Introduction
Carditidae Ferùssac, Reference Férussac1822 constitutes the best-known family among the poorly studied Archiheterodonta (Mollusca: Bivalvia). They are a group of exclusively marine non-siphonate bivalves with a fossil record beginning in the Late Triassic, and a maximum taxonomic and morphological diversity occurring during the late Eocene (Pérez Reference Pérez2019). Subsequently, their taxonomic diversity decreased until the present day. Carditidae shows a high degree of morphological inter- and intraspecific variation, including examples of sexual dimorphism in some species (Dall Reference Dall1902; Heaslip Reference Heaslip and Westermann1969; del Río Reference del Río1995; Pérez et al. Reference Pérez, Alvarez and del Río2017) that are expressed as differences in the convexity of shells, sharpness of radial ribs, or the presence of an invagination on the anterior-ventral margin. The latter structure is called a “marsupium” (Dall Reference Dall1890).
The first proposal of a systematic classification for the Carditidae (Dall Reference Dall1902) split the clade into two subfamilies: Carditinae and Thecaliinae. The main diagnostic character to distinguish both subfamilies is the presence of an incubatory chamber in the females of the second group. As a result of this feature, the Thecaliinae are commonly called “marsupial carditids” by an analogy with the abdominal incubatory pouch present on marsupial mammals (e.g., kangaroos and koalas). This striking structure is present in the two genera included by Dall (Reference Dall1902) in Thecaliinae: Thecalia Adams and Adams, Reference Adams and Adams1857, with only one species, Thecalia concamerata (Gmelin, Reference Gmelin and Gmelin1791) (Fig. 1A–D); and Milneria Dall, Reference Dall1881, with two species, Milneria kelseyi (Dall, Reference Dall1871) (Fig. 1I,J) and Milneria minima Dall, Reference Dall1916 (Fig. 1L–O). Later, Huber (Reference Huber2010) added another taxon to the subfamily, Powellina brookesi (Finlay Reference Finlay1926) (Fig. 1E–H,K), and related it to Thecalia.
Thecaliinae species showing incubatory chamber morphologies. A–D, Thecalia concamerata. E–H, K, Powellina brookesi. I, J, Milneria kelseyi. L–O, Milneria minima. A, Right valve, external view (AMNH-In 292590). B, Right valve, internal view (AMNH-In 292590). C, Right valve, internal view (MACN-In 13778). D, Closed shell, ventral view (AMNH-In 292590). E, Left valve, external view (AMNH-In 133025). F, Right valve, internal view (AMNH-In 133025). G, Right valve, internal view (holotype, AK 70135). H, Closed shell, ventral view (AMNH-In 133025). I, Left valve, external view (lectotype, USNM.Mo 253037). J, Right valve, internal view (lectotype, USNM.Mo 253037). K, Male right valve, internal view (NMR 16419). L, Left valve, external view (NMR 18239). M, Right valve, internal view (NMR 18239). N, Male right valve, internal view (PRI.KSU 8686). O, Closed shell, ventral view (PRI.KSU 8686). Arrows indicate placement of incubatory chamber. Scale bar, 10 mm (A–H, K); 5 mm (I, J, L–O).
The incubatory chamber is an invagination where the eggs hatch and develop during their first stages (Yonge Reference Yonge1969). Nevertheless, there is confusing and overlapping terminology used to describe this structure (e.g., Gmelin Reference Gmelin and Gmelin1791; Adams and Adams Reference Adams and Adams1857; Dall Reference Dall1890; Cox Reference Cox and Moore1969; Yonge Reference Yonge1969; Huber Reference Huber2010). This character represents a defining characteristic for the group and results in a strongly marked sexual dimorphism in this subfamily. It could have a high adaptative value, as it increases the survival of juveniles, but any evolutionary consideration must first tackle the ambiguous terminology in order to accurately describe and compare this structure among taxa.
New classifications and systematic schemes proposed to split the Carditinae into more subfamilies (Carditinae sensu stricto, Venericardiinae, Carditesinae, Palaeocarditinae, Carditamerinae, Miodomeridinae) (Chavan Reference Chavan and Moore1969). A recent phylogenetic study based on morphological characters supported the monophyly of a clade comprising Thecalia + Powellina within carditids, with this clade either strongly related to the subfamily Carditinae or nested within it (Pérez Reference Pérez2019). Also, the relationship between Thecalia and the Carditinae was supported by a molecular phylogenetic study (González and Giribet Reference González and Giribet2015).
