Roy R. Snelling
Natural History Museum of Los Angeles County
900 Exposition Boulevard
Los Angeles, CA 90007, U. S. A.

SUMMARY

Active collecting for social Hymenoptera (the families Apidae, Formicidae, and Vespidae) extended from 20 October to 8 December, 1996. A variety of techniques was employed: active searching for individuals and nests in suitable locations, Malaise traps, litter sifting, and baiting. A total of 271 species of social Hymenoptera was collected, represented by approximately 3000 specimens; 250 of these species belong one family, the Formicidae (ants).

INTRODUCTION

The social Hymenoptera (ants, bees, and wasps) are an influential part of the biotic environment of most animal species in New Guinea, as they are in most tropical rain forest areas. In both absolute numbers and biomass the social Hymenoptera, especially ants, often dominate arthropod faunas of tropical rainforests, both in the canopy and on the ground (Davidson 1997; Wilson 1987). Recent studies estimate that about one-third of the entire animal biomass of the Amazonian terra firme rain forest is composed of ants and termites and that, along with bees and wasps, these insects comprise rather more than 75% of the total insect biomass (Fittkau and Klinge 1973). Other studies have suggested that much the same is true in African rain forest. While comparable studies have yet to be made in New Guinea, it is probable that similar results will be forthcoming.

At the bottom of the "food chain", ants, bees, and wasps represent an almost unlimited, renewable, highly nutritious protein resource for a great many species of insectivorous vertebrates, from frogs and snakes to birds and bats. Humans, too, long ago recognized the importance of these insects as food (Bodenheimer 1951; Taylor 1975).

Ants and social wasps are the chief predators on insects, spiders and other arthropods and, less commonly, on small vertebrates. The larvae of both groups are largely, if not entirely, carnivorous. Since colony populations of some species of ants are often in excess of 50,000 larvae, it follows that considerable quantities of insect prey are collected by the foraging workers in order to feed these larvae.

Ants ". . . form the cemetery squads of creatures their own size, collecting over 90 percent of the dead bodies as fodder to carry back to their nests. By transporting seeds for food and discarding some of them uneaten in and around the nests, they are responsible for the dispersal of large numbers of plant species. They move more soil than earthworms, and in the process circulate vast quantities of nutrients vital to the health of the land ecosystems." (Hölldobler and Wilson 1994)

The social Hymenoptera, together with the termites, in marked contrast to nearly all other terrestrial invertebrates in the Lakekamu area constitute a year-round source for study. Most other invertebrates are more or less seasonal: part of the year only adults are present and part of the year only larvae are present. Such seasonality makes it difficult to accurately survey a fauna in a short period of time. For this reason, the social Hymenoptera, with their stationary, perennial nests are excellent survey subjects. Furthermore, because both ants and wasps are predators, their populations are dependant upon the general ecological health of the region.

Some groups of insects, especially those that include conspicuous or "showy" species are reasonably well collected and well known. This is especially true for such groups as butterflies and certain families of beetles. However, the less gaudy insects are less often and less intensively collected, hence much less well known. The New Guinea Hymenoptera, despite their ecological importance, are relatively poorly known. There are no comprehensive accounts of any of the families including social species for New Guinea as a whole, or for any significant portion of the Island.

METHODS

Non-ants (Apidae and Vespidae) were collected principally by two methods: the netting of foraging individuals and the capture of entire colonies. Since Polistinae, our most speciose subfamily of Vespidae, most frequently nest on the undersides of leaves or on branches of low-growing shrubs, their nests were often discovered by various other groups of researchers and reported to me. I collected complete colonies of those belonging to species not previously collected or that were known to me from only a few netted individuals.

Six Malaise, or flight-intercept, traps were deployed at selected sites along the Bulldog Road and along the Ridge Trail. These traps collect flying insects directly into containers of alcohol and are an excellent method for sampling the fauna of a small area. The traps were placed in various situations: closed canopy forest, clearings, and at the edge of clearings in an effort to secure as great a diversity as possible. Traps were serviced at 10-day intervals and their contents stored in "Whirl- paks" for later examination upon my return to Los Angeles.

Several methods were utilized in the survey for ants. Individual samples resulted from direct examination of rotting logs and branches, searching under loose bark and bark flakes, under stones, splitting stems and twigs of living plants, examination of such myrmecophytes as Myrmecodia and Endospermum and epiphyte root masses, and by searching for foraging columns and individuals.

