Natural history of Holoaden luederwaldti ( Amphibia : Strabomantidae : Holoadeninae ) in southeastern of Brazil

This study reports the rediscovery of Holoaden luederwaldti Miranda-Ribeiro, 1920 and provides information on the distribution, sexual dimorphism, reproduction and vocalization of a population of this species in Campos do Jordao, Sao Paulo (southeastern Brazil). Sampling was carried out in the Parque Estadual de Campos do Jordao (PECJ) from October 2005 through December 2008. Collecting was conducted using pitfall traps with a drift-fence on different altitudinal gradients (1,540 m, 1,780 m and 2,000 m a.s.l.). Fifty-two specimens of H. luederwaldti were collected in the PECJ. The mean snout-vent length (SVL) was 36.17 mm for males and 42.61 mm for females, indicating sexual dimorphism in body size. Holoaden luederwaldti occurred during the warm-rainy months. The population was distributed between 1500 and 2000 m, and the greater abundance was registered in well preserved forest areas. Mature females contained from 36 to 41 oocytes and the mean of oocyte diameter was 3.72 mm. The advertisement call of H. luederwaldti consists of simple notes composed of three harmonics. The record of the population of H. luederwaldti in the PECJ has reinforced the importance of investigating different areas of the forest when conducting faunal surveys.

Jordão, at altitudes ranging from 1,030 to 2,007 m.The PECJ covers an area of approximately 8,300 ha, and occupies onethird of the total area of the municipality of Campos do Jordão.
Collecting was carried out monthly in the PECJ, from October 2005 to December 2008, lasting four days on average.The sampling was conducted using lines of pitfall traps with a drift fence installed at different altitudinal gradients in three areas (spaced about 2.5 km): area I at 2,000 m, area II at 1,780 m, and area III at 1,540 m a.s.l.(Tropical Rainforest).On each altitudinal gradient area, three lines of pitfall traps with a 8 m drift-fence, containing 10 buckets (100 l), were installed, totaling 90 traps.Pitfall lines were installed with a distance of 150 m between each other.
All the specimens collected had their snout-vent length (SVL) measured with a caliper, to a precision of 0.01 mm.Voucher specimens are housed in the Scientific Collection of the Laboratório de Zoologia of the Universidade de Taubaté (CCLZU), Taubaté, São Paulo; in the Célio F.B. Haddad Collection (CFBH) of the Departamento de Zoologia, Universidade Estadual Paulista, UNESP, Rio Claro, São Paulo; and in the Museu de Zoologia da Universidade de São Paulo (MZUSP).
The vocalization of H. luederwaldti was recorded with a Marantz PMD-222 cassette recorder coupled to external semidirectional (ME 66) Sennheiser microphones.All recorded calls were edited with a sampling rate of 44,100 Hz and 16 bit per sample in the mono pattern.The bioacoustic analyses were performed on a computer using the program CoolEdit 96 (Syntryllium Software Corporation), with a 20,050 Hz sampling frequency.The 256 points option (Fast Fourier Transform, FFT) and, when necessary, the 1024 points option was used, mainly in determining fundamental frequencies.Five parameters of the advertisement calls were quantified: frequency bandwidth, fundamental frequency, dominant frequency, note duration, and note repetition rate.For the analyses and bioacoustic interpretations, the terms used follow MARTINS & JIM (2003).
The sex of adult specimens was determined from the snout-vent length (SVL), based on the smallest male with developed testicles.Females were determined by visual inspection, through the transparence of the ventral skin and presence of the oocytes.When necessary, abdominal incisions were made and the specimens observed by means of an optical stereomicroscope, which facilitated the confirmation of the sex.The sex of the juveniles was considered undetermined.
Females were dissected by ventral incision and had their ovaries removed.Contents of the ovaries were spread out in a Petri dish, in order to count the oocytes in each ovary.The largest diameter of the oocytes was measured, using a stereoscopic microscope with an ocular micrometer.For each dissected female, 9 to 15 oocytes were measured.
For the analyses of the abundance of specimens sampled per month in relation to the abiotic factors (total rainfall in the study months and maximum and minimum air temperature) a linear regression was used with a significance level of 5% (ZAR 1999).The parametric t-test was used for the analyses of sexual dimorphism between females and males.

