Ferrisia virgata (Cockerell)
treatment provided by
|Ferrisia virgata (Cockerell)|
( Figs 2G, 21)
Dactylopius virgatus var. farinosus Cockerell, 1893a: 178 . Synonymy by Ferris, 1950: 93.
Dactylopius virgatus var. humilis Cockerell, 1893a: 179 . Synonymy by Ferris, 1950: 93.
Dactylopius segregatus Cockerell, 1893b: 254 . Synonymy by Williams & Granara de Willink, 1992: 183.
Dactylopius ceriferus Newstead, 1894: 24 . Synonymy by Cockerell, 1898: 240.
Dactylopius talini Green, 1896: 7 . Synonymy by Cockerell, 1899: 391.
Dactylopius setosus Hempel, 1900: 386 . Synonymy by Costa Lima, 1939: 3.
Pseudococcus virgatus ; Kirkaldy, 1902: 103. Change of combination.
Dactylopius magnolicida King, 1902: 616 . Synonymy by Williams & Granara de Willink, 1992: 183.
Dactylopius virgatus var. madagascariensis Newstead, 1908: 7 . Synonymy by Mamet, 1951: 216.
Pseudococcus marchali Vayssière, 1912: 366 . Synonymy by Ferris, 1950: 93.
Pseudococcus bicaudatus Keuchenius, 1915: 49 . Synonymy by Green, 1916: 51.
Ferrisia virgata ; Fullaway, 1923: 308. Change of combination.
Ferrisiana virgata ; Takahashi, 1929: 429. Change of combination.
Ferrisia virgata ; Morrison & Morrison, 1966: 78; McKenzie, 1967: 179. Change of combination.
Ferrisia neovirgata Khalid & Shafee, 1988: 71 . Synonymy by Williams, 2004: 268.
For a full synonymy with details of type material and depositories, refer to Williams & Granara de Willink (1992), Williams (1996) or Ben-Dov (2012).
Type material: lectotype (here designated) of Dactylopius virgatus Cockerell : adult ♀ on slide with 4 other adult ♀ and 3 nymphs (mostly in poor condition), lectotype farthest from original label and clearly indicated, slide label: “ Dactylopius / virgatus / Ckll. / Part of type / from Jamaica. / 7”( USNM). Paralectotypes: 4 adult ♀, 1 second-instar nymph & 2 first-instar nymphs, on same slide as lectotype ( USNM); 17 adult ♀ or part of, 1 pharate adult ♀ in cuticle of third instar, 2 adult males, several damaged nymphs (distributed among 3 slides), one slide with 2 labels: “ Dactylopius / virgatus, Ckll MS / BM / On? 1945, 121 / Kingston Jamaica / 1892. T.D.A. Ckll. col.” and “No 105 / B.iv, p. 85 / Type lot / R. Newstead” and other 2 slides with “ Dactylopius / virgatus, Ckll / Type lot. Ex coll Ckll Jamaica / Institute / RN. BM 1945,121” or similar ( BMNH).
Other material from Cockerell (non-type): 2 adult ♀ in poor condition (1 slide), with label: “ Pseudococcus virgatus / (Ckll) / Type / East St. Kingston, Jamaica / Re’ed fr. Ckll. Jan 3, 1984 / 5636” ( USNM); 4 adult ♀ (4 slides) slide-mounted by P.J. Gullan 2002 from dry material in box with two labels inside: “Ckll. Coccidae III. TYPE. / Dactylopius virgatus, Ckll. / East St. Kingston. Jamaica ” and “5636 from Cockerell / Jan 3.94” [latter label in pencil and in different handwriting to main data label] ( USNM); 6 adult ♀ in poor condition (1 slide), with label: Pseudococcus virgatus / (Ckll) / On Guava / Kingston, Jamaica / Ckll. / May 3, 1983 5636” ( USNM) .
