Journal of the Royal Society of Western Australia, 86(3), September 2003 
Materials and methods 
Terminology 
Anurans typically produce an advertisement call that 
is used to attract females and in some species to also 
mediate male-male interactions (Littlejohn 1977). In 
addition to the advertisement call, repertoires often 
include separate signals that function in male-male 
interactions such as an aggressive call, encounter call, or 
rivalry call (hereafter referred to as aggressive calls; 
Gerhardt & Huber 2002). Frogs may also produce short- 
range courtship signals (Gerhardt & Huber 2002) and/or 
distress and release calls (Bogert 1960). The 
advertisement call is usually repeated quasi-periodically 
for many hours per night during the breeding period 
(Narins et al. 2000). The aggressive calls are typically 
emitted in response to physical encounters with other 
males and/or when the calls of conspecific males exceed 
some threshold intensity (Brenowitz et al. 2001). 
Aggressive calls may also be emitted in response to 
physical or acoustic interactions with heterospecifics 
(Gerhardt 1994). 
Males of a number of species have developed 
considerable plasticity in their acoustic responses (cf 
Brenowitz et al. 2001). Nonetheless, the different signals 
produced by a male are typically spectrally and/or 
temporally distinct (Narins et al. 2000) and accordingly 
can be recorded, measured and categorised. We visually 
examined recording sequences and used conventional 
statistical analyses to look for consistently distinct 
acoustic signals with respect to spectral and/or temporal 
structure. 
Recording methods 
Calls of fourteen frogs in total were tape-recorded on 
the 21 August 2000 and 21 September 2000 at Perry 
Lakes, Perth, Western Australia (31° 56’ 40” S, 115° 46' 50" 
E). Calling males were concentrated along approximately 
40 m of shoreline, and the chorus contained about 40 
calling males on both occasions. Frogs called from 
emergent macrophytes and were usually perched 
between 5 and 60 cm above the water level. 
Recordings were made on a Sony Professional 
Walkman (WM-D6C; frequency response for type I 
cassette 40 - 15,000 Hz; ± 0.3% tape speed variation; 
manufacturers specifications) with a Beyer Dynamic 
M88N (C) microphone. Snout-vent length (± 1 mm) was 
obtained for 11 of the 14 frogs by placing the dorsal 
surface of the frog flat against a plastic ruler. Dry-bulb 
air temperature at each call-site (± 0.2 °C) was measured 
using a Miller &: Weber cloacal thermometer directly 
after the calling sequence of each frog was recorded. 
Temperatures ranged from 11 to 16 °C (14.7 ± sd 1.9 °C). 
Frogs were recorded for a minimum of 3.5 minutes 
(mean 5.2 ± sd 0.45). Calls were sampled at 44.1 kHz and 
analyzed using Cool Edit Pro (Syntrillium Software 
Corporation, Phoenix, v 1.2) running on an IBM 
microcomputer. 
Seven call properties were measured following the 
procedures of Littlejohn et al. (1993) and Gerhardt & Huber 
(2002). The call properties measured were call duration 
(ms), pulse number, pulse duration (ms), pulse rate 
(pulses s' 1 ), low frequency peak (kHz), dominant 
Frequency (kHz) 
Figure 1. Power spectrum derived from a type 1 call shows the 
two peaks used to obtain the low frequency peak and dominant 
frequency. 
frequency peak (kHz), and inter-note duration (ms). Pulse 
rate for call type 2 (see results) is reported as the average 
of the two notes. Dominant and low-peak frequencies (Fig 
1) were measured from a power spectrum of the entire 
call. Where a significant relationship was detected, call 
properties were standardized for temperature by 
converting them to the average call site temperature for 
the recordings used in this study (14.7 °C). To allow the 
determination of the direction of relationships, slopes (b ) 
are reported with all regression analyses. The first ten 
males to produce between three and five of each call type 
were used to investigate the differences between call types. 
Variation in each call property between call types 1, 2 and 
3 were analysed with single-factor ANOVA (Diekhoff 
1992). We also performed a discriminant function analysis 
(DFA; Diekhoff 1992) with the call properties as the 
dependent variables and three call types as the 
independent variable. All statistical analyses were 
conducted with Statistica (v 5.0). 
Results 
Call types 
The number of calls recorded for each frog varied 
from 4 to 109 (mean 37.5 ± 6.9). The waveforms revealed 
four distinct calls (Fig 2). Call type 1 was a pulsed call. 
Table 1 
Descriptive statistics for properties of call types 1, 2 and 3 of 
Litoria adelaidensis. Values are mean ± sd; sample size is 10 for 
each call property. 
Call type 1 
Call type 2 
Call type 3 
Call duration (ms) 
105.6 ± 26.5 
165.2 ± 51.0 
46.2 ±11.5 
Pulse number 
9.8 ± 1.6 
6.7 ± 1.8 
2.7 ± 0.41 
Pulse duration (ms) 
8.1 ± 0.96 
9.7 ± 1.5 
10.3 ±3.1 
Pulse rate (pulses s* 1 ) 
86.5 ± 12.0 
66.5 ± 15.3 
41.2 ± 5.7 
Low peak frequency 
1.8 ±0.05 
1.8 ±0.06 
1.8 ± 0.03 
(kHz) 
Dominant frequency 
3.8 ± 0.14 
3.8 ± 0.16 
3.7 ± 0.12 
(kHz) 
Inter-note duration (ms) 
78.2 ± 33.9 
92 
