FAO Home Page French version Spanish version Waicent Free Text Search

Use of attractive devices for tsetse survey and control

Chapter 1: Introduction

Early Work

Attempts to control tsetse flies in Africa have been continuing since the early 1900's. Some control methods, such as bush clearing and game elimination, were widely practised in the past, but have now been discarded mainly for environmental reasons. The use of traps to control tsetse was experimented with in the 1920's, but interest in this method declined following the introduction of insecticide-based techniques in the 1940's. Since then, ground spraying and aerial spraying of insecticides have been the main methods of tsetse control.

However, during the 1970's interest in attractant devices was renewed, in order to reduce dependency on insecticides, and in recognition of the potential of such devices. The basic principle of these methods is to attract the tsetse in large numbers using either an artefact (a trap or a target) or a natural host, and then to kill them using either insecticide or a cage from which they cannot escape.

We will first look briefly at some of the earliest applications, as developed before the introduction of insecticides.

The early devices took many forms, including the traps designed by Harris, Swynnerton, Jack, Morris and Langridge (see Vol. I, 7.2 for details). The Harris trap was used from the 1920's to the 1940's to control G. pallidipes in South Africa (Zululand). By 1937 nearly 9000 traps were in use, and considerable reductions of tsetse numbers had been achieved; however, after 1940, Trypanosomosis caused heavy mortality in cattle, despite over 26,000 traps being in use, and trapping was eventually abandoned as a means of control.

It was realised at this time that the addition of attractive odours would make trapping more effective both for sampling and control. In the 1940's, it was observed that G. morsitans aggregates at elephant resting places, whilst G. pallidipes preferred those of buffaloes. Experiments were then carried out by Chorley in 1948 to show that more flies could be caught at shelters baited with cattle dung and urine than at unbaited shelters. This pioneering work was, however, not followed up for over thirty years.

In the 1950's carbon dioxide was shown to increase catches of G. pallidipes in Morris traps. A few years later, benzene and petroleum-ether pig skin extracts were shown to increase catches of G. pallidipes and G. fuscipes when smeared on Langridge traps. Unfortunately, most of this work ceased in the 1960's, when it became generally felt that the tsetse problem could be solved more effectively by selective insecticide application.

The first use of insecticide-impregnated targets for control was by Rupp in the 1950's in Rwanda and Burundi. The targets were rectangles of black cloth (0.5 x 1.4 m), treated with DDT and suspended with rope over a river. They were apparently successful in reducing the numbers of G. fuscipes martinii.

The idea of using man as a bait to attract and then kill tsetse was first exploited on the island of Principe. Workers on the estates were given dark sticky targets (40 x 50 cm) to wear on their backs and so trap any G. palpalis that landed on them to feed. This method, combined with bush clearing and destruction of hosts, eradicated the tsetse by 1914.

The use of cattle as a bait also has a long history. At the end of the 1940's, short acting arsenical insecticides used to control ticks and other ectoparasites on cattle were gradually replaced by more residual acaricides, and it was noticed that sometimes tsetse numbers were also reduced. This was investigated in more detail and the potential of the method clearly demonstrated. However, due to the non-availability, at that time, of suitable insecticides and formulations to allow practical implementation, the work was discontinued.

Despite these promising early results, the further development of these methods was abandoned in favour of the application of insecticides because:

- the early traps were large and cumbersome, so that making and setting them was very time consuming;

- the traps and targets were used without any odour baits and were not very efficient, high trap densities were, therefore, required;

- suitable formulations of insecticides were not available to ensure that the insecticide used on animals and targets remained effective for a sufficient time;

- it had been demonstrated that direct and selective application of insecticides tsetse resting, refuge and breeding sites could also solve the problem.

Interest in trapping techniques was renewed in the late 1960's, and research in many countries has resulted in the development of bait methods which are now available for practical field application. The aim of this manual is to report on the latest developments and to provide guidelines in their adaptation and use in various local situations.


Recent developments

Odour attractants
One of the major advances has been the discovery of effective odour attractants, especially for morsitans species, and more recently also for some species of the palpalis and fusca groups. These can be used to increase the number of flies either entering traps or alighting on targets.

