Variation in the Temporal Characteristics of the Vocalizations of Bullfinches, Pyrrhula pyrrhula

2. Tierpsychol., 38, 200-211 (1975) @ 1975 VerlJg Paul Parey, Berlin und Hamburg ISSN’ 0044-3573 / ASTM-Codcn: ZETIAG

Department o f Biology, University of Southampton, SO9 fi N H

Variation in the Temporal Characteristics of the Vocalizations of Bullfinches, Pyrrhula pyrrhula

By ROGER WILKINSON and P. E. HOWSE

With 4 figures

Received: 16. 5 . 1974

Introduction

Bullfinches when isolated call repeatedly b e f o r e r e g a i n i n g c o n t a c t with other group members. NICOLAI (1959) distinguishes between two types of call used in intraspecific social contact. According to this author the first of these calls is innate and used only over short distances, but the second call, used over considerable distances, is innate only to the extent that i t is fluting and monosyllabic. NICOLAI (1956) describes this “Lockruf” or luring call as that by which reconnection of lost birds is achieved. H e further states that during the breeding season the birds will only answer the calls of their mates. The fine individual differences between these calls cannot be resolved by the human car.

On the basis of anatomical, physiological and behavioural evidence, PUMPHREY (1961) has concluded that the avian cochlea “is designed to have a speed of response about ten times as high as that of the human ear”. This would imply that a considerable amount of information could be encoded in the temporal structure of bird vocalizations.

HUTCHINSON, STEVENSON and THORPE (1968) suggest that auditory data necessary for individual recognition in the Sandwich Tern (Sterna sandvicen- 52s) may reside in the temporal patterning of the “fish call”. In the Gannet (Sula bassana) there is sufficient information in the profile of the amplitude envelope of the landing call to distinguish one individual from another (WHITE, WHITE and THORPE 1970). JOUVENTIN (1972) has shown that the song of the Emperor Penguin (Aptenodytes forsteri) can be broken down into a series of wave-trains and that the temporal patterning of these is individually characteristic.

Playback experiments have shown that the spacing of the song figures or tempo” of the song may be important for species recognition in some mem-

bers of the Passeriformes (EMLEN 1972; M. SCHUBERT 1971; THOMPSON 1969) although in other species this may be of no importance (G. SCHUBERT 1971).

“

Variation in the Temporal Characteristics of the Vocalizations of Bullfinches 201

There is very little available information on the parameters that may be in- volved in individual recognition of passerine song. EMLEN (1972) suggests that details of “figure morphology” may be of importance to Indigo Buntings (Passerina cyanea) in differentiating between familiar and unfamiliar indi- v i d u a 1 s .

In order to determine whether temporal parameters are available for individual recognition in the Bullfinch an intensive analysis was made of the calls emitted by experimentally isolated individuals.

Methods

A female bullfinch (BTl) , six adult $ $ (B$1, B$4, B$5, B$8, B$9, B $ X ) and a juvenile 8 (BJ1) were maintained in the laboratory. They were individually caged and visually isolated from each other. Contact calls which appear to be comparable with NICOLAI’S second type were produced when one bird was removed to an adjoining room a t a distance of 10 ni f rom its fellows. These were recorded using a Nagra IV tape recorder and a Bruel and Kjaer half inch condensor microphone.

B q A , B $A, B $B and B $C were employed as decoys in Chardonneret traps on a fruit farm a t Swanwick, Hampshire, and were only recorded in the field situation. The laboratory birds also originated from this locality, with the exception of BY1 and B $ 1 which were obtained from County Durham. Each decoy bird was individually separated from the group and its resultant vocalizations recorded on a Tandberg Series I1 tape recorder using a Grampian DP4 dynamic microphone with a 60 cni parabolic reflector unit.

