Smith & Roberts: Call repertoire of Litoria adelaidensis 
(A) 
12 
N ^ 
X 6 
50 ms 
♦ II* •»»•»« 
fc. % % i % * * * * 
50 
100 
150 ms 
(B) 
12 
N 
5 6 
* * 
* fc 
*** 
V 1 *t 
50 
100 
150 ms 
(C) 
12 
N 
X 
6 
* 
w 
0 50 100 150 ms 
(D) 
12 
N c - 
x 6 
0 
ms 
Figure 2. Waveforms (upper) and audiospectrograms (lower) of four types of acoustic signals produced by males of Litoria adelaidensis. 
(A) call type 1, (B) call type 2, (C) call type 3 and (D) call type 4. 
Call type 2 was similar to call type 1, but had two pulsed 
notes with a mean inter-note duration of 78.2 ms 
(maximum = 136.0 ms). All other calls consisted of a 
single note only. Call type 3 was a short pulsed call that 
was similar in duration to one of the notes in call type 2 
(Fig 3). Call type 3 was consistently shorter in duration 
and had fewer pulses (Fig 3) than the other call types. 
Call type 4 was only recorded from one individual but 
was often heard in choruses; it was an unpulsed whistle 
(n = 3, call duration 65.3 ms, low frequency peak 1.7 kHz, 
dominant frequency peak 3.5 kHz). All pulsed call types 
had two distinct frequency peaks (Figs 1 and 2) with the 
higher, or dominant, frequency peak the louder of the 
two (Table 1). 
Variation with temperature 
A significant amount of the variation in the both 
frequency peaks in call types 1, 2, and 3 was explained 
by differences in temperature. For low peak frequencies; 
call type 1: r 2 = 0.66 and b = 0.04 (F, g = 15.6, P < 0.01); 
call type 2: r 2 = 0.62 and b* = 0.04 (F,' 8 = 13.1, P < 0.01); 
and call type 3: r 2 = 0.85 and b y = 0.04 (F 1 8 = 43.6, P < 
0.01). For dominant peak frequencies; call type 1: r 2 = 0.60 
and b y = 0.09 (Fj g = 11.9, P < 0.01); call type 2: r 2 = 0.6 and 
b v = dio (F, a = 13.9, P < 0.01); and call type 3: r 2 = 0.77 
and b y = 0.10 (F, g = 26.7, P < 0.01). 
Variation with body size 
A significant amount of variation in dominant 
frequency was explained by body size for call types 1, 2 
and 3. For call type 1: r 2 = 0.61 and b y = -0.02 (Fj g = 12.6, P 
< 0.01); call type 2: r 2 = 0.53 and b = -0.03 (F l 8 = 9.1, P < 
0.05); call type 3: r 2 = 0.40 and b = -0.02 (F J g = 5.4, P < 
0.05). The low-frequency peak did not vary significantly 
with body size for any of the call types (call type 1 ■ P l 8 = 
0.80, P = 0.40, call type 2: ^ g = 1.3, P = 0.28, call type 3: 
F t g = 4.2, P = 0.07). 
A significant amount of the variation was explained 
by variation in body size for call duration and pulse 
number in call type 1 (r 2 = 0.46, b = -6.8, F \ g = 6.9, P < 
0.05 and r 2 = 0.57, b y = -0.65, g = 10.7, P < 0.05, 
respectively) and call duration and inter-note duration in 
call type 2 (r 2 = 0.55; b = -16.2, F 1 8 = 9.7, P < 0.05 and r 2 = 
0.47, b y = -9.5, F t g = 7.1; P < 0.05, respectively). 
Among call type analyses 
The three pulsed-calls varied significantly from each 
other in call duration (F 236 = 37.3, P < 0.01), pulse number 
(F 236 = 57.8, P < 0.01) and pulse rate (F 236 = 22.4, P < 0.01, 
Fig. 3) only. Call type 2 could be differentiated by the 
presence of two notes. Call types 1 and 3 could be 
differentiated by their pulse number and duration (Fig 
3). 
Call type 1 was produced most commonly (mean of 
the percentage of calls that were call type 1 for each 
individual = 60.3 ± 6.6%) followed by call type 3 (24.0 ± 
5.7%) and call type 2 (15.3 ± 3.7%). The individual that 
produced the type 4 calls did so on three occasions. 
Discriminant analysis 
The results of DFA confirmed the univariate analyses, 
showing that call types 1, 2, and 3 differed from each 
other in call duration, pulse number and pulse rate 
(Wilk's \ = 0.02, approx F 12 u = 25.1, P < 0.01; Fig 4). Even 
though some overlap occurred between call types in call 
duration, pulse number and pulse rate (Fig 3), the DFA 
produced clear separation of each call type (Fig 4). 
93 
