Lagaropsylla signata (Wahlgren, 1903), previously known only from the Island of Java, Indonesia is redescribed and reported for the first time in Deer Cave, Gunung Mulu National Park, Sarawak, Malaysia (west coast of Borneo). Many were found clinging to the earwig Arixenia esau Jordan, 1909. A similar account of a phoretic flea (Lagaropsylla turba Smit, 1958) on the same species of cave-dwelling earwig has been reported in peninsular Malaysia in a well-documented association with the hairless naked bulldog bat, Cheiromeles torquatus Horsfield, 1824. The association of L. signata with A. esau is parallel to the evolution and co-existence with bats in Deer Cave just as in the case of L. turba, A. esau, and C. torquatus. The evidence suggests that L. turba and L. signata are obligate phoretic parasites whose survival depends on A. esau to access a bat host. Arixenia esau is reported for the first time in Deer Cave and the occurrence of L. signata on the island of Borneo represented a new record, previously being found only on the island of Java. Images of L. signata attached to A. esau are provided. Xeniaria jacobsoni (Burr, 1912), often associated with A. esau in other geographical areas, was not present in the material examined from Deer Cave. The natural history of the earwig genera Arixenia Jordan, 1909 and Xeniaria Maa, 1974 are discussed and summarized relative to their associations with phoretic fleas and their bat hosts.

Arixenia esau , Deer Cave, Gunung Mulu National Park, insect phoresy, Lagaropsylla turba , Xeniaria jacobsoni

Phoresy occurs in some insects and arachnids in which one species attaches to another species in a commensal relationship for the purpose of increasing their ability to disperse from one place to another. There is seldom a detrimental effect on the transporting host species. In some cases, attachment to another species is accidental, while others have evolved into vital components of their life history. Although there is an account of a bird flea attaching to a wasp (Rothschild & Clay, 1952), this was attributed to an accidental association of a wasp foraging on a flea-infested avian carcass. Lagaropsylla turba Smit, 1958 is the only documented species of flea that truly demonstrates phoretic behavior for which there was an association between a flea, another insect (A. esau Jordan, 1909), and a bat host (the naken bulldog bat Cheiromeles torquatus Horsfield, 1824). In this scenario, an earwig provides a vehicle for L. turba to come into contact with its only known host, C. torquatus. Nakakta & Maa (1974) provided a summary of the commensal forms of the dermapterid suborder Arixeniina and their associated molossid bat species. Important works cited by Nakata and Maa (1974) included: Jordan (1909), Jacobson (1912), Burr and Jordan (1913), Audey (1952), Cloudsley-Thompson (1957), Medway (1958), Giles (1961), and Popham (1962). Medway (1958) and Marshall (1977) described some crucial bionomical inter-relationships relative to L. turba, A. esau, and C. torquatus, while Marshall (1981) discussed similar parallel behaviors of co-existing associations between bird lice, hippoboscids, and their avian hosts. We reported herein a second flea species [Lagaropsylla signata (Wahlgren, 1903)] that has a similar phoretic association with A. esau. We also redescribed L. signata and further documented a new locality for both L. signata and A. esau.

Earwigs were collected from the floor of Deer Cave while conducting a general insect survey of Gunung Mulu National Park in October 2006 (04°01'18"N, 114°49'24"E) (KBM & GJS) and in January 2009 (04°02'32.8"N, 114°48'49.6"E) (GJS). Specimens were collected in the same location deep within the main gallery of the cave in the early evening. No earwigs were found in the entrance area and path through the first half of the cave, although mixed rock and guano piles were present. Approximately 1000 m and within sight of the “Garden of Eden doline” (a cave ceiling collapse that allows light to enter the main gallery), the path climbed an isolated, cone-shaped hill. Very few earwigs were present at the base of this hill, but numbers increased with elevation. The top of the hill, the observation path, and railings included earwigs. Thousands were present in October 2006 and only a few were present in January 2009. Individuals were observed actively walking, mating, and resting in small cracks. Males, females and nymphs were collected using forceps to overcome their strong grip on the surface of rocks, railings, and clothing. The greatest concentrations of individuals were at the highest points on the hilltop, which suggests a negative geotactic behavior. Individuals were not observed in the area around the base of the hill and nearby cave walls.