Additional information was obtained from tapes loaned by the British Library of Wildlife Sounds (AGF 1 , A G F 2 and alarm calls) and from the Wildlife Sound Recording

I

Society (RM 1 ) . Calls were analvsed on a Kav Electric ComDanv Sound SDectroeraph with the filter

& , v L

set a t the wide band (600 Hz) position. Call length was measured directly from the sonagrams to the nearest half millimetre. The “sequence” calls were separable into discrete elements, each element being defined as the smallest continuous band of sound energy. These elements were generally paired such that two elements occurred together as a single unit. Inter-element and inter-unit intervals were measured from the inception of one element to the inception of thc next. These measurements were subsequently converted into duration in milliseconds.

Modulation rate was estimated by dividing the number of distinct vertical bands within an element by the duration of that element. This procedure was validated by comparison of the values computed in this way with those obtained by oscillographie analysis.

Standard Deviation Mean x 100. Coefficients of variation were computed as

Results

Contact Calls

The contact calls show a high variability in duration within individual birds (Table 1) but a much greater variation between individuals (ANOVA, F 5/71 = 44, p < 0.1 %). Calls were also individually distinctive with regard to the presence or absence of starting transients and inodulation within the call. These variations were characteristic of B 9 1 and B c3 1 between which a call and answer sequence was observed. In other birds the calling relationship was less precise, calling by one bird initiating bouts of calls from the others. The difference in structure of the calls of B 9 1 and B 8 1 could also be inter- preted as a geographical variation, both these originated from County Dur- ham. The other bullfinches were obtained from Swanwick, Hampshire. Typ- ical contact calls for each individual are illustrated in Fig. 1.

202 ROGER WILKINSON and P. E. HOWSE

B ? 1 0 8 1 0.3 L 0 8 5 B 8 8 B d 9

- Mean Duration Standard Sample Coefficient

( msec 1 Deviation Size of Variation 365 ? 1L L L % 356 ? 50 10 14 "I. 229 ! 21 16 9 % L12 2 31 11 8 Ole

L80 2 70 1L 15 " l o LO5 $ 58 2 2 1 L "I.

O J

4 -

KHz

0,

I

9

-b

6 7 8

10 11 12

- I 1 I b Sec 'I 2

Fig. 1: 5. B$5, 6. B$4).

Bullfinch vocalizations. 1-6, Contact calls (1. B$1, 2. B01, 3. B$9, 4. B$S, 10, Alarm call (APR). 7-9, Loud calls (7. B$A, 8. B$B, 9. B$C).

11, ? Location call (B $C). 12, Long call (B $A)

Loud calls

These formed the majority of calls produced by the decoy birds when isolated from their fellows. The 0 (B 0 A) emitted very few loud calls and these were not recorded. The calls of the three 8 6 lacked any transient structure (Fig. 1) but those of B 8 B showed some evidence of modulation. The loud calls differed from the contact calls of the laboratory birds in that they showed a gradual decrease in frequency with duration. With the excep-


tion of B 6 B the loud calls were of shorter duration than the contact calls (Table 2).

Table 2: Mean duration in msec with standard deviation S.D., sample size N and coefficient of variation CV for the loud calls (left) and long calls (right)

(msec )

231 15 "I. * 172 19 5 "lo

BCrC 176 20 5 010

Duration ( m s e c )

315 ' 17 8 5 % 380 43 11 11 "I.

380 1

::. Call duration of B $ B excluding low energy terminal portion.

The high degree of inter-individual variability in call duration (ANOVA, F 2/56 = 49, p < 0.1 %) is greatly reduced if measurement of call duration in B 6 B excludes the low energy terminal portion (ANOVA, F 2/56 = 3.2, 5 % 5 p 5 10%). This terminal portion may not be perceivable by other birds, particularly in conditions of dense foliage or high environmental noise levels. The loud calls of J3 6 A and B 6 C cannot be distinguished from each other with regard to their duration (ANOVA, F 1/36 = 0.9, 20 % point = 1.7).

