Introduction
About 60 years ago, researchers published on the biology and taxonomy of Rhinomusca Malloch, 1932, a species of Stomoxyinae (Diptera: Muscidae) uniquely associated with both black and white rhinoceros (Parsons and Sheldrick Reference Parsons and Sheldrick1964; Zumpt Reference Zumpt1973). Rhinomusca appears to live on rhinos in a lifestyle similar to Haematobia Lepeletier & Serville, 1828. Little else is known about Rhinomusca in Africa, other than anecdotal mention in the rhino literature.
Rhinomusca has always been assumed to be the vector of Stephanofilaria dinniki Round, 1963 (Nematoda: Spirurida: Filariidae) in the black rhino (Round Reference Round1964; Tremlett Reference Tremlett1964). However, this has never been explicitly tested. Entomological studies associated with rhino translocations in the 1990s (Gakuya et al. Reference Gakuya, Kock, Lekolool and Mihok2024; Kock et al. Reference Kock, Mihok, Wambua, Mwanzia and Saigawa1999) provided an opportunity to address Rhinomusca as a potential vector in Nairobi National Park and to a lesser extent at Solio Ranch. Rhinomusca is present in at least two rhino sanctuaries in Kenya and can be trapped in reasonable numbers with Vavoua traps, especially with the use of dry ice (Mihok et al. Reference Mihok, Maramba, Munyoki and Saleh1996a). In the 1990s, Rhinomusca was absent from other areas in East Africa, including the then newly established sanctuary at Ngulia in Tsavo West National Park (Mihok et al. Reference Mihok, Moloo, Oden’y, Brett, Rakwar, Munyoki, Kiilu and Kyorku1996b). We are unaware of any further published surveys for Rhinomusca elsewhere. Nairobi National Park continues to represent the most natural state of black rhinos in Kenya as its population was protected during years of massive poaching (Khayale et al. Reference Khayale, Kanga, Omondi, Kariuki, Brett, Oloo, Mulama, Okita-Ouma, Gaymer and Amin2025). Solio Ranch represents an artificial situation. It was/is a ranch that has evolved as a major highland black and white rhino sanctuary following the first introduction of five rhinos in 1970 (Patton et al. 2007).
Here, we report on the biology of this enigmatic fly and document the patterns of filarial infection in flies from dissections of hundreds of Rhinomusca and thousands of other Stomoxyinae, along with a smaller sample of other Diptera. Contrary to expectation, Rhinomusca was not infected with Stephanofilaria Ihle & Ihle-Landenberg, 1933, but was commonly infected with two forms of Habronema Diesing, 1861, likely new species (Barlaam et al. Reference Barlaam, Traversa, Papini and Giangaspero2020). The morphology of these parasites is described through light microscopy and some electron microscopy. Details of other larval Stephanofilaria unique to Stomoxys spp. are provided in a companion paper (Mihok et al. Reference Mihok, Kock and Krüger2026). The identity of Kenyan Rhinomusca is also revisited.
Materials and methods
Trapping in Kenya
Most flies were captured in Vavoua traps baited with dry ice alone or with 1-octen-3-ol (octenol) in semi-evergreen Croton/Olivia forest by Nairobi National Park headquarters (−1.363° S, 36.772° E, elevation 1774 m) and within Acacia xanthophloea thickets/woodland (−1.412° S, 36.904° E, 1533 m) near the Mbagathi River (Lesilau et al. Reference Lesilau, Verschueren, van’t Zelfde, Musters, de Snoo and de Iongh2021). A much smaller sample of flies was also obtained from Solio Ranch (−0.237° S, 36.885° E, 1940 m) in Laikipia (King’ori et al. Reference King’ori, Waiguchu, Ruoro, Muriithi, Mumbi, Omondi, Aminga, Angwenyi, Mijele and Chiyo2024), along with flies netted from three black rhinos (Mihok et al. Reference Mihok, Moloo, Oden’y, Brett, Rakwar, Munyoki, Kiilu and Kyorku1996b). More details of these two highland sanctuaries with photographs of habitats are provided in Mihok et al. (Reference Mihok, Kock and Krüger2026).
Trapping was also conducted to check for the presence of Rhinomusca dutoiti Zumpt Reference Zumpt1950 in two lowland areas with rhinos: the Ngulia Rhino Sanctuary in Tsavo West National Park (−2.965° S, 38.239° E, 715 m) with 15–25 recently introduced black rhinos (Mihok et al. Reference Mihok, Munyoki, Brett, Jonyo, Röttcher, Majiwa, Kang’ethe, Kaburia and Zweygarth1992, Reference Mihok, Moloo, Oden’y, Brett, Rakwar, Munyoki, Kiilu and Kyorku1996b; Okita-Ouma Reference Okita-Ouma2004), and during boma adaptation near Lugard Falls (−3.040° S, 38.677° E, 341 m) in Tsavo East National Park at the first introduction of eight rhinos to this traditional range (Okita-Ouma Reference Okita-Ouma2004). Many other areas throughout East Africa were trapped in the 1990s during ecological studies without detecting any R. dutoiti among 570,000 flies.
Detection of parasites
Fly proboscides were dissected to document the filariae present in 19 species of Stomoxyinae, with a smaller sample obtained from non-biting muscids, Tabanidae and Sarcophagidae. Squashes were also prepared from the gut and thorax for some flies. Parasites are mainly described through light microscopy of fresh specimens conducted by the first author. Fly parts were teased apart and suspended in insect physiological saline, then lightly pressed between a microscope slide and a cover slip. These preparations were examined at 250× magnification on a Leica compound microscope to detect filariae, which were typically mobile. Representative photographs of parasites were taken at various magnifications with a built-in Leica film camera and digitized for presentation. The parasites were photographed without the use of a clearing agent.
