A NEW SPECIES OF Microgale (LIPOTYPHLA, TENRECIDAE) FROM ISOLATED FOREST IN SW MADAGASCAR 



161 



1 998 in the Analavelona Forest, 1 98 pit-fall bucket days (Raxworthy 

 et al., 1994; Goodman & Rasoloarison, 1997; Goodman, unpub- 

 lished). During the same periods, a combination of Sherman and 

 National Live traps were used for a total of 1,088 trap nights in the 

 Zombitse Forest, 955 trap nights in the Vohibasia Forest, and 535 

 trap nights in the Analavelona Forest. The only individual of M. 

 nasoloi taken in one of these devices in the Vohibasia Forest was in 

 a Sherman Live trap baited with a mixture of peanut butter and 

 ground corn flour placed about 1.5 m off the ground, and set about 

 1 m into the forest from the edge of an old road surrounded by forest 

 habitat. The single individual of this species obtained in the 

 Analavelona Forest was in a pit-fall device placed within 25 m of the 

 forest edge. Given the general efficiency of these two trapping 

 techniques in capturing a wide variety of non-volant small mam- 

 mals, including terrestrial and semi-arboreal lipotyphlans, it appears 

 that M. nasoloi is uncommon or difficult to trap with these devices. 

 Presumably this species also occurs in at least the nearby Zombitse 

 Forest and perhaps other smaller forest satellites that until their 

 recent fragmentation were part of an extensive area of transitional 

 forest. It was not found in the Isoky-Vohimena Forest (22°41.0'S, 

 44°49.8'E), lying between Zombitse and the PN de ITsalo which was 

 inventoried in late January 1996. 



Natural history 



The lipotyphlan fauna of Madagascar is much more diverse in 

 humid areas of the island and only a few species have been recorded 

 in the drier west and southwest. Other than those species mentioned 

 above for the Zombitse-Vohibasia and Analavelona forests, three 

 others have been reported in dry areas of the island. Microgale 

 brevicaudata is known from the northwest possibly as far south as 

 Morondava or Toliara (MacPhee, 1987; Raxworthy & Nussbaum, 

 1994; Ganzhorn et al, 1996); M. pusilla Major, 1896a from the 

 Mahafaly Plateau in the extreme southwest, although this material 

 recovered from owl pellets may date from a period in recent geologi- 

 cal times when this region was more mesic (MacPhee, 1986); and a 

 long-tailed Microgale associated with the longicaudata group from 

 near Morondava (Ade, 1996). 



Little biological data may be gleaned from the capture of the two 

 individuals of M. nasoloi. The Vohibasia animal was a pregnant 

 female with two embryos in the left and one in the right oviduct; the 

 embryos measuring 10 mm in crown to rump length. On the basis of 

 embryo size, the female was near parturition at the time of capture in 

 mid-January. In contrast to the data available for species of Microgale 

 recorded from eastern humid forest, no quantitative information on 

 the reproductive season of small lipotyphlans is available from the 

 southwestern portion of the island. Nevertheless, given that in the 

 eastern humid forest a considerable number of Microgale species 

 give birth during the early portion of the rainy season, which 

 normally commences in late November and early December, a mid- 

 January date for parturition would coincide with the beginning of the 

 rainy season in southwestern Madagascar which tends to occur later 

 than in the east (Donque, 1975). 



The individual from Analavelona was a male with small abdomi- 

 nal testes measuring 3x2 mm and non-convoluted epididymides. 

 Unfortunately, the skull is not available to assess the age of the 

 individual using dental characters, but on the basis of reproductive 

 condition this animal was probably a juvenile. Further evidence to 

 support this supposition is that the male is smaller than the adult 

 female in several external measurements and body mass, all charac- 

 ters that tend to vary with age. The pit-fall bucket in which the male 

 was captured contained the chewed remnants of beetles and cock- 

 roaches, which it presumably fed upon before being removed from 



the trap. 



The Vohibasia specimen was trapped 1 .5 m above the ground on a 

 vine running from the soil surface to the mid-canopy at an angle of 

 about 15° (Fig. 6), suggesting that it must be at least competent at 

 scrambling along supports. Anatomically however, it does not exhibit 

 the features normally associated with arboreality in other members of 

 the genus, since the relatively short tail and hindfoot suggest a greater 

 affinity for a mainly terrestrial lifestyle. In the most extreme cases, M. 

 longicaudata Thomas, 1882 and M. principula Thomas, 1926 have 

 very long, naked-tipped tails approximately twice as long as head and 

 body length, long hindfeet, and are demonstrably able to make use of 

 slender supports above the ground (Goodman & Jenkins, 1998). 

 Caution should be exercised in attributing morphological adaptations 

 to particular lifestyles, since Echinops, which lacks an external tail is 

 nevertheless an adept climber. 



The thesis expounded by Eisenberg & Gould (1970), that species 

 of Microgale may be divided into different locomotory classes 

 based on differences in tail and hindfoot length relative to head and 

 body length, was criticised by MacPhee (1987) because of lack of 

 ecological evidence. Recent direct observation, plus mainly circum- 

 stantial evidence from trap locations, suggest that many species of 

 Microgale are generalists equally at home on the ground as scram- 

 bling amongst lower levels of the understorey; while a few also use 

 additional ecological niches, such as the long-tailed M. longicaudata 

 and M. principula which are adept at exploiting thinner supports 

 above ground level. 



Microgale nasoloi exhibits some features - pale pelage, promi- 

 nent pinnae, short hindfoot relative to head and body length, skull 

 with a broad bimaxillary region, narrow interorbital constriction, 

 flat and broad braincase with pronounced superior articular facets 

 and marked lambdoid crest, well developed anterior dentition and 

 anteroposteriorly compressed M3 - which in combination are unique 

 to this species of Microgale. Many of these features are, however, 

 also present in the Malagasy geogaline tenrec, Geogale aurita, 

 while several are reminiscent of the suite of external, cranial and 

 dental characters which Hutterer (1986) used to define Afrosorex as 

 a subgenus of Crocidura (Lipotyphla: Soricidae). Species assigned 

 to Afrosorex inhabit savanna or forest-fringe areas and the pale 

 dorsal pelage coloration and prominent pinnae, shown also by 

 Geogale and M. nasoloi, are presumably adaptations to semi-xeric 

 conditions. The parallelism in dental features is possibly also an 

 example of similarities in dietary adaptations. One of the other three 

 species of Microgale known to occur in dry habitats is M. 

 brevicaudata, and this species also shows some features converging 

 on M. nasoloi, Geogale and Afrosorex. Externally all of these taxa 

 have prominent ears and short hindfeet, while all but M. nasoloi 

 have a markedly short tail, however M. brevicaudata shows none of 

 the craniodental features shared by M. nasoloi, Geogale, and 

 Afrosorex. This suggests that these shared external features are more 

 plastic than the cranial features and are thus more readily influenced 

 by the dry conditions of savanna or forest fringe habitats, or that 

 species such as M. brevicaudata have been adapting to dry or to less 

 extreme conditions for a shorter evolutionary period than others 

 such as M. nasoloi and Geogale. 



Biogeography 



Just a few kilometers from the Vohibasia Forest there is the 

 paleontological site of Ampoza, which has yielded a remarkable 

 amount of subfossil material that provides insight into environmen- 

 tal change in southwestern Madagascar over the past few millennia. 

 On the basis of current data derived from a pollen core at 

 Andolonomby (75 km SW from Analavelona and 140 km SW from 