Long calls

The long calls have the structural characteristics of the aerial predator alarm calls of other species of birds. They are unmodulated tones which begin and end gradually. These calls were recorded from the three male decoy birds but were also noted from the laboratory birds which were obtained from Swanwick. In each bird they preceded or were interspersed with bouts of sequence calls.

The duration of the long calls is approximately twice that of the loud calls (Table 2). The calls of B 8 A and B 6 B cannot be separated on the basis of their duration (F 1/17 = 0.67, 20 % point = 1.8). The duration of the single recorded call of B 6 C falls within the range for B 8 A and B 6 B.

Location calls

Calls functional for location contain abrupt changes in frequency or transients. One such call was recorded from B 8 C (Fig. 1). This call is of shorter duration than the loud calls. The transients within this call may however be artifacts produced a t the onset of a recording session.

The elements X, Y and 2 of the sequence calls, described in detail below, also showed transient structure which by binaural comparisons would allow for easy location of the calling bird.

Sequence calls

These calls commonly occur in a set order and are almost invariably paired such that individual units consist of two elements. Single elements may also occur but these are very infrequent. The sequence calls were produced by the isolated decoy birds after an initial bout of loud calls. In the laboratory birds they tended to be interspersed with bouts of contact calls. Calls of this type were also incorporated in the songs recorded from two bullfinches. In this


context they were much softer and less easily audible than the loud sequence calls emitted by the isolated birds.

Local Variation i n Sequence Calls. Calls from different localities are illustrated in Fig. 2 .

2

Fig. 2: Local variation of sequence calls. a) AGF 2. b) AGF 1. c) B 8 8. d) 1. One call from B 8 9, 2. calls of R M 1

The Bullfinches from Swanwick, B 8 8 (Fig. 2 ) , B 8 A, B 8 B, B 8 C (Fig. 3), all showed similar sequence calls. Unit A + B consists of two elements of similar structure and precedes the unit C + D in which the second element, D, is of longer duration than the first. An additional unit, G + H, was recorded from B 8 8 such that the sequence A + B, C + D, G + H occurred fre- quently. No sequence calls were recorded from B 8 9 but a single element J, of similar structure to G and H'of B 6 8, was intermixed with bouts of contact calls. Two types of A -t 13 units were recorded from B ? A, but none of the subsequent units were heard.


8:

4 -

0 ,

1

Q

B C D

8'

4 -

0 -

4-

2 A B

- b -

1 A

rt

I V \=

D

,I I. I,

C D

*ryrsl. m*p

2 A B

c I

K H ~ J

Fig . 3 : Sequence calls a) 1. B 6 A , 2. BTA type 1. b) B 8 B. c) 1. B 8 C, 2. B 9 A type 2

Calls of siniilar sequential nature were noted from the tapes loaned by the British Library of Wildlife Sounds and The Wildlife Sound Recording Society. In AGF 2 (location given as Tippings Hill) the elements A and B were replaced by X and Y. These elements were distinct in being comprised of an initial portion showing rapid changes in frequency and a longer unmodulated portion. The uni t C + D was similar in structure to the unit C + D of the Swanwick bullfinches and followed by a shorter unit E + F.

AGF 1 (location given as Sudan allotments) also lacked the uni t A + B. It also showed the two distinct units C + D and E + F, these showed no fixed relationship i n temporal patterning and differed from each other in element duration. Three out of 17 of these units were preceded by the element 2. This element was similar to the long calls of the Swanwick birds but distinguished by the transient structure a t its inception and termination.

One series of calls (RM 2) recorded at Hartshill included paired elements similar to uni t A + B which were followed with great regularity by a call showing some resemblance to the loud calls of the Swanwick decoy Bullfinches.

