Content uploaded by William D. Helenbrook
Author content
All content in this area was uploaded by William D. Helenbrook on Jul 08, 2019
Content may be subject to copyright.
1
W. D. Helenbrook1,2, N. A. Linck3, M. A. Pardo4, and J. A. Suarez2
1
Spatial variation in black-headed night monkey (Aotus nigriceps) vocalizations
2
1State University of New York College of Environmental Science and Forestry (SUNY-ESF), 1 Forestry Drive,
3
Syracuse, New York, United States; 2Tropical Conservation Fund, 760 Parkside Trl NW, Marietta, Georgia, USA;
4
3University of Minnesota Twin-Cities, St. Paul, Minnesota, USA; 4Cornell University, 215 Tower Road, Ithaca,
5
New York, USA
6
Corresponding author: William Helenbrook, 760 Parkside Trl NW, Marietta, Georgia 30064
7
wdhelenb@syr.edu, Phone 239-470-1200, Fax 315- 470-6934; ORCID ID: https://orcid.org/0000-0002-2706-3525
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Acknowledgements
23
We would like to thank the Amazonian Conservation Association (ACA), Villa Carmen Biological Station, and
24
Manu Learning Center (CREES) staff for hosting us, clearing trails, and providing valuable insight into location and
25
behavior of groups. We are indebted to students and staff from the School for Field Studies who assisted with data
26
collection and logistics. We carried out data collection in accordance with the legal requirements of Peru, and with
27
permission of the Amazon Conservation Association and CREES.
28
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
2
Abstract
29
Quantitative acoustic analysis has been used to decipher individual differences, population structure, and taxonomic
30
diversity in numerous primate species. We previously described three distinct call types in wild Aotus nigriceps, and
31
now assess acoustic differences in two of these call types between social groups and spatially distinct populations.
32
Acoustic parameters for both analyzed call types exhibited significant variability between groups. Similarly,
33
geographically distant field sites were acoustically distinct from one another. Several groups also used a variation of
34
a common call: a triplet Ch Ch instead of a duplicate. Other groups made use of ultrasonic frequencies which have
35
not previously been reported in Aotus. Our results suggest that Aotus nigriceps exhibits substantial acoustic
36
variability across sites that could potentially be useful for taxonomic classification, although additional
37
geographically distant populations still need to be sampled. The possibility of individual signatures also exists and
38
will require recording vocalizations from known individuals.
39
Key words: vocalizations, Aotus nigriceps, night monkey, Peru, acoustics
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
3
Introduction
57
Primates use vocalizations to communicate about the presence of predators (Zuberbühler 2002; Schel et al.
58
2009), location of food sources (Slocombe and Zuberbühler 2005), nesting behavior, travel intentions and group
59
cohesion (Boinski 1996), territorial defense (Raemaekers and Raemaekers 1985; Cowlishaw, 1992), mate
60
assessment and pair bonding (Cowlishaw 1996; Geissmann and Orgeldinger 2000). Primate acoustic signals may
61
also be used for kin recognition and can convey information on age, sex, body size, and rank (Salmi and
62
Hammerschmidt 2014). The diversity of these vocal signals makes acoustics a convenient tool to explore differences
63
among primates at various taxonomic levels.
64
Despite the importance of acoustic communication in primates, evidence of population level differences in
65
primate vocalizations are relatively limited (Green 1975; Maeda and Masataka, 1987; Mitani et al., 1992; Fischer et
66
al. 1998; Mitani et al. 1999; Delgrado 2007; Wich et al. 2008; de la Torre and Snowdon 2009; Wich et al. 2012).
67
Acoustic variation between populations can exist for any of several reasons: divergence through cultural drift in
68
species that learn their vocalizations (i.e. inaccurate copying transmitted vertically or horizontally), genetic drift
69
following reproductive isolation, or local adaptation in response to sexual selection, habitat transmission properties,
70
predation pressure, or social selection pressures (Yoktan et al. 2011).
71
Most evidence suggests that non-human primates are not vocal learners; however, several recent studies
72
have found that primates can learn slight modifications to their vocalizations (Watson et al. 2015; Takahashi et al.
73
2017). Examining patterns of geographic variation in call structure can provide some evidence for or against the
74
presence of vocal learning; for example, if there is a sharp acoustic divide between two spatially contiguous areas
75
that show no evidence of genetic divergence (i.e. vocal dialects), this often suggests the presence of vocal learning.
76
Spatial variation in vocal characteristics within a species could be a valuable tool for addressing a variety of
77
ecological questions. For example, primate calls can be used to distinguish species, populations, groups, and
78
individuals (Table 1). If individuals could be distinguished from one another solely using quantitative acoustic
79
analysis, population size could be estimated by combining acoustic analysis and line transect surveys (Terry et al.
