W. KUPPER
Introduction
Description of experiments
Results
Discussion
References
This paper describes how experiments with insecticide-impregnated traps and screens have, during seven years, developed into a new, efficient tool to control tsetse flies (Glossina spp.). The project in Korhogo, northern Cote d'Ivoire, began in late 1980 to investigate the potential of insecticide-impregnated biconical traps or screens to control Glossina populations.
Preliminary experiments, lasting two years, were carried out on the Marahoue Ranch in an area between the rain forest and southern Guinea zones where flies of the morsitans, fusca and palpalis groups exist. During 1981 and 1982 several experiments were performed in gallery forests placing biconical traps at intervals of 100 to 300m over distances of 13 to 85 km, simple blue screens in riverine vegetation every 100m and black screens at a density of 4 per hectare in wooded savanna. Traps and screens were impregnated with deltamethrine at a dose of 400 mg a.i. per trap and 100 mg a.i. per screen and were re-impregnated after two months. The biconical traps rapidly controlled populations of G. palpalis and G. tachinoides, with fly reductions of 98% within two months. Simple blue screens were much less efficient compared with biconical traps and the simple black screens were less efficient than biconical traps against savanna species. In the light of today's knowledge about attraction by colours and fly movements around targets, the screen's failure is more understandable.
After these encouraging results it was decided to carry out two experimental campaigns in larger areas near Korhogo, again comparing screens and biconical traps. Two sectors, Mbengue and Sirasso, were selected. Both sectors were exclusively infested by G. palpalis gambiensis and G. tachinoides, thus necessitating only a linear treatment along the rivers and tributaries with permanent water to control the whole area. In Mbengue and Sirasso, 241 and 163 km of gallery forest, respectively, were treated with traps at 300m intervals and screens at 100m, using a total of 1141 traps at Mbengue and 1630 screens at Sirasso. Traps were again impregnated with 400 mg a.i. of Deltamethrin, but re-impregnated every 4 months and screens with 100 mg a.i. Both areas were protected against re-invasion by several trap barriers of up to 10 km in length with one trap every 100m, this being sufficient, as Politzar (1983) has demonstrated in Burkina Faso, to block any re-invasion by riverine species. Three months in advance of the actual campaign the whole area was thoroughly prospected and the average density per trap per day (ADT) established to calculate fly reductions during the course of the campaign. The ADT was 3.3 for G. palpalis and 2.3 for G. tachinoides for the whole Mbengue area and 1.8 and 1.1, respectively, at Sirasso, but with maximum ADT of up to 30 flies in suitable habitats in each sector. Simultaneously several sentinel herds with a total of 200 head of cattle per sector were examined monthly for their trypanosome infection rates.
The mean age of both species and sexes before and during the campaign was established by means of the ovarian ageing technique and wing fray, as was their host preference through blood meal analysis.
The campaigns were organized by first prospecting a given area using biconical traps once the rains had stopped in October. Traps were placed at 500 to 1000m intervals along all the rivers and tributaries, each day in a different sector, until the whole area was surveyed. All results were plotted on a map of the area enabling areas with higher or lower fly densities to be identified. These surveys were completed by mid-November. The number of traps required was then calculated and they were impregnated, fifty traps at a time, in a metal tray containing the required amount of insecticide and water. The traps were afterwards dried in the sun. Impregnation could take three to four days depending on the number of traps. By the end of November the water level in the gallery forests had dropped sufficiently and the traps were positioned at approximately 300 or 600m intervals. Trees were marked to define the trap positions. During the campaign the fly density was monitored regularly to identify places where, for various reasons, the trap efficiency was insufficient. To do this, several approaches were tested: an intensive survey twice a month over 48 hours (traps emptied daily) at places with the highest fly density during the pre-campaign survey; permanent traps with insecticide impregnated cages positioned over the whole area and emptied fortnightly; or, as used currently, daily control. With the latter a field technician carries five traps during his daily control of the area allocated to him. Walking along the gallery to check campaign traps, he places them at 1000m intervals and collects them again in the afternoon on his way back, allowing a repeated control of every stretch of treated riverine vegetation several times a month. If the fly density did not drop as expected, additional traps were placed. Traps were re-impregnated in mid-March and removed in July to avoid losses through floods.
Entomological Results for Mbengue, 1983-1984
One month after the positioning of the traps the fly reduction had reached 92.1% for G. palpalis and 90.5% for G. tachinoides, followed by a slight recovery during February, but further reductions occurred until the end of the first campaign in mid-July. A disadvantage of the technique was that during the four-month rainy season traps had to be removed from riverbeds to avoid their loss due to flooding. Normally, densities recovered by up to some 3% during this period, partly because of re-emerging flies but mainly due to re-invasions that were no longer contained. This recovery was acceptable since the trap technique does not aim at total eradication, although this may be achieved after several campaigns, as is discussed later.
During the second campaign of 1984 the trap interval was increased to 600m in order to see if an already reduced population could be contained at low density levels by fewer traps. The reduction of both species was kept at between 99-100%, meaning that over long periods perhaps only one single fly was caught per month. G. tachinoides was always reduced faster and to a lower density level than G. palpalis, indicating a higher attractivity of the biconical trap for this fly.