Measurements at various body landmarks were taken of relaxed specimens with a calibrated ocular micrometre from several individual flies. Statistics are presented as the mean and standard deviation. The small form of Habronema in R. dutoiti was also examined with scanning electron microscopy (SEM) and a few proboscides were sectioned by conventional histology. This work was done by support staff at the International Centre of Insect Physiology and Ecology (ICIPE). A subsample of female R. dutoiti was aged by ovarian development (Moobola and Cupp Reference Moobola and Cupp1978). This method defines six age groups up to 7–10 days after eclosion for nonparous flies. The lifespan of Rhinomusca is only a few weeks, given observations on a few flies maintained by Parsons and Sheldrick (Reference Parsons and Sheldrick1964).
Attempts to transmit parasites to laboratory animals
A few R. dutoiti were given the opportunity to feed on the ears of rabbits, and Habronema parasites from proboscides were also injected into a few rabbits. The small form of Habronema from 15 proboscides of R. dutoiti was also mixed with saline (N = 4) or rhino serum (N = 11) and injected separately into cyclophosphamide-immunosuppressed Balb/c mice.
Identification of blood meals
Results for 12 R. dutoiti blood meals from Nairobi typed by ELISA for host blood in a three-tiered system for common African wildlife are reported here; these were not included in Mihok and Clausen (Reference Mihok and Clausen1996). Relevant observations on fly biology not elaborated on in Mihok et al. (Reference Mihok, Moloo, Oden’y, Brett, Rakwar, Munyoki, Kiilu and Kyorku1996b) are also included in this report.
Sequencing of Rhinomusca DNA
We sequenced 798 base pairs at the 3′ end of the cytochrome oxidase gene (COI), for two individual R. dutoiti flies (one each from Kenya and South Africa) to determine if this gene could be used to discern differences between populations. The 3′ end of the gene was chosen because, in tests with other muscids, it showed greater diversity than the 5′ end. The specimen from Kenya was freshly collected in Nairobi National Park in April 1998 using a Nzi trap at −1.36° S, 36.77° E, elevation 1730 m by the first author. The sample from South Africa was collected in Hluhluwe-Imfolozi Park, KwaZulu-Natal, from a black rhino, Diceros bicornis, in April 1998 at −28.22° S, 31.95° E, elevation 349 m by David Cooper (approximate location). To obtain sequences of the COI gene, flies preserved in ethanol were allowed to air dry prior to DNA extraction using the Qiagen DNeasy Blood and Tissue DNA purification kit (Qiagen, Hilden, Germany). The 3′ end of the COI gene was amplified using primers C1-J-2195 and a modification of L2-N-3014 (Simon et al. Reference Simon, Frati, Beckenbach, Crespi, Liu and Flook1994), namely Fly10A 5′-AAT GCA CTA ATC TGC CAT ATT AG-3′, using the following protocol: 95°C 2 min followed by 95°C 1 min, 53°C 1 min, 72°C 1 min for 35 cycles. PCR products were cleaned using Centri-Spin 40 columns from Princeton Separations (Adelphia, New Jersey, USA). PCR products were sequenced using ABI BigDye Terminator Cycle sequencing ready reaction mix (Thermo Fisher Scientific, Foster City, California, USA) and cleaned using G-50 DNA grade Sephadex (Amersham Pharmacia, Piscataway, New Jersey, USA). Reactions were run on an ABI Automated Sequencer. Sequence data were edited using Sequencher (Gene Codes Corporation, Ann Arbor, Michigan, USA), and sequences were aligned manually. The COI sequences for R. dutoiti from Nairobi, Kenya, and South Africa can be retrieved from GenBank using accession numbers PZ733943 and PZ733944.
Addressing the identity of Rhinomusca dutoiti
Trapping at six sites at Skukuza, Kruger National Park (−25.009° S, 31.597° E, 282 m) in April 2024 provided the first author with an opportunity to compare R. dutoiti in Kenya to Rhinomusca in South Africa (Ferreira et al. Reference Ferreira, Crowhurst, Greaver and Simms2024), and to observe flies on two captive white rhinos as well as other wildlife. White rhino dung was also briefly monitored to identify emerging flies. Rhinomusca dutoiti has hardly been studied since the 1960s in South Africa, except for a few flies typed for DNA (Haseyama et al. Reference Haseyama, Wiegmann, Almeida and de Carvalho2015). We have previously used the name R. dutoiti for the flies captured in Kenya following Zumpt (Reference Zumpt1973). We deemed it necessary to carefully consider the recent opinion (based on one fly) that a third species of Rhinomusca may exist in East Africa (Zielke Reference Zielke2022). We addressed that opinion through comparisons with Zielke’s unique characters relative to our specimens, and through colleagues examining a few historical specimens held in museums.
Results
Prevalence of Filaria
Proboscides from 4,031 Stomoxyinae and 788 other Diptera were dissected and examined for filaria (Table 1), along with 220 guts and 93 thoraces. Parasites with the gross morphology of Habronema (Knapp et al. Reference Knapp, Krecek, Horak and Penzhorn1997; Naem Reference Naem2007; Pugh et al. Reference Pugh, Hu and Blagburn2014) were detected in R. dutoiti with two common forms. Overall prevalence was 55% with roughly twice as many proboscis infections in the riverine habitat versus the forest (61% vs 37%, χ2 = 8.88, df = 1, p < 0.005, N = 203). The small form of Habronema was readily differentiated from Stephanofilaria found in parallel in Kenyan Stomoxys (Mihok et al. Reference Mihok, Kock and Krüger2026) by body length versus oesophagus length (Figure 1), along with morphological features. The proportion of oesophagus versus body length was similar in all filariae in R. dutoiti (0.42, linear slope, r2 = 0.87). The small form of Habronema was found on its own in 92 R. dutoiti and one Haematobosca latifrons Malloch, 1932. The large form was found as a mixed infection with the small form in 20 R. dutoiti and on its own in three R. dutoiti, one Haematobosca hirtifrons Malloch, 1932, and one Haematobosca squalida Grünberg, 1913. Haematobosca was the only fly that harboured Stephanofilaria and Habronema, but only as separate infections. A large filaria with unique cephalic features (differentiated as ‘x’ in Figure 1) was also found in one R. dutoiti with 20 larvae present. A fourth nematode was also found as a single parasite in one unidentified Musca. It had an intermediate body and oesophagus length with an indistinct glandular oesophagus, which is not typical of Habronema.