The USNM has a number of specimens of F. virgata sent from Jamaica by T.D.A. Cockerell, including one slide with a number of specimens that we regard as part of Cockerell’s original collection and from which we designate a lectotype (see above). In addition, there are other specimens sent from Jamaica by Cockerell but these are dated 1893 and 1894, whereas the orginal collection was made in 1892 and the species was described in 1893. Although one of these subsequent collections bears the label “Type”, we do not regard these specimens as types since it is clear that they were sent from Jamaica as live specimens in 1894 (as ascertained from notes by Theodore Pergande housed in the Coccoidea collection at SEL). The other subsequent collection was received in May 1893 but the host plant was guava, and this host was not mentioned in Cockerell’s original description of the species. Williams & Granara de Willink (1992) and Williams (1996) discussed syntype material in the BMNH apparently sent by Cockerell to R. Newstead for his opinion. A list of “ Coccidae ” sent from Cockerell to Newstead in May 1892 and a letter from Cockerell to Newstead dated May 28 th 1892 (both held in the BME) show that Newstead was sent specimens of Cockerell’s “ Dactylopius virgatus ” from East Street, Kingston, Jamaica, in 1892, together with notes on the live appearance of the mealybug and the comment: “A very destructive species; on a very glutinous tree with fleshy leaves and long spines”. We have been unable to establish a possible identity for this host plant. Some specimens in the BMNH are labelled: “ Dactylopius virgatus Ckll. , Type lot, ex coll. Ckll. Jamaica Institute” and others “ Dactylopius virgatus Ckll., M.S. On ? Kingston, Jamaica 17892, T.D.A. Ckll.? Type lot”. We regard these specimens as paralectotypes and have listed them above.
Cockerell’s original (1893a: 178) description of this species is accompanied by the comment that the mealybugs were in enormous numbers on an unidentified tree in East Street, Kingston, with the tree referred to as follows: “The leaves are ovate-acuminate, fleshy, entire; stalks reddish, with some long spines, very glutinous”. In the same paper, Cockerell (1893a) described four varieties of his Dactylopius virgatus from Kingston and named two of these varieties, namely D. virgata var. farinosus from Prosopis juliflora ( Fabaceae ) and D. virgatus var. humilis from Tribulus cistoides ( Zygophyllaceae ). One of the unnamed varieties was collected from Acalypha ( Euphorbiaceae ), which we know to still be a host of F. virgata in Kingston, and the other from Annona (sweetsop) ( Annonaceae ). The latter collection may have been F. dasylirii , given that we have a large recent collection of F. dasylirii from this host at the same locality in Kingston. These four varieties apparently differed mainly in the amount and patterns of the wax and bare areas of the dorsum, but there are no museum specimens identified as these varieties and their taxonomic status cannot be assessed. Cockerell (1893b) also listed another species, D. segregatus Cockerell , on grass in East Street, Kingston, but the type material appears to be lost and so it is not possible to determine its identity, especially as F. dasylirii also is known to occur in Jamaica; see Williams & Granara de Willink (1992) for further discussion.
Ferris (1950) synonymised Pseudococcus marchali Vayssière with Dactylopius virgata . The types of P. marchali were collected from mango in Kouroussa, the Republic of Guinea (specifically in Haute-Guinee); a lectotype was designated by Williams (1996) from slide-mounted females in the MNHN. We examined the lectotype and one paralectotype (on the same slide) and both are adult females of F. virgata .
We were able to examine type specimens of Dactylopius ceriferus Newstead from India, Dactylopius talini Green from Sri Lanka, and Dactylopius virgatus var. madagascariensis Newstead (all housed in BMNH) and these adult females were all F. virgata . Williams & Watson (1988) designated a lectotype for D. ceriferus and for D. talini , but not for D. virgatus var. madagascariensis , presumably because most of the syntypes of the latter species are in the Museum für Naturkunde der Humboldt-Universität zu Berlin, Germany ( Williams 1996).
We were unable to examine type material of Dactylopius setosus Hempel from Brazil, but the original description says that the antennae are 0.6–0.7 mm (600–700 mm) long, which is longer than typical for the adult females of F. virgata . Thus Hempel’s species may be F. dasylirii , which also occurs in Brazil, and if so, D. setosus would be a junior synonym of D. dasylirii . Also we have not seen original material of Ferrisia neovirgata Khalid & Shafee from India, but because this species probably was described from immature specimens ( Williams 1996), it would be difficult to allocate to species in any case. In addition, the types of D. magnolicida from Brazil, D. segregatus Cockerell from Jamaica, and Pseudococcus bicaudatus Keuchenius from Java are probably lost ( Williams & Granara de Willink 1992; Williams, 1996) and thus it may never be possible to determine the species status of these names. Although it is possible that one or more of the above names represents a senior synonym of one the new species that we are recognising here, without type specimens it is impossible from the original descriptions to establish the identity of these species.