80
2005; Marques et al. 2013; Kalan et al. 2015). Moreover, variation in specific call characteristics could be used to
81
infer group membership, or be used for taxonomic classification, supplementing morphometric or genetic data.
82
Night monkeys, Aotus spp., are a useful model for investigating patterns of acoustic variation because
83
nocturnal and forest-dwelling species tend to rely heavily on vocalizations to communicate with one another. We
84
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
4
have previously reported on the vocal repertoire of wild Aotus nigriceps, describing three calls: the Squeak, Ch Ch,
85
and Long Trill (Helenbrook et al. 2018). In this study we focus on quantitatively comparing acoustic variation of
86
two of these calls between groups and distant populations.
87
88
Methods
89
Eleven Aotus nigriceps groups were sampled (Fig.1): eight at the Villa Carmen Biological Station in
90
Pilcopata, Peru (12°53'39"S, 71°24'16"W), and three at CREES - the Manu Learning Center, on the edge of Manu
91
National Park (12°47′22″S 71°23′32″W). The two field sites are separated by a low mountain range (~1143m) and
92
are just over 10 km apart at their nearest borders. Villa Carmen has a long history of development, ecotourism and
93
agriculture. The groups sampled near the station lived in secondary forest, often dominated by bamboo or cane,
94
whereas groups sampled at CREES inhabited recovering clear-cut to primary rainforest where bamboo and cane
95
were largely absent. Research groups of 3-8 observers went into the field from 5:30-7:30am and 5:30-7:30pm for a
96
total of 28 days at Villa Carmen and nine days at CREES to collect acoustic data, times when A. nigriceps groups
97
are known to be active near their nesting sites. Several recordings also took place during the day as part of a separate
98
behavioral study.
99
A Zoom H1 Handy Recorder was coupled with a RØDE NTG-2 condenser shotgun microphone and shoe
100
shockmount on a micro boompole at a distance varying from 2-25m. Digital recordings were made at 48 kHz
101
sampling frequency with 16 or 24-bit amplitude resolution. Acoustic analysis was conducted using Raven Pro 1.5
102
sound analysis software (Cornell Lab of Ornithology Bioacoustics Research Program, Ithaca, New York). Calls
103
were digitized and measured spectrographically (DFT size 512, time resolution 3.1 ms, Hann window with 50%
104
overlap). Twenty-four acoustic parameters were measured for each call (Table 2).
105
Inter-group differences were analyzed using non-parametric Kruskal-Wallis tests coupled with post-hoc
106
multiple comparisons of mean ranks tests with a Bonferroni correction. A Mann-Whitney U test was used to assess
107
site differences between Villa Carmen and CREES. Stepwise discriminant function analysis was used to explore
108
acoustic parameters that could be used to classify social groups. For model selection, a stepwise forward method
109
was used (statistic, Wilk’s Lamda) with the criteria Fto enter=3.84 and Fto remove=2.71, and a tolerance level of <0.01
110
(STATISTICA). This process was repeated for both call types separately. Variables that failed a tolerance test where
111
there was an almost exact linear relationship with other variables, did not enter the analysis. We used a 10-fold cross
112
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
5
validation in which 90% of the calls were randomly chosen to calculate discriminant functions, while 10% was
113
excluded for testing. Differences between observed and expected frequencies of duplicate versus triplicate Ch Ch
114
calls was measured using Fisher’s Exact Test. All recordings were conducted non-invasively, minimized impact on
115
behavior, and avoided excessive disturbance, and were therefore deemed exempt from the Institutional Animal Care
116
and Use Committee approval. All applicable international, national, and/or institutional guidelines for the care and
117
use of animals were followed.
118
119
Results
120
Three vocalizations have been described in wild Aotus nigriceps populations: Squeak, Ch Ch, and Trill
121
(Helenbrook et al. 2018). In this study, we analyzed acoustic variability for the two most common calls, the Squeak
122
(N=1302) and the Ch Ch (N=556; Fig. 2). For Squeaks we only measured the dominant harmonic since it was
123
consistently found across all sampled groups. The Trill was not used because of its rarity across most groups. At
124
least ten calls were analzyed from each of seven night monkey groups, ranging in size from 2-5 individuals
125
(Mean=3.7). The dominant harmonic of the Squeak ranged from a mean minimum frequency of 1591 Hz (Range:
126
74-3055 Hz; SD 470) to a mean maximum frequency of 2742 Hz (Range: 2010-4443; SD 252). The Ch Ch call
127
ranged from a mean minimum frequency of 1698 Hz (Range: 44-9092; SD 1182) to a mean maximum frequency of