Occurrence of Habronema and two unidentified nematodes in proboscides of flies caught in traps at Nairobi National Park and Solio Ranch along with flies caught from sedated black rhinos at Solio Ranch
Table 1. Long description
The table is organized into eight columns. The first column lists Fly species or group. The next three columns are grouped under Number of flies dissected (Traps Nairobi, Traps Solio, Rhinos Solio). The final four columns are grouped under Proboscis infections (Small, Large, Small and large, Other).
* Rhinomusca dutoiti Zumpt: 203 from Nairobi traps, 40 from Solio traps, 6 from Solio rhinos. Infections: 88 small, 3 large, 20 small and large, 1 nematode x.
* Haematobosca latifrons Malloch: 244 from Nairobi traps, 4 from Solio traps. Infections: 1 small.
* H. squalida Grünberg: 128 from Nairobi traps, 1 from Solio traps. Infections: 1 large.
* H. hirtifrons Malloch: 117 from Nairobi traps. Infections: 1 large.
* H. aurata Zumpt: 128 from Nairobi traps. No infections recorded.
* Stomoxys (11 species): 2762 from Nairobi traps, 101 from Solio traps. No infections recorded.
* Prostomoxys saegerae Zumpt: 131 from Nairobi traps. No infections recorded.
* Stygeromyia sanguinaria Austen: 115 from Nairobi traps. No infections recorded.
* Stygeromyia maculosa Austen: 47 from Nairobi traps. No infections recorded.
* Haematobia: 1 from Solio rhinos. No infections recorded.
* Non-biting muscids: 631 from Nairobi traps, 50 from Solio rhinos. Infections: 1 nematode.
* Tabanidae (5 species): 71 from Nairobi traps. No infections recorded.
* Sarcophagidae: 36 from Nairobi traps. No infections recorded.
Total length versus oesophagus length in nematode larvae in the proboscis of Rhinomusca dutoiti and in a non-biting muscid in Nairobi National Park, Kenya.
Figure 1. Long description
A scatter plot with the horizontal X axis labeled Body length in mu m, ranging from 800 to 2600. The vertical Y axis is labeled Esophagus length in mu m, ranging from 300 to 1000.
Four distinct data groups are plotted:
1. Small Habronema (solid triangles): A dense cluster located at the bottom left, with body lengths between 1000 and 1400 mu m and esophagus lengths between 350 and 450 mu m.
2. Muscid nematode (single solid diamond): A single data point located at approximately 1550 mu m body length and 560 mu m esophagus length.
3. Large Habronema (open circles): A scattered group in the upper middle area, with body lengths between 1700 and 2200 mu m and esophagus lengths between 620 and 900 mu m.
4. Nematode x (solid circles): Four points at the far right, showing the largest specimens with body lengths between 2250 and 2500 mu m and esophagus lengths between 780 and 880 mu m.
The overall trend shows that as body length increases, esophagus length also increases across the different species groups.
Counts of larval forms in the proboscis with single parasite types were 6.9 ± 6.4 small Habronema (N = 72, max 26, range 1–24), and just one large Habronema (N = 3). When both forms of Habronema were present, counts were 9.1 ± 8.1 (small form, N = 21, range 1–26) and 5.1 ± 6.8 (N = 21, large form, range 1–26). In mixed infections, the number of small and large forms was not correlated (Pearson r = 0.03), with a median of three times as many small versus large forms (distribution skewed to low ratios).
Seven R. dutoiti had simultaneous gut and proboscis infections with one or two small Habronema in the gut, encapsulated or free. One R. dutoiti had a simultaneous infection of four of the large form in the gut, and seven of the small form in the proboscis. No thorax infections were detected. One R. dutoiti had a single small form in the gut but no parasites in the proboscis. None of the flies with gut infections had visible blood.
Proboscides were dissected in 52 females aged by ovarian development. Nearly half (42%) of these flies were in age class V (7–10 days old if nonparous), with the other five age classes varying from 6% to 13%. Fourteen were parous and hence were up to perhaps 20 days old. Parasite prevalence did not differ by age class or sex, with an overall average of 65% in this subsample of additional dissections. Laboratory animals injected with proboscides of R. dutoiti or fed on by R. dutoiti showed no evidence of any filaria establishing after a month (no sign of ill health, no ulcerative skin sores, and no microfilaria appearing in blood).
Filaria morphology with light microscopy
The common forms of Habronema had a minimum of four large rounded sub-median cephalic papillae surrounding a buccal capsule with concave lips (Figure 2). Additional papillae may have been present, but an accurate count was obscured by buccal structures. All filariae had a clearly differentiated glandular oesophagus (Figures 2 and 3) except for the single nematode found in Musca. The large form of Habronema was grossly similar to the small form but with a few minor differences. The cuticle was often prominently striated, but not in all specimens (Figure 3), and the oral cavity was noticeably long (Table 2). The large form also had a shorter glandular portion of the oesophagus (7.8% vs 12.7%).
Small form of Habronema in Rhinomusca dutoiti. A. Head 1000× showing large papillae and indistinct buccal lips; B. Anterior 400× DIC, showing nerve ring (arrow) and transition from narrow muscular to wider glandular oesophagus; C. Junction of oesophagus and intestine 1000×; D. Posterior 1000× showing bulbous curved tail with small papillae, anus/rectum and striations.
Figure 2. Long description
A four-panel set of micrographs labeled A through D.
Panel A, top-left: A high-magnification view of the head showing large papillae and indistinct buccal lips at the anterior tip.
Panel B, top-right: A lower-magnification D I C micrograph of the anterior region. An arrow points to the nerve ring. The image shows the transition from a narrow muscular oesophagus to a wider glandular oesophagus.