RESULTS
I sampled 52 specimens of H. luederwaldti: 33 males, 11 females, and 8 juveniles (undetermined).The mean of SVL of the specimens were: males 36.17 ± 2.42 mm (31.02-40.48mm), females 42.61 ± 2.48 mm (39.37-46.53 mm), and juveniles 26.76 ± 1.43 mm (23.97-28.55 mm).The mean of SVL of females was bigger than males (t = 6.46; 1 df; p< 0.0001), indicating sexual dimorphism in size (Fig. 1).No other sexually dimorphic characters, even in color patterns or morphological structures (Figs 2-4) were observed.Females with mature oocytes were observed in October of 2005, 2006, and 2007, and in February and April of 2008.The number of oocytes in the females varied between 36 and 41 oocytes (37.4 ± 3.57, n = 5).The oocytes are large and unpigmented (Fig. 5).The mean diameter of the oocytes was 3.72 ± 0.8 mm (2.86-4.32mm, n = 96).The oocytes were light yellow, with a large quantity of vitelline (Fig. 5).A positive correlation was observed between the number and diameter of oocytes (Spearman's correlation, r s = 0.95, p = 0.013) (Fig. 6).There was no correlation between the SVL of the females and the number of oocytes (r s = 0.35, p = 0.552), or the diameter of the oocytes (r s = 0.16, p = 0.79).Females with mature oocytes also had some smaller oocytes in the initial stages of development.
Holoaden luederwaldti was registered in the rainy season (Fig. 7), however there was no linear relationship between the number of specimens sampled and the pluviosity (r 2 = 0.001, F = 0.04, p = 0.83).The abundance was slightly associated with the maximum temperature (r 2 = 0.12, F = 5.0, p = 0.03) but not ).The record of specimens at the end of the warm-rainy season revealed that there was still breeding activity, because three adult males and two females with well-developed oocytes were collected at that time.
In the Parque Estadual de Campos do Jordão, H. luederwaldti was distributed between the altitudinal gradient of 1,500 and 2,000 m a.s.l., and was found in all three sampled areas.The species was collected in greater abundance in areas II (n = 32) and I (n = 13), where the status of forest preservation is better, principally in area II, which lies in a depression between two hills, making it more difficult to access and therefore more difficult to impact by deforestation.The area III is In October 2005 and 2007 and April 2008, the calls of the collected individuals of H. luederwaldti (n = 4) which had been kept in plastic bags and in aquaria, were recorded in the field.The advertisement call of H. luederwaldti consists of simple notes composed by three harmonics (H) with a mean frequency range from 936 to 6244 Hz (Figs 8 and 9).The mean frequency ranges of the first harmonic (H1) (minimum and maximum frequencies) is between 936-1377 Hz (863-1455 Hz), with a mean dominant frequency (DF) of 1186 ± 31 Hz (1151-1256 Hz).Most of the energy of these notes is concentrated in the first-order harmonic (fundamental frequency = H1).The second harmonic (H2) shows minimum and maximum frequencies between 3247-3804 ± 94.8 Hz (3157-3995 Hz) and DF = 3557 ± 62 Hz (3428-3675 Hz).The third harmonic (H3) shows mean frequencies between 5626-6244 Hz (5475-6462 Hz) and DF = 5988 ± 83.5 Hz (5846-6166 Hz) (Figs 8 and 9).The duration of the note is 32.6 ± 1.9 ms (30-37 ms) on average, with a mean repetition rate of 252 ± 38.1 (210-308) notes per minute.

