This paper presents the results of investigations conducted between 2011 and 2013 to discover additional populations of Mormotomyia hirsuta Austen. These investigations were conducted primarily in the relatively dry savanna of eastern Kenya, focusing on small hills and rocky outcrops resembling that of Ukasi Hill, the type locality of the “terrible hairy fly”. Investigations were conducted at 144 caves and at ground level, directly below 104 above-ground, narrow, horizontally-oriented fissures, often on near-vertical rock faces. Evidence of Mormotomyia was not found in any of the caves investigated. During the dry season, however, desiccated corpses of Mormotomyia were discovered embedded in a matrix of dried bat guano adhering to the rock face directly below fissures at Ngauluka and Makilu Hills, also located in the Ukasi area. Later, rainy season visits to these two hills revealed populations of living Mormotomyia while, contemporaneously, flies were absent from the type locality. Like the type locality, the rock face directly below the fissures on Ngauluka and Makilu was discolored with pink and purple vertical streaking, presumably stained by bat urine and guano. Using the characteristically stained rock face as a search image, expeditions were expanded to include areas further afield and living flies were found at a third site 187 km to the south. Formerly considered “the rarest fly in the world”, the conservation status of Mormotomyia appears robust. Mormotomyia was actively preyed upon in the field by two species of lizards and remains of the fly were found in a jumping-spider nest. During laboratory observations of five live flies, the single male exhibited lengthy periods of female-guarding, with females being enclosed within the span of the much longer and setulose legs of males for more than 10 minutes.
Mormotomyia hirsuta, the “terrible hairy fly”, was described by E.E. Austen in 1936, based on two male specimens collected in 1933, at Ukasi Hill (as Ukazzi) in the dry eastern Sahelian zone (Coe 1999: 5) of eastern Kenya. The fly is a curious-looking, brachypterous species (Fig. 1), with long, orange-yellow setulae covering the body—a feature particularly well pronounced in males. The non-functional wings are reduced to setulose. strap-like appendages and the halteres are likewise reduced to barely visible nodular processes (Kirk-Spriggs et al. 2011). The species was sufficiently distinctive to warrant the erection of the new family Mormotomyiidae to contain it (Austen 1936). Mormotomyia hirsuta (confined to Kenya), is currently the sole representative of the family and Mormotomyiidae is the only endemic, monotypic family of flies that occurs in the Afrotropical Region.
Mormotomyia was not collected subsequently until 1948, when V.G.L. and G.R.C. van Someren rediscovered the species at the same locality, associated with a vertical fissure in a large rock at the summit of Ukasi Hill (Copeland et al. 2011). On this occasion a large number of flies were collected (including the unknown females) and immature stages were sampled in bat guano that had washed out of the rock fissure. These immature stages (egg, larva and puparium) were subsequently described by van Emden (1950). who further noted that bat guano was the substrate for larval development. Despite sporadic searches, there were no further records of Mormotomyia in the wild until 2010, when the species was again discovered at the type locality (Copeland et al. 2011; Kirk-Spriggs et al. 2011).
Mormotomyia hirsuta has long been of great interest to Diptera systematists and conservation biologists (Courtney et al. 2009: 203; Kirk-Spriggs & Stuckenberg 2009: 172; Oldroyd 1964: 184), as a result of its contentious systematic position and rarity. Due to the aberrant form, reduced head and thoracic setation and wing venation, morphological taxonomists have been unable to resolve the phylogenetic relationship of Mormotomyiidae to other families of Diptera. It has variously been placed in the Calyptratae (Pont 1980: 713); or as a possible transitional family between acalyptrate and calyptrate flies, and probably closest to the Scathophagidae (as Cordyluridae) (van Emden 1950); or as related to families of acalyptrate flies, including Sphaeroceridae (as Borboridae) (Austen 1936) and Heleomyzidae (Sphaeroceroidea) (McAlpine 1989: 1484).
