Insects were collected at 4 sites: Operators’ Camp, plant nursery of Namihombo, Ciramo (Exploration Camp) and Chinjira (Camp Saio). The geographical coordinates of these sites are shown in Table 1 .

Insects were collected by daytime and nightly. For diurnal insects, the manual method described by [19] , which involves the use of a swath net with a 2 mm mesh and a 30 cm diameter circular opening hung on a wooden handle, was used. Insects were captured by-passing the swath net over the grasses, enabling us to capture even insects camouflaged in the grasses. We then collected the captured insects by hand, wearing gloves, and placed them in jars. For insects resting on substrates such as trees, floors or house walls, we used the [20] method of visual observation and monitoring. This method relies on direct observation of insects in their natural habitat, often used for larger or more visible species. For this method, a container or jar was used for capture. We placed the jar on top of the insect; they fled inside, and we closed the jar. Diurnal insects were captured between 8:30 and 11:30 a.m. and between 2:30 and 5:30 p.m.

For nocturnal insects, we used the method described by [21] , which involves the use of light traps. Light traps attract nocturnal insects, such as butterflies and some beetles. These traps are often used to assess species diversity. So, we attached a white sheet to two pieces of tree fixed in the ground, then attached 4 flashlights to the two pieces of tree, i.e. two flashlights per piece of tree, with the light reaching both sides of the sheet for good illumination. Using the jars, when an insect came to rest on the sheet, we held the jar over it forcing it to fly inside the jar, which we then closed. The insects were captured from 6:30 pm at dusk until 9:30 pm, i.e., 3 hours. Counts of the captured individuals took place the following morning.

Insects specimen were identified using the identification keys of [22] - [25] and websites for each genus. Insect identification was made at the genus level.

The data were presented in tables showing, for each genus, the subfamily, family, order, capture site, time, distribution in the Afrotropical region and pictures of the specimen.

Eight locations were sampled for Amphibians and Reptiles. The stations are located between elevations of 1789 m at the Bonded Laydown Yard (BLY) near the confluence of the Lulimbohwe and Mwana rivers, and 2700 m at the Twangiza plant nursery in Chinjira, next to the Chinjira Health Center as shown in Figure 1 .

The characteristics of the sites from which Amphibians and reptiles were collected are presented in Table 2 .

A GARMIN MAP 64s GPS was used to pick the geographical coordinates of the data collection locations. An Android phone and an XP camera were used to capture photos of the specimens and the habitat surveyed. To facilitate capture at night, we used headlamps and jars to transport the specimen. To preserve the museum specimen, we used 10% formalin and 70% ethanol, and placed them in an airtight jar once labelled. We took care to collect a sample for molecular analysis at a later date.

Fieldwork was carried out during the day between 8am - 11am and 3pm - 5pm, and in the evening between 6pm - 9pm. Visual detection, opportunistic method or field raking [25] [26] were used during day and night, associated with the search of potential animal refuges such as litter, loose soil for burrowing species. Some data (lizards) were obtained by pitfall, a method that was used for small mammals’ collection. Eight locations were visited ( Figure 1 ) for the investigations, and their characteristics are listed in Table 1 . The interview technique was used to obtain information on the presence of snakes that we did not have the opportunity to collect during the fieldwork. Once collected, specimen were photographed alive and some were preserved for museum collection. Identification was based on the reptile and amphibian charts of [27] and the species status from [28] .

Some specimen were preserved for later molecular analysis. Specimen were first placed in 10% formalin for 48 hours. They were then washed with running water and placed in 70% ethanol. This procedure ensures long-term preservation of the specimens. Each specimen is labelled with its number in the collection.

Ornithological data were collected by direct observation and mist-netting [29] [30] . Captures were made at 5 different stations ( Figure 1 ). Observations consisted of walking along the tracks of specific birds in the woodlands on non-standardized routes and recording all birds seen or heard. As most birds are most active in the morning, active searches were generally carried out every day between 6.30 and 9.30 am, with afternoon observations around 3 pm at sunset.

Birds were captured using 6 mist nets measuring 12 m in length, 4 m in width and 36 mm in mesh size. The birds captured were first removed alive from the net and placed in a small cloth bag to keep them calm. They were then identified. Before releasing them at their capture station, the specimen were marked by cutting off the end of the first left rectrix with a pair of scissors, to avoid duplication.

Birds were identified according to morphological criteria based on feather, leg and beak coloration; beak shape, color of the region around the iris using the keys of [31] .

