Bioacoustics of Isleria hauxwelli

Study Populations. Isleria ( Bravo et al. 2012 ) is currently considered to comprise two species ( Zimmer & Isler 2003 ; Dickinson & Christidis 2014 ; Clements et al. 2023 ; Fig. 9 ): I. guttata ( Vieillot, 1825 ) and I. hauxwelli ( Sclater, 1857 ) . The phylogeny indicated that I. guttata differs from I. hauxwelli by a sequence divergence of 7.8%. No geographically related vocal or morphological differences have been found within I. guttata , and genetic differences among sampled locations of I. guttata were minimal and not geographically structured. Consequently, the entire clade was identified as study population guttata ( Fig. 10 ). Isleria hauxwelli is currently considered to consist of three subspecies ( Zimmer & Isler 2003 ; Fig. 9 ): I. h. hauxwelli ; I. h. suffusa ( Zimmer, 1932 ); and I. h. hellmayri ( Snethlage, 1906 ). The phylogeny distinguished four main clades within I. hauxwelli , all separated by at least 2.5% sequence divergence. Geographic distributions of these differ from those of the currently accepted subspecies in several respects ( Fig. 10 ). The first main clade covers the range of suffusa and also a portion of the range of hauxwelli in Peru south of the Rio Marañón and west of the Rio Ucayali , and thus includes type localities of both I. h. hauxwelli and I. h. suffusa . This clade was identified as study population hauxwelli- W. Two main clades cover the remaining geographic region of what is currently considered I. h. hauxwelli , south of the Amazon River to the southern extent of Amazonia. The first of these, identified as study population hauxwelli- C, extends from east of the Rio Ucayali and the Andes to the left bank of the lower Rio Madeira and across the middle and upper Madeira to the left bank of the Tapajós and Teles Pires rivers. To the east, the second of these, identified as study population hauxwelli- E, extends from the right banks of the lower Rio Madeira and Tapajós/Teles Pires rivers to the Rio Xingu. Its range includes the type locality of I. h. clarior ( Zimmer, 1932 ) which had been merged with I. h. hauxwelli ( Zimmer & Isler 2003 ) . The final main clade was geographically consistent with I. h. hellmayri (right bank of the Rio Xingu, east) and was identified as study population hellmayri . Study populations hauxwelli- C and hauxwelli- E encompassed subclades that were also considered in the analysis. Two subclades in hauxwelli- C separated by the middle and upper Rio Madeira with a sequence divergence of 0.6% were identified as study populations hauxwelli- C-W and hauxwelli- C-E. Two subclades in hauxwelli- E separated by the Rio Tapajós with a sequence divergence of 0.9% were identified as study populations hauxwelli-E-N and hauxwelli- E-S ( Fig. 10 ). Finally, the phylogeny placed the main clades into two superclades separated by 7.8% sequence divergence. These superclades correspond to the species I. guttata and I. hauxwelli . Vocalizations . Vocalizations were categorized into four types : Song; Alarm Call; Hook Call; and Rattle . The Song was a consistently patterned multiple-note vocalization of unusually long duration for a thamnophilid that was irregularly repeated at often long intervals and thus was especially difficult to record in an unsolicited context. The Alarm Call was a multiple-element note of brief duration that typically was repeated, sometimes in long series. The Hook Call was a continuous note that went up, down, and then up again in frequency. The Rattle consisted of multiple abrupt notes given in a rapid, often downslurred series, the notes of which were barely distinguishable. Songs. Songs of all populations consisted of a long series of similar notes (10–75 notes) that gradually became shorter in duration and wider in frequency band width. Within that general pattern, however, the shape of the notes of Songs of guttata ( Fig. 11A ) differed from those of all populations of hauxwelli . In addition, the overall band width (difference between the highest and lowest frequency) differed significantly between populations of guttata (1099–1926 Hz, ξ = 1459, SD = 256, n = 12) and hauxwelli (1971–4340 Hz, ξ = 3183, SD = 557; n = 55). We found no significant geographic variation among Songs of guttata . Within hauxwelli populations, notes of hauxwelli- W were frequency-modulated ( Fig. 9 , single note enlarged in Fig. 11D ), a characteristic lacking in notes of other populations exemplified by the Song of hauxwelli- C ( Fig. 11C , single note enlarged in Fig. 11E ). Non-overlapping ranges of frequency measurements found among hauxwelli populations did not meet our