Anopheles (Cellia) garnhami Edwards

Harbach, Ralph E. & Wilkerson, Richard C., 2023, The insupportable validity of mosquito subspecies (Diptera: Culicidae) and their exclusion from culicid classification, Zootaxa 5303 (1), pp. 1-184 : 64-67

Anopheles (Cellia) garnhami Edwards

subspecies basilewskyi Leleup, 1957 —original combination: Anopheles garnhami sbsp. basilewskyi . Distribution: Tanzania (Mount Meru) ( Leleup 1957).

Subspecies garnhami Edwards, 1930 —original combination: Anopheles garnhami . Distribution: Burundi, Democratic Republic of the Congo (eastern), Eritrea, Ethiopia, Kenya, Lesotho, Malawi, Mozambique, Rwanda, Somalia, South Africa, South Sudan (southeastern), Tanzania, Uganda, Zimbabwe (derived from Gillies & de Meillon 1968, fig. 52; Somalia from Irish et al. 2020).

Edwards (1930) described the nominotypical subspecies based on specimens from Kenya: Kericho, 6,000 ft. (2♂♂, 3 ♀♀) ; Saiwa (1 ♀) ; Londiani , Kenya, 7,500 ft. (1 ♀) ; Uasin Gishu, 6,000 ft. (11 ♀♀) ; locality not specified (3♂♂, 3 ♀♀), and additionally from Karambo , 6,000 ft. (1 ♀) , a village in Rwanda. Townsend (1990) found 22 of the above 25 specimens in the Natural History Museum, London , and referred to them as “ syntypes ”. Edwards, however, clearly referred to the specimens from Kericho as including the type specimens, but did not state what sort of specimens, which leaves designation of a lectotype, if needed, for the future.

As a basis for comparison to the forms treated here, we quote portions of Edwards’s (1930) original description of garnhami based on the specimens from the localities listed above. All are from highland locations in the general vicinity of Lake Victoria.

♀. Head with the usual [?] dense upright scales, pale yellowish in middle, black at sides; frontal tuft long and pure white.... Palpi... [with] scales on first segment shaggy; a very narrow white ring at tip of first segment [palpomere], a somewhat broader one embracing tip of second and base of third, tip of third and whole of fourth white; third and fourth segments [palpomeres] together about equalling third in length, fourth short, scarcely more than one-third as long as third [i.e. maxillary palpus with three white rings]….

Thorax as seen from above greyish in middle (for about one-third of its width), sides of mesonotum [scutum] very broadly dark brown, but extreme margins again narrowly grey; a small dark brown area immediately in front of scutellum in middle. Median greyish area of mesonotum densely covered with narrow scales (about 6–8 times as long as broad), mostly creamy-white in colour....

Abdomen brownish above, lateral and posterior margins of tergites [terga] dark brown; sternites [sterna] with large and rather conspicuous whitish-grey basal lateral patches. No scales; hairs [setae] pale.

Legs blackish; coxae and trochanters pale; tips of femora and tibiae and of first two or three tarsal segments [tarsomeres] of all legs very narrowly creamy-white.

Wings.... Costa mainly black, with three or four yellowish spots; first very small (sometimes scarcely distinguishable), placed just before base of cubital fork [sector pale and accessory sector pale spots joined on costa (C), subcosta (Sc) and vein R 1]; second larger and always distinct, above stem of radial fork [subcostal pale spot on C, Sc and R 1]; third small (often absent), above middle of radial fork [preapical pale spot on C and R 1]; fourth small, close to tip [apical pale spot at ends of veins R 1 and R 2]. First vein [R] pale at base, and with three pale yellowish spots corresponding with those on costa, otherwise black. Radial fork [veins R 2 and R 3] mainly black. Third vein [R 4+5] pale except narrowly at base and tip. No dark area on fifth vein [CuA] at base of fork [mcu]. Sixth vein [1A] with three dark areas, basal one short (sometimes absent). Small fringe-spots present opposite tips of all veins except sixth [pale fringe spot also illustrated at the middle of cell 1A]. Base of radial fork [R s?] much proximal to that of the median. Knob [capitellum] of halteres black....

♂. …club [fused setae on dorsal lobe of the claspette] long and rather narrow; claspette with long apical hair [seta], but apparently without accessory hair; innermost of the four spines [parabasal setae] of the coxite [gonocoxite] stout on basal half, very slender and curved on apical half.

A great deal of variability is reported for “typical” garnhami from throughout its considerable range (approximately 4,800 km). Examples follow.