128
11636 Hz (Range: 3109-23726 Hz; SD 2538).
129
Acoustic measurements varied significantly between groups (Fig. 3; Table 3). Differentiation between two
130
or more monkey groups was found for all forty-eight independent vocal characteristics (2 call types x 24
131
measurements, p<0.001). Discriminant function analysis distinguished among groups for both call types: Squeak
132
(Wilks' Lambda=0.06, F(60,6706)=81.98, p<0.0000) and Ch Ch (Wilks' Lambda=0.06, F(48,2592)=42.58,
133
p<0.0000) (Table 4 and Fig. 4). Cumulative significant functions were able to explain 87.4% of variance among
134
groups using only Squeak calls, and 87.8% of the variance among groups using only Ch Ch calls. Classification
135
accuracy was similar for both the Squeak (87.4%) and the Ch Ch (76.4%). Duration (90%) and energy parameters
136
for Squeak and Ch Ch, respectively, provided the greatest discriminatory power at the group level (Table 4).
137
Twelve acoustic parameters were significantly different between Villa Carmen and CREES biological field
138
stations (Table 6). Discriminant function analysis identified seven Squeak parameters that significantly distinguished
139
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
6
locations: low and high frequency, bandwidth, duration (90%), delta time, IQR duration, and max time – the first of
140
which contributed the greatest discriminatory power; and eight Ch Ch parameters significantly distinguished
141
locations: low and high frequencies, bandwidth, energy, peak frequency, Q1 frequency, frequency (5%) and center
142
time– the last of which contributed the greatest discriminatory power. The two sites were found to be significantly
143
different based on Squeak (Wilks’ Lambda=0.77, F(13,1281)=29.09; p<0.0000) and Ch Ch (Wilks’ Lambda=0.40,
144
F(13,526)=61.77; p<0.0000). Classification accuracy was 93.8% for Squeak (5 out of 13 CREES measurements and
145
116 out of 116 at Villa Carmen), and 100.0% for Ch Ch. Cumulative significant functions were able to account for
146
47.7% of variance between locations using the Squeak, and 77.7% using the Ch Ch call.
147
Other differences were observed between groups as well. The Ch Ch was predominately found in a series
148
of two (“in duplicate”) (88.3% of cases); however, four groups also produced calls in triplicate (i.e. Ch Ch Ch). Out
149
of 556 total Ch Ch calls, 65 were in triplicate (11.7%), with 2.8% in T2A, 1.3% in A, 50.4% in B, and 2.8% in E.
150
The distribution of triplicate calls across groups differed significantly from even distribution across groups, with
151
Group B exhibiting nearly four times as many triplicates as expected (p=0.0000). In addition, two groups were
152
observed using ultrasonic frequencies as part of the Ch Ch call (>20kHz): group C (N=3) at Villa Carmen and T2A
153
(N=1) at CREES.
154
155
Discussion
156
The majority of acoustic parameters for both calls differed significantly between groups and geographic
157
locations, though single acoustic parameters alone were not sufficient to predict group membership. Variance of
158
acoustic parameters overlapped in nearby groups, making absolute classification difficult. However, there was a
159
consistent pattern whereby calls from the same groups and population tended to cluster together based on similar
160
acoustic measurements. Population level classification was more accurate, largely driven by acoustic parameters of
161
the Ch Ch call. Quantitative analysis of acoustic traits may therefore be useful in elucidating group and population
162
level differences and may provide useful insight into the underlying phylogenetic relationships between groups,
163
populations and potentially species of Aotus. However, additional recordings are needed both at the group and
164
population levels, preferably with more distant populations included.
165
We were unable to investigate individual acoustic variability because of our inability to pair calls to
166
specific individuals in a complex environment at night. Based on various other primate studies it is likely that
167
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
7
individuals can be differentiated based on vocal signatures (Table 1). However, confirmation of vocal individuality
168
will require either analysis in captivity or pairing video and audio recordings in wild nesting groups. If recordings
169
can be attributed to specific individuals, then acoustic analysis could be used to establish whether individual
170
conspecifics vary predictably in their vocalizations. Establishing the ability to vocally differentiate individuals
171
would be particularly useful for a nocturnal species such as the black-headed night monkey, allowing researchers to
172
study group composition solely based on vocal recordings.
