Introduction

Descriptions of specific unidentified avian vocalizations are to our knowledge nonexistent in ornithological literature. Such reports would seem uncalled for regarding diurnal vocalizations, because with enough effort one should be able to follow the sound and identify its source. But an avian nocturnal flight call is different because it is a transient call in the night, typically from a migrating bird that one cannot simply chase down and identify. While early ornithologists in North America were puzzled regarding the identity of many night flight calls, recent advances in spectrographic analysis tools now facilitate identification. Yet even with modern recording and analysis tools, a confounding challenge for determining the identity of an unknown avian nocturnal flight call is that one needs to know all possible candidates for the sound in the region. Documenting a bird call that matches is not enough; one must also ascertain that other species do not give similar calls that need discrimination.

Between 1986-2001, W. Evans and M. O’Brien endeavored to identify Parulidae and Emberizidae nocturnal flight calls in eastern North America. Members of both families typically migrate at night and most are now known to give short (less than 250 ms), high-pitched (5-11 kHz) calls while in nocturnal flight (^{Evans and O’Brien 2002}). Since species identification could not be visually confirmed, the act of presuming identity proceeded inferentially over many years. Progress was primarily made by documenting flight vocalizations from species visually confirmed during the day. The possibility of similar calls from other species was evaluated and distilled over time as call characteristics of more and more species coalesced. The fact that there is no underlying difference between North American parulid and emberizid night flight calls increased the pool of potential contenders for unidentified callers. Both families have species with calls that ascend in pitch; both have species with calls that descend in pitch; both have a variety of frequency modulated calls. The only notable difference now known is that, in eastern North America, emberizids tend to have nocturnal flight calls of longer duration than parulids (^{Evans and O’Brien 2002}).

Other types of songbirds that make annual migratory flights across eastern North America are known to give short nocturnal flight calls in the same high-pitched frequency range as parulids and emberizids. This includes the cardinalids in the genus Passerina: P. caerulea, P. cyanea, and P. ciris. In addition, Regulus satrapa, Turdus migratorius, Anthus rubescens, and Bombycilla cedrorum also give high-pitched flight calls during nocturnal migration. The remaining annual migrant songbirds in eastern North America known to vocalize in nocturnal flight generally utter lower-pitched calls, most below 5 kHz (^{Evans and O’Brien, 2002}).

Since the release of the Flight Calls of Migratory Birds: Eastern North American Landbirds CD-ROM (^{Evans and O’Brien, 2002}), additional flight call identification progress has not been reported for eastern North America. ^{Lanzone et al. (2009)} and ^{Farnsworth (2011)} revealed the first spectrographic portrayal of parulid flight calls in western United States. Emberizid flight calls in the West remain only partially described. Of note is that ^{Watson et al. (2011)} detected Spinus pinus calling in nocturnal migration during an eruption year. While the latter is not call identification progress, it indicates that Fringillidae calls need to be considered when sorting out possible identifications for unidentified night flight calls.

In 2012, William Evans (WRE) was promoting new night flight call monitoring stations in North America and Manuel Grosselet (MG), Georgia Ruiz Michael (GRM) and colleagues were operating a banding station in the state of Veracruz, Mexico, a few kilometers east of Minatitlán. MG had become interested in establishing a night flight call monitoring station to compare with their banding results. Through our mutual collaboration, a night flight call monitoring station began operation at the airport in Minatitlán in late August 2012.

We describe a distinctive nocturnal flight call from an unidentified species we first recorded in Minatitlán during the fall migration period of 2012. Thirty three calls from the unidentified caller were recorded with an automatic acoustic transient detector optimized for extracting short calls of night migrating songbirds in the 6-10 kHz range. The time-frequency characteristics of the call stood out to WRE as a call he was unfamiliar with in more than 25 years studying nocturnal flight calls across eastern North America. The unidentified call was circulated to others studying avian night flight calls in North America as well as several experts on bird sounds in Mexico, and to date remains unidentified.

Methods

Results

We monitored nocturnal flight calls on 72 nights from August 28 through December 3, 2012. Ten nights were missed during this period due to heavy rain, power outage, or equipment problems. In addition, the twelve nights from November 14-25 and 4 nights from November 28 through December 1 were missed. From the results of the Tseep-r detections between October 16 and December 3, 354 calls, most apparently from warblers and sparrows were found by browsing spectrograms with GlassOFire software (Supplement 1: http://oldbird.org/pubs/S1-Evans_et_al.zip). Amidst these data, the unidentified call first began to stand out in the latter half of October 2012 when multiple occurrences were documented.