Panel C, bottom-left: A high-magnification view of the junction where the glandular oesophagus meets the intestine, showing a distinct rounded interface.
Panel D, bottom-right: A high-magnification view of the posterior end. It features a bulbous curved tail with small papillae, the anus and rectum, and visible transverse striations on the cuticle.
Large form of Habronema in two species of Stomoxyinae. Rhinomusca dutoiti: A. Head 1000× in a worm with prominent striations; B. Head 1000× in a worm with a smooth cuticle; C. Nerve ring 1000× in a worm with prominent striations; D. Anterior 100× DIC; E. Anterior and posterior 250× DIC; F. Tail 1000×; Haematobosca latiforns: G. Head 1000× in a worm with a smooth cuticle; H. Tail 1000× DIC in a worm with a lightly striated cuticle.
Figure 3. Long description
The figure consists of eight panels labeled A through H.
Panel A shows a high-magnification view of the head of Rhinomusca dutoiti, featuring a distinct buccal capsule and a cuticle with deep, wavy striations along the outer edge.
Panel B displays the head of another Rhinomusca dutoiti specimen where the cuticle appears smooth without the prominent ridges seen in A.
Panel C focuses on the nerve ring area of a striated Rhinomusca dutoiti, showing internal anatomical structures and the serrated profile of the cuticle.
Panel D is a lower-magnification D I C micrograph showing the elongated anterior portion of the worm curving across the frame.
Panel E shows two separate sections of the worm, one anterior and one posterior, at medium magnification using D I C.
Panel F provides a close-up of the rounded tail end of Rhinomusca dutoiti.
Panel G shows the head of Haematobosca latiforns, characterized by a smooth outer cuticle and internal tubular structures leading from the mouth.
Panel H displays the tail of Haematobosca latiforns using D I C, showing a slightly tapered end with a cuticle that has very fine, light striations compared to the heavy ridges in the first species.
Measurements in μm of larval Habronema in the proboscis of Rhinomusca dutoiti in Kenya
Table 2. Long description
The table presents measurements for two larval forms: Small form and Large form.
Headers are organized into four main categories:
1. Length from tip of head: includes Body, Oral cavity, Total oesophagus, Muscular oesophagus, and Nerve ring.
2. Body width at various landmarks: includes Mouth, Head, Nerve ring, and Junction (intersection of oesophagus and intestine).
3. Width of internal structures: includes Muscular oesophagus, Glandular oesophagus, and Intestine.
4. Tail measurements: includes Width and Anus from tip.
Data for Small form (Mean values):
Body length 1235, Oral cavity 4, Total oesophagus 401, Muscular oesophagus 157, Nerve ring 106. Body widths: Mouth 7, Head 15, Nerve ring 30, Junction 35. Internal widths: Muscular oesophagus 12, Glandular oesophagus 19, Intestine 11. Tail width 10, Anus from tip 44.
Data for Large form (Mean values):
Body length 1968, Oral cavity 87, Total oesophagus 778, Muscular oesophagus 153, Nerve ring 129. Body widths: Mouth 14, Head 26, Nerve ring 53, Junction 58. Internal widths: Muscular oesophagus 20, Glandular oesophagus 32, Intestine 20. Tail width 12, Anus from tip 63.
Standard deviation and sample size (N) are provided for all metrics, showing higher variability in the Large form’s total oesophagus length (Std Dev 84) compared to the Small form (Std Dev 25).
Notes: Junction – body width at intersection of oesophagus and intestine; Std Dev – Standard deviation.
The gut forms of the small Habronema were photographed from several R. dutoiti (Figure 4). They were in advanced development even in flies with no ovarian development and were grossly similar to the small forms in the proboscis. Most were free, except for one coiled, encysted parasite. Up to four parasites were found in a fly.
Gut forms of the small Habronema. A. Head 1000×; B. Tail 1000×; C. Encapsulated filaria 250×; D. Anterior 250×; E. Free whole filaria 100×.
The single nematode found in one Musca was lightly striated and had an undifferentiated oesophagus and prominent lips (not photographed). It had a tapered tail without the typical bulbous tip and terminal papillae found in Habronema. Prominent round anal glands were visible internally. The largest filaria differentiated as ‘nematode x’ in R. dutoiti was represented by 20 specimens in one fly. It had lateral cephalic alae (Figure 5) similar to those illustrated for abnormal Habronema muscae (Rakhshandehroo et al. Reference Rakhshandehroo, Sharifiyazdi, Shayegh and Ahmadi2014 ) and a thin, coiled oesophagus; hence, its identity is not clear. It had prominent cuticular striations and a thinner, tapered tail than the common large form of Habronema. Body size and proportions were otherwise similar (Figure 1, N = 5).
Very large nematode × with lateral cephalic alae (arrow) and a thin coiled oesophagus 1000×.
Sectioning of the proboscis of R. dutoiti yielded one image with a good cross-section of Habronema (size not recorded). Parasites were identifiable through basophilic staining and a thick eosinophilic cuticle (Figure 6). The cuticle was only weakly striated and was underlain by eosinophilic cells. Other internal structures were grossly similar to the few cross sections available for larval Habronema (Schuster and Sivakumar Reference Schuster and Sivakumar2017; Waddell Reference Waddell1969).
Tissue cross-section of Habronema in the proboscis with an oblique transverse section to the left 1000×.
Scanning electron microscopy
SEM images were obtained only for the small form of Habronema (Figure 7) due to unforeseen circumstances. The small papillae on the bulbous tip of the tail were obvious. Four large cephalic papillae were also clearly visible. We were unable to obtain an en face view of the anterior end for a detailed view of the head. Parasites had clear striations mid-body, about 3 μm wide, with less conspicuous striations elsewhere.
SEM images of the small Habronema in the proboscis of Rhinomusca dutoiti. A. Tail showing bulbous tip with terminal papillae and bulge at the anus; B. Head showing cephalic papillae, longitudinal grooves and light striations; C. Closer view of head; D. Mid body showing prominent striations and longitudinal grooves; E. Whole parasite with ring-like orifice near anterior in bottom left; F. Head and tail showing variation in cuticle.
