This is, chronologically, the final part to be published on the Geometridae of Borneo. It is therefore a convenient place for a review of the family as a whole, particularly with regard to generalities such as: the diversity of the Bornean fauna and its biogeographic character relative to that of the Oriental tropics and the world totals for Geometridae; ecological segregation of the various higher taxa that may provide pointers to adaptive trends, including host-plant specialisations; a summary of morphological features that may be important in resolving problems in the higher classification of the family, particularly the position and interrelationship of the two subfamilies treated in detail, the Sterrhinae and Larentiinae.

The diversity of the Bornean Geometridae in a world context
A recent survey of diversity in world Geometridae (Scoble, Gaston & Crook, 1995) enables the richness of the Bornean fauna to be compared, subfamily by subfamily and in total, with that of both the Oriental tropics and the world as a whole. This is of practical interest as it can give an idea of how useful a “localised” faunistic series such as this can be over a wider geographical area to facilitate at least preliminary identification work on neighbouring faunas.

The analysis in the next section also indicates this wider relevance, as it highlights representation of different biogeographic categories in the fauna. Thus, a treatise on Borneo moths should include all or most species: widespread in the Indo-Australian tropics; widespread in the Oriental tropics; with more restricted S.E. Asian plus Sundanian distributions; limited to Sundaland; endemic to Borneo. These categories are more or less nested one within the next in reverse order. In addition there are a small number of species shared between Borneo, the Philippines and Sulawesi, and also extending further east to the Moluccas and New Guinea. These are particularly frequent amongst more montane taxa (Holloway, 1970) as will be seen in this part in the sterrhine genus Cyclophora Hübner and extensively in the Larentiinae.

Some genera and species groups in Larentiinae have their centres of species richness in the Australasian tropics, particularly New Guinea, though one, Poecilasthena Warren, is most diverse in Australia itself. In Tripteridia Warren and Tympanota Warren the species concerned are mostly of the widespread montane type referred to in the previous paragraph. In Scotocyma Turner and the Collix intrepida Prout group the Bornean species is endemic but clearly related to the Australasian taxa. In Poecilasthena there is a Bornean endemic and a Sundanian species, but they belong to a widespread Indo-Australian species group that extends from Burma to Fiji and New Caledonia. The new genus Papuarisme is distinctive in having a small group of Bornean endemics that appear to form a monophyletic group that is possibly sister to a larger Australasian complex, mostly found in the mountains of New Guinea.

A pair of species in this Papuarisme group has the characteristic Bornean duplex pattern noted by Holloway (1970, 1986b, 1993[4]) where a Sundanian or widespread Bornean montane species is sister to one restricted to G. Kinabalu. The latter replaces the former at higher altitudes and is often larger. In addition to the Papuarisme pair, there are further examples in Hypocometa Warren and Poecilasthena in the Larentiinae. Many other geometrid species have only been found on G. Kinabalu, particularly Larentiinae. A list was presented by Holloway (1996) but this is updated in Table 1 in the light of taxonomic changes made in this volume.

Table 1: Endemic Bornean Geometridae only recorded from G. Kinabalu. Those with an asterisk are represented only by single specimens.

