::The Moths of Borneo::

The Lymantriidae are a major noctuoid family that is particularly diverse in the Old World tropics. Current estimates of the number of species worldwide range from 2160 (Holloway, Bradley & Carter, 1987) through 2500 (Ferguson, 1978; Common, 1990) to 2700 (Scoble, 1992). Heppner (1991) broke a global estimate of 2416 down into regions that can be ranked as follows: Afrotropical (1004); Oriental east to Moluccas (742); Australasia including New Guinea and islands eastwards (255); Palaearctic (203); Neotropical (180); Nearctic (32).

The Australasian fauna has been estimated at 74 described species and many undescribed by Nielsen, Edwards & Rangsi (1996). The lymantriid fauna attenuates rapidly east of New Guinea with about 17 species in the Solomons (Holloway, unpublished checklist), four in Fiji (an endemic trio of Calliteara Butler, and the monobasic endemic Adetoneura lentiginosa Collenette) and three in New Caledonia (of general Australasian affinity, (see Orgyiini and Orvasca Sciasticta Collenette comb.n., Orvasca sp.2692, Orvasca rufalba Holloway comb.n.). The family is absent from New Zealand.

Faunal richness decreases northwards from the Oriental tropics, with just over 100 species recorded for Taiwan (Heppner & Inoue, 1992) and just over 150 for Japan (Inoue et al, 1982). The Nearctic fauna (Ferguson, 1978) is essentially an attenuated Palaearctic one, though with the genus Dasychira Hübner probably endemic (See Orgyiini). The Neotropical fauna is distinct from that of the Nearctic and appears to bear no obvious relationships to that of the Indo-Australian tropics (See Introduction).

The faunas of Borneo and Sumatra are similar in size and composition. Schintlmeister (1994) referred to approximately 230 species for Sumatra, with the comment that many more members of the Arctornis Germar (Redoa Walker) group require investigation and description. He noted about 46 of these, a total much lower than that for Borneo with about 81. Outside this group, the Sumatran total of 180 is somewhat less than the Bornean tally of 216. The Bornean fauna therefore contains over one third of the Oriental estimate of 742 by Heppner (1991).

Apart from the coverage by Schintlmeister (1994) of the Sumatran fauna, adding greatly to the earlier treatment by van Eecke (1928), there have been several other publications on island faunas in the region, particularly by Collenette. These listed and described species in Peninsular Malaysia (Collenette, 1932), Java (Collenette, 1949a), Bali (Collenette, 1949b) and Sulawesi (Collenette, 1947), as well as being more broadly based geographically (e.g. Collenette, 1953). More recently there have been accounts of the faunas of Nepal (e.g. Kishida, 1993, 1994, 1995) and Taiwan (Wang, 1993).

The highly polyphagous, but largely arboreal larvae include many agricultural and forestry pests and, when abundant, may also cause medical problems with the urticating properties of their setae (see Nygmiini n and Toxoproctis Gen.n.). On the other hand, their richness in tropical lowland forest in the region and relatively low mobility may enable them to be used as indicators for biodiversity assessment and environmental monitoring (Holloway, 1984, 1985b).

Definition of the family

The Lymantriidae are noctuoids with a thoracic tympanum and show a number of unique features within the superfamily (Ferguson, 1978; Maes, 1984a; Common, 1990; Scoble, 1992), as well as having a prespiracular counter- tympanal hood as in the Arctiidae.

The tongue is usually reduced or absent: the adults do not feed.

The antennae of both sexes are strongly bipectinate, the pectinations being longer in the male, terminating in one to three long spinules (Figs 1, 2), though such spinules also occur in some Arctiinae (Kitching & Rawlins, 1999 [1998]).

Males often have paired, pocket-like tymbal organs on the third abdominal sternite (Figs 3-7). The occurrence of this character was reviewed by Dall' Asta (1988). It is found in many genera of all the tribes occurring in Borneo and in all the genera not placed to tribe except one. It may therefore be part of the family ground-plan. However, it was only found in one (Thagona Möschler) of a sample of four Neotropical genera dissected.

The female genitalia have a small pair of setose lobes ventral to, and distinct from, the ovipositor lobes. These were termed pseudopapillae by Maes (1984a), who suggested they were diagnostic for the family, but they are seen also in aganaine Noctuidae (Holloway, 1988) and some Arctiidae (Kitching & Rawlins, 1999 [1998]). They can be seen, for example, in Figs 81, 214 and 243.

The larvae have conspicuous, round, red or yellow dorsal glands, one on each of abdominal segments 6 and 7.

