THE RUFOUS CACHOLOTE (FURNARIIDAE: PSEUDOSEISURA) IS TWO SPECIES L

7/w Condor IO2:409-422 0 The Cooper Ormthological Somety 2WO

THE RUFOUS CACHOLOTE (FURNARIIDAE: PSEUDOSEISURA) IS TWO SPECIES l

KEVIN J. ZIMMER? Los Angeles County Museum of Nutural History, 900 Exposition Blvd., Los Angeles, CA 90007, e-mail: kjzsrz@ tcmnet

ANDREW WHI~AKER Estrado do Aleixo, Conjunto Acariquaru Sul, Rua da Samaumas 214, Mamas, Amazonas, Brazil 69085

Abstract. New information on the vocalizations and behavior of populations of the Ru- fous Cacholote (Pseudoseisura cristata), combined with a morphometric analysis of museum specimens, reveal that it actually consists of two biological species: a form that is restricted to the arid caatinga region of northeastern Brazil (P. cristata), and a wider-ranging form (P. unirufa) that occupies seasonally flooded savannas and deciduous woodlands in northern and eastern Bolivia, northern Paraguay, and southwestern Brazil. Long considered a subspecies of P. cristata, P. unirufu is shown to be morphologically, vocally, and ecologically distinct from allopatrically distributed populations of P. cristatu. Reciprocal playback experiments of tape-recorded vocalizations indicate that neither form responds to the other’s vocalizations. The two species may also differ in their breeding systems and social struc- tures. P. cristatu is shown to be a cooperative breeder, with nonbreeding helpers assisting in territorial defense, nest-building, and care of nestlings. No evidence of cooperative breeding was found in P. unirufa. Kev words: cooaerative breeder, duets, Furnariidae, Pseudoseisura cristata, Pseudos- eisura unirufa, species’ limits, vocalizations.

INTRODUCTION In this paper we present evidence that the sub- The genus Pseudoseisuru as currently treated is species of Rufous Cacholote are distinct biolog- comprised of three species of large, rather ab- ical species, which, although similar in general errant furnariids (the cacholotes) whose affinities appearance, are morphologically, vocally, and outside of the genus are uncertain (Vaurie 1980). ecologically differentiated and which appear to The three species of cacholotes are more or less differ in their social structure. allopatrically distributed through eastern and METHODS southern South America (Peters 1951, Vaurie 1980, Ridgely and Tudor 1994). Two of the spe- We made observations of “Rufous” Cacholotes cies (P. lophotes and P. gutturalis) are consid- at various sites in Mato Grosso, Brazil (August ered monotypic. The third, P. cristufu (the Ru- and October 1991, each September from 1993- fous Cacholote), is currently treated as a poly- 1998), Pernambuco, Brazil (January 1996-1999 typic species consisting of two disjunctly dis- and February 1996) and Bahia, Brazil (January tributed subspecies: nominate P. c. cristata 1996-1999 and February 1996). All measure- (hereafter cristutu) of northeastern Brazil, and P. ments used in behavioral data (distances, c. unirufa (hereafter unirufu) of northern and heights, etc.) are estimates. Mapped distributions eastern Bolivia, northern Paraguay, and south- (as they appear in this paper) are based entirely western Brazil (Hellmayr 1925, Peters 1951, on label data from specimens that the senior au- Ridgely and Tudor 1994). Although the Rufous thor examined and by more recent records doc- Cacholote is a large, noisy, and generally con- umented by tape recordings. These localities spicuous bird of open habitats, it has received were entered into a Geographic Information Sys- slight attention from ornithologists, and little has tem (Isler 1997) and mapped by Morton Isler been published of its basic biology (Vaurie (Division of Birds, National Museum of Natural 1980). History, Washington, D.C.). We assume that vocalizations of cacholotes, like those of other suboscines, are mostly or en- ’ ’ Received 13 May 1999. Accepted 17 January 2000. 2 Current address: 1665 Garcia Rd., Atascadero, CA tirely inherited (Kroodsma 1984, 1989, Kroods- 93422. ma and Konishi 1991), and as such, provide po-