significance test but indicate the need for further frequency analysis when additional samples are obtained. Calls . The most commonly recorded call of Isleria taxa was termed an Alarm Call because this behavioral context was often noted by field workers. Alarm Calls differed diagnosably between guttata and hauxwelli populations. Alarm Calls of guttata ( Fig. 12B ) were long (5‒7 elements, ξ = 5.8; duration 0.088 ‒0.144 sec, ξ =, 0.111; n = 10), whereas those of hauxwelli populations ( Fig. 12A ) were shorter (2‒4 elements, ξ = 2.6; duration 0.032 ‒0.068 sec, ξ =, 0.043; n = 10). To the human ear, guttata Alarm Calls were harsh and raspy (n = 34), whereas hauxwelli Alarm Calls were burry (n = 71). Alarm Calls were often repeated or delivered in long series. Differences of Rattles ( Fig. 12C and 12D ) and Hook Calls ( Fig. 12E ) among populations were not significant, although the Hook Call was not recorded among our sample of 54 recordings of hauxwelli E–S. Vocalization summary . Characteristics of Songs (note shape and frequency band width) and Alarm Calls (number of elements and duration) distinguished I. guttata from all I. hauxwelli populations. The only vocal character distinguishing populations of I. hauxwelli was the frequency modulation of notes in the Song of hauxwelli- W. Morphology . Comparison of 29 specimens of I. guttata with 142 specimens of I. hauxwelli confirmed that I. guttata and I. hauxwelli (both sexes) differ diagnostically in underpart coloration and in wing edgings and terminal spots as described in the literature (e.g., Zimmer & Isler 2003 ). No plumage differences were found in samples of I. guttata . In comparing plumages of populations of I. hauxwelli , using both visual and spectrophotometry comparisons within the large samples ( 72 males , 70 females ), the only diagnosable difference was the lack of a white interscapular patch distinguishing both sexes of hellmayri . Variation in darkness of male plumage of I. hauxwelli appeared to be clinal, and individuals of adjoining populations were indistinguishable. Zimmer (1932) noted that darker specimens had smaller spots on the wings and tail, but considerable overlap in spot size was found among adjoining populations. In our comparisons the darkest males originated in the northwest ( hauxwelli- W), and the palest males were from the Huanchaca region of northeast Bolivia ( hauxwelli- C). Although not as pale as adjoining populations, males of hellmayri were also relatively pale, and half of the available male specimens (4 of 8) had whitish edges to feathers in the underparts giving a scalloped appearance, but these were present on a few individuals from other populations. These could be signs of immaturity, or transition to definitive plumage. In blind tests, darker hellmayri females were almost perfectly distinguishable from lighter females of hauxwelli-E (both subclades). The darkest females were those of hellmayri and hauxwelli- W, located at the geographic ends of the overall distribution. However, in neither blind arrays nor spectrophotometry measurements could females of hellmayri be separated diagnostically from the paler females of hauxwelli- E, nor could the darker females of hauxwelli- W be separated from the paler females of hauxwelli- C. The palest female specimens of hauxwelli- C originated in the Huanchaca region of northeastern Bolivia , as did the palest males.

Remarks. Although the study populations included in this subspecies exhibited variation in vocal and morphological characters, non-overlapping vocal differences among populations did not meet our criteria for diagnosability. With the assembly and analysis of additional recordings and specimens, some study populations are likely to warrant taxonomic recognition. This supposition is supported by the branching patterns and mtDNA distances in the phylogeographic results. Further vocal analysis should emphasize Song frequency. Frequency measurements suggested that there might be diagnostic differences among hauxwelli populations. Frequency was typically higher east than west of the Tapajós/Teles Pires. Differences were especially noticeable across the Teles Pires at Alta Floresta and across the lower Tapajós. However, Songs of female-plumaged birds appear to be higherpitched than those of males, and such analyses should control for sex. This subspecies differs by 2.5–3.0% in mtDNA from other subspecies of I. hauxwelli , but the genetic data indicate that hauxwelli- C is more closely related to hauxwelli- W than to hauxwelli- E. Nuclear data would provide further insight on the discrepancy between the song and mtDNA data.