The Golden-spectacled Warbler: a complex of sibling species, including a previously undescribed species

lbi3(1999) 141,545-568

The Golden=spectacled Warbler: a complex of sibling species, including a previously undescribed species

PER ALSTRoMl' & URBAN OLSSON2 Department of Systematic Zoology, Evolutionary Biology Centre, Uppsala University, Norbyvagen 18D,

752 36 Uppsala, Sweden *Department of Zoology, University of Gdteborg, Box 463,405 30 Gdteborg, Sweden

The Golden-spectacled Warbler, usually treated as a single species, Seicercus burkii, is widely distributed in mountains of southern Asia. We argue that it should be treated as five different species, one of which is described here for the first time. Two species occur in sympatry in the Himalayas. The other three species are found in eastern Asia, where all three are sympatric in China, and two of these breed sympatrically also in Vietnam. In Burma and adjacent parts of India, one of the Himalayan species is sympatric with one of the east Asian species. We describe differences in morphology, vocalizations, altitudinal distribution and habitat between these five species, as well as results from playback tests, which show that all sympatric taxa differ and appear to be reproductively isolated from each other.

'The Golden-spectacled Warbler, usually considered to he one species, Seicercus burkii, breeds throughout the Himalayas, in the mountain ranges of northeastern India and western Burma, in mountainous regions of much of central and southern China and northern Burma, and in northern Vietnam (Ah & hpley 1973, King et al. 1975, Watson et al. 1986, Cheng 1987, Rohcrts 1992; Fig. 1). In winter it descends to lower altitudes south of the breeding areas, south to central India, southern Burma, northernmost peninsular Thailand, Cambodia, Laos and central Vietnam (Ali 81 Riplcy 1973, King et al. 1975, Watson et al. 1986, Lekagul & Round 1991). Several subspecies have been deqcribed.

In the late 1980s and early 199Os, we gradually hecame aware of discontinuities in vocalizations among Golden-spectacled Warblers breeding on the mountain Emei Shan in Sichuan province, China. We were able to correlate the different vocalizations with ciiscontinuities in morphology among these birds, and wc noted that these discrete forms were segregated hititdinally. Later, we extended our research to other parts of China and other regions in Asia, and also esamined museum specimens. Our goal was to inves- tigate the number of recognizable forms in this

*(;:orresponding author. E t i1 a i I : per.alstrom@evolution . uu .se

complex, their distributions, and to them.

what names to apply

MATERIAL A I D METHODS

Field study sites

We have studied the Golden-spectacled Warbler on its breeding grounds in China, in the Qin Ling Shan (mainly on Taibai Shan) in southern Shaanxi (late June 1994, mid-June 1995), in Jiuzhaigou in southern Gansu (mid-May and early/midJune 1989, mid-June 1994), on Emei Shan in central Sichuan (mid-May 1987, early to mid-May 1989, late April to early May 1992, late May to early June 1994), a t Kunming in central Yunnan (early June 1990) and on Wuyi Shan in Fujian (late April to early May 1993); several places in central Nepal (late March 1994); in Darjeeling and Kalimpong districts in northern West Bengal, northeast Indla (late May 1997); in Chamoli district, Garhwal, northern Uttar Pradesh, northwest India (late May 1998); and on Fan Si Pan, northern Vietnam (late May 1999). See Figure 1.

Morphology

We captured a total of 37 males in China, in Sichuan (n = 32), Gansu (n = 2), Shaanxi (n = 1) and Fujian (n

0 1999 British Ornithologists' Union

Golden-spectacled Warbler 547

USA (loan); Academy of Natural Sciences of Philadelphia, Philadelphia, USA (loan); and Zoological Institute, Academia Sinica, Beijing, China. Of these, 183 males and 85 females were measured (by P.A.).

Nomenclature

We examined all of the type specimens which we were able to locate (some appear to have been lost) in order to assign names to the different populations.

Vocalizations

We have heard several hundred individuals in China and the Himalayas, and tape recorded the songs and/or calls of more than 100 of these. In Vietnam, 30-40 individuals were heard, and songs from nine and calls from three were tape recorded. The tape recordings were made either with a Sony cassette recorder WM-D6, or a Sony DAT recorder TCD-D3, TCD-D7 or TCD-D8, and a Telinga Pro parabolic reflector/microphone.

Additionally, we have obtained tape recordings of songs and calls of several individuals from China by Geoff Carey, Mike Crosby, Hannu Jannes and Ben King; from the Himalayas (northeastern and northwestern India, Nepal and Bhutan) by Tim and Carol Inskipp, Darren Irwin, Jochen Martens, Paul Holt, National Sound Archive (London) and Edward Vercruysse; from northern Vietnam by Craig Robson; from Thailand by Philip D. Round; and from the Chin hills in western Burma by Craig Robson.

Sonagrams of songs of about 70 individuals and of calls of about 40 individuals were produced, using the

strophe elements

software Canary 1.2 (Mitchell et al.. 1995). The song terminology used here is explained in Figure 3.

Playback tests

We carried out a number of playback tests in China. On Emei Shan, Sichuan province, a total of 29 males were exposed to playback of different taxa in 1992 and 1994, and in Fujian province seven males were tested in 1993. A speaker with a 20-m long cable was placed in the territory of a singing male, and the playback was started when we considered the bird to be close enough to the speaker to hear the song clearly. We played a sequence of two to eight different songs, each for two minutes (occasionally 3-8 minutes), with pauses of about one minute between the songs. The songs were generally played in the following sequence: (I) foreign song, (2) own type of song, (3) foreign song, (4) own type of song, etc. The song type of the bird that was exposed to playback was played frequently to test that the bird was still responsive, and to check that a lack of response to a foreign song was not because the bird had got tired of searching for an invisible intruder. The reaction of the territory-holder to the playback was carefully registered. In some cases we noted the response to be weak. This means that the bird approached the speaker, and appeared to show some interest in the foreign song but with little or no aggression. A strong response means that the bird repeatedly flew back and forth over the speaker, alternating between song and ‘aggression calls’ (short or drawn-out dry rattles, e.g. tm), and nervously flicking its wings and/or tail.

All of the songs used in the playback tests were tape

tremolo trill

1 .o 2.0 3.0 4.0 5.0

Time (s)

Figure 3. Sonagram terminology. Five strophes, interspersed by pauses, are shown. Each strophe consists of a variable number of discrete elements (the term note is sometimes used in the text to describe a sound that may or may not consist of more than one element). The elements are often arranged in phrases (two or more different elements given at least twice in succession; in the first strophe two elements are given three times, in the second strophe two elements are given twice, and in the fourth strophe three elements are given twice). A frill is here defined as a phrase of two to three (exceptionally four) different elements which is repeated three times or more in succession, while a tremolo consists of three or more identical elements in succession.

@ 1999 British Ornithologists’ Union, Ibis, 141, 545-568

548 I? Alstrom & U. Olsson

recorded by PA., except for burkii from Nepal, which was recorded by Jochen Martens. White-spectacled Warbler Seicercus ufiinis was included in the playback tests because of its similarity with especially tephrocephalus.

Altitudinal distribution

In most places, we estimated the altitude by using known reference points in the vicinity (e.g. on Emei Shan there are several temples of known altitude). In Shaanxi province, China, in India and in Vietnam, we measured the altitude with a Casio alt-6000 altimeter watch. The altitudinal distribution was recorded mainly along roads and trails. In most places there were only one or two roads or trails running from the bottom to the top of the mountains, and few or no side trails. Accordingly, in these places the altitudinal distribution war recorded from a very narrow transect.

We also noted altitudes given on labels of specimens. All of the values given under Results are translated from feet to the nearest 10 m.

RESULTS

Morphology

Based on morphology, the birds were separated into nine groups. The main plumage differences (Table 1, Figs 4 & 5, Plate I ) between these groups are in the head and tail patterns. Groups 1 and 4/5/6 differ from each other chiefly in the colour of the median crown- stripe and sides of the crown below the dark lateral stripes (mainly green in the former, as opposed to grey in the lattcr three). Both stand out from the others by their more contrasting crown patterns. Groups 4, 5 and 6 are basically similar to each other except for the patterns of the outer rectrices (notably R4 in group 5). Groups 2 and 3 differ from each other mainly in the crown pattern (especially the pigmentation of the dark lateral crown-stripes), and the same is true for groups 8 and 9. The former two groups (2 and 3) are distinguished from the latter two (8 and 9) chiefly by the colour of the median crown-stripe and sides of the crown below the dark lateral stripes (mainly green in the former, as opposed to grey in the latter; analogous to the diffcrences between groups 1 and 4/5/6), and in the pattern of R4 (showing much white in the former, while lacking white in the latter). The extensively white R4 in groups 2 and 3 is a feature otherwise found only in some individuals in group 5 (a few birds in group I also show white on R4, although the pattern

does not quite match that of groups 2 and 3). Group 7 is mainly characterized by the small amount of white on its outer tail-feathers (no overlap with groups 2,3 , 5, 8 and 9 and only a marginal overlap with groups 1, 4 and 6 on R5 and with groups 4, 5 and 6 on R6; note, however, that in the length of the white wedges in relation to tail length, there is a slight overlap with all except group 2 ) . For details, see Table 1 and Figures 4 and 5.

There are more or less pronounced differences in structure between these groups (Tables 2 & 3, Figs 6 & 7). The most striking differences are that groups 1 and 7 are proportionately shorter-tailed (measured as tail/wing ratio) and larger-billed (measured as bill length x bill widthhail length) than the others. Groups 4 and 5 are smaller than the others, though with relatively large bills, and group 5 has proportionately the longest tail of any group. See Table 2 and Figures 6 and 7 for details. On wing formula, group 7 differs from groups 6 and 8 particularly in its shorter P10 and indistinct emargination on P4 (Table 3).

I t should be noted that the groups that differ most prominently in plumage and structure from each other

Group 1 burkii

R6 R5

Group 2 whistler; / Group 3

nemoralis

R 6 R5 R4

R6 R 5 R6 R 5

Group 5 Group 6

tephrocephalus

R6 R 5 R6 R 5

Group 7 Group 8 soror valentini

Group 9 latouchei

Figure 4. Outer tail-feathers. Drawing: Per Alstrorn. *50% ( n = 10) S. tephrocephalus from western Burmdnortheast India (group 5) lack white on R4. #Most S. burkii (group 1) lack white on R4 (see Table 1).

@ 1999 British Ornithologists' Union, Ibis, 141, 545-568


0.90 I 1 u

5

6,

I

0)

0.85 - C .-

5 0.80 0) c 0 - - ._ @ 0.75

c ’6 0

O o u n I

a, B W N - T 4 T 5 T 6 O X S V L O k o

0.70 1 , , .--

r f 4 5 1 b B W N T 4 T 5 T 6 S V L

Group

Figure 6. Box plot of tail-length/wing-length in different taxa. B, Group 1 burkii; W, group 2 whistleri; N, group 3 nemoralis; T4, group 4 tephrocephalus from Yunnan (China) and northeast Burma; T5, group 5 tephrocephalus from northeast India and west Burma; T6, group 6 tephrocephalus from Sichuan (China); S, group 7 soroc V, group 8 valentini; L, group 9 latouchei. Same individuals as in Table 2 plus 4 unsexed burkii (group 1).

2o 3 l5 R4, the description must refer to group 1, since we

have not found any specimens of group 2 or group 3 with only two white outer rectrices and only these three groups occur in the Himalayas (cf. Fig. 8).

Cryptolopha auricapilla Swainson, 1837, was described from ‘India’. The type of auricapilla was originally in the Linnaean Society’s Museum, London,

1 0 1 , 0 W N T 4 T 5 T 6 S V L

Group

Figure 5. Box plot of length of white wedge (in mm) on R5 (a) and R6 (b) in different taxa. B, Group 1 burkii; W, group 2 whistleri; N, group 3 nemoralis; T4, group 4 tephrocephalus from Yunnan (China) and northeast Burma; T5, group 5 tephrocephalus from northeast India and west Burma; T6, group 6 tephrocephalusfrom Sichuan (China); S, group 7 soroc V, group 8 valentini; L, group 9 latouchei. Same individuals as in Table 2 plus 4 unsexed burkii (group 1).