Figure 7. Long description
A collection of six S E M micrographs labeled A through F.
Panel A shows the tail of the parasite with a bulbous tip featuring terminal papillae and a distinct bulge at the anus. A 5 micrometer scale bar is at the bottom.
Panel B displays the head at 1,500 times magnification, revealing cephalic papillae, longitudinal grooves, and light striations. A 10 micrometer scale bar is present.
Panel C provides a closer view of the head at 3,500 times magnification, emphasizing the texture of the cephalic region. A 5 micrometer scale bar is present.
Panel D focuses on the mid-body, showing prominent transverse striations and deep longitudinal grooves. A 10 micrometer scale bar is present.
Panel E shows the whole parasite at 350 times magnification. A white arrow points to a ring-like orifice near the anterior end in the bottom left. A 50 micrometer scale bar is present.
Panel F compares the head and tail at 500 times magnification, highlighting variations in the cuticle texture between the two ends. A 50 micrometer scale bar is present.
Each panel includes technical data at the bottom, such as 15 K V voltage and specific specimen identification numbers.
Multiple longitudinal grooves were pronounced at the anterior end and were reduced at the posterior end. Parasites also had a deep longitudinal groove running from about the intestinal junction to the tail. The overall shape was more oval than round as in the tissue cross section (Figure 6). A ring-like orifice was found near the cephalic end in two of four parasites with SEM images. That structure has not been documented in larval filaria. It was at about the right location for the vulva in an adult worm. We are, however, unaware of any sexual differentiation occurring in larval forms. Considerable shrinkage was evident in fixed specimens prepared for SEM, as they were much shorter than live specimens (585, 622, 630, 795 μm).
Sequencing of Rhinomusca DNA
In the 798 base pair fragment of the COI gene sequenced only two base pairs differed between the two flies from Kenya and South Africa (Supplemental material S1). Namely, at position 1050, relative to the reference COI sequence of Drosophila yakuba Hans Burla, 1954 (GenBank NC_001322; Clary & Wolsthenholme, Reference Clary and Wolsthenholme1985), R. dutoiti from Kenya has a thymine (T) vs a cytosine (C) in the specimen from South Africa. At position 1186, the Kenyan sample had a cytosine (C) versus a thymine (T) in the specimen from South Africa (Supplemental material S2, GenBank accession numbers PZ733943 and PZ733944). Both of these nucleotide base changes are synonymous substitutions and do not translate into a change in amino acids (Supplemental material S3).
A search for additional COI sequences for R. dutoiti in GenBank and BOLD yielded 30 sequences from specimens collected in South Africa from which the 5′ end of the gene representing the ‘barcode’ region of COI had been amplified. There were no base pair differences between the sequences except for a single base difference in sample KMPXN1413-19.COI-5P, which is likely to be a sequencing error (Supplemental material S4). Dorsal and ventral views of fresh specimens of R. dutoiti representative of the fly from Kenya used for COI sequencing are provided in Supplemental material S5 and S6. Metadata for all specimens with sequences are provided in Excel format in Supplemental material 3.
Identification of blood meals
Only one of twelve blood meals from R. dutoiti yielded a positive identification as ‘unidentified Bovidae’. Two meals may have come from black rhino, but results were below the cut-off criterion for the ELISA test (0.21 and 0.27 vs a 0.30 threshold; positive control sera were 0.40–0.60). Rhinomusca dutoiti fed on blood-soaked cotton, but no attempt was made to establish a colony. In assays for fructose, R. dutoiti had only trace levels of fructose in two of 19 flies (10%).
Identity of Rhinomusca
We have retained the name R. dutoiti for our specimens based on the two simple characters provided in Zumpt (Reference Zumpt1973) for ‘easy separation’ relative to R. brucei Malloch, 1932 (mainly black legs, darker body, Figure 8). Six males and seven females (pinned specimens) caught in Kenya in the 1990s were re-examined in 2026 for the characters mentioned for identifying the proposed new species R. alaseta by Zielke (Reference Zielke2022). Our specimens had the ‘new’ characters mentioned by Zielke (Reference Zielke2022) but otherwise matched R. dutoiti. Zielke (Reference Zielke2022) assumed these characters were missed by Zumpt (Reference Zumpt1973) in his species description, but we believe they may have simply not been mentioned given the prominent colour differences between R. dutoiti and R. brucei. Zielke’s special characters also present in our Kenyan specimens were the white band on the first femur in side view (Figure 9), small setulae on the ventral surface of the wing by the convergence of R1 and Sc, and areas without microtrichia on the wings (Figure 10). Frons proportions in our flies (Figure 9) were as described by Zumpt (Reference Zumpt1973).
Rhinomusca dutoiti males and females from Nairobi National Park, Kenya. Fresh specimens photographed after brief freezing in the 1990s with a Nikon macro lens and ring flash on Kodachrome ASA 64 film.
Selected views of pinned specimens of Rhinomusca dutoiti from Nairobi National Park, Kenya, in the 1990s, photographed in 2026 through a Leica stereomicroscope under daylight LED illumination with a Canon digital pocket camera. A. Contrasting white side of first femur in a female (arrow); B. Wide female frons; C. Narrow male frons.
Wing of Rhinomusca dutoiti. A. Overall dorsal view of wing of a female showing pattern of areas with and without microtrichia; B. Close-up ventral view of wing showing two small setulae near convergence of veins R1 and Sc (arrow), many conspicuous setulae along vein R4+5, and distinct areas with microtrichia (m).
We were able to confirm our observations for two specimens of R. dutoiti from South Africa collected in the 1950s, one of which was a paratype collected by Zumpt (Figures 11 and 12). Each specimen had a white side on the fore femur, either one or two small setulae near the convergence of R1 and Sc on the ventral surface of the wing, and areas without microtrichia as mentioned by Zielke (Reference Zielke2022). Photographs of a few R. dutoiti from KwaZulu-Natal from other museums taken recently by colleagues (Pietermaritzburg, Onderstepoort, South Africa), and a fly photographed at Skukuza in 2024, also all had a fore femur with a distinct white side (close-ups of the wings were not available).
