Distribution, diet, and trophic level of Arvicanthis abyssinicus and Tachyoryctes splendens around the area of recently extinct Ethiopian Wolf Canis simiensis on Mount Guna, northwestern Ethiopia

: Abyssinian Grass Rats Arvicanthis abyssinicus and Common Mole Rats Tachyoryctes splendens are preyed upon by the Ethiopian Wolf Canis simiensis . The aim of this study was to assess distribution, diet and trophic level of Arvicanthis abyssinicus and Tachyoryctes splendens on Mount Guna, where wolves have recently become extinct. Rodents were captured with Sherman trap and identified, and samples were taken to Debre Tabor University for dissection and diet analysis via microscopy examination of stomach contents. 110 A. abyssinicus and 52 T. splendens were captured from the study area, and the estimated population sizes of A. abyssinicus and T. splendens in Mt. Guna computed by Peterson-Lincoln Index were 1,364 and 416, respectively. In addition, 379 burrows (203 of A. abyssinicus and 176 of T. splendens ) were counted. Both species were observed to consume plants and arthropods, with plants predominant. We recommend that intensive studies should be carried out to determine the effects of rodent communities upon Mt. Guna afroalpine and subafroalpine ecosystems. abstract: Guduunfaa: Gosoota hantuutaa keessaa tuqaa ( Tachyoryctes splendens ) fi hantuutni huuraa ( Arvicanthis abyssinicus ) soorata jeedala diimtuu ( Canis simiensis ) keessaa isaan tokko. Kaayyoon qo’annoo kanaa bakka duraan jeedalli diimtuu irra jiraachaa turte fi


INTRODUCTION
Rodents exploit a wide range of habitats throughout the world (Lange et al. 2004). Their distribution and abundance is influenced by vegetation structure and composition (Gebresilassie et al. 2004;Kannan & James 2009), which influence micro-climate and provide food and cover against predators (Hansson 1999).
Ethiopia is a physically and biologically diverse country in Africa, where differences in climate and topography have resulted in a wide diversity of habitats and species (Tedla 1995;Abunie 2000). Of 284 mammals identified from Ethiopia, rodents comprise 25% and contribute about 84% of the total endemic species (Yalden & Largen 1992;Bekele et al. 1993;Bekele & Leris 1997;Laverenchenko et al. 1997). Ecological knowledge about these rodent, including their distribution and abundance, is limited (Habtamu & Bekele 2008).
Distribution and abundance of rodents can be estimated by trapping and recording of animals signs, holes, or related elements that infer the presence of animals (Krebs 1978;Rabinovich 1980). The best techniques are those based on the use of capture traps and recording of signs, due to the crepuscular habits of rodents (Aplin et al. 2003).
Rodents are omnivorous, mainly consuming plant parts such as fruits and seeds, as well as small arthropods (Best et al. 1993). Techniques commonly used to assess the diets of rodents include field observations of partially consumed organisms (Meyer & Shiels 2009), captive-feeding trials (Shiels 2011), and stomach content analysis (Ruffino et al. 2011). While effective, stomach content analysis has in a relatively small number of studies (Hope & Parmenter 2007).
The Ethiopian Wolf Canis simensis is a Critically Endangered medium-sized canid highly adapted to live in Ethiopian afroalpine and subafroalpine ecosystems above altitudes of 3,000 m, and it is the almost exclusive predator of high altitude rodents (Marino et al. 2010;Yihune & Bekele 2012 Fluctuations in rodent population density can have major impacts on the dynamics of their food (arthropods and plants) and predators (Ims & Fuglei 2005). If the population density of rodents in an ecosystem exceeds the carrying capacity, prey can be affected by overconsumption and the ecosystem can collapse unless control measures are taken. Conversely, if the population density of rodents is too low to sustain predators, they can become locally extinct. The aim of the present study was to assess distribution, diet and trophic level of two rodent species on Mt. Guna known to be prey of the now-extinct Ethiopian Wolf population.

