«SONG “DIRECTS” IN THREE POPULATIONS OF WHITE-CROWNED SPARROWS By P. MARLERand M. TAMURA The phenomenon of “dialect” variation in bird ...»
368 Vol. 64
SONG “DIRECTS” IN THREE POPULATIONS
OF WHITE-CROWNED SPARROWS
By P. MARLERand M. TAMURA
The phenomenon of “dialect” variation in bird song,appearing as a consistent differ-
ence in the predominant song type between one population and another of the same species, has a special interest for biologists, serving as a focus for attention in discus- sion of such diverse topics as speciation (for example, Huxley, 1942; Mayr, 1942), learning (Thorpe, 1954, 1958) and the mechanisms of social communication (Marler, 1959). The White-crowned Sparrow (Zonotrichia leucophrys) affords one of the best known casesof such “dialect” variation among North American birds, and it has been commented upon by many who have observed this species (Blanchard, 1941; Peterson, 1941). Before the ontogenetic basis of such local song variation can be assessed and before its evolutionary significance can be satisfactorily determined, careful descrip- tions of the nature and extent of the variation are required. This paper seeksto provide some of this necessary information by describing song variation in the individual and in a population, both at one time and from year to year, and also by comparing songs in three populations, two adjacent and one distant.
The three areas where recordings were made are: (1) Sunset Beach State Park, Santa Cruz County, California, about one hundred miles to the south of Berkeley;
(2) the central part of the city of Berkeley, Alameda County, California; (3) the area around Inspiration Point, Contra Costa County, two miles to the northeast from the center of Berkeley. The region between the Berkeley and Inspiration Point areas is more or less continuously occupied by White-crowned Sparrows, while much of that separating these two
Fig. 1. A and B, sonagrams of two song types, illustrating the terminology used in describing different parts of the song. C, three songs from one bird illustrating the two main types of variation found in songs of the same individual. The top song is complete, the middle song lacks one C-syllable, and the bottom one has a curtailed A-phrase in the whistle portion of the song. The sonagrams in this and other figures are photographs of sonagrams which were retouched with white paint to mask inscriptions and some background noise.
istics prevail not far away. Thus birds in Marin County, California, may have a whistle at the end of the song as well as at the beginning, and Peterson mentions other deviations from the general pattern. Two typical examples from the study areas are shown in figures 1A and 1B.
To describe the parts of the song we have arbitrarily divided it into two parts, the whistle portion and the trill portion. As in previous studies the term note is used for any continuous trace on the sonagram, a phrase is a note or group of notes in the whistle portion, which is not immediately repeated, and a syllable is a note or group of notes which is usually repeated one or more times, making up the trill portion (see Marler and Isaac, 1961). Although trill syllables are occasionally only given once we use the same term for them since in other respectsthey appear identical to those which are repeated. The two main parts of the whistle portion have been labeled phrases A and B. The syllables in the trill portion are labeled syllables A to C in the order in which they appear. In no case were there more than four syllable types in a trill.
Songsare often delivered in long sessionsfrom the same perch. During such sessions of singing behavior there is a rather regular pattern of delivery. For example the Berkeley birds, with an average song duration of 1.9 seconds,paused about 11 seconds between songs, thus spending about 17 per cent of singing time in actual song.
VARIATIONS IN INDIVIDUAL BIRDSMany hours of listening to White-crowned Sparrow singing have left us with the impression that, after a period of plasticity in early spring, each bird sings essentially the same song pattern from about March to June. In a series of 3.5 consecutive songs from one bird which was analyzed, the duration of the songs varied somewhat (1.75 2 0.61 seconds), but inspection of the sonagrams revealed the same basic pattern throughout (fig. 1C). Variation in over-all duration occurred either because the long notes in the whistle portion varied in duration in successive renderings, or because syllables were omitted from the last part of the trill, the number of C syllables varying between four and five. We may note that regardless of whether four or five syllables were given, the last one always terminated with a lower frequency than the others.
Thus the variation does not arise simply from omission of the last syllable (fig. 1C).
Variation in the whistle portion of the song seemsto be largely a result of fluctuations in the amplitude of the notes, although this is difficult to distinguish from actual omission of some of the parts. In spite of these variations, the general impression is one of consistent conformity to a single song pattern, at least through one year and, on the basis of studies of captive birds, probably throughout adult life.
VARIATION OF POPULATIONSThe Sunset Beach population in 1959.-Statistics on the songsof 26 birds recorded at Sunset Beach in April, 1959, reveal a relatively homogeneoussample (table 1). The data are derived from one representative sonagram from the recordings of each bird.
The over-all duration ranges from 1.7 to 2.6 seconds with a mean value of 2.1 * 0.2 seconds. Evidently song duration varies relatively little within this population, and the same may be said of the number of notes and syllables plus phrases (31.1 f 6.4 notes, 10.2 f 1.8 syllables and phrases). Similarly the maximum and minimum frequencies are relatively consistent. As has been found in studies of other species,the minimum frequencies represented on the sonagrams vary less than the maximum frequencies (Marler and Isaac, 1960a, 1960& 1961).
