Boto (Inia geoffrensis—Cetacea: Iniidae) aggregations in two provisioning sites in the lower Negro River—Amazonas, Brazil: are they related?

Introduction

Alliances and aggregations are formed by many species of mammals for many different reasons. African elephants (Loxodonta africana) move long distances as a group for safety purposes (Thouless, 1995). Female and juvenile coatis (Nasua narica) form groups to gain access to limited food resources (Gompper, 1997; Gompper, Gittleman & Wayne, 1998). Southern flying squirrels (Glaucomys volans) form groups during the cold season to gain thermoregulatory benefits (Layne & Raymond, 1994). In a number of primate species, males typically form alliances to gain access to reproductive opportunities or to increase their competitive advantage in territorial disputes (Pusey & Packer, 1987; Kappeler & Van Schaik, 2002). Some mammalian females form alliances or social bonds mainly to compete for food resources while nursing, as observed in lions Panthera leo (Heinsohn & Packer, 1995), or for protection of their offspring, such as seen in baboons Papio cynocephalus ursinus (Silk et al., 2009) and feral horses Equus ferus caballus (Cameron, Setsaas & Linklater, 2009).

Many odontocetes form groups that regularly swim together as a unit (Norris & Dohl, 1980). Groups are formed for a variety of purposes, including feeding, breeding, communication, social interaction, learning and defense (Norris & Dohl, 1980). The formation of groups for these reasons, and as basis for alliances, are very common in the family Delphinidae, as is observed in Tursiops sp., Lagenorhynchus acutus, Sotalia guianensis, among others (Parsons et al., 2003; Connor et al., 2011; Mirimin et al., 2011; Lunardi & Ferreira, 2014). However, it is not common for river dolphins of the families Platanistidae, Lipotidae and Iniidae to form stable groups (Braulik, 2014; da Silva & Martin, 2014; Smith, 2014). The South Asian River dolphin, or susu (Platanista gangetica) is generally observed alone or in loose groups of two or three individuals, but may occasionally be observed in aggregations of 20 individuals or more (Braulik, 2014). Very little is known about the social organization of the Yangtze River dolphin, or baiji (Lipotes vexillifer). Group sizes were generally of three or four individuals, with as many as 16 observed in a single location, but it is unknown if these groups were based on stable social affiliations or were ephemeral associations related only to common use of resources (Smith, 2014).

The Amazon River dolphins (Inia geoffrensis), also known as botos, are solitary animals and are rarely seen in groups of more than four individuals; even when found in groups, these groups are not considered to be of stable long-term relationships (Martin & da Silva, 2004). The only stable bonds observed in botos are between a mother and her calf, during the lactation period (Best & da Silva, 1989a). Aggregations of up to several dozen botos often occur at waterway confluences and lake system entrances, but these are ephemeral and based on exploitation of a food resource (da Silva & Martin, 2014). Botos, as other river dolphins, including the susu, usually have fixed home ranges where they may hunt individually as well as cooperatively (Herman, 1980; Martin & da Silva, 2004; da Silva & Martin, 2014).

Cetaceans use a variety of strategies to find and capture their prey (Heithaus & Dill, 2009). While most species of dolphins are highly social, with groups covering large areas in search for food, other species appear to have fixed home ranges and can be seen in the same area day after day (Norris & Dohl, 1980). The latter behavior is common in river dolphins, including the susu, and the boto (Herman, 1980; Martin & da Silva, 2004; da Silva & Martin, 2014). Both species exhibit a wide range of prey capture methods and hunt individually as well as cooperatively in groups (Ballance, 2009).

The boto is the largest freshwater cetacean and is widely distributed throughout the Orinoco, Amazon River basins (Best & da Silva, 1989b) and Araguaia-Tocantins River basins (Hrbek et al., 2014). Despite occurring over an enormous area (da Silva & Martin, 2014), the botos’ home range is still poorly known, which is one of the priorities of its action plan (da Silva & Martin, 2010). Site fidelity was studied (McGuire & Henningsen, 2007); however, results were limited due to behavioral, morphological, and ecological characteristics. Daily movements of up to 20 km are typical, and individuals move at sustained swimming speeds of 3–6 km/h (da Silva & Martin, 2014).