173
Aside from differences in acoustic parameters, two other acoustic differences were discovered among
174
groups. First, a triplet Ch Ch call was found in recordings from groups T2A, A, B, E. Though relatively rare within
175
the sampled populations (11.7% of cases), over half of these cases were found in Group B. The other groups at Villa
176
Carmen that used the triplet call are likely of the same population since they are isolated on all but one side and in
177
relative proximity to group B (<1300m at furthest extent). The prevalence of the triplet call in Group B suggests that
178
this is not an aberration but rather a consistent modification of a common call. The fact that the triplet call only
179
occurred in certain groups could reflect any number of possibilities including increased prevalence of a particular
180
behavioral context, or a vocal innovation (genetic or learned). Alternatively, it is possible that the presence of both
181
duplicate and triplet Ch Ch calls is the ancestral state and the absence of the triplet call is derived. Either way,
182
additional sampling of nearby groups – coupled with underlying population genetics analysis - would confirm
183
whether this is a relatively unique acoustic irregularity which is independent of underlying population structure, or
184
whether this call variation routinely arises and is widespread. Likewise, being able to obtain calls specific to
185
individuals through video and audio pairing in nests would allow us to decipher whether all individuals within a
186
particular group use the triplet call.
187
It is uncertain whether the use of ultrasonic frequencies in night monkeys is rare or whether this is a
188
common response to environmental pressures such as inter-species competition for lower frequencies or predator
189
avoidance. Of course, other nocturnal primates (i.e., Tarsius, Galago, Microcebus, Nycticebus) and some diurnal
190
neotropical primates (i.e., Callithrix and Cebuella) produce calls containing ultrasonic frequencies, though only the
191
tarsiers produce calls entirely within the ultrasonic range, with the other species always producing dominant
192
frequencies in the human audible range (Ramsier et al. 2012). In several species, the use of ultrasound appears to be
193
context specific, often in the presence of predators, including humans (e.g. Rahlfs and Fichtel 2010; Gursky-Doyen
194
2013).
195
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
8
Aotus currently consists of eleven described species based on both phenotypic and genotypic evidence.
196
Night monkey taxonomy has been revised considerably based on differences in karyotypes, morphology, molecular
197
sequencing, malaria sensitivity, immunological responses, and geographic isolation (Menezes et al. 2010). Despite
198
this, few specimens from any one study have come from Aotus nigriceps despite this species having one of the
199
largest ranges of any Aotus species. Moreover, the current taxonomic classification lumps A. nigriceps populations
200
from areas with considerably different elevations and from areas separated by significant river systems. Thus, the
201
possibility remains that further evolutionary and conservation management units may exist. Considering the distinct
202
differences in call types previously described between Aotus species and the use of quantitative acoustic sampling to
203
differentiate many other primate species, we anticipate that further analysis would prove useful in differentiating
204
population-level or species-level taxonomy.
205
Finally, Aotus nigriceps likely produce more than the three described call types since captive Aotus species
206
have exhibited larger vocal repertoires. In captive situations it is easier to record night monkeys at close distances
207
and calls can be induced in different situations, which could facilitate observation of a wider variety of call types.
208
We anticipate that with continued sampling these additional call types could also be recorded in the wild.
209
210
Conflict of Interest
211
The authors declare that they have no conflict of interest.
212
213
References
214
Ambrose L (2003) Three acoustic forms of Allen's galagos (Primates; Galagonidae) in the Central African region.
215
Primates 44:25-39
216
Boinski S (1996) Vocal coordination of troop movement in squirrel monkeys (Saimiri oerstedi and S. sciureus) and
217
white-faced capuchins (Cebus capucinus). Adaptive radiations of Neotropical primates 251-269
218
Braune P, Schmidt S, Zimmermann E (2008) Acoustic divergence in the communication of cryptic species of
219
nocturnal primates (Microcebus ssp.). BMC biology 6:1-10.
220
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
9
Clink DJ, Bernard H, Crofoot MC, Marshall AJ (2017) Investigating individual vocal signatures and small-scale
221
patterns of geographic variation in female Bornean gibbon (Hylobates muelleri) great calls. International Journal
222
of Primatology 38:656-671
223
Clink DJ, Crofoot MC, Marshall AJ (2018) Application of a semi-automated vocal fingerprinting approach to
224
monitor Bornean gibbon females in an experimentally fragmented landscape in Sabah, Malaysia. Bioacoustics 1-
225
17
226
Cowlishaw G (1992) Song function in gibbons. Behaviour 121:131-153
227
Cowlishaw G (1996) Sexual selection and information content in gibbon song bouts. Ethology 102:272-284.
228
https://doi.org/10.1111/j.1439-0310.1996.tb01125.x
229
de la Torre S, Snowdon CT (2009) Dialects in pygmy marmosets? Population variation in call structure. American
230
Journal of Primatology 71: 333–342
231
Delgado RA (2007) Geographic variation in the long calls of male orangutans (Pongo spp). Ethology 113:487–498
232
Fischer J, Hammerschmidt K, Todt D (1998) Local variation in Barbary macaque shrill barks. Animal Behaviour
233
56:623–629
234
Garbino GS, Martins-Junior AM (2018) Phenotypic evolution in marmoset and tamarin monkeys (Cebidae,
235
Callitrichinae) and a revised genus-level classification. Molecular Phylogenetics and Evolution 118:156-171.