In addition, litter samples were taken. The coarse leaf litter was sifted out; the resultant accumulation of fine litter & duff was then divided among several coarse-meshed bags and placed in a collecting bag. As the litter dried out, the arthropods contained therein dropped out of the mesh bags into a container of alcohol at the bottom of the collecting bag. This method makes possible the collecting of samples of minute, (less than 1.5mm long), cryptic, or lethisimulating species inhabiting the leaf litter that would almost certainly be overlooked by direct examination.

Pitfalls were not used as a survey method. The results of limited pitfall traps set out by Kurt Merg's student group included only a few of the most common surface-foraging species. Canopy sampling was limited to tree falls and was, therefore, strictly opportunistic. Tree falls must be examined within about three days of falling, else the resident ants will have relocated to adjacent trees. Only one such tree fall was available during my stay at Ivimka, but it yielded several species not otherwise seen.

Samples of most of the ant species were separated out from the general collection and on December 16 I travelled to Canberra, Australia. Three weeks were spent in Canberra working with the ant collections of the Australian National Insect Collections (ANIC), probably the most extensive collection of New Guinea ants in any institution. There, with access to a microscope I was able to begin the process of sorting the minute species from the litter samples, and to begin the process of species identification. I was able to identify about one-third of my material as a result of having access to the ANIC collections and literature. Further identifications were possible once I returned to Los Angeles.

A primary set of voucher specimens from the Ivimka camp survey will be deposited in the Entomology Section of the Natural History Museum of Los Angeles County. Additional vouchers will be deposited, to the extent that duplicates are available, in the following institutions: Australian National Insect Collections, Canberra; B. P. Bishop Museum, Honolulu; Museum of Comparative Zoology, Cambridge, MA; The Natural History Museum, London.

RESULTS

APIDAE (Social bees)

When one speaks of "bees", most people tend to think of the social honeybees, Apis spp., esp. A. mellifera, not realizing that of the more than 2000 known bee species, most are non-social species. Bees, of course, are recognized for their importance as pollinators, not only of important crop plants, but of most of the flowering plants of the World. The deliberate manipulation of social bees as pollinators is a relatively recent phenomenon. The social species, which are mostly tropical and subtropical in distribution, have long been utilized as an important food resource. In addition to the honey, gathered by people all over the world where these bees occur, the larvae and pupae have also been recognized by traditional peoples as an important source of protein. The wax, of which the combs are constructed, has been utilized as an important household item.

Only four species of social bees were encountered at Ivimka, the Asian honeybee (Apis cerana) and three species of "sweat" or "stingless" bees (Trigona spp.). All were common and routinely attracted to perspiring humans or any other source of salts. These social bees presumably nest in tree hollows high above the ground. No nests were seen. At the time of this survey, there were few plants in bloom. Examination of those that were found yielded no additional species of social bees, although one large (up to 2.5 cm long) black species of leaf-cutter bee, Megachile clotho (F. Smith), was common at flowers of Lotus sp. along the Avi Avi River.

The New Guinea bee fauna is poorly studied, even for such relatively common and conspicuous bees as the Trigona species. At present there is no informed estimate as to the number of Trigona species present in New Guinea. At least some of the "Indo-Malayan" species treated by Schwarz (1939) are also present on New Guinea, as presumably is true of some now known from the Cape York Peninsula of Australia. Michener (1965) lists 10 species for New Guinea, but there are probably more.

Little can be said of the few species collected at Ivimka beyond the observation that all are forest dwelling species. Any activity that adversely affects the forest will probably similarly impact these social bees.

FORMICIDAE (ANTS)

A total of 266 samples of ants was made, 254 of which represent individual collections; the remaining 12 samples are from sifted litter. Two hundred and fifty species of ants belonging to 59 genera were collected at Ivimka, a much higher number than I had anticipated. In an area of lowland rain forest at the lower Busu River, Wilson (1959c) found 171 species in 51 genera. Wilson's study is the only prior base-line New Guinea work available for comparative purposes; the Busu River fauna examined by Wilson is notable for its Melanesian components. Wilson commented that "[t]he ant fauna of the Busu-Bupu area is perhaps the richest ever recorded for a single locality anywhere in the world . . . it has been estimated that at least 59 genera and 172 species occured within a few square kilometers in the collection area." [More recent systematic work has reduced Wilson's 59 genera to 51] The results of Wilson's study and mine are compared in the Appendix, Table 2.

