To comprehend the evolutionary morphological transformations undergone in Chirostyloidea, Aegloidea and Lomisoidea in general and those undergone in Lomis hirta and carcinization in particular, it is necessary to reconstruct the morphology of the last common ancestor of Chirostyloidea, Aegloidea and Lomisoidea. We consider those characters which are either shared by Galatheoidea (as outgroup) and one of the studied species or which are shared by all (or at least Kiwa puravida and Aegla cholchol) species as belonging to the ground pattern of Australopoda.
Since there is no consensus as to whether Aegloidea or Chirostyloidea is the sister group to Lomisoidea, we restrict our discussion largely to the morphological transformations from the Australopoda ground pattern to the representatives of Kiwa, Aegla and Lomis studied herein. Going by the phylogenies proposed by Schnabel et al. (2011) and Bracken-Grissom et al. (2011), it appears possible that the last common ancestor of all anomurans possessed a squat lobster-like habitus (see also Haug and Haug, 2014). The ancestral state reconstruction put forward by Bracken-Grissom et al. (2013) suggests that Australopoda has derived from a squat lobster-like ancestor itself, a hypothesis based on the sister group relationship of the Galatheoidea and Australopoda (for a detailed description of the anatomical ground pattern of Galatheoidea see Keiler et al. 2015a). This is the scenario which we also favour. In conclusion, there can be little doubt that Lomis is derived from a squat lobster-like ancestor.
The anatomy of the last common ancestor of Chirostyloidea, Aegloidea and Lomisoidea - the ground pattern of Australopodanext section
If we accept that a more or less squat lobster-like appearance is present in the ground pattern of Australopoda, we can assume that the external morphology corresponded broadly with that of galatheoid squat lobsters and Aegla in having a carapace slightly longer than it was broad, pleonal segments 1-4 visible from above and a plastron with an almost straight posterior margin possibly not covered by the bent pleon. Although Aegla broadly reflects the external morphology of the ground pattern of Australopoda, certain internal characters seem to be highly derived in the freshwater anomurans (Tab. 1). The tubular structure of the antennal bladder in Aegla is most probably derived since Lomis, Kiwa and Galatheoidea (see Keiler et al., 2015a) possess a lobed bladder, which in turn reflects the condition in the ground pattern of Australopoda. The tubular bladder resembles the tubular intervening duct of the antennal glands in crayfish (see Marchal, 1892), with both appearing to be an adaptation to a freshwater lifestyle (McNamara et al., 2015).
The morphological ground pattern of Chirostyloidea and the evolution of Kiwaidae
On the basis of their overall proportions in the dorsal view, Kiwa and Aegla can both be regarded as squat lobsters. They do differ from each other; however, in the proportions of their carapace and the shape of their pleon (see Fig. 2H). While Aegla displays proportions corresponding to those of galatheoid squat lobsters (Keiler et al., 2015a), Kiwa possesses a strongly bent pleon which is actually a feature of a crab-like habitus. The longer carapace in Kiwa, however, compensates for this shortening and imparts the overall squat lobster-like habitus. The proportions of Kiwaidae are probably derived within Chirostyloidea, since the two other chirostyloid taxa Eumunididae and Chirostylidae (of whose fresh material was unfortunately not available for this study) have proportions more similar to those in galatheoid squat lobsters (see Schnabel and Ahyong, 2010) and reflect the chirostyloid ground pattern (see also phylogeny in Roterman et al., 2013). The triangular posterior emargination of the plastron probably evolved independently in Kiwa and Lomis since a pronounced emargination of the plastron is not present in Eumunididae or Chirostylidae (see e.g. Baba et al., 2009).
The evolution of the sternal plastron and its impact on internal anatomy
The presence of a broad sternal plastron seems to be a prerequisite for the compaction and anterior shift of the cephalothoracic ganglion brought about by the disappearance of the barriers formed by the endosternites. The cephalothoracic ganglion in Aegla, Kiwa, Lomis and thus in the ground pattern of Australopoda is similarly compact as in Galatheoidea, but located more anteriorly than in Galatheoidea (Keiler et al., 2015a). Its compactness sets it apart from the more elongated form found in other decapods such as homarid lobster or axiid mud shrimp (see Bouvier, 1889). In lobsters, crayfish and mud shrimps, in which a broad plastron is not present, the endophragmal skeleton forms a narrow scaffold (see e.g., Secretan, 1998) restricting the leg nerves and thus exerting a constraint on the form of the cephalothoracic ganglion. The leg arteries in the ground pattern of Australopoda emanated separately and at roughly equal distances from each other, similarly to the situation in Kiwa (Fig. 5E) and galatheoid squat lobsters (see Keiler et al., 2015a). The fusion of the posterior leg arteries only appeared in the lineage leading towards Aegla. The position of and distances between the roots of leg arteries p1a-p4a correlate roughly with the position of the cephalothoracic ganglion and the degree of compaction of the neuropils (Fig. 5E). Though hemolymph supply to the cephalothoracic ganglion is affected by the ascending arteries which run between the roots of the leg arteries, the ascending arteries and the position of leg arteries p1a-p4a seem, at least to some degree, to be structurally dependent from, i.e. coherent with, each other. In other words, if the cephalothoracic ganglion and thus the ascending arteries became shifted anteriorly along the ventral vessel, the roots of leg arteries p1a-p4a became also shifted anteriorly. Or alternatively, the anterior shift in the joint between the ventral vessel and the descending artery (which is constrained by the need to puncture the cephalothoracic ganglion) possibly necessitated an anterior shift in the leg artery roots to ensure sufficient hemolymph supply to the legs by shortening the distance the hemolymph is required to flow.