Desmobathrinae
Ozola prouti Holloway	Ozola submontana Holloway
Geometrinae
Paramaxates spinivesica Holloway*	Idiochlora berwicki Holloway*
Idiochlora stictogramma Prout
Sterrhinae
Chrysocraspeda truncipennis Prout*	Scopula brookesae Holloway
Perixera melantroches Prout	Scopula quadratisparsa Holloway
Idaea galsworthyi sp. n.
Larentiinae
Goniopteroloba solivaga Prout	Axinoptera penataran sp.n
Tristeirometa bathylima Prout	Gymnoscelis transapicalis Holloway
Hypocometa titanis Prout	Calluga punctinervis Holloway*
Phthonoloba bracteola Holloway	Micrulia subzebrina sp.n*
Phthonoloba stigmatephora Prout	Symmimetis kolopis sp.n
Phthonoloba caliginosa Holloway*	Poecilasthena nubivaga Prout
Phthonoloba altissima Holloway	Xanthorhoe mesilauensis Holloway
Sauris quassa Prout*	Xanthorhoe liwagu sp.n
Ziridava asterota Prout*	Ecliptopera zaes Prout
Pasiphila coelica Prout	Ecliptopera furvoides Thierry-Mieg
Pasiphila sayata Holloway	Dysstroma pendleburyi Prout
Pasiphila luteata Holloway	Lampropteryx moroessa Prout*
Pasiphila eurystalides Prout	Papuarisme lagadani sp.n
Syncosmia layanga Holloway	Papuarisme maerens Holloway*
Syncosmia discisuffusa Holloway
Ennominae
Achrosis classeyi Holloway*	Milionia pendleburyi Prout
Garaeus altapicata Holloway	Apophyga altapona Holloway
Synegia punctivervis Holloway	Catoria proicyrta Prout
Entomopteryx statheuta Prout	Bornealcis versicolor Prout
Astygisa waterstradti Holloway	Myrioblephara geniculata Prout

The predominantly montane character of the tropical Larentiinae was emphasised by Holloway (1986b), who also reviewed the general biogeography of the group in the Indo-Australian tropics, with mention also of the montane Geometridae. This particular character of the Larentiinae is illustrated in an analysis later in this introductory section. The montane ecosystems of Borneo are notable for a mingling of north-temperate and south-temperate elements, with also a small number of intrinsically tropical montane groups. G. Kinabalu is most notable in this respect (Beaman, 1996; Corner, 1996).

As the Lepidoptera are mainly herbivorous as larvae, often very specifically so, the question arises of whether the specialist herbivorous faunas of the various phytogeographic elements in the Bornean montane biota are of a similar biogeographic character: fellow-travellers rather than local colonists. This still needs to be tested by rigorous studies of the plants themselves, but there are indications of at least some concordance amongst the Larentiinae. The Australasian Tympanota Warren and Poecilasthena appear to be associated with Podocarpaceae and Leptospermum (Myrtaceae) respectively, and the north-temperate subgenus Gymnodisca Warren of Pasiphila Meyrick is associated with the Ericaceae of similar biogeographic character. Typical Pasiphila in New Zealand feed on other plant families, e.g. Rubiaceae and Scrophulariaceae (Dugdale, 1975). Milionia Walker in the Ennominae (Holloway, 1993[4]) is another example of an Australasian tropical group associated with southern conifers. Unfortunately, our knowledge of the host-plant requirements of tropical geometrids is far too scanty to permit a more detailed analysis of such trends.

The rich Sundanian and Oriental tropical lowland fauna, together with this biogeographically varied montane one both contribute to the high representation within Borneo of Indo-Australian geometrid diversity as a whole. Table 2 compares the Bornean fauna with world and Oriental totals given in Scoble, Gaston & Crook (1995). The totals for Oenochrominae are for the old concept of this group and thus include the true Oenochrominae, Desmobathrinae and minor groups suggested to be Ennominae, such as Orthostixinae, in Holloway (1996a). Scoble et al. also attempted to estimate real species richness from considering the history of description of Geometridae and descriptions of new taxa and synonymies from recent revisions of Neotropical groups. Similar data are presented for Borneo in Table 3.

The total geometrid fauna of Borneo is approximately one quarter of that recorded by Scoble et al. for the Oriental Region as a whole and just under half the total for all Australia listed by Nielsen, Edwards & Rangsi (1996). Bornean "Oenochrominae" , Geometrinae and Sterrhinae represent almost a third of the Oriental total, whereas the Larentiinae represent under one fifth. The proportion of endemics to widespread taxa is relatively low in the first three, higher in the Larentiinae, with the bulk of the Oriental larentiine fauna being located in mountain regions of the Himalaya and western China rather than in the tropical lowlands and lower montane zones.