Investigations carried out in the 1970's, demonstrated that cattle odours greatly increased the numbers of tsetse caught. Carbon dioxide had been known for some time to be one of these attractants, but is not practical or economical for use in control because quite large amounts are naturally present in the atmosphere. Hence much more carbon dioxide is required for the fly to sense it above the background level than for other odour baits which are present in only minute quantities.

Other attractants identified showed great potential for control, such as various ketones, especially acetone and 1-octen-3-ol (known as octenol). These greatly increased catches of G. morsitans and G. pallidipes, even when used without carbon dioxide.

In the 1980's, it was reaffirmed that bovid urine was an effective attractant for G. pallidipes, and that cattle urine used in conjunction with acetone gave big increases in catches. The main attractive chemicals in the urine were later identified as two phenols, namely 4-methylphenol and 3-n-propylphenol.

All these advances were directed at morsitans group species, and for some time no attractants could be found for the palpalis and fusca groups. Recently, however, octenol and the urine of various of bovids, especially bushbuck, have been found to increase catches of G. tachinoides; the main active phenols are 3-methylphenol and 4-methylphenol. Effective attractants have also now been identified for several of the fusca species.


Although alternative trap designs were developed for tsetse in the 1950's and 1960's, they were all based on the earlier models and retained many of the unfavourable features i.e. they were large, cumbersome, and difficult to transport.

It was only with the development of the biconical trap by Challier and Laveissi╦re in the early 1970's (see Vol 1. Section that a trap became available that was relatively cheap, collapsible, so that many can be carried in a vehicle, and quickly and easily assembled.

The biconical trap is very effective for species of the palpalis group including G. palpalis and G. tachinoides . It is less effective for the morsitans and fusca groups although good numbers can still be caught of some species, e.g. G. pallidipes.

The biconical trap is widely used throughout Africa for sampling many species of tsetse flies, although it has only been used for control of riverine species when it is usually treated with insecticide to enhance efficiency.

Efforts to reduce cost and increase the efficiency of the biconical trap led to the development of a series of monoconical traps by Lancien in 1981, Mŕrot in 1987, and Laveissi╦re in 1988. These consist of an upper cone of clear plastic or mosquito netting with black and blue screens hanging below the cone. The pyramidal traps of Lancien and Gouteux are similar but consist of an upper triangular pyramid of mosquito netting, again with hanging blue and black screens.

Although these traps are effective for palpalis species, they are less so for morsitans species. Research on improving traps for the latter really started in the 1980's when Vale produced a series of experimental traps, the alpha trap and the beta trap, which greatly improved our understanding of the response behaviour of tsetse to traps. This led to the cubic F2 and F3 traps of Flint which are not only very effective for G. pallidipes but also catch G. morsitans in reasonable numbers.

Although the F3 is suitable for sampling, the function for which it was developed, it is too expensive for control. This led Brightwell to develop a new series of triangular traps in the late 1980's that were effective for both G. pallidipes and G. longipennis (a fusca species), yet were cheap and could be easily constructed by local people. The F3 trap was also subsequently modified to a triangular version, the epsilon trap.


In the late 1970's and 1980's a series of insecticide-treated screens were developed for riverine species, in an attempt to find a cheaper and more practical device than the biconical trap for control. The first of these, produced by Challier and Gouteux in 1978, consisted of a blue screen measuring 150 x 90 cm; later screens, as used by Laveissi╦re in the 1980's, were rather smaller. More recent developments of targets for riverine species have concentrated on combinations of blue and black, often with emphasis on using materials that retain insecticide better. An example of this is the blue and black target of Laveissi╦re.

In the 1980's work started on identifying suitable targets for control of the savanna species, primarily G. pallidipes and G. morsitans. Monitoring with electric nets indicated that many flies flew around a insecticide-treated black target rather than landing on it. Hence insecticide-treated netting invisible to tsetse was positioned close to the black cloth to intercept flies that did not alight on the target. The first version of this was the R-type target. This was made of black cloth and netting and pivoted like a wind vane, so that the netting was always in the most effective position just downwind of the cloth. The targets were baited with 1-octen-3-ol and acetone, and the insecticide-treated netting was protected from rain and sun by a roof of white plastic. They were subsequently used successfully to greatly reduce numbers of tsetse.