AENICTINAE: The Aenictinae includes our only known true legionary or army ants, all belonging to the genus Aenictus. Wilson (1964) recorded seven New Guinea species of Aenictus, known only from workers; an additional two species, based on males only, were not treated. At Ivimka, workers of two species were collected. Based on the males, we have a minimum of five species at Ivimka. Because these male species cannot now be associated with workers, the present list is assumed to be inflated.

Legionary ants are group raiders that do not have established nests. As far as known they are specialized predators of other ant species (Wilson 1964). Colonies have a single queen and may number into the hundreds of thousands. Foraging columns of Aenictus workers were commonly seen during late afternoon as they filed across paths or other cleared areas

CERAPACHYINAE: Wilson (1959b) recorded 11 species of Cerapachys from New Guinea, 6 of which were noted to be present at the Lower Busu River site (not 4 as stated by Wilson 1959c).

Cerapachyines are generally considered to be rare; they are certainly unobtrusive and seldom seen. Colonies usually consist of a single queen and a few dozen workers. As far as known, all are predaceous; some, at least, are specialized predators on other ants, invading the nests and taking the brood as prey; others prey on termites. In the New World, Cerapachys have been found running in foraging columns of army ants, themselves predators on other ant species.

DOLICHODERINAE: At Ivimka, the subfamily Dolichoderinae includes only a few species of non-stinging ants, most of which are arboreal. When crushed, the ants produce a rank, disagreeable, highly repugnant odor, the principle component of which is often butyric acid. Colony defense is further enhanced by the aggressive nature of most of the species. When disturbed, the workers swarm out in large numbers and, while they cannot sting, their bites can be annoying.

Colonies are usually populous and seem always to have several queens. Most of our species nest in almost any suitable plant cavity, especially hollow branches, old termite galleries in dead logs, and under epiphyte root mats; at Ivimka the domatia of the myrmecophyte, Myrmecodia, were almost exclusively monopolized by Anonychomyrma scrutator. The species of Anonychomyrma are among the most often encountered of our ants: the workers forage on vegetation in long, dense files running up the trunks of trees & shrubs and across vines. The several species of both Anonychomyrma and Philidris, in addition to being general predator-scavengers, also tend aphids and mealybugs on various shrubs and trees. These Homoptera are often hidden under small shelters of plant fibers on the host plant. The two species of Leptomyrmex usually nest in soil or in rotten wood on or near the ground; both are aggressive predators.

Two of the species encountered at Ivimka are adventive. Tapinoma melanocephalum is a now tropicopolitan "tramp" species originally from Africa. Technomyrmex albipes is of Asian origin and now widespread from India to Australia and throughout much of Oceania; it is recently established in the United States (California). These are both "insinuators" that have apparently minimal impact on established native species.

FORMICINAE: Like the Dolichoderinae, Formicinae are stingless; they relay almost wholly upon chemical defense: when agitated they spray concentrated formic acid (so-named because it was originally derived from the Palearctic ants of the genus Formica). Formic acid is a powerful insecticidal and fungicidal fumigant. Some of the smaller species, especially those in genera such as Acropyga and Paratrechina rely more upon their small size and cryptic habits. One species of Camponotus, C. dorycus, is the largest ant at Ivimka; colonies are populous and, when disturbed, the ants aggressively defend the colony; the large workers (so-called soldiers) can draw blood when they bite. They add insult to injury when they spray formic acid into the wound.

The species of Acropyga and Paratrechina nest in decaying wood, as do some of the Camponotus and Polyrhachis. Other species of Camponotus nest in living plant stems, where they subsist largely on the exudates of homopterans that feed on the plants from within the stems. Many species of Polyrhachis build nests of masticated plant fibers (carton) or felted plant hairs (felt) that are attached to large leaves of palms or Pandanus. A few Polyrhachis, and Oecophylla smaragdina, build arboreal nests of living leaves joined together by the woven silk of the ant larvae.

Formicinae are mostly generalized predator-scavengers, but with a strong predeliction for plant-derived carbohydrates, either in the form of nectar from flowers or extra-floral nectaries or as secretions ("honeydew") from aphids, mealybugs, and other Homoptera. Larvae of certain families of butterflies (Lepidoptera: Lycaenidae and Noctuidae) are assiduously tended and protected by formicine ants, especially O. smaragdina; the ants thus protect herbivores that may do considerable localized damage to individual trees and shrubs.

