Bibliometric survey of fission-fusion dynamics in mammals

ABSTRACT: The term ‘fission-fusion’ was first coined by Hans Kummer in 1971 and has subsequently been used to describe fluid animal social dynamics in which individuals move between subgroups of variable duration, composition, and cohesion. Fission-fusion dynamics (FFD) have now been recognized as far more widespread than previously thought, making comparative studies a fruitful avenue for assessing the variation in FFD across sites and species. However, the mammalian literature appears to show species- and site-specific biases. Furthermore, the definitions, methods, and parameters used to study FFD are highly variable because of site-specific logistical considerations and species-specific data collection protocols. In this study, we aimed to systematically assess these differences, which are often mentioned in the literature but rarely quantified. To do this, we conducted a bibliometric analysis of the mammalian FFD literature. We collected data from &gt;150 papers and found both inter- and intra-taxonomic differences in terminology, methodology and FFD variables recorded. We show there is a large bias in the literature towards chimpanzees, bonobos and spider monkeys. We found inconsistent reporting of metrics used to describe FFD (i.e., cohesion, temporal patterns, composition) across taxa. By collating this dataset of FFD literature, we hope to promote and help facilitate the comparative research that is important for understanding the evolution of FFD across phylogenetically distant taxa.<br>METHODS: We searched for published studies of FFD in primates and non-primate mammals (hereafter mammals) using various combinations of the keywords ‘fission,’ ‘fusion,’ ‘dynamics,’ ‘subgroup,’ and ‘grouping.’ We searched Google Scholar for peer-reviewed scientific articles, book chapters, and dissertations/theses that were original studies (i.e., not review papers). This bibliometric analysis included 31 variables across six broad areas: (<i>i</i>) basic information about the source; (<i>ii</i>) terminology; (<i>iii</i>) definitions; (<i>iv</i>) group demography; (<i>v</i>) methods; and (<i>vi</i>) subgrouping properties (See 'Variable Explanations' tab in spreadsheet). Basic information (<i>i</i>) included referential information such as year published, source (journal/book), study species, and study site. We used terminology (<i>ii</i>) to mean words that define an entity, such as the word used for subgroup. Definitions (<i>iii</i>) refer to the explicit way that terms were operationalized to collect data, such as the definition of a subgroup. Group demography (<i>iv</i>) refers to overall community/group characteristics, such as group size and sex ratio. With methods (<i>v</i>), we are referring to the types of sampling protocols used in the study such as the data recording interval. By subgrouping properties (<i>vi</i>), we mean the data resulting from employing these methods to assess social dynamics within and between subgroups (e.g., duration of subgroup, percent of solitary animals, etc.). We only included records that presented original fission-fusion data from wild animal populations. We therefore did not include theoretical papers or those based on simulations. Moreover, to be included in our dataset studies had to report at least some of the 31 variables, including at least one from 'subgrouping properties’ and either ‘definitions’ or ‘methods’ (i.e., no rows of ‘not listed’ for everything). <br>Our search yielded 321 publications, which we subset into Dataset A) all publications passing selection criteria (n=157 publications), and Dataset B) publications screened but which did not meet criteria for inclusion (n=164 publications). Publications in Dataset A were checked against each other to ensure there were no duplicates; for example, if a dissertation had the same data as a published paper, then we used the source that contained the most information. Some of these studies included data for multiple sites or species, which we separated out (into their own rows) when extracting data from the study (157 papers yielded 174 records). We highlight records that are commendable examples of same-method, cross-taxonomic studies (rows highlighted in yellow).<br>SUMMARY STATS: Of these records spanning 1968 to 2018, 86% were from primate species and 14% were from mammal species. The sources with the largest number of records were the <i>International Journal of Primatology </i>(29), <i>Primates </i>(21), and <i>Behavioral Ecology and Sociobiology </i>(15). These records represented 36 genera and 48 species that were studied at 82 sites in 39 countries. <br>