1.50

1.30 1 s -I-

0 .- 4 1.20 - - ._ n 0 1.10

5 1.00

> m [r

.- - have sympatric breeding distributions (see Distri- bution, below, and Fig. S), while the ones which are most similar are allopatric.

0.90 Nomenclatural review

Sylvia burkii Burton, 1836, was described from ‘the Himalayas’ (type locality restricted by Ticehurst 1925 to Siklum). The type of burkii was originally deposited in the Museum of the Army Medical Department, Chatham, UK, but the fate of this collection is unknown, and it seems likely that the type has either decayed or been destroyed (Robert Prys-Jones in litt.). The type description states that the tail is brown except for white inner webs to the two outer rectrices (R5-R6). Provided that all of the tail-feathers were present, and that Burton did not overlook any white on

0.80

1 , , , , , , ~ ~ 0.70 B W N T 4 T 5 T 6 S V L

Group

Figure 7. Box plot of relative bill size (bill-length x bill-width/tail- length) in different taxa. B, Group 1 burkii; W, group 2 whistleri; N, group 3 nemoralis; T4, group 4 tephrocephalus from Yunnan (China) and northeast Burma; T5, group 5 tephrocephalus from northeast India and west Burma; T6, group 6 tephrocephalus from Sichuan (China); S, group 7 soror; V, group 8 valentini; L, group 9 latouchei. Same individuals as in Table 2 plus 4 unsexed burkii (group 1).


550 F! Alsfrom & U. Olsson

Table 1. Plumage characteristics. A, B and C denote groups that we treat as conspecific. Group 4 (fephrocephalus) breeds in Yunnan, China, northeast Burma and north Vietnam, group 5 (tephrocephalus) breeds in west Burma and northeast India, and group 6 (fephrocephalus) breeds in Sichuan, China.

A B

Group 1 burkii

Group 2 whistleri

Group 3 nemoralis

Group 4 fephrocephalus

Head

Upperparts

Underparts

Lateral crown-stripes black, usually distinct on forehead (almost to bill); median crown- stripe green, with sparse, thin, pale greyish streaks; sides of crown immediately below dark stripes green; eye-ring thinly broken at rear

On average brighter and more vellowish- green than in gps 2 & 3

On average deeper and more saturated yellow than in gps 2 & 3

Wingbar on greater Distincthather distinct: coverts 22%; faint: 26%; lack-

ing: 52% (n = 27)

R4

R5

R6

Usually no white 0-22.5 (2.6; 6.22; 30)*

White wedge on inner web 13.3-34.0 (25.8; 4.03; 30); divided by tail-length 0.29-0.76 (0.60; 0.10; 29); no white on outer web

White wedge on inner web 27.0-39.0 (31.1; 2.56; 29); divided by tail-length 0.60-0.88 (0.73; 0.06; 29); no white on outer web or only very thin pale stripe along shaft basally

Resembles gp 3, but lateral crown-stripes not quite so dark, especially anteriorly (often very faint over mid- or even rear part of eye), and median crown-stripe paler greenish, with less grey admixed

On average paler than in gp 3

As in gp 3

Distincthather distinct: 100% (n = 17)

White wedge on inner web 22.0-35.0 (30.5; 2.93; 19)3

White wedge on inner web 29.5-40.0 (34.6; 2.44; 19)5; divided by tail-length 0.64-0.85 (0.73; 0.06; 19); often very narrow whitish stripe along shaft basally

White wedge on inner web 33.5-43.0 (36.4; 2.87; 18)9; divided by tail-length 0.68-0.90 (0.77; 0.06; 18); basal part of outer web shows broad white stripe along shaft or is all white

Lateral crown-stripes greyish-black, fading out on anterior crowdforehead; median crown-stripe green, with sparse, thin, pale greyish streaks; sides of crown immediately below dark stripes green; eye-ring complete

Dull greyish-green

Yellow, with variable amount of greenish and brownish on breast(-sides) and flanks

Distincthather distinct: 85%; faint: 13%; lacking: 2% (n = 113)

White wedge on inner web 20.0-36.0 (28.3; 3.10; 87)3

White wedge on inner web 29.0-38.0 (32.9; 1.93; 88)5; divided by tail-length 0.60-0.81 (0.70; 0.05; 86); often very narrow whitish stripe along shaft basally

White wedge on inner web 30.0-40.0 (34.4; 2.08; divided by tail-length 0.58-0.84 (0.73; 0.05; 86); basal part of outer web shows broad white stripe along shaft or is all white

Lateral crown-stripes black, usually distinct on forehead (almost to bill); forehead grey with little green (sometimes more green); median crown-stripe ‘bright’ grey, with no or very little greenish admixed; grey along lower border of lateral crown- stripes; eye-ring complete

On average more yellowish-green than in gps 7, 8 & 9, similar to SP 1

On average more saturated yellow than in gps 7, 8 & 9, similar to SP 1

Distincthather distinct: 2.5%; faint: 22.5%; lacking: 75% (n = 40)

No white

White wedge on inner web 14.5-30.5 (24.8; 3.61; 23)6-7; divided by tail-length 0.33-0.70 (0.57; 0.09; 22); no white on outer web

White wedge on inner web 26.0-33.0 (29.2; 1.96; 22)lO.l’; divided by tail-length

20); no white on outer web, or only very thin pale stripe along shaft basally

0.61-0.80 (0.68; 0.05;

R, rectrix; gp, group. Numbers in brackets show: mean: sd; number (males and females combined, since there are no plumage differences between the sexes; for numbers of males and females, respectively, see Table 2). ’Photos and field notes of individuals from Sichuan indicate that there is an on average difference in head pattern from groups 4 and 5, but that is not supported by the few specimens from Sichuan we have seen (n = 7). 20nly 16.7% (three males, one female and one unsexed; n = 30) show any white, and of these, one male and the female show white on only one side of the tail; in all except one male (17 mm), the white is narrow and irregular in shape and does not reach the tip of the feather, and in none does it match whistleri and nemoralis perfectly (Fig. 4). 3Difference between whistler; and nemoralis


Table 1 continued.

€3

Group 5 Group 6 tephrocephalus tephrocephalus

-

~- - -~

As in gp 4 As in gp 4, though may show on average less contrasting crown pattern'

As in gp 4

As in gp 4

Faint: 20%; lacking: 80% (n = 5)

White wedge on inner web 0-22.5 (10.5;

White wedge on inner web 19.0-30.0 (25.3; 3.59; divided by tail-length 0.44-0.67 (0.57; 0.08; 10); no white on outer web

9.56; 1014

White wedge on inner web 23.5-33.0 (29.1; 2.79; 10)lo,ll; divided by tail-length


0.55-0.78 (0.66; 0.06;

As in gp 4

As in gp 4

Distincthather distinct: 11%; faint: 44.5%; lacking: 44.5% ( n = 9)

No white

White wedge on inner web 12.5-26.0 (20.0; 4.05; 20)6,7; divided by tail-length 0.27-0.57 (0.42; 0.10; 20); no white on outer web

White wedge on inner web 25.5-32.0 (28.2; 1.78; 2O)lo,l1; divided by tail-length


0.52-0.67 (0.59; 0.04;

Group 7 soror

~ ~~ . ~. ~

Lateral crown-stripes greyish-black, fading out on anterior crown/ forehead (often above mid-part of eye); forehead usually green; median crown-stripe dull pale grey with a little pale greenish admixed; some grey along lower border of lateral crown-stripes; eye-ring complete

Dull greyish-green

Yellow, with variable amount of greenish on breast(-sides) and flanks

Distincthather distinct: lo%, faint: 20%; lacking: 70% ( n = 20)

No white

White wedge on inner web 0.5-17.5 (5.9; 4.58; 35)7; divided by tail-length 0.01-0.46 (0.14; 0.1 1; 35); no white on outer web

White wedge on inner web 14.0-26.0 (20.2; 3.08; 35)"; divided by tail-length 0.32-0.67 (0.46; 0.08; 35); no white on outer web

C ~~ ~~

Group 8 Group 9 valentini lafouchei

_. ~ ~

Resembles gp 7, but lateral crown-stripes on average blacker and more extensive on forehead; forehead, especially, and median crown-stripe with on average less greenish

As in gps 7 & 9

As in gps 7 & 9

Distincthather distinct: 64%, faint: 27%; lacking: 9% (n = 55)

No white

White wedge on inner web 20.0-34.0 (27.9; 3.72; 52)*; divided by tail-length 0.39-0.71 (0.57; 0.07; 52); often very narrow whitish stripe along shaft basally


Resembles gp 8, but lateral crown-stripes on average paler and median crown-stripe with on average more greenish tinge (accordingly similar to gp 7, but forehead on average less green)

As in gps 7 & 8

As in gps 7 & 8

Distincthather distinct: 1 00% ( n = 1 2)

No white

White wedge on inner web 18.0-31.0 (25.8; 2.30; 14)*; divided by tail-length 0.40-0.65 (0.54; 0.04; 13); often very narrow whitish stripe along shaft basally


.. ~~~~~

significant (t-test [male]: P < 0.02). 440% (one male and three females: n = 10) show no white at all. Wifference between whistler; and nemoraiis significant (t-test [male]: P < 0.003). Wifference between 4 and 5 n s (t-test [male]: P = 0.3); between 4/5 and 6 significant (t-test [male]: P = 0.0003). 'Only 4.5% (n = 88) of tephrocephalus (all populations combined) and soror are in the overlap zone. 8No significant difference between valentinf and latouche; (t-test [male]: P = 0.44). Wifference between whfstleri and nemoralis significant (t-test [male]: P < 0.005). Wifference between 4 and 5 n s (t-test [male]: P = 0.14); between 4/5 and 6 ns (t-test [male]: P = 0.051). "Only 5.7% (n = 88) of fephrocephalus (all populations combined) and soror are in the overlap zone '*No significant difference between valentini and latouche; (t-test [male]: P = 0.1 1).

0 1999 British Ornithologists' Union, Ibis, 141, 545-568

552 f! Alstrorn & U. Olsson

Table 2. Measurements (field and museums combined). Group 1 (burkil) from throughout Himalayas and lowlands of India; group 2 (whistler/) from northwest India; group 3 (nemoralis) from Nepal, northeast India (most), and west Burma; group 4 (tephrocephalus) from Yunnan province, China and northeast Burma (the birds caught in Vietnam perfectly match this group on measurements, although no breeding specimens examined); group 5 (tephrocephalus) from west Burma and northeast India; group 6 (tephrocephalus) from Sichuan province, China; group 7 (soror) from Sichuan, Yunnan and Fujian provinces, China, and Burma, Thailand, Cambodia and Vietnam; group 8 (valentini) from Yunnan, Sichuan, Gansu and Shaanxi provinces, China; and group 9 (latouche4 from Fujian province, China. Data in brackets are means, standard deviations and numbers. All measurements in mm.