Oecophylla smaragdina ranges from India to Australia, east to Taiwan. This "weaver ant" is well-known as an unusually aggressive predator that builds multidomous arboreal nests that may occupy several trees. In much of the range of this species in eastern Australia, the ants are green, rather than red as in our populations; these are the "green ants" of the Aboriginal "Dream Time".

MYRMICINAE: The Myrmicinae is our most speciose subfamily; while most of the species possess stings, they are mostly too small to deliver a painful sting. Only the species of Podomyrma are capable of an annoying sting. Other myrmicines, however, respond to disturbance in sufficiently large numbers as to be at least moderately annoying.

While some species, especially in the genera Crematogaster, Lordomyrma, and Podomyrma are arboreal, nesting in hollow stems and branches, most of the myrmicines nest in soil and litter or in rotting wood. Some of the species of Strumigenys nest under bark chips on Ficus trunks, while others nest in rotting wood.

Feeding habits are diverse: most Myrmicinae are generalist predator-scavengers and aggressively recruit to newly discovered food resources and drive other ants away. Some genera, such as Strumigenys and Trichoscapa are specialized predators, usually on minute litter-dwelling arthropods. Various species of Pheidole are probably seed-gatherers at Ivimka, as they are elsewhere in the world; some, however, are probably general scavenger- predators. About other ants, we know little, but assume such genera as Rhopalothrix, Eurhopalothrix, Dacetinops, and Pristomyrmex to be specialized predators also, but these are so poorly known that little can be said of them.

Cardiocondyla nuda and Monomorium floricola are exotic species that probably have little or no deleterious impact on the native species. Although originally described from Fiji, C. nuda is assumed to be African. Since the ant was found only near the Avi Avi River, but well removed from the camp area, it seems likely that this does not represent a recent introduction. Perhaps C. nuda was brought into the area during the Second World War on supplies being carried to Wau via the Bulldog Track.

Presumably originally a tropical Asian species, M. floricola is now a widespread tropicopolitan "tramp" species. According to Wilson and Taylor (1967), this ant is "almost wholly arboreal, forming large colonies in trees and bushes in habitats of various degrees of disturbance. It is a prominent urban species in most tropical countries. Colonies seem unable to penetrate undisturbed native forests." While the latter point is debatable, it generally seems to be true. At Ivimka, M. floricola is not a commonly encountered species and was found only in the temporary camp. It is possible that it was brought in recently on supplies carried in from Port Moresby.

PONERINAE: The Ponerinae, together with the Aenictinae and Cerapachyinae, are among the better known of the New Guinean ant fauna, largely due to the several taxonomic studies by Taylor (1967) and Wilson (1958a & b, 1959). Several difficult groups, especially the genera Hypoponera and Myopias, have yet to receive comprehensive taxonomic treatment. A single male specimen, collected in a Malaise trap is evidently the first New Guinea record of the rare genus Probolomyrmex.

Ponerine ants are largely predaceous and a few are evidently specialized predators. Several species of Myopias are specialists on millepedes (Myriapoda) (Willey and Brown 1983; pers. obs.). Others, such as species of Discothyrea and Proceratium (Brown 1958) and of Probolomyrmex (Taylor 1965) are suspected of being predators on eggs of various litter inhabiting arthropods. Most ponerines are solitary foragers, but some, such as Odontomachus, recruit very rapidly to a concentrated prey supply, such as exposed termites in a rotten log. The species of Leptogenys are nomadic and are group raiders, at least some of which (L. breviceps) prey largely on termites. On several occasions, workers of Rhytidoponera nexa were seen carrying moribund leeches; whether the leeches had been actively attacked is uncertain.

Most of our species nest in soil and/or leaf litter, often under a covering object such as a stone or piece of wood, or in rotting wood. A few, such as Diacamma and Odontomachus may be opportunistically "arboreal", nesting in preformed cavities in dead tree branches, usually within 5 meters of the ground. Dense root mats of epiphytes may also be utilized as nesting sites.

Ponerines characteristically are considered to be stinging ants and some of our species, especially in the genera Leptogenys, Odontomachus, and larger Pachycondyla are capable of inflicting quite painful stings. Others of similar size, such as the larger Myopias and Rhytidoponera, have surprisingly mild stings.

PSEUDOMYRMECINAE: This subfamily is largely tropical, with two genera (Pseudomyrmex and Myrcidris) in the New World and one genus (Tetraponera) in the Old World. All are arboreal and all are largely predaceous on other arthropods. Most species nest in hollow twigs and/or branches of trees and shrubs; a few species of Tetraponera are obligate residents of Asian bamboo species. While some of the largest species are capable of delivering a painful sting, most species are timid and innocuous.