Small mammals were collected by three types of traps: 1) spring-loaded metal rod traps; 2) Sherman traps are rectangular aluminum or wooden boxes measuring 7.5 × 9 × 23 cm, fitted with an entrance and a trigger device through which the animal must pass before reaching the bait. Once attracted by the bait, the animal enters the box, triggers this device which closes the entrance, imprisoning it for having stepped on the trigger platform before accessing the bait; before reaching the bait. Clappers and Shermans were used to harvest rodents. 3) Pitfall involves burying buckets with a capacity of 26 × 26 × 19 cm (bucket depth, upper internal diameter, basal internal diameter) at ground level, with bottoms perforated so as not to retain rainwater. The buckets are spaced 5 m apart. They are also crossed at their axes of symmetry by a sheet of tarpaulin around 45 cm high. The sheet forms a barrier to shrews, supported vertically by sticks. The part of the tarpaulin in contact with the ground is sunk to a depth of around 5 cm, to completely block the passage of the shrews underneath. We used 42 buckets, 21 per line, at each sampling site. Pitfalls were used to sample shrews.

The taxonomic list of insect genera is given in Table 3 .

The insects found at the Twangiza mine site belong to 6 orders, 14 families, 20 subfamilies and 22 genera. Several subfamilies here have a single genus; but the Phaneropterinae and Mirinae encompass 2 genera each. The Acrididae family encompasses 4 subfamilies and 4 genera, more than any other family, followed by the Tettigonidae with 2 subfamilies subdivided into 3 genera. The Apidae, Sphingidae and Miridae families include 2 subfamilies each, and each subfamily is divided into 1 genus. The order Orthoptera is the most represented by genus, with 7 genera on its own, followed by the order Lepideroptera with 5 genera, and last is the order Odonata with only one genus.

The high representativeness of Orthoptera in this ecosystem can be explained by their adaptability and habitat diversity, as they are known to adapt to diverse habitats ranging from grasslands to forests and even urban areas; also, their ecological flexibility enables them to colonize various environments and this contributes to their abundance [35] ; this is also so considering the abundance of food resources, as they are herbivores and feed on various plant parts. The availability of food resources in a vegetation-rich ecosystem favors their growth and reproduction [36] . The presence of a diversity of habitats and food resources best explains the representativeness of Orthoptera in Twangiza, but environmental conditions such as temperature, relative humidity and available resources and the scarcity of predators or competitors must also be taken into account [37] [38] . The presence of hymenopterans, lepidopterans and coleopterans in the Twangiza ecosystem can be explained by the presence of flowering plants, as they are major pollinators [39] . During the inventory, despite the dry season, there were still plants with flowers on which these insects were looking for nectar for their diet. Therefore, we cannot undermine the presence of food resources as one of the factors explaining the presence of hymenopterans in an ecosystem; but also their adaptability to different inhabitants ranging from tropical forests to deserts [40] . Note that lepidopterans, coleopterans and hemipterans are also good indicators for assessing ecosystem health, as their diversity and abundance can reflect the state of habitats and environmental impacts such as pollution or climate change [41] .

Odonata are not well represented here, as they are an order of insects that depend on water, as Twangiza is a montane environment with few flowing waters or springs.

16 genera were found at the Operators’ Camp, 14 genera at the Exploration Camp (Ciramo) and 13 genera at each of the plant Nurseries (Namihombo and Cinjira) and at Cinjira (Camp Saio) sites. At the first two sites, there is still a lot of grass to provide insects with a food resources; whereas in the last two sites, the trees are tightly packed together and are already very tall, preventing the grass from developing as a result of their canopy. The number of genera is similar as the entire hill is made up of identical microhabitats. Few genera were captured at night, as a result of the strong draught. Insects do not stand strong drought, as this can either cut off their wings during flight or sweep the whole insect away; consequently, they remain camouflaged in the grass so as not to be destroyed and die.

All of these genera are known in the Afrotropics. They occupy Central and East Africa. Their presence in tropical Africa is due to their adaptability, habitat diversity, availability of food resources and favorable environmental conditions, despite the presence of predators such as birds, reptiles, amphibians and mammals.

Reptiles were represented by the order Squamates with its two Suborders. The Suborder Saurians with four Families: the Family Chamaeleonidae with two species (Trioceros ellioti and Trioceros rudis), the Family Gekkonidae with a single species Hemidactylus mabouia, the Family Lacertidae with the species Congolacerta vauereselli and the Family Scincidae with four species: Trachylepis maculilabris, Leptosiaphos graueri and Leptosiaphos sp.

