go to previous section of this publication

S. gregaria hopper bands showed a habitat-dependent difference in their migratory behaviour, which was important for the selection of the control measure to be applied. In the southern part of the country (large sandy areas with small patches of vegetation) the insects covered up to 3 km per day. In contrast, in the north (large, dense areas of vegetation) the daily migration distance was only a few hundred meters. The consequence for barrier treatments was that in the south wide barriers with up to 1000 m spacing had to be made, while in the north narrow barriers with maximally 100 - 150 m spacing could be used.

Both possibilities were tested in field studies (cf. 2.1.2.1). In large-scale experiments with barrier widths of 100 m and over and distances >1000 m, in 1993 in Boumdeid mortality rates >70 % were obtained, and in 1995 in a large-scale test in the northern part of the country rates of 60 to 90 %.

The determination of the sizes of the hopper bands gave numbers of individuals from ca. 50,000 up to several millions. Initial observations indicated a substantial reduction in the number of individuals in the course of larval development. Whether this equalled the value of 92 % given in the literature, could not be established. This must be quantified in future observations, since this has important consequences for the treatments. With such a high mortality, treatment of smaller hopper bands would be superfluous.

Another aspect of the type of treatment is brought out by the observations on the areas occupied by the larvae during resting periods and during migration, since this is also a means to save insecticide (cf. Table 7).

Similarly, it is an absolute prerequisite for the quantification of the success of treatments to determine the number of individuals in a hopper band. For such estimates, which must always be made both before and after any treatment, three methods have been developed. Which one we use in a specific instance is determined by the behaviour of the larvae and the habitat conditions:

a) Density determination for migrating bands
Small rectangles or squares (generally 10 to 20 m on a side) were traced in the sand in the path of the moving bands. The number of larvae crossing the marked areas was counted. The plots had to be small enough that all entering and leaving larvae could be observed at the same time. The length of the band was measured by pacing it off or driving along the entire length. Then the width was estimated at regular intervals and any obvious changes in the density were noted at each point. From these values the number of individuals can be calculated with sufficient accuracy.

b) Determination of density for larvae resting in bushes
During the night and at mid-day the larvae interrupt their migration and congregate in bushes and trees. In such cases, an exact determination of the number of larvae is impossible, and only a rough estimate can be made. The precision of these estimates can be checked by treating the insects with an insecticide with a high knock-down effect and then counting them.

c) Density determination for larvae resting on sandy areas
In the morning hours the larvae warm themselves on sandy patches, often in a wind-shielded place behind trees and bushes. The number of animals can be calculated if one knows the size of the total area covered, the distance between individuals and the area covered by a single larva. The latter values were quantified for each larval stage, and the number of insects per m² was calculated taking into account the different inter-insect distances (cf. Tab. 6).

Tab. 6 Surface covered by larvae of various stages

With these methods and values, the areas occupied by the hopper bands during the resting and migratory phases were calculated. For bands with larvae in stages L4/L5 this gave a ratio of 1:138 (cf. Tab. 7).

Tab. 7 Surface covered during migration and resting

This ratio of the surfaces is not a constant parameter, instead being dependent on the stages of the insects, abiotic and biotic factors (wind, temperature, habitat conditions, larval age, etc.). For hopper bands of stages L1/L2. a mean ratio of 1:33 was determined and for L3/L4, 1:93.

Feeding preference tests demonstrated a significant preference by S. gregaria for cultivated crops over wild plants. When the animals are offered S. thebaica, T. terrestris, H. bacciferum, V. unguiculata and maize, the last two species are clearly preferred (Fig. 10).

Fig. 10 Food preference of S. gregaria larvae

The percentages are relative to the total amount of food offered and partially consumed each day. The data are the means ±SD for 3 experiments with 20 larvae each in stages L4 to L5. The experiments started at the beginning of stage 4 and ended with the adult moult.

The standard value repeatedly cited in the literature is that S. gregaria larvae can consume their own weight in food daily. Appropriate experiments in Mauritania. showed, however, that this value cannot be substantiated, rather the daily consumption of food is only 30 ±8 % of the body weight. This is an average value over the whole of larval development. The daily food uptake varied from 0 % (just before and after a moult) to 35 - 45 % (around the middle of each stage). The repetition of these experiments with larvae in the 1994/95 season gave comparable, although slightly higher values of 35 ±5 %.

Fig. 11 Gain in body weight by S. gregaria larvae
The larval weights are the mean of 5 weightings each.
The food uptake was determined for L3, L4 and L5-larvae. The measurements took place with the following food being offered:
a) Maize
b) S. thebaica
c) S. thebaica, V. unguiculata, H. bacciferum (mixture)
As a mean, we determined a value for the daily consumption of food equal to 30 ±8 % of the body weight.

Analogous experiments were performed with adult animals up to 10 days after the moult and with animals 3 weeks older. For the former the uptake values were 20 ±5 %, and for the older animals 25±3 % of the respective body weight. However, both groups were sexually immature. These new findings result in a completely new standard value for determining potential crop harvest losses. Indeed, many calculated crop losses existing in the literature are quite likely overestimates-

The high values for food consumption cited in the literature have in many cases been seen as justifying blanket treatments of pasture lands, since S. gregaria is regarded as a food competitor for grazing livestock. Questioning of Mauritanian herdsmen indicated, however, that the grazing stock and S. gregaria actually have quite different food preferences (Fig. 12).

Fig. 12 Responses given by herdsmen concerning food preferences of locusts (L), S. gregaria, camels (C) goats (G) and sheep (S) (Latin and vernacular names cited)

The food preferences were ascertained by questioning 20 (in 1993) or 50 (in 1994) Mauritanian herdsmen. In the survey, only annual plants were considered which are dependent on rainfall for their presence and which represent a proportion of at least 10 % of the total vegetation.

In 1993 in the test area Trewia, 3 different experiments were carried out to determine the vegetation losses caused by S. gregaria hopper bands. For this, first of all the utilizable biomass in the local vegetation (F. olivieri, S. thebaica, Boerhavia spec.) was quantified by performing comprehensive weighings at selected random points and making a final estimate for the whole area. Secondly, the number of larvae present in each region, as well as their stages and duration of stay, were established. The observation areas showed a degree of coverage by the vegetation of 40 - 80 % (S. thebaica or mixed S. thebaica-Fagonia). The insect densities on the 3 areas were 20, 10 and 25 larvae/ m². From these values and the previously determined food uptake values per animal, the vegetation losses were calculated. In the three experiments, which covered between 15 and 36 days, losses determined were 1.4%, 5.7% and 18.2% of the total vegetation. The differences can be explained by the differences in the composition of the vegetation. The losses are greater in the regions where there are more high growing plants, e.g. S. thebaica, present. For this growth form, the "biomass" utilizable to S. gregaria, namely the leaves, is decreased relative to the total biomass (leaves plus stems), i.e. the loss of vegetation through the larvae is a higher percentage (The ratio of leaves to stems varies for Schouwia, depending on the growth form and density, between 1:1.5 and 1:2). Relative to the livestock, however, the losses are lower, since leaves and at least a majority of the stems will be eaten.

go to next section of this publication