General conclusions with regard to the effects of an organophorus insecticide (temephos) on entomocoenoses associated with preimaginal stages of the simulium damnosum theobald complex (diptera: simuliidae)

@The results presented here for a doctorate in natural Paris-Sud, Orsay centre. Research worker at ORSTOM, are reported in detail in a sciences (Elouard, 1983) at 24 tue Bayardr 75008, Paris, ocP/vcu/HYBro / 84 .2 ORIGINAL: FRENCH thesis submitted the University of France. WORLD HEALTH ORGANIZATION ORGANISATION MONDIALE DE LA SANTE GENERAL CONCLUSIONS WITH REGARD TO THE EFFECTS OF AN ORGANOPHOSPHORUS INSECTICIDE (TE}GPHOS) ON ENTOMOCOENOSES ASSOCIATED WITH PREII.,IAGINAL STAGES OF THE SIMULIW DAI.{NOST]M THEOBALD COMPLEX (OIPITM: --SIMULmEEII- by Jean-Marc Elouard2 I. STATEMEM OF THE PROBLEM Human onchocerciasis is a debilitating dermal fi.lariasis which is widespread throughout tropical Africa; it is more serious in clinical, social and economic terms, however, in guinean or sudanian savanna areas. The pathogen responsible for this endemic disease is the filaria Onchocerca volvulus (Leuckart, 1983), which is strictl"y limited to man an-I'-i3-ffismittedliiGt-Africa by female blackfly of the Sirmrlium damnosum complex. The flyisbothvectorandintermediatehost1th@gesiedfroman infected person, has to complete a cycle of development within the fly to reach the infective larva1 stage, which takes about seven days. The diseaseis transmitted only by parasitic forms having compLeted this obligatoryintravectorial cycIe. The invasion of tissues of ectodermal derivation by the microfilariae produces various pathological states, the most serious of r*rich isirreversible blindness. Ihis clinical event causes human populations to emigrate from many valleys infested by blackflyr or at least to avoid livingthere. Thus onchocerciasis constitutes a considerable obstacle to the economic and demographic development of West African countries, particularly savanna zones where the disease is prevalent along most natercourses. The iseue of this documont do€s not constitute formal publication. lt should not be reviewed, abstractod or quoted without the agreement of the World Health Organization. Authors alone are responaible for views expreased in signed articles. Ce document ne constitue pas une publication. ll ne doit faire I'objet d'aucun compte rendu ou r6sum6 ni d'aucune citation sans I'autorisation de I'Organisation Mondiale de la Sant6. Les opinions exprim6es dans les articles signds n'engagent que leurs auteurs. ocP/vcu/HYBrol 84.2 Page 2 In the absence of any effective treatment (prophylaxis, chemotherapy, surgery) suitable for mass application, the only rnethod possible at Present for-prlventing the epread of the disease is vector control. In Ivory Coast the Simulium danrrosum complex comprises six cytotyPes now designated as specffiffi6ffi s.s., g4!39ry, 9. sPnctipauli,.g@' S. "q,r"rosiila-nffiah"ttse,-TlETered Uy ttre term S. da-Itosum s.1. !'hen no a$ffi;E$n-is made between the various membere of the complex. The larval ecophase of the S. damnosum s.1. cycle of develop,m.ent is aquatic and etrictly confinea to raplas. This makes it vulnerable, for its distribution is contenErated in well-defined stretches with clear liurits. For this reason limitation of vector populations by means of larvicides was the method adopted by the Onchocerciasis Control Programe (OCP), which has been operated since 1975 by the l{orld Health Organization (WUO) in the Volta Rivei Basin area. It is at present the most effective and practicable technique in view of the vast area to be protected, with its 764 000 sq km and its 18 000 km of potential breeding sites for S. darnosum s.1. Since the larval development cycle is so short, weekly epraying hae proved necessary for effective control of vector populations. The insecticide eelecEed for a large-scale campaign of this type, due to last for about 20 years, mrust have properties that allow it to meet often contradictory requirements such as effective action against the larvae of fhe vector, ease of application, lowest possible co8t, little reeidue but far-reaching effect, harmlessness for man and marunale and lowest toxicity possible for the rest of the aquatic environment. Temephos (or Abate@, an" organophosphates insecticide used by OCP, was chosen because it was the only substance available strich met the desired criteria. It is not entirely selecEive, however, and has a destructive effect on non-target organisnrs living in the same biotopes as S. damnosum s.1., and even other biotopes in rivers where the species is not present. Apart from depressing the biotic potential of the erea treated, there is a great risk