Study area
Mt. Guna is located in South Gondar Zone at 11.750 0 N, 38.250 0 E, with a peak rises to 4,231 m ( Figure  1). Mt. Guna has afroalpine (3,500-4,231 m) and subafroalpine (3,200-3,500 m) ecosystems suitable to Ethiopian Wolf Canis simensis and mountain rodents. The area coverage is 210 km 2 of land above 3,200 m, and 110 km 2 above 3,400 m, but no more than 40 km 2 above 3,800 m which is good habitat for Ethiopian Wolf (Sillero-Zubiri & Macdonald 1997;Belste et al. 2012) Mt. Guna is surrounded by six districts of South Gondar Zone namely Lay and Tach Gayint, Farta, East Este, Simada, and Dera. The economy of the people living in those districts is mainly dependent on subsistence Agriculture. They cultivate crops like barley, wheat, bean, potato, and own livestock including cattle, sheep, donkey, and horse. The vegetation of Mt. Guna includes different grass species like 'guassa' and tree species with many animal lives including rodents, jackals, gelada baboons, hyenas, and rock hyraxes. The area is also inhabited by giant lobelia tree which is known to be the unique characteristic to afroalpine and subafroalpine ecosystems (ALZR 2006;Belste et al. 2012).

Data collection
Before starting the field work, permission for rodent collection and habitat observation was acquired from South Gondar Zone Wildlife Conservation Authority Office, Ethiopia. A quantitative cross-sectional study was conducted in the dry season of 2015 on Mt. Guna, northwestern Ethiopia. Samples were collected at three localities: bottom, middle and top of Mt. Guna. Three 60 X 60 m trapping grids (1.08 ha) were set at each site, and 36 Sherman traps were placed: 12 at top of afroalpine, 12 at middle, and 12 at bottom of subafroalpine ecosystem in the sampling period. Traps were baited with a mixture of rolled oats and peanut butter. The traps were set in the late afternoon (1500-1600 h) and left in the sampling sites for three consecutive days and nights. They were checked during the morning (0600-1200 h) and afternoon (1600-1800 h). Capture terms consisted of two sampling periods (February and March) separated by a month interval without trapping. All captured rodents were used for species identification, distribution, as well as stomach contents analysis according to the standard (Hope & Parmenter 2007). After identification done through morphological characteristics including differences in body size, shape, fur texture and colour (Aplin et al. 2003;Jonathan 2004), only two species of rodent namely Abyssinian Grass Rat A. abyssinicus, and Common Mole Rat T. splendens were counted, marked and released to their respective habitats with other trapped rodents when encountered. In addition, active burrows (fresh burrows that have marks of rodents and freshly excavated soil and cut parts of various plants) of the above rodents were identified and counted from the same grids for further determination of distribution and abundances. Percentage of active burrows in the area was estimated as: Population of active burrows %= (Number of active burrows/Total burrows examined all over) * 100 (Feliciano et al. 2002;Desoky 2007).
Eight rodents (four for each species) were taken to Debre Tabor University, laboratory of Department of Biology, for dissection to remove the gastro-intestinal contents. Diet analysis was carried out using the method of DeBlasé & Martin (1981). Accordingly, contents of the stomach were placed in a petri dish and thoroughly mixed to loosen material to give all constituents a uniform distribution. Then the contents were examined under a light microscope at 20x magnification to identify food items. Four fields of observation were examined. The diagram of food web was also drawn and trophic level of rodents was shown for Mt. Guna based on current result and the literature.
Since Belste et al. (2012) reported the presence of Ethiopian Wolf in Mt. Guna after the report of species extinction (IUCN/SSC 2011), interview questions were administered to check the presence-absence of the species. Fifty local people were purposively selected and interviewed to know when the wolf was observed for the last time in the study area and the reason for extinction. The local people were purposively selected based on the knowledge of the wolf, distance from the study area (nearby) and residence in the adjacent villages.

Data analysis
Distribution of A. abyssinicus and T. splendens was calculated with descriptive statistics. Number of rodents and their burrows collected from the three sites was shown on table and graph. Absolute number of the rodents in a population was calculated by Peterson-Lincoln index (Seber 1982) as follows.
, where N= population size estimate, M= marked individuals released, S= size of second sample captured, and R= marked animals recaptured.
Food web for the overall study area was shown by table and diagram. Independent two-sample t-test on Minitab software was used to compare food items eaten by both species. In addition, Independent two-sample t-test on R-software was used to compare distribution of both species among the three sites. Statistical value of 0.05 was taken as significance level.