[4091 410 KEVIN J. ZIMMER AND ANDREW WHITTAKER tentially important characters for systematic calizations, or were generally nonterritorial dur- study as in other suboscines (Bierregaard et al. ing the trial period and therefore unresponsive 1997, Isler et al. 1997, Krabbe and Schulenberg to playback of any kind. Failure to account for 1997). To analyze vocalizations, we assembled seasonal differences in responsivenesshas been recordings of cristutu (n = 41 cuts involving 30 cited as a potential design weakness in other individuals, pairs, groups) and unirufu (n = 23 playback experiments (Kroodsma 1986). Re- cuts involving 17 individuals or pairs). A “cut” sponses to playback were characterized as contains one or more consecutive vocalizations strong, moderate, weak, or none. A “strong” re- from a given individual, pair, or group of cach- sponse involved immediate and repeated vocal- olotes. Locations and recordists for all record- izations (from previously nonvocalizing birds) ings examined are listed in Appendix 1. For as well as approach toward the sound source. A comparison, vocalizations were categorized as “moderate” responseinvolved either a cautious loudsongs, duets, chatter, or calls. “Loudsongs” approach without vocalizing, or sustained vo- were consistently patterned multi-note vocali- calizing without approach. A “weak” response zations (Isler et al. 1997) given by an individual involved single or unsustained vocalizations bird, seemingly in the context of territorial ad- (from previously nonvocalizing birds) without vertisement. These were given individually by approach. The “none” category includes in- both sexes (most commonly when individuals stances in which birds remained silent (in the were separated by considerable distances), but case of previously nonvocalizing birds) and did were also incorporated into duets. “Duets” in- not approach the sound source, as well as involved simultaneous singing by a pair or family stancesin which already vocalizing birds neither group of cacholotes. Duets consisted of typical changed the delivery or rate of their vocaliza- loudsongsof both sexes,combined with a “chat- tions, nor approachedthe sound source. ter” that was not given outside of a duet context. The senior author examined representative Vocalizations categorized as calls were of low specimens of cristata (17) and unirufu (11). amplitude and structurally simple (usually in- These specimens are housed at the Academy of volving single notes) and were seemingly given Natural Sciences of Philadelphia, Philadelphia in the context of contact vocalizations between (ANSP), Carnegie Museum, Pittsburgh (CM), mates or family members. Our tape recordings Field Museum of Natural History, Chicago were made with Sony TCM-5000 recorders,and (FMNH), Los Angeles County Museum of Nat- Sennheiser MKH-70, ME-66, and ME-SO micro- ural History, Los Angeles (LACM), the Louisi- phones. All recordings have been or will be ar- ana State University Museum of Natural Sci- chived at the Library of Natural Sounds, Cornell ence, Baton Rouge (LSUMZ), and the National University, Ithaca, New York. Spectrograms Museum of Natural History, Washington, D.C. were made by Phyllis Isler (Division of Birds, (USNM). A list of specimens examined is pro- National Museum of Natural History, Washing- vided in Appendix 2. A wing rule with a per- ton, D.C.) on a Power Macintosh 7500 computer pendicular stop at zero was used to measure flat- using Canary version 1.2.1 (Bioacoustics Re- tened wing chord (wing), tail length (tail), and search Program, Cornell Laboratory of Omi- hind toe length without the claw (hind toe), and thology, Ithaca, New York). dial calipers were used to measure tarsus length Playback experiments were conducted to de- (tarsus), culmen length from the anterior end of termine the reactions of cacholotesof each sub- the nares to the tip (culmen), bill depth at the species to vocalizations of the other. In each anterior end of the nares (bill depth), bill width case, an individual, pair, or group of cacholotes at the anterior end of the nares (bill width), and was first presented with 5-10 tape recordings of hind claw length (hind claw). All measurements the other taxon, given at typical intervals of made with calipers were made to the nearest about 15 sec. After a “buffer” period of about 0.01 mm, those taken with the wing rule were 2 min, the cacholotes were then presented with made to the nearest 0.5 mm. Measurement ter- 5-10 tape recordings of their own vocal type. minology conforms with that used by Pyle et al. This was done to account for potential seasonal (1987). General linear models were used to in- influences by resolving the question of whether vestigate age- and gender-corrected differences birds that were unresponsive in the first trials between the two subspeciesfor each of the eight were discriminating between the different vo- characters measured. Residual diagnostics indi- THE RUFOUS CACHOLOTE IS TWO SPECIES 411

FIGURE 1. Distribution of Rufous Cacholote (Pseudoseisurucr&ztu) as confirmed by examination of specimens or tape recordings:black squares= P. c. cristatu; black triangles = P. c. min.&u. “T” next to a symbol indicates the type locality for that taxon. The Moxos, Bolivia type locality for P. c. unirufa pertainsto a region, not a locality. Because this region was historically centered in the Trinidad area, Departamento Beni, the “T” is placed there. cated no violations of the general linear model PLUMAGE assumptions of error normality and constant var- Both forms of “Rufous” Cacholotes are some iance. Plumage was described from museum shade of uniform rufous-brown throughout, ex- specimens and compared to a standard color ref- cept for a contrasting grayish crown found in erence (Smithe 1975). most unirufa (see below). The crown feathers of all individuals are highly elongate, forming a RESULTS distinctive crest. As is typical of fumariids in DISTRIBUTION general, males and females are not sexually di- The two forms of “Rufous” Cacholote are al- morphic in plumage. Adults of both cristutu and lopatrically distributed (Fig. 1). Nominate cris- unirufu have bright yellow eyes (pers. observ.). tutu is endemic to the Brazilian states of Bahia, Juvenile cristutu have dusky scaling or barring Cear& extreme eastern MaranhHo, Minas Gerais, on the sides of the face, throat, and breast, and Paraiba, Pernambuco, and Piaui (Peters 1951, show variable amounts of dusky streaking or Ridgely and Tudor 1994). P. c. unirufu occurs barring on the crown (Vaurie 1980; pers. ob- in two disjunct populations, one in northern Bo- serv.). Juvenile speiimens of cristutu from livia (Departamento Beni), and the other extend- LACM also showed whitish tips or fringes to ing from eastern Bolivia (Departamento Santa some feathers of the throat and abdomen, a char- Cruz) eastward through northern Paraguay (Alto acter which would be difficult to discern in the Paraguay) to southern Mato Gross0 and Mato field. Juveniles of cristutu are further distin- Gross0 do Sul, Brazil (Peters 1951, Ridgely and guished in the field by their bluish-gray eyes Tudor 1994). (pers. observ.). We do not know the extent to 412 KEVIN J. ZIMMER AND ANDREW WHITTAKER

TABLE 1. Ranges (sample sizes) and means k SD of selected measurements (mm) of “Rufous” Cacholotes (Pseudoseisura cristata), followed by a statistical comparison of the means. Values of P derived from Analysis of Variance (ANOVA).

Characters cristata unirufa P

Culmen 14.1-19.8 (15) 13.1-15.5 (11)