Entomological results for Sirasso, 1983-1984
Sirasso was chosen for a further experiment with simple impregnated screens at 100m intervals. It was expected to protect 1200 km2 by treating 163 km. As in Mbengue this sector harboured only G. palpalis and G. tachinoides, but at lower densities (1.11.8 ADT). Screens were impregnated with mg a.i. of Deltamethrin. The reduction one month later was 87.2% for G. palpalis and 82.5% for G. tachinoides. While the reduction in G. tachinoides density continued to 96.7%, the G. palpalis population recovered and the subsequent maximum reduction was never above 80.8%. Over the whole campaign of 7 months the mean reductions were 70 and 90.5%, respectively, for the two species. Clearly the screens were less efficient than biconical traps and they were therefore replaced during the second campaign of 1984 by biconical traps at 300m intervals. As a result the reduction quickly passed 95% and no flies of either species were caught during three months, providing the final evidence that the impregnated biconical trap was more efficient against riverine species than the impregnated simple screen.
Haematological results
The pre-campaign parasitological survey of the selected 200 head of cattle in each sector revealed a mean herd trypanosome prevalence of 23.3% at Mbengue (T. vivax dominating) and 41.1% at Sirasso. All infected animals were from then on treated monthly with 7 mg/kg Berenil. During the first campaign at Mbengue the mean prevalence decreased to 3.8% and further during the second campaign to 2.9% and to 8.0% and 4.9%, respectively, at Sirasso, showing how through tsetse control the risk of Glossina transmitted trypanosomiasis can be reduced to an acceptable level.
Parasitological results
The higher herd trypanosome prevalence at Sirasso reflected the higher fly infection rates in both tsetse species, 27 and 21%, compared to 6.2 and 16.4%, respectively, at Mbengue. These infection rates were considered to be fairly high for riverine species. After the positioning of the traps only very few flies could be dissected, but during the two years, due to the age of the flies caught, no infected ones were found.
Physiological age
During the pre-campaign survey and after the positioning of traps or screens, all flies that were alive were age-determined, females by the ovarian dissection technique and males by the wing fray technique. The mean age during the survey was 2.2 and 2.3 months for G. palpalis and G. tachinoides, respectively, at Mbengue and dropped rapidly to 1.0 and 0.7 months, respectively, in the early part of the campaign. The corresponding figures for Sirasso were 2.7 and 2.9 months during the survey and between 1.0-1.5 months after the positioning of the screens. Also the sex ratio changed in favour of male flies, confirming earlier findings that the biconical trap attracts older flies better than young ones and females more than males, being thus ideally suited for trypanosomiasis control. At Mbengue, where only traps were used, it can be said that the mean population age had been reduced to such an extent that acquisition and development of trypanosome infections by the flies, with subsequent transmission to domestic stock, became almost impossible. Results from a private ranch situated on the left bank of the Bandama river confirmed this. During the two years of the campaign not a single animal was lost due to trypanosomiasis and the herd has not been treated since with any trypanocidal drug. Regular examinations have revealed only one T. congolense infection in two years.
Extension of the campaign, 1985-1987
After these two promising experimental campaigns, it was decided to extend the treated area to 13,400 km2 around Korhogo where only riverine species occur. The Mbengue and Sirasso sectors were enlarged in such a way that all the main streams were treated up to their sources and the watershed between the rivers Bandama, Bou and Bagoue formed the western border. Impregnated traps were used at 600m intervals previously in the treated parts and at 300m in the extension area. The treated area in 1985 measured 6309 km2 for both sectors and formed one big block. In the 1986 campaign Sinematiali I with 1445 km2 was added, thus increasing the treated area to 7754 km2. In 1987 Sinematiali II with 1400 km2 was added as well as parts of Napie, increasing the treated area to 12, 000 km2. In 1988 the remaining part of Napie will be added, thus bringing the treated area to a total 13,400 km2.
Mbengue I
After the rainy season of 1984 the reduction for G. palpalis was still 98.8% but reached a maximum level of 99.8% in the 1985 campaign. At the end of the 1985 rainy season, the reduction had once again dropped to 98.8% but reached between 99.4 and 100% in the 1986 campaign. Again the rainy season allowed a certain population increase. In 1987 the reductions varied between 98.5 and 100%. For G. tachinoides the picture was very similar but the reductions were much more pronounced and reached almost eradication levels. Only in early 1987 were very few G. tachinoides caught, showing that this species still persisted at a very low level. The subsequent survey was negative with not one fly caught in the old sector, Mbengue I, after five consecutive campaigns. Consequently it will not be treated during the forthcoming campaign (1988), but it will be closely watched with regular surveys to see if eradication has finally been achieved.
Mbengue II
In the first campaign, 1985, the trap interval was 300m and G. palpalis was reduced by 99.4%. This sector has some very dense, almost impenetrable, gallery forests with very poor visibility. This might explain why, with traps at 600m in 1986 and 1987, fly reductions were generally less marked than during the first campaign. For G. tachinoides the results were much better and this fly was not caught in 1986 and only once in 1987. The 1987 survey revealed small residual populations in certain well-defined river parts or tributaries. The trap density will be slightly increased in those areas during the 1988 campaign.