236
https://doi.org/10.1016/j.ympev.2017.10.002
237
Geissmann T, Orgeldinger M (2000) The relationship between duet songs and pair bonds in siamangs, Hylobates
238
syndactylus. Animal Behaviour 60:805-809. https://doi.org/10.1006/anbe.2000.1540
239
Green S (1975) Dialects in Japanese monkeys: vocal learning and cultural transmission of locale‐specific vocal
240
behavior? Zeitschrift für Tierpsychologie 38:304-314
241
Gursky-Doyen S (2013) Acoustic characterization of ultrasonic vocalizations by a nocturnal primate Tarsius
242
syrichta. Primates 54:293-299
243
Hafen T, Neveu H, Rumpler Y, Wilden I, Zimmermann E (1998) Acoustically dimorphic advertisement calls
244
separate morphologically and genetically homogenous populations of the grey mouse lemur (Microcebus
245
murinus). Folia Primatologica 69:342-356
246
Helenbrook W, Preston L, Link N, Quirk M, Suarez J (2018) Bioacoustics of the black-headed night monkey, Aotus
247
nigriceps. Neotropical Primates 24:29-33
248
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
10
Kalan A, Mundry R, Wagner OJ, Heinicke S, Boesch C, Kühl HS (2015) Towards the automated detection and
249
occupancy estimation of primates using passive acoustic monitoring. Ecological Indicators 54:217-226.
250
https://doi.org/10.1016/j.ecolind.2015.02.023
251
Macedonia JM, Taylor LL (1985) Subspecific divergence in a loud call of the ruffed lemur (Varecia
252
variegata). American Journal of Primatology 9:295-304. https://doi.org/10.1002/ajp.1350090406
253
Maeda T, Masataka N (1987) Locale‐specific vocal behaviour of the tamarin (Saguinus I. labiatus). Ethology 75:25-
254
30
255
Marques TA, Thomas L, Martin SW, Mellinger DK, Ward JA, Moretti DJ, Harris D, Tyack PL (2013) Estimating
256
animal population density using passive acoustics. Biological Reviews 88:287-309.
257
https://doi.org/10.1111/brv.12001
258
Masters, JC (1991) Loud calls of Galago crassicaudatus and G. garnettii and their relation to habitat structure.
259
Primates 32:153-167
260
Menezes AN, Bonvicino CR, Seuánez HN (2010) Identification, classification and evolution of owl monkeys
261
(Aotus, Illiger 1811). BMC Evolutionary Biology 10:248. https://doi.org/10.1186/1471-2148-10-248
262
Mitani JC, Hasegawa T, Gros‐Louis J, Marler P, Byrne R (1992) Dialects in wild chimpanzees? American Journal
263
of Primatology 27:233-243
264
Mitani JC, Hunley KL, Murdoch ME (1999) Geographic variation in the calls of wild chimpanzees: a reassessment.
265
American Journal of Primatology 47:133-151
266
Nietsch A (1999) Duet vocalizations among different populations of Sulawesi tarsiers. International Journal of
267
Primatology 20:567-583
268
Raemaekers JJ, Raemaekers PM (1985) Field playback of loud calls to gibbons (Hylobates lar): territorial, sex-
269
specific and species-specific responses. Animal Behaviour 33:481-493. https://doi.org/10.1016/S0003-
270
3472(85)80071-3
271
Rahlfs M, Fichtel, C (2010) Anti‐predator behaviour in a nocturnal primate, the grey mouse lemur (Microcebus
272
murinus). Ethology 116:429-439. https://doi.org/10.1111/j.1439-0310.2010.01756.x
273
Ramsier MA, Cunningham AJ, Moritz GL, Finneran JJ, Williams CV, Ong PS, Gursky-Doyen SL, Dominy NJ
274
(2012) Primate communication in the pure ultrasound. Biol Lett 8:508-511. http:doi.org/10.1098/rsbl.2011.1149
275
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
11
Salmi R, Hammerschmidt K (2014) Individual distinctiveness in call types of wild western female gorillas. PLoS
276
ONE 9:1-13. https://doi.org/10.1371/journal.pone.0101940
277
Schel A, Tranquilli S, Zuberbühler K (2009) The alarm call system of two species of black-and-white colobus
278
monkeys (Colobus polykomos and Colobus guereza). Journal of Comparative Psychology 123:136-150
279
Slocombe K, Zuberbühler K (2005) Functionally referential communication in a chimpanzee. Current Biology
280
15:1779-1784. https://doi.org/10.1016/j.cub.2005.08.068
281
Takahashi DY, Liao DA, Ghazanfar AA (2017) Vocal learning via social reinforcement by infant marmoset
282
monkeys. Current Biology 27:1844-1852
283
Terry AM, Peake TM, McGregor PK (2005) The role of vocal individuality in conservation. Frontiers in Zoology
284
2:10-26. https://doi.org/10.1186/1742-9994-2-10
285
Watson SK, Townsend SW, Schel AM et al (2015) Vocal learning in the functionally referential food grunts of
286
chimpanzees. Current Biology 25:495-499
287
Wich SA, Schel AM, De Vries H (2008) Geographic variation in Thomas langur (Presbytis thomasi) loud calls.