Selected views of a Rhinomusca dutoiti paratype from the Florida museum collected in 1950 by Zumpt from KwaZulu-Natal in South Africa. A. Dorsal view of whole fly; B. Side view of head showing white side of fore femur; C. Dorsal view of wing showing pattern of microtrichia; D. Close-up ventral view of wing near convergence of veins R1 and Sc with one small setula on R1; E. Museum label.
Figure 11. Long description
The composite contains five panels labeled A through E.
Panel A: A dorsal view of the entire fly specimen. A vertical scale bar to the left indicates 3 mm. The fly has a grey thorax with dark longitudinal stripes and translucent wings spread outward.
Panel B: A lateral close-up of the head and anterior thorax. The large compound eye is dark brown. Below the head, the fore femur is visible, showing a distinct white coloration on its side.
Panel C: A dorsal view of a single wing. The wing is translucent with a network of dark veins. The surface is covered in a fine pattern of microtrichia, appearing as tiny dark specks.
Panel D: A high-magnification ventral view of the wing edge where veins R 1 and S c converge. A single small hair-like setula is visible on the R 1 vein.
Panel E: Three stacked museum labels. The top label reads Mtubatuba, Zululand, leg. Zumpt, 15.1.50. The middle label is yellow and reads Paratypoid. The bottom label reads Rhinomusca male symbol, dutoiti. n. sp., Zumpt. 1950.
Selected views of a Rhinomusca dutoiti from the Florida Museum collected in 1954 by Paterson from KwaZulu-Natal in South Africa. A. Dorsal view of whole fly; B. Side view of head showing white side of fore femur; C. Dorsal view of wing showing pattern of microtrichia; D. Close-up ventral view of wing near convergence of veins R1 and Sc showing two distinct small setulae; E. Museum label.
Figure 12. Long description
The composite contains five panels labeled A through E.
* Panel A: A dorsal photo of the entire fly specimen. A vertical scale bar on the left indicates 7 mm. The fly has a grey thorax with dark longitudinal stripes and translucent wings folded over its abdomen.
* Panel B: A lateral close-up photo of the head and anterior thorax. It shows a large reddish-brown compound eye and a pale, white-colored fore femur extending downward.
* Panel C: A dorsal photo of a single wing. The wing is translucent with a network of brown veins and a fine pattern of microtrichia across the membrane.
* Panel D: A high-magnification ventral photo of the wing base. Near the convergence of veins R 1 and S c, two distinct small black setulae are visible as dark dots against the membrane.
* Panel E: Two rectangular museum labels. The top label is printed and reads Hluhluwe Natal 3 I I I 54. The bottom label is handwritten and reads Rhinomusca female symbol dutoiti Zpt det Paterson 54.
Association of Rhinomusca with rhinoceros
Large numbers of medium-sized flies assumed to be R. dutoiti were observed on a captive black rhino at Solio (Figure 13). Clouds of flies were also sometimes observed on wild rhinos; many flies unfortunately departed quickly on arrival at sedated animals after darting. Vavoua traps baited with 1-octen-3-ol at Solio caught up to 51 R. dutoiti per day. In contrast, catches at Nairobi were rarely more than a few per day (when only this bait was used). High numbers of R. dutoiti at Solio were also noticeable in terms of flies chasing cars, something not seen elsewhere. Carbon dioxide as a gas was also released during trapping in two lowland areas (Nguruman Acacia woodland and Shimba Hills coastal forest), but R. dutoiti was not caught.
Photographs of Stomoxyinae closely associated with wildlife. A. Rhinomusca dutoiti on the flanks of a captive black rhino at Solio Ranch in 1994; B. Rhinomusca dutoiti on a wild white rhino at Hluhluwe-iMfolozi in 1985; C. Absence of flies on captive white rhinos in Kruger National Park in 2024; D. Haematobia on the backs of wild lions at Kruger National Park. Identities of flies are assumed from size and ecology.
Figure 13. Long description
The four panels are arranged in a two-by-two grid.
* Panel A, top-left: A side-profile photo of a captive black rhino at Solio Ranch. Its grey flanks are heavily speckled with hundreds of small dark spots representing Rhinomusca dutoiti flies.
* Panel B, top-right: A close-up, front-facing photo of a wild white rhino at Hluhluwe-iMfolozi. Numerous small dark flies are clustered on its face and the side of its head.
* Panel C, bottom-left: A photo of two captive white rhinos standing in an enclosure at Kruger National Park. Their skin appears smooth and uniform, showing a total absence of fly clusters.
* Panel D, bottom-right: A photo of two wild lions lying down at Kruger National Park. One lion is seen from the back and the other is lying on its side. Their golden fur is covered in dense patches of small dark Haematobia flies, particularly concentrated along the spine and upper back.
At Kruger National Park, animal keepers mentioned that clouds of flies have not been noticeable in recent years on either captive or wild rhinos. Almost no flies of any kind were seen during daily visits to observe two captive white rhinos over a week in 2024 (Figure 13). Catches of biting flies other than Stomoxys spp. in traps were low at this time. The few R. dutoiti captured at Skukuza were caught next to captive buffalo. In contrast, small flies (presumed to be Haematobia) were abundant on buffalo (Supplemental material 2), as noted in the literature (Fay and Doube Reference Fay and Doube1987), but were absent in traps, including those set next to a boma with captive buffalo. A colleague captured about 150 Haematobia [H. meriadani Zumpt, Reference Zumpt1973, H. thirouxi (Roubaud, 1906), H. spinigera] by hand net from a passing wild buffalo. Some lions were also covered in small flies, likely Haematobia given their size and position (Figure 11). Only 270 Sphaeroceridae emerged from two netting cages placed on about 0.5 m2 of the white rhino waste piles for one day.
