Shrews (Soricidae) of the lowland forests around Kisangani (DR Congo)

Abstract Background The Congo Basin rainforest is the second largest rainforest in the world and one of the most biodiverse regions on Earth. Nevertheless, the Congo Basin biodiversity remains to be fully mapped, with many species awaiting discovery or official description. In recent years, much effort has been put into research on shrews (Soricidae), particularly in the region around Kisangani (D.R. Congo). Shrews are opportunistic feeders that are able to forage on a large diversity of invertebrate prey and therefore play an important role in the forest ecosystem. Furthermore, as they largely depend on forest habitats and have limited dispersal capacities, shrews form an interesting model group to study biogeographic patterns in the Congo Basin. New information This paper collates the efforts on shrew research from the wider region around Kisangani, in the centre of the Congo Basin. Apart from sampling information, the dataset includes morphological measures, DNA sequences and photographs. This dataset is therefore critical in the study of the taxonomy and ecology of Soricidae in the Congo Basin lowland rainforests.


Introduction
The Congo basin rainforest is the second largest in the world and one of the most biodiverse regions on earth (Mittermeier et al. 2003, Lewis 2005. Both the forest and its biodiversity are threatened by forest loss and bushmeat hunting (Malhi et al. 2013). Despite its importance for climate change mitigation and biodiversity conservation, knowledge on the impact of forest loss and degradation and bushmeat hunting on local biodiversity is currently lacking (Gibson et al. 2011, Alroy 2017, Phillips et al. 2017. In fact, the biodiversity of the Congo basin is generally understudied, with several new species being discovered every year (e.g. Colyn et al. 2010, Stanley et al. 2013).
Our knowledge on occurrence, ecology and taxonomy of shrews (Soricidae) in the Congo basin is currently incomplete (Mukinzi et al. 2005, Gambalemoke 2014, Jacquet et al. 2015. Shrews represent critical food web links via their role as predators of small vertebrates and invertebrates and as prey for several vertebrate predators (Churchfield et al. 2004). Apart from their importance in the ecosystem, shrews are also a suitable model taxon for evaluating biogeographic and historical hypotheses (Quérouil et al. 2003). Indeed, the central Congo basin (i.e. the lowland forests south of the Congo river) harbours less terrestrial vertebrate species than the northern part which is due to the smaller habitat area and isolated position of the central Congolian lowland forests (Van de Perre et al. 2019). As richness is limited by habitat area and isolation, this implies that diversity differences amongst Congolian lowland forests are mostly due to forest-associated taxa with limited dispersal capacities, such as shrews.
Therefore, this paper assembles shrew occurrences from three studies in the central Congo Basin (Gambalemoke et al. 2008b, Van de Perre et al. 2018) and includes the metadata of the captured specimens (morphological measurements, DNA and sampling details) which allows for the advancement of the knowledge on taxonomy and ecology of shrews in the central Congo basin.

General description
Purpose: This paper assembles data collected in the framework of the PhD theses of Jean-Claude Mukinzi Itoka, Sylvestre Gambalemoke Mbalitini and Frederik Van de Perre. These three theses were executed at or in collaboration with the University of Kisangani, the University of Antwerp and the Royal Belgian Institute for Natural Sciences. Although the design and purpose of the three studies differ, the sampling design (the method in which shrews were collected) is equal across studies, which justifies the publication of the dataset as a whole.

Project description
Study area description: We compiled data from three studies in the region around Kisangani (Gambalemoke et al. 2008b, Van de Perre et al. 2018. The combined data represent 36 sampling sites in which sampling effort was equal and distributed within 6 localities in the Tshopo Province ( Fig. 1

Sampling methods
Sampling description: In all localities, shrews were sampled using the paceline method, which involved placing 20 pitfall traps at 5 m intervals on transects (Nicolas et al. 2003).
Pitfall traps consisted of non-baited buckets (10-litre, 30×30×23 cm) that were buried in the ground, with rims even with the ground surface. A plastic drift fence (100 m) was set to increase capture effectiveness by guiding shrews toward traps. Pitfall traps were maintained at their locations for 21 days and were checked daily. Only in Yoko, these pitfall Sampling localities (dots) in the environs of Kisangani (see Table 1 for additional details). The city of Kisangani (square) is surrounded by a mosaic of agricultural land and regrowth forest (light green), while old-growth forests (dark green) can be found throughout the area. Blue lines represent the Congo River and its tributaries. The map on the left shows the situation of the study area within Africa.
Shrews (Soricidae) of the lowland forests around Kisangani (DR Congo) lines were set for multiple sessions at the same location. In all other locations, trapping was only conducted once (Table 1).
Quality control: Species were identified based on external morphology and cranio-dental characteristics. In addition, species assignments were confirmed for several specimens of each species by molecular analysis (16s rRNA). Taxonomic nomenclature follows Hutterer (2005). Specimens belonging to problematic species complexes that are in need of revision were provisionally labelled with cheironyms, pending formal description.
Step description: Field measurements Sex and sexual condition were noted for each specimen:  Palumbi et al. 1991) and 16S-Hm (reverse: 5′AGATCACGTAGGACTTTAAT-3′, Quérouil et al. 2001). PCR amplification was performed in 15-µl reaction mixtures that contained 7.5 µl Qiagen Multiplex, 0.2 µM of each primer, 1.5 µl DNA template and 5.4 µl sterile deionised water. The reaction mixtures were preheated at 95°C for 15 min, followed by 42 amplification cycles (95°C for 30 s, 46°C for 90 s and 72°C for 90 s), with a final 10 min extension at 72°C. The samples were purified and sequenced in both directions at VIB Genetic Service Facility (University of Antwerp). Sequences were aligned using the Geneious software (Drummond et al. 2015).

Geographic coverage
Description: Lowland forests of the Kisangani, Isangi and Ubundu territories of the Tshopo province (former province Orientale), Democratic Republic of Congo.

Taxonomic coverage
Description: All species belong to the family Soricidae, particularly the subfamily Crocidurinae. The dataset contains species from 5 genera: Crocidura (14 species), Paracrocidura (1), Scutisorex (2), Suncus (1) and Sylvisorex (4) (  The dataset contains a number of specimens that likely belong to species new to science. Specimens morphologically resembling known species but found far outside the distribution of the known species have been identified using a cf. statement, others were named using a cheironym. Crocidura sp1 yoko has easily distinguishable characteristics: small size (4-6 g), brownish on the back, greyish-brown on the belly, brownish tail that is completely glabrous, except from the base which is covered with few small vibrissae, the down side of the tail clear, almost white at the base and around the anus and its small paws are equally light coloured (Fig. 2). Its skull resembles that of Crocidura ludia but is smaller (Mukinzi-Itoka 2014). Sylvisorex n.sp. is a small and rare species. Brown greyish on the back and silvery grey on the belly. It has a long tail covered with small hairs that grow longer and are more numerous towards the tip, forming a white brush. The tail is brown-black on top and slightly lighter on the bottom (Fig. 3). The species resembles Crocidura polia Hollister, 1916. However, the new species has little vibrissae on its tail while, in C. polia, half of the tail is covered with vibrissae .
Scutisorex n. sp. (description in progress, J. Hulselmans pers. comm.) was found in RF Yoko and its distribution seems to be limited to the forest bloc between the Lomami and Lualaba. Specimen preservation method: Samples of liver, spleen and kidney were stored in 96% alcohol and RNA-later (only kidney). Blood samples were transferred to filter paper.