The adult resting posture is tectiform, or flattened tectiform, with the hairy front and middle legs displayed, the forelegs held out in front, the middle legs laterally (Tweedie & Emmet, 1991). However, some lithosiine arctiids exhibit a similar resting posture (Kitching & Rawlins, 1999 [1998]).

General features

The following observations are drawn from accounts of the family cited in the previous section. The adults show a great diversity of wing facies, though it is usually only the forewing that is patterned. This can range from the rather typical noctuoid patterning of most Orgyiini, through the more zig-zag fasciation of the Lymantriini, to the bolder, more colourful patterns of the new tribe Nygmiini or the satiny white of most species in the new tribes Leucomini and Arctornithini.

Ocelli and chaetosemata are absent. There are often long, hair-like scales on the thorax and abdomen. Both forewings and hindwings have M2 arising closer to M3 than to M1, the quadrifid state of the former being common to all noctuoids except the Notodontidae. A forewing areole can be present or absent.

Sexual dimorphism is moderate to extreme, females tending to be much larger, particularly in the abdomen which, with the non-feeding habit, is full of eggs at eclosion. Females either have much larger wings than males or, in the case of some Orgyiini, are flightless and brachypterous or apterous. Even when winged, most lymantriid females fly only weakly and are much rarer than males in light-trap samples.

Most female lymantriids have a scale tuft (corethrogyne) at the apex of the abdomen, and this is used to cover the egg mass to protect it. The scales can be irritant. Stereoscan photographs of these scales from egg masses of a small sample of lymantriid species made at the Australian National Insect Collection suggest that these may vary amongst the tribes. Those of Lymantria are narrow, long, longitudinally ridged, the ridges finely serrate such that they are barbed away from the apex. Between the ridges are frequent septa, with some perforations between the septa. The scales of Orgyia are similar, but mostly broader and of more variable breadth. In the Nygmiini they vary from ridged to irregularly and strongly barbed rods, the barbs of variable length and frequency, more like thorns on a stem: the latter could be larval hairs derived from the cocoon, as described by Kitching & Rawlins (1999 [1998]). Weakly flighted or apterous females may lay their egg mass on the cocoon from which they have just emerged. The extent to which this occurs is variable but is most developed in the Nygmiini and Orgyia. The Arctornithini are exceptional in laying eggs singly, the females lacking a corethrogyne. The Leucomini are intermediate, laying eggs in rows or groups, lacking a corethrogyne, but with scales attached to the egg mass in Leucoma (de Worms, 1979).

The male genitalia are highly variable in structure. The eighth abdominal sternite is rarely modified to a large extent (an exception being Dura Moore), though the tergite is occasionally so, as in Dura, Calliteara of the Orgyiini, some genera of the Nygmiini and in one of the unassigned genera. The female genitalia usually have distinctly setose ovipositor lobes, short and deep, as well as the pseudopapillae. The signum is bicornute or a flange in all tribes except the Arctornithini (see below and Arctornithini n). The development of the sterigma is highly variable, but reaches its greatest complexity in some Nygmiini.

The eggs are of the upright type, circular, domed or slightly flattened, laid usually in tight masses, though the Arctornithini and Leucomini are exceptions. The larvae can remain gregarious (e.g. some Nygmiini) or disperse, often by ballooning on silken threads (Orgyiini in particular, where the female is flightless or only weakly so).

The larvae have dense and often long secondary setae on verrucae as in other 'hairy' noctuoids such as Arctiidae. The urticating properties of these have been referred to. Denser brushes or longer hair pencils occur in some groups. Crochets are numerous on the prolegs, in a uniordinal, homoideous mesoseries. The disposition and size of the verrucae and other larval characteristics will probably prove of value in elucidating relationships within the family, as indicated by the key of Gardner (1938) and discussed in later sections. Gardner presented a very detailed account of Indian species from all five of the tribes recognised here.

Pupation is usually in a dense silken cocoon that incorporates larval setae, though this can be very weak or vestigial in the Leucomini and Arctornithini where the pupa is exposed in a cradle of silk. The pupa also has dense secondary setae, mostly arising from the verrucal scars of the larva. A cremaster is present, with short, hooked setae at its apex.

Classification within the family

Ferguson (1978) considered the possibility of segregating subfamilies within the Lymantriidae but felt that knowledge of tropical faunas was too limited to permit this. He contented himself with segregating Nearctic taxa into two tribes, Lymantriini and Orgyiini.

Though this Bornean survey has probably covered most of the morphological variety present in Indo-Australian representatives of the family, there are sections of the Afrotropical fauna that require further study and the whole of the small Neotropical fauna. Dissection of a sample of Neotropical genera (Caviria Walker, Desmoloma Felder, Sarsina Walker and Thagona Möschler) has shown no definite relationships with any of the groupings or isolated genera recognised for Borneo. Only Thagona males had tymbal organs and none of the females examined had a signum. A number of Neotropical genera thought to have been Lymantriidae are now placed in the Bombycidae as subfamily Phiditiinae (Lemaire & Minet, 1999 [1998]).

