Carcinization: an evolutionary reality?
Existing hypothesesnext section
The term carcinization was first coined by Borradaile (1916), but a relationship between hermit crabs and lithodids had been proposed much earlier by H. Milne Edwards (1837), De Haan (1849), Boas (1880a, b), and Bouvier (1894a, b, 1895, 1897). While H. Milne Edwards (1837) and De Haan (1849) related the two groups through the coenobitids, Boas (1880a, b) and Bouvier (1894a, b, 1895, 1897) hypothesized a link between lithodids and the genus Pagurus Fabricius, 1775 (as Eupagurus Brandt, 1851). Boas (1880b) specifically compared mouthparts, gill structure, musculature of the chelipeds, and pleonal tergites of a pagurid and a lithodid and concluded that of all genera known at the time, the Lithodidae were closest to the genus Pagurus. In a subsequent study, Boas (1924), disputed the transitional role of the Hapalogastrinae in Bouvier’s (1894a, b, 1895, 1897) theory of pleomere plate evolution that had been put forth in the interim. Boas (1880b, 1924) was of the opinion that the membranous condition and plate reduction seen in Pagurus tergites represented decalcification of the median regions, which he attributed to an evolution from the ancestral form of early reptant decapods. Boas explained the differences seen in the well-calcified second tergite of Paralithodes brevipes (H. Milne Edwards & Lucas, 1841) (as Lithodes Latreille, 1806) simply as a subsequent reinforcement of the membranous area between the two tergal plates of Pagurus and with the addition of newly developed marginal plates. He equated the median regions of tergites 3-5 to the membranous areas between the tergites of Pagurus; marginally on each a row of little calcified plates had formed. His biggest earlier difficulty had been explaining the presence of paired first pleopods in females of Paralitho-des Brandt, 1848 (as Lithodes). However, when he learned of certain pagurid genera that also had female paired first pleopods, he placed the lithodid ancestor in or close to the pagurid genera Nematopagurus A. Milne-Edwards & Bouvier, 1892 and/or Pylopagurus A. Milne-Edwards & Bouvier, 1892, rather than Pagurus.
Although both Boas (1880a, b, 1924) and Bouvier (1894a, b, 1895, 1897) arrived at the same conclusion regarding the evolution of the Lithodidae from the Paguridae, Bouvier’s evolutionary concept was quite different. His hypothesis provided a most explicit pathway for the evolutionary transformation of a membranous hermit crab pleon to that of a heavily calcified and apparently plated lithodid pleon. McLaughlin & Lemaitre (1997) reviewed Bouvier’s (1894a, b, 1895, 1897) carcinization theory in depth. Nonetheless, it is helpful to reiterate the particularly pertinent aspects of that transformation in light of what is now known about the actual transition. In Bouvier’s view (Fig. 2), the membranous pleon of an ancestral hermit crab (Fig. 2a) progressively was invaded by calcified nodules, and ultimately these nodules fused to form calcified plates. Initial fusion and plate formation occurred principally in the first and second tergites (Fig. 2b), with the third through fifth invaded first with granules and successively with increasingly larger and more numerous nodules (Fig. 2c). Gradually, over evolutionary time, the plates of the first and second tergites fused completely (Figs. 2e, g), while the initial fusion of invading nodules in the third through fifth tergites formed the lateral and marginal plates (Figs. 2d, e). Ultimately the median nodules also fused to form solid plates (Figs. 2f, g).
Fig. 2. Summary of stages in the hypothetical transformation of the pleon from a pagurid ancestor to a lithodid as proposed by Bouvier (1894a, b, 1895, 1897), showing membranous (green) and calcified (yellow) portions, and calcified nodules (pink): a, pagurid ancestor, dorsal view; b, Hapalogaster Brandt, 1850, dorsal (top) and ventral (bottom) views; c, Neolithodes A. Milne-Edwards & Bouvier, 1894, posterior (top) and ventral (bottom) views; d, Lithodes Latreille, 1806, posterior (top) and ventral (bottom) views; e, Lopholithodes Brandt, 1848, posterior (top) and ventral (bottom) views; f, Paralomis White, 1856, ventral view; g, Cryptolithodes Brandt, 1848, ventral view. All pleons are of adult males. Abbreviations: ac, accessory plates; la, lateral plates; m’, “pièces marginales” of Bouvier (1897: 42, figs 5-9); Me, unpaired median plate; t, telson. Numbers indicate pleomeres 1-6. (Modified from Bouvier 1897, and McLaughlin & Lemaitre 2001). Not to scale.