Detailed study of the specimens collected in 2010 allowed redescription of the thirdinstar larva and puparium, using stereoscan microscopy, and description of the female reproductive tract (Kirk-Spriggs et al. 2011). This study revealed that the structure of the female reproductive tract suggested that Mormotomyia should be ascribed to the acalyptrate superfamily Ephydroidea (Kirk-Spriggs et al. 2011) and later McAlpine (2011) noted that the general structure of the antenna of Mormotomyia concurred with this view. Recent advances in molecular phylogenetics (e.g. Wiegmann et al. 2011) presented an additional technique to help resolve the phylogenetic placement of Mormotomyia, and live flies, collected into 96% ethanol, for the first time were used for such an analysis (Copeland et al. 2011). Results of this study confirm placement of the Mormotomyiidae in the Ephydroidea. as sister to the remainder of the Ephydroidea. except the Ephydridae (Winkler et al., in prep.).
While clarification of the phylogenetic placement of Mormotomyiidae represents an important milestone, other issues of biological importance remain unresolved, particularly those related to reproductive and dispersal behaviours and species conservation. Questions that still need to be posed are. for example: what is the biological significance of the pronounced male-biased sexual-size dimorphism observed in the Ukasi population (Copeland et al. 2011)?; and was Mormotomyia restricted to a single relict population, with the attendant problems of managing the conservation of an endangered species (Courtney et al. 2009: 203), or were there other, as yet undiscovered, populations? Analysis of both mitochondrial and nuclear DNA of individuals from the population of flies collected in 2010 presented evidence of outbreeding, suggesting that the Ukasi flies were probably part of a metapopulation of Mormotomyia (Copeland et al. 2011). If other populations do exist, how does a fly with non-functional wings (Fig. 1) disperse? Examination of the tarsi of Mormotomyia (Kirk-Spriggs et al. 2011) revealed none of the modifications of the tarsal claws found in other bat-associated fly species known either to be phoretic on Chiroptera. i.e. My stacinobiidae (Holloway 1976), or ectoparasitic on them, i.e. Streblidae and Nycteribiidae (Oldroyd 1964: 184). To address these questions, further investigations were conducted. In this paper the discovery of three additional populations of Mormotomyia in eastern Kenya is reported, together with additional observations on natural history.
MATERIAL AND METHODS
For the purposes of this paper a differentiation is made between caves and fissures (rock fissures). The former refer to openings into a rock system, often relatively large, with a floor that is either continuous with ground level, or descends below it. Fissures (Fig. 2) are defined as rock fissures above ground level, usually narrow and horizontally-oriented (i.e. much wider than high), often present on vertical or near-vertical rock faces, making access difficult for mammalian and avian predators. Virtually all fissures located during this study were horizontally-oriented. As is the case with caves, fissures are commonly inhabited by bats.
Between 19 April 2011 and 9 January 2013 expeditions were conducted to sites in Kenya, falling approximately along a northwest-southeast transect, from coordinates 0.4501°N 36.8852°E to 4.6154°S 39.3532°E (Fig. 3). Geological formations similar to those occurring at the type locality of Mormotomyia were examined (i.e. small inselbergs/kopjes to medium-sized hills that represent remnants of ancient basement rocks). With the exception of three caves near the Kenyan coast, all were located in dry habitats. The presence of bats was indicated by the occurrence of fresh bat guano within caves or below fissures, and by aural and visual evidence. Evidence of fly presence was sought by searching bat guano on rock faces and on the ground directly below horizontal fissures and by examination of cave floors and walls, using battery operated torches (flashlights). During dry-season expeditions evidence of recent fly occurrence was sought by searching accumulated dry guano deposits for Mormotomyia corpses. On six days (2, 3, 5, 7, 8 and 9 December 2012) at Makilu and Ngauluka Hills, lizard predation of Mormotomyia was observed through binoculars between 08h30 and 09h00. Lizards were identified using Sprawls et al. (2002).
During observations of living flies in the field, the relative population size of Mormotomyia was monitored below one fissure at the Makilu Hill site over a period of 38 days (29 November 2012 to 7 January 2013). Living flies were counted every second day. between 08h00 and 08h30, both on guano and the rock face, to a height of c. 4 m above ground level.
When Mormotomyia were active, fresh guano was collected from below a fissure on Makilu Hill and placed in two one-litre plastic containers to a depth of c. 4 cm. Containers were covered with their original plastic lids, from which a large rectangular section had been cut and replaced with fine-meshed cloth. During transport to the laboratory, cotton material was placed over the containers. Containers were placed in a 60 cm × 45 cm × 45 cm Perspex cage and their covers removed. Emerging adults were provided with cotton wool soaked in sugar water. Containers were left in the cage after the emergence of adults to serve as possible oviposition sites. Twenty adults captured at the Makilu Hill site were placed in a two-litre plastic container containing moistened paper towelling. These were also transported to the laboratory.