Group 1 burkii

Group 2 whistleri

Group 3 nemoralis

Group 4 tephrocephalus

Males Wing

Tail

Bill-length

Bill-width

Tarsus

Hind-claw

TaiVwing

Bill-length x bill-widthhail

Females Wing

Tail

Bill-length

Bill-width

Tarsus

Hind-claw

Tail/wing

Bill-length x bill-widthhail

53.0-59.5 (56.6; 1.77; 20)

40.5-47.0 (43.2; 1.72; 21)

(13.6; 0.59; 21)

3.3-4.0 (3.6; 0.18; 16)

(17.7; 0.48; 16)

(5.6; 0.36; 19)

(0.76; 0.02; 20)

(1.12; 0.10; 16)

12.3-1 5.1

17.0-18.5

5.0-6.4

0.72-0.79

1.01-1.40

51.5-55.0 (53.5; 1.26; 6)

39.5-42.5 (40.9; 1.29; 5)

(13.5; 0.54; 6)

3.5-3.9 (3.7; 0.17; 4)

(17.6; 0.39; 4)

(5.6; 0.06; 3)

(0.76; 0.02; 5)

(1.20; 0.07; 3)

12.9-14.0

17.1-18.0

5.6-5.7

0.72-0.78

1.14-1.28

55.0-62.0 (58.0; 1.64; 15)’

44.5-51.5 (47.9; 1.79; 15)4

12.3-13.7 (12.9; 0.42; 15)7

(3.1; 0.13; 9)’O

18.1-19.0 (18.4; 0.34; 9)i3

(5.5; 0.34; 10)l6

(0.82; 0.02; 15)i9

(0.84; 0.06; 9)22

2.9-3.3

5.1-6.2

0.80-0.86

0.76-0.92

53.5-56 .O (54.9; 1.03; 4)

44.0-46.5 (44.9; 1 .la; 4)

12.4-1 3.3 (12.9; 0.41; 4)

(3.2; 0.22; 4)

(1 8.3; 0.72; 3)

(5.5; 0.34; 4)

0.80-0.84 (0.82; 0.02; 4)

0.85-0.99 (0.92; 0.07; 4)

3.0-3.5

17.8-19.1

5.1-5.8

55.0-62.0 (58.3; 1.85; 58)’

(48.0; 2.19; 5 ~ 7 ) ~

(1 3.2; 0.41 ; 5 ~ 7 ) ~

(3.3; 0.14; 42)’O

(18.7; 0.57; 35)j3

(5.5; 0.36; 45)i6

0.75-0.86 (0.82; 0.02; 57)19

(0.90; 0.06; 41)22

43.5-53.0

12.3-14.4

3.0-3.6

17.1-20.0

4.9-6.2

0.77-1.08

53.0-57.5 (54.6; 1.20; 29)

42.0-47.5 (44.6; 1.50; 30)

(12.8; 0.42; 29)

(3.3; 0.15; 21)

(18.6; 0.67; 21)

5.0-6.1 (5.6; 0.32; 20)

0.76-0.87 (0.82; 0.03; 29)

(0.96; 0.07; 21)

12.1 -1 3.8

3.0-3.5

17.2-19.8

0.86-1.08

52.5-57.0 (54.1; 1.00; 17)‘

41.0-47.0 (43.8; 1.75; 16)5

(13.2; 0.43; 16)8

3.3-3.7

12.2-1 4.0

(3.5; 0.1 1; 12)”

16.7-1 7.4 (17.1; 0.31; 5)i4

4.7-5.9 (5.2; 0.34; 8)’’

0.78484 (0.81; 0.02; 16)20

0.95-1.20 (1.06; 0.08; i1y3

49.5-53.0 (50.6; 1.22; 7)

39.5-44.0 (41.3; 2.14; 6)

(13.1; 0.55; 6)

(3.6; 0.26; 5)

16.3-1 7.7 (17.0; 0.57; 4)

(5.2; 0.22; 5)

0.78-0.89 (0.81 ; 0.04; 6)

12.5-14.0

3.2-3.8

4.8-5.3

1.01-1.26 (1.10; 1.12; 4)

’No significant difference between whistleri and nemoralis (t-test [male]: P = 0.58). 2No significant difference between 4 and 5 (Mann-Whitney Utesl [male]: P = 0.42), but significant difference between 4/5 and 6 (Mann-Whitney U-test [male]: P c 0.0001). 3No significant difference between valentiniand latouchei (Mann-Whitney U-test [male]: P = 0.1 1). 4No significant difference between whistleri and nemoralis (t-test [male]: P = 0.76). 5No significant difference between 4 and 5 (t-test [male]: P = 0.07), but significant difference between 4/5 and 6 (t-test [male]: P c 0.0001). %Significant difference between valentini and latouchei (t-test [male]: P = 0.03). ’Significant difference between whistleri and nemoralis (t-test [male]: P = 0.01). 8No significant difference between 4 and 5 (t-test [male]: P = 0.46), and no significant difference between 4/5 and 6 (t-test [male]: P = 0.61). QSignificant difference between valentini and latouchei (t-test [male]: P = 0.01). losignificant difference between whistleri and nemoralis (t-test [male]: P = 0.0007). llNo significant difference between 4 and 5 (t-test [male]: P = 0.10), but significant difference between 4/51 and 6 (t-test [male]: P = 0.03). ”No significant difference between valentini and latouchei (t-test [male]: P = 0.98). 13No significant difference between whistleri and nemoralis (1-test [male]: P = 0.1 9). 14No significant difference



Table 2. continued.



Group 7 soror

Group 8 valentini

Males 51.5-56.5

(53.5; 1.98; 6)‘

(45.8; 1.70; 6)5

(1 3.2; 0.46; 6)8

(3.4; 0.18; 6)”

(17.2; 0.37; 6)14

(5.4; 0.25; 6)”

(0.86; 0.03; 6)’O

(0.98; 0.05; 6)23

42.5-47.0

12.4-1 3.7

3.1 -3.6

16.8-1 7.8

5.0-5.7

0.82-0.90

0.90-1.06

Females

48 .O-49.5 (48.8; 0.64; 4)

(42.1 ; 2.02; 4)

(13.1; 0.32; 3)

3.3-3.9 (3.6; 0.25; 4)

(16.9; 0.20; 3)

(5.3; 0.31; 3)

(0.86; 0.04; 4)

40.0-44.5

12.9-13.5

16.7-1 7.1

5.0-5.6

0.82-0.91

0.96-1.16 (1.09; 0.11; 3)

53.5-61.5 (58.3; 2.44; 1

(48.0; 2.38;

(13.3; 0.38; 17)8

(3.3; 0.21; 17)”

(18.0; 0.50; 14)14

(5.5; 0.42; 15)”

(0.82; 0.03; 1

(0.92; 0.10; 17)23

42.5-51 .O

12.6-14.2

3 .O-3.7

17.0-19.0

4.7-6.6

0.73-0.86

0.76-1.18

51.5-54.5 (53.5; 1.73; 3)

43.0-45.5 (44.5; 1.32; 3)

(12.8; 0.07; 2)

(3.4; 0.35; 2)

(16.9; 0.14; 2)

(5.3; 0.29; 3)

(0.83; 0.005; 3)

0.89-1.1 0 (1.00; 0.15; 2)

12.7-1 2.8

3.2-3.7

16.8-17.0

5.0-5.5

0.83-0.84

56 .O-62 .O (58.4; 1.42; 27)

41.047.5 (44.6; 1.76; 27)

(14.1; 0.46; 25)

3.3-4.0 (3.8; 0.19; 23)

(17.4; 0.35; 24)

(5.4; 0.32; 25)

(0.76; 0.02; 27)

(1.17; 0.09; 22)

13.3-15.4

16.7-18.0

4.8-5.9

0.7 1 -0.80

0.99-1.30

53.0-57.5 (54.6; 1.59; 8)

(40.6; 1.48; 8)

(13.6; 0.28; 8)

(3.5; 0.17; 8)

(17.3; 0.63; 7)

(5.7; 0.29; 7)

(0.74; 0.02; 8)

0.99-1.34

38.5-42.5

13.2-14.1

3.2-3.7

16.5-18.0

5.1-6.0

0.71-0.78

(1.18; 0.11; 8)

57.5-64.5 (61 .O; 1.75; 42)3

(49.9; 1.95; 42)6

(13.4; 0.55; 41)9

(3.4; 0.14; 38)”

(18.3; 0.59; 35)15

(5.8; 0.31; 40)18

(0.82; 0.03; 40)”

(0.92; 0.06; 36)24

46 .O-54.5

12.4-14.6

3.1-3.7

16.8-1 9.4

5.1-6.6

0.76-0.86

0.78-1.06

53.0-60.5 (57.0; 2.28; 10)

42.5-47.5 (45.0; 1.76; 10)

(13.3; 0.34; 10)

3.3-3.9 (3.6; 0.18; 10)

(18.1; 0.60; 6)

(5.8; 0.20; 10)

(0.79; 0.03; 10)

(1.04; 0.10; 10)

12.6-13.7

17.3-1 8.7

5.5-6.1

0.72-0.83

0.80-1.1 5

Group 9 latouchei

59.0-61.5 (60.0; 0.85;

(48.4; 0.79;

(13.0; 0.44; 9)9

(3.4; 0.15; 8)12

(18.1: 0.45; 8)15

(5.2; 0.25; 8)18

(0.80; 0.03; 10)”

(0.92; 0.04;

45.5-51 .O

12.2-13.6

3.2-3.7

17.6-1 8.8

5 .O-5.7

0.75-0.85

0.86-0.96

52 .O-57.5 (55.5; 2.42; 4)

(45.3; 1.16; 3)

(12.6; 0.32; 4)

(3.3; 0.38; 3)

(17.4; 1.12; 3)

(5.4; 0.35; 2)

(0.80; 0.02; 3)

0.82-0.97

44.0-46.0

12.3-1 3.0

2.9-3.6

16.1-18.2

5.2-5.7

0.78-0.82

(0.89; 0.10; 2)

between 4 and 5 (Mann-Whitney U-test [male]: P = 0.53), but significant difference between 4/5 and 6 (t-test [male]: P = 0.0001). 15No significant difference between valenfini and lafouchei (f-test [male]: P = 0.52). ’“0 significant difference between whistleri and nernoralis (t-test [male]: P = 0.92). 17No significant difference between 4 and 5 (f-test [male]: P = 0.18), and no significant difference between 4/5 and 6 (t-test [male]: P = 0.13). ’*Significant difference between valentini and lafouchei (t-test [male]: P < 0.0001). 19No significant difference between whistleri and nernoralis (Mann-Whitney U-test [male]: P = 0.87). 20Significant difference between 4 and 5 (Mann-Whitney U-test [male]: P = 0.005), but no significant difference between 4/5 and 6 (Mann-Whitney U-test [male]: P = 0.60). 2 1 N ~ significant difference between valentini and lafouchei (Mann-Whitney U-test [male]: P = 0.16). 22Significant difference between whisfleri and nernoralis (Mann-Whitney U-test [male]: P = 0.02). Z 3 N ~ significant difference between 4 and 5 (Mann-Whitney U-test [male]: P = 0.06), but significant difference between 4/5 and 6 (Mann-Whitney U-test [male]: P = 0.001). *4No significant difference between valentini and latouchei (Mann-Whitney U-test [male]: P = 0.92).



Table 3. Wing formula (field, central China; emargination on P4 also museums). P10 is measured in relation to the tips of the primary coverts, the others in relation to the wing tip. Numbers in brackets: mean; sd. We found no differences in wing formula between burkii and nemoralis; wing formulae of whistleri and latouchei not checked.


(n = 9)

Group 7 soror

(n = 10)

Group 8 valen tini (n = 8)

P4 P5 P6 P7 P8 P9 P10 Emargination on P4

3.0-4.5 (3.83; 0.50) 0.5-1.0 (0.56; 0.17)

0-0 (0; 0 ) 0-1 .O (0.44; 0.39)

2.0-4.0 (2.94; 0.73) 9.0-13.0 (11.28; 1.18) 10.5-13.0 (1 1.67; 0.87) lacking: 3%; faint: 43%; distinct: 53% ( n = 30)

4.0-7.0 (5.1 1; 0.86) 1 .O-2.0 (1.55; 0.50) 0-0.5 (0.05; 0.16) 0-0.5 (0.15; 0.24)

2.0-2.5 (2.15; 0.24) 9.0-10.5 (9.70; 0.54) 7.0-10.0 (8.45; 0.86)

lacking: 70%; faint: 30% ( n = 20)

2.0-4.5 (3.44; 0.82) 0.5-1 .O (0.69; 0.26)

0-0 (0; 0) 0-0.5 (0.44; 0.18)

2.0-4.0 (2.81 ; 0.65) 8.0-13.0 (11.12; 1.58) 9.0-12.5 (11.06; 1.12)

lacking: 12%; faint: 45%; distinct: 43% ( n = 42)

UK, but it was probably sold in 1863, and it is not known where it went (Gina Douglas in litt.). The original description is inconclusive with regard to which species it refers to. Accordingly, we consider auricapilla a nomen dubium.

