Surveys of small mammals on carefully selected mountains and mountain ranges on Luzon Island, Philippines, since 2000 have led to the discovery of seven previously unknown species of forest mice, Apomys, a remarkable radiation on just a portion of one island. On the basis of morphological and cytochrome (cyt) b DNA sequence data presented here, we propose a new subgenus, Megapomys, to include the large-bodied members of the genus, which form a monophyletic unit of relatively large mice (averaging ca. 65–110 g) with tails about as long as or slightly shorter than the length of the head and body; all of these species forage on the ground. Other members of the genus are assigned to the subgenus Apomys; they are smaller (ca. 18–41 g), have long tails, and usually or often forage above the ground surface. Members of the subgenus Megapomys include four previously recognized species (A. abrae, A. datae, A. gracilirostris, and A. sacobianus) and the seven new species described here (A. aurorae, A. banahao, A. brownorum, A. magnus, A minganensis, A. sierrae, and A. zambalensis). All occur in northern and central Luzon Island, with the exception of one species that occurs on Mindoro Island; none is present in southern Luzon. Each species can be distinguished both morphologically and genetically. Although there are few records of Megapomys below 500 m elevation, they are common above about 1000 m, and some species occur near the peaks of the highest mountains on Luzon (i.e., up to nearly 2900 m). On four mountain ranges, two species of the subgenus co-occur, one at lower and one at higher elevations, although there is usually some syntopic overlap. Sister-species usually occur allopatrically in different mountain ranges, with one possible exception. Some of these species occur in areas not previously known to support endemic mammals, indicating that these areas are previously unrecognized areas of mammalian endemism where further study is warranted.

In 2005 and 2007, we conducted surveys of mammals along an elevational transect on Mount Palali (peak 1707 m) in the Caraballo Mountains, a poorly known mountain range on Luzon Island, Philippines. The surveys covered eight localities representing habitats from lowland agroforest and regenerating disturbed lowland rainforest at 780 m to mossy forest near the peak. We recorded 24 species, including one native shrew, one non-native shrew, five fruit bats, seven insectivorous bats, one monkey, six native rodents, two civets, and one pig. One species of Apomys is the newly described A. sierrae, and two species of Chrotomys are potentially undescribed species. Elevational patterns varied among mammals: bats were most diverse in the lowlands, native nonvolant small mammals had almost equal richness along the entire elevational transect, and most species of large mammals were present at all elevations. Bait attractiveness and diel activity pattern differed among native nonvolant small mammals: Apomys microdon and Rattus everetti were nocturnal and attracted to coconut baits, Chrotomys sp. 1 and 2 were mostly nocturnal and favored earthworms, Crocidura grayi showed no bait preference and was active during day and night, and A. sierrae showed no bait preference and was most active at night. The non-native Suncus murinus was restricted to highly disturbed areas at 780 m, whereas all native nonvolant small mammals were present in both degraded and undisturbed forest, supporting the hypothesis that non-native small mammals are not successful in invading native habitats on oceanic islands when the native community of small mammals is diverse. We conclude that all three rainforest types at all elevations, including newly regenerating forest, provide important habitat for mammals on Mount Palali.

We conducted an elevational transect survey of mammals on the highest peak in the Mingan Mountains of Aurora and Nueva Ecija provinces in the central Sierra Madre Range of Luzon from May to August 2006 and documented 35 species of mammals. These included one shrew (Soricidae), six fruit bats (Pteropodidae), one ghost bat (Megadermatidae), three horseshoe bats (Rhinolophidae), three roundleaf bats (Hipposideridae), four evening bats (Vespertilionidae), one macaque (Cercopithecidae), 12 mice and rats (Muridae), two civets (Viverridae), one wild pig (Suidae), and one deer (Cervidae). Our survey of this small, poorly known mountain range included the discovery of two new species of forest mice (Apomys), a new species of shrew-mouse (Archboldomys), a probable new species of shrew-rat (Rhynchomys), and an unidentified species of tube-nosed bat (Murina). Species richness of bats decreased with increasing elevation, whereas that for native non-volant small mammals increased with increasing elevation up to 1677 m, then declined at 1681 and 1785 m. Statistically significant diel activity and bait preferences were observed among the native non-volant small mammals. Only the two species of small forest mice (Apomys microdon and A. musculus) were captured exclusively above ground; the other non-volant small mammals were nearly always captured on the ground. The Mingan Mountains, with either three or four endemic species, are clearly a significant center of mammalian endemism, deserving of conservation.

We surveyed non-volant mammals in montane (1300 and 1400 m) and mossy forest (1500 and 1550 m) on Mount Cetaceo in the northern Sierra Madre of northeast Luzon in 2004 and 2005. We recorded a total of 12 species of mammals, including one shrew (Soricidae), seven murid rodents (Muridae), one deer (Cervidae), one wild pig (Suidae), one macaque (Cercopithecidae), and one civet (Viverridae). In this paper, we present the first ecological data on two recently discovered species endemic to northeast Luzon, Archboldomys musseri and Apomys sierrae. We captured only three species at 1300 m; captured five species at 1400 m; captured or inferred six species at 1500 m; and documented four species at 1550 m. Combined with an earlier study at 960 m (3 species), these results imply that species richness might increase from the lowlands to ca. 1500 m. The small number of sampling localities prevented strong statistical inference, but percent trap success with roasted coconut bait declined with increased elevation along our short transect, and trap success with live earthworm bait increased along the transect, so that overall trap success varied little along the transect. The most abundant species, A. sierrae, significantly preferred earthworm bait and was significantly more abundant in mossy forest than in montane forest. The percentage of diurnal captures along the transect increased with increased elevation, from a low of 0% at 1300 m to nearly 11% at 1550 m, even though no species showed significantly more diurnal than nocturnal/crepuscular activity. These patterns are similar to those on other mountain peaks on Luzon. The number of native species recorded on Mt. Cetaceo (7), plus one species recorded in the nearby lowlands, is similar to the number expected on the basis of the peak elevation of Mt. Cetaceo, supporting the hypothesis that the diversity of small mammals on a given mountain on Luzon is correlated with the elevation of the peak. Large mammals on Mt. Cetaceo are actively hunted; because they use both lowland and montane/mossy forest habitats, they may be threatened by the combination of agricultural expansion in the lowlands and unregulated hunting.

We used a combination of capture methods—harp traps, a tunnel trap, and mist nets—to sample the bat assemblage at two locations (disturbed lowland forest at 620 m and old-growth montane forest at 1450 m) on Mount Banahaw, Luzon Island, Philippines. We placed harp traps and mist nets on ridges and across trails and streams in ravines. The tunnel trap was placed over streams. Over 13 nights, we captured 300 individuals representing 23 species, including two possibly undescribed rhinolophid species. Seventy percent of all mist net captures were frugivores (family Pteropodidae), 95% of harp trap captures were members of the narrow-space insectivore ensemble (the families Rhinolophidae and Hipposideridae and the vespertilionid Kerivoula whiteheadi), and 89% of all bats captured in the tunnel trap were edge-and-gap insectivores, comprising vespertilionids, excluding the subfamily Kerivoulinae. In the lowland forest location, harp traps set along a ridge captured 10 times as many bats as an equal number of harp-trap-nights in a nearby valley. The rapid accumulation of individuals and species over only 13 nights, and effective sampling of insectivorous bats of the families Rhinolophidae, Hipposideridae, and Vespertilionidae, including several poorly known species, indicate that the methods used in this study may be used successfully to further our knowledge of tropical bat assemblages.