This superfamily, as currently recognised, groups together many families of large moths mainly on characters of absence (Common, 1970; Franclemont, 1973) : of the M vein stem in the cells of both wings (though this is present in the forewings of some genera in the Bombyx lineage of the Bombycidae, and the hind-wing of some Brahmaeidae); of the tongue and frenulum (though the former is strongly developed in many Sphingidae, as is the latter (also in Apatelodidae), and the frenulum is weakly present in other families, e.g. Bombycidae).

The strongly bipectinate male antennae in most families, associated with strong sembling behaviour to female pheromones and the spinning of a fine silken cocoon, though widely distributed and highly developed in the Bombycoidea, are not unique to them: the Lymantriidae, in the Noctuoidea, also have strongly bipectinate antennae in the male, sembling behaviour to a static female, and lack a tongue. Lack of adult feeding and male sembling to a static female may be biologically linked, as described for the Saturniidae by Janzen (1984).

On these characters alone, therefore, there is no indication the superfamily is monophyletic: a natural grouping.

However, there are a number of features that link the families, or a larger portion of them, together and indicate a relationship with the Neotropical Mimallonoidea (Mimallonidae), treated as a separate superfamily by Franclemont (1973). In the Mimallonoidea ocelli, tongue and palps are absent, and the larva is striking in that later instars build a portable case of silk incorporating fragments of vegetation and frass.

The families of Bombycoidea represented in South-East Asia are the Lasiocampidae, Eupterotidae, Bombycidae, Brahmaeidae, Saturniidae and
Sphingidae plus a new, anomalous taxon,  Kenguichardia gen. nov., tentatively associated with the Bombycoidea by the loss of the stem of vein M in the wing cells, lack of tympanal organs, and presence of the character of the forewing radial veins discussed in the next section. Tongue and frenulum are strong, though ocelli and chaetosemata are absent.

Other bombycoid families are the Australasian Anthelidae and Carthaeidae, the Palaearctic Lemoniidae, Mirinidae (Minet, 1986) and Endromidae, and the New World Oxytenidae and Cercophanidae (allied to Saturniidae), and Apatelodidae. The South African genus Spiramiopsis Hampson is of uncertain placement, though currently resting in the Bombycidae (Fletcher & Nye, 1982), and has a number of interesting features.


The absence, and occasional presence, of the M stem in wing cells has already been mentioned. One feature common to many bombycoid lineages is the independent stalking of the second and third radial veins of the forewing. This is seen in some homoneurous non-ditrysian groups such as the Hepialidae, but is very rarely found in Ditrysia (present in Dalceridae).

To illustrate the distribution of this feature simply amongst the various bombycoid groups, the configuration of the radial veins as they arise from the cell can be expressed by means of a shorthand system. A semicolon separates veins or branching systems that arise independently from the cell; where they are connate (not stalked but arising from the same point on the cell) a hyphen is used. A system of brackets can be used to indicate order of separation of veins in a branching system and whether they are anterior or posterior. For example, the Megalopygidae are distinguished from the Limacodidae in the key to families in part 1 of this series (Holloway,1986) by whether R4 arises anteriorly or posteriorly from Rs. This can be expressed as (R2(R3(R4,R5))) for Megalopygidae and (R2((R3,R4)R5)) for Limacodidae, those veins listed after the comma arising posteriorly, those before anteriorly, and the order of bracketing from the distal dichotomy indicating the order of branching off from Rs basad. A fusion, e.g. in the hindwing of Brahmaeidae, can be indicated as R1>Sc, and an anastomois (fusion then separation), as in the formation of the forewing areole in the Anthelidae, as R3><R4, or, as in formation of the Dalceridae areole (Holloway, 1986,fig.13), R2 + 3><R4 + 5.

When the number of radial veins is reduced, it is sometimes apparent which have become fused or lost, as in the Sphingidae, but often this not the case. In the latter instance it is perhaps least confusing to leave the R veins unnumbered.

The following examples illustrate some of the variety of venation found in the Bombycoidea, from groups with a full complement of veins to those with one or two of the radials lost. Note the recurrence of R2, R3 in the groups with all radials present except for the Bombyx lineage of the Bombycidae and the Apatelodidae:

(R1); (R2, R1); ((R4, R5)M1)
or (R4, R5)-(M1)

Lasiocampidae: Alompra group
Sphingidae: some Smerinthini

(R1); (R2 + 3); R4, R5) - (M1)

Most Sphingidae

(R1); (R2, R3); (R4(R5, M1))

Most Lasiocampidae

(R1); ((R2, R3)R4); (R5, M1)

Lasiocampidae: Bhima

(R1); (R2, R3); (R4, R5); (M1)