Members of Carditinae and Thecaliinae share features such as mytiliform shells and a byssate epifaunal mode of life. They are both frequently attached to hard substrates (e.g., stones, corals, mollusk shells) by a proteinaceus byssus emerging from the anterior-ventral margin. Most Carditinae species have a byssal gape (with a variable degree of development) for the passage of this byssus. The Thecaliinae usually have tiny shells, and their incubatory chamber is positioned in the same area as this byssal gape. Is the byssal gape a necessary precursor of an incubatory chamber? Is there a selective process guiding the shaping of these traits? How these structures are related and have evolved was a matter of speculation (Yonge Reference Yonge1969; Pérez Reference Pérez2019) that now can be explored in a modern phylogenetic context.
Darwin (Reference Darwin1872) used the concept of “pre-adaptation” as an attempt to solve the problem presented by Mivart (Reference Mivart1871) about “incipient stages of useful structures,” a classical challenge for the Darwinian evolutionary theory that implies the non-adaptability of intermediate stages of phenotypes. “Pre-adaptation” became an unfortunate term in evolutionary biology with very different usages (Gould Reference Gould2002). Gould and Vrba (Reference Gould and Vrba1982) proposed the term “exaptation” to define part of the phenomenon included in the more extensive and variable usage of pre-adaptation. An exaptation is a character that was evolutionary co-opted to exert a new function unrelated to previous evolutionary forces acting on it (natural selection for a particular function or neutral evolution) (Gould and Vrba Reference Gould and Vrba1982). This implies a character acquiring a new function throughout the evolutionary history of the lineage rather than (or irrespective of) an enhancement of the previous one. Some popular and oft-mentioned examples of exaptation are the use of feathers for flight known in birds and other theropod dinosaurs, or the transition in the mammalian lineage, found between mammaliform and mammals, of several bones from the articulation of the mandible to auditory elements in the middle ear (Gould and Vrba Reference Gould and Vrba1982; Gould Reference Gould2002).
The main objective of this work is to test whether the incubatory chamber of Thecaliinae is related to characters previously present in the Carditinae and therefore to discuss whether this structure could be the result of an exaptation event in the evolutionary history of the Carditidae. Until the present study, this discussion was unclear, because it was precluded by confounding and/or ambiguous terminology in the specialized literature. Therefore, and as a first necessary step, a more precise definition and terminology of the studied structures is provided. Subsequently, using the most recent phylogenetic hypotheses, we will seek to establish the putative evolutionary transition between a byssal gape and the incubatory chamber and the possible implied evolutionary pathways. Acknowledging these events would allow an easier establishment of homologies for subsequent analyses.
Materials and Methods
Materials
We revised 46 fossil and extant carditid species (listed in the Appendix) from the following collections: Palaeontological Research Institution (PRI), Ithaca, New York, USA; American Museum of Natural History (AMNH-Fi and AMNH-In), New York, USA; University of Florida (UF), Gainesville, Florida; Yale Peabody Museum (YPM), New Haven, Connecticut, USA; National Museum of Natural History of the Smithsonian Institution (UNSM.Mo), Washington, D.C., USA; Natural History Museum of Rotterdam (NMR), Rotterdam, Netherlands; Australian Museum (AUK), Sydney, Australia; Auckland War Memorial Museum (AK), Auckland, New Zealand; the Natural History Museum of London (NHMUK), London, UK; Museo Argentino de Ciencias Naturales “Bernardino Rivadavia” (MACN-In and MACN-Pi), Buenos Aires, Argentina; Cátedra de Paleontología de la Universidad de Buenos Aires (CPBA), Buenos Aires, Argentina; and Museo de La Plata (MLP), La Plata, Argentina. We follow the systematic proposal of Pérez (Reference Pérez2019), which includes the subfamilies Carditamerinae, Thecaliinae and Carditinae in the large group Cardiobyssata Pérez and Giachetti (Reference Pérez and Giachetti2020) (= Carditida Pérez, Reference Pérez2019). Specimens belonging to the genera Cardita Bruguière, Reference Bruguière1792, Beguina Röding, Reference Röding1798, Birkelundita Heinberg, Reference Heinberg1993, Thecalia, Milneria, and Powellina were considered as the in-group of the analyses (subfamilies Carditinae and Thecaliinae). Specimens belonging to the genera Carditamera Conrad, Reference Conrad1838, Byssomera Olsson, Reference Olsson1961 (considered as a genus according to Vermeij Reference Vermeij2013), Glans Megerle von Mühlfeld, Reference Megerle von Mühlfeld1811, and Centrocardita Sacco, Reference Sacco1899 were considered out-groups (subfamily Carditamerinae).