More recent authors similarly have pointed to the close evolutionary relationship between the Paguridae and Lithodidae, but none have been as definitive in tracing pathways. Borradaile (1916) and Wolff (1961), for example, while deriving the lithodids from a Pagurus-like ancestor, cited other examples of quasi-carcinization (e.g., Porcellanopagurus Filhol, 1885a, Birgus Leach, 1815, and Probeebei) as convergent events within the Paguroidea. Cunningham et al. (1992), strongly supported by Gould (1992), not only evolved two lithodid genera from Pagurus, but also these authors, in a clear presumption of monophyly, proposed that those lithodid genera were so closely allied to the latter genus that they should be taxonomically included in it. Richter & Scholtz (1994) did not suggest a specific pagurid ancestor. They simply concluded that the Lithodidae represented a group of secondarily free-living “asymmetrical” hermit crabs. However, they did point to a series of characters they believed supported their conclusion. A decidedly different pathway, which suggested that alternatives to shell dependence led to carcinization, was proposed by Blackstone (1989) and reiterated by Gherardi (1995, 1996b). In yet another, potentially controversial hypothesis, von Sternberg (1996) adapted Sluys’ (1989) “rampant process of parallelism” in proposing a homeotic gene model of c-loci regulatory elements to account for carcinization in the Anomura. Most recently, Morrison et al. (2002), using mitochondrial gene rearrangement, purportedly confirmed the Cunningham et al. (1992) hypothesis.
In a series of studies of several lithodid genera, Crain & McLaughlin (2000a, b), McLaughlin & Lemaitre (2001), McLaughlin et al. (2001, 2003 ), and McLaughlin & Paul (2002) have demonstrated what morphological changes actually do occur in the pleonal tergites during the transition from the megalopa through early lithodid crab stages. In the megalopae of all species that have been investigated, the six pleomeres are represented dorsally by six undivided tergites. These tergites are typically chitinous, as in many, but not all pagurids, in the genera Hapalogaster Brandt, 1850, Oedignathus Benedict, 1895, and Placetron Schalfeew, 1892, of the lithodid subfamily Hapalogastrinae, and in the genera Lithodes, one species of Paralithodes, and Cryptolithodes Brandt, 1848, of the subfamily Lithodinae. Alternatively, these tergites may be at least weakly calcified as in the Hapalogastrinae genus Acantholithodes Holmes, 1895, and in the Lithodinae genera Lopholithodes Brandt, 1848, Paralithodes (one species), Phyllolithodes Brandt, 1848, and Paralomis White, 1856. As may be seen in the following summation, the evidence presented by the above cited authors confirms McLaughlin & Lemaitre (1997), who postulate changes in the lithodid pleon resulted at least initially from division rather than fusion. The most complete data are available for two species of the genus Lithodes and one of Paralomis, with more limited information for species of eight other genera. Although the patterns of changes during the early crab stages are similar, these changes do not appear to occur at the same rate among or even within genera.