288
American Journal of Primatology 70:566-574
289
Wich SA, Krützen M, Lameira AR et al (2012) Call cultures in orang-utans? PLoS One 7:e36180.
290
https://doi.org/10.1371/journal.pone.0036180
291
Yoktan K, Geffen E, Ilany A, Yom-Tov Y, Naor A, Leader N (2011) Vocal dialect and genetic subdivisions along a
292
geographic gradient in the orange-tufted sunbird. Behavioral Ecology and Sociobiology 65:1389-1402
293
Zimmermann E, Bearder SK, Doyle GA, Andersson AB (1988) Variations in vocal patterns of Senegal and South
294
African lesser bushbabies and their implications for taxonomic relationships. Folia Primatologica 51:87-105
295
Zimmermann E (1990) Differentiation of vocalizations in bushbabies (Galaginae, Prosimiae, Primates) and the
296
significance for assessing phylogenetic relationships. Journal of Zoological Systematics and Evolutionary
297
Research 28:217-239. https://doi.org/10.1111/j.1439-0469.1990.tb00377.x
298
Zuberbühler K (2002) A syntactic rule in forest monkey communication. Animal Behaviour 63:293-299.
299
https://doi.org/10.1006/anbe.2001.1914
300
301
302
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
12
Fig. 1 Field research conducted at Villa Carmen Biological Station (eight black-headed night monkey groups) and
303
CREES – Manu Learning Centre (three groups), in southeastern Peru. A mountain ridge separates the two field
304
stations (10.1km at nearest points)
305
306
307
308
Fig. 2 Spectrogram representing both calls quantitatively analyzed in this study: the Squeak (left) and the Ch Ch
309
(right). Both calls have been previously described (Helenbrook et al. 2018). In this example, a duplicate Ch Ch is
310
depicted. The upper range of the spectrogram is faint, partially a result of minimizing backgound noise in the 7-8
311
kHz from insects
312
313
314
315
316
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
13
Fig 3 Boxplot plate of four descriptive acoustic measurements (e.g., duration, maximum frequency, minimum
317
frequency, and bandwidth) for the Squeak and Ch Ch across all groups. Median represented by solid box, box plot is
318
25-75% of call variability, and whiskers are minimum and maximum which do not signify significance, rather
319
distribution of values is depicted. Note that inter quartile ranges are depicted instead of standard error or deviation
320
because of the non-parametric nature. T2A, TSH, and T9A groups are from CREES while A-H are from Villa
321
Carmen. Significant differences illustrated in Table 3
322
323
324
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
14
325
Fig 4 Canonical score scatterplot. The biplot axes are the first two canonical variables. These define the two
326
dimensions that provide maximum separation among groups. We used a 10-fold cross validation in which 90% of
327
the calls were randomly chosen to calculate discriminant functions. Here we present the results of the 10% excluded
328
for testing
329
330
331
332
333
334
335
336
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
15
Table 1 Evidence of primate acoustic variation at various taxonomic levels.
337
Primate species (N)
Individual
Group
Population
Subspecies
Species
Call
description
Study
Cebidae,
Callitrichinae (28)
-
-
-
-
Yes
Long call
Garbino 2018
Galago spp. (8)
-
-
-
-
Yes
Loud call
Zimmermann
1990
Galago alleni
-
-
Yes
Loud call
Ambrose 2003
Galago
crassicaudatus, G.
garnettii
-
-
-
-
Yes
Loud call
Masters 1991
Galago
senegalensis, G.
moholi
-
-
-
-
Yes
14 calls
Zimmermann
1988
Gorilla gorilla
Yes
-
-
-
-
8 calls
Salmi et al.
2014
Hylobates muelleri
Yes
-
No
-
-
Great call
Clink et al. 2017
Hylobates muelleri
Yes
-
Yes
-
-
Great call
Clink et al. 2018
Lepilemur spp. (10)
-
-
Yes
-
-
Loud call
Méndez-
Cárdenas et al.
2008
Microcebus spp. (3)
-
-
-
-
Yes
Whistle, Purr,
Chitter
Hending et al.