that the insecticide may modify the natural balance in one or more waya 'causing the disappearance of economically important species orr on the other hand, causing che undesirable proliferation of other organiems. In the aquatic environment such an imbalance usually takes the form of a break in the food chain wtrich, starting with the initial input of solar energy, leads to the most economically important link: fish. llydrobiological studies were therefore essential to ensure that no ecological catestrophe would result from the long-term and large-scale weekly spraying of temephos. Since the first year of operation of OCP, three groups have been monitoring the rivers in Upper Volta, Ghana and Ivory Coast. Unfortunately, it was not possible to collect enough bioecological data, to serve as a reference for fauna, before spraying began. The planned extension of the Progranrme area to the west and south (Phases II and III), however, provided an opportunity to remedy this omission. Ihe aquatic monitoring protocol was therefore put into effect on the Bagou6, the Maraou6 and the Sassandra, which had remained untreated until subjected to ocP /vcu IHYBT1/ 84 .2Page 3 repeated applications of temephos. The protocol, however, vras based on European or American experience and it was not certain that it wouLd be appropriate and sensitive enough to detect all important modifications in thefauna of African rivers. rt t,herefore seemed useful to carry out complementary investigations to determine Ehe real impact of the toxic substance in the short-, medium- and long-term, and to understand by what mechanism, whether direct or indirect, the insecticide acts on different species of insects and on their associations in cormnunities. The data were collected at Bouak6 by the ORSTOM Hydrobiological Laboratory(rvory coast) from 1975 to 1980. The wLrk was supported throughout by wHogrants, under the terms of an agreement reached with OCp. II. DISTRIBUTION OF S. DAMNOSTIM S.L. AND ASSOCIATED FAUT.IA To determine the effect of temephos on fauna associated with thepreinaginal stages of s. damnosum s.1., it lras necessary to establish adetailed picture of rtr-Edo-iifi]-of the larvae and nymphs of the complex,identify the species associated with it and define the conditions under which such associations occur. Preimaginal stages of S. damnosum s.1. have a relatively marginal ecoLogy among lotic speciesl they-aTEEin-tin fast-flowing reaches and showpreference for floating substrates where available. On fLoating substrates sub.jected to severe rheological conditions the onchocerciasis vector is the most abundant species, outnumbering by far the few lotic species that succeedin becoming established. Thus, due to a range of ecological conditions, the entomocoenosis of these substrates lacks richness and diversity. The optimum rate of flow (preferred current speed) and the distribution ofpopulations around this optimum (rheophilic valence) vary with the preimaginalstage. Early stages tend to congregate in swift currents, while older stages can adapt to more varied and generally less rapid flow rates. Distribution of the species is not Limited to floating substrates infast-flowing reaches, however. rt is modified mainly by iariations in thedischarge of the river corresponding to seasonal changes, and by the tendencyto disperse due to pressure of larval overpopulation. These factors obligelarvae to live in biotopes falling far shoit-of the optiumm. It may be accepted that at any given season the density of S. damnosums.1. larvae decreases along with two major occurrencesr tire sf6iffi!-6FlIi rate of flow and the transformation of floating, undulating subitraies intoinnnobile substrates attached to the river bed or banks. The females of S, damnosum s.1. lay their eggs in contiguous clusters on substrates best exp6ffiEfid currents and upstream of breeding sites.This gives a very high density of the early stales of the insect on ttre substrate concerned, and makes it difficult to monitor their colonizationdynamics. The older stages, on the other hand, due to their tendency tomigrate, carried by the drift, and their greater capacity to aaapt il -fluctuations in the current' present colonization curves showing a progressiveincrease corresponding to the number of days of inrmersion of the suistfates. Fauna of associated species are divided between these two types ofdynamics. rn swift currents, however, competition from the preiominantS. danrnosum s.1. restricts their colonization. Spatial associations between preimaginal stages of S. damnosum s.1. and ocP/vcu/EYBto I 84 .2Page 4 lotic insects are subject to three major greatest abundance of each species; the velocity of