In the state of Amazonas—Brazil, in the lower Negro River near Manaus, there are currently seven places where botos are provisioned and tourists are allowed to swim with them. These activities, here entitled boto aggregations, occur around wooden floating houses that have a submerged platform where tourists, along with keepers of the establishments interact with the animals (Alves et al., 2013). In general, aggregations are on average made up of eight individuals per site (VMFS, pers. obs., 2018). Each individual is identifiable by scars, pigmentation patterns and other marks, with each animal receiving a name from the keeper. Interactions between botos and tourists occur daily and with no pre-established time. The keepers feed the animals every day even when no tourist interactions occur, but as soon as a group of tourists arrives at the floating house larger quantities of food are offered. During the interactions, not all animals are fed at the same time and some do not get any food at all. This has been demonstrated in the aggregation at Novo Airão (Barezani, 2005; Romagnoli, 2010), but has also been observed in all other aggregations. Therefore there is no apparent feeding pattern, some animals come to feed every day for days at a time, while others come sporadically and disappear for days or even months (Barezani, 2005; Angélica Nunes, pers. com., 2015) while new animals appear at the aggregation and begin to interact with tourists (VMFS, pers. com., 2018). One of the provisioning sites that has been running continuously since 1998 is in Novo Airão (number 1 on Fig. 1) (Barezani, 2005; Romagnoli, 2010; Alves et al., 2011). In 2004 another aggregation was formed at the mouth of the Ariaú River, at the Ariaú Amazon Towers Hotel in Iranduba county (number 2 on Fig. 1) ( Romagnoli, 2010). Six other aggregations are shown in Fig. 1. In the Novo Airão aggregation, the keepers believe that the botos belong to a family group containing males and females, as well as offspring born within the group, but there are no data to confirm this supposition. Considering the importance of kinship in the formation of social bonds and taking into consideration that boto populations of the Negro River have not been genetically characterized, the goals of this study were to (1) genetically characterize individuals in the Novo Airão and Ariaú aggregations, (2) estimate genetic diversity within and between these aggregations, and, (3) determine genetic relationships between individuals (e.g., sibling, parent–offspring).

Materials and Methods

The two floating houses we sampled were Flutuante Boto-cor-de-rosa, currently known as AMA BOTO in Novo Airão (180 km north west of Manaus) in October 2004 and July and December 2005, and Flutuante do Boto, located in the mouth of the Ariaú River in Iranduba County (60 km northwest of Manaus) in March 2006. The two sites are 90 km away from each other (Fig. 1) and were the only ones in operation at the time samples were collected . Collecting permits were provided to IPF by IBAMA/ICMBIO (No. 11325-1 and No. 13462-1). The protocol for handling and removing tissue samples from live animals was approved by the National Research Institute of the Amazon (INPA) (No. 001/2013).

We led three different sampling expeditions in Novo Airão and one expedition in Ariaú. Collection of all samples followed the same methodology. For two consecutive days, we collected samples from all the animals that visited the aggregation, totaling around six hours of interaction with the animals per day. A total of 17 samples were analyzed—ten from botos of Novo Airão and seven from botos from the Ariaú River.

Epithelial tissue from the botos was collected using a modified thimble (Cunha, Azevedo & Lailson-Brito Jr, 2010) designed specifically to collect dolphin skin samples, and adapted by us for use on animals that allow human contact. We scraped the thimble along the pectoral fluke, removing long and thin superficial threads of skin. The animals demonstrated no visible reaction to sampling, and we observed no scarring later on. All samples were stored in 95% ethanol and deposited in the official animal tissue collection of the Amazonas Federal University (Coleção de Tecidos de Genética Animal—CTGA/UFAM).

We used skin fragments of approximately 3 mm for DNA extraction. Total DNA was isolated using the Genomic Prep Cells and Tissue DNA Isolation Kit (GE Healthcare, Chicago, IL, USA). The extraction was assessed both qualitatively and quantitatively through electrophoresis in 1% agarose gel.

Twelve microsatellite primers were used for the kinship analysis, five of which had been developed for Tursiops truncatus, Ttr11, Ttr48, Ttr63, TtrRC12 (Rosel, Forgetta & Dewar, 2005), TtruAAT40 (Caldwell, Gaines & Hughes, 2002) and one which was developed for Tursiops aduncus, MK9 (Krützen et al., 2001). These primers have proven to be informative for Inia geoffrensis (Gravena et al., 2009). The other six primers (Ig3A1, Ig11B1, Ig2G1, Ig1F1, Ig2B1, Ig1B1) were developed for the focal species (Gravena et al., 2009). PCR conditions follow those described in Gravena et al. (2009). All primers were labeled with a fluorescent dye following Schuelke (2000). Amplified microsatellite fragments were analysed in the MegaBACE™ 1000 automatic sequencer (GE Healthcare). We inferred the size of the amplified fragments using the ET 400 ROX size standard (Amersham BioSciences, Piscataway, NJ, USA) in the Genetic Profiler and Fragment Profiler programs (GE Healthcare).

Sex of all individuals was determined via PCR using one pair of primers for the ZFX gene and one pair for the SRY gene found on the X and Y chromosomes, respectively (Rosel, 2003). Although the primers were initially developed for Tursiops truncatus, they have been tested for other species of cetaceans, including botos (Rosel, 2003). PCRs were carried out simultaneously with both primer pairs using standard PCR conditions (Rosel, 2003). We used a minimum of two PCR reactions for each sample in order to assess reliability of the results. In each PCR we also ran two control animals of known sex, a male and a female, and a negative control. PCR results were observed on 2.5% agarose gel and compared with a 100 bp DNA size standard. An individual was considered male if two different bands were observed in the gel, one representing the ZFX gene amplicon (382 bp) and one the SRY gene amplicon (339 bp), whereas females had only one band representing the ZFX gene amplicon.