2017
Microcebus spp. (3)
-
-
-
-
Yes
Advertisement
call
Braune et al.
2008
Microcebus
murinus
-
-
Yes
-
-
Trill
Hafen et al.
1998
Tarsius spp. (4)
-
-
-
-
Yes
Loud call
(Duet call)
Nietsch 1999
Varecia variegata
-
-
-
Yes
-
Loud call
Macedonia and
Taylor 1985
338
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
16
Table 2 Name and description of acoustic parameters measured. Not all parameters were included in discriminant
339
function analysis because of redundancy.
340
Parameters
Description
Low frequency (Hz) The lower frequency bound of the selection
341
High frequency (Hz) The upper frequency bound of the selection
342
Bandwidth 90% The difference between the 5% and 95% frequencies
343
Energy (dB) The total energy within the selection bounds
344
Dur90% The difference between 5% and 95% times
345
Delta frequency (Hz) The difference between the upper and lower frequency limits of the selection
346
Peak frequency The frequency at which max power occurs within the selection
347
Delta time (s) The difference between the begin and end time for the selection
348
Center frequency (Hz) The frequency that divides the selection into two frequency intervals of equal energy
349
Q1 frequency (Hz) The frequency that divides the selection into two frequency intervals containing 25% and
350
75% of the energy in the selection
351
Q3 frequency (Hz) The frequency that divides the selection into two frequency intervals containing 75% and
352
25% of the energy in the selection
353
Max power The maximum power in the selection.
354
Frequency 5% The frequency that divides the selection into two frequency intervals containing 5% and
355
95% of the energy in the selection
356
Frequency 95% The frequency that divides the selection into two frequency intervals containing 95% and
357
5% of the energy in the selection
358
Center time The point in time at which the selection is divided into two time intervals of equal energy
359
Q1 time The point in time that divides the selection into two time intervals containing 25% and
360
75% of the energy in the selection
361
Q3 time The point in time that divides the selection into two time intervals containing 75% and
362
25% of the energy in the selection
363
IQR duration (s) The difference between the 1st and 3rd quartile times
364
Time 5% The point in time that divides the selection into two time intervals containing 5% and 95%
365
of the energy in the selection
366
Time 95% The point in time that divides the selection into two time intervals containing 95% and 5%
367
of the energy in the selection
368
Max time The first time in the selection at which a spectrogram point with power equal to max
369
power occurs
370
371
Quantitative acoustic characteristics analyzed (Raven).
372
373
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
17
Table 3 All Aotus nigriceps vocal measurements were found to have significant (p<0.001) inter-group differences.
374
Post-hoc analysis was conducted using multiple comparisons of mean ranks with Bonferroni adjustment. All
375
associations significant at p<0.05.
376
377
Call Measurement Group Significant Associations (p value)
378
Ch Ch
379
Dur90% A C(0.00); D(0.04)
380
B C(0.00); D(0.01)
381
C E(0.00); F(0.01); T2A(0.00)
382
High frequency A C(0.00); D(0.02); T2A(0.00)
383
B C(0.00); D(0.04); T2A(0.00)
384
C E(0.00); F(0.00); H(0.01)
385
D E(0.01); F(0.01); T9A(0.05)
386
E T2A(0.00)
387
F T2A(0.00)
388
Low frequency A B(0.00); D(0.03); T2A(0.00)
389
B C(0.00); D(0.00); E(0.00); T2A(0.00)
390
C T2A(0.02)
391
Bandwidth 90% A B(0.00); E(0.00)
392
B C(0.00); E(0.00); F(0.00); H(0.00); T2A(0.00); TSH(0.00)
393
E T2A(0.00)
394
TSH T9A(0.04)
395
396
Squeak
397
Dur90% A B(0.00); E(0.01); TSH(0.00)
398
B C(0.02); E(0.00); T2A(0.00); TSH(0.00)
399
C E(0.00); TSH(0.00)
400
E H(0.00); T2A(0.00); TSH(0.00); T9A(0.05)
401
H TSH(0.00)
402
T2A TSH(0.00)
403
High frequency A C(0.03); E(0.00); H(0.00); TSH(0.02)
404
B C(0.00); E(0.00); H(0.00); TSH(0.00)
405
C E(0.00); H(0.00); T2A (0.00)
406
D TSH(0.04)
407
E H(0.00); T2A(0.00); TSH(0.00)
408
H T2A(0.00); TSH(0.00)
409
Low frequency A B(0.00); E(0.00)
410
B C(0.00); D(0.01); E(0.00); H(0.00); T2A(0.00); TSH(0.00); T9A(0.00)
411
C T2A(0.00)
412
E T2A(0.00)
413
H T2A(0.00)
414
Bandwidth 90% A B(0.00); E(0.00); H(0.00); TSH(0.01)
415
B C(0.00); E(0.00); H(0.00); T2A(0.00)
416
C E(0.00); H(0.00); T2A(0.00)
417
E T2A(0.00); TSH(0.00)
418
H T2A(0.01); TSH(0.00)
419
420
421
422
423
424
425
426
427
428
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
18
Table 4 Stepwise discriminant function analysis of all non-redundant measured acoustic parameters.