Other material examined and used for decription. 1 adult ♀, ex unidentified ornamental plant, BRAZIL, Espírito Santo, Vítoria , INCAPER library, 6.iv.2005, M. Culik ( BME) ; 9 adult ♀ (9 slides, including DNA voucher FBK005), ex Acacia sp. on seed pod, GHANA, Greater Accra region, Shai Hills Resource Reserve , 05° 53’N, 00°03’E, 75 m, 12. vi. 2005, T. Kondo ( BME) GoogleMaps ; 4 adult ♀ (4 slides, including DNA voucher LGC01078 View Materials ), ex Acalypha wilkesiana on foliage, JAMAICA, Kingston, East Street , 7. ix. 2003, T. Kondo ( BME) ; 13 adult ♀ (13 slides, including DNA voucher TK0461), ex Coccoloba uvifera ( Polygonaceae ), MEXICO, Guerrero (state), Acapulco, near Fiesta Inn Hotel , 16°51’N, 99°51’W, 5 m, 16.i.2007, T. Kondo ( BME) GoogleMaps ; 8 adult ♀ (8 slides, including DNA voucher FBK004), ex culture of sprouted potato, SOUTH AFRICA, Stellenboch , 1. i. 2002, W. Wakgari (orginally from an exotic ornamental, Cestrum aurantiacum in 2000, reared in culture at CRI) ( BME) ; 7 adult ♀ (7 slides, including DNA voucher NH164), ex Acalypha wilkesiana , THAILAND, Nakhon Pathom, Kasetsart Univ. Kampaengsaen campus, 14°02’S, 99°58’E, 17 m, 1. vi. 2007, T. Kondo ( BME) GoogleMaps .
Ferrisia virgata is more widely distributed than any other species of the genus. In addition to the above list, we examined, but did not measure, specimens of F. virgata from the following countries: Australia, Brazil, Cambodia, China, India, Indonesia (Sumba - Lesser Sunda Islands), Kenya, Malaysia, Philippines, Singapore, Tahiti, Uganda, USA (Hawai’i) and Zambia, on the following plant species: Acalypha sp. ( Euphorbiaceae ), Amaranthus sp. ( Amaranthaceae ), Annona muricata , A. squamosa and A. reticulata ( Annonaceae ), Calophyllum inophyllum ( Calophyllaceae ), Carica papaya ( Caricaceae ), Citrus sp. ( Rutaceae ), Codiaeum sp. ( Euphorbiaceae ), Coffea sp. ( Rubiaceae ), Gardenia sp. , ( Rubiaceae ), Hibiscus sp. ( Malvaceae ), Hoya sp. ( Apocynaceae ), Indigofera sp. ( Fabaceae ), Leucaena sp. ( Fabaceae ), Mangifera indica ( Anacardiaceae ), Morus sp. ( Moraceae ), Pisonia sp. ( Nyctaginaceae ), Psidium guajava ( Myrtaceae ) and Rhipsalis baccifera ( Cactaceae ). The above host-plant list is not comprehensive as F.virgata is polyphagous and can feed on numerous plant species. It occurs widely in Central and South America (see below under “Biological notes”), but we have not made detailed lists of localities and host plants from this region.
The native range of F. virgata is unknown and it is possible that Jamaica is not part of the natural distribution of this species because, when it was described, Cockerell (1893a) referred to this mealybug as a destructive species and reported it on a number of hosts in Kingston, Jamaica. Thus F. virgata might have been introduced to Jamaica, where it became a pest; however, this mealybug still occurs at the type locality in Jamaica as it was collected there on Acalypha wilkesiana in 2003. See “Biological notes” below for further discussion.
ADULT FEMALE. Diagnosis. Ferrisia virgata can be diagnosed by the following combination of features: presence of clusters of small oral-collar tubular ducts on ventral margins of last 2 or 3 abdominal segments; ventral oral-collar tubular ducts frequently associated with a minute discoidal pore which rarely touches rim of duct, usually distant from rim by at least half length of duct; dorsal enlarged tubular ducts totalling 69–101 throughout dorsum, sclerotised area around rim of each duct often with 1 or 2 oval discoidal pores usually not touching outer margin of sclerotised area and almost never projecting out from margin; multilocular disc-pores on venter of abdominal segments VI (11–28) (usually more than 15 in double row), VII (22–41) and VIII + IX (14–29); anal lobe cerarii each with 2–3 (generally 3) conical setae; both pairs of ostioles present; antennae usually ≤600 µm long; hind coxa with translucent pores.