The microsatellites were analyzed for the presence of null alleles, allelic dropout and stutter using the program MICRO-CHECKER v2.2 (Van Oosterhout et al., 2004). Matrix of genotypes is available at https://github.com/legalLab/datasets. Deviation tests of Hardy–Weinberg equilibrium proportions and linkage disequilibrium were performed in ARLEQUIN v3.5 (Excoffier & Lischer, 2010). Genetic variability was assessed in terms of the number of alleles per locus, heterozygosity (observed and expected) and the estimated gene diversity within each aggregation. Considering that some of these estimates suffer influence of sample size (Leberg, 2002), we implemented a rarefaction analysis and calculated allelic richness (AR) and private allelic richness (PAR) in the program HP-Rare (Kalinowski, 2005), so that the number of alleles and allele richness estimates could be compared between localities. To show the power of the group of chosen markers we also estimated the joint probability of genetic identity (IC) according to Paetkau et al. (1995) and the joint probability of paternity exclusion at all loci (QC) using Weir’s (1996) method.

To infer the most likely number of biological groups existing within our sample, we used the program STRUCTURE 2.3.4 (Pritchard, Stephens & Donnelly, 2000) to generate posterior probabilities for different numbers of groups using the ‘admixture’ and ‘correlated-allelic-frequencies’ models. We explored the possibility of our sample containing from one to three biological groups. Assignment space was explored with 1,000,000 MCMC chains, preceded by 100,000 MCMC chains discarded as burn-in. Each analysis was repeated 20 times from a different randomly selected starting point. The most likely number of biological groups (K) was inferred using Bayes’ Rule (Suchard, Weiss & Sinsheimer, 2001; Pericchi, 2005).

Analysis of molecular variance (AMOVA) (Excoffier, Smouse & Quattro, 1992) was employed to describe the amount of differentiation and gene flow (Neigel, 2002) between the two aggregations. We analyzed the data using Wright’s F_ST (Wright, 1969) rather than Slatkin’s R_ST (Slatkin, 1995), as the variance of the latter tends to be greater when less than 20 loci are analyzed (Gaggiotti et al., 1999).

Estimates of recent migration rates between the Novo Airão and Ariaú aggregations were calculated using the program BayessAss v3.0 (Wilson & Rannala, 2003). Analyses were carried out assuming two geographic groups. Each MCMC chain had 5 ×10⁶ steps, with 10⁴ discarded as burn-in during posterior analyses.

Additionally, the inbreeding coefficients (F_IS) (Weir & Cockerham, 1984) using ARLEQUIN v3.5 (Excoffier & Lischer, 2010) were obtained. Analyses were carried out assuming two geographic groups. F_IS represents the average probability that any particular individual has two copies of an allele that is identical-by-descent at any particular locus, i.e., the average probability of expected homozygosity within each group.

We estimated the levels of relationship between pairs of individuals from the two aggregations using the relatedness coefficient (r) of Lynch and Ritland rLR99 (Lynch & Ritland, 1999) and Queller and Goodnight rQG (Queller & Goodnight, 1989) using the programs KINGROUP (Konovalov, Manning & Henshaw, 2004) and IDENTIX (Belkhir, Castric & Bonhomme, 2002). Assuming that the study group is not inbred (sensu pedigree inbreeding), r values of 0.12, 0.25, 0.50 or greater than 0.50, were considered as third order relatives (cousins, CO), second-order relatives (half siblings, HS) and first-order relatives (full siblings, FS, or parents and offspring, PO), respectively. Both positive and negative values near zero are representative of unrelated individuals.

Results

Relatedness

KINGROUP analysis showed low mean r values (relationship coefficients), both within and between aggregations (−0.03 within Novo Airão, -0.05 within Ariaú and −0.08 between the two locations). Although average values are negative, we can find a certain degree of relatedness of some individuals, both within and between aggregations (Fig. 2). Various coefficient values indicated first, second as well as third order level relationships between individuals. However, after analyzing the r values obtained for the pairs of individuals using both softwares, only 23% of comparisons, 34 pairs, had any kind of relationship confirmed. Thus, based on the described criteria to classify relationships, six were considered third order relationships (CO) and seven second order relationships (HS); 21 pairs showed disagreement in relationships inferred in the two programs, and 102 pairs were unrelated (Table S1).

Discussion

Currently there are seven floating houses provisioning Amazonian river dolphins in the lower Negro River near the city of Manaus. Little is known about the animals that frequent these sites. This was the first study that involved the use of molecular markers to understand the genetic structure and relationship of two of the seven aggregations, and to genetically characterize botos from the Negro River.

Conclusion

Of the 136 relationship comparisons, only 34 comparisons showed levels of kinship between botos in and between aggregations. Among them, we observed six pair that are considered cousins, seven pairs that are half siblings, and 21 pairs where assignments disagree between the two programs used. Although we observed few instances of familiar relationships, these results do not contradict the paradigm that aggregations of botos are not familial groups, and that aggregations are formed exclusively to receive food. The effects of this type of interaction between botos and humans are still unknown, and therefore basic biological, ecological and ethological research is necessary to develop a comprehensive understanding of the potential consequences of this newfound interaction. Efforts in this direction have already begun within the scope of the Projeto Botos do Rio Negro.

Supplemental Information