Rodent species identification and distribution
In the present study, two species of rodent namely Abyssinian Grass Rat and Common Mole Rat were collected from afroalpine and subafroalpine ecosystems of Mt. Guna (Image 1). From all sites, a total of 162 rodents were captured during the first session of study period. From these, 110 (52.73% from the top, and 47.27% from the middle) were A. abyssinicus and 52 (23.08% from the top, 26.92% from the middle, and 50% from the bottom) were T. splendens. Regarding to the distribution, high number of A. abyssinicus (52.73%) was collected from the peak of the mountain following the middle (47.27%); however, low number of T. splendens (23.08%) was collected from the top, even though there was no significant differences among the three sites (p= 0.41, df= 2.22). On the other hand, high number of T. splendens (50%) was sampled from the bottom of Mt. Guna from where no single A. abyssinicus was collected as indicated on Table 1. In addition, 124 A. abyssinicus and 48 T. splendens were captured in the second session of the study period with almost similar distribution as of first session ( Table 2). The overall estimated population number of A. abyssinicus and T. splendens in the study area computed by Peterson-Lincoln Index was 1,364 and 416, respectively.
A total of 379 burrows, of which 179 were active, were counted from the same grids laid in total of 3.24 ha to sample rodents. As shown on Table 1, there was no A. abyssinicus burrow counted from the bottom of the study area whereas 104 (59.09 %) T. splendens burrows were counted from the same site. However, there was high population of A. abyssinicus at the top (n= 107, 52.63 %) followed by the middle (n= 96, 47.37 %) of the mountain. These distribution differences of burrows among the three sites were also not significant (p= 0.84, df= 2). Out of the total burrows of each rodent, 51.14 % of them were active for T. splendens, whereas 43.84 % were active for A. abyssinicus. Such high percentage of active burrows supports the existence of significant numbers of both rodents in the study area.

Stomach content analysis
Eight rodents: four T. splendens and four A. abyssinicus were dissected and stomach contents were taken. Diets of each rodent were identified into plant and arthropod types under the light microscope. Accordingly, T. splendens and A. abyssinicus have been identified to consume both plants and arthropods (Table  3). In contrast, there was no significant difference on the consumption of arthropods between both species (p= 0.466, df= 4). In addition to the stomach content analysis, the remains of leftover of plant parts and arthropod parts after their feeding were observed on the way to their burrows during field work.

Habitat observation
During data collection, the authors have repeatedly observed Mt. Guna afroalpine and subafroalpine ecosystems. Destruction of grasses and other vegetation due to high population of rodents, rock hyraxes and livestock at the bottom and middle of the ecosystems was observed (Figure 2). Furthermore, the same condition was seen at the top of afroalpine due to high populations of rodents and some groups of Gelada Theropithecus gelada, endemic to Mt. Guna, foraging in the same habitat. Moreover, rodents living adjacent to agricultural lands were observed to migrate to agricultural fields and damage farmers' crops.

Assessment on the extinction of Ethiopian wolf
Side by side with data collection of rodents in Mt. Guna, 50 respondents were interviewed to check whether Ethiopian Wolf is present or locally extinct. All interviewees assured the local extinction of the wolf by habitat destruction in spite of other factors such as human killing, rabies virus and climate change they were asked. Most of the respondents had observed the wolf between 6-8 years followed by before 10 years. All of them responded that they have never seen the wolf since four years (Figure 2). They said that a single wolf has been observed for years until total disappearance from the study area in 2011. Environmental degradation due to high population of rodents as a result of the extinction of wolf predator from Mt. Guna was clearly seen through field observation. In support to this, study shows that cyclic fluctuations in population density of rodents have major impacts on the dynamics of their food and vertebrate predators (Ims & Fuglei 2005). In addition, high population of livestock was observed in the ecosystems. Such pressures can lead to massive destruction, depletion and degradation of wildlife habitats as well as severe reduction in wildlife population (Hillman 1993). Furthermore, field observation and interviews of local farmers showed that those rodents living in the study area migrated to nearby agricultural fields for searching of additional foods. Such migration observed to cause crop damage in agricultural fields even though it was not quantified by the present study.
Out of the total diets consumed by both species, there was high frequency of plant parts (not identified into species level) in their stomach contents which assures the fact that the majority of their diet is plant. However, more plant parts were consumed by Common Mole Rats than by Abyssinian Grass Rats. Even though both species consume arthropods (not identified into species level), it was significantly less when compared to plant parts.
In agreement to the current study, Best et al. (1993) identified that rodents are omnivorous animals mainly consuming large quantities of plant parts. This might be due to the availability, selectivity, and palatability of plant diets in afroalpine and subafroalpine habitat as already stated by Beyene (1986) for other species of the Tachyoryctes genera.
Based on the current result of stomach content analysis and other published literatures, trophic levels and food web have been diagrammatically shown for the study area. Besides to the stomach content analysis,   Lopes et al. (2016) identified all diet types and food categories for tropical birds. The study revealed 23 food types frequently eaten by birds and proposed standard names for birds (e.g., granivore, frugivore, and insectivore). Accordingly, plant diet consumed by birds mainly are fruits, seeds, grains, plant fluids, leaves, buds, grasses, nectar, pollen, etc. In addition, birds also consume insects including ants, cockroaches, termites, locusts, bugs, beetles, flies, butterflies, dragonflies, bees, wasps, stoneflies, and mayflies. Birds of prey eat animal flesh including fish, reptiles,mammals, birds, insects, and molluscs.

