It is abundantly clear that there are a multitude of avenues of investigation remaining to be pursued before we can unravel the evolution of, and relationships within, the Anomura. Each bit of information, be it molecular, developmental, or morphological, contributes to our overall comprehension. Unquestionably, the evolution of a crab-like body form has occurred more than once. Whether each occurrence is equivalent to each other occurrence has yet to be ascertained. From our limited data, it appears that this body form arose independently in Lomis, the Porcellanidae, and the Lithodidae. It is most unlikely that there was a common precursor in the recent ancestral past, but certain morphological and developmental common trends weave a tantalizing thread.
In addition to the unsolved puzzles of asymmetry, pleopod loss and uropod specialization, this study also has not resolved the question of the plesiomorphy or apomorphy of direct development. Knowlton (1974) suggested that the basic unifiable sequence of metamorphic development in crustaceans could be reduced to three basic forms: nauplius, protozoea, and zoea. There has been considerable emphasis in recent years on the principal role of the nauplius larva in crustacean evolution (Björnberg 1986; Walossek & Müller 1990, 1998; Slack et al. 1993; Walossek 1993; Dahms 2000). However, Scholtz (2000), after an extensive review of crustacean nauplii, has suggested that the free living nauplius of euphausiids and penaeids well may have evolved from an ancestral egg-nauplius, thus being secondarily derived. If Scholtz’s hypothesis, in contrast to the opinions of Felgenhauer & Abele (1983) and Martin & Abele (1986), is substantiated, the free living penaeid nauplius and other attributes of the Dendrobranchiata, very well may not exhibit the most primitive of larval, megalopal and juvenile decapod characters. The concept of a secondarily derived, free-living nauplius also lends some credibility to Wolpert’s (1994) hypothesis that direct development is ancestral. Similarly, some of the developmental trends that have surfaced during the present investigation might be more readily explained through the primitive nature of direct development.
From this study we have been able to demonstrate that pleonal tergal plate development does not proceed as had been previously hypothesized. And for some genera we have been able to trace what actually is occurring in these tergites in early developmental stages. This information, however, leads immediately to the question: why is it such a complicated pattern, and how did it evolve? For an answer, we can look only to the data currently available. The lithodid megalopal tergites are entire, as they are in all decapods known, and as they are in most adults. Why then does all the sundering occur in lithodids? Two or possibly three processes seem to be involved: division, decalcification/dechitinization, and/or lack of calcium deposition. It is these processes that lead from the king to the hermit.
As has been demonstrated, in the most heavily calcified lithodid genera, Paralomis, Lopholithodes, Cryptolithodes, the pattern of tergal subdivision appears to restricted in tergites 3-5. The first and second tergites do not divide but instead fuse to form a single plate. In Paralomis minimal divisions in tergites 3-5 result in the tergites each represented by a single median and pair of lateral plates. There is no apparent decalcification except at the sutures delineating the plates. In the second crab stage of Lopholithodes, in addition to the division of tergites 3-5 into similar median and paired lateral plates, the accessory plates of tergite 3 have become delineated. Judging from the adult condition, we can assume that a similar division occurs in tergite 3 in Cryptolithodes; but at a later juvenile stage. In tergite 3 there is decalcification not only at the sutures separating the median and lateral plates, but in the area encircling each accessory plate. In both genera, tergites 4 and 5 undergo no further division and the only apparent areas of decalcification are at the sutures. We interpret these initial partitions and limited areas of calcium loss as the first of many steps on the evolutionary path to complete decalcification of pleonal tergites 2-5.
The steps along this pathway of decalcification are more apparent in the other genera of the Lithodinae as have been described in detail for species of Lithodes. In the Hapalogastrinae genus, Acantholithodes, where the megalopal tergites are also calcified, the process of decalcification is even further underway in crab stage 1 with prominently uncalcified areas separating tergites 1 and 2 and the latter now represented by a median and well-separated paired lateral and marginal plates. Decalcification is even more apparent in tergites 3-5, with the integument of median and portions of the lateral plates now predominantely chitinous. Further decalcification and tergal identity loss can be followed in other genera.