Ferrisia virgata is most similar to F. dasylirii as both species have more than one size of ventral oral-collar tubular ducts, with clusters of small oral-collar tubular ducts on the ventral margins of the last 2–3 abdominal segments, and a minute discoidal pore usually near rim of each enlarged dorsal tubular duct and most ventral oral-collar tubular ducts but never touching the rim of the duct opening. F. virgata can be distinguished readily from F. dasylirii by the following features: (i) discoidal pores associated with the sclerotised area around the rim of the dorsal enlarged tubular ducts ( Fig. 4 Aa) on the abdomen usually not touching the outer margin of the sclerotised area and almost never projecting out from margin (in F. dasylirii most discoidal pores are on the outer margin of the sclerotised area and often each pore and its surrounding sclerotisation projects out from the margin; Fig. 4 Ab); (ii) small oral-collar tubular ducts in clusters on the posterior abdomen with distal end of each duct rounded (in F. dasylirii the distal end of each duct is slightly tapered towards the attachment of inner ductile); (iii) antennae usually ≤600 µm long with apical segment 105–125 µ m long (usually ≥600 µ m long with apical segment 120–150 µm (usually>125 µm) long in F. dasylirii ); (iv) usually ≥15 multilocular disc pores on the venter of abdominal segment VI, typically forming at least a partial double row medially (in F. dasylirii usually ≤15 pores and typically forming a single, sometimes irregular, row); (v) translucent pores present on dorsal surface of hind coxa, especially posterolaterally, although often few in number (usually absent in F. dasylirii ); (vi) one or both anal lobes sometimes with an extra 1–2 conical cerarian seta that are more slender than the other 2 setae (in F. dasylirii only 2 cerarian setae except some specimens from Dasylirion ). F. virgata also is similar to F. cristinae , F. kondoi and F. williamsi (which all have at least two sizes of ventral oral-collar tubular ducts as well as clusters of small oral-collar tubular ducts on the ventral margins of the last 2–3 abdominal segments). However, F. virgata can be readily distinguished from the other three species by the position of the minute discoidal pores associated with ducts which, in F. virgata , are always near the enlarged tubular ducts and ventral oral-collar tubular ducts but never touch the rim of the duct opening (discoidal pores always adjacent to duct openings in F. kondoi and F. williamsi , and mostly adjacent in F. cristinae ). F. virgata shows some similarity to F. milleri and F. ecuadorensis but can be separated easily from these two species by the absence of ventral marginal clusters of small oral-collar tubular ducts on the head, thorax and anterior abdominal segments ( F. milleri has these clusters around the entire body whereas F. ecuadorensis has clusters on all abdominal segments).
Description of slide-mounted specimens (based on selected type material of D virgatus and other specimens listed above; Fig. 21). Body elongate oval, 2.10–4.48 mm long, 0.94–2.52 mm wide. Eye marginal, 50–75 µm wide. Antenna 8 segmented, 480–640 µm (generally ≤600 µm, often ≤560µm) long; apical segment 105–125 µm long, 30–35 µm wide. Clypeolabral shield 150–188 µm long, 135–200 µm wide. Labium 180–225 µm long, 90–130 µm wide. Anterior spiracles 60–85 µm long, 37–53 µm wide across atrium; posterior spiracles 70–113 µm long, 55–80 µm wide across atrium. Circulus quadrate, 115–195 µm wide, divided by an intersegmental line. Legs well developed; hind trochanter + femur 390–515 µm long, hind tibia + tarsus 410–530 µm long, hind claw 32–43 µm long. Ratio of lengths of hind tibia + tarsus to hind trochanter + femur 1.03–1.14, ratio of lengths of hind tibia to tarsus 2.68–3.25, ratio of length of hind trochanter + femur to greatest width of femur 3.73–5.25. Tarsal digitules subequal, each 50–60 µm long. Claw digitules subequal, each 35–43 µm long. Translucent pores present on hind legs, totalling 19–70 on coxa, femur and tibia combined; with 10–30 on dorsal surface of each hind coxa. Ostioles: both pairs present; each anterior ostiole poorly developed, with 22–28 trilocular pores and 7–11 setae; each posterior ostiole with 41–49 trilocular pores and 11–14 setae. Anal ring 120–195 µm wide, with 6 anal ring setae, each seta 170–235 µm long.
