go to previous section of this publication
2.1.3 Ecotoxicological investigations
To complement the field and semi-field trials on the impacts of IGRs (benzoyl phenyl ureas), natural insecticides (M. volkensii) and fungal preparations (B. bassiana and M. flavoviride) on desert locusts, a series of ecotoxicological laboratory and field tests were carried out in Mauritania, starting in 1992. These were designed to investigate potential side-effects of the said products on non-target arthropods, and assess possible risks under normal conditions of application in the field, by way of comparison to conventional products. The laboratory tests focused on species which are typical of the prevailing locust biotopes in Mauritania, and which are at high risk of exposure in the event of insecticide application. In addition, a beneficial organism - Pharoscymnus anchorago (Coleoptera: Coccinellidae) - was also included in the investigations which, by virtue of its status of natural antagonist of date scale insects, is of major economic significance. The small-scale field tests focused on terrestrial arthropods.
2.1.3.1 Laboratory tests
Methods and classification of risks
The test organisms were exposed via three different routes, namely the topical, residual and oral routes. In the case of topical application, a microapplicator or a ULV sprayer was used to apply a defined quantity (µg a.i./animal or µg a.i./g body weight) of active ingredient (a.i.) dissolved or suspended in organic solvents, directly to the cuticles. This is equivalent to direct exposure to insecticide aerosols on application in the field. With residual exposure, the cuticle itself was not directly treated, but an artificial (e.g. petri dish) or natural (e.g. leaves) substrate (µg a.i./cm2) on which the test organisms were placed for a certain period after application. Contamination or infection also takes place here via the cuticle through contact with insecticide residues on the treated surface. Under field conditions - especially with remanent insecticides - this is the most probable mode of exposure. With oral exposure, the active ingredient is ingested via the food. For this purpose the preparations are e.g. injected into prey or applied to plant foods and subsequently fed to the animals. When the animals are fed with prey, which are usually only partially consumed, the precise intake of active ingredient is often not known. Consequently, in the tests in Mauritania the preparations were dosed such as to ensure an active ingredient intake equivalent to or higher than that in the field. In addition, the test organisms were also often fed several times.
Where no side-effects are observed under laboratory conditions, it can as a rule be assumed that none will occur in the field, provided that the safety margins are kept sufficiently tight. On the other hand, if effects are demonstrated in the laboratory, this does not automatically mean that they will also occur in the field, since exposure in the field is usually lower. If and when effects are observed in the laboratory, further tests under semi-field or field conditions are required. Having said that, an initial risk assessment can be made on the basis of laboratory tests. Where dosages tested in the bioassays are equivalent to the maximum recommended field doses (maximum deposition of active ingredient relative to unit area), Hassan et al. (I985) propose the following risk classification for beneficial arthropods:
|
Risk class |
effect1 |
rating |
|
I |
<50% |
harmless |
|
II |
> 50 % |
slightly harmful |
|
III |
> 80 % |
moderately harmful |
|
IV |
> 99 % |
harmful |
1
e.g. mortality, reduction in parasitisation or predation rate etc.Although the boundaries of the individual risk classes are not generally accepted, it is broadly agreed that effects > 50% are to some degree or other harmful. The maximum concentration of active ingredient to which an organism is exposed in the field upon insecticide application is a function on the one hand of the biology of the affected organism (diurnal or nocturnal, living concealed or exposed etc.), and on the other hand of the recommended field dose for effective locust control, which varies from product to product (Tab. 1). The maximum exposure (Emax) of an organism, or the maximum deposition of active ingredient on a substrate (soil, vegetation), can be approximated on the basis of those data. A field dose of 100 g a.i./ha would, for instance, result in a maximum deposition of 1 µg a.i. on the approximately 1 cm2 dorsal surface of a freely exposed beetle, and on smaller surfaces correspondingly less. Freely exposed vegetation or bare soil would be contaminated to the same degree. In reality the value is usually lower, as the total surface (of all vegetation, soil particles etc.) to which insecticide droplets can adhere is larger than the ground surface area. In fruit crops, for instance, only 40 - 50% of a product is deposited in the canopy (Barrett et al. 1994). Consequently, simple calculation of the insecticide dosage relative to the ground surface area always represents a worst case scenario.