Therefore, it was decided to continue the process started by Ferguson and to identify tribal groupings rather than subfamilies. The result is the recognition of three new tribes, segregated from Ferguson's concept of the Lymantriini. They are discussed in more detail in the relevant parts of the systematic section, but are listed below with their principal diagnostic features.

LYMANTRIINI	The forewing facies is characteristic, with zig-zag or lunulate fasciation, a V-shaped discal spot and an orbicular spot.
ORGYIINI	The larva has dorsal brushes on the first four abdominal segments.
NYGMIINI	The seventh abdominal segment is expanded and membranous in the female.
LEUCOMINI	The male genitalia are asymmetric.
ARCTORNITHINI	The signum of the female bursa is a broad, sparely but evenly scobinate plate.

In addition, there are six genera treated at the end that could not be assigned with confidence to any of these groups.

Relationships within the Noctuoidea

Kitching & Rawlins (1999 [1998]) have suggested that the Lymantriidae are sister-group to a paraphyletic Pantheidae, sharing apomorphic features such as the presence of secondary setae in first instar larvae, including, on the prothorax, distinctive, enlarged prespiracular verrucae, making that segment the widest part of the body. The character is otherwise seen only in some Acronicta.

As both the Pantheidae and Acronicta have already been covered in this series, the former as a subfamily of the Noctuidae (Holloway, 1985a), and the latter within a narrow concept of the noctuid subfamily Acronictinae (Holloway, 1989a), a brief review of their classificatory position is merited.

There is widespread disagreement on this position. Poole (1995) placed the pantheines and acronictines in a group with the noctuid subfamily Bryophilinae on the grounds of the presence of secondary setae in first instars, a character he interpreted as a reversal to a primitive state. Otherwise, he stated that the adults of the three groups are ‘clearly “trifid” noctuids on the basis of larval characters and the structure of the tympanal region’, though later he asserted that ‘I have an unsupported feeling that the Pantheinae and Acronictinae are not closely related despite the similarity of their larvae’.

Speidel, Fänger & Naumann (1996) also treated the pantheines as a noctuid subfamily, grouping them in a trichotomy with the Nolidae (as defined by Kitching & Rawlins (1999 [1998]) and by Holloway (1998a)) and a large clade that includes the trifine subfamilies covered by Holloway (1989a) with the Plusiinae and a few other small groups. There is an emerging consensus on the broad composition of this clade, though disagreement over the membership of some of the groups, particularly the Pantheinae (idae), though all place the Acronictinae sensu stricto within it.

It is likely that the Pantheinae do not represent a monophyletic group, and there is disagreement over which genera should be included. Speidel, Fänger & Naumann (1996) included Dilobinae and Mominae in their concept: neither of these groups is represented in Borneo. Kitching & Rawlins (1999 [1998]) included all the Bornean genera in their Pantheidae, but placed Dilobinae in the Noctuidae.

The Bornean pantheine species covered by Holloway (1985a) consist of three rather dissimilar species in different genera, all with quadrifine hindwing venation and conspicuously hairy eyes. Trichosea Grote may be related to Acronicta, as discussed by Holloway. Elydnodes Hampson has genitalic features in common with Trichosea and might best be grouped with that genus, wherever it is finally placed.

The other two species, in the genera Anepholcia Prout & Talbot and Antitrisuloides Holloway, together with Trisuloides Butler and Smilepholcia Prout & Talbot, belong to a predominantly montane Indo-Australian assemblage of rather large species with densely scaled, variegated forewings and orbicular stigmata. The hindwings are yellow, bordered broadly with black interior to yellow fringes. The relationships of this group require further investigation, particularly with regard to the north temperate pantheine type genus, Panthea Hübner. This will probably require discovery of the early stages. Kobes (1985) has covered the slightly more diverse Sumatran pantheine fauna.

In conclusion, the precise relationships of the Lymantriidae within the Noctuoidea are still unclear, but they may group with the Arctiidae, Nolidae and possibly part of the Pantheidae in a sister-relationship with the bulk of the groups currently placed in the Noctuidae (Kitching & Rawlins, 1999 [1998]). Recent molecular studies (Mitchell, 1998) strongly support a sister-relationship between Lymantriidae and Arctiidae, but place these as sister to another pair of taxa currently treated as Noctuidae subfamilies by Kitching & Rawlins (Aganainae and Herminiinae) but by other authors (e.g. Common, 1990) as good families. The molecular data also placed the Pantheinae with the trifine Noctuidae in a monophyletic grouping.