Sirasso I
By the end of the 1984 rainy season, G. palpalis had strongly recovered to only 90.4% reduction, but reductions quickly reached a maximum of 99.7% in the 1985 campaign. Again, by the end of 1985, the reduction was only 91.4% but once more quickly reached 99.4 and 100% during the last 3 months of the 1986 campaign. G. palpalis was only caught once during the 1987 campaign. The recovery during the rainy seasons was largely a result of the southern barrier not being maintained during the rains and the fact that the area borders a classified forest. The results for G. tachinoides followed closely those for G. palpalis but recovery was much less during the rainy seasons. As in Mbengue I the area was free of flies by the end of the 1987 rains and eradication seems feasible.
Sirasso II
The reduction of G. palpalis followed the pattern in Sirasso I but with a less marked recovery during the rainy seasons because this area borders a treated area with hardly any re-invasion problems. Reductions throughout the three campaigns, 1985 to 1987, were close to 100%, but in 1987, during the control catch of January, a small, untreated tributary to the Bandama River was discovered, which brought the overall reduction down to 67.7%. Once this gallery was treated, the usual high reduction levels were achieved. G. tachinoides behaved as in Sirasso I and its original population level was usually reduced by 98 to 100%.
Sinematiali I
This sector covering 1445 km2 mainly along the Bandama River and its tributaries, with some very dense gallery forest was added in 1986. During the first part of the 1986 campaign, traps were positioned at 600m intervals and baited with octenol/acetone, the hypothesis being that if the odours increased the catch by a factor of 2, traps at 600m would reduce flies in the same way as traps at 300m intervals. Unfortunately this was not the case, suggesting that neither of these attractants were effective for riverine tsetse species. Additional traps were positioned in March, bringing the reduction up to the usual 98% level. The 1987 campaign taught us that, under certain conditions, such as very dense and deep gallery forests, the trap interval cannot always be increased to 600m. Only after additional traps were placed did the reductions for G. palpalis reach the usual level of over 90%. G. tachinoides was initially present at a very low density and after the beginning of the campaign was not caught again.
Sinematiali II
In 1987 this area of 1400 km2 was added. Unsatisfactory results during the first two months were improved by positioning additional traps and through better supervision during the second half of the campaign. The importance of a permanent and thorough trap control was clearly demonstrated. G. tachinoides was not caught during the preliminary survey, nor during the control catches.
Napie
In 1987 this area was only partly treated, mainly because of demands by SODEPRA, the Animal Production Agency of Cote d'Ivoire. As there was almost no protection against invading flies the reduction of G. palpalis was not outstanding, varying between 80 and 97% during the last four months of the campaign. G. tachinoides existed at a very low density and had disappeared after the first control. During the forthcoming campaign (1988) this sector will be properly treated and barriers established, controlling tsetse within 13,400 km2 and eradication, hopefully, will be partly achieved.
That any tsetse-control technique has strong and weak points is self-evident; the trap technique is no exception. Its effectiveness depends to a large extent on the thoroughness, the exactitude and the perseverance of the staff, as traps must be controlled several times a month during the whole campaign. As shown in Sinematiali, where traps were lost without being reported, any lack of supervision greatly influences the impact of the technique. Late bushfires, theft, destruction through cattle or human beings, or traps that have dropped to the ground are hazards that must be corrected all the time. Losses and repairs representing up to 15% of the cost of the campaign material per year, are to be expected. Traps must be positioned as closely as possible to the water's edge to be efficient. Early, usually heavy, rains in March or April with subsequent sudden floods are a special danger of which the staff must be aware. At this time traps must be placed further up the river bank. Places with insufficient density reductions, for whatever reasons, must be found and additional traps positioned.
In Cote d'Ivoire the sacred forests of nearby villages, which are often very close to water courses, are a special problem because they serve for traditional rites of the local Senoufo. These forests, consisting of the original vegetation, cannot be altered and often harbour flies at very high densities. Access of non-initiated people is normally strictly forbidden. Only recently could we persuade some of the traditional chiefs to allow us into these forests and treat them permanently throughout the year. The first success has, in the meantime, convinced others to follow suit. We hope that in a few years all sacred forests will be treated and this permanent threat eliminated.
After applicability it is usually the cost-effectiveness which determines whether or not a technique is chosen for a tsetse campaign. Several methods, the sterile male technique, helicopter application and impregnated traps, have been analyzed recently by an economist. The total expense for the trap technique, using for this calculation the size of a treated area like Korhogo, is a quarter dollar per hectare per year or, with the livestock density we have in the treated area, less than the cost of a single Berenil treatment per animal per year. There is no doubt that, even if eradication is not achieved, (and where has it ever been achieved over longer periods?) this technique, perhaps with the use of attractants, is by far the least expensive approach to tsetse control and certainly the least damaging to the environment.
Politzar, H. and F. Cuisance. 1983. A trap-barrier to block reinvasion of a river system by riverine tsetse species. Rev. Elev Med. vet. Pays trop. 36 (4): 364-370.