Individual Variation in Sequence Calls. The calls were analysed in detail with regard to:

1. Element duration 2. Inter-element interval 3 . Inter-unit interval 4. Modulation rate.


B d ' A

B d ' 0

Bd ' C

B d 8

The results for the Swanwick birds are presented below. The constancy in temporal features of the sequence calls was also determined for AGF 1 and AGF 2. These showed an individual variability in element duration and modulation rate comparable to that of the Swanwick birds. The inter-unit intervals were again highly variable within the individual and the interelement intervals showed a variability similar to that of element duration.

1. Element Duration. Swanwick birds (Table 3 ) . Table 3 shows the mean duration in milliseconds of each element for

individual bullfinches. The birds were divided into three groups according to the situations in which the sequence calls were recorded. An analysis of variance was conducted on the raw data (element length in mm obtained from the sonograms). The laboratory bird (B 8 8) showed significantly more intra- individual variation in element duration than the other birds. This was shown by the differences in the average mean sums of squares (MSS) for these three groups.

MSS for B 8 8 = 14.67 for 42 df. MSS for B 8 A, B 8 B and B 8 C = 2.57 for 92 df. MSS for B J 1 and B 8 X = 4.45 for 16 df.

Average Average F tests showed the variance ratios between B 8 8 and each of the other

B 8 8/B 8 X and B J1, F4.JI6 = 3.3, 0,l % < p < 1 % B 8 8/B 8 A, B 8 B and B 8 C, F4JgL7 = 5.7, p 5 0.1 %

groups to be significant:

The variation within the decoy group (B 8 A, B 8 B and B 8 C) was analysed in detail. The inter-individual variation was high for elements A and B:

Element A, MSS = 183.74 for 2 df Element B, MSS = 136.75 for 2 df

but low for elements C and D: Element C, MSS = 26.82 for 2 df Element D, MSS = 6.35 for 2 df Inter-individual variation was significantly greater than intra-individual

variation for all elements.

A

130 2 11 ( 9 ) 8 010

154 2 10 [ 7 ) 6 ' l o

7 5 , 3 ( 6 ) L '10

159 2 23 (9 ) 1L %

Table 3: Mean duration in msec with standard deviation, (sample size), and coefficient of variation for each element A through D of sequence calls

126 2 12 ( 8 ) 1 0 %

1L5 2 22 ( 9 1 15 '10

7 9 2 L

256 2 10 136 f 10 ( 9 ) L % ( 9 ) 7 "I.

237 2 1L 12L t 17 (11) 6 " I o (11) 1L %

232 2 9 126 t 5

I

( 6 ) 5 % I (10) L O1.a ( 9 ) L %

0 I C I - 0 I

162 f 18 428 2 35 317 2 32

B J l

B d X

260 2 23 ( L ) 9 %

3L2 2 0 2L7 2 27 1 % ( 2 ) 0 % ( 2 ) 11"l.

196 2 15 ( L ) 7 010

196 2 3 ( 2 I 2


Inter -element i n te rva l s

A - B C - D

The between-bird variances for elements A and B were homogeneous: F2/? = 1.34, p > 1 0 % (A/B)

as were thc variances for the elements C + D: F2/? = 4.22, p > 10 % (C/D) The variation for A + B was significantly greater than that for C + D:

Inter - unit interval

B - c

A + B FJ4 = 9.66, 1 % < p < 5 %, +

323 ? 12

323 *_ 17

( 9 ) A " lo

(11) 5 %

328 f 13 ( 9 ) 4 " lo

L L 5 f LO (14) 9"I.a

308 ? 22 ( L ) 7 %

370 ? LO 12) 11" l .

The greater amount of variation in elements A and B suggests that the duration of these elements could be important parameters on which individual recognition might be based in the group of decoy birds.

2. Inter-element Interval (Table 4) Cocfficients of individual inter-element intervals were of similar magni-

tude to those for element duration. The raw data (before conversion to msec) was divided into three groups and analysed as in the previous section.

588 t 188

8OL ? 351

( 7 ) 32 '10

( 8 L L "10

3LO t 106 ( 6 ) 31"Io

L20 f 100 (11) 24 "I.