the current. ecological factors: the period of type of substrate colonized and the According to our observations, with the exception of S. unicornutum and Nanocladius sp., the density of lotic insects is at its rEfiffi-ffiE-Ee lreter level is falling, at 1o!, lrater or when the 1evel is beginning to rise again(wittr the sampling techniques used this could not be ascertained with regard to abundance). The non-target insects also prefer moderate or slow currents(0.30-0.60 m/s). The seme holds true for other species of Simulium found on stretches of the lower Maraoud (S. adersi, S. tridens, S. unicornYtum). As soc iat ions be tween thes e spec iili-GrviilffiE-on66EErEfaEf s vec tor occur only in moderate currents (0.40-0.80 m/s), and mainly involve older larvae. It may therefore be stated that despite the year-round presence of the onchocerciasis vector, very few fauna are in spatial association with it, particularly at the early, more rheobiontic stages. In general, a smal1 number of S. damnosum s.1. is found in association with large numbers of non-target-fi8;E;; rocks, wtrile on floating substrates a sma1l number of 1oticinsectsisfoundinassociationwith1argenumbersofE@s.1. larvae. Functional associations do exist, however, between the latter and non-target fauna. The presence of S. darrrosum s.1. limits rheophilic valences andthedistributionofS.adersiaiI@eintheenvironment'and prevents the establishnent of S. echoutedeni. The preirnaginal stages of S. damnosum s.1. also coastitute a link in the aquatic food chain. Most predators, holrever, are not found in cloee association with S. damrosum s.1.; they are mainly nonselective, passively entomophagouaorffi;ffirous,preyingoPportunistica11yon@, living in biotopes ntrere larvae of the onchocerciasis vector ere rare and feeding on them only when they are carried along by the drift. These passive predators are mainly the Hydropsychidae (Trichoptera) and are found on or under rocks, depending on the species or larval stage. III. EFFECT OF TEMEPHOS To limit the transmission of onchocerciasis to an economically acceptable 1eve1,a11@s.1.1arvaemr8tbedestroyedwithinthetar8etarea' in aLi ttte Fot-ere they occur. It would have been na'ive to assume that the ingecticide had a strictly specific action against this species. Even if this had been the case, it was inconceivable Ehat the eliurination of S. darnosum s.1. should not have more or less direct consequences for the rest offi'ffi=E"na, since the species occupies a niche in space and serves as prey, thus constituting a link in the food chain. In view of the rapid development of larvae of the onchocerciasis vector, destruction by chemi""l ,"..t" must be carried out in weekly spraying operations. fni" schedule makes it possible to distinguish betrreen irunediate or short-term effects and the delayed, ,ore far-reaching effects resulting from repeated treatment, known as long-term effects' 1. Short-term effect As expected, temephos is toxic in the short term for most lotic insects in the doses used by OCP. This has been established by the study of rates of drift, by LC56 dlterminations carried out in gutters and by quantification of the fiuna-f..".nt on substrates (see Annexes). These methods are not ocP/vcu/HYBto/ 84.2 Page 5 equally precise and reliable, however, and vary in terms of the time and means required. A routine spraying operation carried out by OCP (theoretical concentration within the range 0.05 to 0.1 mg/l calculated over 10 minutes) produces, after a period of latency varying from 15 to 45 rninutes, a massive detachment offauna reflected by a rise in the drift index and a fal1 in the density offauna present on substrates. Although all taxonomic groups are affected, the larvicide is selective to some degree; whereas Tricorythidae, chimarra and some Baetidae areparticul-arly susceptible and disappeaffi.r:il-fhe biotopes almost completely, other species such as S. schoutedeni or the Chironomidae manage to resist thedoses applied by oCP. EEtTEEeT;ces in srrsceptibility among organisms, wtrich allow some fauna to survive in the lotic environment, bring about profound biocoenetic modifications in the long term. It has also been observed that there are variations in susceptibility among the different larval Btages of lotic species. Early stages are much more seriously affected by temephos than older organisms, the difference in susceptibiliEy reaching a factor of 10. This phenomenon prompts two observations: - The effect of spraying will vary according to the mean age of larvae; the older they are on the treatment day, the higher the number of survivors. rn line with this theory, species with a short cycle (chironomidae, for example) have