Much of my compositional work in recent years has focused on animal communication. I am strongly committed to understanding bioacoustic communication in living animals through personal observation, ongoing review of the scientific literature, and consultation with biologists and animal care technicians. I use this research in a very direct way to shape, organize, and inspire my recent compositions.
The Duke Lemur Center is the largest lemur reserve in the world outside of Madagascar.
For more information on this wonderful, unique research facility and the charismatic animals who live there, visit the DLC website.
I am currently working on a large-scale 8-channel composition entitled The Audible Phylogeny of Lemurs. The catalyst for this work was the article "Phylogeny of the Lemuridae Revisited: Evidence from Communication Signals" by Joseph Macdeonia and Kathrin Stanger [Folia Primatol 1994; 63: 1-43]. The authors produced a lemur phylogeny from communication evidence and compared it to then-current DNA-based phylogenies. In comparing the Macedonia-Stanger cladistic analysis to a more recent molecular phylogeny [Yoder, Anne D. and Yang, Ziheng. "Divergence dates for Malagasy lemurs estimated from multiple gene loci: geological and evolutionary context," Molecular Ecology 2004; 13: 757-773], I was struck by how well the earlier communication-based analysis held up. Excellent news for an artist interested in animal communication. Although I still take creative license in my new piece (otherwise I wouldn't be doing my job as an artist!), my approach in The Audible Phylogeny of Lemurs is much more documentarian in nature than in the comparatively science-fictional Echoes of Lemurs. I am attempting to demonstrate species relationships by drawing out similarities in the acoustic structures of calls and in contexts of emission, so that the evolutionary story encoded in the vocalizations of modern animals can be heard and appreciated by the listener.
In the summer of 2008, I conducted a pair of preliminary playback studies at the DLC as part of my research for The Audible Phylogeny of Lemurs.
Playback study: Alarm Responsiveness
A study of alarm responsiveness across multiple genera. The species in the study were: Propithecus coquereli, Varecia variegata rubra, Eulemur rufus, Hapalemur griseus, and Lemur catta.
[Note: I observe taxonomic ranks found in the Conservation International Tropical Field Guide Series: Lemurs of Madagascar, Second Edition. Conservation International has raised many lemurs from subspecies to full species status. Some of these changes of status are not observed by the Duke Lemur Center. I have no authority to comment on this matter; I have simply chosen, for consistency, to adopt Conservation International's ranks, since I already use the field guide for other information.]
In this study, I wanted to investigate whether alarm responsiveness follows phylogeny in at least a broad sense. From observations of interspecific alarm contagiousness at the DLC, I had guessed that more closely related (to be precise, cladistically related) species would be more responsive to one another's calls. The results were promising, but not conclusive, and there are numerous confounding factors, especially habituation and housing proximity. It is difficult to establish whether the animals have simply adapted to one another's calls in captivity, or whether the responses really are hard-wired (or based on acoustic similarity of calls, itself related to phylogeny).
Even so, some of the responses were encouraging. Counting only actual alarm vocalizations as positive responses:
In the first video (below left), Varecia respondedafter a tense pausewith aerial alarms to playback of Propithecus aerial alarms. These animals appear in the same clade in both my source phylogenies. Though they are believed to have diverged 40 million years ago, their aerial alarms are strikingly similar. In the second video (below right), Propithecus responded to playback of their own aerial alarm, followed by an aerial alarm response from a group of nearby Varecia.
Male (above) and female (below) crowned lemurs (E. Cornonatus)
Hapalemur griseus [ more on H. griseus ] responded to the terrestrial alarms of their sister species Lemur catta with their own terrestrial alarm. In the following recording, the first sound is the L. catta alarm playback (yapping); this is followed by a marked increase in arousal clicks and click series in H. griseus; finally, H. griseus emits three terrestrial alarms.
Eulemur rufus responded to the pulsed alarm of Eulemur rubriventer (a species in the same genus) with an alarm of their own. The first sound on the following video is an E. rubriventer alarm playback, followed by the more rapid pulsed alarm response of E. rufus. Note the excited tail-waving and scanning of the ground and sky.
In the course of the study, there were no unexpected alarm responses. Only animals in either the same genus (E. rufus and E. rubriventer) or clade (Propithecus and Varecia;H. griseus and L. catta) were responsive to one another's calls.
Still, the results were not entirely satisfying. For one thing, some species are more responsive in general. Propithecus, in particular, were very cooperative test subjects, responding with appropriate groans, "shee-faks," sky- and ground-scanning, and tail-waving to almost every call they heard. However, only their own aerial alarm elicited an alarm vocalization in response.