The original description of Acanthiza arrogans Sundevall, 1838, was based on one specimen from Calcutta, West Bengal, India (22"30", 88"20'E), collected on 9 February 1828. This specimen was originally housed in the private collection of Baron Gyllenkrok, Sweden, but appears to have been lost

a 40"

30"

20"

10"

(another type specimen from the same collection, Phylloscopus t. trochiloides, is in the Swedish Museum of Natural History). The original description states that the two outer rectrices on each side are white on the inner webs. Provided that all of the rectrices were present and were examined, arrogans should be considered a junior synonym of burkii.

Muscicapa bilineata Lesson, 1839, was described from 'Indiae orientales'. The original description is inconclusive, and we have not been able to trace the type specimen (originally in the private collection of

70" 80" 90" 100" 110" 120" 130"

Figure 8. Breeding distributions based on our own observations (field and museums). (a) Places where sympatry between two or more taxa has been observed. w, whistleri; n, nemoralis; t, tephrocephalus; s, soror; v, valentini; I, latouchei. (b-f) Detailed maps of the different taxa. continued



M. Charles Brelay, Bordeaux, France). Accordingly, we consider bilineata a nomen dubium.

The type locality of Culicipeta tephrocephalus Anderson, 1871, was given as Bhamo, northeastern Burma (24"15'N, 97'15'E). The type, which was originally in the Indian Museum of Calcutta, India is presumed to have been lost during floods during World War I1 (J.B. Alfred, communicated through S. Sathykumar and Pratap Singh). The description is extremely brief. However, groups 1, 2 and 3 can be

Figure 8 continued. (b) burkii; (c) 0 whistleri, nemoralis; (d) tetrocephalus; (e) soroc (f) + valentini, ? refers to the fact that we have not seen any specimens, and therefore cannot rule out latouchei with certainty; A latouchei.

eliminated by the fact that it has 'the head and nape cinereous, with a black band along their lateral mar- gin'. The measurements, and the statement that it is slightly brighter green than burkii (meaning Himalayan birds), lead us to believe that tephrocephalus refers to group 4, 5 or 6 rather than any of groups 7, 8 or 9. Group 4 breeds at Bhamo (see below), while neither group 5 nor group 6 are known to have occurred in that part of Burma, so we conclude that tephrocephalus represents group 4.

C yptolopha birmanica Berezowski & Bianchi, 189 1, was described based on one specimen collected in Lower Pegu, Burma (c. 18"N, 96"E) on 5 November 1879. The holotype is in the Zoological Museum, Russian Academy of Sciences, St Peterburg, Russia (No. 106160). I t has been examined on our behalf by Roald L. Potapov, who kindly sent us measurements



and colour photographs of it. I t matches group 4. The two syntypes of Cryptolopha burkii distincta La

Touche, 1923, were collected at Mengtz (Mengzi), Yunnan, China (23"20'N, 103'21'E) on 25 March and 2 1 April 192 1, respectively. We have examined these (Nos 128843 and 128844 in the Museum of Comparative Zoology, Boston, USA), and they match group 4.

The holotype of Cryptolopha burkii valentini Hartert, 1907, was collected on Taibai Shan, Qin Ling Shan, Shaanxi province, China (peak: 33"58'N, 107'42'E) on 2 June 1905. We have examined this (No. 450459 in the American Museum of Natural History, New York, USA) and it fits group 8.

The holotype of Seicercus burkii latouchei Bangs, 1929, was taken a t Kuatun, Fujian province, China (c. 27"4'N, 117'5'E) on 11 May 1898. We have examined this (No. 128830 in the Museum of Comparative Zoology, Boston, USA) and it matches Group 9.

The holotype of Seicercus burkii whistleri Ticehurst, 1925, was collected at Dharmsala, Himachal Pradesh, India (32"14'N, 76'24'E) on 22 October 1925. We have examined this (No. 1926.1.4.1; in The Natural History Museum, Tring, UK) and it matches Group 2.

The holotype of Seicercus burkii nemoralis Koelz, 1954, was collected a t Sangau, Lushai hills, Mizoram, India (22"44'N,93"04'E) on 23 February 1953. We have examined this (No. 147865 in the Museum of Zoology, University of Michigan, Ann Arbor, USA) and it matches group 3.

In addition to these names, we have seen specimens of group 3 and group 5 labelled Seicercus burkii pectoralis (in the Museum of Comparative Zoology, Boston, USA, and in the Academy of Natural Sciences of Philadelphia, Philadelphia, USA), but we have not been able to find any published reference to this name.

Taxonomic and nomenclatural conclusions

Since five of the original types may be irretrievably lost, and since the original descriptions of these taxa are very brief, the nomenclatural situation is compli- cated. We suggest the following arrangement: group 1 should be named burkii (arrogans should be considered a junior synonym); group 2 whistleri; group 3 nemoralis; group 4 tephrocephalus (birmanica and distinctus are considered junior synonyms); group 8 valentini; and group 9 latouchei. The names auricapilla and bilineata are considered nomina dubia. Groups 5 and 6 are unnamed, but our sample of these is too small to judge whether or not they deserve to be described (see Discussion) and we provisionally treat

them as tephrocephalus. Birds belonging to group 7 are undescribed and we therefore propose that they be named:

Seicercus soror, sp. nov.

Holotype The Natural History Museum, Tring, UK, No. 1914.5.6.292, male, Hopiachen, Sichuan province, China, 16 September 1902, collected by F.W. Styan.

Diagnosis of taxon Very similar to tephrocephalus and, especially, valentini and latouchei, but differs from all three by considerably shorter white wedges on inner webs of two outer pairs of rectrices (usually <14 mm on T5 and <24 mm on T6; very slight overlap with tephrocephalus; cf Table 1, Figs 4 & 5). Further distinguished from tephrocephalus by less contrasting crown pattern (lateral crown-stripes shorter and more diffuse anteriorly, and less deep blackish throughout; median crown-stripe less pure grey; forehead usually greenish).

Description of holotype Plumage. Forehead dull green. Distinct median crown-stripe

rather pale dull grey with a little green streaking mixed in. Prominent lateral crown-stripes brownish-black, reaching from above anterior part of eye (where slightly diffuse) to nape. Sides of crown below lateral crown-stripes dull green with diffuse thin rather pale dull grey (with a little green mixed in) stripe immediately below dark stripes. Lores mottled yellowish-white and dull green. Ear-coverts dull green with a few very thin pale streaks, and slightly diffuse lower border. Complete eye-ring pale yellow (very slightly dusky anteriorly above eye). Neck-sides dull green. Rictal bristles blackish, with pale very bases. Throat, breast, flanks, belly and undertail-coverts yellow, with pale dull green wash on breast-sides and flanks (especially). Mantle, scapulars, back, rump and uppertail-coverts dull green. Lesser coverts dull green. Median coverts brown-grey with diffuse dull green fringes (slight contrast; feathers appear essentially green). Greater coverts and carpal covert brown-grey with diffuse dull green outer edges and tips (slight contrast; feathers appear essentially green when in order). Alula brown-grey with greenish outer webs (least bright on largest feather). Primary coverts medium dark brown-grey with diffuse dull green outer edges (indistinctly reaching tips). Remiges medium dark brown-grey with slightly diffuse greenish edges to outer webs; greenish edges do not reach beyond emarginations on emarginated primaries and barely to tips of other remiges (pale outer edges slightly paler and more yellowish-tinged than upperparts, especially on primaries); inner webs show narrow (<1 mm) greyish-white edges which do not reach beyond notches on notched primaries and barely reach tips of other remiges. Tertials brown-grey with faint greenish tinge and diffuse dull green edges to outer webs (tertials do not contrast markedly with



upperparts even though they are less green). Undenving- coverts yellow, greater under primary coverts with brown-grey centres. Rectrices brown-grey with diffuse dull green edges to outer webs which are widest basally and barely reach tips; virtually absent on R6. R5 shows c. 10-mm-long off-white wedge on inner web on right side (washed with pale brown-grey and diffuse proximally; also narrow greyish-white edge to inner web (R5 lacking on left side) (cf Fig. 4). Inner web of R6 shows c. 22-mm-long off-white wedge (distal c. 13 mm of feather all off-white); proximal part of wedge slightly diffuse); also narrow off- white edge to inner web (cf Fig. 4). Outer web of R6 shows indistinct narrow whitish stripe along shaft basally (cf Fig. 4).

Bare parts. Upper mandible dark, paler on very tip and narrowly along cutting edge. Lower mandible pale. Tarsi, toes and claws pale. Structure. Bill relatively long and broad-based, with rather

straight edges seen from above, and distinct ridge along culmen (cf Fig. 2); relatively deep. Nostrils elongate, slit-like, partly covered by feathers basally (cf. Fig. 2). Rictal bristles prominent (longest bristle 6.5 mm from tip to whitish base). Tarsus and toes rather thin.

Measurements. Wing 60.0 mm; tail 46.0 mm; bill-length 14.0 mm; bill-width 3.8 mm; bill-depth (where widest; slightly distal to tips of nostrils) 2.8 mm; tarsus 17.5 mm; hind-claw 5.5 mm.

Wing formuh.Wing-point P7-P6; P8 -2.0 mm; P9 -10.0 mm; P10 -29.5 mm (9.0 mm > tips of primary coverts); P5 -2.5 mm; P4 -6.0 mm; P3 -8.0 mm; P2 -9.5 mm; P1 -10.5 mm. P5-P8 emarginated on outer webs, P9-c. P7 notched on inner webs.

Paratype The Natural History Museum, Tring, UK, No. 1923.3.10.9, female, Mengtz (Mengzi; 23"20'N, 103"21'E), Yunnan province, China, 15 September 1920, collected by J.D. La Touche.

Paratypic variation The main differences from the holotype are: marginally blacker lateral crown-stripes; marginally less green on median crown-stripe; green anteriormost crown; white wedge on R5 3.5 mm on right side and 7.5 m m on left side; and white wedge on R6 20 m m on right side and 19 mm on left side.

Etymology The name soror means sister, alluding to its close similarity with the other taxa in this complex.

Neotype of Sylvia burkii Burton, 1836

Since the type of this taxon probably no longer exists, we designate the following specimen a neotype: The Natural History Museum, Tring, UK, No. 97.12.10. 2801, unsexed (could be either sex according to

measurements), Siklum, November 1875, collected by J. Biddulph.

Diagnosis Median crown-stripe predominantly green, with very little grey admixed (as in whisthri and nemoralis, but unlike in soror, tephrocephalus, valentini and latouchei). Most important difference from whistleri and nemoralis is lack of white on R4 or, uncommonly, a relatively short, narrow, irregular white pattern (whist& and nemoralis invariably show a prominent white wedge on R4; see Table 1, Fig. 4).

Description of neotype Similar to holotype of Seicerms soror with the following exceptions.