which juveniles of unirufu share these plumage from FMNH (D. Stotz, pers. comm.). Two other or soft-part characters(if at all). What appeared specimens of cristutu had the crown feathers to be a food-begging juvenile bird accompany- concolor with the mantle but had contrasting ing two adults of unirufu near Pixafm (Mato dark streaks along the shafts of some feathers. Grosso, Brazil) had yellow eyes similar to those These birds also had some pale barring on the of the adults, and we saw no evidence of dusky throat and abdomen, and had bill measurements barring on the face or underparts. intermediate between those of juveniles and Nominate cristutu are noticeably paler above adults examined. These two birds were pre- than are unirufu. Most specimens examined ap- sumed to represent older sub-adults. Most uni- proximated Tawny (color #38; capitalized color rufu (8 of 11 specimens examined) had most names from Smithe 1975) and several speci- crown feathers distinctly Glaucous (color #79), mens were paler still (closer to Cinnamon, color with a darker, more brownish-gray longitudinal #39). Upperparts of unirufa are a darker, redder streak along the shaft. Three of these specimens brown, closer to Ferruginous (color #41), with had the crowns somewhat variegated, the over- the tail being darker still (between Ferruginous lying feathers being Glaucous, but with many and Chestnut, color #32). Only one specimen underlying or rear-crown feathers (the longest (USNM 16846) was contradictory to these pat- feathers of the crest) concolor with the mantle. terns. That individual was identified as cristatu, These three specimens were from Departamento but had the upperpartsdarker and closer to uni- Beni (LSUMZ 124081) and Departamento Santa rz& in color. The label for this specimen lacked Cruz (LSUMZ 37649), Bolivia, and from Mato locality data. Contrary to Ridgely and Tudor Grosso. Brazil (ANS 1 I 1616). The three re- (1994) there was no apparent difference in col- maining unirufu examined (LSUMZ 80296, oration between the two specimensfrom Depar- FMNH 334462, and FMNH 180345) had the tamento Beni, Bolivia, and other unirufu, apart crown concolor with the mantle, with no trace from crown color (see below). of gray. These specimenswere from Mato Gros- Crown colors differed significantly between so, Brazil, Departamento Beni, Bolivia, and De- cristutu and unirufu. All adult specimensof cris- partamento Santa Cruz, Bolivia, respectively. tutu (n = 12) had crowns concolor with the man- tles. This was consistent with our field obser- SIZE vations. Three juvenile specimens of cristutu In most measurable characters,cristutu averaged had many crown feathers contrastingly grayish larger than unimfu (Table 1). Several of these with dark longitudinal shaft streaks and dusky differences were statistically significant, al- terminal bars, as did another juvenile specimen though ranges for each character overlapped. THE RUFOUS CACHOLOTE IS TWO SPECIES 413

I o-1 I 0 0.5 I 0 1.5 20 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 0 05 10 1.5 20 Time (seconds)

FIGURE 2. Loudsongs of Rufous Cacholotes. (A) P. c. crisfafa (presumed male) from Pernambuco, Brazil (K. J. Zimmer recording). (B) P. c. unirufa (presumed male) from Mato Grosso, Brazil (K. J. Zimmer recording). (C) P. c. crisfafa (presumed female) from Pernambuco, Brazil (K. J. Zimmer recording). (D) P. c. unirufa (presumed female), partial song from Departamento Beni, Bolivia (T A. Parker recording, LNS # 38375).

There was little gender effect in any of the mod- is similar in pattern, but is shorter in length and els, but there was a significant (P < 0.05) age differs in the shape of individual notes and in effect introduced by the inclusion of three ju- frequency (Fig. 2B). The presumed female loud- venile cristafu in the analysis. These juvenile song of cristata is an evenly paced series of 15- birds had the smallest bills of any of the cristutu 35 well-spaced, frequency modulated “zjeep” specimens. When only adult specimens were notes with a particularly strident, buzzy quality considered, culmen size was diagnostic. Culmen (Fig. 2C). The presumed female loudsong of lengths (mm) of adults ranged from 15.7 to 19.8 unirufa (Fig. 2D) is a similarly evenly spaced, for cristata and 13.1 to 15.5 for unirufa. Mea- long series of identical “CHEP” notes (“chep- surements from our sample fit well with those chepchepchepchep . .“). all at essentially one from an identically sized sample (n = 17 [cris- pitch, but delivered at a much faster rate, and tuta], n = 11 [unirufa]) reported by Hellmayr without the notably buzzy quality found in songs (1925). of female cristuta. Both tonal differences in individual notes and rate differences in the loud- VOCALIZATIONS songs of these two taxa are immediately obvious Spectrographic analysis confirmed our field im- to the human ear. pressions that cristata and unirufu differed strik- Because “Rufous” Cacholotes are sexually ingly and consistently in several aspects of their monomorphic, it was not possible in the absence vocal repertoires. Both forms appear to have ho- of hand-sexed, color-marked birds to assign a mologous, sex-specific loudsongs, which are de- particular vocalization to one sex with absolute livered independently and as part of duets. What certainty. However, birds of both forms which we believe to be the male loudsong in cristutu gave the more complex “bouncing ball” loud- is a descending series of notes that begins with song also were consistently the first members of 4-10 evenly spaced notes and then accelerates a pair or group to respond aggressively to tape into a short rattle (“&up chllp chllp chup chup playback. These loudsongs also were more fre- chachachchch .“) (Fig. 2A). These loudsongs quently delivered independent of a duet. Within have a distinct “bouncing ball” pattern, in duets by mated pairs, one individual consistently which introductory notes are widely spaced but gave the bouncing ball loudsong, whereas the between-note intervals rapidly diminish. What other individual consistently gave the repetitive we believe to be the male loudsong of unirufa long series. Thus, we can safely assume that the 414 KEVIN J. ZIMMER AND ANDREW WHI’II-AKER

5 --

4 --

3 --

2 --

1 --

0 I I I I I I I I I I I 1 I 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 Time (seconds) FIGURE 3. Duet vocalizationsof Pseudoseisura crisata cristata. (A) partial mate-femaleduet from Bahia, Brazil (T A. Parkerrecording, LNS #35711). (B) partial male-femaleduet from Pemambuco,Brazil (K. J. Zimmer recording).(C) isolatedduet chatter (presumed male) from Pemambuco, Brazil (K. J. Zimmerrecording).

different types of loudsongs are specific to one simultaneous duets. Duets were initiated by ei- sex or the other. The greater structural complex- ther sex, with the partner joining in immediately. ity of bouncing ball songs, along with their The endings to duets were usually well-coordi- greater frequency of occurrenceoutside of duets nated, and it was rare for a bird to sing more and the more aggressive territorial response of than one loudsong (males) or a few notes (fe- individuals which gave them point to these be- males) after its mate had stopped. ing male songs. The pattern of duets was similar between cris- The most frequently heard vocalizations from tutu and unirufa. In cristutu (Fig. 3A, B), a typ- all “Rufous” Cacholotes were duets (Fig. 3 and ical duet involved the male and female simul- 4). As in some other fumariids (e.g., Crunioleu- taneously delivering their respective loudsongs, ca vulpina; Zimmer 1997), a song from one with the male filling many of his between-song cacholote usually elicited an immediate duet re- intervals with a continuous, low, buzzy chatter sponse from its mate. Based on field observa- (Fig. 3C). During periods in which the male was tions, we would estimate that as much as 75% delivering the chatter, the strident notes of the of male and female song bouts overlapped as female’s loudsong stood out as the most con- THE RUFOUS CACHOLOTE IS TWO SPECIES 415

A 7-

6 -~

5--

4 --

3 -~

2 --

1 --

4 --

1 -.