3 L 3 t 21 ( L ) 6 " lo

372 t 0 ( 2 0 %

ruble 4 : Inter-element and inter-unit interval duration in msec of sequence calls, with standard deviation, (sample size), and coefficient of variation

B d A

B d B

B d C

0d 8

BJ 1

B d X

290 t 19 ( 8 ) 7 '10

298 t 19 [ 7 1 6 " lo

239 f 8 [ 6 ) 3 '10

267 38 I 9 1 1L0/.

309 t 2L ( L ) 8 OIo

231 f 13 [ 2 ) 5 " lo

Analysis of variance showed a larger amount of intra-individual variation for element intervals in B 8 8 than in either of the other groups. This is shown below for the interval A - B:

MSS for B 8 8 = 25.23 for 8 df Average MSS for B J 1 & B 8 X = 4.88 for 18 df Average MSS for B 8 A, B 8 B & B 8 C = 7.97 for 4 df

MSS for B 8 8 = 27.88 for 13 df Average MSS for B 8 A, B 8 B & B 8 C = 3.50 for 26 df Average MSS for B J 1 & B 8 X = 13.28 for 4 df Thc variation within the decoy group was again analysed in detail. The

inter-individual variation was significantly greater than the intra-individual variation for the interval A - B (FJI8 = 22.80, p < 0.1 %) but not for the

and for the interval C - D:


A 0 C

( 9 ) 7 "/. ( 8 ) 9 "1. ( 9 ) la"/.

( 7 ) 13"/0 [ 8 ) 21 "10 (11 9

06C 179 t 9 166 t 22 120 t 25 [ 6 ) 5 "10 ( 6 ) 13"/o (10) 21 "1.

B 8 A 180 2 13 173 ? 16 115 ? 21

080 132 ? 17 112 2 2L 10L t 10

interval C - D (F2/?,, < 1 .O, non-significant). This inter-individual variation was significantly higher for the interval A - B than for the interval C - D:

Interval A - B, MSS = 111.26 for 2 df Interval C - D, MSS = 1.42 for 2 df F2/2 = 78.4, p (0.1 "/o ( A - B / C - D )

This suggests that the interval between the beginning of elements A and B in the unit A + B could be an additional cue for individual recognition in the group of decoy birds.

3 . Inter-unit Zntervul (Table 4). In the songs of B J 1 and B 8 X (Fig. 4) there is a great degree of constancy in the inter-unit interval. However, the sequence calls of the isolated birds show such a high variability within individuals that t.his parameter can be excluded as one upon which individual recognition can be based. This does not exclude it from having this function within the song.

D 87 2 32

[ 9 ) 37 %

103 21 (11) 20 "10

1L2 2 L7 19 ) 33 %

A B C D

I I I I I 0 Sec 1 2

F i g . 4 : Sequence calls in thc song. a) B$X. b) BJ1

Variation in thc Tcmporal Characteristics of the Vocalizations of Bullfinchcs 209

4. Modulation Rate. This was determined for B 8 A, B 8 B and B 8 C (Tablc 5). There is a distinct tendency for the intra-individual variability to increase froin a fairly low level in clement A to a vcry high level in elcmcnt D. It is unlikely that the modulation rates in the uni t C + D could code for individual recognition. In the unit A + B it is more probable that modulation rate could be an individually distinctive parameter; B 8 A and B 8 B can be better scparated by the modulation rate of element A (ANOVA, F 1/14 = 42.8, p < 0.1 %) than by any other teinporal feature of the sequence calls.