Character Definitions and Description
In the original description of Thecalia concamerata, Gmelin (Reference Gmelin and Gmelin1791) mentioned the presence of an inner “chamber” in its shells. Subsequent authors followed this description (Bruguière Reference Bruguière1792; Dillwyn Reference Dillwyn1817). Gray (Reference Gray1828) was the first to describe this structure by mentioning its probable function in incubation and used the term “nursing pouch.” Adams and Adams (Reference Adams and Adams1857) and Adams and Angas (Reference Adams and Angas1864) described this structure as a “cup-shaped appendage.” Dall (Reference Dall1890) introduced the term “marsupium” to name imprecisely the incubatory structure observed in the genera Milneria and Thecalia. He retained this term in later works (Dall Reference Dall1902, Reference Dall1903, Reference Dall1916). Dall (Reference Dall1902) distinguished the two occurrences of this structure using the terms “funicular infold” for that in Thecalia and “dome-like indentation” for that in Milneria. Afterward, Yonge (Reference Yonge1969) used the terms “brood pouch” and repeated “funicular infold” to establish the same distinction. Chavan (Reference Chavan and Moore1969) vaguely distinguished both structures as “external” and “internal” incubatory chambers. These and other terms were used through time as indistinct synonyms (e.g., “cup-like process,” “ventral impocketing,” “shelly pocket,” “ventral infolding,” “brood chamber,” “marsupial chamber,” “rudimentary pouch”; Smith Reference Smith1885; Thiele Reference Thiele1935; Rudwick Reference Rudwick1964; Cox Reference Cox and Moore1969; Coan Reference Coan1974; Harvey Reference Harvey1995; Huber Reference Huber2010). Jones (Reference Jones1963) and others (e.g., Carter et al. Reference Carter, Harries, Malchus, Sartori, Anderson, Bieler, Bogan, Coan, Cope, Cragg, García-March, Hylleberg, Kelley, Kleeman, Kříž, McRoberts, Mikkelsen, Pojeta, Tëmkin, Yancey and Zieritz2012) used the terms “brood chamber” and “marsupium” for the region of gills or ctenidia of other bivalves (including other carditid lineages) with internal brooding of juveniles, but without an associated shelly structure. More recently, Carter et al. (Reference Carter, Harries, Malchus, Sartori, Anderson, Bieler, Bogan, Coan, Cope, Cragg, García-March, Hylleberg, Kelley, Kleeman, Kříž, McRoberts, Mikkelsen, Pojeta, Tëmkin, Yancey and Zieritz2012) used “marsupium” for both soft tissue and shelly chambers. Hence, a clear definition and description of this structure is needed.
The incubatory chamber of the thecaliine carditids is an inward invagination of the shell margin at its ventral or anterior-ventral area. In interior view, this shelly invagination varies from a shallow, exteriorly opened concavity to a funnel-shaped, exteriorly closed structure with a circular section. The exteriorly opened invagination is covered sometimes by a periostracum in living specimens (as in Milneria). The degree of development of the concavity can vary in the externally open incubatory chamber. The closed incubatory chamber may vary in its proportional development. It is unknown whether these two cases are extremes of a developmental gradation. Eggs and young individuals are retained in the incubatory chamber during prodissoconch stages (probably attached by byssal threads). According to Yonge (Reference Yonge1969), there are no observations in living specimens of the entrance of younglings into the incubatory chamber, but this author suggested that they must develop within the ctenidia and pass through it into the inhalant cavity before entering into the chamber. Males have a small byssal gape in the same position where the incubatory chamber is positioned in females. This byssal gape is not recognized in females.
The homology between the two incubatory chamber morphologies remains unclear. In this work, we prefer to distinguish both structures with two different terms. The structure represented by a rudimentary fold of the shell margin and externally opened is named the “external pouch” (as is present in Milneria and Powellina; Fig. 1F,J,L,M,O). By contrast, we use the term “funicular infold” for the completely closed structure, a name well established by Dall (Reference Dall1902) and Yonge (Reference Yonge1969) (as is present in Thecalia; Fig. 1B,C). The term “incubatory chamber” is used with no distinction for both morphologies.
Another character considered in this study is the gape present in the ventral or anterior-ventral region of the shell margin in members of Carditinae and Thecaliinae. Dall (Reference Dall1871) did not mention this structure when describing Ceropsis (= Milneria) for the first time. Terms used include “byssal excavation” (Finlay Reference Finlay1926) and “byssal gape” (Dall Reference Dall1902, Reference Dall1903; Yonge Reference Yonge1969; Carter et al. Reference Carter, Harries, Malchus, Sartori, Anderson, Bieler, Bogan, Coan, Cope, Cragg, García-March, Hylleberg, Kelley, Kleeman, Kříž, McRoberts, Mikkelsen, Pojeta, Tëmkin, Yancey and Zieritz2012). This structure is generally included as part of the shell outline in descriptions of bivalves, or it is even ignored. We used the term “byssal gape” in this work, following the terminology of Carter et al. (Reference Carter, Harries, Malchus, Sartori, Anderson, Bieler, Bogan, Coan, Cope, Cragg, García-March, Hylleberg, Kelley, Kleeman, Kříž, McRoberts, Mikkelsen, Pojeta, Tëmkin, Yancey and Zieritz2012). These authors indicated the presence of a byssal notch or byssal sinus accompanying the gape. The byssal gape of carditids is restricted to a gape or a space with an associated embayment toward the interior of the shell, positioned at the ventral or anterior-ventral margin of both valves (Fig. 2). This gape is related to the passage of byssal threads and it is a non-closed space between valve margins in ventral view (Fig. 2I,O). The byssal gape and associated embayments possess different degrees of development, from a slight curve (Fig. 2A–G) to a well-defined concavity (Fig. 2J–L) or a deep notch (Fig. 2H,M–O).