The ecological stratification of the ants of the lower Busu River was described by Wilson (1959). He recognized three reasonably distinctive strata: (1) The Ground stratum comprises those species that nest in the soil, leaf litter, and all rotting wood on the ground, up to and including the largest rotting logs. The majority of both species and genera live at this level and relatively few species leave it to forage in the arboreal zones. (2) The low arboreal stratum includes species that nest in herbaceous and shrubby ground vegetation and up to the first several meters of the trunks and branches of larger trees. A small number of species nest primarily or exclusively in this zone. Species nesting in this stratum commonly forage down to the ground stratum and some species from the ground stratum forage into this level. (3) The high arboreal stratum species nest in the upper trunks and canopy of the A- and B-stratum trees; ants nests within epiphytes, abandoned termitaria, and preformed cavities in both living and dead wood, including twigs. Although most of these species confine their foraging activities to the high canopy, a few do forage all the way to the ground. In the following list of Ivimka ants, each species is characterized according to this same plan, the number following the name indicating 1-Ground stratum; 2-Low arboreal stratum; 2- High arboreal stratum.

Ants may be further characterized by their general feeding habits. Thus some ants may be specialist predators (S), general predators (G), pastoralists (P), or seed harvesters (H). As noted above, most species of the ponerine genus Myopias are presumed predators on millipedes; some species of Leptogenys are specialist predators, but no data are available for most of our species. Within the Myrmicinae, Strumigenys tend to specialize on Collembola, but apparently will capture other arthropods as well. Most of our species are general predators and scavengers, readily attacking living arthropods when feasible, often taking dead or dying prey, as well. These ants also will take other food items, including seeds, nectar, and other general scavenge. Pastoralists include our species of Acropyga, Philidris, Anonychomyrma, and at least one species of Podomyrma. None of the species observed at Ivimka was found to be a confirmed seed harvester, but it seems likely that one or more of the species of Pheidole may fall into this category.

In the list of the ants collected at Ivimka (Appendix, Table 1), some of the unidentified species are represented by alates (winged individuals) only. Except in the case of the several Aenictus species, these are winged females, usually attracted to lights at night. It is usually possible to associate females with their respective worker forms on the basis of similarities in sculpture and pilosity. In the absence of clear indications of likely associations, some of these females have been listed here as separate species. These belong to genera that include mostly high-arboreal species and are assumed to represent additional species.

VESPIDAE (Social Wasps)

All PNG social and subsocial wasps belong to the family Vespidae and are divided among three subfamilies, the Polistinae, Stenogastrinae and Vespinae (Snelling 1981). All three of these subfamilies are present at the Ivimka study site, with the cosmopolitan subfamily Polistinae clearly the dominant group in both species and general abundance. As has been noted by various authors (e.g., Gressitt 1959, 1961), the insect fauna of New Guinea is largely Oriental in its affinities. Three of our social wasp genera, Parapolybia, Polybioides and Vespa, are primarily Oriental and reach their southern terminus in New Guinea.

All of the social wasps are predators on other arthropods. Species Polistinae are mainly predaceous on larval Lepidoptera (caterpillars), but occasionally will take other prey as well. The one vespine genus at Lakekamu, Vespa, attacks the nests of other social Hymenoptera, especially the various Polistinae; the adult wasps are driven away from the nest; the Vespa worker then removes the larvae and pupae. Although Stenogastrinae are predatory, little detailed information is available. The hunting wasps are most often seen flying slowly about in dark and secluded areas. Some species are known to prey on chironomid midges (Diptera: Chironomidae) and, apparently, on spiders.

Four genera of Polistinae have been recorded from New Guinea: Parapolybia, Polistes, Polybioides, and Ropalidia, and all except Polybioides have been collected at Ivimka. Of these, Ropalidia is the most speciose, represented by 16 species.

Parapolybia has not been previously recorded from Papua New Guinea, although our one species was common at Ivimka. Vecht (1966) described P. varia ssp. furva from a few specimens collected in the Kebar Valley, Vogelkop Mountains of Irian Jaya. Our specimens are apparently very similar in color pattern, both darker and less abundantly yellow marked than in the more "typical" form which ranges from India to China and south to Borneo, Sumatra, and the Philippines. This wasp constructs single comb nests on the underside of leaves; fully mature nests may contain up to about 200 workers. These wasps are moderately aggressive and capable of delivering a painful sting.

