Control of human African trypanosomiasis: trap and odour preference of tsetse flies (Glossina morsitans submorsitans) in the upper Didessa river valley of Ethiopia


H. Belete andW. A. Oyibo (corresponding author), Department of Microbiology, Immunology and Parasitology, Faculty of Medicine, Addis Ababa University, Addis Ababa, Ethiopia. Tel.: 251 1 533197; Fax: 251 1 513099; E-mail:
G. Tikubet, ICIPE-Ethiopia Collaborative Project, PO Box 3853, Addis Ababa, Ethiopia.
B. Petros, Department of Biology, Addis Ababa University, PO Box 1176, Addis Ababa, Ethiopia.
I. N. Otigbuo, Department of Biology, Montgomery College, 7600 Takoma Avenue, Takoma Park, MD 20912, USA. E-mail:


Ethiopia is one of the endemic countries for Human African Trypanosomiasis (HAT) as over 100 000 people are at risk of having the disease. The control of HAT using odour preference of tsetse flies (Glossina morsitans submorsitans) was studied in upper Didessa river valley of Ethiopia. No information exists on the effectiveness of attractants for these species of tsetse flies in Ethiopia. Three attractants and their combinations namely: acetone, octenol, cow urine, acetone + octenol, acetone + octenol + cow urine, were evaluated using biconical and NGU traps for their efficacy as a first step in developing a sustainable community-based HAT control initiative. The biconical traps baited with acetone, octenol or cow urine, or when combined, were more effective in catching G. m. submorsitans than the NGU traps (P < 0.05). However, the NGU traps caught more female tsetse flies than the biconical traps (P < 0.05). The acetone, octenol and cow urine combination was the most effective in attracting tsetse flies in both the biconical and NGU traps. Acetone was the best attractant while octenol was the least effective. Cow urine showed great promise for possible use in community-based HAT control activities, especially urine that has been kept for several days. The use of cow urine in HAT control in Ethiopia is likely to succeed in the future because of its sustainability.


Human African Trypanosomiasis (HAT) is a serious public health problem causing sleeping sickness in tropical Africa in addition to fatal consequences in humans and animals especially cattle. The wide occurrence of this disease in people and their livestock poses a great constraint to agricultural and economic development on the continent. The major economic impact of African trypanosomiasis is in losses of cattle and the countries mostly affected are those with the least number of veterinarians (Braend 1979). However, the full economic consequences of trypanosomiasis go far beyond figures on disease occurrence or livestock mortality. In areas affected by the disease, such as Ethiopia, it is difficult for the local farmers to keep livestock at all.

Scientists have been working for decades in order to develop effective, economic and safe ways to protect livestock and people against trypanosomiasis. The task proved extremely difficult. Yet, given the wide prevalence of the disease and its major economic importance, the potential reward for success is enormous. The control programmes so far employed have not been very satisfactory (Jordan 1995). The reasons are based on the fact that there are more than 20 species of Glossina responsible for the transmission of the disease, and they are adapted to a wide range of habitats, and this accounts for the endemicity of the disease.

The disease is found only in Sub-Saharan Africa between latitudes, 14°N and 29°S, which are the limits of the geographical distribution of the tsetse fly vector. WHO (1998) reported that there are 200 disease-endemic foci in 36 countries; 60 million people at risk of infection and an estimated 300 000 new cases each year.

Ethiopia is one of the endemic countries with over 100,000 people at risk of having Trypanosoma brucei rhodesiense infection. One of the commonest species of tsetse flies that is incriminated as an important vector of animal trypanosomiasis in Ethiopia is Glossina morsitans submorsitans Newstead 1924. G. m. submorsitans have been reported in the river valleys of western Ethiopia, Mui River, around Lake Abaya in the Rift Valley, Didessa river valley, Metekel, Angar, Baro and Gilo rivers (Balis & Bergeon 1970; Ford 1971). Bovine trypanosomiasis has also been reported in Kindo Koisha district, Wollaita zone, in south Ethiopia (Kidaemariam et al. 2002).