a clear advantage over species whose development tak; several months (Trichoptera), since early stages of the Latter may be subjected to several spraying cycles. - l,ltren comparing the toxicity of different insecticides or following changes in susceptibility of organisms to a certain insecticide, it is essential that the age of individual organisms or the structure of populations be taken into account. In the short-term, temephos does not have too serious an effect on the physiology and etholoW of surviving organisms. These observations result from a study of rates of exuviation and of the nyctohemeral pattern of drift. The values for these qwo phenomena do not vary greetly in the hours before and after spraying. Although in possession of the quantitative results reported above, we should be wary of advancing a figure for the short-term toxicity of temephos. since the toxicity varies among species, according to preimaginal stage and with different experiments, the mortality recorded obviously depends upon the faunaL composition of the colonies observed and the structure of the populations concerned. Depending on whether susceptible or lese susceptible species predominate, the effect of temephos may be described as catastrophic or slight. Nevertheless, in comparison with the results obtained with other insecticides (DDT, Lindane, deltanethrin, malathion, etc.) ternephos toxicity can be regarded as quite acceptable. At present it is the only available chemical rreapon that is effective against s. danqosum s.1.1, moderately toxic 1 On the Bandama, the twin species resistance to temephos. S. soubrense and S. sanctipauli show ocP/vcu/HYBro I u .2Page 6 for non-tar:get fauna and not too costLy. Only Bacillus thuringiensis H-14 seems capable of replacing temephos by virtue of its selective toxicity for Simulium. Large-scale application of BaciLlus thuringlgglg is stil1 somewhat pi"uGatica1becausetheformu1atio',"ffitab1eforairborne spraying. 2. Long-term effect Study of the short-term effect shows the irmnediate impact of an insecticide when spraying continues over several years, however, the long-term effect is nrore important since it shows the modifications brought about in populations and food chains. Long-term studies should be more general and include the overdosing and underdosing that occur in routine treatment. The latter are inevitable, owing to wide hydrological fluctuations and the time lapse, sometimes lasting several days, between lrater leve1 readings and calculation of the quantity of insecticide to be applied, based on rates of discharge. Among the stressed the fricorythidae prol i feration most obvious indicators of long-term modifications, we have eventual disappearance of S. adersi, the rarefction of and the Bastidae CentroptffimJi-. 4 and Pseudocleon sp. and the of S. schoutedeni and the Chironomidae. Some species are present on the river reaches both during periods of treatment with temephos and outside guch periods. On this time scale, modifications in population structures are easily observed, whereas variations in the numbers living on natural substrates are more difficult to monitor. Evaluation of the quantitative variations produced by temephos in the long term has been possible only in the case of studies of dietribution within certain reaches, that is, where experimental conditions limited variations due to seasonal change or type of substrate. The overall value of 437. reduction in the quantity of fauna should be regarded as relative because it includes insects which have proliferated due to the insecticide (Chironomidae, S. schoutedeni), those which have greatly diminished (Ephemeroptera, someffiu-ffi-Fecies) and others which have shown little variation(Hydeopsychidae). Simulium is a special case because the disappearance of S. danrrosum s.1. and S. adersi has been partly compensated by the F?6T-iteration of s. sE6ffii6ti. In a monitoring programrne applied on such a vest scale, with such wide seasonal fluctuations and variations in distribution, it seems difficult to quantify the long-term effects of femephos in terms of variations in population densities. This approach woul-d require an excessive amount of sampling and would be too costly in time and equipnrent. On the other hand in the long-term structuraL variations in the colonies present are more easily identified, whereas interpretation of this type of data is inconclusive in the short-term. This is due to the establishment of stable coenotic structures which are typical of temephos treatment periods and very different from those observed during periods rrithout treatment. Changes in structure brought out by multifactor analysis are detectable whatever