Polistes is a cosmopolitan genus of social wasps that is best represented in the Afrotropical and in the Neotropical regions. The Polistes fauna of New Guinea is not well studied and the number of species present here is uncertain but probably there are about 12-15 species. Colonies usually are rather small, seldom exceeding 50 individuals. The nest consists of a single comb suspended from a branch or other supporting object and is never enclosed within an envelope. Lepidopterous larvae are the usual prey but the foraging workers will often attack other relatively soft-bodied insects. The wasps are generally not especially aggressive but are capable of delivering a painful sting.

One of our species is P. tepidus, a large black species with rich yellow markings. This species was originally described from Australia; Vecht (1971) treated the New Guinean form as a separate subspecies, P. tepidus malayanus Cameron, which, despite its name, was described from New Guinea (Manokwari, Irian Jaya). Pending a detailed study of more material from both regions I have opted to ignore any attempt at subspecific recognition. Another of the Ivimka species is P. bambusae described by Richards (1978) from Wau; according to Richards, this species nests in the hollow stems of large bamboo. I collected only two foraging workers. Two additional species remain unidentifiable at this time. One of these is apparently similar to P. laevigatissimus G. Soika (known only from the types from Broome, West Australia) and the other similar to P. riekii Richards, described from the Cape York Peninsula, Queensland.

The most speciose social wasp genus at Ivimka is Ropalidia. Unfortunately, the systematics of the New Guinean fauna is not well worked out. The pioneering work of Vecht (1941, 1962) on the Ropalidia of the Indo-Australian region was never completed, leaving the large and complex New Guinean fauna essentially untouched. One paper by Cheesman (1952) is virtually worthless and Richards (1978) treated a few of our forms in a cursory fashion. A thorough taxonomic study of the New Guinean representatives of this genus would be worthwhile. I estimate that there are perhaps as many as 50 species in New Guinea; with 16 species presently known from Ivimka, this fauna seems unusually rich and deserving of more intensive study.

Ropalidia is not only the most abundantly represented social wasp genus, it is the most diverse in nesting habits. Some species construct small, single comb nests that lack an envelope under leaves or tree branches. Others (R. pratti and R. cf. pratti) construct single comb nests, again lacking an envelope, in hollow logs or tree trunks or cavities in soil. At least one species (R. cf. zonata) constructs multiple comb nests that lack an envelope. Another species, R. deminutiva, attaches its enveloped, single comb nest to the underside of large leaves. A football-size multiple comb nest that is enclosed by an envelope is made by Ropalidia sp. D; these nests are found hanging on shrubs or low branches. Mature nests of this species may contain several hundred workers.

The subfamily Stenogastrinae is a small one with about half a dozen genera and fewer than 50 species. Various genera of stenogastrines range from India to China, south to New Guinea; we have the genus Anischnogaster represented in New Guinea by several species. Two species are present at Lakekamu, A. iridipennis (F. Smith) and a presently undescribed species; the subfamily is currently being revised.

Stenogastrinae are rare, secretive wasps that are seldom collected. The few species are primitively social. Several females may be present on a nest, but each constructs her own cells and is responsible for the larva that develops therein. The nests are very simple: a single series of cells arranged in linear fashion along a thread-like stalk. Nests are situated in dark locations, such as the overhangs along river banks, in caves, etc., where they are concealed among hanging rootlets. The wasps are timid and make no effort to defend the nests.

The subfamily Vespinae reaches its southern terminus in New Guinea, where there are two species of Vespa, V. affinis (Linn‚) and V. trimeres Vecht. At Ivimka, V. trimeres was present, but seldom encountered. Little is known of the ecology of this ground-nesting species, but it is presumably a predator on other social Hymenoptera, a trait common to many species of Vespa (Spradbery 1973; Starr 1992; Vecht 1957). Several polistine nests with only very small larval cells were seen that were occupied by a queen and up to six workers; such nests apparently are initiated by the survivors of a nest depredated by a foraging Vespa worker. I observed other polistine nests in Hong Kong that had been attacked by foraging Vespa: the polistines made no effort to defend the nest, but abandoned it to the much larger hornet, which proceeded to remove the larvae. The survivors initiated a new nest at a nearby location.

Foraging Vespa workers were occasionally seen flying slowly through dense vegetation near ground level, apparently searching for prey. One nest was located in the ground at the base of a tree, but was not collected.