Except for the Pylochelidae, for which we still do not have any substantive larval, megalopal or juvenile data, the pleons of non-lithodid paguroids progress from chitinous, identifiable, entire tergal plates in the megalopae to partially chitinous, and ultimately, in many taxa, completely membranous integuments in subsequent crab stages, often with tergal identity lost, particularly on pleomeres 3-5. In some species at least, what have been referred to as lateral tergal plates, truly are remnants of the megalopal tergites, while in others the lateral “plates” on the left side of the pleon are simply thickenings of the integument for the attachment of the acetabulae of the pleopods. The range of variation in loss of tergal identity has been only meagerly investigated, but despite our present limited knowledge, it can confidently be said to be considerable. The direction unquestionably is toward loss, environmental influences not withstanding.
The megalopal/early juvenile data have shown additional directional changes as well. Ocular acicles, for example, are entirely absent in lithodid megalopae, as they are in a few, but not most other paguroids. In these few other paguroids, acicles develop in crab stages 1 or 2, and they are never subsequently lost. Acicles, as seen in other paguroids, do not subsequently develop in lithodids, although some minor calcifications on the second ocular segments occasionally have been reported (cf. McLaughlin 1983). When ocular acicles are viewed from the developmental prospective, it is clear that it is not loss of acicles in the lithodids, but initial absence, with acicles being gained in other paguroids.
With the metamorphic molt to megalopa, the fourth pereiopods, which have begun developing in the zoeal phase, emerge, in all lithodids, in the primitive condition of functional walking legs. In other paguroids, the fourth pereiopods, if they have begun forming in the zoeal phase, emerge at the megalopal molt as reduced, modified, and specialized appendages.
In both lithodids and other paguroids, a primitive unarmed telson is seen in the megalopa. In lithodids it remains unarmed in crab stage 1, but will in some species develop spinules on the dorsal surface in crab stage 2. In contrast, other paguroids progress from the unarmed megalopal telson frequently to a telson armed with marginal spines or spinules with the molt to crab stage 1. While the lithodid telson continues to be reduced in subsequent crab stages, it remains a well developed, often prominently armed structure in other paguroid crab stages. Clearly the transformation again is from simple (plesiomorphic) to complex (apomorphic).
From the larval phase data, general directional inferences also can be made. Although variable, the lithodid pleon in the last zoeal stage frequently still exhibits the presumably more primitive conditions of no separation or only partial separation of the sixth pleomere. Most frequently, at least in the Paguridae, distinct delineation of this pleomere is complete by the third zoeal stage.
Telsonal processes, if they change in number during the zoeal stages, usually increase by adding processes medially. However, the evolutionary transformation in telson processes appears to progress from “more to less”, the higher initial (ZI) number indicative of the more primitive condition. Lithodid zoeas most frequently hatch with eight to ten processes, whereas, the number in pagurids is commonly seven. These processes, primitively articulated in early zoeal stages, for the most part remain articulated in the last zoeal stages in lithodids, but frequently become fused (apomorphic) in paguroids, particularly in members of the Paguridae.
The evidence obtained from the megalopal/early juvenile development of lithodid pleonal tergites has demonstrated unequivocally that a transition from a shell-dwelling hermit crab to a fully calcified lithodid crab could not possibly have happened in the ways previously hypothesized. The supplemental juvenile, megalopal, and larval data presented here also bolster this finding. Both support the sundering and decalcification hypothesis of McLaughlin & Lemaitre (1997). In contrast, evidence purportedly confirming the more recent transitional hypothesis of “hermit to king” has been shown to be inaccurate and/or incorrectly interpreted.
Admittedly, the current investigation has not fully traced the complex evolutionary pathways involved in decalcification, or adequately addressed loss of symmetry. Nevertheless, to ignore its findings and blindly insist on acceptance of the “... [erroneous] hypotheses proposed by several generations of taxonomists ...” (Morrison et al. 2002: 349), albeit great minds of their times, would be a major step backwards for anomuran systematics. As the earth has been shown not to be flat, we have shown that the lowly hermit crab did not evolve into a mighty king.