Carditinae species showing different developments of byssal gape. A, B, Cardita variegata. C, Cardita muricata. D, Cardita crassa. E, F, Beguina gubernaculum. G, Cardita excavata. H, I, Cardita rufescens. J, K, Cardita crassicosta. L, Cardita aoteana. M–O, Cardita leana. A, Left valve, external view (AMNH-In 29300). B, Left valve, internal view (AMNH-In 29300). C, Right valve, internal view (AMNH-In 29308). D, Right valve, internal view (MACN-Pi 2353). E, Right valve, external view (AMNH-In 29240). F, Left valve, internal view (AMNH-In 29240). G, Right valve, internal view (AMNH-In 19685). H, Right valve, internal view (AMNH-In 29277). I, Closed shell, ventral view (AMNH-In 29277). J, Left valve, external view (AMNH-In 29242). K, Left valve, internal view (AMNH-In 29242). L, Right valve, internal view (USNM.Mo 610611). M, Right valve, external view (AMNH-In 120152). N, Right valve, internal view (AMNH-In 120152). O, Closed shell, ventral view (AMNH-In 120152). Arrows indicate placement of byssal gape. Scale bar, 10 mm.
Character Mapping
The presence of a byssal gape and an incubatory chamber were coded with two approaches: (A) as two characters (1, presence/absence of byssal gape; 2, presence/absence of incubatory chamber); (B) as three characters (1, presence/absence of byssal gape; 2, presence/absence of external pouch; 3, presence/absence of funicular infold).
Characters were optimized in three different phylogenetic tree topologies (Fig. 3): González and Giribet (Reference González and Giribet2015; based on molecular data, GG) and two topologies from Pérez (Reference Pérez2019; based on morphological data). Two of the equally parsimonious tree topologies recovered by Pérez (Reference Pérez2019) were used because of the alternative placements of Thecaliinae. The first topology (P1) has Thecaliinae and Carditinae as sister groups, and the second one (P2) has Thecaliinae nested within Carditinae. González and Giribet (Reference González and Giribet2015) only included Thecalia concamerata, and Pérez (Reference Pérez2019) included T. concamerata and Powellina brookesi in their phylogenetic analyses. Despite this, all four Thecaliinae species were included (adding those not considered in the phylogenetic analyses mentioned). Thecalia concamerata and P. brookesi were considered sister taxa, according to Huber (Reference Huber2010) and Pérez (Reference Pérez2019) and were separated from Milneria minima and M. kelseyi. Huber (Reference Huber2010) considered Milneria as a different lineage, but he did not propose a new placement; because of this, the possibility of a non-monophyletic Thecaliinae is considered in the discussion. Maximum parsimony was used as the optimality criterion. This procedure was performed using the software Mesquite v. 3.2 with the Trace Character History function (Maddison and Maddison Reference Maddison and Maddison2017).
Alternative phylogenetic hypotheses for Thecaliinae placement. 1, Cardiobyssata clade; 2, subfamily Carditamerinae; 3, subfamily Carditinae; 4, subfamily Thecaliinae. Tree topologies taken from González and Giribet (Reference González and Giribet2015) and Pérez (Reference Pérez2019).
Results
The mapping including only two characters (Fig. 4, characters 1 and 2), presence/absence of byssal gape and presence/absence of incubatory chamber, shows congruence between the co-occurrence of both characters. In the molecular tree of GG, the presence of a byssal gape is restricted to a clade composed of species of the genus Cardita and the Thecaliinae. In the morphological trees, P1 and P2, this presence is shared by two clades: the Carditamerinae sister genera Carditamera and Byssomera, and the Carditinae and Thecaliinae. The byssal gape is not described for the Cretaceous carditine Birkelundita (Heinberg Reference Heinberg1993). All these taxa are included in the clade Cardiobyssata. On the other hand, the incubatory chamber is present only in the Thecaliinae. In the GG tree, the incubatory chamber is acquired in a taxonomically more restrictive region of the tree of taxa that already have a byssal gape. In the P1 and P2 trees, the result is similar when considering one of the lineages among Cardiobyssata, the Carditinae + Thecaliinae, except for the Cretaceous genus Birkelundita. Among Carditamerinae, both characters are absent in the clade composed of the sister genera Centrocardita and Glans. Thus, we can consider the presence of a byssal gape as a plesiomorphic character state for the Thecaliinae.