Plumage. Median crown-stripe bright green with very little rather pale grey streaking mixed in. Lateral crown-stripes black, reaching onto forehead (in front of eyes); rather clear- cut anteriorly. Sides of crown below lateral crown-stripes bright green, lacking grey. Lores marginally darker than in soror, mottled yellowish and dull green. Ear-coverts duskier along upper border and below eye-ring. Eye-ring deeper yellow than in soror, very thinly broken at rear. Throat, breast, flanks, belly and undertail-coverts deeper yellow, with pale dull green wash on breast-sides and flanks (especially) and a pale brown tinge which is most obvious on the central part of the breast. Mantle, scapulars, back, rump and uppertail-coverts slightly brighter and more yellowish-green than in soror. Base colour of remiges and rectrices slightly darker brown-grey than in soror. White wedge on R5 much longer (26.5 and 26.0 mm on right and left side, respectively), more clear-cut and more contrasting (due to dark portion of feather being darker than in soror; cf. Fig. 4). White wedge on inner web of R6 longer (29.0 mm on right side; lacking on left), broader, more clear-cut and more contrasting (due to dark portion of feather being darker than in soror). Whitish stripe along shaft on outer web of R6 even narrower and less distinct than in soror.

Structure. Rictal bristles shorter and thinner (5 mm from tip to whitish base).

Measurements. Wing 54.0 mm; tail 43.0 mm; bill-length 13.8 mm; bill-width 3.7 mm; bill-depth (where widest; slightly distal to tips of nostrils) 2.9 mm; tarsus 18.1 mm; hind-claw 5.7 mm.

Wingformula. Wing-point P6-P5; P7 -0.5 mm; P8 -2.0 mm; P9 -8.5 mm; P10 -29.5 mm (12.0 mm > tips of primary coverts); P4 -1.5 mm; P3 -3.5 mm; P2 -5.5 mm; PI -6.5 mm.

Neotype of Culicipeta tephrocephalus Anderson, 1871

Since the type of this taxon probably no longer exists, we designate the following specimen a neotype: American Museum of Natural History, No. 450451, female, Sinlum, Bhamo (24"15'N, 97'1 5'E), northeast


558 FI Alstrom & U. Olsson

Burma, 6000 feet, 12 May 1908, collected by H.H. Harington. It is written on the label that it had a nest with eggs.

Diagnosis of taxon More contrasting crown pattern than in whistleri, nemoralis, soror, valentini and htouchei: pure grey or very slightly greenish-streaked median crown-stripe, black lateral crown-stripes which are distinct and relatively clear-cut on forehead (reach close to base of bill; occasionally slightly less distinct on forehead), and generally greyer forehead; lateral crown-stripes similar to burkii, but colour of median crown-stripe different (mainly green in burkii). Prominent white wedge on R5 (unlike soror; cf. Table 1, Figs 4 & 5 ) , but none on R4.

Description of neotype Similar to holotype of Seicercus soror with the following exceptions.

Plumage. Forehead less extensively green (only very small area just above bill greenish). Median crown-stripe more pure grey with hardly any green mixed in. Lateral crown-stripes black, reaching from close to base of bill to nape (longer and more clear-cut on forehead, and blacker throughout length than in soror). Grey stripe on sides of crown immediately below lateral crown-stripes broader and more pure grey. Ear- coverts marginally brighter yellowish-green. Throat, breast, flanks, belly and undertail-coverts deeper yellow. Mantle, scapulars, back, rump and uppertail-coverts brighter yellowish-green. Remiges very slightly darker brown-grey than in soror. Rectrices slightly darker brown-grey than in soror. White wedge on R5 much longer (30.0 and 28.0 mm on right and left side, respectively), broader, more clear-cut and more contrasting (due to dark portion of feather being darker than in soror); also outer web shows small white tip (cf. Fig. 4). Inner web of R6 shows >3 1 .O mm long off-white wedge (con- tinues narrowly to base of feather; distal c. 22 mm of feather all off-white) - i.e. considerably longer, broader, more clear- cut and more contrasting than in soror (due to dark portion of feather being darker than in soror); small off-white tip to outer web.

Structure. Rictal bristles even longer than in soror, especially proportionately (longest bristle 7.5 mm from tip to whitish base).

Measurements. Wing 51 .O mm; tail 40.5 mm; bill-length c. 13.0 mm (very tip of bill broken); bill-width 3.4 mm; bill-depth (where widest; slightly distal to tips of nostrils) 2.7 mm; tarsus 17.4 mm; hind-claw 5.3 mm.

Wing fornula. Wing-point P7-P5; P8 --c. 2 mm; P9 -6.5 mm; P10 -20.5 mm (12.0 mm > tips of primary coverts); P4 -1.5 mm; P3 -3 mm; P2 -4 mm; P1-5.5 mm. P4-P8 emarginated on outer webs.

Distributions

The breeding ranges of the different taxa are shown in Figure 8. As is evident from Figure 8a, some of the taxa

are sympatric, while the distributions of the others are parapatric or disjunct. In winter, burkii was found within its breeding range and in the lowlands of Bangladesh and India (south to the Eastern Ghats in northeastern Andhra Pradesh, Inda); whistleri and nemoralis within their respective breeding ranges; tephrocephalus in Burma (south to Tenasserim), Thailand (except peninsula), Laos and Vietnam (south to Cochinchina); soror in southern Burma (including Tenasserim), Thailand (except peninsula), Cambodia and southern Vietnam; and valentini in southern China, northwestern Thailand, Laos and Vietnam. We have not seen any specimens of latouchei from outside its breeding range. It should be pointed out that we have concentrated on the breeding distributions, and not studied the winter ranges in detail.

Vocalizations

Songs The songs of whistleri (Fig. 9b) and nemoralis (Figs 9c & 1 Od, Table 4) are simple, consisting of short strophes of fewer than ten short, soft, whistled elements. There is almost invariably a short introductory note, which is rather consistent both intra- and inter-individually. Usually, the rest of the elements are arranged in a phrase that is repeated once, either in full or in part. The strophes are interspersed by pauses of several seconds. Individual strophes are often repeated, either a few times in succession or after some time. Although our sample of songs of whistleri is small (two tape recordings and some ten further individuals heard in the field), we consider it to be indistinguishable from the song of nemoralis. The songs of these two are highly similar to those of valentini (Figs 9g & lOc, Table 4) and htouchei (Fig. 9h). Since we have no tape recordings of valentini from the area in China adjacent to the Himalayas, we cannot say whether the minor differences between valentini and whistlednemoralis shown in the sonagrams and Table 4 are clear-cut or whether the songs grade into each other.

The song of burkii (Fig. 9a, Table 4) is easily distinguishable from whistlednemoralis, both audibly and on sonagrams, by the presence of tremolos and trills (in about 50% of unique strophes; Table 4). There are several additional differences contributing to the dstinctness of the song of burkii (Table 4). It is, however, very similar to the song of tephrocephalus (Figs 9d, 9e & lob, Table 4). Because of insufficient sampling, we cannot say whether or not the differences between burkii and tephrocephalus shown in the sonagrams and Table 4 are valid also where their ranges abut.


Freq

uen

cy (kH

z) F

requ

ency (k

Hz)

Freq

uen

cy (kH

z) m

0

2

0

0 1

(D

2

8

2

x -0 2

0

v 0 0 0

2 g a, 0

-4 0

2

gg

.a, 0

4

mP

0

0

‘‘i

0

2

0

Oi ‘r 0

m 0

z 8 2

4

2

0

A

z

-.- -5 -- i-

lo 0

2

8 =! 3

(D

2

2

0

. * s c 5 v)

v

A

z I

-- - -- 0

;I =”

-.

2

P 0

Freq

uen

cy (kHz)

Freq

uen

cy (kHz)

2

gg

2

gg

‘2

2

0

v 0 P 0

0, 0

‘r 0

0 :I

D

>-

=.=-

P 0.

2.

P, 0.

c

5 \

ii e

--- c t - c

-2

560 F! Alstrom & U. Olsson

1.0 2.0 3.0 4.0 5.0 6.0 7.0 - Q 10.0 p

I 14.0 15.0

Time (s)

Figure 9 continued.

The song of tephrocephalus (Figs 9d, 9e & lob) is easily told from those of soror, vulentini and latouchei, both by ear and on spectrograms, by the presence of tremolos and trills (in about 40% of unique strophes in birds from Sichuan, China, and in about 90% of strophes in birds from Vietnam; Table 4). It also differs in several other respects (Table 4).

There is actually rather pronounced geographical variation in tephrocephalus. Birds in Vietnam differ from those in Sichuan, China, in more frequently giving an introductory note and in having considerably more trills and a larger proportion of phrases (Table 4). We only have tape recordings of one tephrocephalus from Yunnan, China, and one from western Burma, but both these match Vietnamese birds very well.

The songs of valentini (Figs 9g & lOc, Table 4) and latouchei (Fig. 9h) are indistinguishable (although our sample of latouchei is limited to tape recordings of two individuals). The song of soror (Fig. 9f) resembles those of vulentini and latouchei, although with experience it is distinguishable both audibly and on sonagrams. The most easily heard differences are that the song of soror is higher-pitched than those of valentini and latouchei. This applies also to the introductory note, which is a soft chip in soror, and a soft chu in valentini and latouchei. On spectrograms (Fig. 10a & lOc, respectively), the appearance of the introductory note is markedly different in soror, on the one hand, and valentini and latouchei, on the other hand, and the strophes cover a wider frequency range in the former than in the latter two. For further differences see Table 4.

Except in tephrocephalus (see above), we have not

h g 10.0 1 Y

1 .o 2.0 3.0 4.0 5.0 6.0 7.0

L 10.0

8.0 9.0 10.0 11.0

14.0 15.0

Time (s)

12.0 13.0 14.0

noted any geographical variation in the different taxa (although the sample is too small from some areas to evaluate this properly; see Table 4).

To conclude, all sympatric taxa have different songs, while the songs of some of the allopatric ones are very similar or inlstinguishable. With the allopatric taxa, there is a positive correlation between morphological similarity and similarity in song.

Calls Throughout the Himalayas, burlzii calls with a soft, 'whipping' huit (somewhat reminiscent of a high-

a

4 b

t C

0 *

d

'c,

0 0.1 0.2 0.3 0.4

Time (s)

Figure 10. Sonagrams of introductory element. (a) soroc (b) tephrocephalus (Sichuan, China); (c) valenfini; (d) nernoralis. Same individuals as in Figure 9, except for tephrocephalus, which, however, was taped at the same location as the one in Figure 9.



4 6:

2 -

pitched Common Chaffinch Fringilla coelebs), less commonly doubled (Fig. 1 le) . In West Bengal nemoralis consistently called with a softer, lower- pitched double tiu(-)du (Fig. llf), while all of the whistleri we heard in Uttar Pradesh called with single tiu or, differently transcribed, chip (Fig. 1 lg). The latter call has also been heard and tape-recorded in central Nepal, both in the breeding (Jochen Martens) and in the non-breeding season (P.A.), and in northern Uttar Pradesh in winter (Edward Vercruysse).

The call of soror is a short, rather high-pitched, thin tsi(-)dit or, differently transcribed, tsi(-)rit, occasionally a single tsrit (Fig. 1 la). In Sichuan, China, tephrocephalus calls with a short, comparatively faint chup and a double chu(-)du or, differently transcribed, tu(-)ru (Fig. I lb ) . However, we have only heard this call a few times (it apparently does not call so much as the others), and only tape-recorded two individuals, so we cannot say how consistent it is. The calls of tephrocephalus in Vietnam are closely similar to those in Sichuan, both audibly and on sonagrams, although the double (most commonly heard) note sounds thinner and less clearly disyllabic (due to shorter interval between the two elements than in Sichuan), and could be transcribed as trrup or turup. The call of vulentini in central China and northern Vietnam is a short, soft, deflected tiu or, differently transcribed, heu, which is occasionally doubled, tiu-tiu (or heu-heu) (Fig. 1 lc). The only latouchei we tape-recorded in Fujian, south- eastern China, uttered a higher-pitched, sharper tiu, frequently doubled; the sonagram in Figure 1 Id is markedly different from valentini (note that the first part of the element resembles the call of nemorulis, while the second part matches valentini).

In summary, all taxa have distinct calls, although the differences between whistleri and nemoralis are not clear-cut, and the sample of latouchei is too small to be sure that its call is consistently different from those of the others.