The resurgence of sleeping sickness in historical foci in East, West and Central Africa (WHO 1998) has made it imperative for cost-effective and sustainable vector control measures to be evaluated. One of the most common vector control techniques is the trapping of tsetse flies and these traps decrease the population of flies and hence decrease man–fly contact (Laveissiere et al. 1994). Vector control using traps has the advantage of being managed by the communities for sustainability especially when materials for the traps could be sourced locally with minimum cost.

Monoconical and biconical traps create a visual stimulus to which tsetse flies respond by flying into them. In the early 1970s, the biconical traps were used successfully against riverine species such as G. palpalis (Allsopp 1984). Attractants have greatly increased the efficiency of traps over the years (Vale & Hargroove 1978; Vale 1980). Carbon dioxide and acetone were found to be powerful attractants of tsetse; butanol, 1-acetone-3-α, p-cresol, 4-methylphenol and 3-n-propylphenol are other known attractants. Attractive substances in the host breath are known to include acetone and 1-octen-3-ol; and in urine, 4-methyl phenol and 3-n-prophylphenol (Jordan 1995; Spath 1995). However, the most effective cocktails of attractants are still <50% as attractive to tsetse as natural cattle odours. Thus, in order to increase the efficiency and reduce dependency on imported chemicals, cow urine has been used with some success (Holmes 1997).

The efficiency of traps and attractants in the control of tsetse fly have been reported in some countries (Jordan 1995; Spath 1995; Dransfield & Brightwell 2001; Joja & Okoli 2001) but information on the efficiency of attractants and traps in Ethiopia is not available. The efficiency of attractants is known to vary between species and within species at various locations (Jordan 1995).

In this study, we evaluated the efficiency of biconical and NGU traps and attractants, especially the use of cow urine as an effective attractant for G. m. submorsitans in the field as a first step towards the development of a sustainable community-based tsetse fly control strategy in rural Ethiopia.

Materials and methods

Study area

The study was conducted at the upper Didessa river valley, in south-western Ethiopia, where the incidence of animal trypanosomiasis is very high. The area is infested with both G. m. submorsitans and G. tachinoides. G. tachinoides, a riverine species, is restricted to the lower course of Didessa River. Belete et al. (2001) reported that warthog is the major host of G. m. submorsitans in this area.

The climate is marked by wet and dry seasons with peak rainfall between June and September. The studied area has an average rainfall of 1946 mm monthly and mean minimum and maximum temperature of 12.5 and 25 °C respectively. The altitude of the valley is in the range of 1250–1900 m above sea level.

The vegetation of the area is characterized by wooded savanna land with grass species of Hyperrhenia, which can grow to a height of 2 m. Tree species such as Terminalia spp. and Combretem spp. are also dominant in this area. During the dry season, most of the trees and grass except those found along the watercourse of the river become leafless as a result of annual bush burning, which compels G. m. submorsitans to seek refuge in areas not affected by burning, especially strips of vegetation along the watercourse or in isolated evergreen vegetation.

Wild animals reported in the area include warthogs (Phacochoerus aethiopicus), roan antelopes (Hippotragus equinus), bush pigs (Patamochoerus porcus), monkeys (Colobus polykomos), baboons (Paio anubis), hyaenas (Crocuta crocuta), rabbits (Dipodina spp.), hippopotamuses (Hippopotamus amphibious), buffaloes (Syncerus caffer) and bush bucks (Tragelaptamus amphibius).

Odour baits/attractants and traps

The following odour baits and their combinations were evaluated for their efficiencies as tsetse attractant: (i) acetone (A), (ii) octenol (O), (iii) cow urine (CU), (iv) acetone + octenol (AO) and (v) acetone + octenol +cow urine (AOCU). The attractants were mixed in equal proportions.