Mapping of characters on tree topologies. Numbers below branches indicate ancestral state reconstructions for the subsequent node. Question mark next to a number indicates ambiguous reconstruction. Only in-group is shown, and all Thecaliinae species are included. Black squares, character present; white squares, character absent. Tree topologies taken from González and Giribet (Reference González and Giribet2015) and Pérez (Reference Pérez2019).
Results using the three-character approaches do not differ from those of the previous analysis (Fig. 4, characters 1, 3, and 4). In the three topologies (GG, P1, and P2), a funicular infold is only present in Thecalia, which constitutes a subgroup within Thecaliinae. The external pouch is exclusive to the other Thecaliinae: the monotypic genus Powellina and the two species of Milneria. These results are expected, because Thecalia is the only taxon where the funicular infold is present.
Discussion
Byssal Gape
There is a well-established relationship between the presence of a byssal gape in adult shells and an epibyssate attached mode of life (Yonge Reference Yonge1969; Stanley Reference Stanley1975). The byssal gape is distributed among particular groups of carditids, including members of the subfamilies Thecaliinae, Carditinae, and Carditamerinae, which are mainly epibyssate, attached to different substrates. As previously stated, all male Thecaliinae present this character (Dall Reference Dall1902, Reference Dall1903; Yonge Reference Yonge1969), but it is not recognized in females. Among the Carditinae, this character is conspicuous in all species of Cardita and Beguina, but it is not mentioned for the Cretaceous genus Birkelundita. Heinberg (Reference Heinberg1993) described Birkelundita as an infaunal burrower according to their shell outline and, hence, a byssal gape is unnoticed but may have been present in juvenile individuals. This is inferred for postlarval stages of most living burrowing bivalves according to Stanley (Reference Stanley1975).
The byssal gape is usually small in carditids, and their associated embayment shows different degrees of development among species. In living Cardita variegata Bruguière, Reference Bruguière1792, Cardita muricata G.B. Sowerby I, Reference Sowerby1833, Beguina gubernaculum (Reeve, Reference Reeve1843), and in the Eocene taxon Cardita crassa Lamarck, Reference Lamarck1819, the latter structure is poorly developed (Fig. 2A–D). The living Cardita calyculata (Linnaeus, Reference Linnaeus1758) and Beguina semiorbiculata (Linnaeus, Reference Linnaeus1758) have a more developed embayment associated with the gape. Living Cardita excavata Deshayes, Reference Deshayes1854, Cardita crassicosta Lamarck, Reference Lamarck1819, and Cardita aoteana Finlay, Reference Finlay1926 show a more conspicuous concavity (Fig. 2G,J,K). On Recent shells of Cardita rufescens Lamarck, Reference Lamarck1819 and Cardita leana Dunker, Reference Dunker1860 (Fig. 2H,I,M–O), the gape is deep, with a significant embayment. Despite this, its placement on the ventral or anterior-ventral region of the shell margin shows low variability, reinforcing the idea of a homologous structure that was proposed based on the phylogenetic relationship between the three byssate subfamilies within Carditidae (Pérez Reference Pérez2019). Thus, the presence of a byssal gape is a plesiomorphic condition among Cardiobyssata carditids, and two subsequent apomorphic losses have occurred (in the clade Glans + Centrocardita and in Birkelundita). Also, the position of the byssal gape is strongly generalized for all adult byssate bivalves and would represent an ancestral state.
Incubatory Chamber
The presence of a shelly incubatory chamber is exclusive to the Thecaliinae among bivalves. The two configurations adopted by this structure (external pouch and funicular infold) could be extremes in a continuous developmental spectrum, from a shallow concavity to a deep funnel-shaped structure with a circular section in interior view, and from exteriorly opened to closed. Internal relationships among Thecaliinae are unknown, and there are no intermediate forms between these two morphologies. According to Huber (Reference Huber2010), Thecalia and Powellina are closely related, and Milneria could be from another lineage, not included within the other Thecaliinae. In this case, both morphologies would not be homologous. Our results are insufficient to irrefutably establish the homology between external pouch and funicular infold.
Some confusion about the identity of the byssal gape and incubatory chamber occurred in the past. Adams and Angas (Reference Adams and Angas1864) distinguished another species of the genus Thecalia and named it T. macrotheca due to the presence of a more developed incubatory chamber than T. concamerata. Both species were posteriorly synonymized (Smith Reference Smith1885). According to Adams and Angas (Reference Adams and Angas1864), the ventral margin of males is also folded on itself in this species, but not truly developed as in females. They seem to be actually referring to the byssal gape. Finlay (Reference Finlay1926) described the incubatory chamber of Powellina brookesi (= Cardita brookesi) as a “byssal excavation” and confused it with the byssal gape present in other Carditinae species. This is the reason why this structure was not recognized in P. brookesi until Huber (Reference Huber2010).
