P. falciparumand, for over 80% of the infected population, gameto-
cytaemia was under 0.0016% (Burkotet al., 1989). Burkotet al. (1989)
suggested that the increased attraction in laboratory experiments (Day and
Edman, 1983; Dayet al., 1983) was due to the excessively high levels of
parasitaemia in these animals.
Freier and Friedman (1976) demonstrated thatA. aegypti probed
less often and attempted to feed less often on chickens infected with
Plasmodium gallinaceum than on uninfected chickens. Their experi-
ments showed that this differential response was not due to temperature
variation between infected and uninfected chickens. Mahon and Gibbs
(1982) showed that arbovirus-infected hens attracted more mosquitoes
than uninfected hens.
Only a small amount of work has been done to explore these possible
parasite influences in other vector–host associations. Coleman et al.
(1988) reported thatLeishmania-infected dogs were more attractive to
sandflies than uninfected dogs. Baylis and Mbwabi (1995) obtained field
evidence from Kenya that oxen infected withTrypanosoma congolense
were more attractive toGlossina pallidipes than uninfected oxen or
oxen infected withTrypanosoma vivax. Although the results approached
significance, it is worth remembering that oxen are not the natural host of
eitherT. congolenseorT. vivaxand thatG. pallidipesmay prefer to feed
on other bovids, such as bushbuck (Leak, 1999), and thus any observed
effect may or may not mimic the natural situation.
None of these experiments were specifically designed to measure the
effect of infection on the attractiveness of the host; specifically, they did
not measure attraction over distance nor did their experiments take into
account differences in the innate attractiveness of individual hosts prior
to infection or the stage of parasite development.
Three recent sets of experiments have attempted to address these
issues using a dual-port wind-tunnel olfactometer to demonstrate
attraction over 1 m. In the first series of experiments the odour of Syrian
hamsters infected withLeishmania infantum(chagasi) was shown to be
more attractive than non-infected hamster odour to femaleLutzomyia
longipalpis (the natural vector of L. infantum in South America)
(Rebollar-Téllez, 1999). Female L. longipalpis were given a choice
between infected hamster odour and uninfected hamster odour; a
significantly greater number of femaleL. longipalpiswere attracted to
the odour ofL. infantum-infected hamsters. Subsequently, O’Sheaet al.
(2001) repeated and extended these observations and confirmed the
results obtained by Rebollar-Téllez (1999). O’Sheaet al. (2001) also
collected the odours from infected and non-infected hamsters by
air entrainment; when these odours were tested in the olfactometer the
same results were obtained. Taylor (2001) showed that, given a choice of
infected mouse odour and uninfected mouse odour, a significantly greater
number of females were attracted to the odour of mice infected with
P. yoelii nigeriensis. These preliminary experiments also showed that
attractiveness varied throughout the cycle of malaria infection. Mice that268 J.G.C. Hamilton and H. Hurd