Two new cost-effective trap designs for capturing arthropod flower visitors are described. Small flower visitors such as Thysanoptera and Diptera, hardly noticeable during direct flower observations, were dominant in trap catches. These trapping techniques combined with direct observations allow a more comprehensive picture of the flower visitors.

Although the pollination biology of many individual plant species has been investigated in the Oriental region, there have been very few community-level studies. The two most comprehensive of these were in the primary mixed dipterocarp forest of Lambir Hills National Park, Sarawak (4°20′N: Momose et al. 1998, Sakai et al. 1999) and in the warm temperate evergreen broad-leaved forest and cool temperate mixed forest on Yakushima Island (30°N: Yumoto 1987, 1988). Hong Kong (22°17′N) lies midway between these sites, at the northern margin of the tropics, where winter temperatures fall below 10 °C at sea-level for a few days every year and there are occasional frosts above 400 m (Dudgeon & Corlett 1994). Latitudinal effects, however, are compounded in comparisons with other well-studied East Asian sites, by centuries of massive human impact, leaving a degraded landscape of steep, eroded hillsides, covered in fire-maintained grassland, secondary shrublands and, locally, secondary forests (Zhuang & Corlett 1997). This history has left a relatively impoverished fauna but a surprisingly diverse flora, including 400 native tree species (Corlett & Turner 1997). In these circumstances, failures of pollination and dispersal mutualisms might be expected to accelerate the loss of plant species from the landscape (Bond 1994, Kearns & Inouye 1997). Previous studies have shown that most woody vegetation in Hong Kong is dominated by species whose seeds can be dispersed by the commonest avian frugivores, the light-vented and red-whiskered bulbuls (Pycnonotus sinensis (Gmelin) and P. jocosus (Linn.)) and the Japanese white-eye (Zosterops japonicus Swinhoe) (Corlett 1996, 1998), but there is no equivalent information available on pollination biology.

Primary flower resources to bees, i.e. those that satisfy basic needs, such asfeeding and reproduction (Faegri & van der Pijl 1979), include pollen, nectar,oil, resin and gums. Resins are used in nest construction as a waterproof andstructural material (Simpson & Neff 1981). It is also excellent antimicrobialcompound, and a sticky defence against ant attacks (Armbruster 1984, Dressler1982, Lokvam & Braddock 1999, Michener 1990, Roubik 1989, Sakagami etal. 1989). Plant species with flowers which produce terpenoid resins are rare(Armbruster & Webster 1979). Clusia L. (Clusiaceae), Clusiella Planch. &Triana (Clusiaceae), and Dalechampia L. (Euphorbiaceae) are the only knownplants that secrete resins from floral structures (Armbruster 1984,Armbruster & Webster 1979, Bittrich & Amaral 1996). Althoughresin-producing plants provide an essential resource for a large number ofbees in tropical communities, there are few studies about resin-collectingbees on Clusia spp. Skutch (1971) was the first to report visits onClusia by stingless bees to collect floral resin. Other reportsinclude observations on C. odorata Seem. (Croat 1978, Ramirez &Gomez 1978), C. alata Pl. & Tr. (Ramirez & Gomez 1978), andC. nemorosa G. F. W. Meyer (Armbruster 1984, Bittrich & Amaral1996, Lopes & Machado 1998, Mesquita & Fransciscon 1995). Recently, Bittrich& Amaral (1996, 1997) have described the flower morphology and pollination biologyof other Clusia species from Central America and Venezuela.