Fig. [120].—Dorsal view of abdomen, A, of Cenobita, sp.; B, of Birgus latro. T, Telson; 1–6, 1st–6th abdominal segments.

If we look at the dorsal surface of the abdomen we find that, unlike that of the Hermit-crabs, it is completely protected by a number of hard plates (Fig. [120], B). Beneath the carapace can be seen a number of small plates belonging to the last thoracic segment; following these there are four large plates (1–4) representing the terga of the first four abdominal segments; the fifth, sixth, and the telson are, as has been stated, carried on the under side of the abdomen, but they are represented diagrammatically (5, 6, T) in the dorsal view. Besides the large terga, there are a number of small plates laterally, usually two to each segment, but they show a tendency to subdivide and increase in the largest specimens. This condition of affairs is very different to that in the naked fleshy abdomen of an ordinary Pagurid, but it can easily be deduced from that of the genus Cenobita, ordinary Hermit-crabs found in the Indo-Pacific Oceans, from which the Robber-crab has evidently descended. In Cenobita (Fig. [120], A) we see the same system of plates upon the dorsal surface of the abdomen, but they are much smaller, and the lateral plates are not so numerous; indeed, the greater part of the abdomen remains fleshy and uncalcified. The under surface of the abdomen shows the same rugosity as is found in Birgus, and from a number of other anatomical characters it is evident that the Robber-crab is a highly modified Cenobita that has deserted its shell and developed a symmetrical abdomen protected by expanded and hardened plates which represent those found in a reduced condition in Cenobita. The species of Cenobita although they inhabit shells and have normal branchiae, live on the shore, and have not been seen to descend actually into the sea.

The Lithodidae, which are found in temperate seas, especially on the Northern Pacific coasts (though Lithodes maia occurs in the North Sea, and certain species inhabit deep water in the Indian Ocean), have a deceptively Brachyuran appearance, the thorax being much shortened and the abdomen being much reduced and carried tightly flexed on to the ventral surface of the thorax. They live a free, unprotected existence, and are highly calcified. They are, however, certainly Pagurids, as is evidenced by a number of anatomical characters, but most clearly by the asymmetry of the abdomen, especially in the female, which is not only markedly asymmetrical in the arrangement of its dorsal plates (Fig. [121]), but also in the presence of three pleopods upon the left side only, as in Birgus. The male is without these appendages, and the sixth pair of pleopods is absent in both sexes. The remarkable calcified plates upon the abdomen bear a superficial resemblance to those in Birgus, but their evolution is traced, not from a Cenobite, but from an Eupagurine stock.[[144]]

Fig. [121].—Lithodes maia, ♀, in ventral view, × ¼. The abdomen is flexed on the thorax, so that its dorsal surface is seen. l.3, Lateral plates of third abdominal segment; l.5, left lateral plate of fifth abdominal segment; m, marginal plate; T, brush-like last pereiopod; Te.6, telson and sixth abdominal segment.

In some of the Eupagurinae, e.g. Pylopagurus, feebly calcified plates are present upon the segments of the abdomen (Fig. [122], A).

In the most primitive of the Lithodidae we witness the reduction (Fig. [122], B) and disappearance (C) of these original plates, their place being taken first by a number of irregularly situated small spines and warts, which, however, subsequently fuse up to form definite segmental plates. In Lithodes maia, ♂ (D), there are a series of lateral and marginal plates, while in Acantholithus (E) a number of median plates appear, presumably by the fusion of the small spines present in the median line in Lithodes maia; finally, a fusion of the marginal and lateral plates may take place, so that each abdominal segment is covered by a median and two paired lateral plates.

Fig. [122].—Diagrams of abdomen: A, of Pylopagurus, sp.; B, of Hapalogaster cavicauda; C, of Dermaturus hispidus; D, of Lithodes maia, ♂; E, of Acantholithus hystrix. c, Central plates; l, lateral plates; m, marginal plates; T, telson; 1–6, 1st–6th abdominal segments. (After Bouvier.)

It is to be noted that the males and females of the various species do not follow a parallel course of development, the plates in the male being symmetrical, while those of the female are often highly asymmetrical (compare Figs. [122], D, and 121), thus giving the strongest evidence of a Pagurid ancestry.