Call description

The unidentified call consists of two contemporaneous, rising frequency components, one higher than the other. What distinguishes the call from other “double ups” known to occur across eastern North America is the combination of the call’s broad frequency expanse, the relatively wide frequency gap between its component tones, and its longer overall duration. Figure 1 illustrates some of the clearest examples.

Figure 1 Spectrographic examples of the unidentified nocturnal flight call. Calls displayed occurred on nine different nights. Those that occurred on the same night were more than one hour apart. 

Of 16 calls loud enough to reveal the beginning of the call, the mean initial frequency is 6330 ± 865 Hz. The high frequency at the end of these calls is not measurable for 10 calls because it ascends above 11 kHz, the highest frequency we can measure with our 22050 Hz sampling rate. Six of the 16 calls exhibit high frequency points below 11 kHz with the lowest around 9575 Hz. Of 10 calls where the beginning and end are well defined, the mean duration is 86 ± 10 ms. The longest of the 10 durations is 102 ms and the shortest is 67 ms. Of five calls where the frequency bandwidth of the whole call is measurable, the mean bandwidth is 4890 ± 545 Hz. Of these five, the greatest bandwidth is 5930 Hz and the least is 4505 Hz. The mean maximum frequency between the individual rising tone components for the 12 calls shown in Figure 1 is 1805 ± 350 Hz. The largest frequency difference between contemporaneous points on the two tone components of these 12 calls is 2380 Hz and the least is 1155 Hz.

Seasonal occurrence

We documented the unidentified call type 32 times between October 16 and December 3 -- a call recorded on September 19 (nearly a month earlier than the main sequence of detections) brings the total to 33. The unidentified call comprised 9% of all avian nocturnal flight calls in the 6-10 kHz range detected over the Minatitlán Airport terminal from October 16 through December 3 (Figure 2).

Figure 2 Nightly totals of emberizid and parulid flight calls (gray bars) and comparative numbers of the unidentified call (black bars). 

Hourly occurrence

We detected the unidentified call in all 10 monitoring hours of the night, with a notable concentration (52%) between 01:00-04:00 h (Figure 3). Calling activity of emberizids and parulids also showed an increase after 01:00 h. Of the 33 total calls, 27 occurred more than 15 minutes apart from one another. There were three cases where calls occurred within a minute of another. From November 5-11, when the majority of calling from the unidentified species was detected, apparent sunset was at 18:43 h (+/- 1 min) CDT and apparent sunrise was at 07:21 h (+/- 1 min) CDT.

Figure 3 Percent calling by hour of the night for the unidentified caller, emberizids, and parulids from October 16 - December 3. 

Discussion

In Minatitlán, Veracruz during fall 2012, we recorded a type of nocturnal flight call with parameters unrecognized from our extensive nocturnal flight call investigations in the USA. While it is possible that more than one bird species is involved, there is enough apparent commonality among our sample to suspect that all the calls are from one species. It is not out of question that the unidentified call might belong to a local nonmigratory species, but its limited seasonal period of detection and its record of occurrence on nights, and in hours of the night, when other neotropical migrants were moving suggest it is from a species that was in southbound migration.

WRE has not detected the unidentified call in 25 years of spectrographic analysis of nocturnal flight calling from across the eastern USA. This includes large migration season datasets from Florida, south Texas (lower Rio Grande Valley), northwestern Louisiana, eastern Missouri, north-central Nebraska, southeastern Minnesota, northeastern Wisconsin, northern Indiana, northern Ohio, multiple sites across New York, north-central Pennsylvania, Massachusetts, and eastern Vermont. In addition, the call has not been noted by WRE in the western USA in full migration season datasets from south-central Arizona (Ajo, AZ) and southeastern Arizona (Nogales, AZ). Our consultation with other night flight call researchers in North America has not led to any information about the call and we are aware of no other cases where it has been documented. Furthermore, our consultation with several Mexican bird sound experts has not led to any clues to the species’ identity.

Circumstantial reasoning suggests the call may be from a neotropical migrant species that breeds somewhere in the western half of North America north of our Minatitlán monitoring site. A review of ^{Howell and Webb (1995)} reveals only a small number of species that meet this criterion and that are known to migrate through the Minatitlán region. Ten of these are kingbirds and flycatchers in the family Tyrannidae. After extensive study of night flight calls in eastern North America, ^{Evans and O’Brien (2002)} reached a conclusion suspected previously by others that members of the tyrannids of eastern North America do not often call in nocturnal migration. Furthermore, except for the distinctive “ka-zeer” of the Eastern Kingbird (Tyrannus tyrannus), the flight calls they have been documented or suspected to give are below 6 kHz. So, it seems unlikely that the unidentified caller would be one of the 10 species of western tyrannids that migrate through the Minatitlán region.