6 - I kl 3

Playback tests

On Emei Shan, Sichuan, China, we carried out playback tests on 29 males: 13 soror, ten tephrocephalus and six valentini (Table 5a). All except two of these showed no interest whatsoever in the song of any sympatric taxon (although they did respond very strongly to song of their own taxon). The two birds, both of which were tephrocephalus, that reacted towards foreign song did so very weakly and with no apparent aggression; we considered the response to be an over-reaction as a result of the strong stimulus from repeated playback of

I 1 0 0.2 0.4 0 0.2 0.4

0 0.2 0.4 0.6

8i e

4 4

0 0.2 0.4 0.6 0 0.2 0.4 0.6

Time (s)

Figure 11. Sonagrams of calls. (a) soror, Emei Shan, Sichuan, China, late May (one individual); (b) tephrocephalus, Emei Shan, late May (two individuals); (c) valentini, Emei Shan, late May ; (d) latouchei, Wuyi Shan, Fujian, China, early May; (e) burkii, Kalimpong district, West Bengal, India, late May (two individuals); (f) nemoralis, Kalimpong district, late May; (9) whistleri, Chamoli district, Uttar Pradesh, India, late May (two individuals). All tape recordings by P.A.

song of the own taxon. On one occasion, both a male tephrocephalus and a male valentini were in turn attracted to the speaker when their respective song type was played, although there was no reaction to the other's type of song. None of the birds tested on Emei Shan reacted strongly towards song of allopatric taxa, except for two of the valentini which responded strongly towards song of latouchei from Fujian (Table 5a). On Wuyi Shan, Fujian, China, we tested seven lutouchei. None of these responded aggressively towards foreign song, although unfortunately we did not play song of valentini (Table 5b). We did not carry out any proper playback tests in the Himalayas, but in West Bengal we once brought in both burkii and nemoralis to the same spot, one shortly after the other, when their respective song types were played. On another occasion we observed both taxa singing in the


562 f? Alstrom & U. Olsson

Table 4. Differences in vocalizations between taxa.

tephrocephalus Sichuan, China

tephrocephalus Vietnam

soror China (Shaanxi [n = 11,

Sichuan [n = 71, Guizhou [n = 11, Fujian [n = 11)

10 ind.; 5-24 (mean 13.9) strophes per ind.; 81 unique strophes; 45 unique phrases

Numbers 8 ind.; 8-28 (mean 16.6) strophes per ind.; 133 unique

strophes (1 2 without tremolos/trills); 14 unique

phrases’

9 ind.; 6-19 (mean 13.1) strophes per ind.; 120 unique

strophes (73 without tremolos/trills); 44 unique

phrases’

Presence of introductory note in strophe

Characteristics of introductory note (Fig. 10)

Presentz (97.5% of unique strophes)

Occasionally present3 Present2 (93.2% of unique strophes)

Consistent structure, both intra- and inter-individually; 1 el.; top freq. 5.3-5.8 kHz (mean 5.6

kHz); bottom freq. 3.2-3.4 kHz (mean 3.3 kHz); freq. range 1.9-2.4 kHz (mean 2.2 kHz);

mean freq. 4.3-3.6 kHz (mean 4.5 kHz); duration 46-53 ms

(mean 49 ms)

Not present5

Somewhat variable structure3; 2 el.; top freq. 5.9-7.2 kHz (mean 6.5 kHz); bottom freq. 2.2-2.7

kHz (mean 2.4 kHz); freq. range 3.3-5.0 kHz (mean 4.1 kHz);

mean freq. 4.2-4.7 kHz (mean 4.5 kHz); duration 37-69 ms

(mean 52 ms)

Variable structure, both intra- and inter-individually (narrower frequency span; generally

simpler structure; and more widely separated temporally

compared to other el. in strophe); 1 or 2 el., sometimes given 2

(or 3) times

Present (14-5870 [mean 33.0%] of

unique strophes; 3-1 1 [mean 4.81 el./tremolo)

Present (0-27% [mean 6.2%] of unique

strophes; 6-18 [mean 12.21 el h i l l )

Top freq. 4.8-7.7 kHz (mean 6.8 kHz); bottom freq. 2.0-2.9 kHz

(mean 2.4 kHz); freq. range 2.4-5.5 kHz (mean 4.4 kHz);

mean freq. 3.6-5.1 kHz (mean 4.6 kHz)

0.6-1.3 s (mean 0.9 s)

4-1 2 (mean 6.9)

Present (9-62.5% [mean 31.1%] of

unique strophes; 3-1 3 [mean 7.31 el./tremolo)

Present (37.5-91% [mean 61.9”/.] of

unique strophes; 6-24 [mean 10.41 el./trill)




0.8-1.7 s (mean 1.2 s)

6-12 (mean 8.6)

Tremolo in strophe

Trills in strophe Not present7

Frequency range of strophes’O




0.7-1.5 s (mean 1.1 s)

3-1 1 (mean 6.6)

Duration of strophes

Number of elements per strophe’,’’

Phrases’ ,Iz Present in 80-100% (mean 96.5%) of unique strophes; 2-5 (mean 3.2) ellphrase, phrase given 2-4 (mean 2.1) times

25-100% (mean 54.6%)

Present in 25-89% (mean 60.3%) of unique strophes; 2-4 (mean 2.6) ellphrase, phrase given 2-3 (mean 2.1) times

33-100% (mean 63.4%)

Present in 80-1 00% (mean 95.0%) of unique strophes; 2-3 (mean 2.3) ellphrase, phrase

given 2.0 times

50-75% (mean 58.4%) Proportion of unique elements in strophe1.l1

Proportion of unique strophes in song

35-100% (mean 60.3%) 69-1 00% (mean 94.1 Yo) 77-100% (mean 87.3%)

Calls between strophes Not present2 Not present Not presentz

ind., Individuals; el., element; freq., frequency. ’Strophes with tremolos/trills excluded. Zlntroductory note occasionally replaced by call note. 3Eight out of nine individuals have a rather consistent introductory note in 5-33% (mean 13.1%) of the strophes per individual. 4However, in one individual the call is used as an introductory note in 50% of the strophes (n = 14 unique strophes). 5However, one individual has one strophe with four identical elements in sequence. “owever, one individual has one strophe with three identical elements in sequence. ’However, in one individual two unique strophes had two different elements given three and four, respectively, times each: another individual had one strophe with two elements given three times. 8However, four



Table 4 continued.

valentini, latouchei China (Gansu [n = 51, Sichuan [n = 31,

Guangxi [n = 11, Fujian [n = l]), Vietnam (n = 2)

12 ind.; 4-23 (mean 15.2) strophes/ind.; 98 unique strophes; 57 unique phrases

Present* (98.0% of unique strophes)

Consistent structure, both intra- and inter-individually; 1 el.; top freq. 3.2-3.8

kHz (mean 3.5 kHz); bottom freq. 2.5-2.9 kHz (mean 2.7 kHz); freq. range 0.6-1.1 kHz (mean 0.8 kHz); mean freq. 2.9-3.3 kHz (mean 3.1 kHz); duration 40-57 ms

(mean 46 ms)

Not present6

Not present8

Top freq. 4.0-5.5 kHz (mean 4.8 kHz); bottom freq. 1.9-2.8 kHz (mean 2.3 kHz); freq. range 1.6-3.3 kHz (mean 2.5 kHz); mean freq. 3.0-4.0 kHz (mean 3.5 kHz)

0.7-1.2 s (mean 0.9 s)

3-8 (mean 5.2)

Present in 89-100% (mean 99.OY0) of unique strophes; 2-4 (mean 2.4)

ellphrase, phrase given 2-3 (mean 2.1) times

25-100% (mean 54.0%)

25-100% (mean 55.5%)

Present in 50% of ind.

burkii whis tleri, nemoralis India (Uttar Pradesh [n = 21,

West Bengal [n = 5]), Nepal (n = l), Bhutan (n = 2)

10 ind.; 9-21 (mean 15.1) strophes per ind.; 130 unique strophes (63 without

India (Uttar Pradesh [n = 11, West Bengal [n = 6]), Nepal (n = 3),

Bhutan (n = l), W Burma (n = 1)

12 ind.; 5-24 (mean 15.0) strophes per ind.; 106 unique strophes; 77 unique

tremolos/trills); 79 unique phrases' phrases

Not present4

Present (0-62% [mean 27.5%] of unique stro-

phes; 3-10 [mean 4.21 el./tremolo

Present (6-44% [mean 25.0%] of unique stro-

phes; 6-12 [mean 7.11 el./trill)

Present (99.0% of unique strophes)

Consistent structure, both intra- and inter-individually; 1 el.; top freq. 3.6-3.9

kHz (mean 3.7 kHz); bottom freq. 3.1-3.3 kHz (mean 3.2 kHz); freq. range 0.4-0.7 kHz (mean 0.6 kHz); mean freq. 3.3-3.6 kHz (mean 3.5 kHz); duration 27-42 ms

(mean 35 ms)

Not present

Not presentg

Top freq. 5.4-7.9 kHz (mean 6.8 kHz); bottom freq. 1.9-3.6 kHz (mean 2.8

kHz); freq. range 2.6-5.2 kHz (mean 4.0 kHz); mean freq. 4.0-5.4 kHz (mean 4.8

kHz)

Top freq. 4.1-5.9 kHz (mean 5.1 kHz); bottom freq. 1.7-3.3 kHz (mean 2.6 kHz); freq. range 1.4-3.7 kHz (mean 2.5 kHz); mean freq. 3.1-4.4 kHz (mean 3.9 kHz)

0.6-1 .O s (mean 0.8 s)

4-12 (mean 6.1)

0.6-1.2 s (mean 0.8 s)

3-8 (mean 4.9)

Present in 67-100% (mean 90.1%) of unique strophes; 2-4 (mean 2.3)

el./phrase, phrase given 2-4 (mean 2.4)

25-100% (mean 47.7%)

Present in 90-100% (mean 99.2%) of unique strophes; 2-3 (mean 2.3)

ellphrase, phrase given 2-3 (mean 2.0)

33-1 00% (mean 51.4%)

times times

43-100% (mean 82.8%) 18-100% (mean 64.2%)

Not present4 Not present

individuals have one strophe each with two different elements given three times. gHowever, three individuals have one strophe each with two different elements given three times. "JMean frequency of strophes calculated on complete strophe (elements not measured separately). lllntroductory elements not counted. 12When counting number of elements per phrase and number of times phrase repeated, only two or more consequtive identical sequences of notes were considered to be phrases, but when estimating presence of phrases, sequences with one or two elements missing at the end the second (or successive) time given were treated as phrases.


564 i? Alstrorn & U. Olsson

Table 5. Results from playback tests on (a) Emei Shan, Sichuan, China and (b) Wuyi Shan, Fujian, China. The data in the columns ‘None’, ‘Weak and ‘Strong’ are the percentage of the number of periods of playback that yielded any response. The data in the column ‘No.’ are: number of individuals tested; number of periods of playback; total duration of playback in minutes.

Response of

Soror tephrocephalus valentini

Song played None Weak Strong No. None Weak Strong No. None Weak Strong No.

a soror (Sichuan) tephrocephalus (Sichuan) burkii (Nepal) valentini (Sichuan) latouchei (Fujian) whistleri (Nepal) Seicercus affinis (Sichuan) Seicercus affinis (Fujian)

0 0 100 13;40;81 85 100 0 0 13;19;42 0 100 0 0 2;4;8 100

- 100 100 0 0 6;7; 14 100 83 17 0 5;6; 12 100 75 25 0 3;4; 10 - 100 0 0 3;3;6 -

- - -

Response of latouchei

15 0 0 0 0 0 - -

0 100 0 0 0 0 - - -

Song played None Weak Strong No.