DISCUSSION

While the social bee fauna was within expected limits of abundance and diversity, the ants and social wasps far exceeded expectations. With respect to the ants, I can paraphrase Wilson's (1959) comment on the ant fauna of the lower Busu River: The Ivimka ant fauna is possibly the richest ever recorded for a single locality anywhere in the world! This is all the more impressive when one considers that all the material collected at Ivimka was taken no more than 50 meters (and mostly within 10 meters) from any trail and that the aggregate area surveyed was almost certainly less than one km2. Equally impressive is that the acquisition of additional species never ceased; I was acquiring additional ant species up to the last collecting day at Ivimka, a virtual guarantee that not all the species available at that site had been collected.

Similarly, the social wasp fauna seems to be extraordinarily rich, at least within the Old World tropics. This is no doubt due, at least in part, to the fact that no concerted surveys have been undertaken to determine the diversity of social wasps in any limited area of New Guinea. However, during the period from 5 June-15 July 1996, I collected social wasps in Hong Kong, a much larger area; a total of 17 species was collected. This number could possibly be extended by from 1-3 additional species that might ultimately be found there.

Ivimka supports a rich and varied fauna of ants and social wasps, certainly the richest documented for any locality in New Guinea. In large part this diversity is due to the relatively undisturbed nature of Ivimka and the nearly complete lack of adventive species. Species exotic to the area are apparently unable to thrive in such habitats. In contrast, introduced ants predominate at Tekedu, a bare 12 km away.

Important follow-up studies at Ivimka should focus on, and monitor, the influence of non-native (adventive) ant species. I predict that so long as the area remains no more disturbed than it is at present, the few introduced species should have little or no impact on the native ant fauna, nor on other faunal or floristic constituents at Ivimka. The current mix of non-native species consists entirely of "insinuator" species that occupy marginal niches and do not normally offer serious threats to native species so long as the area is minimally impacted by human activities.

However, great care must be exercised to assure that potentially more destructive species, such as Fire Ants (Solenopsis geminata, S. wagneri), do not become introduced into the Basin. These polygynous, highly aggressive New World species not only compete successfully with native ants, they often may be a threat to nestling birds, especially those on or near the ground; similarly, immature mammals may also be endangered. S. geminata, in particular, can be a problem for two reasons. First, it is already widespread in the Oriental and Melanesian regions. Secondly, it is a tropical forest species, hence pre- adapted to the conditions existing at Ivimka and, indeed, throughout lowland areas of New Guinea.

Although not as conspicuous as vertebrates, ants are a dominant element in any tropical ecosystem and the overall health of any given habitat is dependant on the ants resident there. However, the r"le of the many species and/or feeding guilds is poorly understood. Although I managed to collect at least a dozen species of Pheidole, nests of most were not located and I do not know which, if any, are seed-gathering species. Almost two dozen species of Polyrhachis are present at Ivimka. Of their biologies, almost nothing is known. The ants at Ivimka could provide a study resource for ecologists for years to come.