We evaluated 10 6 × 6 NGU (NG2G) and 10 biconical traps. The attractant was introduced into the traps with the aid of standard dispensers. Traps without attractant served as controls (C). Both the tested and control traps were randomly positioned throughout the valley to avoid bias. The traps were set from 8.00 am in the morning to 6.00 pm in the evening, when attracted flies were collected, counted and gendered. This was carried out daily for 2 weeks.

The index factor computed for the tested attractants/baits and traps was defined as the factor by which the mean number of catches using the various attractants was compared relative to the control (unbaited traps). It was computed by dividing the mean/mean total of male and/or female flies caught with the control trap by the mean/mean total of the flies caught using different attractants. The data obtained were analysed using the F-test (two-factor variance analysis) and Student's t-test.


The biconical trap was superior to the NGU trap with all attractants/baits tested (Tables 1–3). The acetone, octenol and cow urine combination (AOCU) was the most efficient in attracting tsetse flies in both the biconical and NGU traps. Octenol was the least effective among the attractants/baits in both traps. The efficiency of these attractants was significant compared with controls (P < 0.05). Among the single attractants, acetone was the best.

Table 1.  Mean number of tsetse flies (Glossina morsitans submorsitans)/trap/day with several baits/attractants and their combinations using biconical trap
Attractants/baitMaleFemaleTotalF:M% FIndex
  1. AOCU = acetone + octenol + cow urine; F = female tsetse; M = male tsetse.

Control (unbaited)6.16.913.01.1353
Acetone + octenol19.234.553.71.80644.14×
Cow urine16.119.435.51.21552.73×
Table 2.  Mean number of tsetse flies (Glossina morsitans submorsitans)/trap/day with several baits/attractants and their combinations using NGU trap
Attractants/baitMaleFemaleTotalF:M% FIndex
  1. AOCU = acetone + octenol + cow urine; F = female tsetse; M = male tsetse.

Control (unbaited)
Acetone + octenol16.528.845.21.75644.56×
Cow urine8.915.023.91.69632.48×
Table 3.  Mean total catches of flies (Glossina morsitans submorsitans) with biconical and NGU using various odour attractants
Odour baitsFlies per trap/day (mean ± SD)Index of increase
BiconicalNGU(Biconical over NGU)
Unbaited (control)13.0 ± 8.09.9 ± 3.531.31×
Acetone45.6 ± 23.327.6 ± 5.651.65×
Octenol31.2 ± 17.9611.0 ± 1.412.48×
Acetone + octenol53.7 ± 2.8245.2 ± 6.361.19×
AOCU81.9 ± 10.667.2 ± 2.121.22×
Cow urine35.5 ± 10.523.9 ± 2.111.49×

Comparison between the traps showed that the mean number of flies caught using biconical traps was significantly higher than the NGU traps (P < 0.05) for all attractants. Biconical traps with cow urine as the attractant increased the tsetse fly catch size by 2.73 times (2.73×) (Table 1) against unbaited controls. Similarly, the same odour attractant with NGU trap was 2.48 times (2.48×) more efficient than the control (Table 2). It is important to note that a sample of cow urine kept in a container for up to 6 days trapped more tsetse flies than the fresh sample at day 1 (Table 4).

Table 4.  Mean number of tsetse flies (Glossina morsitans submorsitans)/trap/day with NGU and biconical traps based on the age of cow urine (1–6 days) used as bait
DaysMean no. of flies/trap/day
117.7 ± 5.78.3 ± 3.4
230.8 ± 9.8717.7 ± 8.14
340.0 ± 10.2828.5 ± 3.4
440.5 ± 11.4226.8 ± 6.8
554.8 ± 5.1457.7 ± 4.57
677.5 ± 12.5764.57 ± 10.28

Although the biconical trap caught more flies overall (male and females), the NGU trap caught more female G. m. submorsitans than males with all the attractants used except octenol in the biconical trap (Tables 1 and 2). The difference in female catch between NGU and the biconical trap was significant (F = 11; P < 0.05). NGU traps were therefore more effective in trapping female flies than males by an index factor of 1.18 (1.18×).