DISCUSSION
The geographical distribution of H. luederwaldti is somewhat uncertain.LUTZ (1958) reported a specimen of H. luederwaldti from the Serra da Bocaina in the state of Rio de Janeiro.HEYER et al. (1990)  Specimens analyzed at the Museu de Zoologia da Universidade de São Paulo and identified as H. luederwaldti (= Holoaden sp.) which had been collected in state of São Paulo from: Paranapiacaba (Santo André), Estação Ecológica da Boracéia, Campo de Fruticultura in the Serra da Bocaina (São José do Barreiro), and the Estação Ecológica da de Bananal, displayed differences in morphological characters when compared with specimens collected from the type locality (Campos do Jordão).POMBAL-JR et al. (2008) provided a map of the geographical distribution of the species of Holoaden, and referred to the species collected in the Atlantic Forest (between Boracéia and Bananal) as Holoaden sp.The present study reaffirms the endemic distribution of H. luederwaldti in the Mantiqueira range, in high-elevation forests (1,500 to 2,200 m a.s.l.) from Campos do Jordão, state of São Paulo, and the Parque Nacional do Itatiaia (PNI) located in the states of Minas Gerais and Rio de Janeiro.
The sampling of specimens of H. luederwaldti from the PECJ provided information about the sexual dimorphism in that population.Females are bigger than the males.This information on sexual dimorphism in the snout-vent length can be used for identification of the sex of specimens belonging to  No description of the vocalization of species of Holoaden had been published in the literature.Information on the emission of calls by species of Holoaden was reported by LUTZ (1958), who noted that specimens of H. luederwaldti collected in Parque Nacional do Itatiaia and the Serra da Bocaina (= Holoaden sp.) emitted a cry when caught.The same behavior was reported for H. bradei (LUTZ 1958).I did not observe, or record, this behavior for the specimens of H. luederwaldti caught in the PECJ It is possible that specimens emit this vocalization, probably a distress call (sensu BOGERT 1960), when they are manipulated in certain ways, but not in others. .The description of the advertisement call of H. luederwaldti given in this study will serve as a comparative basis for future studies with other species of Holoaden.
The record of a population of H. luederwaldti in the PECJ, after 40 years with no information on this species, indicates that we must survey different types of environments and diversify our sampling techniques.The record of H. luederwaldti was made possible by the use of pitfall traps, which have shown to be very efficient in recording species with cryptic habitats (often referred to as "rare").The use of pitfall traps in anurofauna surveys has revealed new species records (e.g., MARTINS et al. 2006) and the rediscovery of species considered threatened (e.g., ZAHER et al. 2005, BRASILEIRO et al. 2008) highlighting the importance of using these traps to widen our knowledge about the distributional patterns of anurans.
The scarcity of information about the geographical distribution, natural history, and ecology of anurans in the Serra da Mantiqueira limits our knowledge of their diversity, population density and size of the area of occurrence of the various woodland species, hindering the development of adequate public policies for their conservation (ROSSA-FERES et al. 2008).This situation has resulted in the compilation of controversial lists of threatened species (see PIMENTA et al. 2005), the great majority of which are considered data-deficient species, occasioning equivocal interpretations of the actual situation of different populations of Brazilian anurans.

Figure 1 .
Figure 1.Snout-vent length (SVL) of juvenile and adult male and female specimens of H. luederwaldti recorded in the Parque Estadual de Campos do Jordão, state of São Paulo, southeastern Brazil.

Figure 7 .
Figure 7. Abundance and temporal distribution of juveniles, males, and females of H. luederwaldti, and the maximum and minimum air temperature and total rainfall in the study months, recorded in the Parque Estadual de Campos do Jordão.
reported a registered specimen col-lected by A. Stanley Rand and Patricia Rand in 1963 at the Estação Ecológica da Boracéia, and ZAHER et al. (2005) registered H. luederwaldti in Estação Ecológica de Bananal, both in the state of São Paulo.