As for the Ophidian suborder (snakes), we learned about local species by talking to the local people. The species cited are: Philothamnus sp. and Naja melanoleuca are the most frequently encountered, while Bitis arietans and Afrotyphlops sp. have become rare in the region. These are listed in Table 4 .

It shall be noted that caution must be made on the reliance on local interviews for snake species identification. However, it is clear that they are available as observed at several instances. What justifies the use of their scientific identification is the local knowledge of their vernacular names in local languages (Mashi, and Kilega).

The IUCN Red List [32] shows the endangered Leptosiaphos graueri, a burrowing lizard that lives in litter and loose soil in less disturbed areas and which deserves special attention for conservation.

Amphibians were represented by the Order Anurans. The specimens collected belonged to two families: Pyxicephalidae, with two species: Amietia desaegeri and Amietia sp, and Pipidae, with two species: Xenopus laevis and Xenopus wittei (photos in appendix). These are presented in Table 5 .

Figure 2 displays Amphibians and lizards (reptiles) average counts per site.

The species Xenopus laevis predominates in average numbers, followed by Amieta desaegeri, and Xenopus wittei. Their maxima respectively reach 29, 21 and 15 vs all the others (max 4 - 8). Regarding abundance-dominance, high similarities exist between Xenopus species and Amieta (>97%) on the one hand, and between Triceros rudis, (unlike Triceros ellioti) vs both Leptosiaph species (85% - 90%) on the other hand. Hemidactylus mabouia and Trachylepis maculilabris are very similarly distributed in abundance (>97%) and they are very distinct from the other species (55%). Congolacerta vauereselli and Amieta show low similarities with any other species. This is displayed in Figure 3 .

Regarding abundance, the occurrence of some dominant amphibian species diminishes in the distance descriptor as we move from the Operators’ camp towards Cinjira Camp Saio (Dominance D). Thus, in decreasing order: Operators’ camp, Namihombo, Laydown Yard, Nyorha, Raw water, Bridge, Cinjira Camp Saio and Cinjira Nursery. This is inversely illustrated by Simpson 1-D. The species with many individuals occur at Nyorha, followed with Chinjira Camp Saio and Laydown Yard. Elsewhere, the counts for various species are similar and lower. Evenness is lowest at Nyorha, as confirmed by the Menhinick index too. The equitability was slightly similar everywhere, except for the Mwiyadrira Bridge with minimum equitability, as only Amieta sp. was encountered there, as presented in Table 6 .

The highest similarity in species presence-absence only occur between Xenopus species and Amieta desa (>80%). More species diversity was found at Nyorha (5), followed with Laydown yard and Chinjira Camp Saio (3). Elsewhere, the species richness was lower and similarly (1 - 2 species). This is shown in Figure 4 .

The lowest diversity in terms of richness (S) was found at Mwiyadrira Bridge, as confirmed by Chao, ACE, Shannon, evenness, Brillioun and Margalef as presented in Table 7 .

Ecosystem restoration is now globally recognized as a key component in conservation programs and essential to the quest for the long-term sustainability of our human-dominated planet. The avifauna inventoried at four sites of the Twangiza mine is divided into 43 species, 22 families and 8 Orders (photos of some species in appendix).

The passeriformes order is the richest as far as families and species are concerned (15 families or around 68.2% and 36 species or 84%). This order is the most represented in the world and in tropical Africa [42] - [44] . Passeriformes (or passerines) is the largest order of birds, with over 6000 species representing 60% of birds.

The Muscicapidae family is the most represented in terms of species (6 species), followed by the Estrildidae (5 species) and three families with 3 species each (Cisticolidae, Nectarinidae and Malaconidae). Several studies have shown that the family of Muscicapidae is widely distributed in various habitats across the world. Birds of this family exhibit great diversity in morphology, behaviors, vocalizations, and life history, which makes Muscicapidae a great study group to address various questions on evolution, diversity and biogeography [45] - [47] . The systematic list of bird species recorded in the Twangiza Mining woodlands is shown in Table 8 .

Figure 5 shows the bird species richness of the sites investigated. bird species recorded.

It can be seen that the largest number of species (19) was recorded at the Operators’ Camp and surrounding areas. The Namihombo plant Nursery site was next, with 14 species, and last site was Ciramo, with only 9 species.