Carcinization had an impact on the morphology of the cephalothorax in Lomis
In Lomis, endosternites 7/8 form a narrow passage, promoting the more posterior branch-off of the fifth leg arteries and thereby extending the distance between the fourth and fifth leg arteries, a situation similar to that in porcelain crabs (see Keiler et al., 2015a). In both Lomis and porcelain crabs, the narrow passage is caused by the posterior emargination of the plastron necessitated by the bend in the pleon which, in turn, is a direct result of carcinization. Curiously enough, the same posterior emargination and resulting narrow passage between the endosternites is found in Kiwa, too. The pattern of leg arteries in Kiwa, however, only slightly resembles that in Lomis, mainly as a result of the larger cephalothoracic ganglion in Kiwa whose larger neuropils (Fig. 5E) necessitate greater distances between the ascending arteries and thus between the leg arteries (see discussion above). The small but numerous endophragmal keels (together forming the median ridge) in Lomis serve as a structure for extra muscle attachment. A medial keel like that in Kiwa is also present in porcelain crabs (Keiler et al., 2015a), and both have a similar function to the median ridge in Lomis, compensating the loss of attachment area caused by the posterior emargination and the shortening of the midline of the plastron. In Kiwa and Lomis, the posterior margin of the plastron is elevated, forcing the ventral vessel to take a slightly elevated course. We assume that the cephalothoracic ganglion in the ground patttern of Australopoda was already located so far anterior that the emergence of a posterior emargination of the plastron did not have a significant effect on its position in Kiwa or Lomis, unlike in the evolution of porcelain crabs (see Keiler et al., 2015a).
Morphological transformations in the pleon – pleonal muscle reduction in Lomis is a result of carcinization
Besides the shape of the ventral vessel system, Lomis possesses several other internal anatomical features whose morphology obviously changed in coherence with external morphological transformations in the course of carcinization. The most obvious of these changes are associated with the pleon. While the ground pattern of Australopoda exhibits a pleon similar to that in Aegla or galatheoid squat lobsters, i.e. incompletely bent and featuring strong muscular bundles (Fig. 3D), Lomis evolved a flattened, more strongly bent pleon whose muscles are largely reduced (Fig. 3F), implying that it sees less active movement than in the ground pattern of Australopoda. The pleon in Aegla still permits a caridoid escape reaction (‘tail flipping’; our own observations; Martin and Abele, 1988) similar to that found in macruran decapods. In Lomis, however, the escape reaction has been lost completely in accordance with the species’ lifestyle hidden under stones in the intertidal zone (our own observations). A similar evolutionary scenario has been described for porcelain crabs (Keiler et al., 2015a), though the pleonal muscles in porcelain crabs are less drastically reduced than in Lomis and some porcelain crab species seem to have retained the ability to rapidly flip their pleon (Števčić, 1971; Hiller et al., 2010). The reduction of the pleonal muscles in Lomis in turn provided more space for the ovaries, a similar phenomenon to that which occurred in the evolution of porcelain crabs (Keiler et al., 2015a).
The anterior shift and compaction of the pleonal ganglia in Lomis is coherent with a crab-like habitus
As well as the pleonal muscles, the pleonal ganglia in Lomis also underwent changes in morphology in the course of carcinization. In the ground pattern of Australopoda, pleonal ganglia pn2-pn6 were probably arranged segmentally, one per segment, and at roughly the same distances from each other, a condition seen in Kiwa and galatheoid squat lobsters (see Keiler et al., 2015a). Pleonal ganglia play an important role in controlling the caridoid escape reaction and other rapid pleonal movements (e.g. Paul, 1989, 2003; Faulkes, 2008). Decapods which exhibit a high level of pleonal activity all possess segmentally arranged pleonal ganglia (see e.g. Bouvier, 1889; Jackson, 1913; Siller and Heitler, 1985; Paul, 2004). The pleonal ganglia in Lomis are proportionally smaller than in Aegla and Kiwa and situated much more anteriorly in the rear cephalothorax, a shift which possibly occurred in response to the loss of the caridoid escape reaction and to the hidden lifestyle. This transformation in the pleonal ganglia may thus also have been (indirectly) affected by carcinization, a theory corroborated by findings in other crab-like taxa (see Keiler et al., 2015a, b). A comprehensive review covering the similar morphological transformations in the independently evolved crab-like habitus, associated coherences and their relevance for convergent evolution is in preparation (Keiler et al., in prep.).