7-c

6 -~

5 --

3 --

2 --

0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0

Time (seconds) FIGURE 4. Duet vocalizations of fseudoseisura crisfafa unirufa. (A) partial male-female duet from Mato Grosso, Brazil (K. _I.Zimmer recording). (B) partial male-female duet from DepartamentoBeni, Bolivia (T A. Parker recording, LNS #38375). (C) isolated duet chatter (presumedmale) from DepartamentoBeni, Bolivia (T A. Parker recording, LNS #38375). spicuous element of the duet. When the male switched vocalizations also varied from one duet delivered his loudsong, the female continued to to the next. One duet might involve regular sing at the same rate, but her notes then became switching (loudsong, chatter, loudsong, chatter, background to the male’s louder song. Thus, al- etc.), whereas the next might be dominated by though the male and female song bouts were si- chatter with only one or two irregularly spaced multaneous, the alternation by the male in songs. The male chatter was frequently a sub- switching from a far-carrying loudsong to a low- dued prelude to a duet, particularly after sepa- er-amplitude chatter had the effect of making the rated birds had just reunited in responseto tape duet sound antiphonal. This effect was accen- playback. In these instances, a few seconds of tuated by the evenly paced, relatively lengthy male chatter appeared to be the synchronizing gaps between female notes. The male’s ability cue that precipitated a duet. to alternate from loudsong to chatter and back When juveniles joined parents in duets (see to loudsong without detectable pauses was re- below), the adult male and female sang their re- markable. The pattern with which males spective songs, and the juvenile most frequently 416 KEVIN J. ZIMMER AND ANDREW WHITTAKER provided the chattering backdrop that is nor- cance of vocal differences among the taxa, and mally delivered by the adult male. When groups of their role as potential isolating mechanisms. of four-six cristata defended their territories We performed tape playback experiments on with song bouts, it was common for two males pairs of unirufa at various points along the and two females to sing typical loudsongs, with Trans.Pantaneira Highway near the settlement of one or more birds providing constant chatter or Pixaim, Mato Grosso, Brazil (seven pairs in Sep- for singing males to alternate between chatter tember 1996 and nine pairs in September 1998) and loudsong. In such instances,the two females and near Cuiaba, Mato Grosso, Brazil (three displayed everything from neat-perfect synchro- pairs in September 1996). Fifteen of the 19 pairs nization (complete overlap) of their notes to an- showed no responseto tape playback of cristata tiphonal coordination. Qualitative differences duets, yet respondedstrongly to playback of uni- could usually be heard between songs of two rufa duets. In most of these casespairs of cach- females in a single song bout, one female giving olotes flew in from distancesof 50-100 m away, the typical buzzy “zjeep,” and the other a more landed next to us, and duetted repeatedly in re- guttural “zherp.” Songs of two males within a sponse to tape playback. Three pairs responded single song bout usually only partially over- weakly to playback of nominate cristata. In each lapped or not at all. case, the pair of cacholotes responded to initial The duet pattern of unirufa (Fig. 4A, B) was playback by moving to a higher perch in the similar to that of nominate birds, in that the male same tree and singing a single duet. Additional and female simultaneously delivered their re- playbacks failed to elicit further response.When spective loudsongs,with the male frequently fill- we switched to playback of unirufa duets, each ing his between-song intervals with a continuous of the pairs approachedus aggressively and du- chatter (Fig. 4C). As in cristata, the male pattern etted repeatedly. Another pair approachedclose- of switching between songs and chatter varied ly (without vocalizing) once after a single play- from one duet to the next. Duets of the two taxa back of nominate cristata, but did not respond sound distinctly different. Duets of unirufa to additional playbacks. This same pair respond- sound like a uniformly loud, fast-paced jumble ed immediately and aggressively five consecu- of chattering notes as the “chippering” female tive times to playback of unirufa vocalizations. loudsong and the alternating male loudsong and In each case, we delayed the next trial until after chatter blur together. By contrast, the duets of the pair had ceased vocalizing and had left the cristata are distinctly slower and more antipho- immediate area. nal, and are characterized by the unique strident We performed tape playback experiments on notes of the females, the tonal quality of which pairs or groups of nominate cristata in January is unlike anything delivered by unirufa. We have 1997 (two pairs-groups near Exu and nine pairs- not encountered more than two unirufa singing together. groups near Lagoa Grande, Pemambuco, Brazil; Other calls by both taxa were limited to single and two pairs-groups near Itaberaba and one “chuk” notes and subdued, short chatter,both of group at Lencois, Bahia, Brazil), January 1998 which appeared to be given by both sexes and (one group near Exu and another near Lagoa all ages. These seemingly functioned as contact Grande), and January 1999 (one pair near Exu notes and in coordinating behaviors between and nine pairs-groups near Lagoa Grande, and mates and/or within family groups. one pair near Canudos, Bahia, Brazil). Three of Tape recordings of the disjunct population of these pairs were completely unresponsive to unirufa from Departamento Beni, Bolivia (Fig. playbacks of any kind. One pair gave a single 4B) sound identical to those of the more wide- duet after several playbacks of unirufa vocali- spread population of eastern Bolivia, northern zations, but did not approachor sing further, and Paraguay, and southwestern Brazil. Similarly, the timing of their duet was such that we suspect we could find no geographic variation in the it was incidental to our playback. When we voice of nominate cristata. switched to playback of cristata vocalizations, the pair flew directly at us from more than 100 PLAYBACK TRIALS m, landed in an adjacent tree, and gave several Playback experiments with Pseudoseisura re- loud duets. The remaining 23 pairs or groups cordings offer further evidence of the signifi- showed no response to repeated playbacks of THE RUFOUS CACHOLOTE IS TWO SPECIES 417

TABLE 2. Summary of playback trials (as detailed in the text) involving the two subspecies of Pseudoseisura cristata (P. c. cristata and P. c. unirufa). “Subject taxon (locality)” identifies the taxon of the bird on which the trial was being conducted, and the site. Locations where these trials were conducted were: Exu and Lagoa Grande (Pernambuco, Brazil); Itaberaba, Lencois, and Canudos (Bahia, Brazil); and Pixaim and Cuiaba (Mato Grosso, Brazil). Codes for Brazilian states are as follows: PE = Pernambuco, BA = Bahia, and MG = Mato Grosso.

