DIESEL AND HORST: BREEDING IN A SNAIL SHELL: SESARMA JARVISI 191 Fig. 16. Sesarma jarvisi. Scanning electron micrographs of plumose setae between the lower ridge of bran- chiostegite and the dorsal coxae. were found from March to August (Table Materials and Methods; Fig. 11) suggests a 4). minor reproductive activity in early spring If the development of these juveniles took followed by a higher peak of breeding ac- about 1-2 months, they must have hatched tivity in late summer (Fig. 17). from January to March, that is, before the spring rainy season. DISCUSSION We collected a sample of 93 brood-shells Sesarma jarvisi has adopted an extraor- in March and April 1991, containing ju- dinary breeding behavior in an adverse en- veniles with CW ranging from 1.3-8.5 mm vironment, the dry karst hills of central Ja- (see Table 5), and representing all juvenile maica. After the bromeliad crab Metopaulias age classes. The calculated development pe- depressus, it is the second species of Ja- riod, using the CW of these juveniles (see maican endemic Sesarminae for which the breeding behavior is known. So far, these Table 4. Number of ovigerous females, brood-shells two species are the only land-breeding with megalopae, and shells with juveniles <3-mm CW brachyuran crabs for which a detailed ac- found in Windsor, Jamaica, from March-December count of the breeding behavior and juvenile 1991. The search activities for shells differed within months, and brood-shells were only occasionally development in the field is available. Al- searched for in August and December, n = no obser- vations. Table 5. Number of juveniles of different size classes found in brood-shells in Windsor, Jamaica, 1991 (mean Mar Apr May Jun Jul Aug Sep Oct Nov Dec* CW in mm). Ovigerous females 1 2 0 0 1 0 n n n 1 1.3-2 2-3 3-4 4-5 5-6 6-7 7-8.5 N Megalopae 0 2 0 0 1 0 n n n 0 1 n n n 0 March 1 6 1 2 7 12 4 33 Juveniles 6 4 3 2 5 10 31 11 60 * R. Diesel, unpublished observations. April 2 3 0 3 barrier by dividing Jamaica's limestone lay- er into a small eastern plate (John Crow Mountains) and a large western plate. Se- sarma jarvisi is restricted to the western part, where it occurs from 100-m altitude at Windsor up to 930-m at Mount Diablo, at lower altitudes than originally recorded by Abele and Means (300 m; 1977). The wet limestone forests in the east are inhabited by the same ecomorph, Sesarma cookei Hartnoll, a species that is larger (CW > 20 mm) than S. jarvisi but lives in a similar habitat (Abele and Means, 1977; R. Diesel Fig. 17. Distribution of Sesarma jarvisi hatching dates calculated from the average juvenile size from broods and D. Horst, unpublished data). found in March and April 1991. Sesarma jarvisi is most common on shad- ed slopes and in moist valleys and is less abundant in karst depressions (cockpits) or though some information on breeding is re- on top of karst hills. The depressions consist ported for other terrestrial or fresh-water mainly of eroded bauxite soil. They receive crabs (e.g., Parathelphusa, see Fernando, more precipitation than hill tops (Aub, 1960; reviewed by Rabalais and Gore, 1985; 1969a), are frequently flooded (Aub, 1969b), Burggren and McMahon, 1988), no specific and lack a crevice system. Therefore, these information on their early development and "cockpits" are unlikely to support popula- parental behavior under field conditions is tions of S. jarvisi. The hill tops are prone available. to desiccation by radiation, and their crev- The behavior of S. jarvisi shows several ice systems may be too dry for the crabs. interesting features. First, this species uses The crevice system inhabited by S. jarvisi shells of land snails for larval and juvenile provides a relatively constant microclimate, development. Second, females manipulate with less extreme temperatures and a con- the microhabitat in the shells before releas- stant high humidity, compared to the forest ing their larvae, by carrying water into the floor. These microclimate conditions re- shells. Third, data on family composition semble those that prevail in the self-dug show that the mother crab and the juveniles burrows of other terrestrial crustaceans, e.g., remain together in the shell for at least 3 Hemilepistus reaumuri Audouin and Savig- months. The food items found in the brood- ny (see Edney, 1958); Uca panacea Novak shells, in particular the remains of mille- and Salmon (see Powers and Cole, 1976); pedes, suggest that the mother crabs also Holthuisana transversa Martens (see Green- feed their young. away and MacMillen, 1978). An extended crevice system also exists in Habitat and Distribution areas with lower annual precipitation (e.g., The distribution ofS. jarvisi is closely