Lithodes aequispinus Benedict, 1895 (Figs. 3M-C12). The six megalopal tergites are chitinous and distinct (Fig. 3M). Accompanying the molt to crab stage 1 is pronounced pleonal flexion. Tergites of the first, second and sixth pleomeres remain chitinous or become weakly calcified. Tergite one is distinct and undivided, whereas, tergite two now has the marginal plates partially to entirely separated. Tergite three also shows complete or partial separation of weakly calcified marginal plates (Fig. 3C1), as identified by the marginal spines of megalopa. Similar divisions of the marginal plates of tergites four and five are indicated by partial to complete sutures. The lateral plates of tergites three to five are faintly delineated, and sometimes slightly calcified. The median regions remain chitinous or become thinly membranous; megalopal spines are sometimes still present, but more frequently only indicated by setal projections. With the molt to crab stage 2, tergites one, two, and six, as well as the telson are moderately well calcified. The first and second tergites are generally unchanged except for armature. Tergites three to five are entirely membranous medially but have irregular lateral plates delineated and weakly to moderately well-calcified (Fig. 3C2); similarly well calcified marginal plates are usually separated from the lateral plates, at least on pleomeres three and four. Crab stage 3 is accompanied by increased calcification in tergites one, two and six, as well as changes in strength and numbers of spines. More importantly, the lateral and marginal plates of tergites three to five may be undivided, completely and widely separated, separated but contiguous, or partially rejoined (Fig. 3C3). The median area of tergite three frequently will have two or sometimes an entire row of small calcified nodules anteriorly, whereas these regions of tergites four and five remain completely membranous or show only one or two pairs of minute, slightly calcified granules. However, two or three accessory, and usually spinose, calcified nodules often can be detected developing in the marginal integument on one or both sides. At crab stage 4, the lateral and marginal plates of tergites three to five exhibit considerable variation, but usually are partially to completely rejoined. The median area of each tergite generally has a few calcified nodules, most numerous and largest on tergite three. Much more prominent and usually spiniform accessory nodules are now present on both, or only on the right sides of the marginal integument (Fig. 3C4). At crab stage 5, the first indications of sexual dimorphism appear. Fusion of the marginal and lateral plate elements of tergites three to five in males may or may not be complete, and while their shapes are irregular, the pairs are generally symmetrical in size (Fig. 3C5 male). Accessory spiniform nodules can be observed developing in the marginal integument on both sides of the pleon. The median areas remain membranous but with a slight increase in the number of calcified nodules. At this stage in females, the usually, now entirely fused lateral and marginal plates of tergites three to five are somewhat to appreciably larger on the left side of the pleon than on the right. However on the right side, several accessory spiniform nodules have developed in the integumental margin (Fig. 3C5 female). Similar nodular development is completely lacking on the left side. As in males, the median areas of these tergites, while still membranous, show slight increases in the number of calcified nodules present, some of which are minutely spinulose. Crab stages 6 and 7 are marked by increases in the number of nodules on the median areas of tergites three to five in both sexes (Fig. 3C6). In males, the size and number of accessory spiniform nodules increases in the marginal integument on both sides, whereas development of these nodules remains restricted to the right side in females (Fig. 3C7). By crab stage 12, the accessory marginal plates on the right side of the female pleon have begun to fuse (Fig. 3C12) and form the so called “marginal” plates of the adult, but in the males such fusion is not as apparent.
Fig. 3. Pleonal tergite development of megalopa and juveniles to crab stage 12 of Lithodes aequispinus Benedict, 1895: M, megalopa; C1-C7 + C12, crab stages. Colors indicate portions of pleomeres 1-6 that are chitinous (orange), membranous (green), calcified (yellow), variably chitinous or membranous (orange hatched green), and variably calcified or chitinous (yellow hatched orange). Megalopa pleon shown in dorsal view; crab stages each shown with tergites of pleomeres 1 and 2 (top), dorsal view, separated from those of pleomeres 3-6 and telson (bottom), ventral view. Views of pleomeres 3-6 and telson in crab stages include part of second tergite (uncolored). Abbreviations as in Fig. 3 with the addition of: n, calcified nodules; sn, spiniform nodules; amp, accessory marginal plates. Not to scale. (Modified from McLaughlin & Paul 2002).
Lithodes santolla (Molina, 1782) (Figs. 4M-C5). Megalopal tergites all are chitinous (Fig. 4M). With the molt to crab stage 1, the tergites of the first, second and sixth pleomeres become weakly calcified, whereas the other tergites remain primarily chitinous. The first tergite is distinct and undivided. In contrast, the second now has marginal plates clearly separated. The tergite of pleomere three also shows complete or partial separation of weakly calcified marginal plates, as identified by the marginal spines of the megalopa. Similar divisions of marginal plates of the fourth and fifth tergites are indicated by partial to complete sutures. The lateral plates of tergites three to five are faintly delineated (Fig. 4C1), and sometimes slightly calcified, but with marked reduction of the posterodorsal spines of the megalopa. The median regions remain chitinous but with megalopal spines indicated by setal projections. Changes accompanying crab stage 2 include the flexure of the triangular, symmetrical pleon against the thorax, with the telson and tergites of the first, second and sixth pleomeres moderately well calcified. The first and second tergites remain unchanged except for armature; whereas, the lateral plates of tergites three to five are clearly delimited, weakly calcified, and usually each now has two or three small spines. The marginal plates of these tergites are still delineated or now partially or completely refused with the laterals (Fig. 4C2). The median areas remain simply chitinous but with few spines of the megalopa distinguishable as low, sometimes weakly calcified, tiny nodules or spinulose protuberances, each with one or two short setae. In crab stage 3, the fused marginal and lateral plates of tergites three to five have become moderately calcified, while the adjacent, chitinous marginal integument now has a few spiniform nodules developing (Fig. 4C3). Although the median areas of these tergites still remain predominantly chitinous, nodular areas of calcification are often more definitive here. During crab stages 4 and 5, the median areas exhibit increases in the number of calcified nodules, sometimes with the nodules fusing to form transverse rows. At crab stage 5, sexual dimorphism is evident, with the composite marginal and lateral plates increasing appreciably in size on the left side in some individuals (females) (Fig. 4C5 female) but not in others (males) (Fig. 4C5 male). Additionally, the increasing larger accessory nodules in the marginal integument develop on both the right and left sides (males) or only on the right side (females).