The homology between the incubatory chamber and byssal gape has been proposed by previous authors (Dall Reference Dall1903; Yonge Reference Yonge1969). These authors interpreted the incubatory chamber as an enlarged byssal gape. According to the phylogenetic placement of the Thecaliinae (González and Giribet Reference González and Giribet2015; Pérez Reference Pérez2019) and the anatomy and topological placement of both structures, the homology between the byssal gape and incubatory chamber could be strongly supported.
A Possible Evolutionary Pathway for the Incubatory Chamber
Among bivalves, 96% of the species are dioecious (have separate sexes), whereas only 4% present other categories (e.g., strict or consecutive hermaphroditism, protandric hermaphrodistism) (Coe Reference Coe1943). Nevertheless, sexual dimorphism is poorly developed on bivalve shells (Kauffman and Buddenhagen Reference Kauffman, Buddenhagen and Westermann1969). Internal brooding of larvae was reported for several groups of carditids without any specialized structure associated, such as some species of the genera Cyclocardia Conrad, Reference Conrad1867, Coanicardita Pérez and Giachetti, Reference Pérez and Giachetti2020, Crassicardia Conrad, Reference Conrad1838, Hippocampocardia Pérez and Giachetti, Reference Pérez and Giachetti2020, Glans, and Cardita (Dall Reference Dall1903; Lamy Reference Lamy1922; Jones Reference Jones1963; Yonge Reference Yonge1969; Schneider Reference Schneider1993). In some taxa (e.g., species of the Rotundicardia Heaslip, Reference Heaslip1968, Claibornicardia Stenzel and Krause, Reference Stenzel and Krause1957, and Darwinicardia Pérez and del Río, Reference Pérez and del Río2017) shell differences could be noticed between males and females based on the presence of more inflated shells near to the umbonal region (Heaslip Reference Heaslip1968, Reference Heaslip and Westermann1969; Pérez et al. Reference Pérez, Alvarez and del Río2017). This morphology has been attributed to the presence of brooding of larvae on females. To this point, brooding of larvae is known or inferred for members of the carditid clades Cyclocardida, Carditamerinae, Carditinae, and Venericardiinae, beyond from Thecaliinae. Dall (Reference Dall1903) suggested the possibility that all members of the Carditidae may brood their larvae, and Yonge (Reference Yonge1969), Coan (Reference Coan1977), and Schneider (Reference Schneider1993) also followed this proposal. If this is the case, brooding may be the plesiomorphic condition for all carditids.
The origin of an exaptation could be considered a three-step process (Gould and Vrba Reference Gould and Vrba1982; McLennan Reference McLennan2008). First, a primary adaptation is present in an organism, with a particular associated function. Second, this adaptation becomes a primary exaptation, through a co-optation event, by changing its original function or gaining an extra one. Any co-opted structure will not arise perfectly suited for the new function. Therefore, in this early stage, there is room for further adaptive tweaking. Finally, the third step involves selective processes modifying the trait according to the acquired function into a well-developed adaptation. As was stated earlier, the presence of a byssal gape is a plesiomorphic condition for all members of the Cardiobyssata, and the presence of an incubatory chamber is a synapomorphy of Thecaliinae. In our case, the primary adaptation and the derived function co-occurred in the Cardiobyssata: byssal gape used for byssus passage and brooding of larvae (Fig. 5A). Subsequently, the byssal gape could have been co-opted to the incubatory function at some point near to the origin of Thecaliinae (if considered a natural group; Fig. 5B). There are not known examples from living species, but Harvey (Reference Harvey1995) described the life cycle of Carditamera floridana Conrad, Reference Conrad1838 (Cardiobyssata: Carditamerinae) and indicated the presence of external incubation of eggs. In this species, eggs and young individuals are attached to the byssate adult shell. Also, Harvey mentioned the presence of sexual dimorphism in C. floridana, which is denoted by larger shells in females, a condition correlated to internal brooding in other carditids (Heaslip Reference Heaslip1968), suggesting the possible occurrence of ovoviviparity (or perhaps internal hatchling) in the past of this lineage. This example could represent a possible intermediate scenario, as Carditamerinae are members of the Cardiobyssata and they constitute the sister clade of the Carditinae + Thecaliinae lineage.
