they are not discussed here. The two concepts are named pronymphal theory and heterochrony theory in this article. Lubbock (1873) attempted to explain the morphological diversity of insect juveniles by hypothesising that members of different insect groups hatch from the egg at different embryonic developmental stages. This idea was elaborated in detail by Berlese (1913) and hence it is frequently referred to as ‘Berlese's theory’. According to this concept, juveniles of holometabolous insects hatch precociously in comparison with those of the hemimetabolous groups (‘de-embryonization’) and therefore they are equivalent with the prolarva (= ‘pronymph’) (the hatching stage still surrounded by embryonic cuticles) of the hemimetabolans. This ‘de-embryonized’ prolarva (practically a free-living embryo) resumes its postembryonic development during the pupal stage, which therefore is viewed as a compression of all of the hemimetabolan larval (= nymphal) stages. A consequence of the pronymphal theory is that the juveniles of the hemimetabolans and holometabolans are considered nonhomologous, and therefore special terms are necessary to be introduced for them. Accordingly, Berlese (1913) restricted the term ‘larva’ for prepupal stages of holometabolans and introduced the following terminology for hemimetabolans: first instar = ‘prosopon’; juvenile instars without external winglets = ‘pronymph’ (proninfa); juvenile instars with external winglets = ‘nymph’ (ninfa). At the same time he used the term ‘nymph’ (ninfa) (taken from the French term ‘nymphe’ meaning pupa) for the last, resting instar (the pupa) of holometabolous insects too. Modern proponents of this theory use the terms ‘larva’ and ‘nymph’ for all of the juvenile stages (except of pupa) of holometabolans and hemimetabolans, respectively. This terminology became widely used particularly after Comstock (1918, 1920). According to this concept, juveniles of hemimetabolous and holometabolous insects hatch at nearly the same developmental stage, so the juvenile postembryonic stages of the two groups are homologous (although particular instars are not necessarily homologizable between different evolutionary lineages, and often not even intraspecifically when their number varies). Hemimetaboly is a relatively plesiomorphic type of development with a gradual organogenesis and larvae morphologically more or less similar to adults. Holometaboly differs from hemimetaboly in the strong specialization of larvae; as a result, larval instars diverge morphologically from the adult and as a consequence the development of adult organs is delayed (heterochronous displacement). This necessitates an accelerated morphogenesis of organs in the last pre-adult stage, which explains the emergence of the pupa homologous with the last larval stage of hemimetabolans. Although some of its assumptions were implicitly accepted among entomologists for a long time, the theory was developed in detail only as a response to the pronymphal theory. Notable proponents include Hinton (1963), Polivanova (1979) and Sehnal et al. (1996). The traditionally used continental European terminology and some important works published in the U.S.A. (Stehr, 1987, 1991) reflect this concept implicitly, given that all juvenile stages of both hemimetabolans and holometabolans were uniformly termed ‘larva’ (‘larve’ in French, ‘Larve’ in German etc.). Several authors use the term ‘nymph’ and its equivalents in a functional sense, reserving it for older preimaginal stages where the external winglets (‘wing buds’, ‘wing pads’) or other provisional organs are already present (von Kéler, 1963; Siewing, 1969), sometimes restricted to neometabolous paraneopterans (Weber, 1933; Moritz, 1995), but frequently used for other kinds of hemimetabolous (including ‘paurometabolous’) insects in the same sense (e.g. Carpenter, 1921; Inglisch, 1967; Landa, 1969; Kunkel, 1981; Richter et al., 1995; Kluge, 2004; Eggleton, 2011). In the modern French terminology the term for pupa is ‘nymphe’, the French equivalent of ‘nymph’ (Joly, 1977). Although significant differences might exist between the embryogenesis of different insects, and exceptions are known, based on the available scarce data it seems that the last embryonic instars of at least most Pterygota can be safely homologised due to the similarities of their embryonic cuticles (Polivanova, 1979; Konopová & Zrzavý, 2005). After the hatch and shed of this instar, the development of the insect body is continuous to the adult stage and the regular moults can be viewed merely as a process of periodical removal of the cuticle that is primarily necessary for growth (Sehnal et al., 1996). Actual growth trajectories can, however, differ significantly, and heterochronic changes in the development of particular organs are frequent. Because of these, homologization of particular postembryonic stages of different insect groups, especially if they are not closely related phylogenetically, is frequently impossible. The question of whether the holometabolan larva represents a strongly specialised form of the hemimetabolan larva, or rather that it can be viewed as a hemimetabolan prolarva of extended life span and capability of periodic moulting similar to a larva, is