The most abundant species at all sites are the Striped Collie, Colius striatus; followed by the Baglafecht Weaver, Ploceus baglafesht and the Yellow-bellied Astrid, Coccopygia cartinia.

Some species were only observed or captured at some sites, such as Linurgus olivaceus, Fraseria caerulescens, Cinnyris chloropygius and Estrilda kandti at the the Operators’ Camp site, and Saxicola rubetra and Camaroptera brachiura at the Namihombo plant Nursery site. The species Batis molitor and Cossypha heuglini were observed only at Ciramo, and Pternestes afer and Anthus cinnamomeus at the Cinjira site.

Prior to the beginning of Twangiza Mining exploitation activities and 4 years before the ecological rehabilitation in the area, environmental and social impact studies were carried out. Thirty-eight bird species were recorded [48] . During our survey, 22 more species were added to the Twangiza bird list as encountered during these studies. These results clearly demonstrate the positive impact of the decade-fold reforestation activity on bird diversity in this area. Ecosystems restored by man are of great importance for the reception and refuge of various forms of life [4] .

The following species presented in Table 9 reached at least 2 counts at each site.

Hirundo angolensis and Colius striatus outnumbered any other species at the Namihombo site. The latter also predominated at the Operators’ camp and at the Ciramo camp.

Table 10 shows the confidence intervals between sampling/observation sites for bird species.

Birds were mostly observed in large numbers at Namihombo, followed by the Operators’ camp, where the respective maxima were up to 98 and 61. It shall be noted that Namihombo is an enclosed nursery with a very large number of plant species and flowers that are believed to attract bird species in a very conducive environment. The Operators’ Camp on the other hand has a very rich soil with lots of trees and a few flowers growing therein. Both Namihombo and Operators’ Camp have altitudes just above 2300 m asl.

Elsewhere, the maxima reached 19 - 21. The variations were so high whatever the site (cv = 331% - 357%); with many zero counts. The highest number of birds occurred at Namihombo and the lowest at Cinjira (high D and N; low equitability J and Pielou alpha). The numerical importance of the most abundant species is illustrated at Namihombo by the Berger-Parker Dominance index.

Table 11 shows the abundance-dominance index.

Regarding the presence-absence values, the number of species was highest at the Operator’s camp followed by Namihombo as illustrated by the diversity indices (high richness S, individuals N, Shannon H, Brillioun, Menhinick, Chiao-1, ACE vs low equitability J and Berger-Parker), as shown in Table 12 .

Eleven species of small mammals were captured in the Twangiza mine area, including 6 rodent species, 4 shrew species and one chiropteran species as shown in Table 13 .

These species are divided into 9 genera and 3 families. Crocidura bicolor is the most abundant species, followed by Lophuromus flavopuctatus, dendromus insignis, Dasymys incomtus and Epomophorus labiatus. In addition, two taxa have not been identified to species level: Crocidura sp and Myosorex sp.; highlighting the need for further taxonomic work. A molecular study is therefore essential for the complete identification of these two specimen.

[48] noted that the remaining indigenous vegetation of Twangiza was limited in extent. Thirty-eight species of birds and forty species of butterflies, but no mammals or reptiles, were recorded within the footprint area. In general, the bird fauna was depauperate both in numbers of species and in abundance. Birds recorded were mainly those of grasslands, and mobile species, which were using scrub along valley streams as a refuge. By far the most common bird was the pied crow (Corvus albus) a bird which is almost always associated with human habitation. No rare, threatened, endemic, specialist, protected or red data book listed bird species were encountered. Pockets of vegetation housed quite a variety of butterflies. Most of these were widespread nomadic or Afromontane species. There was one endemic, Albertine Satyrid (Bicyclus aurivillii) living along a valley stream. However, this species is found throughout the montane forests of the Albertine Rift region. Apart from this endemic species, no rare, threatened, specialist, protected or red data book listed butterflies were encountered.

Due to the high population density, there was also an increase in number of cattle and the little remaining grassland in the footprint area was overgrazed. Although there was some artisanal mining going on the flanks of the hill, the river itself was the main site of artisanal mining. All riverine vegetation had disappeared along the Twangiza River for several kilometers downstream from the pit. Even without the proposed development of the mine, it was clear that the remaining natural vegetation in the footprint area (apart perhaps from the sacred forest) was under threat and that it is only matter of time before this is also turned into farmland or gets totally degraded and eroded through overgrazing. An estimated 90% of the land had been transformed by anthropogenic activities. From a zoological point of view the little remaining biodiversity was of little importance and required no special attention.