10; 20; 53 10; 27; 53

3; 3; 6 3; 6; 16 4; 5; 10 3; 4; 8 - -

100 100

0 78 50

-

- - -

0 0

0 11 50

-

- -

0 0

100 11 0

-

- - -

6; 9; 18 6; 14; 28

6; 12; 24 4; 9; 18 2; 2; 4

-

- -

b soror (Sichuan) 100 0 0 9;7; 18 tephrocephalus (Sichuan) 100 0 0 3; 2; 4 latouchei (Fujian) 0 0 100 7;6, 12

same bush a t the same time without showing any apparent aggression towards each other.

To conclude, males of all taxa discriminate between their own and foreign song and neglect song of sympatric taxa. Playback of songs of allopatric taxa are generally disregarded, although not all taxa have been tested, and the sample sizes are very small.

Altitudinal distributions in breeding season

The altitudinal distributions are summarized in Figure 12. In Uttar Pradesh in northwestern Inha, we found burkii at c. 2085 to c. 2600 m (we did not visit any areas lower down) and whistleri at c. 2775 to c. 3050 m (top altitude just below the upper tree-limit). Accordingly, we found no altitudinal overlap between them, but in the area we visited there was no suitable habitat for either taxon between c. 2650 and 2775 m. The taxon burkii was numerous, whereas less than 15 whistleri were observed, presumably because they had not yet started breeding and were only singing rather sporadically. In Kalimpong district in northeastern India we observed burkii at c. 1700 to c. 2050 m (we did not get below about 1700 m), while we found nemoralis at c. 2000 to c. 2300 m (as high as we got in that area). Accordingly, we found both taxa together in

a narrow altitudinal zone, although nemoralis was more numerous in the overlap zone (both taxa were very common within their respective altitudmal ranges). In Darjeeling district in northwestern India we found nemoralis at c. 2300 to c. 3500 m (sparsely between c. 3300 and c. 3500 m), but we saw no burkii (though we did not visit any areas below c. 2300 m). In museum collections, we found three burkii from 2130-2440 m (Uttar Pradesh, India) and 1770 m (Sikkim), respectively, two whistleri from 3050 m (Uttar Pradesh, India) and 25 nemoralis from 2130-3500 m (central Nepal to Arunachal Pradesh, India).

On Emei Shan, Sichuan, China, we observed soror at c. 700 to c. 1300 m; tephrocephalus a t c. 1300/1400 to c. 2100/2200 m (one male at c. 1000 m in early May 1992 may have been a migrant); and valentini at c. 2100/2200 to c. 2900 m (all taxa were common within their respective altitudinal ranges). In one area we heard two inchiduals of valentini just below the uppermost tephrocephalus, while in another part of the same mountain there was a clear-cut demarcation line between them with no overlap at all. In Jiuzhaigou, southern Gansu province, China, one tephrocephalus was observed at c. 2000 m (we did not visit any places below that altitude), while valentini was common at c. 2300 to c. 3000/3100 m. On Taibai Shan in the Qin



3500

3000

2500

h

E a, 2000

3

v

-0 3 c c .-

1500

1000

500

a

I I

b

? I ? I ?

C

I ?

S T V B W B N Group

Figure 12. Altitudinal distributions (field and museums combined) in (a) Sichuan, Gansu and Shaanxi provinces, China (S, soror; T, tephrocephalus; V, valentinl); (b) Uttar Pradesh, India (B, burkii; W, whistler& and (c) West Bengal (India), Sikkim and Bhutan (B, burkit N, nemoralis). Lower limit of burkii is unknown, as is distribution of this and whistleri in the range 2600 to 2775 m in the western Himalayas.

Ling Shan range, southern Shaanxi province, China we noted soror at c. 600 to c. 1000 m; tephrocephalus at c. 1600 to c. 2200 m (there was no suitable habitat between c. 1000 m and c. 1600 m in the area we visited); and valentini at c. 2200 to c. 2700 m (all taxa were rather numerous within their respective altitudinal range).

We observed only one vahtini below the uppermost tephrocephalus. Elsewhere in the Qin Ling Shan we observed soror at c. 1000 to c. 1300 m, and tephrocephalus at c. 1200 to c. 1300 m (we only visited places between c. 1000 and c. 1300 m). In Yunnan, China two tephrocephah were collected at 3050-3350 m and another one at 3050-3660 m, while four valentini were taken by the same person in the same area at 3960-4570 m, and another one at 3050-3660 m; the altitudes are likely to be wrong for both these groups, since the tree-limit is at about 4000 m (Magnus Liden, pers. comm.), but they nevertheless indicate an altitudinal segregation (only one specimen of each species from the same 600 m interval). Another tephrocephalus from Yunnan was collected at 2740 m, while a vahtini was obtained by the same person

in the same area at 3660 m. Two tephrocephalus from Hubei, China, were collected at 1740-1 830 m, while one vahtini from the same locality was taken a t 2130 m.

Five tephrocephalus from northeastern Burma were collected at 1830-2380 m, but we have not seen any other taxon from that area in the breeding season. From the Chin hilb in western Burma we have seen three specimens of tephrocephalus from 1400-2 100 m and one from 2500 m, and seven nemoralis from 2440-2740 m. Craig Robson's tape recordings from the same area are of tephrocephalus from c. 2100 m and nemoralis from c. 2000 m (one individual of each). On Fan Si Pan, Vietnam, tephrocephalus was common at c. 1400-1780 m, while vahtini was found at 1760 to c. 1900 m; one tephrocephalus was also heard at c. 1850 m (no suitable habitat was visited below 1400 m or above 1900 m).

To summarize, sympatric taxa are segregated altitu- dinally, and the demarcation line between two taxa is often remarkably clear-cut, with hardly any altitudinal overlap. Allopatric taxa may have fairly similar altitudinal distributions.



Habitat

Seicercus soror occurs in the warm temperate evergreen broad-leaf forest zone in lush undergrowth and bamboo, and in secondary growth of rather low to moderate height (bushes and low trees) in relatively recently cleared areas. We found tephrocephalus in similar habitat in the same zone, and also in the lower part of the cool temperate zone with mainly lush deciduous broad-leaf forest (on Emei Shan, the border between these two zones is a t c. 1800 m; Lancaster 1989). In the latter zone tephrocephalus also occurs in more open forest with scattered stands of bushes, ferns and bamboo. We encountered valentini in lush undergrowth in the upper part of the cool temperate forest zone, where there is a mixture of deciduous broad- leaf and evergreen broad-leaf (e.g. Quercus and Rhododendron) and conifers (e.g. Abies). Higher up we also found it in moss-covered bamboo and deciduous scrub in coniferous forest. Latouchei was found in similar habitat as valentini. The taxon bur& occurs in much the same habitat as soror and tephrocephalus, while both whistleri and nemoralis were found in similar habitat as valentini and latouchei.

To conclude, sympatric taxa occur in at least partly different habitats, while allopatric taxa (burkii-sorod tephrocephalus; whistleri-nemoralis-valentini-latouchei) occur in similar habitats.

DISCUSSION

In our opinion, the taxa burkii, whistleri, nemoralis, tephrocephalus, soror, valentini and latouchei are suffi- ciently different morphologically to be recognized. Watson et al. (1 986) treat latouchei as a junior synonym of valentini, but in our opinion latouchei is recognizable, a t least in series, when directly compared with valentini (Table 1; eight latouchei from Fujian and four from Hubei directly compared with long series of valentini from Shaanxi and Yunnan). There are statistically significant differences between valentini and latouchei in length of tail, bill and hind-claw (Table 2 ) , and the ctfferences in tail and hind-claw, but not in bill, are significant also if only the three specimens of latouchei from Hubei province (as near valentini as possible in our sample) are included (tail [male]: Mann-Whitney U test P < 0.01; hind- claw [male]: t-test P < 0.03; bill [male]: t-test ns), suggesting that these differences may be consistent throughout the range of latouchei.

Although there are statistically significant morphological differences between our samples of whist&

and nemoralis (Tables 1 & 2) , nearly all of the nemoralis are from the eastern Himalayas (from Sikkim eastward), so the variation could be clinal. There are minor morphological differences between tephrocephalus from Sichuan (China), Yunnan (China)/ northeast BurmaNietnam and western Burmahorth- east India (Tables 1 & 2, Figs 4-7), as well as vocal differences between the first and the second (and third?) groups (Table 4, Fig. 9). If future research on a larger sample shows these differences to be consistent, it may be warranted to describe two new taxa: one from Sichuan and one from western Burmahortheast India (since the holotypes of the junior synonyms birmanica and distinctus match birds from Yunnanhortheast Burma, i.e. topotypical tephrocephalus). It is, however, possible that the name pectoralis applies to the population in western Burma/northeast India, since we have seen four specimens of tephrocephalus from Mount Victoria in western Burma labelled pectoralis (added in pencil later than the original writing). However, we have also examined two specimens of nemoralis labelled pectoralis (also from Mount Victoria), and on one of these (Academy of Natural Sciences of Philadelphia No. 137720) the word ‘Typus’ is written, but crossed out. We have not found any published reference to pectoralis, so it is possible that this name has never been used in print.

Some of the taxa (burkii-whistlen; burkii-nemoralis; nemoralis-tephrocephalus; tephrocephalus-soror-vahtini; tephrocephalus-latouchei; and soror-latouchei) are sympatric (Fig. 8). These are separate species because they differ in plumage (Table 1, Figs 4 & 5, Plate l), structure (Tables 2 & 3, Figs 6 & 73, vocalizations (Table 4, Figs. 9-1 l) , and altitudinal distribution/ habitat (Fig. 12), and also do not respond to playback of each other’s songs (Table 5).

The taxonomic ranlung of the allopatric taxa is more problematic, as is generally the case with allopatric taxa under any species concept (Alstrom in press). Whether whistleri, nemoralis, valentini and latouchei, which we believe form a monophyletic group (based on similarities in plumage [especially pattern of dark lateral crown-stripes, greater coverts and outer web of R6], structure, song, altitude and habitat choice), should be treated as separate species or as conspecific is debatable. Under the biological species concept (Mayr 1942), they are best treated as conspecific, because they can be assumed to have no intrinsic restrictions to interbreeding. However, in the case of nemoralis and vahtini, this is contralcted by the playback tests that have been carried out (Table s),


https://www.researchgate.net/publication/243771830_Systematics_and_Origin_of_Species?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz


although the sample is very small, and only one taxon and one sex have been tested. It is possible that the birds themselves consider the songs more different than we do. This is apparently true in several sympatric species of Phylloscopus (e.g. P. reguloides, P. occipitalis, and P davisoni [Martens 1980, Alstrom & Olsson 19931; P sindianus and P. collybita [Martens 1982, Helbig et al. 1996]), among which the songs are diffi- cult to distinguish. Moreover, it should be pointed out that we have not come across any intermediate specimens between nemoralis and valentini (there is a tray in the Natural History Museum, Tring, UK with supposed intermediates, but in our opinion these are all pure nemoralis). This suggests that either their ranges do not meet, or there are restrictions to the gene flow between them (alternatively, we have not come across any specimens from the area of geographical overlap).

Under the phylogenetic species concept sensu e.g. Cracraft (1983, 1989) and Nixon & Wheeler (1990), whistlerVnemoralis and valentindatouchei are separate species because they are diagnosably different (they are easily distinguishable, e.g. by the colour of the median crown-stripe and pattern of R4). It is, however, doubtful if nemoralis and latouchei can be safely distinguished from whistleri and valentini, respectively, and because of insufficient sampling, we cannot say whether the variation is clinal or clear-cut. The call note of latouchei (Fig. l l d ) may prove to be the diagnostic feature of that taxon, but due to the small sample we do not know whether it is consistent in the population where it was recorded, whether it is given throughout the taxon's range and whether it is uttered by both sexes. Under a phylogenetic species concept which accepts arbitrary inclusiveness (e.g. Mishler & Donoghue 1982, Lidkn & Oxelman 1989, Alstrom in press), these four taxa can also be treated as a single species because they are believed to form a monophyletic group.