Discussion

It rcmains uncertain whether the vocalizations described as contact calls and loud calls are individual variations of one particular type of call or re- presciit distinct categories of vocalizations. The loud calls show a remarkable similarity to the alarm calls recorded by Philip ]<ADFORD (Fig. 1). These were obtaincd from a Bullfinch i n Pembrokeshire in response to a stuffed Kestrel (Falco tinnuncrtlus). Analysis showed them to be of shorter duration (mean = 139 msec) than the loud calls. I t is possible that this difference in duration rc- prcscnts a local variation in a functionally similar call or that the sanie call could function for both contact and alarm. In both situations the result is to attract birds towards the vicinity of the caller; i n one case for mobbing and in the other to re-establish social contact. It would be unnecessary for a call of this nature to be individually distinctive; the loud calls of B 8 A and B 8 C lack any distinctive structure and cannot be distinguished from each other in terms of their duration.

NICOLAI (1959) describes the luring call or "Lockruf" as monosyllabic and fluting. This description would apply equally well to the contact calls and to the loud calls. Because of their individual distinctiveness it is thought that of all the vocalizations recorded those designated as contact calls are the most likely to be synonyinous with the "Lockruf".

NEWTON (1972) states that Bullfinches are remarkably sedentary. This would assist thc formation of local dialects and could explain the differences in the structure of sequence calls in birds from different areas. It might also explain the difference in contact calls between the Swanwick and County Durham Bullfinches. The cumulative evidence for all the birds suggests that a high degree of individual distinctivcness is encoded in the teinporal features of the sequence calls. The unit A + B shows greatest constancy within individuals and a great variability between individuals. In all birds except B 8 X the elements A and B are almost identical. If, as it would seem, there is sufficient information in element A to identify the caller then the repetition of this i n elenient B could be of advantage to the receiver. Additional information may be encoded i n t h r inter-e!cinent interval.

The similarity i n the uni t C + D of the decoy birds could be a result of convergence through vocal imitation. MUNDINGER (1970) has reported in- stances of vocal imitation in American cardueline finches. No critical period existed for this learning of the calls from other individuals or other species. NICOLAI (1959) has shown that young Bullfinches learn the song and "Lock- ruf" of their father. This does not necessarily exclude the possibility of later modification of some of the song eleinents through vocal imitation. The convergence of the uni t C + D in the decoy birds could be related to their un- natural confincnient and may never occur in the wild. The learning ability

Z rlerp%)rhol Ed 2 5 , heir 2 14


of female Bullfinches is not so great as that of the d 6. In addition to learning elements of their fathers’ song they may also acquire elements from their partner’s song when they form the first true pair bond. The two types of unit A + B recorded from B 0 A might be explained by these observations of NICOLAI (1959). It is however interesting to note the similarity between type 1 of this unit and that of B CT A, and between type 2 and unit A + B of B 6 C.

The intra-individual variation in element duration is higher than that in many other species of birds. GREENEWALT (1968) has used intra-individual constancy in vocalizations as an estimate of the temporal resolving power of the bird. A lack of observed teinporal precision however does not automati- cally imply that the temporal resolution in Bullfinches is less acute than in other species.

MARLER (1 960) suggests that variation in song within the individual could contribute to the clues on which individual recognition is based. If this is true then an acute power of temporal resolution is necessary for the bird to be able to discriminate these differences. It would be interesting to determine if any relationship exists between the plasticity in vocal output of a bird, as shown by the intra-individual variability for a single call, and the amount of vocal imitation in the species to which it belongs.

The temporal resolving power of Bullfinches is currently being investi- gated using operant conditioning techniques to determine the ability of the bird to discriminate between single and dual click stimuli.

Acknowledgements

The authors are indebted to Ron KETTLE of the British Institute of Recorded Sound, Richard MARCOSCHIS of the Wildlife Sound Recording Society and the sound recordists, A. G. FIELDS and A. P. RADFORD. We wish to thank Dr . P. COOKE for statistical advice, Miss Gudula SCHULZ-PRUSS for the German Summary and Mr. E. WRIGHT for his valuable com- ments on the manuscript.

Summary

Temporal characteristics of the vocalizations of bullfinches were investi- gated.