Fig. 4. Pleonal tergite development of megalopa and juveniles to crab stage 5 of Lithodes santolla (Molina, 1782) and Paralomis granulosa (Jacquinot in Hombron & Jacquinot, 1846): M, megalopa; C1-C3 + C5, crab stages. For C5 of L. santolla, only left half of tergites 3-6 and telson shown. Views of pleomeres 3-6 and telson in crab stages of L. santolla include part of second tergite (uncolored). Color-coding and abbreviations as in Figs 2 and 3. Not to scale. (L. santolla, modified from McLaughlin et al. 2001; P. granulosa, modified from McLaughlin et al. 2003).
Paralomis granulosa (Jacquinot, in Hombron & Jacquinot, 1846) (Figs. 4M-C5): All six megalopal tergites and telson are well calcified, and each of the tergites is armed with one or more pairs of identifying spines (Fig. 4M). With the molt to crab stage 1, the pleon may or may not be strongly flexed against cephalothorax (not illustrated as flexed); the tergites of the first and second pleomeres are entire, distinct or partially to almost entirely fused. The tergites of the third through fifth pleomeres each has developed partial to complete lateral sutures dividing each into median and lateral plates (Fig. 4C1). At crab stage 2, the tergite of the first pleomere is usually partially to completely fused with the tergite of the second. The lateral plates of tergites three to five, respectively, are entirely separated from median plates (Fig. 4C2), but there is no delineation of marginal plates. In crab stage 3, the adult condition of fused tergites one and two is complete (Fig. 4C3); still no marginal plates have been delineated on tergites three to five. In crab stage 4 there may be the initial early development of accessory, small, calcified nodules in the integumental margins adjacent to the lateral plates of tergites three to five. By crab 5 stage, sexual dimorphism is clearly apparent, with quite small, but well defined accessory marginal plates present on the right side in females (Fig. 4C5 female) and on both sides in males (Fig. 4C5 male). Similarly, the onset of female lateral plate asymmetry is unmistakable, particularly in tergite 5.
Lopholithodes (Figs. 5M-C2): The megalopa of L. mandtii Brandt, 1848, has the six pleonal tergites represented by individually distinct, moderately calcified plates, each with several identifying spines (Fig. 5M). With the molt to the crab stage 1, the first and second tergites remain entire; the third through fifth tergites now show incomplete or complete lateral sutures, dividing each tergite into median and lateral plates (Fig. 5C1). The median plates are each provided with spines corresponding to those seen in the megalopa. In crab stage 2, the first and second tergites fuse, either partially or completely; the third tergite is now divided into one median, two small accessory plates and two lateral plates, all separated by membranous areas (Fig. 5C2). The fourth tergite is now represented by a median and two lateral plates, as is the fifth, although the membranous areas of separations are more apparent in the fourth.
Fig. 5. Pleonal development of megalopa and juveniles to crab stage 1 or 2 of Lopholithodes mandtii Brandt, 1848, Phyllolithodes papillosus Brandt, 1848, Paralithodes brevipes H. Milne Edwards & Lucas, 1841, and Cryptolithodes sitchensis Brandt, 1853. Adult male pleon shown for all except L. mandtii. Color-coding and abbreviations as in Figs 2 and 3.