(R1); (R2, R3); (R4, R5) - (M1)




(R1); (R2, R3); (R4); (R5); (M1)


(R1)-(((R2, R3)R4)R5)) - (M1)


(R1); ((R2, R3)(R4, R5)) - (M1)


(R1); ((R2(R3(R4, R5)))M1)

Bombycidae: Bombyx group

(R1); ((R2(R3(R4, R5))) - (M1)

Apatelodidae: Apatelodes

(R1); ((R, R)R) - (M1)


or (((R, R)R)M1)

Lemoniidae: Lemonia
Bombycidae: Mustilia group
Saturniidae: Citheronia

((R1(R, R))R); (M1)

Saturniidae: Actias, Loepa, Cricula

(R1); ((R, R) R); (M1)

Saturniidae: Samia, Attacus

(R1(R, R)); (M1)

Saturniidae: Antheraea

(R1) - ((R,R)M1)

Saturniidae: Anisota

(R1); (R, R); (M1)

Saturniidae: Hemileuca

Broad, deep-winged groups such as most Saturniidae, the Eupterotidae and Brahmaeidae rest with the wings flat and both wings exposed. Therefore the hindwing is usually as strongly and extensively patterned as the forewing. Lasiocampidae rest with the wings steeply tented over the body, the expanded humeral area of the hindwing protruding from under the costa of the forewing, giving a very disrupted, irregular outline that combines with cryptic patterning to give a dead leaf effect. In most Lasiocampidae the hindwings are most strongly patterned costally because of this posture. Sphingidae: Macroglossinae have a swept back, jet fighter type of posture. The tribe Sphingini of the Sphinginae hold their wings at a more swept back angle, and angled towards the substrate, usually a tree trunk; the forewing pattern is often patterned for crypsis on bark, sometimes with flash coloration on the hindwing. None of these indicate any linkage between bombycoid families.

However, some of the Sphinginae: Smerinthini (eg. Smerinthus, Daphnusa), the Bombycidae, the South Africa Spiramiopsis (N. Duke, pers. comm.) and Apatelodidae share a peculiar resting posture not seen in other macrolepidoptera. The forewings are held out at right angles to the body, with strongly patterned dorsum of the hindwing scrolled up and over that of the forewing. The abdomen is usually flexed upwards or sideways.


In most groups the uncus of many of the taxa is bifid or bilobed;  when entire, as in most Sphingidae: Macroglossinae, this appears to be a derived character.

Kuznetzov & Stekolnikov (I985) studied genital musculature in several of the bombycoid families, concluding they  were interrelated and suggesting the interrelationships were as in Fig. 1. All except the Lasiocampidae share a secondary type of muscle insertion. These authors split the superfamily into the three: Lasiocampoidea, Sphingoidea and Bombycoidea sensu stricto.

Figure 1. Diagram of relationships of some Bombycoidea families according to Kuznetzov & Stekolnikov (1985).

The generalised condition of a macrolepidopteran larva might be taken to be possession of three pairs of verrucae or scoli on each segment, invested with secondary setae. These are situated dorsally, subdorsally, and laterally below the spiracles. This condition is seen in many Saturniidae and in the verrucae, when present, of Lasiocampidae.

In all Bombycoidea there are secondary setae not associated with verrucae or scoli, often very numerous, and with the superfamily there is a trend either towards great hairiness (Lasiocampidae, Eupterotidae, Anthelidae, Lemoniidae some Apatelodidae, some Saturniidae), or reduction of the setae to fine spines or granulation, and associated reduction of the scoli in number, yet often with enlargement of a few of them (other families, a few Apatelodidae and many Saturniidae). In all these families with non-hairy larvae, the eighth abdominal segment has a central scolus or horn dorsally rather than a pair, though some Saturniidae (especially Saturniidae) have a pair or show signs of partial fusion of a pair (W. Nassig, pers. comm.). Most Bombycidae, the Sphingdae,  a few
Apatelodidae, the Oxytenidae, Endromidae and early instars of the Carthaeidae have the caudal horn only. In Bombycidae the thoracidae segments are not ornamented; dorsal pairs of filaments occur on segments A1-7, A9 in one species of Bombyx and members of the Ocinara group and Gunda often have transverse swellings dorsally on A2 and A5. However, many other groups have paired filaments or horns on T2 and T3 as well as the horn on A8 but rarely on other abdominal segments, such as the Brahmaeidae (a small pair also on A9), Spiramiopsis, several subfamilies of the Saturniidae (Agliinae, Rhescyntinae, Citheroniinae); the Cercophanidae have caudal and thoracic areas produced into tapering processes. Thus there are rather unusual larval features linking many of the bombycoid families. Snake mimicry is also a recurrent feature (Macroglossinae, final instar Brahmaeidae, some Bombycidae).