The R-type target has since been replaced by the simpler S-type target which consists of a central panel of black cloth flanked by netting side panels. Although the netting was formerly thought to be essential, it now appears that a larger, all black cloth target may be more economical, practical and similarly effective, and this is now being increasingly used for savanna species.


Insecticide-treated livestock
This is the most recent technique which has been revived for tsetse/Trypanosomosis control, and one which is now generating considerable interest. The introduction of synthetic pyrethroids in suitable residual formulations to apply to cattle as a dip or pour-on has made it feasible, in some situations, to combine the control of ticks and tsetse. Successful trials have been carried out in which many thousands of cattle were dipped with deltamethrin, and the fly eradicated over much of the area. The technique is now being evaluated in several countries and promising results recorded.


Biological considerations

In order to control any pest species it is necessary to have a good understanding of its biology, behaviour and population dynamics, that is the natural factors which affect distribution and abundance.

Using simple population models, it is possible to predict by how much the natural death rate has to be increased to artificially induce a population decline. Because tsetse have a very low birth rate, about one larva every ten days, the natural adult mortality only has to be raised by between 2 and 4 per cent a day to reduce numbers. For several species of tsetse, such death rates can be readily imposed with traps, targets or bait animals.

If such levels of mortality can be applied evenly to a population, and are sustained, the tsetse will inevitably be eradicated, simply because the death rate will be higher than the maximum possible birth rate. Unfortunately these conditions can seldom be met, and consequently the objective is usually at least a 90 per cent reduction in numbers rather than eradication.

The ease with which the total adult death rate can be increased will vary depending on the causes of natural mortality. These may be classified into two categories:

Density independent factors
These are factors which may kill a proportion of the population regardless of the density, such as extreme temperatures, low humidity, rainfall etc. These determine changes in tsetse numbers, especially seasonal changes.

Density dependent factors
These are factors which act more severely when tsetse numbers are high than when they are low. One characteristic of tsetse populations is that they tend to be very stable compared to other insects; this means that certain factors are 'regulating' the tsetse populations at a particular level. Such factors are not well known but could include predation, pupal parasitism, dispersal, and competition for hosts. They may work indirectly. Thus if there are large numbers of tsetse in an area, wild animals may move away, making it more difficult for tsetse to find a host to feed on. As a result flies may die of starvation, or, more likely, the pregnant females will be unable to adequately nourish their larvae, thus affecting the next generation.

If a tsetse population is near the limits of its natural distribution, it is likely that most of the natural causes of death will be density independent, and control measures will simply add to this, making control easier. However in areas where conditions for tsetse are optimal, as the death rate is increased by control methods, the natural death rate resulting from density-dependent factors will decline because the density of flies is dropping. Hence the death rate caused by the control measures has to be higher to counteract this.

Tsetse numbers are affected not only by the birth rate and the death rate, but also by the rate of immigration and emigration from an area. How much the flies move, their mobility will also affect how easy it is to control them. Odour attractants are only effective over a relatively short distance (<100 m) and the efficacy of odour baited devices depends largely on the mobility of the flies, rather than on the devices attracting flies from afar. If the flies move a lot, even with a few traps in an area, sooner or later they will encounter a trap and be killed. This is generally true for savanna flies, especially G. pallidipes. If they do not move much but remain localised, a higher density of traps will be needed to ensure that pockets of flies do not remain.


The mobility of the flies also affects the level of reinvasion. With savanna flies only a few traps are needed per square kilometre,to also reduce the numbers outside the trapping zone. However, large areas have to be treated (at least 500 km2) or excessive reinvasion will occur. With less mobile tsetse species, control can be effective over a few square kilometres, but tsetse numbers outside the trapping zone will be little affected.