If the over-all quantitative properties of the songs give an impression of homogeneity, this is further reinforced by a survey of their qualitative characteristics. Figure 2 shows the songs of eighteen of these birds from 1959 selected for illustration because of the lack of background noise. Consider first the trill portions, where there is a
remarkable degree of correspondence in the structure of the individual syllables.
Syllable A has a similar structure in all of the songs, with minor variations in the shape and number of notes (see Song 0). There are usually two groups of notes in each syllable (for example, A-F, I-M, P-R) but occasionally the groups constitute individual syllables (H, N, 0). In one case only half of the second syllable A is given (Song G). Syllable B which is usually given once, and occasionally two or three times (Songs M, N and 0), has an almost identical structure in all of the birds, although the number of notes may vary between two and four. Syllable C also varies little from bird to bird (compare A, B, J, L, M, 0, P and Q, for example).
In the whistle portion of the song, phrases A and B are present in every case.
Phrase A consists of one long tone of almost constant frequency, sometimes divided into two or more notes, either by complete breaks or by more moderate reductions of intensity. Phrase B is more variable. Sometimes it is a pure tone (O-R). Usually it is a vibrato note (A-N), the rate of which varies from about 50 to about 200 oscillations per second in different birds. In some individuals in which phrase B is not a vibrato there is a note or group of notes between phrases A and B (0 and P), adding to the impression of greater qualitative variability in the whistle portion of the song, as compared with the trill portion.
The Sunset Beach Birds in 1960.-In order to gain some impression of the stability of the song types from year to year, we returned to exactly the same area in May, 1960, and recorded songs of 20 birds, 16 of which are shown in figure 3. The birds were not banded so that we have no idea how many individuals are represented in both samples.
The correspondencein song types between the two years is striking. There is a similar proportion of vibrati in phrase B of the whistle portion. The structure of syllables A, B and C in the trill portion of the songsis closely similar. The over-all statistics for song duration, numbers of notes and syllables, and maximum and minimum frequencies vary little from the previous year (table 1). A few birds in which phrase B is a pure tone also have short notes interposed between phrases A and B, as in the 1959 sample (fig. 2, 0, P; fig. 3, P). Our conclusion is that the song characteristics are stable within this population over a period of two years. On the basis of what is known of the longevity of passerines (Lack, 1954), it seems reasonable to assume that some newcomers were represented in the 1960 population, and we may infer that young birds develop a song type which resembles that of the population within which they are living.
Songs of the Berkeley population.-In view of the lack of variability between the Sunset Beach samples of 1959 and 1960, the Berkeley records for the two years have been merged, giving a total of 13 birds. The over-all statistics are similar to those for the Sunset Beach birds (table I), except for the somewhat larger number of syllables in the trill portion of the song. Once more the sample is relatively homogeneous. The pattern of the songs is similar in the two areas, with the whistle portion preceding the trill. Once more the B phrase may be either a whistle or a vibrato note. However it is clear from qualitative inspection that there are striking differences between the two populations in the structure of the syllables which make up the trill (fig. 4, I-P).
Of the three basic syllable types in each trill, syllable C is most clearly and consistently different. Each syllable consists of only one or two notes, instead of the three notes in the Sunset Beach songs,and the notes have a different shape. In addition the last C syllable in the trill of the Berkeley birds often descendsto a lower frequency than those which precede it, a phenomenon never encountered in the Sunset Beach songs.The Berkeley songshave no equivalent of the elaborate B-syllable of the Sunset Beach sample. However the Berkeley A and B syllables, which are generally fairly similar to each other, do bear some resemblance to the A syllables of the Sunset Beach songs. In this and other respects the songs of the two populations have characteristics in common. Nevertheless those differences which exist are unambiguous and consistent, so that the home locality of a single bird chosen at random could be accurately assigned on the basis of its song pattern.
Songs of the Inspiration Point population.-The ten birds recorded in May of 1960 at Inspiration Point were living in a restricted area about two miles from the center of Berkeley. Measurements of the sonagrams once more reveal a very homogeneous sample, similar in most respects to those from the other populations (table 1). There were more syllables and phrases per song than in Sunset Beach songs. In this they resembled the Berkeley songs, and the resemblance is still more evident in the qualitative properties of the songs (fig. 4, A-H). Syllables A and B in the trill are very similar in the two samples. Only the C syllables of the trill show any consistent difference, with a slightly more elaborate structure than those in the Berkeley songs, and with a maximum frequency which is closer to that in the earlier parts of the trill. The last C syllable often descendsto a lower frequency than the preceding ones, as in the Berkeley songs.
The trill portion of the song is remarkably consistent throughout the sample. Both phrases A and B had a very similar structure in all birds, all of the B phrases being Vol. 64 THE. CONDOR 374 Fig. 4. A-H, songs of eight birds recorded at Inspiration Point, Contra Costa County, in May,
1960. I-P, songs of eight birds recorded in Berkeley, Alameda County, in April, 1959, and May, 1960.
The B phrasesof the Berkeley songs,which may be either vibrato or non-vibrato, vibrati.
are so variable in this respect that it is difficult to be sure whether the difference is consistent. All of the vibrati in our sample of Berkeley songshad a narrower frequency span than those in the Inspiration PO nt songs.