Larvae of Sphingidae, Saturniidae and Brahmaeidae have a 'sphinx-like' alarm posture with head and thoracic segments retracted and lifted off the substrate; the dorsum of the thorax appears expanded  and directed forwards

Thus, with unusual features of venation, resting posture, male genitalia and larvae found to be common to diverse families within the Bombycoidea, there are some grounds for considering that the group may be a natural one (see also Minet (1986)). The diagram of family relationships in Fig. 1 would require reversal of the dorsal pair of scoli on A8 in some lineages of Saturniidae.


Host-plant specialisation at the family level is only seen in some of the smaller Bombycoid families, with the Brahmaeidae mostly recorded from Oleaceae, the Bombyx lineage of the Bombycidae from Moraceae and the Mustilia lineage from the Symplocaceae and the Theaceae, two closely related families. In the Eupterotidae the genus Ganisa has only been recorded from Oleaceae, as has the genus Dolbina in the Sphingidae. Also amongst the Sphingidae, the genera Clanis and Sataspes would appear to be restricted to Leguminosae.

Most taxa in the other families appear to have a range of diet covering several plant families. It is instructive to note the frequency with which various families have been recorded as host-plants in the various major bombycoid groups. Janzen (1984) drew attention to dietary differences between the Saturniidae and Sphingidae in seasonal forest in Costa Rica. But there are even more striking differences between the major groupings with  the Sphingidae. There is no dietary overlap between the Sphingini (12 spp), Smerinthini (3 spp) and Macroglossinae (17 spp), as can be seen from the Table 1 below, illustrating just the more frequently recorded families.

There is a very similar dietary segregation within the Bornean sphingids (Table 2), each subfamily having much the same plant family preferences as in Costa Rica, and only slight overlap between the groups, eg. Leguminosae in Sphingini and Smerinthini and Convolvulaceae for Sphingini and Macroglossinae. The Smerinthini are much more diverse in Borneo than Costa Rica.

Table 1.
Most frequently recorded host-plant families for Sphingidae higher taxa in Costa Rica; data from Janzen (1984). The number of genera in each family recorded for each sphingid species are summed for all species (i.e. the same plant genus will be counted more than once if fed on by more than one species). 





Bignoniaceae (10)

Anacardiaceae (2)

Rubiaceae (22)

Verbenaceae (5)

Lauraceae (2)

Dilleniaceae (5)

Compositae (5)


Apocynaceae (5)

Solanaceae (5)


Moraceae (4)

Convolvulaceae (4)


Vitidaceae (3)

Annonaceae (2)


Euphorbiaceae (2)

Boraginaceae (2)


Asclepiadaceae (1)

Labiatae (2)


Sapotaceae (1)

Table 2. Most frequently recorded host-plant families for Bornean Sphingdae, collated from the full geographical range of the species and presented as in Table 1.




Bignoniaceae (18)

Leguminosae (13)

Araceae (40)

Oleaceae (17)

Fagaceae (8)

Vitidaceae (38)

Verbenaceae (16)

Anacardiaceae (6)

Rubiaceae (29)

Leguminosae (15)

Sterculiaceae (4)

Dilleniaceae (10)

Labiatae (5)

Bombacaceae (4)

Onagraceae (8)

Convolvulaceae (5)

Lauraceae (3)

Leeaceae (7)

Scrophulariaceae (4)

Burseraceae (2)

Convolvulaceae (6)

Pedaliaceae (3)

Sapindaceae (2)

Balsaminaceae (6)

Malvaceae (2)

Guttiferae (2)

Actinidiaceae (6)

Meliaceae (2)


Nyctaginaceae (5)

Araceae (2)


Polygonaceae (4)

Caprifoliaceae (2)


Begoniaceae (3)

There are some major differences however, that may reflect ecological differences between the two regions. Also the host data for the Bornean species are collated from the whole Indo-Australian ranges of many of the species, but those for Costa Rica are from a very restricted locality. Oleaceae are strongly represented for the Sphingini, Leguminosae for Sphingini and Smerinthini, Fagaceae for the latter, and Araceae and Vitidaceae for the Macroglossinae. These plant families are utilised only infrequently or not at all in Costa Rica.

The other major group considered by Janzen was the Saturniidae contrasting both their host-plant range and adult behaviour with that of the Sphingidae. Many saturniid subfamilies are represented in Costa Rica, but only the Saturniinae in Borneo. The host-plant data for this family are presented in Table 3.