Response to playback of Subject taxon (locality) n P. c. cristata P. c. unirufa P. c. cristata Exu, PE 1 strong weak Exu, PE 3 strong none Lagoa Grande, PE 17 strong none Lagoa Grande, PE 2 none none Ltaberaba, BA I strong none Itaberaba, BA I none none Lencois,BA I strong none Canudos,BA 1 strong none Totals 27 strong = 24 none = 26 none = 3 weak = 1 P. c. unirufa Pixaim, MG 13 none strong Pixaim, MG 2 weak strong Pixaim, MG 1 moderate strong Cuiaba, MG 2 none strong Cuiaba, MG 1 weak strong Totals 19 none = 15 strong = 19 weak = 3 moderate = 1 univufa, yet respondedstrongly and immediately new nest. Whittaker found a second territory to playbacks of cristutu (Table 2). along the same road which was defended by four birds, all apparently adult in eye color and plum- SOCIALITY AND BEHAVIOR age. In January 1996 near Lagoa Grande (Pemam- In January 1997 and 1998, we performed buco, Brazil), we found on a telephone pole an playback experiments in 16 cristutu territories active cristata nest which was being attended by (as detailed above) in Pemambuco and Bahia. four birds. The open terrain allowed us to vi- Birds at two of the territories were unresponsive sually track the movements of foraging birds as to playback. At five of the territories, tape play- they made trips to and from the nest to feed back resulted in territorial defense by a single young. After verifying that each of the birds was pair of birds, although in some instances other bringing arthropod prey to the same hole in the cacholotes (which may have been occupants of large stick nest, we played tape of cristuta in an the same territory) also responded vocally from attempt to confirm our count of nest attendants. more than 100 m distant without approaching. Four birds, each of which appeared to be adult, At five territories, playback resulted in close ap- responded by immediately flying to the crossbar proach and repeated song bouts from four-six of the pole on which the nest was placed. They birds (two groups of four, two groups of five, gave several excited song bouts before flying and one group of six). At four other territories, off. In February 1996, A. Whittaker returned to playback resulted in defense by two adults and the same spot and found a group of five birds, a single juvenile. In all instances, the vocalizing at least two of which were juveniles, engaged in birds climbed around in the same trees and bush- constructing a new nest less than 100 m from es while responding to our tapes and exhibited the pole nest, which appearedto be unoccupied. no agonistic behaviors toward one another. After One of the juveniles was seen carrying a stick they ceasedcalling, each group flew off together. ca. 30 cm in length from the pole nest to the Some groups with four or more birds included 418 KEVIN J. ZIMMER AND ANDREW WHITTAKER one or more juveniles, although others appeared homesteads and ranches, often placing their to be comprised of all adults. The subtle char- nests close to human habitations. acters that distinguish juveniles are difficult to see except at close range and in good light, and DISCUSSION thus, exact group composition was frequently Nominate cristata and unirufa differ significant- difficult to determine. In all cases of group ly in their male and female loudsongs and con- (more than two birds) territorial defense, all sequently, in their duets. None of these vocali- birds actively joined in the posturing and sing- zations are known to vary geographically within ing. Juveniles often responded by flying in and the two forms. The northern Bolivian population landing immediately next to one adult bird and of unirufa, which is nearly as geographically appeared somewhat more timid than did adults. isolated from other unirufa as are the latter from In January 1999, playback experiments at 11 cristatu, appears to be vocally identical to east- territories in Pernambuco and Bahia resulted in ern populations of unirufa. A primary challenge only one case of group territorial defense, that in applying the biological speciesconcept to al- by three birds. All other territories were defend- lopatrically distributed taxa is the need to judge ed by either pairs or lone birds. whether or not the two taxa are sufficiently dif- In contrast to our experiences with cristata, ferentiated as to prevent extensive hybridization we have never had more than two unirufa in a in the event of secondary contact. Carefully de- territory respond to tape playback. More than signed and systematically conducted playback 90% of all of our encounters with this form have experiments provide convincing evidence of involved two closely associatedbirds that were whether vocal differentiation confirmed by spec- strongly territorial. On a few occasions in Au- trographic analysis is a sufficient barrier to in- gust-October visits to the Pantanal, we have en- terbreeding. countered three-four unirufa foraging closely In our study, these experiments strongly sug- together in what we took to be family groups, gest that neither taxon of “Rufous” Cacholote but these groups were nonvocal and unrespon- recognizes nor reacts to the vocalizations of the sive to tape playback. other. The five inter-taxon responses (four of which involved only a single song or duet with- HABITAT out approachor subsequentvocalization, and the The two types of “Rufous” Cacholotes occupy other, which involved unsustained approach markedly different habitats. Nominate birds are without vocal response) are almost certainly tax- inhabitants of arid and semi-arid caatinga scrub onomically insignificant. During our observa- and woodlands (Vaurie 1980, Ridgely and Tudor tions of cacholotes, we found that many of the 1994, pers. observ.). Within this formation, the duetting species of furnariids in Mato Gross0 densest populations of cacholotes appear to be and northeastern Brazil seem to share an inter- in some of the most arid regions of Pernambuco generic response to any loud, chattering duet. and Bahia. Even there, the distribution of cach- On several occasionswe found that a duet from olotes appears somewhat clumped, with some a pair of Greater Thornbirds (Phacellodomus seemingly suitable areas unoccupied. Remark- ruber), Rufous Horneros (Fumarius rufus), or ably, areas subjectedto relatively severe grazing Pale-legged Homeros (F. Zeucopus) would result pressure by goats often have larger cacholote in an immediate vocal response from a pair of populations than more pristine areas in the same P. c. unirufa that we were watching. The hor- region (pers. observ.). neros (and to a lesser extent, the thornbirds) P. c. unirufa occupies deciduous woodland, were similarly responsive to any duets by the gallery forest edge, and seasonally flooded sa- cacholotes.Similarly, P. c. cristatu in northeast- vanna with “islands” of swampy woodland em Brazil routinely responded to noisy duets of (Ridgely and Tudor 1994, Parker et al. 1996). Band-tailed Homeros (F. $gulus) by delivering We have found it to be relatively evenly distrib- a single duet of their own. All of these species uted throughout these habitats in the northern (particularly the cacholotes and horneros) have Pantanal of Mato Grosso, Brazil, but at lower some structural similarities to their duets, pos- densities than we have found cristata in its cen- sibly evolved in response to acoustic character- ter of abundance. istics of their common environment. Interspecif- Both types of “Rufous” Cacholotes frequent ic recognition of disturbance duets from neigh- THE RUFOUS CACHOLOTE IS TWO SPECIES 419 boring