LITERATURE CITED 

• Bodenheimer, F. S. 1951. Insects as human food. Dr. W. Junk Publ., The Hague, 352 pp.
• Bolton, B. 1977. The ant tribe Tetramoriini (Hymenoptera: Formicidae). The genus Tetramorium Mayr in the Oriental and Indo-Australian regions, and in Australia. Bull. Brit. Mus. (Nat. Hist.) (Entomol.) 36:67-151.
• Bolton, B. 1995. A new general catalogue of the ants of the World. Harvard University Press, Cambridge, MA., 504 pp.
• Brown, W. L., Jr. 1958. Predation of arthropod eggs by the ant genera Proceratium and Discothyrea. Psyche 64:115.
• Cheesman, L. E. 1951. A collection of Polistes from Papuasia in the British Museum. Ann. Mag. Nat. Hist. (12) 4:982- 993.
• Cheesman, L. E. 1952. Ropalidia of Papuasia. Ann. Mag. Nat. Hist. (12) 5:1-26.
• Davidson, D. W. 1997. The role of resource imbalances in the evolutionary ecology of tropical arboreal ants. Biol. J. Linn. Soc. (in press)
• Fittkau, E. J., and H. Klinge. 1973. On biomass and trophic structure of the central Amazonian rain forest ecosystem. Biotropica 5:2-14.
• Gressitt, J. L. 1959. Wallace's line and insect distribution. Proc. XVth. Internatl. Congr. Zool. 1958, pp. 66-68.
• Gressitt, J. L.  1961. Problems in the zoogeography of Pacific and Antarctic insects. Pacific Insect Monogr. 2:1-94.
• Hölldobler, B., and E. O. Wilson. 1990. The ants. Belknap Press, Harvard Univ., Cambridge, MA., 732 pp.
• Hölldobler, B., and E. O. Wilson. 1994. Journey to the ants. Belknap Press, Harvard Univ., Cambridge, MA., 228 pp.
• Michener, C. D. 1965. A classification of the bees of the Australian and South Pacific Regions. Bull. Amer. Mus. Nat. Hist. 130:1-362.
• Richards, O. W. 1978. The Australian social wasps (Hymenoptera: Vespidae). Austr. Jour. Zool., suppl. ser., no. 61, 1-132.
• Schwarz, H. F. 1939. The Indo-Malayan species of Trigona. Bull. Amer. Mus. Nat. Hist. 76:83-141.
• Shattuck, S. O. 1992. Generic revision of the ant subfamily Dolichoderinae (Hymenoptera: Formicidae). Sociobiol. 21:1- 181.
• Snelling, R. R. 1981. Systematics of social Hymenoptera, IN: Social Insects (H. Hermann, ed.) 2:369-453. Academic Press, Inc., New York.
• Spradbery, J. P. 1973. Wasps: an account of the biology and natural history of solitary and social wasps. Univ. Washington Press, Seattle, WA, 408 pp.
• Starr, C. K. 1992. The social wasps (Hymenoptera: Vespidae) of Taiwan. Bull. Natl. Mus. Nat. Sci. (Taiwan) 3:93-138.
• Taylor, R. L. 1975. Butterflies in my stomach. Or: insects in human nutrition. Woodbridge Press Publ. Co., Santa Barbara, CA, 224 pp.
• Taylor, R. W. 1965. A monographic revision of the rare tropicopolitan ant genus Probolomyrmex Mayr (Hymenoptera: Formicidae). Trans. Roy. Entomol. Soc. Lond. 117:345-365.
• Taylor, R. W. 1967. A monographic revision of the ant genus Ponera Latreille (Hymenoptera: Formicidae). Pac. Insects Monogr. 13, 112 pp.
• Vecht, J. van der. 1957. The Vespinae of the Indo-Malayan and Papuan areas (Hymenoptera, Vespidae). Zool. Verh. Rijkmus. Nat. Hist. Leiden 34:1-83.
• Vecht, J. van der. 1966. The East-Asiatic and Indo-Australian species of Polybioides Buysson and Parapolybia Saussure (Hym., Vespidae). Zool. Verhandl. 82:1-42.
• Vecht, J. van der. 1971. The subgenera Megapolistes and Stenopolistes in the Solomon Islands (Hymenoptera, Vespidae, Polistes Latreille). Entomol. Essays Comm. Ret. Prof. K. Yasumatsu, pp. 87-106.
• Willey, R. B., and W. L. Brown, Jr. 1983. New species of the ant genus Myopias (Hymenoptera: Formicidae: Ponerinae). Psyche 90:249-285.
• Wilson, E. O. 1958a. Studies on the ant fauna of Melanesia. I. The tribe Leptogenyini. II. The tribes Amblyoponini and Platythyreini. Bull. Mus. Comp. Zool. 118:101-153.
• Wilson, E. O. 1958b. Studies on the ant fauna of Melanesia. III. Rhytidoponera in western Melanesia and the Moluccas. IV. The tribe Ponerini. Bull. Mus. Comp. Zool. 119:303-371.
• Wilson, E. O. 1959a. Studies on the ant fauna of Melanesia. V. The tribe Odontomachini. Bull. Mus. Comp. Zool. 120:483-510.
• Wilson, E. O. 1959b. Studies on the ant fauna of Melansia. VI. The tribe Cerapachyini. Pac. Insects 1:39-57.
• Wilson, E. O. 1959c. Some ecological characteristics of ants in New Guinea rain forests. Ecology 40:437-447.
• Wilson, E. O. 1964. The true army ants of the Indo-Australian area (Hymenoptera: Formicidae: Dorylinae). Pac. Insects 6:427- 483.
• Wilson, E. O. 1987. The arboreal ant fauna of Peruvian Amazon forests: a first assessment. Biotrop. 19:245-251.
• Wilson, E. O., and R. W. Taylor. 1967. The ants of Polynesia. Pac. Insects Monogr. No. 14, 109 pp. 

APPENDIX

Table 1. Social Hymenoptera collected at Ivimka camp, Lakekamu Basin, Papua New Guinea.