tied at Discovery Bay), but during dry and hot to areas with high precipitation (> 2,000 mm periods, diel fluctuations in the relative hu- annually) and wet limestone forests in west- midity affect the microclimate in the crevice ern Jamaica. In pre-Columbian times, dense system. The occurrence of erratic drops in forests covered almost the entire island (As- humidity might be one reason why S. jarvisi prey and Robbins, 1953), and the crabs does not live in such areas, but, the distri- probably had a wider distribution. After the bution of the species is affected by several colonization of Jamaica by man and the interconnected factors: (1) an extended rock- subsequent deforestation, in many areas only rubble layer of eroded limestone, common isolated islands of wet limestone forest re- west of the Blue Mountair.s and less so in mained (e.g., Mount Diablo and Dolphin the John Crow Mountains, a small lime- Head; Morrissey, 1983), leading to a partly stone plate in eastern Jamaica; (2) relatively disjunct distribution of the species today. constant high humidity in the crevice sys- The metamorphic Blue Mountains run from tem, which depends on a high average pre- north to south, forming a zoogeographic cipitation and forest cover above; and (3) the occurrence of large shells of land snails. Sesarma jarvisi is the smallest of the Ja- Jamaica supports 400-450 land-snail spe- maican endemic Sesarminae. Although egg cies, 95% of them endemic to Jamaica size is comparable to that of other species (Goodfriend, 1986a), whose distribution (see Hartnoll, 1964a; Abele and Means, depends on the occurrence of limestone (the 1977; R. Diesel, unpublished data), it pro- metamorphic and igneous Blue Mountains duces a very small clutch. Both small body support only few species, and population size and small clutch size appear to be ad- densities are low; R. Diesel, unpublished aptations to breeding in the limited space data). The snail-shell size and the aperture of snail shells. area vary within a species depending on lo- Another feature unique to the species is cal annual precipitation (Goodfriend, 1986b, their ability to transport and release water 1987). Shells in drier and hotter areas are into the brood shells. Water uptake from smaller than those in wetter locations and the substrate for gas exchange, drinking, or have a relatively smaller aperture area (Fig. feeding using tufts of setae is a common 14); shell volumes of P. lucerna are an av- feature in terrestrial and semiterrestrial crabs erage of 2.5 times higher in Windsor than (cf. Gross et al., 1966; Bliss, 1968; Hartnoll, in the Hellshire Hills (Fig. 3). Within the 1973; Quinn, 1980; Felgenhauer and Abele, distribution range of S. jarvisi the apertures 1983; Wolcott, 1984), but transport to and of shells of P. lucerna at Heron's Hill are release of the water into a "nest" (or burrow) almost too small for female crabs to enter has not been reported previously for crabs. (D. Horst, unpublished observation), which In addition to an external water film, S. may explain why shells of P. jamaicensis, jarvisi may also transport water in its bran- in spite of their large aperture area, were chial chambers. The mechanism by which more frequently used for breeding at Her- the crabs release the water into the shell on's Hill (Table 1). remains unknown. The water held within The preference of S. jarvisi for large shells the branchial chambers may be dumped out with relatively small apertures implies that into the shell, by means of a water-dumping retention of moisture in the shell is a sig- behavior like that reported for the ocypodid nificant microclimate factor for the species. crab Heloecius cordiformis (H. Milne Ed- Juveniles probably require a higher ambient wards) (see Maitland, 1990). humidity for survival than do adults. This Since juveniles remain in the brood-shell could be the reason that juveniles smaller after metamorphosis, they are not able to than 7-mm CW were never found outside forage outside and thus depend on the food their moist brood-shells. Future experi- resources found in the shell. We frequently ments on the transpiratory water loss of dif- observed food items of plant and animal ferent-sized juveniles under various humid- origin in shells with young up to 5-mm CW. ity regimes and measurements of the We have not seen mother crabs carrying humidity within various types of shells could food into the shell (experiments are under reveal the size at which juveniles are able way), but it is reasonable to assume that the to leave the brood-shell for extended activ- food is provided by the mother, which may ities in the crevice system, e.g., to forage. care for young until they reach 6.5-mm CW.