speciescould be important in decreasing tutu, Vaurie (1980) made no subspecificdistinc- the risks of predation for species which have tions, and reported that “the distal half, or more, permanent pair-bonds and territories. During our of the longer crest feathers has grayish edges, playback trials, we frequently set-off duet and is brown or brownish at its center or along “chain reactions” of neighboring pairs of hor- the shaft, the base of the crest feathers being neros simply by playing back cacholote tape. rufous.” Ridgely and Tudor (1994) describe Ru- The four “weak” responsesthat we observed in fous Cacholote (without distinguishing between our trials were no stronger than the responses cristata and unirufa) as having the “crest feath- given by the same cacholotesto neighboring fur- ers ashy gray.” They go on to state that birds nariids of different genera. The one inter-taxon from Beni, Bolivia “appear paler and more cin- response (out of 46 trials) that was rated as namon and lack gray in crest.” The implication “moderate” came when a pair of unirufa ap- of this description is that all populations of Ru- proached briefly at the first playback of a cris- fous Cacholotesexcept those from Beni, Bolivia tata duet, but did not vocalize and quickly re- are gray-crested. treated. Subsequentinter-taxon playbacks to this There are two obvious sources for the con- pair produced no response. flicting literature regarding crest color. The first Duets serve multiple functions, including co- is that crest color is variable in at least some ordinated defense of territory and pair-bonding. populations of unirufa, perhaps especially in the In “Rufous” Cacholotes, as in the majority of Beni population. Although in Mato Grosso, Bra- duetting species, the same types of duets are zil we have encountered only gray-crested uni- used during bouts of territorial defense, as a pre- rufa, examination of specimens revealed that lude to copulation or nest building, and when birds from Departamentos Beni and Santa Cruz pairs reunite following visual separation (Fara- (Bolivia), and Mato Gross0 could have crests baugh 1982, pers. observ.). In sexually mono- that were rufous-brown, gray, or variegated chromatic species (most Furnariids), sexual brown-and-gray. Examination of a small sample specificity of duet contributions probably serves of specimens could easily suggestthat all birds in mate attraction as well as in identifying the from Departamento Beni were brown-crested, or sex of a territorial trespasser(Farabaugh 1982). that all other unirufa were gray-crested.The sec- A likely by-product of sexual specificity of duet ond sourceof confusion is that juveniles of nom- roles is the strengthening of reproductive isola- inate cristata are gray-crested, which could lead tion. In an habitually duetting genus such as investigators to the conclusion that all popula- Pseudoseisura, vocal differences between taxa tions of cristata and unirufa were equally vari- would seem to present major obstaclesto repro- able with respect to crest color. ductive or social compatibility. In addition to crest color, cristata and unirufa Morphological distinctions between cristata differ more subtly in overall color, and in several and unirufa are less obvious than the vocal dif- mensural characters(notably bill size). ferences, but are nonetheless significant. Crest Particularly intriguing is the evidence of dif- color is the most conspicuous distinguishing fering social structuresbetween cristata and uni- character,and one that has been confused in the rufa. Our field work demonstrates that for at literature. Hellmayr (1925) stated that unirufa least part of the year it is common for nominate was “very similar to cristata, but averaging birds to live in groups, in which there appears smaller with slenderer bill; general coloration of to be an alpha-pair with l-4 helpers. These plumage deeper cinnamon rufous, and feathers helpers not only assist in territorial defense by of pileum always more or less tipped (or edged) joining in cooperative duets, but, at least in some with ashy.” This statement leaves as ambiguous instances, actively share in nest building and the crest color of nominate cristata, particularly care of young. The presence of juveniles in in light of Hellmayr’s (1925) following comment some of the groups that we encountered suggests that “I find the western race to be separableafter that helpers are offspring from previous nestings all, although single examples are not always dis- that remain on their natal territories to assist tinguishable.” Vaurie (1980) noted only that their parents in raising younger siblings. Similar cristata was “somewhat paler and brighter than cooperative breeding systems have been docu- the birds of southwesternBrazil, northern Para- mented for a number of Neotropical taxa guay and Bolivia.” In his description of P. cris- (Brown 1978). We found juvenile P. c. cristata 420 KEVIN J. ZIMMER AND ANDREW WHITTAKER joining adults in duetting in response to play- titularly oversaturation of preferred habitat by back, and in construction of new nests. Hudson established breeders (Koenig and Pitelka 1981). (1920) reported that after fledging, juvenile Species that live in stable environments are bet- White-throated Cacholotes (P. gutturalis) re- ter candidates for the evolution of cooperative mained with their parents until the next breeding breeding systems than are species that occupy season,and that these birds were most frequent- less stable environments, where periodic re- ly found in family groups of five or six. Simi- source bottle-necks result in lower densities and larly, Nores and Nores (1994) found that juve- more vacant potential territories (Gill 1990). Our nile Brown Cacholotes (P. lophotes) in Argen- observations of unirufa are from the Pantanal tina remained in the natal territories for 5-13 region of Mato Grosso, a 150,000 km2 alluvial months, frequently joining in the defense of the area in the upper Paraguay Basin. The region parental nest, and occasionally aiding in nest has a highly seasonalclimate, with a pronounced construction or repair. Although this was consid- midwinter dry season (July-October) followed ered as a possible “step toward cooperative by a November-April rainy seasonwith l,OOO- breeding,” no care of younger siblings was re- 1,400 mm of rain (Dubs 1992). Each year the ported (Nores and Nores 1994). Brown Cach- major rivers in the region overflow their banks olotes also were reported to breed before becom- during the rainy season, flooding much of the ing adult (Nores and Nores 1994). basin and constantly altering the topography of Our observations of cristutu suggestsome dif- the landscape. Annual flooding and drying cy- ferences from the social structurereported for P. cles in this region may create sufficient resource gutturalis and P. lophotes. We did not find evi- instability to inhibit the development of coop- dence of paired juvenal-plumaged birds inde- erative breeding systemsin unirujii. By compar- pendent of adults. Also, groups defending terri- ison, the caatinga inhabited by P. c. cristata, the tories did not always include juvenal-plumaged arid lowland scrub inhabited by P. gutturalis birds. In some groups, all of the birds appeared (Parker et al., 1996) and the chaco woodland to be adult-plumaged with yellow eyes. This was and scrub occupied by P. lophotes (Ridgely and the case in the group of four birds that were Tudor 1994) would appear to be more seasonally cooperatively feeding nestlings near Lagoa stable