The Arctiidae, Nolidae (Holloway, 1998) and Lymantriidae all have sound- producing organs in all or a large number of their taxa, though the location of these is different in each case. Kitching & Rawlins suggested that these features might be serially homologous (sternal) and therefore provide some evidence of relationship. Sound production in Noctuidae is mostly restricted to trifine groups (Heliothinae, Agaristinae), where it involves foveate areas of the forewings (Matthews, 1991); foveae also occur at the base of the forewings of Amyna Guenée species (Acontiinae). Amongst the Catocalinae, Arcte Kollar has a conspicuous strigil on the sixth abdominal tergite (Robinson, 1975).

Host-plant specialisation

As stated earlier, lymantriid larvae are, as a rule, highly polyphagous, and this is general for all the tribes recognised here. The larvae are predominantly arboreal defoliators, but a few orgyiines and nygmiines also attack herbaceous plants. In the former, the genera Psalis Hübner, Laelia Stephens and Pantana Walker have become specialist on grasses and reeds, the facies of the adults also reflecting this in its rather straw-like character.

The other instance of specialisation is by the leucomine genus Perina Walker that has, apart from a few dubious records, been recorded entirely from Moraceae, particularly the banyan, Ficus religiosa.

Some lineages of Nygmia Hübner may favour Loranthaceae, and the diversity of Arctornis Germar in Sundaland may reflect an ability to cope with the challenges of dipterocarp foliage rarely seen elsewhere in the Lepidoptera (See Arctornithini n).

Ecological and biogeographic representation

Table 1 presents summaries of the biogeographic and ecological composition of the four largest tribes in Borneo in the form of two-way tables, biogeographic categories against ecological ones. This format has been applied in previous volumes of the series and originally by Holloway & Barlow (1992) and Holloway (1994).

Well over half the species in all tribes except Orgyiini fall into the Sundaland or endemic lowland forest categories. Endemism is highest in the Arctornithini, though this may decline as the faunas of other parts of Sundaland become better known. It is also substantial in the Nygmiini but low in the Lymantriini. The Orgyiini have more species in the widespread biogeographic categories and the disturbed, open habitat category, a reflection of the high level of pest species in the group and the adaptation of many to feeding on herbaceous vegetation.

Montane species are infrequent except in Orgyiini, where these are again of a more widespread nature, and Nygmiini, where endemism is high. All four groups, but particularly the Nygmiini and Arctornithini, are well represented in the ecological category for high ecological amplitude over altitude, with many species occurring in both lowland and montane forest.

The family generally is extremely rich in primary lowland forest and declines in diversity with latitude, altitude and disturbance. Chao (1981) demonstrated a similar decline with altitude on Mt. Emei in western China, with 93% of species recorded below 1100m and only 21% above 1900m. Three- quarters of this fauna is Oriental tropical and subtropical, the rest Palaearctic, with the latter increasing in representation with altitude.

Holloway (1984) noted a gradient in lymantriid diversity in forest types sampled during his survey of the G. Mulu National Park, with diversity increasing from young secondary forest through alluvial forest regenerated from farmland abandoned a century before, to undisturbed lowland forests. Data for secondary forest and various types of softwood plantation in Sabah were obtained by Chey (1994). The general trends of moth diversity observed in these sites (Chey, Holloway & Speight, 1997) are also evident in just the Lymantriidae, with the samples from secondary forest, Eucalyptus deglupta plantation (where there was a much more diverse secondary forest component in the understorey) and one of Pinus caribaea having higher diversity than the others, the lowest being from the two legume tree plantations (Acacia mangium, Paraserianthes falcataria), where high numbers of Orgyia postica Walker and, in the former only, Rhypotoses glebula Swinhoe were recorded. Values of the Fisher, Corbet & Williams (1943) a measure for a selection of these sites are given in Table 2. The Mulu survey values and those from the data of Chey are not strictly comparable, as the former were from pooling two to four consecutive nights of samples and the latter from pooling samples made over a whole year. Mean values of a for pooled monthly samples from the plantations were half (or less) those for the same sites for the whole year for all Macrolepidoptera (Holloway, 1998b). The gradient of diversity appears to be much stronger in the Arctornithini than in the Nygmiini.

Table 1. Percentage of species for the major tribes of Bornean Lymantriidae amongst various biogeographic and ecological categories as discussed in the text.

Table 2. Measures of lymantriid diversity from light-trap samples made in lowland forest types during the survey of the G. Mulu National Park, Sarawak (Holloway, 1984) and by Chey (1994) in softwood plantations near Brumas in Sabah.