Table 2. Bornean Geometridae as a percentage of Oriental and world totals determined by Scoble, Gaston & Crook (1995). The concept of Oenochrominae is 'traditional' and includes Desmobathrinae and other minor groups (Holloway, 1996a).

	Borneo	Oriental	World	%Oriental	%World
Ennominae	432	1846	9710	23	4
"Oenochrominae"	54	136	610	40	9
Geometrinae	218	584	2296	37	9
Larentiinae	199	1042	5749	19	3
Sterrhinae	176	543	2763	33	6
TOTAL	1079	4151	21144	26	5

Table 3. New species described, new synonymies made and Bornean species revived from synonymy in parts of The Moths of Borneo. Figures in brackets express these as percentages of the total species.

	Total species	New species (%)	New synonyms (%)	Revived species (%)
Ennominae	432	83 (19)	56 (13)	13 (3)
"Oenochrominae"	54	5 (9)	2 (4)	1 (2)
Geometrinae	217	48 (22)	18 (8)	9 (4)
Larentiinae	199	29 (15)	21(10)	6 (3)
Sterrhinae	176	21 (12)	24 (13)	10 (6)
TOTAL	1079	186 (17)	121 (11)	39 (4)

The percentage of new species described plus revived species is approximately equivalent to the mean level noted by Scoble et al. for a limited number of Neotropical geometrid genera that have been reviewed recently. The total for Bornean Larentiinae would have been higher but for the large number already described in the survey of the fauna of G. Kinabalu by Holloway (1976). However, the level of synonymy is less than half that for the Neotropical reviews. These statistics support the assertion by Scoble et al. that “there are no signs that the total number of species is set to rise by anything remotely like an order of magnitude” The indications are that the global total of just over 21000 noted by Scoble et al. would be unlikely to even double in the event of a complete inventory of species.

The ecological and biogeographic representation of Bornean geometrid groups
Holloway & Barlow (1992) and Holloway (1994) presented analyses comparing biogeographic and ecological representation across various Bornean moth groups that portrayed the proportions of, for example, endemic species restricted to lowland forest and the same for montane forests, versus, at the opposite extreme, species widespread through the Indo-Australian tropics that favour open or disturbed vegetation types.

The raw data were presented in the form of two-way tables, the rows representing biogeographic categories and the columns ecological ones. Tables for various ennomine groups were included in the publications cited: the Ennominae as a whole, groups covered by Holloway (1996a) and in this part are tabulated in Fig 1. Summary data, with some categories combined (all montane species together for ecological categories; Sunda with Wallacean and Himalayan with wide Oriental for biogeographic categories), are presented for all major geometrid groups in Tables 4 and 5, ranked according to montane representation (Table 4) and proportions of endemics (Table 5).

Table 4. Percentages of broad habitat categories for various geometrid groups in Borneo, ranked according to percentage of montane species. The figures have been rounded to one decimal place so may not exactly total 100.

	Lowland	Montane	Lowland and montane	Secondary vegetation and open habitats
Larentiinae: Trichopterygini	11.7	78.5	9.5	0
Larentiinae excluding Trichopterygini	25.6	56.0	17.1	1.3
Ennominae: Boarmiini	39.9	45.2	13.5	1.4
All Ennominae	43.9	39.0	15.7	1.4
Ennominae: Hypochrosini	53.3	34.2	11.7	0.8
Desmobathrinae	54.8	30.9	7.1	7.1
Geometrinae: Hemitheiti	52.4	29.2	13.6	4.9
Sterrhinae: Cosymbiini lineage	63.3	25.5	5.6	5.6
Geometrinae excluding Hemitheiti	53.2	24.7	18.3	3.6
Ennominae: Cassymini	62.2	19.0	18.9	0
Sterrhinae: Sterrhini lineage	60.8	16.3	13.5	9.5

Table 5. Percentages of biogeographic categories for various geometrid groups in Borneo. The Sunda / Wallacea category includes Sundanian species and those shared between Borneo or Sundaland and Wallacea. The wide Oriental category includes all species shared with mainland Asia. The groups are ranked according to percentage endemism. The figures have been rounded to one decimal place so may not exactly total 100 in each row.