Seasonal dispersal of flies is another important factor to consider. Whilst some forest and riverine species show little seasonal change in distribution, others become more widely dispersed during the rainy season. Whereas in the dry season the flies are concentrated in dense vegetation, during and after the rains they may spread many kilometres into open woodland and bush (dispersal zones).

Monitoring such changes in tsetse distribution during control is essential to decide upon the most effective distribution of traps or targets. If control is the objective, it will be best to site the majority of traps in or near the dry season refuges; to avoid "hot spots" there should be some traps in all such areas, as there may be little movement between them in hot dry weather. This strategy will greatly reduce the number of flies available to move into the dispersal zones during and after rains.

For eradication, however, a more even distribution of traps is essential because not all flies may leave the deciduous woodland during the dry season, this is especially true for G. morsitans and G. longipennis. Similarly to maintain areas of local eradication, it may be necessary to extend barriers of targets or traps several kilometres out into normally unfavourable open areas to prevent reinvasion during the rains.

Such considerations of the biology, ecology and behaviour of tsetse flies, enable the most appropriate strategy for tsetse control to be chosen.


Control Strategies

There are several options:

Complete eradication
Flies are eliminated and no measures need be taken to maintain this situation. This can really only be done by removing the tsetse over its entire distribution. This is possible in a few situations such as islands, e.g. Zanzibar, or isolated belts. It is much easier to eradicate near the limits of natural distribution, e.g. northern Nigeria or southern Africa, because seasonally adverse climatic conditions increase the natural mortality rate in populations, so less effort is needed to raise the death rate above the birth rate, if operations are conducted at these times.

Local eradication
Flies are eliminated but a deliberate man-made barrier of traps, targets, insecticide or cleared vegetation has to be maintained to prevent reinvasion. Unfortunately, such barriers are not perfect, and in practice surveys and a control capacity have to be retained to detect and deal with any reinvasion that occurs.

Flies are reduced to very low level and maintained at that level so that the disease challenge is reduced. In practice this is the only realistic strategy over much of Africa. The level of reduction that can be maintained will normally be set by the level of reinvasion, which may vary seasonally. This approach is only viable economically if costs of control are kept to a minimum and are below the level of the benefits gained.

Several factors determine the area of tsetse infestation that must be covered in a control operation. If eradication is to be attempted, the entire tsetse belt must eventually be covered. If local eradication is decided upon, then natural barriers should be used as much as possible, with the length of artificial barrier kept to a minimum to minimise costs and reduce the chances of reinvasion.

With control, other factors are important. For relatively immobile fly populations, control may be achieved over an area of a few square kilometres, however, as the area is increased, so the control becomes more effective and more economical, because barriers, if they are necessary, make up a smaller proportion of the total area. The logistical demands of organisation, however, become more complex. A balance has, therefore, to be reached between these conflicting factors.


Seasonal or year-round control
When control rather than eradication is undertaken, it may not be necessary to exert this all year round. In some situations e.g. riverine species in the wetter zones of West Africa, it may be very difficult to maintain traps or targets during the rainy season. If the rate of reinvasion is low, it may be feasible to remove the targets at the start of the rains, to prevent flood loss or damage, and replace them when the rains are over.

If the rate of reinvasion is high during the rains, as it is for savanna species, traps or targets must be retained in the area all year round. However, if the main tsetse areas are only used for dry season grazing, it may not matter if control is less effective during the rains. A low incidence of Trypanosomosis, resulting from small numbers of tsetse spreading into the rainy season grazing areas, can be dealt with using chemotherapy and may not even be a problem if the cattle are well nourished and show some trypanotolerance.


Rate of reduction of tsetse numbers
The rate of reduction of the tsetse population is determined by the effective trap/target density, i.e. those deployed in optimal sites, and by the natural rate of tsetse increase, the net effect of birth/death rates and immigration/ emigration rates. The higher the effective trap/target density, the faster the tsetse population will decline.

If there is seasonal reinvasion, clearly there must be a sufficient density of traps to reduce numbers fast enough before cattle again move into the area. In general, economic considerations will mainly determine what rate of reduction it is feasible to impose on the tsetse population.


Table of Contents

Other Chapters


Comments: AGA-Webmaster