429
430
Call Measurement Wilks’ Lambda Partial Lambda p value
431
Squeak
432
Bandwidth 90% 0.07 0.85 0.00
433
Energy 0.06 0.92 0.00
434
Dur90% 0.07 0.82 0.00
435
Frequency 95% 0.06 0.88 0.00
436
Max power 0.06 0.89 0.00
437
IQR duration 0.06 0.96 0.00
438
Center frequency 0.06 0.98 0.00
439
Q1 frequency 0.06 0.93 0.00
440
IQR bandwidth 0.06 0.90 0.00
441
Max time 0.06 0.88 0.00
442
443
Ch Ch
444
Energy 0.17 0.33 0.00
445
Dur90% 0.06 0.92 0.00
446
Bandwidth 90% 0.05 0.94 0.00
447
IQR duration 0.06 0.96 0.00
448
Frequency 95% 0.06 0.89 0.00
449
Center frequency 0.09 0.65 0.00
450
Delta frequency 0.07 0.77 0.00
451
Q1 frequency 0.07 0.79 0.00
452
453
454
455
456
Table 5 Discriminant function analysis correct classification utilizing 10-fold cross validation at group level.
457
458
Squeak alone
459
Group
% Correct
A
B
C
E
H
T2A
TSH
A
66.7
6
0
3
0
0
0
0
B
75.0
4
12
0
0
0
0
0
C
78.6
0
2
11
1
0
0
0
E
89.9
4
1
0
62
1
1
0
H
75.0
0
0
0
2
6
0
0
T2A
44.4
3
0
0
1
0
4
1
TSH
75.0
0
0
1
0
0
0
3
Ch Ch alone
460
Group
% Correct
A
B
C
E
F
H
T2A
A
100
7
0
0
0
0
0
0
B
83.3
2
10
0
0
0
0
0
C
83.3
0
1
5
0
0
0
0
E
66.7
3
1
0
10
0
0
1
F
0
1
1
0
0
0
0
0
H
0
0
0
0
1
0
0
1
T2A
90.9
0
0
0
1
0
0
10
461
462
463
464
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;
19
Table 6 Mann Whitney results comparing group acoustic parameters between distant locations (~10km). Sample
465
size (Ch Ch: CREES 108 and VC 441) and (Squeak: CREES 136 and VC 1150). *All associations significant at
466
p<0.05.
467
468
Call Measurement CREES Mean VC Mean U P-value
469
Squeak
470
Low Frequency 1896 1557 41532 0.00*
471
High Frequency 2905 2715 48217 0.00*
472
Bandwidth 90% 737 633 57946 0.00*
473
Energy (dB) 75 89 51471 0.00*
474
Dur90% 0.036 0.28 50231 0.00*
475
Peak Frequency 2577 2395 47752 0.00*
476
Delta Time 0.049 0.041 51588 0.00*
477
Center Frequency 2525 2370 40555 0.00*
478
Q1 Frequency 2341 2242 56039 0.00*
479
Q3 Frequency 2687 2477 37523 0.00*
480
Frequency 5% 2079 1951 71783 0.11
481
IQR Duration 0.02 0.015 51031 0.00*
482
Max Time 99 222 47045 0.00*
483
484
Ch Ch
485
Low Frequency 2504 1497 12490 0.00*
486
High Frequency 12901 11268 13405 0.00*
487
Bandwidth 90% 4991 4263 16333 0.00*
488
Energy (dB) 112 88 11554 0.00*
489
Dur90% 0.03 0.03 23617 0.89
490
Peak Frequency 5904 5596 20661 0.03*
491
Delta Time 0.04 0.04 23239 0.70
492
Center Frequency 5703 5541 23653 0.91
493
Q1 Frequency 4585 4740 17581 0.00*
494
Q3 Frequency 6810 6235 18448 0.00*
495
Center Time 376 98 8585 0.00*
496
IQR Duration (s) 0.02 0.02 23220 0.68
497
Frequency 5% 3255 3215 20142 0.01*
498
499
500
501
502
503
504
505
506
507
508
509
All rights reserved. No reuse allowed without permission.
was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which. http://dx.doi.org/10.1101/688333doi: bioRxiv preprint first posted online Jul. 8, 2019;