Dorsum. Anal lobe cerarii each with 2 conical setae, 37–43 µm long (sometimes an additional 1–2 robust setae, 20–36 µm long, slightly thinner than typical conical cerarian setae), with 36–60 trilocular pores and 4–6 auxiliary setae. Dorsal body setae slender, each 15–53 µm long. Trilocular pores each 4–5 µm in diameter. Enlarged tubular ducts totalling 69–122 on dorsum, each duct 30–35 µm long, 5.0–7.5 µm (generally 7.0–7.5 µm) wide at mid-length, rim of duct opening sclerotised, 10–13 µm wide, surrounded by a sclerotised circular area 15–35 µm wide, usually enclosing 1 or 2 oval discoidal pores (not touching outer margin of sclerotised area and almost never projecting out from margin) and with 1–5 (generally 2 or 3) setae, each 12–35 µm long, usually either within sclerotised area (especially on abdomen) or on edge of sclerotised area (especially on head); ducts distributed marginally in clusters of 2–6 on head and thorax, on margins of all abdominal segments in groups of 1–4, but with 5–9 ducts on each side of abdominal segment VII, and also 3–7 medially to submarginally on head and thorax, 5–10 medially to submedially on abdominal segments.
Venter. Body setae slender, each 12–197 µm long, longest setae medially on head; apical setae of anal lobe 280–325 µm long. Multilocular disc pores present on posterior abdominal segments only: 11–28 pores on segment VI (usually more than 15 and in at least partial double row), 19–41 on segment VII, 14–29 on segments VIII + IX; each pore 7–11 µm in diameter. Trilocular pores each 3–4 µm in diameter. Minute discodial pores each 2.0–2.5 µm in diameter, almost always associated with oral-collar tubular ducts, but never touching rim of ducts. Oral-collar tubular ducts on most of venter (excluding margins of posterior abdomen) each 10–13 µm long, 2.5–4 µm wide, totalling 53–95, distributed as follows: 33–43 on head and thorax, and on abdominal segments: 14–19 total on segments I–III; 4–7 on IV; 4–11 on V; 7–16 on VI; 8–18 on VII; none on VIII. Small oral-collar tubular ducts each 7.0–9.0 µm long, 3–4 µm wide, distributed as follows on margin of abdominal segments: 0–2 on each side of VI; 3–9 on each side of VII; 2–9 on each side of VIII.
Biological notes. The first description of F. virgata ( Cockerell 1893a) records this mealybug as numerous on a range of host plants in Jamaica. Shortly afterwards, this mealybug was described (under several different names) from India ( Newstead 1894), Sri Lanka ( Green 1896), Madagascar ( Newstead 1908), West Africa ( Vayssière 1912) and Java ( Keuchenius 1915). This pattern of spread and abundance at new localities is typical for newly invading insect species ( Liebhold & Tobin 2008) and suggests that F. virgata was carried around the world within about 10 years of its first discovery in Jamaica. However, the high abundance of F. virgata in Jamaica in 1892 suggests that it was not native to that island. It occurs throughout Central and South America ( Williams & Granara de Willink 1992; Culik et al. 2006; Kondo et al. 2008) and may be the species described as Dactylopius setosus by Hempel (1900) at the very beginning of his publishing career (however F. dasylirii also occurs in Brazil).