Evans (1938) separated garnhami into “typical” and “atypical” forms and Gillies & de Meillon (1968) resolved garnhami in two places in their keys to adult females (Sections VIII and X).

Apparently, there is bionomical variability of possible consequence. Horsfall (1955) noted: “ De Meillon (1947 a) [1947 herein] noted that variations in reported sites appear in the literature. The form in southern Africa does not inhabit houses, and, similarly, D. B. Wilson (1938) [not seen by us] regards the form in Tanganyika [ Tanzania] as no inhabitant of houses. Symes (1931) [not seen by us], on the other hand, found large numbers of adults in huts in Kenya where goats as well as humans were housed.”

Corradetti (1940) described variation of the nominotypical subspecies throughout its range in Africa to include specimens from Dessiè , Ethiopia, which is approximately 400 km north of the capital Addis Ababa and about 1,300 km north of the type locality of garnhami . He explained the variation as follows (translated from the Italian, some intervening text omitted) .

During my research on malaria in Abissinia [ Ethiopia] I had occasion to observe An. garnhami in Dessiè and in some surrounding areas around this city. The study of the Ethiopian material caused me to observe that the variability of some characters is considerably wider than it had previously been described.

[Heading] Morphological variability of A. garnhami in the Dessiè area.

Adult. ...In the wings the sector pale spot is complete or incomplete (limited to the subcosta and the 1st longitudinal vein [Sc, vein R] and totally absent on the costa). The subapical dark spot may be complete, incomplete or absent [reflects presence and/or size of preapical pale spot]....

Pupa. Considerable variability is observed in the bristles B [seta 5] of segments V–VII, they may appear simple [single], with a small lateral branch, or divided into various branches. Consistency may not be observed in this respect in the various segments of the same nymph [pupa] and sometimes not even in the two halves of the same segment....

[Heading] Considerations on the systematic value of the morphological variants of A. garnhami .

From the study of the literature hitherto known on this anopheline it is evident that the limits of variability of the characters presented by the species appear markedly different in different parts of Africa. Knowledge in this regard can be summarized as follows:

1) Kenyan material (typical). Wing length mm. 5–5.5 (5.5– 6 in Uasin Gishu). Female palpi with 3 spots. Strong variability in the wing spots, and absence of basal spots on the costa. Bristle B [seta 5] of the pupa on segments VI–VII with 3–7 branches. Absence of small accessory tergal plates on the abdominal segments of the larva.

2). Material from Uganda (Mt. Elgon). Female palpi with 4 spots. Strong variability in wing spots and absence of basal spots on the costa. Bristle B [seta 5] of the pupa on segments V–VII sometimes simple [single]. Absence of small accessory tergal plates on the abdominal segments of the larva.

3) Variety walshi ( South Africa). Wing length about 4 mm. Female palpi with 4 spots. Very little variability in the wing spots: little tendency for sector [pale] and subapical [pale] spots to disappear; presence of basal spots on the costa. Marked tendency for the development of small accessory tergal plates on segments VI–VII of the larva.

4) Dessiè material. Wing length: males mm. 3.4‒4.1; female mm. 3.7–4.5. Female palpi with 3 spots. Strong variability in the wing spots and absence of basal [pale] spots on the costal vein. Strong variability in the characters of the bristle B [seta 5] of the pupa. Presence of small plates on the metathorax of the larva. Strong variability regarding the presence or absence of accessory tergal plates on abdominal segments of the larva from [segment] IV onwards.

From these observations [there is] the need to extend the research on A. garnhami of the various African regions in order to determine the limits of variability of the species in the various zones from the study of sufficient material from each locality.

