Sounds produced by human activities are often loud and may mask acoustic signals used by other species for communication. To circumvent this, animals may use various strategies, including shifting modality completely or complementing acoustic information by also using another modality of communication. Here we tested the overlooked multimodal shift hypothesis using pied tamarins (Saguinus bicolor) as models. We predicted that in noisier areas the species would exhibit more scent marking behaviour (i.e., olfactory communication), while reducing the emission of long calls (i.e., acoustic communication). We collected information on vocal and scent marking behaviour in nine groups of wild pied tamarins in Manaus, Brazil. We found that scent marking occurrence increased with noise amplitude, though the number of long calls did not change. Thus, our results do not suggest a shift between channels but complementation of information, where scent marking may counteract against the compromising of the acoustic channel by the urban environment. This is an interesting result from a conservation perspective and supports ongoing efforts to create ecological corridors in Manaus, Brazil; once habitat connection is established, pied tamarins may be capable of coping with city noise to communicate with conspecifics, a key tenet of species survival.

The study was conducted in the Central Brazilian Amazon, in the city of Manaus, Amazonas state, Brazil. We followed nine groups of pied tamarins in different forest fragments within the city limits of Manaus. The study area comprised five forest fragments ranging from 24 to ~700 ha, where other primate species also occur, including Saimiri sciureus and Pithecia crysocephala. The most common source of anthropogenic noise in all areas is road traffic, though there is also air traffic, park visitors, talking and screaming from children and university students, and singing and gunshots from military activities. The studied tamarin groups varied from 3 to 13 individuals; two of these groups were relatively well-habituated to human presence due to frequent public visitation to the parks in which they occurred. We followed non-habituated groups by radio-tracking. We captured the groups using baited Tomahawk TH105 (10 x 10 x 40 cm) live traps to attach the radio collars onto the animals, placing the traps ~1.60 m above the ground. Once we captured the animals we sedated them with 0.2 mg/kg Ketamine^® anesthesia and attached a SOM 2380 transmitter (164.00-164.99 MHz) (Wildlife Materials) to the alpha female of each group. The frequencies emitted by the transmitter were detected with a two or three-element directional antenna and an ATS^® receiver (164-168 MHz).  Data collection took place from November 2018 to December 2019. We followed each tamarin group for ten complete days from 6:00 a.m. to 5:00 p.m. In the field, we collected data during five-minute behavioral bouts (followed by five-minute intervals), recording all occurrences of scent marking and counting the number of long calls emitted from all individuals in the group in each behavioral bout. During these 5 minutes, we also took a 1-minute sample of noise amplitude and calculated the equivalent continuous sound levels (LCeq; time-averaged level of sound) (dbC), using a calibrated CEL-246 sound level meter (Casela Solutions). Calling and scent marking activity were recorded from all nine groups, but we only detected scent-marking behaviour in five groups.

There are missing data related to the coordinate of the place where tamarins scent marked or called when GPS error was greater to 15 m os mistakenly non-taken. There are also some missing values of amplitude estimates that were due to equipment space limit achievement during field activity or values of amplitude that were extravagant and related to some human-induced activity.