Table 3. Most frequently recorded host-plant families for Saturniidae in Costa Rica, and for Saturniinae (CR) and Borneo (B), compiled as described for Tables 1 and 2.

Saturniidae (CR)

Saturniinae (CR)

Saturniinae (B)

Leguminosae (34)

Rubiaceae (3)

Rosaceae (14)

Rubiaceae (9)

Anacardiaceae (3)

Lauraceae (8)

Flacourtiaceae (7)

Flacourtiaceae (2)

Anacardiaceae (7)

Bombacaceae (5)

Rutaceae (1)

Rutaceae (6)

Sapindaceae (5)

Lauraceae (1)

Euphorbiaceae (5)

Sterculiaceae (5)


Meliaceae (5)

Anacardiaceae (4)


Fagaceae (4)

Fagaceae (3)


Lythraceae (4)

Myrtaceae (3)


Simaroubaceae (4)

Rutaceae (3)


Rhamnaceae (4)

Bignoniaceae (3)


Leguminosae (3)



Corylaceae (3)

The six most popular families are not the same in the two regions, but overall four families are shared, and three families between the Saturniinae in each locality.

Janzen (1984) divided the Costa Rica 'big moths' of the Bombycoidea into two behavioural categories: (1) those  that do not feed as adults are shortlived, the male searching for a static female calling with a pheromone, have marked sexual dimorphism, oviposit in masses, usually in tree crowns, the larvae aposematic or mimicking urticating forms, feeding on foliage often low in nutrients, high in phenolics (Janzen & Waterman, 1984); (2) both sexes feeding, highly active, streamlined, long-lived with sexual dimorphism weak or absent, ovipositing singly on small plants, saplings, shrubs, vines or tree crowns, the larvae cryptic, with a high rate of nutrient assimilation from young or old foliage that is more nutrient rich though often containing toxic alkaloids, but being in low in phenolics (Janzen & Waterman, 1984).

Though the Sphingini and Macroglossinae of Costa Rica and Borneo show close parallels, the representation of weak-tongued Smerinthini and other non-feeding big moths is somewhat different, with a very diverse saturniid fauna in Costa Rica, but only weak representation of Smerinthini and, for the Neotropics as a whole, Lasiocampidae. In Borneo the Smerinthini and Lasiocampidae are diverse, Eupterotidae are present, yet the Saturniidae are only weakly represented and that by only one subfamily. It could therefore be that the non-feeding 'big moth' niche is occupied by Lasiocampidae, Eupterotidae, Saturniinae and Smerinthini for Borneo. This contention is supported when the host-plant preferences of Eupterotidae and Lasiocampidae are added to those of Sarturniinae and Smerinthini for Borneo and compared with those for Saturniidae plus Smerinthini for Costa Rica. (Table 4). There are nine families in common, with Rubiaceae, Flacourtiaceae and Bignoniaceae peculiar to Costa Rica, and Rosaceae, Meliaceae, Combretaceae, Lythraceae and Rhamnaceae to Borneo.

Table 4.
Most frequently recorded host-plants for non-feeding 'big moths' in Costa Rica (NFBM - CR) and Borneo (NFBM - B), calculated as discussed in Tables 1 and 2.




Leguminosae (15)

Leguminosae (34)

Leguminosae (35)

Myrtaceae (15)

Rubiaceae (9)

Rosaceae (18)

Combretaceae (5)

Flacourtiaceae (7)

Fagaceae (16)

Euphorbiaceae (4)

Anacardiaceae (6)

Lauraceae (16)

Lauraceae (3)

Sapindaceae (5)

Anacardiaceae (14)

Sapotaceae (3)

Sterculiaceae (5)

Rutaceae (9)

Theaceae (3)

Lauraceae (4)

Meliaceae (8)

Verbenaceae (3)

Fagaceae (3)

Combretaceae (8)

Rosaceae (3)

Myrtaceae (3)

Lythraceae (7)

Rutaceae (3)

Rutaceae (3)

Bombacaceae (6)

Dipterocarpaceae (3)

Bignoniaceae (3)

Sterculiaceae (6)



Rhamnaceae (6)



Sapindaceae (5)



Leguminosae (4)



Lauraceae (2)



Oleaceae (2)



Combretacea (2)



Whilst all Bornean Saturniinae are deep-winged, Costa Rica Saturniidae have a diversity of wing forms. However, the inclusion of Eupterotidae and Lasiocampidae as Bornean non-feeding big moths brings the Bornean situation into closer parallel with that in Costa Rica. The similarity of the Lebeda to Kunugia group of lasiocampid genera in form to that of the Costa Rican Ceratocampinae is  particularly striking.

<<Return to Contents page

Copyright Southdene Sdn. Bhd. All rights reserved.