Relationship of caves and fissures to the presence of bats (Chi-square = 20.22; p < 0.001).
Identification and deposition
Newly-sampled specimens of Mormotomyia from Mbuinzau and Makilu Hills were identified by dissection and comparison of the male terminalia with specimens from the original Ukasi Hill population. These were found to be conspecific with M. hirsuta. Voucher specimens are deposited in the National Museums of Kenya, Nairobi, the International Centre of Insect Physiology and Ecology, Nairobi and the National Museum, Bloemfontein, South Africa.
Exploration of caves and fissures
A total of 48 sites (Fig. 3) were examined between April 2011 and January 2013. Three of these were subterranean caves; one was a large hole in a Baobab tree, Adansonia digitata L. (Malvaceae), that housed numerous fruit bats; and the remainder were rocky hills with caves and fissures. Many of these hills had multiple caves and fissures. A total of 248 caves/fissures were investigated. Some fissures and caves, particularly those in the area of the type locality (Ukasi Hill), were investigated more than once and the total number of visits to caves/fissures was 337. Ukasi is one of a small chain of hills located within 2 km of each other, the others being Ngauluka Hill and Makilu Hill (Fig. 4). These three hills were visited on 31 separate days, comprising a total of 88 investigations of caves and fissures.
Distribution of bats and flies among caves and fissures
Appendix 1 provides a list of the areas and caves/fissures examined that contained resident bat populations, including those that also housed Mormotomyia. Caves and fissures that contained neither bats nor flies when visited are listed in Appendix 2. These data were used to test the following two hypotheses: firstly, that the distribution of bats was independent of habitat type (i.e., caves and fissures), and secondly, that the distribution of Mormotomyia was also independent of habitat type. During the survey, fissures were significantly more likely than caves to harbour bats. Bats were present in 18.8% (n = 144) of unique caves and in 46.2% (n = 104) of unique fissures (Table 1). Fissures were also significantly more likely to house Mormotomyia. Flies were found in 3.8% of fissures, while they were never observed in caves (Table 2).
Exploration in the area of Ukasi Hill (the type locality)
Twenty-three hills within 20 km of the type locality were investigated (Fig. 3). Although many caves and fissures were found to have resident bat populations (Appendix 1), evidence of Mormotomyia was found on only the two hills nearest the type locality, Makilu and Ngauluka Hills (Fig. 4). Living flies were not discovered at the type locality.
Relationship of caves and fissures to the presence of Mormotomyia (p = 0.03, Fisher's Exact Test).
When visited on 11 August 2012, substantial amounts of dry guano were found below Fissure 3 at Makilu Hill (Fig. 2), much of this adhering to the steep rock face near its base. Close examination revealed what appeared to be the desiccated remains of three dead flies, embedded in a matrix of dried guano. These were highly fragile and were sampled together with the guano and taken to the laboratory for cleaning and microscopical examination, where the insect remains were confirmed as that of M. hirsuta (Figs 6, 7). Similarly, during examination of fissures and caves on Ngauluka Hill on 26 September 2012, a single corpse was found, also embedded in dried guano, at the base of the rock beneath Fissure 14. During removal, however, this corpse became detached from the guano and was lost. The rock face below both Mormotomyia-positive fissures was stained purple with pinkish streaks (e.g. Fig. 2), as was the rock face below the fissure at the type locality on Ukasi Hill, where Mormotomyia was collected in 2010 (Fig. 8). This trend suggested that bat residence in fissures was associated with a characteristic discoloration of the rock face, a feature easily observed at a distance with binoculars. This visual cue was used to focus investigations more finely on fissures with similar characteristics.
