Materials and methods
Vocalizations were recorded using portable tape recorders with external microphones and were analyzed either with the MEDAV sound analyzing system Spektro 3.2 (M. malagasius, M. striatus, M. ventrimaculatus) or on a PC using the software CoolEdit (Syntrillium Software Corp.) (M. horridus).
The following morphological measurements were taken with a calliper to the nearest 0.1 millimeter: SVL (snout-vent length), HW (head width), HL (head length), ED (horizontal eye diameter), END (eye-nostril distance), NSD (nostril-snout tip distance), NND (nostril-nostril distance), TD (horizontal tympanum diameter), HAL (hand length), FORL (forelimb length), HIL (hindlimb length), FOL (foot length), FOTL (foot length including tarsus), IMTL, IMTH (length and height of inner metatarsal tubercle), TL1 (length of first toe).
Institutional abbreviations are as follows: BMNH (The Natural History Museum, London); FAZC (Franco Andreone Zoological collection; preliminary numeration of specimens which will be deposited in MRSN); FMNH (Field Museum, Chicago); MNHN (Muséum National d’Histoire Naturelle, Paris); MRSN (Museo Regionale di Scienze Naturali, Torino); MSNG (Museo di Storia Naturale di Genova); NMBE (Naturhistorisches Museum Bern); PBZT (Parc Botanique et Zoologique de Tsimbazaza, Antananarivo); TM (Transvaal Museum, Pretoria); UADBA (Université d’Antananarivo, Département de Biologie Animale); ZFMK (Zoologisches Forschungsinstitut und Museum Alexander Koenig, Bonn); ZMA (Zoölogisch Museum, Amsterdam); ZSM (Zoologische Staatssammlung München).
Statistical analyses were carried out using SPSS for Windows, version 9. We performed Mann-Whitney U-tests to assess significance of intersexual and interspecific differences in size and morphometric ratios (relative tympanum length and head width, ratios TD/SVL and HW/SVL; relative size of inner metatarsal tubercle, IMTL/SVL and IMTH/SVL). Temporal and metric measurements are given as range, with mean ± standard deviation in parentheses. Terminology for the description of femoral and tibial glands follows Glaw et al. (2000). Number of crossbands on hand and foot are given to the tip of the longest finger or toe, respectively. The term “tibial gland” is here coined for the prominent gland structures on the dorsal surface of the tibiae of Mantidactylus horridus and M. malagasius (see descriptions in the corresponding sections). The terms femur, tibia, humerus and radius, in the descriptive species accounts, are used to refer to the corresponding external limb segments, and not to the bones.
Coordinates and altitude of collecting localities are listed in Vences and Glaw (submitted). Full names and geographic references of the campsites visited by F. Andreone and quoted in the text are as follows: (1) Masoala Campsite 2 = Forêt de Beanjada, 15°16.8’S, 49°59.8’E, 620 m, Masoala Peninsula, Antalaha Fivondronana, Antsiranana (Diégo Suarez) Faritany (Province); (2) Masoala Campsite 3 = Forêt d‘Andasin‘i Governera, 15°18.3‘S, 50°01.2‘E, 700 m, Masoala Peninsula, Antalaha Fivondronana, Antsiranana (Diégo Suarez) Faritany (Province); (3) Ilampy = Masoala Peninsula, Campsite 4 (Antsarahan’Amba-rarato), 15°23.52’S, 50°02.82’E, 510 m, Antalaha Fivondronana, Antsiranana Faritany (Diégo Suarez Province); (4) Tsararano Campsite 1 = Tsararano Forest, Campsite 1 (Antsarahan’ny Tsararano), Andapa Fivondronana, Antsiranana Faritany (Diégo Suarez Province), 14°54.4’S, 49°41.2’E, 700 m.
Osteological data refer to the cleared and stained specimens listed in Table III. Terminology and character definitions largely follow the accounts of Clarke (1981), Drewes (1984), Blommers-Schlösser (1993), Glaw and Vences (1994), and Glaw et al. (1998). Characters were analyzed with PAUP*, version 4 beta (Swofford 1998). We performed Maximum parsimony (MP) analyses (heuristic searches with TBR branch swapping), coding all characters as unordered. Multistate characters were coded as polymorphisms. Species of Heterixalus (Hyperoliidae) were used as outgroup (see Blommers-Schlösser, 1993; Glaw et al., 1998).