therefore largely philosophical: in both cases we can see a continuous development of the prolarva to adult with multiple sheds of the cuticle, the difference of the two concepts resting on a subjective periodization of the ontogeny. Our opinion is that although the problem perhaps cannot be considered as finally settled, we are not aware of any facts providing a convincing support for the very special assumptions of the pronymphal theory. The more simple heterochrony theory, however, has an equal explanatory power, and no empirical finding seems to seriously conflict with it. It is conspicuous that Bybee et al. (2015) did not elaborate their terminology for all insects. They vaguely divide insects into hemimetabolous, paurometabolous and holometabolous groups, and give some trivial examples for each, but they do not say anything about the less trivial cases. Their classification gives the impression that hemimetaboly, paurometaboly and holometaboly are three independent types of postembryogenesis of equal ‘rank’. As it will shown below, this is far from reality. The traditionally recognised developmental types of Pterygota are certainly useful generalizations, but all of them can be considered as modifications of a basic pattern. In accordance with the fossil record (e.g. Kukalova-Peck, 1978), the first pterygotes were possibly ametabolous, characterised by a high and unstable number of larval instars, probably with the presence of articulated wings in several of the later instars, and perhaps with the ability to moult after attaining sexual maturation. Both prometaboly (among recent insects found exclusively in the order Ephemeroptera) and hemimetaboly can be derived from the alate ametaboly by delay of the wing development and fixation of a single adult instar. Alternatively, hemimetaboly might have evolved from prometaboly either by suppression of the subimago or rather suppression of the final moult and fixation of the neotenic subimago as the reproductive stage. Probably all of these scenarios have actually happened, possibly multiple times. The evolutionary relationships of the main developmental types and the various specialised developmental situations occurring in different hexapod clades were discussed by Sehnal et al. (1996) and Štys & Šobotník (1999). Although typical cases of hemimetaboly and holometaboly seem strikingly different, several cases can be found which arguably obscure the clear distinction between the two developmental types and also between the corresponding juvenile instars. In some highly specialised paraneopteran groups (thrips, whiteflies, several aphids, male coccids) one or several late larval instars evolved to a resting, nonfeeding exopterous form which, although it is certainly a homoplastic acquisition to the holometabolan pupa, still might readily be termed ‘pupa’ on morphological and functional bases (neometaboly) (Weber, 1933; Joly, 1977; Moritz, 1995; Sehnal et al., 1996). A definition of the pupa in this sense will result in the uniting of the thrips and male coccids with the endopterygotes into a broadly understood Holometamorpha as it was proposed by Nüesch (1987) but apparently has not been followed by any subsequent author. Conversely, the pupa of several plesiomorphic holometabolans (Raphidioptera and several other neuropteroids, Trichoptera) are relatively active and the pharate adult is capable of crawling or swimming within the pupal skin (Tauber, 1991; Sehnal et al., 1996; Kristensen, 1999), and in this respect they somewhat resemble the active last preimaginal instars of most hemimetabolans. It is also impossible to find obvious differences between the morphogenetic processes in hemimetabolans and holometabolans. The imaginal discs, once viewed as a complex of ‘embryonic’ cells which remain undifferentiated during larval development (Robertson, 1936; Pearson, 1972), are not de novo acquisitions of Endopterygota, but they can be derived from analogous portions of the epidermis of hemimetabolous larvae; the extensive histolysis and histogenesis taking place in them is a consequence of a strong morphological difference between the larval and adult stages (Švácha, 1992; Sehnal et al., 1996). The delay of histogenesis in certain organs of holometabolans is also not an exclusive apomorphy of the Endopterygota, but rather a general trend within the clade Phalloneoptera (Paraneoptera + Endopterygota). For instance, Paraneoptera differ from the majority of Polyneoptera in that the wing pads are not progressively formed from the first instar but appear in late instars. The Aleyrodomorpha are in this respect more ‘holometabolan’ than the Endopterygota, the exopterygotous instar being limited to the adult stage. Larvae of various insects live in a habitat different from that of the adult. The adaptation of the larval and adult stages to different habitats resulted in a divergence between their morphologies, compelling a delay of histogenesis of various organs and their relatively rapid formation displaced to the end of the postembryonic development (Sehnal et al., 1996). Although the adaptation to specialised habitats and therefore the delay of organ formation is carried to the extreme in the highly specialised endopterygotan clades, still