All specimens were collected into and stored in 95% ethanol and later examined in the laboratory. During microscopic examination, many fleas were noted attached to the earwigs. One species of earwig was present and the mouthparts and genitalia were dissected to facilitate and confirm our identification. Earwigs were photographed in ethanol with the aid of a Canon 6D DSLR camera and Visionary Digital Passport II imaging system. Image stacks were montaged with Zerene Stacker v.1.04. Fleas were mounted on microscope slides in accordance with procedures outlined by Hastriter and Whiting (2003:1043) and were illustrated with the aid of an Olympus BX61 Compound Microscope and an Olympus CC12 digital camera accompanied with an Olympus Microsuite B3SV program. All images were edited in Adobe Photoshop CC 2015. Earwigs and fleas were deposited in insect collections at the Monte L. Bean Life Science Museum, Brigham Young University, Provo, Utah (three A. esau; MH-905: slides 7♂, 3♀ fleas, many fleas in alcohol), the Museum of Southwestern Biology, University of New Mexico, Albuquerque, New Mexico (99 A. esau; MH-905: slides 2♂, 2♀ fleas,), and the Cleveland Museum of Natural History, Cleveland, Ohio (9 A. esau; MH-905: slides 2♂, 2♀ fleas).

Lagaropsylla signata was previously known only from the type series. Although the host was listed as Ch. plicata, no information of how the bat was collected (mist net), or its habitat (cave, open field, hollow tree, human dwelling, etc.) was provided. Banjuwangi, Java is only a general locality, since modern Banjuwangi is a sprawling metropolis at the eastern extreme of Java along the Straights of Bali. In our study, a total of 72 L. signata were collected from the bodies of A. esau (7♂, 15♀) and from bat guano on the cave floor (15♂, 35♀). Cheiromeles torquatus was not found in Deer Cave in our study. Evidence of its historical presence was documented by Cranbrook (2010) in Niah Cave from archaeological records from the Pleistocene (40,000 years ago). Niah cave is approximately 115 km from Deer Cave. The finding of both L. signata and L. turba on A. esau implies that they have both adopted the same phoretic vehicle to access a bat host. Such evidence suggests that C. torquatus is present in Deer Cave and may serve as the principle host for L. signata but the definitive bat host for L. signata is unknown.

Because of the association of A. esau (Fig. 10) and Xeniaria jacobsoni (Burr, 1912) (two closely related earwigs) and their potential bat hosts, a discussion of the known biology of each follows. Nakata and Maa (1974) and Marshall (1977) provided assessments of the dermapteran suborder Arixeniina and their associated molossid bats. The phylogenetic position of A. esau as sister group to one lineage of free-living spongiphorid earwigs in the genus Marava Burr, 1911 has recently been proposed (Naegle, et al. 2016). Medway (1958) documented colonies of C. torquatus in a “cave at Niah” (Subis Cave) in Sarawak, approximately 115 km from Deer Cave. From Subis Cave, Medway collected 23 fleas from a juvenile (non-flying) C. torquatus that had fallen from the roost 250 feet above the chamber floor. He noted that fleas were also widespread on the floor and were often attached to living Arixenia. These fleas, plus two additional males of L. turba that were collected in a hollow tree in Ulu Gombak, Selangor, Malaysia, comprised the type series from which Smit described L. turba. A bat was not associated with the Ulu Gombak collection. Arixenia esau was documented on the bodies of C. torquatus by Jordan (1909) and Medway (1958), while Cloudsley-Thompson (1957) reported “17 specimens of A. esau from bats harvested from a durian (Durio sp.) tree cavity” in addition to 168 specimens roaming free within the same tree cavity. Nineteen specimens of X. jacobsoni were also reported by Cloudsley-Thompson (1957) from the same durian tree cavity, but none were collected from bats. Medway and Chong (1969) documented A. esau on four mist-netted specimens of Ch. plicata in Pulai, Kelantan, and Borneo. Four specimens of X. jacobsoni were also reported by Burr (1912) from a cave near the shore of Babakan, Banjoumas Residency, Java. Further historic records of X. jacobsoni documented by Cloudsley-Thompson (1957) included two or three specimens from Mops mops (de Blainville, 1840) in Kuala Lumpur, Malaysia in July 1919, seven specimens from M. mops in “Malaya” in 1920, an unspecified number from Mindanao, Philippines, and 35 specimens from South Java (Banjoemos) from cave near Babakan in 1933. The presence of two earwig species occurring in the same cave environments and both documented on different bat species remains an enigma. It appears that A. esau is associated only with C. torquatus, while X. jacobsoni has been found on M. mops, Ch. plicata, and C. torquatus. Xeniaria jacobsoni has been collected on only one occasion on the body of C. torquatus but is found together with A. esau in common environs. The discovery of multiple specimens of L. signata attached to A. esau (behavior similar to L. turba) with no evidence of L. signata on any bat species will require further identification of bats in Deer Cave (and other caves) that may prove to harbor L. signata. The known distribution of bat hosts of Arixeniina species and the geographic distribution of Lagaropsylla spp. known from these areas are summarized in Table 1. Arixenia esau and X. jacobsoni have been associated in the hollows of trees but not in caves, while Xeniaria bicornis Maa, 1974 and Arixenia camura Maa, 1974 have been found together in the hollows of trees but neither have been found in caves (Nakata and Maa, 1974). The latter two species found only in Mindanao were reported by Nakata and Maa (1974) from Ch. plicata and C. torquatus, respectively. Note: Cheiromeles torquatus does not occur in Mindanao, although Cheiromeles parvidens, a close relative, does. Arixenia camura, identified by Nakata and Maa (1974) would better be referred to C. parvidens and not C. torquatus.