Phyllolithodes (Figs. 5M-adult):Only the megalopa and crab stage 1 of P. papillosus Brandt, 1848 have been examined. The six tergites of the megalopa are somewhat calcified, and all are provided with two or more spines (Fig. 5M). With the molt to crab stage 1, considerable decalcification and division has begun. The median and lateral plates of the second tergite are well separated, and evidence of the presumed upcoming separation of marginal plates is indicated (Fig. 5C1). The fifth tergite is entire, although a weak indication of the presumed separation of a median plate can be observed. The division of the median plate of tergite seen in the adult (Fig. 5 adult) is already indicated in crab stage 1.
Paralithodes (Figs. 5M-adult):The six pleomeres of the megalopa of P. brevipes are chitinous (Fig. 5M) and no true spines are developed, but possibly incipient spines are indicated on the first and second tergites by slight protuberances and setae. In contrast, these tergites are very weakly and partially calcified in P. camtschaticus Tilesius, 1815, with very prominent spines present on tergites 1 and 2 and smaller spines on tergites 3-5. With the molt to crab stage 1 in P. brevipes, the first, second, and sixth pleonal tergites and telson exhibit integumental calcification. In the first and sixth tergites calcification is complete, whereas calcification of the second is partial to complete. The second tergite also has now partially to completely divided into median, lateral, and marginal plates (Fig. 5C1). Tergites of pleomeres 3-5 remain chitinous; however, indications of lateral and marginal plate divisions are already apparent. Similar tergite division was illustrated by Kurata (1964) for crab stage 1 of P. camtschaticus. The decalcification and nodular development of the median plates seen in the adult (Fig. 5 adult) has not begun at the first crab stage in this species.
Cryptolithodes (Fig. 5M-adult): Megalopal and first crab stages of two species have been examined, C. sitchensis Brandt, 1853 and C. typicus Brandt, 1848. The megalopae of both have unarmed, chitinous tergites (Fig. 5M). Although some flexion in the pleon is apparent in the first crab stage, no marked changes in tergite development have occurred, except for very faint lateral depressions. Clearly the fusion of the first and second tergites and sundering of the third to fifth tergites seen in the adult (Fig. 5adult) proceeds much more slowly in Cryptolithodes.
Acantholithodes (cf. McLaughlin & Lemaitre 2001, figs. 4a-e): The megalopa of A. hispidus (Stimpson, 1860) has weakly calcified tergal plates, each provided with prominent spines. By the first crab stage, the second tergite has already divided into a single median, and paired lateral and marginal plates. On the marginal and lateral plates most of the megalopal spines still can be recognized. Marginally, tergites three and four are still distinguishable; however, decalcification is well underway centrally, with only a few chitinous patches and still calcified spines of the third still visible.
Hapalogaster (cf. McLaughlin & Lemaitre 2001, figs. 4f-k): Development of the megalopa and first crab stage of H. dentata (De Haan, 1844), and of the megalopal, crab stages 1 and 2 of H. mertensii Brandt, 1850, have been examined. In both species, chitinous tergites are present in both the megalopa and crab stage 1. In crab stage 1 of H. mertensii the surface is provided with scattered short setae, but granules are not apparent, and just slight lateral thickenings can be observed in the second tergite. Onlythe first and sixth tergites and telson are clearly delineated in crab stage 2.Segmentation of the primarily membranous second through fifth tergites is apparent only laterally. Very faint indications of incipient lateral and possibly marginal plates can be detected on the second tergite, and very slight lateral thickenings mark the third and fourth tergites.
Placetron (cf. McLaughlin & Lemaitre 2001, figs. 5a-d): In P. wosnessenskii Schalfeew, 1892, the six distinct tergites are chitinous and unarmed in both the megalopa and crab stage 1. Changes from the megalopa pleon are seen in early crab stage 1 with the third through fifth tergites becoming centrally narrower, with later crab stage 1 specimens showing indications of lateral plate delineation.
Oedignathus (cf. McLaughlin & Lemaitre 2001, figs. 5e-i): As in Placetron, the tergites of the megalopa of O. inermis (Stimpson, 1860) are simply chitinous. However, in crab stage 1, while the first and second tergites are quite distinct, tergites three to five are only clearly distinguishable laterally. The first tergite is still chitinous; the second has the marginal regions noticeably thickened, and the lateral areas are faintly indicated. The tergites of the third through fifth pleomeres have become completely membranous, although each is still faintly delineated. The dorsal pleonal surface is covered with scattered, tiny, short bristles and occasional minute spinules.
