RESEARCH NOTE Both Lutzomyia intermedia sensu lato (Lutz & Neiva) and its closest relative Lutzomyia whitmani (Autunes & Countinho, 1939) are incriminated vectors of Leishmania (Viannia) braziliensis (Vianna, 1911) in Brazil (LM Deane & G Grimaldi 1985, p. 247-281. In K-P Chang & RS Bray (eds), Leishmaniasis, Elsevier, Amsterdam). Although the geographical ranges of the species overlap considerably, work over the last 50 years has shown that Lu. intermedia is more synanthropic and independent from forest (e.g. OP Forattini et al. 1976 Rev Saúde Públ S Paulo 10: 31-43, AC Gomes et al. 1980 Rev Saúde Públ S Paulo 14: 54-56, 1989 Rev Inst Med Trop S Paulo 31: 32-39). In addition, Lu. intermedia is highly endophagic (Forrattini et al. 1976 loc. cit., AC Gomes et al. 1986 Rev Saúde Públ S Paulo 20: 280-287) while Lu. whitmani, although often abundant peridomes-tically, does not enter houses to such a great degree (HH Taniguchi et al. 1991 Rev Inst Adolfo Lutz 51: 23-30, ACR Azevedo et al. 1996 Mem Inst Oswaldo Cruz 91: 75-79). Endophagy is epidemiologically important as it increases both human-sand fly contact and the proportion of bites on children, and raises the possibility of vector control with insecticides via house-spraying or impregnated bednets. Results obtained during a recent study of sand fly behaviour in Paraná suggest a possible mechanism for the relatively greater endophagy of Lu. intermedia over Lu. whitmani. On 16 nights in May 1996, the attraction of sand flies to four baits (Table) was assessed in the backyard of a house in Fazenda Palmital, Terra Boa, Paraná (U Teodoro 1991 Rev Saúde Públ S Paulo 25: 129-133). Baits were placed under sand fly-proof nets 5 m apart with CDC miniature light traps suspended from a roofed frame above them, from 18:00 to 21:00 hr. On the first four and the last four nights the CDC-traps functioned with light bulbs, while on the other eight nights the light bulbs were removed. Flies caught during the trials were mounted in Berlese solution for clarification, and identified according to genital morphology. A total of 7,854 Lu. intermedia and Lu. whitmani were caught, distributed among the captures as shown in the Table. Although Lu. whitmani is the most abundant sand fly in both types of capture, adding light to the traps more than doubled the proportion of Lu. intermedia in the catches of female Lu. intermedia + Lu. whitmani (for all baits) from 1.4% to 3.8% (Yates corrected c2 = 5.47 on 1 d.f., P< 0.02); and the representation of Lu. intermedia in the male catches increased by more than four times, from 3.4% to 18.9% (Yates c2 = 179.2 on 1 d.f., P< 0.001). Light is clearly attractive over at least 2.5 m, as for both species it significantly reduces the difference in catches between the baits. Incidentally, these data also show that for both species females are more attracted to light than males, as has been shown for Lu. longipalpis (C Dye et al.1991 Animal Behaviour 42: 183-192). Adding light increased the proportion of females from 22.3% to 76.3% for Lu. whitmani (Yates c2 = 1716 on 1 d.f., P< 0.001), and from 10.5% to 35.4% for Lu. intermedia (Yates c2 = 13.2 on 1 d.f., P< 0.001). Light is thought to attract other species of sand fly into houses, such as Phlebotomus smirnovi, vector of visceral leishmaniasis in the former USSR (WHO 1990 Technical Report series 793). We suggest that the greater positive phototaxis of Lu. intermedia shown here, may suggest why Lu. intermedia are more endophagic than Lu. whitmani.