On the other hand, L. catta didn't respond to anything, not even (as we had fully hoped and expected) to H. griseus.
The animals were also not completely consistent in responding to their own alarms.
I am considering a follow-up to this preliminary study with a more thorough study, possibly with the help a biologist or psychologist. By the end of the study, I had identified a number of factors that need to be taken into account in playback situations:
• Playbacks should occur only after provisioning; the animals appeared distracted and preoccupied before they had eaten. Pre-provisioning playbacks were noticeably less likely to elicit responses.
• Animals that have just vocalized or are currently vocalizing when approached will be too distracted. On one occasion we approached a group of Varecia, already involved in an alarm chorus, hoping we might influence the chorus. They ignored us entirely.
• Plenty of time should be taken between individual playbacks and between playback sessions. I am concerned that a succession of alarm calls, without enough time in between, may elicit an alarm response simply by way of its cumulative effect on the animals' arousal levels.
• No more than two sessions total per group. We found that even the highly responsive Propithecus had lost interest by the third round of playbacks, even with several days between sessions.
• Larger speakers for more convincing acoustic simulation of alarm calls. This is especially important for animals like Varecia that have extremely loud calls.
• Recordings used for playback should be at the highest possible resolution. (In one case, I had only a low resolution recording of a particular call, and gave it try without result. When I obtained a new recording and swapped it out, the animals were far more responsive.) Low resolution audio (less than 16-bit/44.1KHz) may not be easily recognized by the animals.
• Recordings used for playback should be close-miked for ideal proximity simulation. Distant-miked recordings were generally ineffective. The animals at the DLC hear so many calls around the facility in the course of a day that distant calls may not necessitate a response.
Playback study: Contact calls in ring-tails
I performed a second playback study involving contact calls in L. catta. I wanted to evaluate robustness of call recognition under minor audio processing. Basically, I wanted to ask: Can ring-tails recognize their own contact calls, even when the calls have been digitally altered?
The first question is, do they react to straight playback of contact calls? The answer is: Yes, they seem to, but the responses are highly variable and difficult to interpret.
As to whether they recognize digitally altered calls: They reacted strongly to high intensity calls with the first or second partial removed and to time-stretched (2X and 3X) calls in general; they reacted less strongly to pitch-shifted (up or down a perfect fourth) calls.
But because we couldn't establish normative responses to straight playback, it was difficult to interpret responses to digitally altered calls.
Nevertheless, the responses were fascinating to watch and there were some aspects of the study that have led me to consider conducting a more focused study on call recognition at a later time. In particular, we noticed a trend to suggest that playback caller ID might have an impact on the response:
The caller used for the playbacks was a female named Sosiphanes, who had been ousted from the Natural Habitat 4 group the previous summer. She was also known to the NHE 2 group through boundary encounters. Interestingly, the group to which she had previously belonged reacted very strongly to her calls (group movement, agonism, vocalization); the group that knew her through boundary encounters reacted somewhat strongly; the NHE 6 group, who do not know her at all, had a very minimal reaction; and the small group to which she currently belongs (with the male Aracus and their daughter Lilah) had almost no reaction. Possibly worth a follow-up.
Below are some examples of the animals' reactions to straight vs. processed playbacks.
The first video (below left) uses a high-arousal "late-high wail." The first playback is the unaltered call. It elicits some contact calls in response, followed by a flurry of chirps which initiate rapid group movement. This was a surprising response and the only such full group relocation we saw in response to any playback (including alarms). The second playback is the same late-high wail, but this time the first harmonic has been digitally removed. The response was comparatively muted, though it did elicit a similar set of contact calls including multiple chirps.
The second video (below right) uses a slightly lower intensity late-high wail. In this case, it is the unaltered call that receives a minimal response. The straight playback is followed by a time-stretched playback of the same call, which elicits some excited movement, searching for the sound source, and an agonistic exchange.
Recording and Observation
Since December 2006, I have made regular visits to the Duke Lemur Center to observe and record the animals. I am collecting a library of recordings for use in my own compositional work and to share with other researchers. My main reference in this endeavor is the work of Joseph Macedonia, who authored a series of seminal articles on lemur vocalization in the 80s and 90s. In addition to making my own recordings, I am in the process of digitally transferring many of Macedonia's original recordings.
I am still in the process of organizing and documenting the many hours of recordings I have made at the DLC. I will soon post a selection of representative soundfiles on this page, so check back for updates. In the meantime, see the Wikipedia page Calls of the Ring-Tailed Lemur for a definitive list of L. catta calls and their emissions of context, featuring recordings by Macedonia and myself.
