Co-occurrence patterns
Co-occurrence analyses at large geographic scales give information on patterns issued from historical processes, often shaped by life history traits of the species involved (see Davis et al. 2018 for an example on Carnivores). Concerning West African commensal rodents, the only significant interspecific segregation pattern found by Hima et al. (2019) among the four dominant species (M. natalensis , R. rattus , Crocidura spp., and R. norvegicus ) along the Cotonou (Benin) – Niamey (Niger) corridor was between R. norvegicus and M. natalensis . Conversely, the two Rattusspecies and the pair R. norvegicus / Crocidura spp. showed significant aggregation at this spatial scale (i.e. they were found more often than expected by chance in the same localities). The authors did not propose any explanation of these trends, that may typically result from a mixture of historic and stochastic processes on the one hand, and behavioural ones on the other hand, especially when the co-occurrence event do correspond to real co-existence / syntopy at the microhabitat scale. In Senegal, Dalecky et al. (2015) showed that aggregative patterns between native species of rodents seem to be disrupted by the presence of Mus musculus in commensal assemblages. At the scale of the city of Niamey and using different methodological approaches, Garba et al. (2014) found strong segregation patterns between nativeM. natalensis and both invasive R. rattus and M. musculus , whereas the latter two species showed either random or slightly aggregated (depending on the set of districts considered) co-occurrence patterns. Invasive rats and mice were found associated with urban areas characterized by intense commercial and exchange activities (markets, coach stations and stores) that lies in the heart of town. In these habitats, they probably replaced native M. natalensis which has been formerly present, leading to the native / invasive segregation patterns observed.
Here, we were able to tackle the species co-occurrence questions at two different scales, thanks to our standardized sampling protocol. At the global scale, aggregation cases were only observed between native species, that probably share the commensal space for long. Both cases involved M. erythroleucus , with a relatively closely-related rodent species (P. daltoni ) on the one hand, and with the shrew (C. olivieri ) on the other hand. Interestingly, these three species can be considered as the most prone to live as commensals of humans among native ones, to the exception of M. natalensis , which may partly explain their regular associations in the localities sampled. Most of the segregation patterns observed cannot be discussed as they are likely biased by distribution differences between the species concerned. Conversely, the segregation observed between the invasive M. musculus and the native M. erythroleucus , well supported using both randomization schemes, is especially interesting as it echoes the situation observed in northern Senegal where the house mouse is progressively, and apparently rapidly, replacing native rodents (and especially M. erythroleucus ; Dalecky et al., 2015; Diagne et al., 2021). The processes underlying this invasion success are not yet fully understood, but they may include parasitological (Diagne et al., 2016, 2020, 2021) and/or immunological (Diagne et al., 2017) aspects. The speed of this replacement, which was estimated to cover a few dozens of years in Dalecky et al. (2015), is here highlighted by the segregation pattern observed, which tends to indicate that once the house mouse has colonized a new locality, M. erythroleucusrapidly declines in abundance, until it disappears. New samplings in sites where M. erythroleucus was still present in this 2013-2015 time window, especially along the Tambacounda – Kidira axis and along the Mauritania – Senegal border, would confirm this trend if it showed that the house mouse had become the dominant, or even the unique, rodent species present.
At the local scale, various patterns were observed between the commensal species in southern Senegal. The segregation observed in Tambacounda as well as in some of its districts between the invasive R. rattusand M. musculus was among the most significant. This trend toward a mutually exclusive distribution in separate housing or working plots may be the result of direct or indirect interactions between these two species. Such interactions have been documented in outdoor habitats of Pacific islands as in the Galapagos or in New-Zealand (Harper & Cabrera; 2010, Bridgman et al., 2018). In these cases, a negative impact of R. rattus on M. musculus was suspected, based more on indirect (risk effect) than direct / exploitation competition (possibly including predation by R. rattus on M. musculus ). The processes at work in complex commensal environments such as those found in large cities may be different, and the outcome of the interactions may not systematically benefit to the larger species (here the black rat). Instead, the house mouse may well be favoured in urbanized environments such as those that are developing in sub-Saharan Africa, as exemplified by the situation observed in Dakar (Stragier et al., 2020) and in most of the cities from the western part of Senegal North of the Gambia, i.e. the area which has benefitted from the groundnut trade for its early and accelerated development since the 1960s’ (Lombard et al., 2020). In such habitats, the small size of the house mouse could represent a real advantage to i) better hide from predators (including humans), ii) more easily slip in well-protected buildings and rooms and iii) subsist on less abundant food resources. From there, competition with larger rodent species (including native ones) may not represent a hindrance to the house mouse range expansion, contrary to what has been hypothesized from results obtained on experimental vs control grids in a 150,000 inhabitant city of central Argentina (Gomez et al., 2008). The continuous development of urbanization according to modern standards along the West-to-East major communication axes (mainly roads) should therefore lead to the continued invasion of the country by M. musculus , a trend that could be confirmed in the future by re-sampling localities where the species is either absent or sharing the space withR. rattus .
Other major cases of segregation at the local scale involve M. natalensis in Kédougou, with both C. olivieri and R. rattus (only with one randomization scheme, however). Here also, interactions probably occur regularly between these species that appear abundant in this city, which may have led to some kind of mutual exclusion at the scale of the housing or working units sampled. Competition between M. natalensis and R. rattus is regularly proposed to be at work, or to have occurred in situations where they were confronted: in Eastern RDC villages, it turned to the advantage of the black rat who replaced M. natalensis in a number of villages during the first half of the 20th century (Misonne, 1959). In Tanzania and Swaziland, the fact that M. natalensis rarely entered houses was associated to the dissuasive presence of R. rattus (or R. tanezumi ) in this habitat, a hypothesis which was strengthened by the regular observation of M. natalensis in commensal habitat in Namibia where no Rattusspecies occurs (Monadjem et al., 2011). Trying to find out which process may underlie this potential exclusion of M. natalensis byR. rattus in commensal habitats, Cuypers et al. (2017) failed to demonstrate an avoidance behaviour mediated by scent markings. Additional works are necessary to understand the processes at work, but the relative stability of the ratio R. rattus / M. natalensis in Kédougou since the arrival of the former species in this city more than 25 years ago (Bâ, 2000) advocate for good competitive skills of M. natalensis in this context. This is all the more apparent as, immediately around the distribution area of M. natalensis in southern Senegal, the black rat is very well installed and often dominant (Duplantier et al., 1997; Dalecky et al., 2015; Lucaccioni et al., 2016; this study).
At the same time, it has to be noticed that R. rattus andC. olivieri , in Kédougou as in other localities (namely Rufisque and Dielyani), show a clear aggregative pattern suggesting that they apparently co-habit quite easily in the commensal space. The fact that the black rat partly forages and lives in upper parts of buildings when the shrew exclusively lives at ground level may explain such cohabitation. These two species were also the most regularly involved in aggregative associations with native rodent species, and especially withM. erythroleucus . This may testify for an ancient cohabitation history between these species (more ancient than with M. musculus , in particular), and / or be linked with less overall niche overlap between them. The latter is not apparent when looking only at the microhabitat dimension, but may involve dietary, space use or other niche components.