Medway (1958) observed negative geotropic behavior of A. esau, many falling from associated roosting bats at heights 250 feet above, and that their principle foods were insects and glandular skin of C. torquatus. Medway made no mention of earwigs on the cave walls. Burr and Jordan (1913) observed X. jacobsoni in countless numbers on the surface of guano and everywhere on the rocky walls at Gouwa Lawa (bat-cave), at Babakan, Java. This would suggest that X. jacobsoni (and likely A. esau as well) could serve as the principle vehicle for fleas (developing in the guano on the cave floor, or in tree hollows) to ride to the roosting bats located high on the ceilings of huge caves. It is highly unlikely that adult fleas would be capable of making the journey from the cave floor to the bat hosts hundreds of feet overhead. In the massive caves that harbor C. torquatus, these bats never make contact with the cave floor unless they fall to the floor as juveniles (which is fatal to the pups), or adults that bump into walls and are temporarily stunned. The inappropriate behavior of bumping into things (attributed to the presence of artificial light) was observed by Kirk-Spriggs (1989). Both earwig species feed on insects but X. jacobsoni feeds more voraciously on insects than does A. esau, as their diet also includes cannibalism of their own kind, especially those that are vulnerable during molting (Burr and Jordan 1913). Popham (1962) studied the morphology of the mouthparts of both species and concluded that the mouthparts of A. esau are more specialized for feeding (grazing) on the peculiar hairless, glandular exudates on the skin of C. torquatus, whereas the mouthparts of X. jacobsoni are less specialized and more suited to feeding on insects (although it has been found on M. mops, Ch. plicata, and C. torquatus). The differences in the recorded occurrence of A. esau and X. jacobsoni on different bat species may be a reflection of their differences in mouthpart morphology, dictating an ability to feed on the different skin (and hair) types.

Jordan (1909) made reference to A. esau occurring in the sack formed by the membrane of the wings of C. torquatus. Not to be confused with the gular pouch, Kitchener (1954) described the function of these pouches. With the wings inserted into these pouches (by manipulations of the hind legs and feet), they are neatly folded around the bat, facilitating a quadrupedal mode of mobility within their roost areas (caves, crevasses, and hollow trees). Schutt and Simmons (2001) also addressed the function of the sub-axillary pouches and their relation to quadrupedal mobility. Arixenia esau might have a lesser chance of falling from the highly active aerobatics of their bat hosts if they seek these protected areas during bat flight. According to Medway (1958), the gular pouch is not sufficiently large enough to accommodate Arixenia.

Cheiromeles torquatus is distributed in southern peninsular Thailand, Malaysia, the insular portions of Indonesia (Java, Sumatra, Borneo), and Palawan, Philippines (Schutt and Simmons, 2001). The distribution of A. esau likely follows that of this bat host (although the distribution of A. esau is less well defined than its host). Although C. torquatus is listed in the IUCN Red List of Threatened Species, version 3.1, as “Least Concern”, studies and assessments are particularly urgent because it is rare in some areas of its range and destruction of many habitats (hollow trees in forested areas) are being destroyed along with logging and human encroachment. The distribution of X. jacobsoni appears to be much broader, but may be a reflection of greater collecting activities.

Both L. turba and L. signata appear to require earwigs to transport them to a viable bat host that would otherwise be inaccessible. Such obligate phoretic behaviors requires additional studies, especially to elucidate the association of L. signata and its yet unknown bat host species.

We thank Michael Naegle for his phylogenetic work on Arixenia which directly led to the discovery of the phoretic fleas. Specimens were collected under permits from the Government of Malaysia; Economic Planning Unit of Malaysia, UPE: 40 ⁄200 ⁄19 SJ.1040 (permit ID no. 1389) and 40 ⁄200 ⁄19 ⁄1481 (permit ID no. 1933). We thank Brian Clark and Gunung Mulu National Park and staff; Fatimah Abang and the University of Malaysia Sarawak; Haji Ali Bin Yusop and Sarawak Forests Department; Lucy Chong, Sarawak Forestry Corporation and the State Government of Sarawak. We are also grateful to Glen D. Chilton and Terry D. Galloway for providing valuable suggestions to improve this paper. Portions of this project were funded with NSF Grants #DEB-1557114 (MFW), #DEB-0515924 (KBM), #DEB-0738179 (KBM), and #DEB-0845984 (KBM).