Rainy season expeditions
Based on the evidence from dry-season collections of desiccated Mormotomyia corpses from Makilu and Naguluka Hills, these and the Ukasi Hill site were closely monitored following the onset of the rains, which began in November 2012. During this time populations of live flies were discovered to be active on fresh guano covering the ground beneath fissures in the rock faces on Makilu and Ngauluka Hills (Appendix 1). Guano had recently been washed out of these fissures by precipitation. Mormotomyia were not. however, found on Ukasi Hill. Shortly thereafter, searches in similar habitat, far to the south, revealed a third population on Mbuinzau Hill (Figs 3, 5, 9, 10; Appendix 1), at a distance of 187 km from the type locality. Living flies from below fissures Makilu 3, Ngauluka 16 and Mbuinzau 12 (see Appendix 1) were collected separately into 95 % ethanol, providing suitable genetic material for DNA analysis and allowing investigations of whether outbreeding had occurred between the original Ukasi Hill population and the newly discovered ones.
Phenology of Mormotomyia
At the Makilu Hill site, live flies were first observed on 29 November 2012, gradually increasing in number, until peaking abruptly on 9 December 2012 when > 800 individuals were counted (Fig. 12). Live flies were last observed on 3 January 2013.
Predation on Monnotomyia
During the period that Mormotomyia were active on Makilu and Ngauluka Hills, lizards were observed attacking flies. Observations with binoculars revealed that male and female Red-headed Rock Agamas, Agama agama L. (Agamidae), and the Five-lined Skink, Trachylepis quinquetaeniata (Lichtenstein) (Scincidae), were actively feeding on Mormotomyia at the entrance to Ngauluka fissure 16 and Makilu fissure 3 (Fig. 2). Eight lizards were observed feeding at Makilu Hill on 9 December when Mormotomyia numbers were highest. Lizards were not observed to enter the fissure itself.
In addition, the expedition to Ukasi Hill on 18 December 2011 revealed a jumping spider (Araneae: Salticidae) nest c. 2 m above the ground, mostly obscured by a tiny crack in the rock face, below the large fissure that produced Mormotomyia in late 2010. The nest was removed and. while one spider escaped capture, the nest and its remaining occupants were placed in a vial containing 95 % ethanol. Subsequent examination in the laboratory revealed three nymphal and one adult salticid (probably Menemerus sp.; C. Haddad pers. comm. 2014); one mostly intact corpse of Mormotomyia embedded in spider silk along with other Mormotomyia body parts, including one head and 11 legs; and the head and partial thorax of an acridid (Orthoptera) nymph.
Of the 20 adults collected in the field only one female and one male Mormotomyia survived the journey to the laboratory. The remainder were either moribund or dead. The female specimen outlived the male specimen and died after 11 days.
Four female and one male Mormotomyia emerged in the laboratory from guano collected at the base of Makilu Hill. On two separate days the single male was found to have enclosed a female within the span of his legs (Fig. 11). In neither case was the beginning of the interaction observed and it was not possible to determine whether the same, or a different female, was “guarded”. Both “guarding” periods lasted at least 10 inmutes. Short video footage was recorded as part of one encounter, and a single, possible attempt at copulation appears in the film between 30 to 32 seconds (pleiocarpal 2014a). Three other short videos of the same male illustrate feeding and grooming behaviour as it sponges sugar water placed on the surface of a rock (pleiocarpal 2014b, c, d).
Previous successful and unsuccessful expeditions to the type locality of Ukasi Hill in search of Mormotomyia suggested that the appearance of flies was unpredictable, except for an apparent relationship with heavy precipitation events, when moist guano is washed out of fissures, thus providing a suitable medium for larval development (Austen 1936; Copeland et al. 2011; Kirk-Spriggs et al. 2011; van Emden 1950). Up to now, the dark interior of these fissures has not been examined and it is possible that breeding of Mormotomyia continues uninterrupted, even during dry periods, providing bats are in residence.
During recent expeditions, conducted between 2011 and 2013, living flies were not detected at the type locality on Ukasi Hill, where they were encountered in 2010. The absence of living flies at Ukasi Hill, although discovered contemporaneously on two nearby hills, provides further evidence of the ephemeral nature of the presence of living flies. That notwithstanding, when living flies are present at a site they may be active, and readily detected, for a considerable period of time. The presence of Mormotomyia at Makilu Hill was documented for a period lasting at least 36 days (29 November 2012 to 3 January 2013), suggesting that properly timed future expeditions would allow more extensive behavioural studies of Mormotomyia to be conducted under natural field conditions.