environments. More detailed investigation Grande (Pernambuco, Brazil). Based on speci- into the reproductive biology of P. c. unirufa is mens examined at LACM, we believe that there needed to confirm the absence of cooperative is at least one transitional plumage between ju- breeding. Of particular interest would be com- venal and adult, in which birds have slight parisons of populations of unirufa from dry amounts of dusky streaking along the shafts of woodlands to those of the Pantanal. crown feathers and have some pale-fringed feathers on the throat and breast. If this is the CONCLUSIONS case, then observed groups of 4-6 apparent The two described forms of “Rufous” Cachol- adults could be composed of an alpha pair plus ote are allopatric in their distribution and occupy older sub-adults. Such birds would be difficult different habitats. P. c. cristutu are diagnosable to distinguish from adults in the field. This from all P. c. unirufu by marked differences in would suggestthat sub-adult P. c. cristata delay male and female loudsongs and duets, and adult breeding for longer than young of other cach- cristata may be separable by culmen length olotes. They also would appear to play a larger alone. Nominate birds are further generally dis- role in breeding activities by assisting in the tinguished by their overall larger size, paler feeding of nestlings. However, the general lack Tawny or Cinnamon plumage (as opposed to of group territorial responseto playback trials in Ferruginous), and crestswhich are concolor with 1999 (only 1 of 11 territories being defended by the body. Juveniles of cristatu have plumage more than two birds) may be indication that cris- characters and eye color different from adults. tutu group structure is not permanent. There is as yet no indication that juveniles of The apparent lack of cooperative breeding in unirufu exhibit such differences. Extensive tape P. c. unirufu is interesting given that some form playback experiments demonstrated that neither of this behavior is found in every other taxon in form of “Rufous” Cacholote respondsto vocal- the genus. Cooperative breeding in many species izations of the other. Cooperative breeding has has been linked with ecological constraints, par- been determined to be common in cristata, THE RUFOUS CACHOLOTE IS TWO SPECIES 421 whereas it appears to be absent in unirufa pop- DUBS, B. 1992. Birds of southwestern Brazil. Schel- ulations studied to date. The sum of these dif- lenberg Druck, Kiisnacht, Switzerland. FARABAUG~, S. M. 1982. The ecological and social sig- ferences points to cristata and unirufa as being nificance of duetting. D. 85-l 24. In D. E. Kroods- separatespecies. Thus, we submit that the com- ma and E. H. Miller [EDS.], Acoustic communi- plex consists of two species as follows: cation in birds. Vol. 2. Academic Press, New Pseudoseisuru cristuta (Spix), the Caatinga York. GILL, E B. 1990. Ornithology. W. H. Freeman, New Cacholote York. Pseudoseisura unirufa (d’orbigny and Lafres- HELLMAYR, C. E. 1925. Catalogue of birds of the naye), the Gray-crested Cacholote Americas. Field Mus. Nat. Hist. Publ. Zool. Ser. The traditional English name of “Rufous Vol. 13, part IV. Cacholote” is not appropriately applied to one HUDSON, W. H. 1920. Birds of La Plata. E. F? Dutton, New York. form over the other. Therefore, we suggest the ISLER, M. L. 1997. A sector-based ornithological geo- English name of “Caatinga Cacholote” for P. graphic information system for the Neotropics. cristata (highlighting the unique habitat in Ornithol. Monogr. 48:345-354. which it occurs) and “Gray-crested Cacholote” ISLER, M. L., I? R. ISLER, AND B. M. WHITNEY. 1997. Biogeography and systematics of the Thamnophi- for P. unirufa, which emphasizesthe best plum- lus punctutus (Thamnophilidae) complex. Orni- age character for separating most individuals of thol. Monogr. 48:355-382. the two forms. In choosing the latter name we KOENIG, W. D., AND E A. PITELKA. 1981. Ecological recognize that not all individuals of unirufa have factors and kin selection in the evolution of co- gray crests, but attempts to incorporate the geo- operative breeding in birds, p. 261-280. In R. D. Alexander and D. W. Tinkle [EDS.], Natural selec- graphic range or habitat preferencesof this form tion and social behavior: recent research and new into a name were unsatisfactory. theory. Chiron Press, New York. KRABBE, N., AND T S. SCHULENBERG.1997. Species ACKNOWLEDGMENTS limits and natural history of Scytalopustapaculos (Rhinocryptidae), with descriptions of the Ecu- Specialthanks must go to M. and I? lsler, who gen- adorian taxa, including three new species. Orni- erously donated their time and talents to providing the thol. Monogr. 48:47-88. map and spectrograms (respectively) used in this pa- KROODSMA, D. E. 1984. Songs of the Alder Flycatcher per. K. Garrett (LACM) was most helpful in coordi- (Empidonax alnorum) and Willow Flycatcher nating specimen loans from various institutions. We (Empidonau traillii) are innate. Auk 101: 13-24. also thank J. V. Remsen and S. Cardiff (LSUMZ); T KROODSMA, D. E. 1986. Design of song playback ex- Schulenberg, D. Willard, and S. Hackett (FMNH); R. periments. Auk 103:640-642. K. Panza (CM); S. L. Olson and P Angle (USNM); KROODSMA, D. E. 1989. Male Eastern Phoebes (Say- and R. Ridgely and D. Agro (ANSP) for the loan of ornis phoebe, Tyrannidae, Passeriformes) fail to specimens from their respective institutions. D. Stotz imitate songs. J. Comp. Psych. 103:227-232. (FMNH) and R. Ridgely (ANSP) also kindly provided KROODSMA, D. E., AND M. KONISHI. 1991. A suboscine information regarding additional specimens in their bird (eastern phoebe, Sayornis phoebe) develops collections. A. Schaffner assisted with the statistical normal song without auditory feedback. Anim. analysis of morphological data. G. Budney, A. Priori, Behav. 421477-487. and the rest of the staff at the Library of Natural LANYON, W. E. 1978. Revision of the Myiarchus fly- Sounds, Cornell University were of great help in al- catchers. Bull. Mus. Am. Nat. Hist. 161:429-627. lowing access to the LNS collection of recordings. T. NORES, A. I., AND M. NORES. 1994. Nest building and Schulenberg has been a constant source of advice and nesting behavior of the Brown Cacholote. Wilson stimulating conversation on all taxonomic matters, and Bull. 106:106-120. also provided assistance in obtaining certain referenc- PARKER, T A., Ill, D. E STOTZ, AND J. W. FITZPATRICK. es. Finally, we gratefully acknowledge V Emanuel for 1996. Ecological and distributional databases, p. providing us with travel opportunities that have made I 15-436. In-D. E Stotz, J. W. Fitzpatrick, T ..&. much of our work possible. Parker Ill. and D. K. Moscovits IEDS.~. Neotroni- cal birds:’ ecology and conserva;ion. >bniv. CLi- LITERATURE CITED cage Press, Chicago. PETERS, J. L. 1951. Checklist of birds of the world. BIERREGAARD,R. O., M. COHN-HAF~, AND D. E STOTZ. Vol. VII. Harvard Univ. Press, Cambridge, MA. 1997. Cryptic biodiversity: an overlooked species PYLE, P, S. N. G. HOWELL, R. F? Y~INICK, AND D. E and new subspecies of antbird (Aves: Formicari- DESANTE. 1987. Identification guide to North idae) with a revision of Cercomacra ~runnina in American passerines. Slate Creek Press, Bolinas, northeastern South America. Ornithol. Monogr. CA. 48:111-128. RIDGELY, R. S., AND G. TUDOR. 1994. The birds of BROWN, J. L. 1978. Avian communal breeding sys- South America. Vol. 2. Univ. Texas Press, Austin, tems. Annu. Rev. Ecol. Syst. 9:123-l%. TX. 422 KEVIN J. ZIMMER AND ANDREW WHITTAKER