	Endemic	Sunda/Wallacea	Wide Oriental	Indo-Australia
Larentiinae: Trichopterygini	40.5	45.2	9.6	4.8
Ennominae: Hypochrosini	31.7	40.0	25.0	3.3
Larentiinae excluding Trichopterygini	31.5	30.9	21.7	15.8
Ennominae: Boarmiini	31.1	39.2	23.0	6.8
All Ennominae	30.0	41.2	23.0	5.8
Geometrinae: Hemitheiti	29.1	30.1	24.3	16.5
Sterrhinae: Sterrhini lineage	27.0	33.9	31.1	8.1
Sterrhinae: Cosymbiini lineage	26.7	34.5	23.4	15.6
Desmobathrinae	23.8	40.4	30.0	4.8
Geometrinae excluding Hemitheiti	9.2	50.4	33.9	6.4

Figure 1. Percentage of species for geometrid higher taxa amongst various biogeographic and ecological categories as discussed in the text.

The publications cited also included cluster analyses of the groups derived from coefficients comparing their similarities of representation across the various categories in the two-way table. This analysis showed that most Bornean moth groups had a high proportion of lowland forest species, either endemic or with general Sundanian distributions.

Groups departing from this general trend were either those with increased representation of montane species, such as the Arctiinae, or those with high representation of widespread species, with low levels of endemism, and greater representation in the disturbed and open habitat category, such as the Sphingidae and stictopterine Noctuidae. The higher trifine subfamilies of the Noctuidae occupied an extreme position, having high proportions of both montane endemics and of widespread open habitat species, including numerous crop pests, both perhaps arising through a trend towards feeding on herbaceous plants enabling them to radiate into habitats at high latitudes and altitudes where such plants are more frequent, and also early stage successional habitats in the lowlands.

A similar analysis of a range of geometrid groups, based on the ennomine tables published by Holloway & Barlow (1992) and those in Fig 1 again showed a large general cluster of groups with a high proportion of lowland species, endemic or Sundanian, with exclusion of the two larentiine groups. These are distinguished by very high representation of montane species, and clustered in to the main grouping via the Boarmiini where montane species are also numerous (see also Table 4). They also have a high proportion of endemic species. The non-trichopterygine Larentiinae resemble the higher trifine noctuids in having strong representation at high latitudes also, facilitated in part by adoption of herbaceous feeding in the larvae.

Some minor trends of interest can be seen in Tables 4 and 5. The highest proportions of very widespread species are seen in the nontrichopterygine larentiines, the Hemitheiti in the geometrines and the cosymbiine lineage of the Sterrhinae. The first two groups include a high proportion of species where the larvae are flower-feeding, e.g. the Eupitheciini in the larentiine group. There may be some correlation between generalist flower-feeding and dispersal ability, though only the eupitheciines have been successful in colonising remote oceanic islands. A few cosymbiines have achieved this and species of Scopula Schrank in the other sterrhine lineage. This lineage also has the highest proportion of species in the disturbed and open habitat category in Table 4, perhaps a reflection on the herbaceous larval feeding habits in many species of Scopula and Idaea Treitschke, genera that have also been successful in open habitats at higher latitudes though this is not accompanied, in Scopula at least, by high representation in tropical montane systems.

A tentative classification of the Geometridae

Fig 2 presents a tentative phylogeny of the family Geometridae, biased somewhat towards groups represented in the Oriental tropics but including also (square brackets) groups from elsewhere that were at least partially investigated or where data in the literature provided an indication of affinities. Detail within the subfamilies is as suggested by Holloway (1993 [4], 1996a) and in the subfamily accounts of this volume. Relationships between them are the topic of this section.

Figure 2. Tentative phylogeny for the Geometridae. Non-Bornean tribes included are indicated by square brackets. The subfamilies Alsophilinae and Archiearinae are not represented in Borneo.

