240
Discussion. The posterior group continued to approach all stimuli except T9
(song) longer than the anterior group (concordance P < 0.001). The posterior
group had difficulty in distinguishing the test stimulus from the expected call,
but the degree of generalisation extended only to the trill-like rhythms. As
motivation levels and learning are relatively unaffected by the lesion, it is sug-
gested that the selection of stimulus characteristics for the recognition pro-
cess might have been impaired, (i.e. more information was allowed through
for the matching process). When the received and expected stimulus are rela-
tively similar the anterior group also had difficulty in making a selective dis-
tinction (e.g. to T2).
Frequency of behavioural change
Results. I will first consider the frequency of change of the behavioural com-
ponents shown by all three groups to all nine stimuli. The mean number of
changes was 3.3 higher in the post-stimulus period than in the pre-stimulus
period. After T6 (arousal) the mean increase for all birds was 6.0. This in-
crease for a call with clear communicative content was significantly more
than for T5 (filter-trill, P < 0.025) and T2 (slow trill, P < 0.05) and twice as
high as the average post-stimulus increase. This shows (cf. Behaviour during
extinction, p. 235) that the frequency of behavioural change may be an indi-
cator of communication received, at least in terms of arousal elicited.
The intergroup differences according to behaviour changes (Table III) dis-
tinguish the stimulus group T135 from the other stimuli (concordance P <
0.01). The posterior group showed less behavioural change after Ti (P < 0.047)
and T3 (P < 0.02) in the postoperative poststimulus period than the anterior
group, and less, if compared to prestimulus levels, after T5 (P < 0.05). After
all other stimuli there were no marked differences, except for the iambus call
(T9). To T9 the posterior group showed more behavioural change than the an-
terior and sham-control groups (P < 0.025). (The posterior group also show-
ed non-significant increases to T2 and T6.)
Discussion. One may consider three categories of stimuli on the basis of ap-
proach duration and frequency of behavioural change. The first consists of
filter-trills (T1;3)5). The second consists of ‘novel’ signals (T4j7j8) that differ
in several respects from the train-trill (e.g. frequency of trill elements). The
third category (T2j6j9) possibly contains specific communicative properties
separate from the train-trill, (T6 — arousal, T9- song and T2, the slow trill
has similarities with a warning trill — Maier, 1980).
One should consider what the responses may represent in terms of the
nature of the stimuli. The arousal trill has structural properties similar to the
train-trill, but it has other message properties (Maier, 1980). In all birds it
elicited a high frequency of behavioural change (Table III). As approach be-
haviour seems to indicate that the stimulus has been recognised as similar to
the learned stimulus, it is suggested that a high degree of behavioural change