Discussion
The common larval filaria found in R. dutoiti matched the genus Habronema in morphology and in development, with mature larvae present in recently emerged flies. Cephalic features were most similar to SEM images of adult H. microstoma (Schneider, 1866) (Naem Reference Naem2007). Habronema microstoma is found in Stomoxys calcitrans. The more common Habronema muscae (Carter, 1861) is found in Musca domestica (Traversa et al. Reference Traversa, Otranto, Iorio, Carluccio, Contri, Paoletti, Bartolini and Giangaspero2008). Prominent longitudinal grooves in our specimens were similar to those in SEM images of larval H. muscae (Buzzell et al. Reference Buzzell, Tariq, Traversa and Schuster2011; Schuster and Sivakumar Reference Schuster and Sivakumar2013). The degree of striation of the cuticle varied among morphologically similar specimens, and hence we are uncertain about assigning significance to this feature. A smooth cuticle is typical of Draschia megastoma (Rudolphi, 1819) (formerly H. megastoma), whereas a striated cuticle is typical of H. muscae (Waddell Reference Waddell1969). The presence of gut forms was consistent with the overall life cycle of Habronema (Amado et al. Reference Amado, Silveira, Vieira and Traversa2014; Ransom Reference Ransom1913). However, it would be inappropriate to assign species names to our findings given the lack of information on larvae from vectors in areas with diverse wildlife. Our main collection in Nairobi National Park was from an area where domestic livestock were absent.
Although having features in common with Habronema spp. from domestic equines, the parasites in R. dutoiti were distinctly shorter than H. muscae L3 larvae in M. domestica (range 1701–2207 vs 2274–3527 μm) (Amado et al. Reference Amado, Silveira, Vieira and Traversa2014). Larvae of H. muscae and D. megastoma in M. domestica have been measured at between 2,850 and 3,100 μm, but this information cannot be used to differentiate these two species in the vector (Schuster and Sivakumar Reference Schuster and Sivakumar2013). We have been unable to find similar measurements of L3 larvae for H. microstoma and H. majus (Creplin, 1849) (Jian et al. Reference Jian, Wang, W-X and Zhang2017). The species of Habronema have been defined from the morphology of adult worms, with much less information on larval forms.
Domestic equine hosts of Habronema are not present in rhino conservation areas in Africa. In periurban areas to the north and west of Nairobi National Park, small numbers of domestic horses were present, but only 5–10 km distant from trapping sites. Zebras were present within a few km of traps and were common in the open grasslands. However, the two species were never observed in close proximity (black rhinos are browsers and zebras are grazers). The only area in Kenya where the two species may co-exist at high numbers is the Lewa Wildlife Conservancy. Entomological and parasitological studies would be worth pursuing at Lewa as it now contains many black and white rhinos as well as Burchell’s and Grevy’s zebra. Zebras harbour unique Habronema species, but not in large numbers (Krecek Reference Krecek1989; Wambwa et al. Reference Wambwa, Ogara and Mudakha2004). However, no blood meals were identified as zebra in R. dutoiti, or in Stomoxys where more host information was available. Researchers often study tabanids biting zebras but have never commented on harassment by smaller biting flies (Caro et al. Reference Caro, Argueta, Briolat, Bruggink, Kasprowsky, Lake, Mitchell, Richardson and How2019).
Habronema is not a common nematode in rhino (Knapp et al. Reference Knapp, Krecek, Horak and Penzhorn1997; Penzhorn et al. Reference Penzhorn, Krecek, Horak, Verster, Walker, Boomker, Knapp and Quandt1994). This is puzzling given past observations of Rhinomusca associated with rhino, and the high prevalence of two forms of Habronema in Rhinomusca. The original description of two adult worms of Habronema khalili in a captive black rhino is not a good match for the larval forms we found (Ezzat Reference Ezzat1945). There is also a report of conjunctival habronemiasis in a white rhino, inferred from histological sections, but it was not possible to identify the filaria responsible (Horowitz et al. Reference Horowitz, Dubielzig, Botero-Anug, Lucio-Forster, Bowman, Rosenzweig, Frenkel and Ofri2016). Larval forms in R. dutoiti also did not match the morphology of adult Parabronema spp. in rhino (Fitzsimmons Reference Fitzsimmons1962; Junker et al. Reference Junker, Horak and Penzhorn2015) or the larval Parabronema species in Haematobia that is transmitted to camels in China (Liu et al. Reference Liu, Shi, Luo, Li, Wang, Yang, Wang, Zhai, X-Y and Yang2021).
Biology of Rhinomusca
The only form of Rhinomusca that we have trapped has matched the original description of R. dutoiti (Zumpt Reference Zumpt1950), despite the fact that R. brucei was discovered in Kenya in the 1930s. Rhinomusca was abundant on rhinos in Tsavo in the 1960s, but it is difficult to determine if the flies caught by Parsons and Sheldrick (Reference Parsons and Sheldrick1964) were in fact R. brucei based on the drawing in their paper. Zumpt (Reference Zumpt1973) noted that the few specimens of R. brucei available for taxonomic comparisons were not in good shape; hence he had ‘doubts as to whether a specific separation is justified’ for R. dutoiti / R. brucei.
Our COI sequencing data indicates that the COI gene does not have sufficient variation to differentiate populations of R. dutoiti. Populations as distant as those sampled in Kenya and South Africa differed by only two base pairs. These could possibly be used as markers for future work, but additional samples would need to be tested to determine if these differences are consistent among populations. Another 30 flies in GenBank and BOLD sequenced from South Africa indicate that the 5′ end of the COI gene does not differentiate individuals within that population. A more informative means to differentiate R. dutoiti populations in the future would be via the use of single nucleotide polymorphisms (SNPs) at the nuclear genome level.