There were also variations in the number of fed flies caught with biconical and NGU traps: NGU traps caught a higher proportion of fed flies (data not shown).


The control of HAT using vector control still remains practicable especially when affected communities are involved. The use of flytraps, chemotherapy and other vector control activities would add up to the use of an integrated method in controlling HAT. Treatment failures, availability and affordability of drugs for HAT (Brun et al. 2001; Etchegorry et al. 2001) make evaluations of cheaper and sustainable control methods imperative.

Our study in the evaluation of traps and odour attractants for the control of G. m. submorsitans in Ethiopia has shown the strength and weaknesses of both the biconical and the NGU traps. Based on the overall number of flies (male and females) caught, the biconical trap was more efficient. However, the NGU traps caught more female flies. Preference for trap design and colours among Glossina species is common (Flient 1985). Brightwell et al. (1987) found that NGU traps are three times as effective as biconical traps for catching G. pallidipes, which is one of the Morsitans groups of tsetse. For the Palpalis group of species, the vertically oriented biconical trap and its derivative are highly effective, whereas for the Morsitans group, compact or horizontally oriented shapes are more attractive (Jordan 1995).

The large number of gravid and engorged females trapped in the NGU trap suggested that some females may have been seeking larviposition sites. The trap's shape and its closeness to the ground might have been responsible for attracting flies, which had last instar larva to be deposited. This selective advantage of NGU in trapping more female than male flies has a strategic implication for tsetse control in Ethiopia. Calculations based on basic life table data have shown that it is only necessary to catch some 1–4% of the female population per day in order to achieve effective control (Jordan 1995).

Our evaluation of odour attractants showed that the catch size increases as the number of attractants/odour baits in the cocktail increase. A cocktail of all known attractants, except carbon dioxide, can increase trap captures of G. pallidipes by 15–20 times (Jordan 1995). The mixture of acetone, octenol and cow urine attracted more flies than the combination of acetone and octenol or with the use of individual attractant. Acetone was the best attractant among the attractants evaluated. Electroantennographic responses, which are used to detect the potency of olfactory stimulants, had shown that the dose–response curve to octenol is 106 times higher than acetone for G. pallidipes and G. m. morsitans (Hall et al. 1984). Although the potency of octenol to olfactory sensation towards G. pallidipes and G. m. morsitans in the laboratory was reported to be higher, our study showed that its efficiency as odour bait was lower than that of acetone and cow urine.

Vale (1980) reported that acetone alone increased up to six times the catches of G. pallidipes and G. m. morsitans with the biconical trap. This is more than recorded for G. m. submorsitans in this study. The proportion of female and male catches of G. m. submorsitans was also affected by odour type. The use of acetone as attractant either alone or in a cocktail with other odour attractants increased the catch of female flies in both traps.

Spath (1995) reported that ox urine increased the catch of G. tachinoides significantly by 1.2 times in odour-baited biconical traps. The phenolic fraction of cow urine gave increases of up to 1.6 for G. longipalpis and 1.4 for G. tachinoides while the addition of acetone and/or 1-octen-3-ol to the phenolic fraction increased attraction of G. longipalpis and G. tachinoides significantly up to 1.8 and 1.3 times respectively (Spath 1995). However, acetone alone, in combination with octenol or with the phenolic fraction, reduced the catch of G. medicorum significantly.

The efficiency of cow urine, especially when left to stand for up to 6 days, implies that cow urine could be employed in community-based HAT control using any of the traps due to its ready availability especially in poor countries. The evaluation of cow urine for a sustainable HAT control in at-risk communities of Ethiopia is suggested. The improved efficiency of aged cow urine could be due to the bacterial activities in urine samples, the action of which may have been responsible for releasing compounds such as ammonia, phenols, aldehydes, ketones, etc. that serve as attractants. Our finding infers that large fly catches are possible using cow urine baited-traps for the control of G. m. submorsitans.


We thank the Department of Biology, Addis Ababa University, Addis Ababa, for the provision of materials used in the field, and SIDA/SAREC, Sweden, for financial support.