Possible evolutionary pathways. A, Primary adaptation stage, with byssal threads traversing the byssal gape, represented in Cardita species. B, Primary exaptation stage, with eggs or juveniles attached to byssal gape. At this stage, both functions (byssal passage and brooding) could have been present. C, Secondary adaptation stage, with eggs or juveniles in an externally opened incubatory chamber, represented in Powellina and Milneria. D, Refined secondary adaptation stage (or an alternative secondary adaptation stage), with eggs or juveniles in an externally closed incubatory chamber, represented in Thecalia. Co-optation event occurred between stages A and B.
The two phylogenetic hypotheses tested here are congruent with this evolutionary pathway, with Thecaliinae as the sister taxa of Carditinae (Pérez Reference Pérez2019), or nested within it (González and Giribet Reference González and Giribet2015; Pérez Reference Pérez2019). In addition, the carditine species Cardita leana (Fig. 2M–O) and Cardita thaanumi (Dall, Bartsch and Rehder, Reference Dall, Bartsch and Rehder1938) are the closest relatives to Thecaliinae in the results of González and Giribet (Reference González and Giribet2015), and they show a very invaginated byssal gape. From here, two pathways could be achievable in an adaptive context. Considering the monophyly of the Thecaliinae, the primary exaptation would have allowed the evolution of the incubatory chamber as a secondary adaptation (Fig. 5C,D). The external pouch morphology described for the genera Milneria and Powellina could represent the first stages, and the funicular infold of Thecalia could represent a more derived stage refining the structure by closing it to the external environment. This structure could have been positively selected if the survival chances of larvae were increased.
In an alternative scenario, the co-optation event could have happened twice. Huber (Reference Huber2010) considered that the genera Thecalia and Milneria belong to distinct lineages, with the first one closer to Carditinae, and the phylogenetic relationships of the second unknown; sexual dimorphism occurs in both taxa and would be a result of convergence with the acquisition of the incubatory chamber occurring twice independently. Unfortunately, there are no phylogenetic studies of the whole family (molecular or morphological) including all Thecaliinae to discern between these two possibilities. The most parsimonious interpretation is a single origin, but current evidence does not allow favoring one option over the other. Considering these evolutionary pathways, our exaptation case constitutes an “addition exaptation” (at least in its first stage) according to Arnold (Reference Arnold, Eggleton and Vane-Wright1994), because the new use was added to the first one (brooding added to byssus passage).
Another alternative hypothesis implies that the incubatory chamber originated from an interspace between two external radial ribs (Pérez Reference Pérez2019). Interspaces are concave zones among the convex expressions of radial ribs on the shell surface. They are a synapomorphy of the Carditidae and are retained in all members with very few modifications (e.g., Miodomeris cossmanni, where interspaces are extremely reduced). The ventral extreme of an interspace could have infolded to the interior of the shell and transformed into a precursor of the chamber. Despite this, and as can be observed in some examples of Thecalia (Fig. 1A), the exterior slit determined by the invagination obliquely crosses more than one radial rib and their interspaces. However, even if the ventral extreme of an interspace is related to the origin of the incubatory chamber or associated to the co-optation of the byssal gape, this only applies to the funicular infold morphology of Thecalia, and not to the external pouch of Powellina and Milneria, implying two independent origins for both morphotypes of incubatory chamber. Close relationship between Powellina and Thecalia has been proposed by Huber (Reference Huber2010) and Pérez (Reference Pérez2019). Therefore, the role of an interspace in the origination of the funicular infold only applies to Thecalia.
The possibility of the incubatory chamber as a newly developed structure, present as a synapomorphy of Thecaliinae cannot be completely discarded, but if we considered the broad presence of incubatory habits among carditids, the anatomical and topological evidence for the structures, the phylogenetic relationships between groups, and the presence of possible intermediate stages among other Cardiobyssata, the evidence strengthens our exaptationary hypothesis.
The earliest records of Carditinae are from the Late Cretaceous but correspond to the non-byssate Birkelundita (Heinberg Reference Heinberg1993), hence the timing of these evolutionary events cannot be currently established. Birkelundita is recovered by Pérez (Reference Pérez2019) alternatively as the basalmost Carditinae or as a sister taxon of Beguina. The first byssate Carditinae are recorded in the Eocene of the Paris Basin (Cardita crassa) (Cossmann and Peyrot Reference Cossmann and Peyrot1912). Our ancestral-state reconstruction places the emergence of the byssal gape of adults in the Cardiobyssata node. Byssate Cardiobyssata are known from the late Eocene (Carditamera williamsi) (Dockery Reference Dockery1997) and are well-established during the Miocene (e.g., species Carditamera apotegea, C. dasytes, C. arata, among others). The fossil record of the Thecaliinae is unknown, but the divergence time between Thecalia and the genus Cardita is inferred close to the Cretaceous/Paleogene boundary according to the results of González and Giribet (Reference González and Giribet2015). Therefore, the character ordination emergence suggested by our results does not contradict the known fossil record.