There are four other neotropical migrant songbirds that meet the previously noted criterion: Icterus cucullatus, Vireo flavoviridis, Cyanerpes cyaneus and Peucaea botteri. The flight call of I. cucullatus is a distinctive lower pitched “veek”, which WRE is familiar with and is widely known. Regarding V. flavoviridis, after extensive study of night flight calls in the eastern USA, ^{Evans and O’Brien (2002)} came to the conclusion that if Vireonids do call in nocturnal migration, the calls are uncommon and below 6 kHz. So V. flavoviridis is not a likely candidate. The northern population of C. cyaneus reportedly migrates southward out of regions to the north of Minatitlán. The species has not been banded or observed by MG, GRM and their team at banding sites in the Minatitlán region, and little is reported on its migration behavior. ^{Schaldach (2003)} indicated the species migrates diurnally, but this does not rule out that it may also make nocturnal movements. No calls of this species in the Macaulay Library at the Cornell Lab of Ornithology (ML) and Xeno-canto (XC) audio archives matched our unidentified call.

Regarding the secretive P. botteri, we found little concrete information to ascertain the migratory status of the three subspecies that range north of Minatitlán (arizonae, botterii, and petenica). The evidence that some northern populations are migratory seems to have originated from the paucity of detections in northern locales during the winter months. That may have changed in southeastern Arizona where P. b. arizonae now appears to winter in numbers, perhaps due to the incursion of non-native grass species (^{Schmierer 2013}). A night flight call monitoring station that operated near Nogales, AZ from April 17 through June 19 and August 7 through November 18, 2015 logged 20,000 small passerine night flight calls; however, our unidentified call was not among them though P. b. arizonae was a common nester in proximity. No calls of this species in the ML and XC audio archives matched our unidentified call.

The banding data involved 25 species that have flight calls unknown to us, but all these species are present and banded year round in Minatitlán and none are listed as migratory in ^{Howell and Webb (1995)}. Two of these species are now known to undertake migrations, at least among their northern populations, Crotophaga sulcirostris (^{Bowen 2002}) and Turdus grayi (MG personal observation). No calls of either species in the ML and XC archives matched our unidentified call.

These latter two species bring up the fact that it is difficult to gain migration information in Mexico for Mexican breeding species in which a portion of the population may migrate relatively short distances at night to a region where the species has permanent residents. Changes in local populations due to influxes of migrants may not be easily noticed for species that tend to wander about or which are secretive. And there is little public information from collisions with tall man-made structures in Mexico to give us clues to nocturnal movements, so it is possible there are migration patterns yet to be discovered.

Though we obtained a small sample of calls, the apparent pattern of more calling in the latter hours of the night for the unidentified was shared by migrant emberizids and parulids. This is consistent with the idea that the unidentified call is from a migrant bird and not nocturnal flight behavior from a local nonmigrant species. This late-night calling pattern might correspond with increased density of low-altitude migration associated with a diminishing sea breeze in the coastal region of eastern Mexico as the night progresses. Overall, the quantity of emberizid and parulid calling detected from our quiet Minatitlán recording site was relatively low compared to similar acoustic monitoring stations across inland eastern USA. Yet, banding data by ^{Rodríguez Ramos et al. (2015)} indicate that large numbers of many species of neotropical migrants cross the region. Perhaps the main flow of southbound migrants occurs at altitudes too high to be detected with our acoustic monitoring system, and the increased calling toward the end of the night was from individuals transitioning down from nocturnal migration.

Besides the work represented in ^{Evans and O’Brien 2002}, the only previous publication tackling the identification of a cryptic nocturnal flight call is a paper by WRE speculating and providing supporting evidence that a set of nocturnal flight calls recorded from eastern Florida in 1989 and 1991 were from Catharus bicknelli (^{Evans 1994}).Of course, in some cases, the night flight call of a species is so distinctive that identification is largely not in question, for example the flight call of Spiza americana. ^{Larkin et al. (2002)} developed a simple algorithm incorporating parameters of this call’s bandwidth, modulation rate, and duration to automatically detect it with a high rate of efficiency. This enabled a 200-km east-west running transect of 10 flight call monitoring stations targeting S. americana calls to be operated in near real-time across the Rio Grande Valley of Texas.

While acoustic monitoring does not provide precise information on the numbers of a species passing, monitoring calling activity can be useful for elucidating periods and geographic scope of migration, as well as potentially serving as an index for abundance. Therein lays the utility in learning the identity of our unknown nocturnal flight call.