there seems to be no justification for a claim that the difference between the two developmental types is qualitative rather than merely quantitative. It is apparent that no universally consistent definition of larva and nymph can be formulated on either a morphological or a developmental basis. The traditionally recognised developmental types, including hemimetaboly and holometaboly, therefore represent merely evolutionary grades rather than strictly distinct character states. Similarly, the traditional grouping of pterygote larvae propagated by Bybee et al. (2015) is not more than a practical convenience without deeper biological relevance. Particularly the distinction between the ‘nymph’ and ‘naiad’ is controversial. The category ‘naiad’ uniting larvae of Ephemeroptera, Odonata and Plecoptera based on the claim that they develop via a narrowly defined ‘hemimetaboly’ (contrasting ‘paurometaboly’) obscures the unique plesiomorphic development (prometaboly) of the vast majority of Ephemeroptera, and puts an unjustified emphasis on the probably independently evolved larval morphological adaptations to aquatic habitats, ignoring the fact that otherwise the larvae differ in many respects across the three orders. Finally it is noted that ‘larva’ is a general zoological term which is used in a broad sense for juveniles of not only most arthropod groups, but members of several other metazoan phyla, sometimes even vertebrates (fish, Amphibia) which display indirect development, many times without implying homology between ‘larvae’ of distantly related clades (Siewing, 1969; Stehr, 1987; Hall, 1999). Usage of this term in a similar general sense for all juvenile stages (except of the pupa) of insects fits well into this concept. Bybee et al. (2015) state that ‘there is not a requirement for scientific terms to reflect homology (e.g. insect wing venation and genitalia). Thus, the use of each term and descriptive scientific terms in general is not restricted to any idea of monophyly, homology […]. Instead, we argue that these terms should be used to define both general habitat and postembryonic developmental similarities’ (p. 669). We strongly disagree with this view. Any morphological terminology should be based on homologies as much as it is practicable. If homologies of particular organs are not yet fully understood, as might be the case in certain details of wing venation or genitalia in some insect groups, then descriptive terms might be introduced tentatively in order to formulate hypotheses or simply to be able to discuss the structures in concern, but this does not mean that such terms have any scientific value. The current concept of the homologies of preimaginal stages of at least the ectognathous hexapods as it was reviewed, for example, by Sehnal et al. (1996) can be considered satisfyingly accurate. Therefore, from a scientific point of view we cannot see a raison d'être for the idiosyncratic, outdated terminology advocated by Bybee et al. (2015). We reject the statement that contemporary entomologists ‘struggle with the appropriate [emphasis by us] use of the terms larva, nymph and naiad’ (p. 667) and we believe that such a notion is rooted in a superficial and inaccurate understanding of the terminology. The ‘confusion between terms’ ‘larva’, ‘nymph’ and ‘naiad’ that Bybee et al. (2015) intended to clear up does not exist. Evidence shows that the terms ‘larva’, ‘nymph’ and ‘naiad’ – particularly the latter two – refer merely to evolutionary grades rather than to distinct, nonhomologous types of juveniles. This does not mean that if they are defined on a descriptive basis they are not recognizable, and certainly the majority of juveniles are easy to be classified into one of them. They, however, do not have a biological meaning deeper than that of the term ‘maggot’ used for the characteristic larval form of higher Diptera, or ‘grub’ for that of chafers. Such terms are frequent in entomology and there is no need of eliminating them from our terminology, but one should not think that they have a precise biological definition and therefore a ‘correct’ (i.e. scientifically meaningful) use. Because the term ‘nymph’ – much less so ‘naiad’ – is already widely used in descriptive entomology, and can perhaps have some educational value when used for particular kinds of larvae, we find their usage acceptable, albeit not particularly useful. We encourage editors and reviewers to let the authors decide whether they want to adopt the established idiosyncratic terminology fixed by tradition or rather follow the homology-based terminology – none of the ways seem to cause any confusion. The worst practice would be to force the authors to use a particular kind of allegedly ‘correct’ terminology and thus prevent its natural semantic evolution. We are grateful to Attilio Carapezza (University of Palermo) for his help in translating literature from Italian and to Cameron McKay (Nankai University) for revising the English of the manuscript. This study received financial support from the National Natural Science Foundation of China (grant no. 31472024) and the One Hundred Young Academic Leaders Program of Nankai University to DR, and the Czech National Science Foundation (grant P505/11/1459) to PŠ.