Under the rules of the International Commission on Zoological Nomenclature (1 985), the species should be named S. valentini if all four taxa are considered conspecific, or S. whistleri and S. valentini, respectively, if only two species are recognized.

There is an analogous problem with burkii and tephrocephalus, which are similar in plumage (especially pattern of dark lateral crown-stripes, greater coverts and tail), song and habitat choice, but which are also easily separable by the colour of the median crown-stripe. The playback tests that have been carried out yielded no response (Table 5 ) , indicating that, despite the similarities in song, interbreeding might

not take place in sympatry (however, the number of playbacks is very small, and to be fully conclusive, the reaction of females should also be tested). If they are treated as conspecific, the name burkii has priority over tephrocephalus.

As has been shown above, the classification of whistleri/nemoralis/valentini/latouchei and burkii/ tephrocephalus is arbitrary, irrespective of which species concept one adheres to (whereas the classification of the sympatric taxa is straightforward). We prefer a classification based on relationships (monophyly) rather than on actual or potential interbreeding. We consider it most appropriate to treat whistleri and valentini as separate species, and we prefer to treat nemoralis and latouchei as subspecies of whistleri and valentini, respectively, to emphasize that they differ only on average from their sister taxa. Accordingly, we suggest the following taxonomic arrangement:

Seicercus burkii (Burton, 1836) Monotypic (junior synonym: Acanthiza arrogans Sundevall, 1838).

Seicercus tephrocephalus (Anderson, 187 1) Monotypic (junior synonyms: Cyptolopha birmanica Berezowslu & Bianchi, 1891; and Cyptolopha burkii distincta La Touche, 1922)

Seicercus soror Alstrom & Olsson, sp. nov. Monotypic.

Seicercus whisthi Ticehurst, 1925 Seicercus whistleri whistleri Ticehurst, 1925 Seicercus whistleri nemoralis Koelz, 1954

Seicercus v u h t i n i (Hartert, 1907) Seicercus valentini valentini (Hartert, 1907) Seicercus valentini latouchei Bangs, 1929

We are indebted to Ian Lewington for producing the excellent colour plate; to Walter J. Bock for advice on nomenclatural matters; to Roald L. Potapov for examining, measuring and photographing the holotype of S. b. bimanica for us; to Robert Prys-Jones for his efforts in trying to track down the type specimen of S. b. burkii; and to J.B. Alfred, S. Sathykumar and Pratap Singh for their help in finding out the fate of the type of S. b. tepkrocephalus. We are most grateful to Walter Bock, Robert Prys-Jones, Alan Knox, Pamela C. Rasmussen, Fredrik Ronquist and Per Sundberg for their comments on the manuscript. Invaluable assistance was provided by the following persons during our museum visits or in connection with loans: Mark Adams, Robert Prys-Jones, Michael Walters and F.E. Warr at the Natural History Museum, Tring, UK; Paul Sweet and Mary LeCroy at the American Museum of Natural History, New York, USA; Pamela C. Rasmussen a t the National Museum of Natural History, Washington, DC, USA; Raymond A. Paynter, Jr. and


https://www.researchgate.net/publication/238324781_Species_Concepts_and_Speciation_Analysis?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz

https://www.researchgate.net/publication/264567829_Phylogeny_and_species_limits_in_the_Palaearctic_Chiffchaff_Phylloscopus_collybita_complex_Mitochondrial_genetic_differentiation_and_bioacoustic_evidence?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz


https://www.researchgate.net/publication/269906227_Species_Concepts_A_Case_for_Pluralism?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz


568 F! Alstrom & U. Olsson

Alison Pirie at the Museum of Comparative Zoology, Harvard University, Boston, USA; Robert B. Payne and Janet Hinshaw at the Museum of Zoology, University of Michigan, Ann Arbor, USA; David Willard at the Field Museum of Natural History, Chicago, USA; David J. Agro at the Academy of Natural Sciences of Philadelphia, Philadelphia, USA; and Tan Yao-kuang at the Zoological Institute, Academia Sinica, Beijing, China. We also sincerely acknowledge the financial support which has been given to this project over the years from: Elis Wides Fond (Swedish Ornithological Society), Stiftelsen Erik och Lily Philipsons Minnesfond, Janneman Schmidts Stipendiefond, Kungliga och Hvitfeldtska Stipendieinrattningen, Olle och Signhild Engkvists stiftelser, and Uddenberg-Nordingska Stiftelsen. We are most grateful to Geoff Carey, Mike Crosby, Paul Holt, Tim and Carol Inskipp, Darren Irwin, Hannu Jannes, Ben King, Jochen Martens, Richard RanWNational Sound Archive, London, Craig Robson, Philip D. Round and Edward Vercruysse for lending us tape recordings; to Josef Redfors and Stefan Rizell for translating texts in Latin and Russian, respectively; to Bjorn Johansson for assistance in the field; to Hannu Jannes for catching and measuring some birds for us in China; to Bjorn Alden, Thor-Bjorn Engelmark and Magnus Liden for information on botany; to Johan Wallander for assistance with some of the statistics; and to Ante Strand for digging out many of the references for us.

REFERENCES

Ali, S. & Ripley, S.D. 1973. Handbook of the Birds of lndia and Pak- istan, Vol. 8. Delhi: Oxford University Press.

Alstrom, P. in press. Biological or phylogenetic species? Competing species concepts in ornithology. Alula.

Alstrom, P. & Olsson, U. 1993. Blyths Leaf Warbler Phylloscopus reguloides found breeding in Thailand. Forktail 9: 150-1 52.

Anderson, J. 1871. On eight new species of birds from Western Yunan, China. Proc. Zoo/. Soc. Lond. p. 213.

Bangs, 0. 1929. The Chinese forms of Seicercus of the burkiiform- circle. Proc. New England Zool. Cl. 11 : 1-5.

Berezowski, M. & Bianchi, V. 1891. Ptitsi Gansuiskago Putesh- estvia G.N. Potanina 1884-1 887.

Burton, E. 1836. Characters of several birds from the Himalayan mountains. Proc. Zool. SOC. Lond. Part 3 (1 835): 152-1 54.

Cheng, T-h. 1987. A Synopsis of the Avifauna of China. Beijing, Hamburg & Berlin: Paul Parey.

Cracraft, J. 1983. Species concepts and speciation analysis. In Johnston, R.F. (ed.) Current Ornithology: 159-1 87. New York: Plenum Press.

Cracraft, J.1989. Speciation and its ontology: the empirical consequences of alternative species concepts for understanding patterns and processes of differentiation. In Otte, D. & Endler, J.A. (eds) Speciation and Its Consequences: 23-59. Sunderland, MA: Sinauer.

Cramp, S. (ed.) 1988. The Birds of the Western Palearctic, Vol. 5. Oxford: Oxford University Press.

Grant, P. & Mullarney, K. 1989. The New Approach to Identification. Ashford, privately published.

Hartert, E. 1907. Die Vogel der Palaarktischen Fauna, Vol. 1: 497. Berlin: Friedlander.

Helbig, A., Martens, J., Seibold, I., Henning, F., Schottler, 8. & Wink, M. 1996. Phylogeny and species limits in the Palearctic chiffchaff Phylloscopus collybita complex: mitochondria1 genetic differentiation and bioacoustic evidence. lbis 138: 650-666.

lnternational Commission on Zoological Nomenclature 1985. lnternational Code of Zoological Nomenclature, 3rd edn. Berkley: University of California Press.

King, B., Dickinson, E.C. & Woodcock, M. 1975. A Field Guide to the Birds of South-East Asia. London: Collins.

Koelz, W.N. 1954. Ornithological studies. Contrib. lnst. Regional Exploration, 1: 17.

Lancaster, R. 1989. Travels in China, a Plantsman’s Paradise. Woodbridge, Suffolk: Antique Collectors’ Club.

La Touche, J.D. 1922. Bull. Br. Ornithol. Club 43: 41. Lekagul, B. & Round, P.D. 1991. Birds of Thailand. Bangkok: Saha

Lesson, R.P. 1839. Avium species novae. Rev. Zool. Paris2: 104. Liden, M. 81 Oxelman, B. 1989. Species - pattern or process?

Taxon 38: 228-232. Martens, J. 1980. Lautausserungen, verwandtschaftliche Beziehun-

gen und Verbreitungsgeschichte asiatischer Laubsanger (Phylloscopus). Berlin: Verlag Paul Parey.

Martens, J. 1982. Ringformige Arealuberschneidung und Artbildung bei m Zi I pzal p, Phylloscopus collybita. Das lorenzii- Problem. Z. Zool. Syst. Evolutionsforsch. 20: 82-1 00.

Mayr, E. 1942. Systematics and the Origin of Species. Cambridge, MA: Harvard University Press.

Mishler, B.D. & Donoghue, M.J. 1982. Species concepts: a case of pluralism. Syst. Zool. 31: 491-503.

Mitchell, S., Cunningham, S., McClellan, J. & Montgomery, J. 1995. Canary 1.2. Ithaka: Cornell Laboratory of Ornithology.

Nixon, K.C. & Wheeler, Q.D. 1990. An amplification of the phylogenetic species concept. Cladistics 6: 21 1-223.

Roberts, T.J. 1992. The Birds of Pakistan, Vol. 2 Passeriformes. Karachi: Oxford University Press.

Sundevall, C.J. 1838. Foglar frAn Calcutta, samlade och beskrifne af Carl J. Sundevall. Physiogr. Sallskap. Tidskr. 1: 77.

Swainson, W. 1837. On The Natural Hisfory and Classification of BirdsVol. 2: 259-260. London: Longman, Brown, Green & Long- mans.

Karn Bhaet Co.

Ticehurst, C.B. 1925. Bull. Br. Ornithol. Club 46: 22-23. Watson, G.E. , Traylor, M.A., Jr. & Mayr, E. 1986. Family Sylviidae,

Old World Warblers. In Mayr, E. & Cottrell, G.W. (eds) Checklist of Birds of the World, Vol. 11: 3-294. Cambridge, MA: Museum of Comparative Zoology.

Received 6 November 1998; revision accepted 10 June 1999

@ 1999 British Ornithologists’ Union, lbis, 141, 545-568

https://www.researchgate.net/publication/221959261_An_Amplification_of_the_Phylogenetic_Species_Concept?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz

https://www.researchgate.net/publication/221959261_An_Amplification_of_the_Phylogenetic_Species_Concept?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz




https://www.researchgate.net/publication/233820782_The_Birds_of_The_Western_Palearctic?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz

https://www.researchgate.net/publication/233820782_The_Birds_of_The_Western_Palearctic?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz



https://www.researchgate.net/publication/272551989_Species_Pattern_or_Process?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz

https://www.researchgate.net/publication/272551989_Species_Pattern_or_Process?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz

https://www.researchgate.net/publication/266141640_Blyth's_Leaf_Warbler_Phylloscopus_reguloides_found_breeding_in_Thailand?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz

https://www.researchgate.net/publication/266141640_Blyth's_Leaf_Warbler_Phylloscopus_reguloides_found_breeding_in_Thailand?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz








https://www.researchgate.net/publication/267763057_A_Field_Guide_to_the_Birds_of_South-East_Asia?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz

https://www.researchgate.net/publication/267763057_A_Field_Guide_to_the_Birds_of_South-East_Asia?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz



https://www.researchgate.net/publication/272810116_The_Birds_of_Pakistan?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz

https://www.researchgate.net/publication/272810116_The_Birds_of_Pakistan?el=1_x_8&enrichId=rgreq-1f79a9d5ad346359561e2e14641424b5-XXX&enrichSource=Y292ZXJQYWdlOzIyNzU2MzM1MDtBUzo5OTg3NjU2NjIwODUxM0AxNDAwODIzODM3MDgz

The Golden-spectacled Warbler: a complex of sibling species, including a previously undescribed species

Documents