The calls produced by experimentally isolated bullfinches were subjected to sonagraphic analysis. Within individual, between individual, and local variation in these calls was demonstrated. Several categories of calls were described, of which those termed “sequence calls” contained the greatest amount of information on which individual recognition could be based. Contact calls were individually distinctive in terms of their duration but with the exception of those produced by one 0 these lacked the structural complexity and information content of the sequence calls.

Zusammenfassung

Untersucht wurden zeitliche Eigenschaften von Gimpelrufen. Rufspektro- gramme von experimentell voneinander getrennt gehaltenen Gimpeln wurden analysiert. Die Rufe zeigen intraindividuelle, interindividuelle und lokale Unterscliiede. Mehrere Rufarten werden beschrieben, von denen ,,Sequenzrufe“

Variation in the Temporal Characteristics of the Vocalizations of Bullfinches 21 1

den groflten Iiiformatioiisgehalt haben; sie konnen individuellem Erkcnnen diencn. Kontaktrufe sind individuell verschieden lang; ihnen fehlt (ausgenom- men bei eineni 0 ) die strukturelle Komplexitat untl der Inforinationsgehalt dcr Sequenzrufe.

Literature cited

EMLEN, S. T. (1972): An experimental analysis of the parameters of bird song cliciting species recognition. Bchav. 41, 130-171 GREENEWALT, C. H. (1968): Bird song: Acoustics and Physiology. Smithsonian Institute Press, Washington, D . C . HUTCHINSON, R. E., J. G. STEVENSON and W. H. THORPE (1968): The basis for individual recognition by voice in the sandwich tcrn (Stcrna sandviccnsis). Bchav. 32, 150-157 JOUVENTIN, P. (1972): Un nouvcau systhnie de reccnnaissancc acoustique chcz les oiseaux. Bchav. 43, 176-185 MARLER, P. (1960): Bird Songs and Mate Selection. In : Animal Sounds and Communication (Ed. by W. E. LANYON and W. N . TAVOLGA), p.321-347. Washington, D . C . : American Institute of Biological Sciences e M U N D I N G E R , P. C. (1970): Vocal imitation and individual recognition of finch calls. Science 168, 480-482 NEWTON, 1. (1972): Finches. Collins: London N I C O L A I , J. (1956): Zur Biologie und Ethologic des Gimpcls (Pyrrhula pyrrhula L.). 2 . Tierpsychol. 13. 93-132 N I C O L A I , .J. (1959): Familientradition in der Gesangsentwicklung des Gimpels (Pyrrhula pyrrhula L.). J. Orn. 100, 39-46 PUMPHREY, R. J. (1961): Sensory Organs: Hearing. In: Biology and Comparative Physiology of Birds, Vol. 2 (Ed. by A. 1. MARSHALL), p. 69-86. Acadcmic Prcss: N e w York SCHUBERT, G. (1971): Experimentclle Untcrsuchungcn iiber die artkcnnzeichncndcn Parameter im Gcsang dcs Zilpzalps (Phylloscopus c. collybitu). Bchav. 38, 289-314 SCHUBERT, M. (1971): Untcrsuchungcn ubcr die reaktionsausloscndcn Signnlstrukturcn des Fitisgesanges, Phylloscopus t . trochilus, und das Verhaltcn gcgenuber artcigcnen Rufcn. Behav. 38, 250-288 THOMPSON, W. L. (1969): Song rccognition by territorial male buntings (Passerina). Anim. Behav. 17, 658-663 WHITE, S. J., R. E. C. WHITE and W. H. THORPE (1970): Acoustic basis for individual recognition in the Gannet. Naturc 225, 1156-1158.

Authors' address: R. WILKINSON and P. E. HOWSE, Department of B:ology, Southampton University, Southampton, S 0 9 . 5 N H , England

14::.

Documents

Variation in the Temporal Characteristics of the Vocalizations of Bullfinches, Pyrrhula pyrrhula