SMITHE, E B. 1975. Naturalist’s color guide. Am. Mus. ing institutions: Carnegie Museum, Pittsburgh, Penn- Nat. Hist., New York. sylvania (CM); Field Museum of Natural History, Chi- VAURIE, C. 1980. Taxonomy and geographical distri- cago, Illinois (FMNH); Los Angeles County Museum, bution of the Furnariidae (Aves, Passeriformes). Los Angeles, California (LACM); Louisiana State Bull. Am. Mus. Nat. Hist. 166:1-357. University Museum of Natural Science, Baton Rouge, WILLIS, E. 0. 1992. Three Chamaeza antthrushes in Louisiana (LSUMZ); Academy of Natural Sciences of eastern Brazil (Formicariidae). Condor 94: 1 lo- Philadelphia, Philadelphia, Pennsylvania (ANSP); and 116. the National Museum of Natural History, Washington, ZIMMER, K. J. 1997. Species limits in Cranioleuca vul- D.C. (USNM). pina. Ornithol. Monogr. 48:849-864. cristutu (10 males, 5 females, 1 sex unknown)- BRAZIL: Bahia, Barra (LACM, 2 males, 1 female), APPENDIX 1 Rio do Peixe (CM, 1 female; FMNH, 3 males), un- Recording locations and recordists. Numbers following specified locality (USNM, 1 sex unknown). Ceara each name represent the number of recordings from (USNM, 1 male). Paraiba, Coremas (LACM, 2 males, the recordist it each site. 1 female), Santa Luisa (USNM, 1 female), Souza cristuta- BRAZIL: Boa Nova. Bahia CT. A. Parker. (USNM, 1 male, 1 female). Piaui, Ibiapaba (FMNH, 1 1); Lencois, Bahia (K. J. Zimmer, 6); Uricuri, Pernam: male). buco (K. J. Zimmer, 1); Lagoa Grande, Pernambuco (A. Whittaker, 2; K. J. Zimmer, 33). unirufa (5 males, 5 females, 1 sex unknown)-BO- unirufa-BOLIVIA: 8 km south of Trinidad, De- LIVIA: Departamento Beni, 6 km southeast of Trini- partamento Beni (T A. Parker. 1). BRAZIL: Trans- dad (LSUMZ, 1 female), 10 km southwest of Trinidad Pantaneira Hwy. between Pixaii and Porto Jofre, (LSUMZ, 1 male). Departamento Santa Cruz, Buena- Mato Gross0 (K. J. Zimmer, 22). vista (FMNH, 1 male), Puerto Suarez (CM, 1 female, 1 sex unknown), unspecified locality (LSUMZ, 1 fe- APPENDIX 2 male). BRAZIL: Mato Grosso, Descalvados (ANSP, 1 List of localities and lending institutions for specimens male, 1 female; FMNH, 1 male, 1 female), Fazenda examined. All specimens were from one of the follow- Santa Rosa (LSUMZ, 1 male).