In addition to any future sampling of Rhinomusca caught in traps or emerging from dung, the chasing of cars could provide a practical sampling method using sticky targets (Muyobela et al. Reference Muyobela, Pirk, Yusuf, Mbewe and Sole2021). Hand net catches of Rhinomusca from rhinos would be the ideal method for obtaining definitive biological insights, but this is difficult for wild animals. For tame animals in bomas, the use of walk-through ‘traps’ designed for Haematobia may also provide a way of catching flies without any sampling bias introduced during hand net catches (Rodriguez-Vivas et al. Reference Rodriguez-Vivas, Cutolo, ATMd, Cuore, Molento, López-Osorio, Rodrigues, Spina, Borges, Lopes, Pulido-Medellin, Fiel, Costa-Junior, Anziani, Marques San Martín and Sabatini2026).
Historically, flies assumed to be R. dutoiti were abundant on rhinos at the type locality of this species, Hluhluwe-iMfolozi Park (Figure 13); the current situation in KwaZulu-Natal is undocumented. Zumpt (Reference Zumpt1950) commented that R. dutoiti were abundant at Imfolozi around white rhinos, but not around black rhinos at Hluhluwe. The presence of clouds of flies on rhinos is sometimes mentioned in the literature (Shrader et al. Reference Shrader, Emslie, Hillman-Smith, Kretzschmar, Marneweck, Melletti, Owen-Smith, Rookmaaker, Slater, Melletti, Talukdar and Balfour2025), but most researchers have not identified the flies seen. Our results for blood and nectar feeding, along with the ages of females determined by ovarian development, suggest that R. dutoiti may spend part of its life on other hosts. This is also supported by catches of R. dutoiti in biting fly traps and its tendency to chase cars. In contrast, Haematobia is rarely caught in traps and only seems to leave to oviposit or when the host enters or passes through a very dark area. Haematobia emerging from cattle or buffalo dung presumably are able to find hosts rapidly within a herd context, whereas Rhinomusca may have to search for hosts more widely distributed in the environment (e.g. black rhino density at Nairobi National Park is about one rhino per two km2).
Rhinomusca dutoiti in South Africa occurs in diverse habitats in South Africa, and hence any relationships with the habitats sampled in Kenya may simply be related to the local activity patterns of rhino. In a matched subset of trapping with dry ice with or without 1-3-octen-ol (34 Vavoua trap-days per habitat during November–December, 1993), total catches of R. dutoiti were much higher in the riverine versus forest habitat (296 vs 41). Rhinos were observed in all areas while travelling 50 km each day across the park, but we are unaware of any quantitative surveys of rhino habitat use in the park.
Although we never found R. dutoiti in rhino dung, Parsons and Sheldrick (Reference Parsons and Sheldrick1964) documented a close relationship between Rhinomusca and dung at Tsavo in the 1960s. They observed flies presumably depositing eggs on dung and reared adult Rhinomusca from larvae and pupae in the moist interior of middens. They were able to maintain two male flies with daily feeding on blood till death at 17 or 19 days but were less successful in maintaining females. At high air temperatures in Voi, development from egg to emergence was 17–20 days with three larval stages, with a longer development time in Nairobi at cooler temperatures. They found one fly with 160 mature eggs and with undeveloped follicles, suggesting that a second batch of eggs may be produced. Unfortunately, they never attempted to rear eggs on any artificial medium.
Zumpt (Reference Zumpt1950, Reference Zumpt1973) also commented briefly on the association between Rhinomusca and rhino dung. We recently found several pinned specimens of R. dutoiti from Hluhlwe in the 1970s in the ARC-Onderstepoort Veterinary Research Campus collection labelled as coming from rhino dung. Hence, our inability to find any R. dutoiti emerging from black rhino dung in Kenya is puzzling relative to the historical record. If this species was breeding in rhino dung, then we should also have detected it in traps set next to translocated rhinos kept in bomas for long periods after translocation.
Rhino dung is a suitable substrate for rearing Musca domestica and Stomoxys calcitrans (Wright et al. Reference Wright, Oehler and Johnson1975). A few species of wild Stomoxyinae have now been successfully reared (Muita et al. Reference Muita, Bargul, Makwatta, Ngatia, Tawich, Masiga and Getahun2025; Sumba et al. Reference Sumba, Mihok and Oyieke1998), so it should be straightforward to follow up on our studies with life cycle studies using laboratory-bred flies to validate our assumption that the filaria seen in R. dutoiti are transmissible to host(s). Given the many translocations of rhinos over the last 30 years, it also seems likely that R. dutoiti could now be present in lowland sanctuaries, and this can also facilitate further work. Lastly, the availability of PCR techniques for identifying filariae in vectors (Lui et al. Reference Lui, Kulpa, Verocai, Armién, Edwards, Wiener and Rech2023; Traversa et al. Reference Traversa, Otranto, Iorio, Carluccio, Contri, Paoletti, Bartolini and Giangaspero2008) now provides a more efficient way to detect and confirm the identity of the parasites found.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0022149X26102144.
Acknowledgements
The authors thank K. McHugh for editorial assistance, the Kenya Wildlife Service (KWS) and South Africa National Parks for permission to work in conservation areas, and M. Taoie, K. Williams, and A. Pont for assistance with Rhinomusca taxonomy, and David Cooper for collection of a R. dutoiti from South Africa. Felipe Soto-Adames and Andrew Levy at the Florida State Collection of Arthropods, Florida Department of Agriculture and Consumer Services, provided excellent photographs of museum specimens of R. dutoiti. This study was primarily funded by a European Union grant to the International Centre of Insect Physiology and Ecology while the first author was a research scientist with ICIPE. Rodrigo Krüger is supported by a research fellowship from the National Council for Scientific and Technological Development (CNPq – process no. 300899/2025-4). J. Murithi and P. Lisamula prepared histological and SEM material. R. Giordano sequenced and analysed mitochondrial DNA data from Rhinomusca. J. Esterhuizen, M. Taioe, and K. Labuschagne participated in fieldwork in South Africa. Numerous KWS and ICIPE staff provided field and logistical support in Kenya. This is contribution #2172 from the Institute of Environment at Florida International University, Miami, FL.
Competing interests
The authors declare no competing interests.