Research at the Brookfield Zoo
In the summer of 2008, I spent two weeks at the Brookfield Zoo on a research protocol approved by the zoo's Biological Research and Steering Committee. I recorded and observed the Tropic World primates off exhibit, including several non-exhibit animals. The species to which I had access were:
Black handed spider monkeys - Ateles geoffroyi and Ateles paniscus
Orangutan - Pongo pygmaeus, 3 hybrid and 5 Bornean
Pygmy slow loris - Nycticebus pygmaeus
Mandrill - Mandrillus sphinx
Sooty mangabey - Cercocebus atys, and sooty mangabey-mandrill hybrid
Red-capped mangabey - Cercocebus torquatus
Black and white colobus - Colobus guereza
Sykes guenon - Cercopithecus mitis albogularis
Guinea baboon - Papio hamadryous papio
Western lowland gorilla - Gorilla gorilla gorilla
One of the great pleasures of working at the Brookfield Zoo was recording Benny and Indah, a male-female pair of white-cheeked gibbons (Hylobates concolor leucogenys). They vocalized quite reliably each morning 6-8 AM before going on exhibit. Their spectacular duet is deafeningly loud in person and is nearly as impressive visually as it is sonically, as the climax of Indah's part is accompanied by a frenetic, body-shaking display.
Here is a complete duet (over nine minutes). It is initiated by Indah ("whooo"); Benny first responds (high two-note call) after Indah's first climax.
Lemurs gather at the Cold Shelter Box during the heat of the day. Left to right: Propithecus coquereli, 2 Lemur catta, Eulemur rufus
Benny, male white-cheeked gibbon, on exhibit at the Brookfield Zoo. Photo courtesy of the Brookfield Zoo.
Indah, female white-cheeked gibbon, on exhibit at the Brookfield Zoo. Photo courtesy of the Brookfield Zoo.
Male (left) and female (right) golden lion tamarins at the Brookfield Zoo. Photos courtesy of the Brookfield Zoo.
Occasionally, in the course of animal observation at zoos and research facilities, one is charmed by quirkily talkative individuals. Evita, a female spider monkey (Ateles geoffroyi), was such an individual. Displaying none of the shyness of her companions, Evita followed me as I recorded and supplied me with a steady stream of "winnies," a spider monkey greeting call.
A spider monkey, hanging by its prehensile tail, on exhibit at the Brookfield Zoo. Photo courtesy of the Brookfield Zoo.
Quite possibly the most complex vocalizations I encountered at the Brookfield Zoo were those of the callimicos (Callimico goeldii). While lacking the fanfare of their nearby GLT neighbors (above), callimico vocalizations possess a much greater degree of precision and specificity, which only gradually became apparent to me during the course of recording and observation. The following edited example gives an idea of the nuance and subtlety of callimico calls and represents only a handful of the many distinct, context-specific calls in their vocal repertoire.
Research at the Wisconsin National Primate Research Center
In the summer of 2007, I spent a week at the Wisconsin National Primate Research Center on a research protocol approved by the IACUC (Institutional Animal Care and Use Committee) of the University of Wisconsin-Madison, investigating vocalization of rhesus monkeys, long-tailed macaques, and common marmosets. In addition, Charles Snowdon generously allowed me to record the colony of cotton-top tamarins maintained by the university's psychology department. I refer the reader to his website for more information.
Recording Alexander, while two lemurs on the ground are grooming
Callimicos (Callimico goeldii), also called Goeldi's monkey or Goeldi's marmoset, on exhibit at the Brookfield Zoo. Photo courtesy of the Brookfield Zoo.
Much of my research has involved the production of instrumental music modeled both acoustically and formally on animal communication. Rather than continue the project of physically modifying traditional instruments or devising extended playing techniques for them, I have begun to develop an instrumentarium for the specific purpose of emulating animal vocalization, stridulation, or "mechanical" communication. In many cases, the instruments are mechanically simple, but with a combination of precise playing techniques, contact miking and amplification, and real-time digital signal processing, the sound world they produce is convincing. For more information on the instruments and their associated playing techniques, download the Swampoid electronic score.
The Doorstop Box
A lidded wooden box with two doorstops screwed into the top, amplified by a high-quality large-diaphragm condenser microphone placed directly in front of the opening at about two inches for maximum proximity effect. The doorstops are usually bowed at various points along shaft with variable bow speed and pressure to produce sounds that range from raspy growls to shrieks. While bowing the doorstops, the performer opens and closes the lid, producing a strong impression of vowel formants shaped over bowed "vocalizations."