The survey revealed three additional sites with Mormotomyia populations, making a total of four sites, one 187 km from the other three. Together, they indicate a much wider geographical distribution than was previously thought. The survey also indicated that a similar habitat profile (fissures and not caves) was common to all sites at which the fly was found, which should facilitate the location of additional Mormotomyia sites.
Fissures were significantly more likely than caves to harbour both flies and bats (Tables 1, 2). Most caves and fissures were visited infrequently, however, and the absence of bats and flies at a site examined on one or a few occasions does not necessarily imply that they are not present at other times. Additionally, access to some recesses of investigated caves was problematic, or impossible, and in general, considerably more challenging than closely examining the well-lit ground and rock faces below fissures. These factors may have affected the detection of bats and Mormotomyia in caves and, perhaps, biased the results. Nonetheless, sampling included multiple caves and fissures across an extensive geographical range and these data suggest strongly that fissures provide a preferred habitat for both Mormotomyia and the bats on which the flies depend.
Predation on Mormotomyia
Mormotomyia adults are preyed upon opportunistically by lizards and these may play a role in limiting the population size of Mormotomyia. It is likely that jumping spiders also prey opportunistically on Mormotomyia. Insects associated with the examined salticid nest included a few Mormotomyia and a nymphal acridid grasshopper. Although no predation by spiders was observed, the presence of multiple insect bodies suggests that the salticids carry insect prey back to their nest, to consume it within the safety of the crack in the rock face. Similar behaviour has been observed in another salticid species, Heliophanus termiophagus Wesolowska & Haddad, that carries its prey into the safety of tunnels within abandoned termitaria prior to consumption (Wesolowska & Haddad 2002).
Lizards were not observed entering fissures, and the association of Mormotomyia with narrow fissures on rock faces that are difficult to reach may offer considerable protection for this species, particularly against predation by small mammals. Predation of Mormotomyia by small mammals or birds was not observed, nor was there any evidence of predation on immature stages. Vertebrate spoor were not observed on, or near, the moist guano that had accumulated below the fissures.
Mormotomyia exhibits pronounced male-biased sexual size dimorphism, with seven different body part measurements significantly larger (by 33–61 %) in males than females (Copeland et al. 2011). Sexual size differences in insects usually favour females and often correlate with increased fecundity. Nonetheless, there are numerous instances of the reverse being true (Copeland et al. 2011, and references therein). Larger size in males may be driven by sexual selection, size being a proxy for male fitness. Larger males may be more likely to win battles for territoriality and access to females. A form of female-guarding behaviour, whereby the male stands above the female, enclosing her within the span of the legs, may also drive increases in male size (Bonduriansky 2006).
Female-guarding behaviour of this type by a Mormotomyia male was observed, as reported in some other Diptera families. Adler and Adler (1991), for example, studied three species of Tipulidae, the males of which guard females following copulation, by standing above them, normally maintaining this position until females oviposit, or until dislodged in conflicts with conspecific males. As is the case with Mormotomyia (Copeland et al. 2011), these tipulid males had longer legs than did the females, significantly so for two of the investigated species. Interestingly, in the tipulid Limonia simulans (Walker), sexual dimorphism was also observed in the shape of the last tarsal segments that in males are sinuate on their inner surface, while in females these segments are linear (Adler & Adler 1991). A similar condition occurs in Mormotomyia, although in this case the sexual dimorphism is confined to the first tarsal segment of the mid tarsus (Austen 1936; Kirk-Spriggs et al. 2011; van Emden 1950). For both species modified tarsal segments may be involved in guarding and also mating behaviour, although direct evidence for this has not been observed. Kirk-Spriggs et al. (2011) suggested that the sexually dimorphic tarsal segments of male Mormotomyia may serve such a clasping function during copulation. Post-copulatory female-guarding behaviour is also relatively common in flies of the family Neriidae (Bonduriansky 2006; Mangan 1979; Preston-Mafham 2001) and has obvious benefits in circumstances where male competition for females is high. Even if another male successfully drives off a guarding male before oviposition is completed, engaging the interloper in battle may be enough to allow the now-unprotected female to deposit all, or most of her eggs (Adler & Adler 1991).