Subspecies basilewskyi was described by Leleup (1957) from a subalpine prairie characterized by the monotypic plant genus Hagenia [ Hagenia abyssinica (Bruce) , which is found at high elevations in Afromontane regions in East Africa]. The collection site was a marsh at 2,900 m elevation on a vegetated volcanic mountain (Mount Meru) isolated by a large arid area. Leleup (1960) re-published the brief 1957 description and provided further detail and illustrations of a wing, dorsal lobe of the claspette and aedeagal leaflets. The type series consists of a female holotype, a male allotype and seven female and three male paratypes. The author was unable to preserve immature stages for study. Leleup reported several unambiguous diagnostic characters that distinguish basilewskyi from the nominotypical form: Wings lack pale fringe spots, except for the pale scales of the apical pale spot, and pale spots at the bifurcation of R 2 and R 3 and at the base of M 3+4 (crossvein mcu) (pale spots present in the nominotypical form); the dorsal lobe of the claspette with a distinctly broadened club with a shorter accessory seta (longer narrow club with a longer accessory seta in nominotypical garnhami ); aedeagal leaflets about equal in number but thinner with less serration in basilewskyi . Leleup (1960) summarized the purported differences by stating: “All zoologists will agree in admitting that when all the individuals of an isolated colony offer common characteristics differentiating them from the typical form, it is indeed a subspecies and not of a simple variety [translated from the French].” Stone et al. (1959) and Lips (1960) also treated basilewskyi as a subspecies but Gillies & de Meillon (1968) considered variation throughout the range of garnhami and concluded that “…it appears impossible to define basilewskyi in any satisfactory way, and it must be regarded simply as a variant at below the subspecific level.” Given the inconclusive and poorly documented variation of garnhami over such a large range, we think Gilles & de Meillon’s decision to synonymize basilewskyi with garnhami was incorrect. Overlapping characters are not a valid criterion to combine different forms. This opinion was shared by Brunhes et al. (1998) who resurrected basilewskyi from synonymy with garnhami , stating that “The geographical isolation of this population, the absence of apparent sympatry, with An. garnhami , and the concern not to attribute to An. garnhami too large a set of characters [great variability], lead us to consider An. g. basilewskyi as a valid taxon. It has been very frequently argued that mountain species are genetically isolated from each other, as are island populations. The gene flow these populations receive is probably much reduced, which causes faster speciation. The many morphological variations observed in the orophilic [thriving in mountainous or subalpine regions] Anopheles would, in this hypothesis, be the consequence of a speciation in progress or already carried out [emphasis ours]. As Corradetti (1940) suggested, all available information should be re-examined on orophilic species and in particular for An. garnhami [translated from the French].”

We also consider the distinct differences found for basilewskyi , along with its apparent isolation on a volcanic mountain surrounded by a large arid area, to provide more than sufficient evidence of an independently evolving species and therefore accord basilewskyi species status: Anopheles (Cellia) basilewskyi Leleup, 1957 . Anopheles basilewskyi is currently listed as a species in the Encyclopedia of Life.

Anopheles garnhami has one synonym, An. garnhami var. walshi Evans & de Meillon, 1933 . From its original description, it is “Only known up to the present from Drakensberg, 3,000 to 4,000 ft., near Tzaneen, Transvaal, and Tzaneen itself, 2,380 ft., 15 miles from the mountains, where a few specimens have been collected. Cotype ♂♂ and ♀♀ from Magoeba’s Kloof, Tzaneen, North Transvaal...”.

Evans & de Meillon (1933) described and contrasted walshi to the “ type form” of garnhami . However, some illustrated wing spot characters of their “ type form” do not entirely agree with the original description of Edwards (1930). Present on the South African “ type form”, not found on nominotypical garnhami , are a discrete accessory sector pale spot (only on vein R); base of M 3+4 (crossvein mcu) dark, as in basilewskyi ; pale fringe spot on the border of cell 1A not depicted. Variety walshi was described as differing from the South African “ type form” by having prehumeral and humeral pale spots and a discrete accessory sector pale spot on vein R. Subsequently, de Meillon (1947), while still retaining the validity of variety walshi , stated: “Resembles garnhami very closely and it is doubtful if it deserves even varietal rank. Re-examination of a series of cotypes leads me to believe that it cannot be separated from the type form with certainty.” He, however, then listed distinguishing characters. Stone et al. (1959) listed walshi as a synonym of garnhami but Lips (1960) treated it as a variety. Gillies & de Meillon (1968) stated: “As pointed out by De Meillon (1947), garnhami var. walshi grades into typical garnhami in all characters and cannot be separated with any certainty. It is accordingly regarded here as falling within the limits of variation of the normal form.”

We think the status of walshi and the “ type form” in southern Africa merit further examination since walshi is inadequately compared to “normal” garnhami , which has not been characterized throughout its range. Synonymous variety walshi and nominotypical garnhami have both been reported from the Drakensburg Mountains in Transvaal, South Africa. We note that the high elevation habitat typical of nominotypical garnhami is disjunctive with garnhami and the purported synonymous walshi from South Africa found in a mountain range a very great distance from the Kenyan highlands. This suggests to us the probability that a species complex related to garnhami exists in East Africa south of Ethiopia.