Importance for Character Building and Coding
An essential aspect of exaptations is their highly contingent nature (Gould and Vrba Reference Gould and Vrba1982), which binds their study to historical analyses (Baum and Larson Reference Baum and Larson1991). Armbruster (Reference Armbruster1997) and Blackburn (Reference Blackburn2002) already denoted the importance of including a phylogenetic context to study the adaptive/exaptive nature of traits. For instance, Baum and Larson (Reference Baum and Larson1991) mapped different characters of fingers and otic crests related to arboreality in order to describe exaptation cases in the macroevolutionary history of the salamander genus Aneides. Armbruster (Reference Armbruster1997) used a phylogenetic context and character mapping to describe the exaptationary origin of pollination by resin-collecting bees in the dicot Dalechampia (Euphorbiaceae). To assess the importance of exaptationary scenarios in macroevolutionary studies, it is strictly necessary to consider phylogenetic frameworks and historical approaches in order to avoid Panglossian narrations (Gould and Lewontin Reference Gould and Lewontin1979; Baum and Larson Reference Baum and Larson1991; Blackburn Reference Blackburn2002). This strategy also influences the identification of homologies, a fundamental step for well-based phylogenetic reconstruction (de Pinna Reference de Pinna1991; Grant and Kluge Reference Grant and Kluge2004; Ramirez Reference Ramirez2007). Exaptations are not usually considered in evolutionary studies of invertebrates, and this constitutes a possible example. Determination of evolutionary relationships among characters will help to place trait evolution within lineages in a broader context. Failure to recognize exaptations can lead to both wrong conclusions about historical processes and macroevolutionary pathways and spurious reconstructions of phylogenetic relationships. Natural history studies of different taxa could be considerably enriched by these approaches.
Conclusion and Perspectives
A possible exaptationary origin of the incubatory chamber of marsupial carditids is linked to the byssal gape present in the Cardiobyssata carditids, following our results. A three-step evolutionary pathway is proposed: the byssal gape acted as primary adaptation, then it was co-opted to a new function (brooding of larvae), becoming a primary exaptation at some point during the transition from the Carditinae to the Thecaliinae lineage. Subsequently, the adaptive processes modified this structure into the incubatory chamber constituting a secondary adaptation (putatively in an external pouch morphology first, and in a funicular infold later). Until internal relationships between Thecaliinae are established, the uniqueness of this evolutionary event cannot be determined. Further phylogenetic studies that include a broad sampling of Carditinae and all Thecaliinae species would be needed to answer these questions. It would not be surprising to find internal brooding or eggs attached to the shell in other species of the Carditinae.
There remain unanswered questions as an outcome of this proposal that would require developmental and soft tissue anatomy studies. For example, how the presence of an incubatory chamber disturbs the byssally attached mode of life in the Thecaliinae females, or what the early ontogeny of this striking structure might be. Other possible origins, such as via an interspace, could also be analyzed through ontogenetic studies.
The identification of the exaptationary origin for the incubatory chamber of marsupial carditids is a key step in the study of the Carditidae. Sexual dimorphism expressed in shells was recognized early in the study of the Thecaliinae, but generally disregarded as a curiosity. Nowadays, with the expansion of the use of phylogenetic tools for analyzing the evolutionary history of invertebrates, the inclusion of non-adaptative evolutionary scenarios could generate new and exciting possibilities.
Acknowledgments
The authors are indebted to those curators and researchers who facilitated access to collections: M. Longobucco, C. del Río, and A. Tablado (Museo Argentino de Ciencias Naturales “Bernardino Rivadavia,” Buenos Aires); M. Tanuz (Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires); A. Riccardi (Museo de La Plata); M. Siddall, L. Berniker, and B. Hussaini (American Museum of Natural History, New York); and G. Dietl, L. Skibinski, and V. Wang (Paleontological Research Institution, Ithaca). We especially thank the Club Social Sorgo de Aleppo for an interdisciplinary and pleasant space that allowed exchanges of opinion about evolutionary thinking. We also thank P. Milla Carmona (Instituto de Estudios Andinos “Don Pablo Groeber”) for his valuable opinions and comments and P. Picasso and M. Ezcurra for improving the language. Editor K. Boyce, J. Echevarría, and two anonymous reviewers are acknowledged for their constructive comments and review. Special thanks to the Paleontological Research Institution and the Paleontological Society for partially funding this project by the grants awarded to D.E.P.: the John W. Wells Grant-in-Aid of Research and the PalSIRP Sepkoski Grant, respectively. This work was also partially supported by the National Research Council of Argentina (CONICET), the National Agency for Scientific and Technological Promotion (PICT 2017-0220), and University of Buenos Aires grants (UBACyT 2018) to I.M.S.
Appendix. List of studied specimens. A dagger (†) indicates a fossil species.