McLaughlin & Lemaitre (1997) expressed the belief that decalcification and sundering were major components in lithodid pleonal plate development, and to a certain extent the above summation has shown that this is true. McLaughlin & Lemaitre (2001) commented that while they did not have direct information on marginal plate development in Lopholithodes, it could be seen to a certain extent in Phyllolithodes. It also was very apparent that in Acantholithodes the marginal plates correspondingly arose from division of the existing lateral plates. Similar division of the lateral plates in Lithodes aequispinus (Fig. 3C2) and L. santolla (Fig. 4C1) has been observed at the first and/or second crab stage; however, contrary to expectations, by the third or fourth stage the marginal and lateral plates of the third through fifth tergites refuse (Figs. 3C3, 3C4; 4C2, 4C3), and the development of supplemental marginal nodules begins (Figs. 3C4; 4C3). Clearly this peripheral development leads to the “marginal” plates seen in the adult pleon of L. santolla (cf. Macpherson 1988: pl. 10, fig. B) and L. aequispinus, and these peripheral plates are not homologous with the marginal plates of the second tergites in these species.
These documented tergal transformations unequivocally refute Bouvier’s (1894a, b, 1895, 1897) hypothesis of lithodid plate formation, and his concept of lithodid carcinization. However, Boas’ (1880a, b, 1924) hypothesis, as it pertains to pleonal tergite development, did not require the invasion of an entirely membranous pleon by calcified nodules and their subsequent fusion. Boas thought, correctly, that the membranous condition and plate reduction seen in pagurid tergites represented decalcification. Nevertheless, the descriptions of ter-gite development in several lithodids by Crain & McLaughlin (2000a, b), McLaughlin & Lemaitre (2001), McLaughlin et al. (2001, 2003 ), and McLaughlin & Paul (2002) negate the validity of Boas’ (1924) explanation that the well calcified second tergite of Lithodes was a subsequent reinforcement of the membranous area between the two tergal plates of Pagurus. In all paguroid megalopae, the second tergite is a single, entire, commonly chitinous plate. As shown in the descriptions of megalopal and early crab stages of several Pagurus species (Carvacho 1988; McLaughlin et al. 1989, 1992, 1993; Crain & McLaughlin 1994) and of a species of Anapagurus Henderson, 1886 (Ingle 1990), the tergite of the second pleomere begins to lose its median distinctiveness only with the molt to crab stage 1, and this through decalcification and/or dechitinization, not division. In the genera of the subfamily Lithodinae, the second tergite, if not at least partially calcified at the megalopal stage, quickly becomes so in crab stage 1 or 2 in most genera. Concurrently or subsequently, this tergite may undergo division but not decalcification. In Cryptolithodes and some genera of the Hapalogastrinae calcification of the second tergite is lacking in the first or second juveniles stages, but this condition reflects heterochronic calcification, not loss. Boas (1924), like Bouvier (1894a, b, 1895, 1897) considered that the marginal plates of the Lithodinae tergites were new additions to the original tergites.
As has been shown, the marginal plates of the second tergite, when they are present, like the lateral plates, result of divisions of the original megalopal tergal plate. The marginal plates of tergites 3-5 do separate from the lateral plates in the second crab stage in Lithodes santolla and L. aequispinus and then refuse in the following stage(s). In these species, and in Paralomis granulosa, the subsequent development, in the adjacent marginal integument, of nodular areas of calcification that frequently fuse and form “accessory marginal plates” are new additions, but these are not homologous with the marginal plates of the second tergite. The marginal plates of the third through fifth tergites seen, or indicated in Phyllolithodes and Acantholithodes (McLaughlin & Lemaitre 2001: figs. 2g, 4b), similarly arise from division with the lateral plates at crab stage 1 or 2. That these refuse, as they do in Lithodes is most probable, given the adult structure of the marginal calcifications, but until additional stages are documented we can not be absolutely certain.