Tab. 4 Dosages recommended by FAO or manufacturers for the locust control agents investigated in the bioassays (relative to the ground surface area in a.i./cm2).
|
Active ingredient |
Recommended dosage |
Source |
Max. a.i. deposition |
|
IGRs |
|||
|
Diflubenzuron |
60 g a.i./ha |
FAO |
0.6 µg a.i./cm2 |
|
Triflumuron |
50 g a.i./ha |
manufacturer |
0.5 µg a.i./ cm2 |
|
Teflubenzuron |
50 g a.i./ha |
manufacturer |
0.5 µg a.i./ cm2 |
|
Entomopathogenic fungi |
|||
|
Beauveria bassiana |
2.5 x 1013 spores/ha |
manufacturer |
2.5 x 105 spores/ cm2 |
|
Metarhizium flavoviride |
3-5 x 1012 spores/ha |
manufacturer |
3-5 x 104 spores/ cm2 |
|
Metarhizium flavoviride |
2.5 x 1012 spores/ha |
manufacturer |
2.5 x 104 spores/ cm2 |
|
Botanicals |
|||
|
Melia volkensii |
10 g a.i./ha |
manufacturer |
0.1 µg a.i./ cm2 |
|
Combined products |
|||
|
Fenitrothion + Esfenvalerate |
250 g a.i./ha |
FAO |
2.5 µg a.i./ cm2 |
|
Chlorpyrifos + Cypermethrin |
220 g a.i./ha |
FAO |
2.2 µg a.i./ cm2 |
The evaluation of the bioassays summarised below is based on the system proposed by Hassan et al. In a number of tests, not only the maximum recommended dosage was investigated, but also a range of different dosages or concentrations. Thus LD/LC50 and/or ED/EC50 values including the 95% confidence interval were calculated, providing a more accurate picture of the dosage-effect relationships of the respective products. A product is classed as harmful whenever the LD/ LC50 or ED/ EC50 is lower than the maximum expected environmental concentration in the field. However, a classification into the risk classes as defined above based on LD/ LC50 or ED/ EC50 values is not possible, due to the different quality of the data. Instead, we adopt the classification of acute toxicity which has been proposed by FAO-LOCUSTOX for Sahelian epigeal non-target arthropods:
|
Risk class |
LD/ LC50 or ED/ EC50 |
rating |
|
I |
> 100 |
very slightly toxic |
|
II |
10-100 |
slightly toxic |
|
III |
l-10 |
moderately toxic |
|
IV |
< 1 |
highly toxic |
Results
I. IGRs
Three products from this group were tested in various formulations: triflumuron (Alsystin), teflubenzuron (Nomolt) and diflubenzuron (Dimilin). IGRs act as stomach poisons and, to a lesser extent, contact poisons against the larval stages of various insects. The effect is based on an impairment of chitin synthesis, as a result of which a normal moult in the course of metamorphosis is prevented. IGRs are suitable for barrier application due to their relatively high remanence. In the bioassays, ULV and OF formulations were used for topical and residual applications, and WP and SC formulations for oral applications (exposure via food).
As expected, relevant side-effects were observed only in the larval and nymphal stages (Tab. 5a). In the case of the assassin bug C. arenaceus, the calculated LD50 values involved a high degree of experimental error, as indicated by the wide 95% confidence interval, and the classification is provisional. In 1992, C. arenaceus was considered a potential test species for standard toxicity testing, due to its ecological importance as a predator of cicadellids and abundance in the field. However, the species turned out to be rare in subsequent years and was not used any more. Therefore, the preliminary results with IGRs (and M. volkensii, see below) could not yet be confirmed. During exposure to contaminated prey, adult bugs displayed impaired fecundity, and a reduced egg hatching rate. Both effects were reversible. In some cases, ULV formulations showed acute toxicities, which were attributable to the carrier substances.
Diflubenzuron had marked side-effects on larvae of the coccinellid P. anchorago, the OF formulation being more toxic (86% mortality at 0.6 µg / cm2) than the WP formulation. The EC50 (parameter: adult development) of the WP formulation (0.61 µg / cm2) was within the range of the maximum expected environmental concentration of 0.6 µg / cm2. The toxicity to the larvae would probably have been greater in the field, because due to the high remanence of diflubenzuron the duration of exposure would be longer there than in the laboratory experiment, where it was limited to 2 days. Since experience has shown that the various IGRs have similar effects, it is in principle not recommended that they be used in or around date palm plantations. Surprisingly, subadult spiders with at least one moult still to come subsequent to the point of application were not sensitive. This finding has since been confirmed by further investigations (Peveling et al. 1997).