These relationships are established mostly on characters of the adult male and female abdomen. Characteristics of early stages have been of value for recognising groupings within subfamilies, but data are too scattered and the approaches in the literature to documentation too varied to permit an overview of the whole family, a research project well beyond the scope of this series. Patocka (1994) has provided a key to the pupae of a high proportion of European geometrids.

The two major lineages in Fig 2 are segregated primarily on male secondary sexual characters. In the Larentiinae and Sterrhinae lineage, modifications to the abdomen, when present, occur on the second sternite (Trichopterygini, Sterrhini lineage) or in the form of coremata more distally on the abdomen (Cosymbiini, other Larentiinae). In the Ennominae, Geometrinae and Desmobathrinae lineage they occur on the third sternite in the form of a transverse comb of setae (Ennominae) or a pair of lateral patches of setae (the other two groups). These features usually occur in conjunction with a hair-pencil sheathed in the hind-tibia, though this is absent from most larentiines (the trichopterygine genus Phthonoloba Warren is an exception) and the Cosymbiini lineage.

Features of the tympanal organs may also indicate a relationship between the Larentiinae and Sterrhinae, such as the hammer-headed ansa and the absence of a tympanic lacinia in most (Cook & Scoble, 1992). The bursa copulatrix in the two subfamilies tends to be generally scobinate, spined or rugose, rarely with a definite signum (this, when it occurs, possibly arising through a coalescence of the general spining). In most Ennominae, Geometrinae and the Archiearinae there is a definite signum. Exceptions within the Ennominae/Geometrinae grouping are the Desmobathrinae, where a great diversity of bursa ornamentation occurs, and the ennomine tribe Plutodini where spining is general.

The sister-relationship of the Geometrinae and Desmobathrinae was discussed by Holloway (1996a). Apart from the setal character on the male third sternite, three out of four tribes included show very high concentrations of geoverdin wing pigment.

The Orthostixinae are placed as sister-group to the rest of the Ennominae, the latter grouped on absence of vein M2 in the hindwing. The Orthostixinae have been given subfamily status formally by Hausmann (1996). The genus Heteralex Warren may likewise be a basal branch of the Ennominae (Holloway, 1996a).

Relationships between the two major lineages discussed above and the small subfamilies Oenochrominae and Alsophilinae are shown unresolved in the phylogeny. These smaller families all exhibit incomplete reduction of abdominal prolegs to a single pair on segment A6 (as well as the anal claspers), but incomplete reduction also occurs in some southern hemisphere Ennominae (Common, 1990; Scoble, 1992).

In the Archiearinae, shown in a sister-relationship to the rest of the family, the labial palps are exposed in the pupa (Nakamura, 1987), in contrast to the rest of the family (Scoble, 1992), and the cremaster is distinctively T-shaped, the setae two-segmented (illustrated by Nakamura). Mosher (1916) noted a T-shaped cremaster also in Alsophilinae, but Nakamura suggested the latter were closer in general pupal characters to the Ennominae. It is possible, therefore, that the Archiearinae are sister-group to the rest of the Geometridae if exposure of the labial palpi is plesiomorphic. This relationship is also supported by the lack in Archiearinae of the accessory tympanum seen in all other Geometridae (Cook & Scoble, 1992).

The family as a whole is defined by the structure of the tympanal organs, in particular the presence of the tympanic handle or ansa (Minet, 1983; Cook & Scoble, 1992; Scoble, 1992). The Geometridae and Uranioidea are set apart from the Drepanoidea, Bombycoidea and Noctuoidea by possession of chaetosemata, though these are found widely amongst other ditrysian superfamilies and in the butterflies. The inchworm geometer larva, with abdominal prolegs reduced to those on A6 and A10, is also characteristic of the family, though not universal as discussed above.

More detailed descriptions of the family may be found in Common (1990) and Scoble (1992). Comment here is restricted to diagnostic features.