Having found no pleonal developmental evidence to support the prevailing doctrine of transformation of the lithodid pleon from that of a shell inhabiting pagurid, it is now necessary to examine the data available to see if it is possible to ascertain which of the two more publicized recent hypotheses, Cunningham et al.’s (1992) “from a hermit to a king” or McLaughlin & Lemaitre’s (1997) “from king to hermit” is most credible. Despite the fact that the former was based on molecular evidence and the latter on morphological evidence, both hypotheses acknowledged the close relationship between pagurids and lithodids. Cunningham et al. (1992) supplemented their DNA evidence with the shared lithodid-pagurid character of pleonal asymmetry, the presumed geological age of Pagurus as opposed to lithodids, and the purportedly comparable ontogeny of the coconut crab, Birgus latro (Linnaeus, 1758). McLaughlin and Lemaitre (1997) discussed asymmetry, with particular emphasis on pleopod and uropod asymmetries, from the stand point of adult morphology. They did not mention geologic age or address the presumed carcinization in Birgus latro. These latter two points can be clarified by perusal of the existing literature.
With regard to geologic age, suffice it to say that many of the fossil claws attributed to hermit crabs may or may not be accurately identified (Aquirre-Urreta & Olivero 1992). Factual fossil shell-dwelling pagurids appear to have been reported only by Mertin (1941), Hyden & Forest (1980) and Aquirre-Urreta & Olivero (1992), the latter confirming the presence of a pagurid as far back as the Cretaceous. However, the claim by Cunningham et al. (1992) of the relative youth of lithodids was recently challenged by Feldmann (1998) who described a species of Paralomis from the Miocene of New Zealand. Although the Miocene is appreciably younger than the Cretaceous, Feldmann suggested that the absence of lithodids in the fossil record probably was the result of the habit of most species to reside in deep water, and that the lithodid fossil record would go well beyond the Miocene when fossils were accurately recognized.
Borradaile (1916), Wolff (1961), Cunningham et al. (1992) and Morrison et al. (2002) all indicated that carcinization in Birgus latro represented a parallel evolutionary event, distinct from carcinization in the lithodids. We concur that carapace development in Birgus is distinct from that of lithodids. However, the broadening observed in the posterior carapace of B. latro was shown by Harms (1932) to result from changes in the animal’s respiratory mechanism as an adaptation to terrestrial life, and thus really is not homologous with the broadening of the brachyuran or even lithodid carapace. Harms (1932: 263) implied that the observed shortening of the pleon of young Birgus reflected muscular contraction, as he noted that in preserved specimens, the pleon was considerably more extended. The transformation of pleonal tergites from the soft condition seen in Coenobita species to the heavily thickened, chitinous or somewhat calcified plates that protect the adult pleon of Birgus required more than two years in the specimen that Harms observed and appeared to have been environmentally triggered. Other specimens that did not abandon their shells showed no such thickenings. Although descriptions of zoeal and megalopal stages are adequate for a number of coenobitids and Birgus, juvenile stages are not. Held (1963) and Reese & Kinzie (1968) presented interesting accounts of megalopal and early juvenile activities of B. latro, as did Brodie (1999) for first crab stage Coenobita compressus H. Milne Edwards, 1837, while Harms (1932, 1938) supplied limited information on pleonal tergite development and asymmetry. Unfortunately none of these accounts provided adequate data to permit the inclusion of either B. latro or any Coenobita species in the present analysis.
The mitochondrial gene rearrangement study of Morrison et al. (2002) purportedly supported Cunningham et al.’s (1992) conclusion of hermit crab ancestry, but what Morrison et al. (2002) actually demonstrated was just parallel evolution of the crab-like body form, a conclusion similarly reached by Borradaile (1916), Wolff (1961), and McLaughlin & Lemaitre (1997). Genetic information is without doubt a vital tool in investigating phylogenetic relationships, but the interpretations and conclusions drawn from supplemental information must be based on equally accurate facts. For example, Morrison et al. (2002) cited Blackstone (1989) to support their theory that changes in relative size and shape could easily be generated by a heterochronic shift in developmental timing. However, Blackstone’s data for ‘displacement heterochrony’ and the associated carcinization presumably exhibited by two populations of a species of a west coast hermit crab were subsequently shown by Crain & McLaughlin (1994) to simply reflect morphological differences between two distinct species. Morrison et al. (2002) offered no additional evidence, other than their interpreted “support” to contradict the McLaughlin and Lemaitre (1997) hypothesis.