A closed wooden box with rubber bands (two small, one large, high quality rubber) threaded through the top, amplified by a contact microphone attaced directly to the surface of the box. Four springs are suspended across the inside of the box to produce sympathetic vibrations (a pitched reverb) when the surface of the box is stroked with a rubber ball or the rubber bands are stroked with wet fingers. The sounds range from glottal pulse trains to mid-range shrieks.
A metal plate, mounted on wood, with guitar pegs affixed to the sides, supporting a pair of guitar strings across the plate. The guitar strings are attached directly to metal brackets, which are in turn screwed directly to the plate. This conducts the string vibrations to the plate, which has a contact microphone mounted in the center. The contact mic picks up a surprising amount of detail this way. The strings are usually bowed, and the instrument is used to model insect stridulation and frog-like croaks. When mounted with proper shock absorption, the string plate also boasts an impressively long bow bounce.
A set of five cheap hearing aids, mounted on a table, with volumes set at maximum. The headphones are planted in the fingers of a pair of gloves. The performer can then move the gloves over the hearing aids to produce feedback, which, if properly performed, resembles birdsong. The gloves are also used to emulate cricket stridulation.
I had to trick an aggressive male red-capped mangabey (Cercocebus torquatus) into vocalizing for me by planting my equipment in front of him, out of sight, and walking to the other end of the enclosure, also out of sight. He could still hear that I was in the enclosure, and that is, I believe, what elicited these grunting and barking vocalizations. The calls are so loud that it took a few tries to get the levels right, since I couldn't monitor the equipment during recording. It took multiple sessions at two hours or more each to capture these few minutes of audio, but clearly it was worth it.
Male Red-Capped Mangabey (Cercocebus torquatus) on exhibit at the Brookfield Zoo. Photos courtesy of the Brookfield Zoo.
Rhesus Macaques (Macaca mulatta)
I recorded a considerable number of the 1000 or so rhesus monkeys housed at the WNPRC, including adults, juveniles, and infants. The following soundfiles represent only a tiny sample of the many unique voices I captured.
Some of the most vocal animals were mothers with infants. When approached with the microphone, the mothers reacted with fear grimacing, lip smacking, and a variety of submissive vocalizations. In the second example, the infant vocalizes (the lower voice) along with the mother.
Female rhesus monkey with infant. This is a public domain image.
Juveniles (post-weaning) were also extremely vocal. In the first example below, a roomful of excited juveniles (male and female) react to the care technicians' preparations for room cleaning. In the second example, a pair of female juveniles, housed together, react to a novel stimulus, i.e. the microphone.
Some of my favorite recordings of rhesus monkeys were those of the adult males' "happy" sounds. In the following recording, adult males react to a playful exchange with a care technician involving distribution of treats. The monkeys were too apprehensive to emit these vocalizations with me in the room, therefore, we hid the microphone in the bottom of a food container and sent in the care technician without me.
Young rhesus monkey. This is a public domain image.
Cynomolgus Macaques (Macaca fascicularis)
My time with the 20 or so cynomolgus macaques was somewhat limited. Here also it was sometimes necessary to hide both myself and my equipment from the animals in order to get certain vocalizations.
The first recording below is an adult female in estrus who protected a toy as if it were an infant. The second recording is a six-year-old male reacting to the microphone and to me, both novel stimuli, as we had just met.
Cynomolgous (or crab-eating, or long-tailed) macaque (Macaca fascicularis). Photo by John J. Mosesso. This is a public domain image.
Cynomolgous (Macaca fascicularis). Photo by John J. Mosesso. This is a public domain image.
I recorded a considerable number of the 200 or so common marmosets housed at the WNPRC. I discovered in the course of recording them that they would nearly cease to vocalize when I ceased to move. This meant constantly slowly moving up and down enclosures, and when forced to stop, keeping an arm swinging in circles, which is a challenging way to record delicate high frequency vocalizations like these.
One of the highlights, in the first recording below, is an infant, upside down on its mother's stomach, vocalizing as its mother crawled overhead. The infant is the lower voice, and it is surrounded by a variety of high-frequency adult vocalizations.
The second recording is one of many lucky close-miked recordings I was able to capture of these not-so-shy monkeys. It is a recording of a long call.
Another highlight was recording the zoo's off-exhibit golden lion tamarins (Leontopithecus rosalia), a male-female pair who reacted with a magnificent burst of territorial and alarm vocalizations when I approached their enclosure. These displays decreased in intensity over the course of a given recording session, during which I approached them multiple times. But the next day's session would invariably begin with another high-intensity display. Here is an excerpt:
Snapshot of common marmoset from the Callicam. Image courtesy of Primate Info Net.