II. Entomopathogenic fungi
The bioassays conducted in Mauritania with entomopathogenic fungi were designed to investigate the host range. Two species were tested, B. bassiana (conidiospores, MYCOTECH Co., USA) and M. flavoviride (blastospores, BBA Germany, and conidiospores, LUBILOSA project Benin). The applied spore dosage of B. bassiana corresponded to the recommended equivalent field dose of 2.5 x 1013 spores/ha. In the case of M. flavoviride, higher doses than recommended were tested - depending on the test organism. B. bassiana displayed a very broad host spectrum. Infections were identified in all test species following topical application. Spiders (P. viridis) and assassin bugs (C. arenaceus) were especially sensitive, showing mortality rates of 90 and 100%, respectively. Assassin bugs also contracted secondary infections following contact with treated prey. Even some of the robust tenebrionids suffered infections (Tab. 5b). By contrast, blastospores of M. flavoviride were not pathogenic for any of the test organisms. This also applies to the oil-kerosene formulation of conidia, although at a high dosage (8 l/ha) this product did produce acutely toxic effects in the bug Cosmopleurus sp. (Lygaeidae). As regards a comparison of the pathogenicity of the two fungi, it should be noted that the tests were not carried out simultaneously and, with the exception of Trachyderma hispida, the same test species were not available. Nevertheless, it can be concluded from the tests that the B. bassiana strain had a wider host spectrum than the two M. flavoviride strains. However, more recent investigations in Senegal have demonstrated that the latter are under laboratory conditions pathogenic for Hymenoptera, an order not investigated in Mauritania to date.
III. Melia volkensii
Similar to the IGRs, the effect of M. volkensii is based on an impairment of the larval development of insects. Consequently, in the tests to date only larval and nymphal stages have been investigated (Tab. 2c). M. volkensii was moderately toxic to P. anchorago. At higher dose rates, the larval development was delayed by up to several weeks, or was prevented entirely. However, the EC50 was well above the expected initial environmental concentration, and hence even at a higher dosage lasting hazards to the coccinellid fauna are unlikely, especially since the active ingredients of M. volkensii are broken down very rapidly under field conditions. In the case of the assassin bug C. arenaceus, whilst the ED50 was low, it was associated with a high degree of experimental error (very large confidence interval), and no statistically sound risk classification could be conducted.
IV. Conventional insecticides
in the tests, the acute toxicity of two broad-spectrum combined products - each consisting of one organophosphate and one pyrethroid - was tested: Profenofos/cypermethrin (Polytrin C) on spiders and assassin bugs, and fenitrothion/esfenvalerate (Sumicombi) on tenebrionids and coccinellids (Tab. 5d). With the exception of tenebrionid beetles (moderately toxic), both insecticides were highly toxic to the non-target arthropods tested. Consequently, application of these products over large areas is likely to result in considerable side-effects. In particular, on application in or around date plantations, direct contamination of the palms should be avoided - as with the IGRs - to prevent hazards to the beneficial fauna.
V. Comparative evaluation of the various preparations
IGRs present a low risk to terrestrial arthropod fauna. One exception are coccinellids, and probably also chrysopids which, as natural antagonists of scale insects, are of major significance in Mauritanian date production. It is therefore recommended that these products not be applied on date plantations. Side-effects on aquatic arthropods have already been investigated to a sufficient degree in other countries (e.g. FAO-LOCUSTOX, Senegal). IGRs in principle pose a high risk to these species. Contamination of temporary bodies of water, such as occur in Mauritania following the rainy season, must always be avoided.
Of the two fungi investigated, B. bassiana displayed a wider host range than M. flavoviride. The latter would therefore seem more appropriate for potential application, provided that its effect on desert locusts is equally pronounced. However, pathogenicity and virulence can even vary within one species from isolate to isolate. According to the present state of knowledge, it can be assumed in principle that the risk of infection for terrestrial arthropods in the field is low, since the spores quickly lose their viability under the prevailing climatic conditions in Mauritania.
The first trials with M. volkensii produced no evidence of harmful side-effects. However, its efficacy against locusts at the dosages applied to date (10 g a.i./ha) was unsatisfactory. It is therefore to be assumed that the efficacy will need to be improved considerably in order to achieve adequate control results. In this case, side-effects on beneficials such as P. anchorago would have to be re-assessed.
All the alternative products tested here have one thing in common, namely that their effect is extremely delayed and - with the exception of the fungi - that they are appropriate only for the control of hopper bands. The advantage of the high degree of selectivity thus contrasts with the drawback of low application potential. Consequently, when crops are acutely threatened it will still be necessary to continue using unspecific synthetic insecticides. In the latter case, compliance with the recommended dosages is absolutely imperative. Hopper bands in date plantations should be treated exclusively on the ground using hand sprayers, in order to keep contamination of the leaves and damage to the beneficial fauna as low as possible.