With the exception of records of F. malvastra , all other reports of Ferrisia from outside of the Nearctic and Neotropical regions are of F. virgata . Thus descriptions of immature stages ( Awadallah et al. 1979a; Paul & Ghose 1989) and studies of the biology (e.g., Ghose & Paul 1972; Awadallah et al. 1979b) of F. virgata in Egypt and India probably do refer to this species. Ferrisia virgata has been reported widely as a destructive pest on ornamental plants in India (reviewed by Ghose & Paul 1972). Awadallah et al. (1979b) found that F. virgata produced five generations per year if maintained on sprouting potato tubers at temperatures ranging from 16–29°C and 54–71% RH. These authors reported production of males if the mealybugs were maintained at high density on either potato sprouts or on Acalypha foliage, and claimed the occurrence of parthenogenetic reproduction for females maintained singly (but not isolated). Parthenogenesis is highly unlikely, especially as Ghose & Paul (1972) found that isolated virgin females did not produce any eggs. Adult females do not produce a definite ovisac but rest upon a pad of cottony white filaments on which they deposit eggs ( Fig. 2G). Ferrisia virgata is clearly ovoviparous because eggs hatch usually within 30 minutes of being laid ( Ghose & Paul 1972). Fecundity per adult female has been reported as 64– 78 eggs ( Awadallah et al. 1979b) to 222– 737 eggs ( Ghose & Paul 1972) under similar rearing conditions in the laboratory. In nature, the lower surface of leaves and the junction of the petiole with the stem were preferred oviposition sites on its main host plant in Egypt, A. macrophylla [this may refer to the cultivar A. wilkesiana “macrophylla’] ( Awadallah et al. 1979b). Three species of parasitoid wasps, Leptomastix sp. (Encrytidae) and two Tetrastichus spp. ( Eulophidae ), three species of Coccinellidae and two species of Neuroptera were reported as natural enemies in Egypt ( Awadallah et al. 1979b), although numbers were low relative to the high numbers of mealybugs. In India, Ghose & Paul (1972) recorded only a coccinellid predator ( Scymnus sp. ), of which the larvae were effective at controlling a population of the mealybug in culture. Ferrisia virgata is polyphagous, although Acalypha appears to be a favoured host plant in many countries, based on data from collections that we have examined for this revision.
Ferrisia virgata has been reported to transmit swollen shoot virus (CSSV) in cacao ( Thresh & Tinsley 1959; Roivainen 1976; Thorold 1975; ICTVdB Management 2006) and citrus tristeza virus (reported as lime dieback) on Citrus aurantifolia in Ghana ( Hughes & Lister 1953). In Africa, F. virgata has been shown to prefer cocoa leaf and reproductive parts (flowers and pods) to stem tissue, and although it has been reported to produce honeydew, it is rarely ant attended (Entwistle 1972), probably because its stylets reach the phloem infrequently ( Entwhistle & Longworth 1963).
Smithsonian Institution, National Museum of Natural History
Marie Selby Botanical Gardens
Universidade do Extremo Sul Catarinense, Bairro Universitário
No known copyright restrictions apply. See Agosti, D., Egloff, W., 2009. Taxonomic information exchange and copyright: the Plazi approach. BMC Research Notes 2009, 2:53 for further explanation.
Ferrisia virgata (Cockerell)
|Kaydan, M. B. & Gullan, P. J. 2012|
|Williams, D. J. 2004: 268|
|Khalid, M. & Shafee, A. S. 1988: 71|
|McKenzie, H. L. 1967: 179|
|Morrison H. & Morrison, E. R. 1966: 78|
|Takahashi, R. 1929: 429|
|Fullaway, D. T. 1923: 308|
|Green, E. E. 1916: 51|
|Keuchenius, P. E. 1915: 49|
Pseudococcus marchali Vayssière, 1912: 366
|Ferris, G. F. 1950: 93|
|Vayssiere, P. 1912: 366|
Dactylopius virgatus var. madagascariensis
|Mamet, J. R. 1951: 216|
|Newstead, R. 1908: 7|
|Kirkaldy, G. W. 1902: 103|
|Williams, D. J. & Granara de Willink, M. C. 1992: 183|
|King, G. B. 1902: 616|
|Costa Lima, A. 1939: 3|
|Hempel, A. 1900: 386|
|Ferris, G. F. 1953: 362|
|Cockerell, T. D. A. 1896: 202|
|Cockerell, T. D. A. 1899: 391|
|Green, E. E. 1896: 7|
|Cockerell, T. D. A. 1898: 240|
|Newstead, R. 1894: 24|
|Cockerell, T. D. A. 1893: 178|
Dactylopius virgatus var. farinosus
|Ferris, G. F. 1950: 93|
|Cockerell, T. D. A. 1893: 178|
Dactylopius virgatus var. humilis
|Ferris, G. F. 1950: 93|
|Cockerell, T. D. A. 1893: 179|
|Williams, D. J. & Granara de Willink, M. C. 1992: 183|
|Cockerell, T. D. A. 1893: 254|