Arrup Chamberlin, 1912

FODDAI, DONATELLA, BONATO, LUCIO, PEREIRA, LUIS ALBERTO & MINELLI, ALESSANDRO, 2003, Phylogeny and systematics of the Arrupinae (Chilopoda Geophilomorpha Mecistocephalidae) with the description of a new dwarfed species, Journal of Natural History 37 (10), pp. 1247-1267 : 1261-1265

Genus Arrup Chamberlin, 1912

Arrup Chamberlin, 1912: 654 .

= Prolamnonyx Silvestri, 1919: 84 (type species: Geophilus ? holstii Pocock, 1895 , by original designation) (synonymy: Crabill, 1964).

= Nodocephalus Attems, 1928: 115 (type species: Mecistocephalus edentulus Attems, 1904 , by original designation) (synonymy: Crabill, 1964).

Type species. Arrup pylorus Chamberlin, 1912 , by original designation.

Species included. A. holstii (Pocock, 1895) , A. edentulus (Attems, 1904) , A. pylorus Chamberlin, 1912 , A. sauteri (Silvestri, 1919) , A. dentatus (Takakuwa, 1934) , A. obtusus (Takakuwa, 1934) , A. pauroporus (Takakuwa, 1936) , A. doii (Takakuwa, 1940) , A. areolatus (Shinohara, 1957) , A. asiaticus (Titova, 1975) , A. mamaevi (Titova, 1975) .

Diagnosis. Forty-one pairs of legs. Body colour homogeneous, without darker patches. Body length 18–50 mm. Ratio of length to width of cephalic plate ca 1.2–1.5. Frontal line present. Two clypeal plagulae separated by a mid-longitudinal stripe, covering less than one-quarter of the clypeus. Clypeal setae roughly arranged in two lateral groups. Buccae without setae. Spiculum absent. Side-pieces of labrum fully divided into an anterior and a posterior ala. Internal margin of each anterior ala reduced to a point. Posterior margin of labrum not hairy. Coxosternum of the first maxillae undivided, areolation present at least in the median part. Coxosternum of the second maxillae undivided. Metameric pores placed in posterior position. Telopodite of the second maxillae slightly overreaching those of the first maxillae. Claws of second maxillae absent. Forcipular trochanteropraefemur with a distal tooth only. Intermediate articles of forcipules with or without teeth. Forcipular tergum without median sulcus. Sternal rhachides anteriorly not furcate. Last sternum subtriangular, longer than wide. Number of coxal pores different in different species. Anal pores present.

Geographic distribution. Central Asia; east China; Japan: Hokkaido, Hondo; Taiwan; west North America.

Remarks. New combinations are proposed here for the following species: Arrup areolatus (Shinohara, 1957) for Nodocephalus areolatus Shinohara, 1957 , Arrup asiaticus (Titova, 1975) for Nodocephalus asiaticus Titova, 1975 and Arrup mamaevi (Titova, 1975) for Prolamnonyx mamaevi Titova, 1975 . The new combinations, potentially deriving from Crabill’s (1964) synonymization of Nodocephalus Attems, 1928 and Prolamnonyx Silvestri, 1919 with Arrup Chamberlin, 1912 , are proposed here on the basis of a critical evaluation of each species and in the light of our cladistic analysis.

Biogeography

Most Arrupinae are endemic to small geographic regions. The basal taxon Partygarrupius is only known from very few localities on the Japanese island of Hokkaido and the representatives of Agnostrup occur from north-eastern China to Honshu; the geographic ranges of both these genera are thus contiguous but not overlapping. The genus Arrup instead has a wider distribution, centred in the Japanese region, but one species occurs along the western coast of North America (California) and two species in Central Asia, suggesting a fragmentation of a former wider holarctic distribution. This genus hence shows a distribution very similar to that of the genus Dicellophilus Cook, 1895 , but for the presence of the latter in Europe rather than in Central Asia.

Overall, the diversity of the Arrupinae is centred in north-eastern Asia, where most probably the centre of origin of the subfamily, possibly also that of the whole family Mecistocephalidae , may be located.

The distribution of the Arrupinae shows a distinctly vicariant pattern within the boreal temperate region, with a close parallel in the Dicellophilinae . A similar vicariant pattern is also shown by the basal Mecistocephalinae , but shifted towards the tropical zone. This distribution contrasts with that of the higher Mecistocephalinae , which include most of the mecistocephalid species, that show a distinct dispersal pattern through the tropics (Bonato et al., submitted).

The type material of the new species Nannarrup hoffmani was collected in Central Park, New York City, associated with numerous specimens (both adults and juveniles) of Henia vesuviana (Newport, 1845) and Schendyla nemorensis (C. L. Koch, 1836) , two introduced species from Europe already recorded from other North American sites. We can exclude a European origin for N. hoffmani , but we are pretty sure that its presence in New York City was due to introduction by human agency. This locality is widely separated from the distribution range of all other species of Arrupinae and of all Mecistocephalidae at large. The presence of this family in North America is limited to two species of Dicellophilus and Arrup pylorus , from California. The real provenance of N. hoffmani hence remains obscure, although either a west American or an east Asiatic origin may be guessed.

Another case of introduction by human activity is represented by Tygarrup intermedius Chamberlin, 1914 , the type species of Tygarrup Chamberlin, 1914 , that was described as coming from British Guyana but actually was ‘taken in Wash., D.C. in pots of plants imported from that country’ (Chamberlin, 1914). An additional case of introduction concerns the mecistocephalid specimen from coastal South America identified by Bücherl (1942) as Partygarrupius moiwaensis and considered by the author himself as imported from Japan. In both cases, single specimens were collected without any evidence of viable populations, but in the case of N. hoffmani a breeding population seems to have taken a foothold in New York, as the type series comprises both adults and juveniles, collected in different seasons.