In laboratory observations of confined Mormotomyia, events prior to the initiation of guarding were not observed. As a result, it is not possible to distinguish between post copulatory guarding and pre-copulatory persistence of the male, in the face of one. or more, instances of rejection by the guarded female. The duration of female-guarding was considerably longer for Mormotomyia (> 10 min, n = 2) than that reported for Tipulidae, where post-copulatory guarding episodes that were not interrupted by another male lasted for 2.3 ± 0.78 min (n = 25, Dactylobis montana (Osten Sacken)) and for 0.9 ± 0.23 min (n = 44, Limonia simulans (Walker)) (Adler & Adler 1991). Mormotomyia's longer guarding period suggests that pre-copulatory guarding may have been observed. The apparently rejected attempt at mating by the male in the video footage cited above appears to support this interpretation. The recently emerged females may not have been sufficiently sexually mature to mate. Alternatively, it is conceivable that guarding behaviour may be different when other males are present. The observations of female guarding behaviour by Mormotomyia are based on that of a single male that may have behaved differently in the presence of potential challengers. More extensive observations of non-teneral flies in the laboratory and the field are necessary to clarify the type(s) and significance of female-guarding in this species.
Conservation status of Mormotomyia
Although Mormotomyia hirsuta is stenoecious, distributional data presented here indicates that the species is more widespread than previously thought and probably not uncommon in sites resembling those in which it has already been found. Both the macrohabitat (inselbergs) and microhabitat (fissures) in which the species occurs are little affected by human activity. Similar rocky outcrops and small hills are widely distributed in the drier areas of Kenya, particularly in the eastern and northern parts of the country. It is likely that the species will also be found to occur in the expanse of Tsavo East and Tsavo West National Parks in Kenya, through to the border with Tanzania and across it into Mkomazi National Park, the southeastern-most extension of the Sahel (Coe 1999: 5). Hitherto, Mormotomyia has been considered “the rarest fly in the world” (F.C. Thompson, pers. comm. 2010). The results reported here suggest that the conservation status of Mormotomyia is robust and that no special efforts are required to ensure its continuing survival.
Adult Mormotomyia appear to be relatively fragile insects, frequently losing legs in nature and suffering high mortality while being transported from the field to the laboratory (a c. four-hour trip over mostly smooth roads). Collection of moist guano at Mormotomyia sites is easier, however, and the transport of a reasonably small amount of guano (perhaps 2–3 litres) should produce sufficient adults to study interactions between the sexes in the laboratory, including mating and guarding behaviour. Additionally (as indicated above), at certain times the number of active flies may be substantial and observations made in the field may yield useful information, although activity of adult flies below fissures appears to be limited to individuals that have recently emerged and whose behaviour is limited to ascending the rock face and entering the fissure.
The means by which Mormotomyia disperses remain unknown. For practical reasons it was not possible to examine the interior of fissures, or undertake trapping of bats to ascertain whether the flies are phoretic as adults, a possibility that appears unlikely given that Mormotomyia lacks the modified tarsal claws apparent in fly families known to be phoretic or ectoparasitic on bats (Kirk-Spriggs et al. 2011). Answering this vexing question should be the primary objective of future research.
Finally, the discovery of three additional Mormotomyia sites provides the opportunity to compare the genetic makeup of individuals among and within multiple populations of the species. Preliminary results of mitochondrial DNA-barcoding of five individuals each from Makilu. Ukasi. and Mbuinzau Hills yielded two clusters. One cluster included all five Makilu specimens and single specimens from Ukasi and Mbuinzau. The other cluster included the other four Ukasi specimens and the final four specimens from Mbuinzau Hill, nearly 200 km away (R.S.C., unpub. data). These data hint at considerable dispersal of Mormotomyia and support earlier suggestions, based on genetic analyses, that the Ukasi population is not a genetically isolated one (Copeland et al. 2011). Analysis of nuclear DNA from the three new populations is underway (W. Booth, pers. comm. 2014) and should shed light on the question of dispersal and gene exchange of this interesting fly.
This research was funded through a grant (to R.S.C), from the Mohamed Bin Zayed Species Conservation Fund. We are grateful to P.M. Kabiro of ICIPE's Earth Observation Unit who contributed the map in Fig. 3. We thank Chief B. Musoo for permission to work in the Ukasi area. C.M. Maithya, J. Muriuki, M. Musingila, K. Ngalu, M.K. Kirk-Spriggs and M. Kodheki assisted in the field.