Red-throated Bee-eater Merops bullocki
Blue-winged Goose Cyanochen cyanopterus Wattled Ibis Bostrychia carunculata diets of animals existing in Mt. Guna were reviewed and summarized in Table 4. Based on the summary, a diagram of food web was drawn ( Figure 3). As shown on Figure 4, T. splendens and A. abyssinicus consume both plants and arthropods, hence grouped under omnivorous animals (2 nd , 3 rd and even more trophic levels). However, it is difficult to determine specific position of omnivores' trophic level since they consume materials from different trophic levels of the food web (Williams & Martinez 2004).

Thick-billed Raven Corvus crassirostris
Ethiopian Wolf Canis simensis, which is endemic to Ethiopia, is the rarest canid in the world with a total global population of 500 individuals and classified as 'Endangered' on the IUCN Red List of Threatened Species (Marino & Sillero-Zubiri 2013;Marino et al. 2017). The wolf highly adapted to live in afroalpine and subafroalpine ecosystems above altitudes of 3,000 m, and is almost exclusively the predator of high altitude rodents (Marino et al. 2010;Yihune & Bekele 2012). Study shows that Ethiopian Wolf is an important flagship species for conservation of the afroalpine biodiversity (IUCN/SSC 2011), but faces serious threats to its survival in its ecosystem. In the present study, all the interviewees agreed the local extinction of Ethiopian Wolf Canis simensis due to habitat degradation. According to the respondents the single wolf was seen in 2011 for the last time. This agrees with the report by IUCN/SSC (2011) on its extinction from Mt. Guna although Belste et al. (2012) claimed to identify the wolf from the same study area referring to the local community. For long, the ecosystems of Ethiopian highlands had been threatened by overpopulation, overgrazing, and crop cultivation (Leipzig 1996). Similarly, Mt. Guna is currently under human induced threats from agricultural expansion, livestock overstocking, overharvesting of natural resources and settlements (Belste et al. 2012) that might mainly resulted in the local extinction of Ethiopian Wolf.

CONCLUSION
In the current study, the distribution, diets and trophic level of two rodent species namely Abyssinian Grass Rat and Common Mole Rat was assessed from Mt. Guna afroalpine and subafroalpine ecosystems. The results from Sherman traps and burrows count showed that both species were identified to be highly populated in the study area. Stomach content analysis revealed that both species consume both plants and arthropods, hence grouped under omnivorous animals (2 nd , 3 rd and even more trophic levels). Furthermore, Ethiopian Wolf was also confirmed to be locally extinct from Mt. Guna due to habitat degradation.
Based on the current study, we recommend that intensive studies should be carried out to analyse the effect of rodent community on Mt. Guna ecosystems due to the extinction of Ethiopian Wolf. To reduce effects of rodent pests and environmental degradation in the present study area, the long run rodent prey control by their predator, and habitats conservation should www.threatenedtaxa.org The Journal of Threatened Taxa (JoTT) is dedicated to building evidence for conservation globally by publishing peer-reviewed articles online every month at a reasonably rapid rate at www.threatenedtaxa.org. All articles published in JoTT are registered under Creative Commons Attribution 4.0 International License unless otherwise mentioned. JoTT allows allows unrestricted use, reproduction, and distribution of articles in any medium by providing adequate credit to the author(s) and the source of publication.