Nannarrup hoffmani : a case of miniaturization?

Most mecistocephalids are medium-size geophilomorphs, with adult body length ranging from ca 3 to 13 cm, but for some Arrup species with adult body length just about 2 cm. Nannarrup hoffmani , with its much smaller size (1 cm), represents the absolute minimum reported in the family (figure 28).

Comparably reduced size, within geophilomorphs, is only known to occur in some genera of the Geophilidae (within Geophilus Leach, 1814 , Hyphydrophilus Pereira, Minelli and Barbieri, 1994 , Ribautia Brölemann, 1909 and Dinogeophilus Silvestri, 1919 , where D. oligopodus Pereira, 1984 , with 29 pairs of legs, is just 4.5 mm long), the Schendylidae (within Schendyla Bergsoe and Meinert, 1866 and Schendylops Cook, 1899 , where Schendylops oligopus Pereira, Minelli and Barbieri, 1995 , with (27)–29 pairs of legs, is 8–9 mm long) and the Ballophilidae (within Taeniolinum Pocock, 1893 and Ityphilus Cook, 1899 , where I. calinus , with 41–43 pairs of legs, is 9 mm long). Reduction in adult body size hence evolved independently in several derived lineages, often coupled with a secondary reduction to the smallest number of pairs of legs actually found in the respective family. No reduction in the number of body segments is involved in the case of Nannarrup hoffmani , but 41 pairs of legs is nevertheless the minimum and possibly the ancestral number within the family.

Similarly, an extreme increase in adult body size evolved independently in different lineages within the Geophilomorpha (in Mecistocephalidae , Himantariidae , Gonibregmatidae and Oryidae ), always coupled with the achievement of the highest number of segments found in the respective lineage. Within mecistocephalids, in particular, this is the case of Mecistocephalus Newport, 1843 , where the largest species are those with the highest numbers of pairs of legs (13 cm in both M. microporus Haase, 1887 with ca 100 pairs of legs and M. mirandus Pocock, 1895 with 65 pairs of legs).

To date, no particular attention has been devoted to the very reduced body size of some centipedes and to its possible interpretation as examples of miniaturization. Our new finding, however, gives stimulus for some thoughts.

Mainly with reference to vertebrates, Hanken and Wake (1993: 502) pointed to the difference between simple small body size and true miniaturization within a lineage, in that ‘miniaturisation involves not only small body size per se, but also the consequent and often dramatic effects of extreme size reduction on anatomy, physiology, ecology, life history, and behavior—the costs of size decrease, and its compensations’. According to these authors there is a ‘critical size’ for miniaturization, defined as ‘that at which important physiological or ecological functions such as feeding, locomotion or reproductive biology are affected, necessitating a major change in the way an organism deals with its ancestral adaptive zone’. In vertebrates, the most common effect of miniaturization is reduction and structural simplification (Hanken and Wake, 1993).

Some features of Nannarrup hoffmani are worth commenting upon as possibly due to miniaturization, similarly resulting in reduction and simplification, when contrasted to its closest relatives (figure 29). In particular, the number of ordinary setae and that of specialized sensilla covering the antennae are very low, clypeal setae are very few, the subdivision of the labral side-pieces into two alae is incomplete and extremely weak, the poison calyx is quite short and bears a low number of secondary ducts. For all these structures we can hypothesize a reduced expression due to the lack of organ-forming cell material or of suitable space. Conversely, the extreme reduction of the maxillary complex and the forcipular segment (figure 29) is hardly explained as simply correlated to the reduction of body size; mechanical adaptations related to the feeding function are probably involved.

Evidence from different animal groups shows that while structures such as muscles could be further miniaturized, the reduction of the brain cannot go beyond a minimum. A negative allometric relationship between body size and relative size of the brain have been described for vertebrates by Hanken and Wake (1993) and a similar effect is also known for beetle larvae and other miniaturized insects (Beutel and Haas, 1998).

All geophilomorphs lack eyes, their main sensory organs are the antennae, provided with ordinary setae and specialized sensilla of different shape. While the overall size of the antennae is proportional to body size both in N. hoffmani and its larger relatives, in the former the antennal sensilla (particularly the claviform sensilla) and the ordinary setae are larger and longer than expected (figure 29). These structures show a partial resistance to reduction, probably due to the fact the their size is close to the minimum size at which they can still work. The same explanation may be offered for the size of the ducts forming the poison calyx of N. hoffmani , which remain quite large in comparison with those of larger relatives.

The extreme reduction in body size does not seem to have affected the stability of features such as the setae on the clypeus and the cephalic plate, which still remain symmetrically patterned. In this respect, their behaviour is different from what was observed in the tiny spider Comaroma simonii Bertkau, 1878 ( Araneae : Anapidae ), just 1.6 mm body length, where the slit sense organs on the prosomal appendages have lost their usual patterned and symmetrical distribution (Kropf, 1997).

The complete lack of a frontal line on the cephalic plate apparently contrasts with other and possibly more conservative traits such as the presence of a homogeneous and well-marked areolation on the whole cephalic plate and the number and relative size of coxal pores. The cuticular cells (scutes) of some body regions (e.g. clypeus and antenna), in particular, maintain the same width (ca 10 m m) in all the three compared species, despite the variation in body size.