the type of substrate sampled. During the months wtren the water level is falling and at low water, periods of treatment wifh temephos and periods without temephos each have their own typological structures. In all cases, the Chironomidae group and S. schoutedeni seem to figure largely in periods of spraying activity, proliferating at the expense of S. damnosum s.1., q. 44eEl, the Tricorythidae and many other Ephemeroptera. ocPlvcvlHYBro/84.2 Page 7 rt should be noted that the proliferation of some chironomidae is not specificaLly l-inked to pollution by temephos or any other insecticide, but seems to occur with all types of pollution. The proliferation of S. schoutedeni and the Chironomidae could contributeto the maintenance of uyai6ffiElffiae-f,6pulations on reaches treated wirh temephos. lrle have shovrn that these Tricoptera are only slightly or moderately susceptible to temephos at the doses used by OCp and that they afe passive Predators of S. damnosum s.1. larvae. The disappearance of the latter has thus had little effect on Hydropsychidae populations since it has been compensated by the extra food provided by Chironomidae and S. schoutedeni, which are relatively insusceptible to temephos. I^Ie should also mention that we found few variations in Ehe ecology of mostgpecies subjected to repeated spraying with ternephos. This appl,i.s not or,lyto seasons of high density of the species but also to their preferred current speed and theophilic valence. The only exceptions are Centroptilum sp. 4 andCricotopus quadrifasciatus, whose period of highest denffid s. tridens, which has narrowed the range of its preferred rate of flow. The nyctohemeral pattern of drift, on the other hand, does appear to be affected by temephos, being less marked and even becoming patternles6 after several months of treatment. Tire overall effect of temephos can therefore be sununed up a9 follows: inthe short-terrn it produces numerical variations within populations due to itstoxicity; and in the long-term it modifies the structure of the colonies ofinsects living in the reaches concerned. However, short-term variations indensities obviously lead to variations in biocoenetic structure, just as changes in the structure of colonies noted in the long-term are a reeult of numerical changes within populations. The difference observed between short-term and long-term effects is linked to variations in natural perameters and the techniques used in the investigations. rn the long-term, norrnal seasonal fluctuations can mask numerical changes due to temephos, while in the short-term a sPraying operation merely reduces the numbers oi ali taxa to agreater or legeer extent without causing any proliferation of some species; structural differences are thus less marked. IV. PROPOSAL FOR A NEW PROTOCOL FOR MONITORING THE EFFECTS OF TEI.{EPHOS In all investigations the quality and precision of the resuLts dependlargely on the techniques used. with regard to the monitoring of riverstreated with temephos by ocP, most of the methods used were ""le"t.a uyextrapolation of data obtained outside the tropics and r*,hose applicabiiity totropical rivers had not been thoroughly examined. we therefore feel it oPPortune to carry out a critical review of the results obtained and to drawfrom them certain suggestions for improvement. The hydrobiological monitoring techniques used by ocp are as fo1lows: - collection of saxicolous fauna by means of Surber sampler and artificial substrates (cement blocks)! - sampling of lentic fauna by means of E. grab; - study of day and night drift. Apart from the surber sampler, the other methods, which were not ar1 covered in our study, were not able to show that temephos had any effect at ocP/vcu/HYBrol 84 .2 Page 8 all on non-target fauna. Therefore it would not seem useful to keep them in the progranme orr if they are kept, they must be made more effective. In general terms, determination of the effects of an anti-blackfly larvicide on non-target fauna should be carried out in three stagesl the first two designed to assess short-term toxicity and the third to assess long-term toxicity. The stages are as followsl - rapid short-term toxicity tests on a new insecticide or a new formulat ion I - determination of the short-term susceptibility of the different species to the pesticide used and some monitoring of how this susceptibility changes over timel - long-term surveillance, thaL biocoenotic changes resulting from 1. The Drift is, determination of Lhe profound repeated treatment. Rapid assessment of the toxicity of a new insecticide can be made by studying the drift. This method, however, is more qualitative than quantitative end it is difficult to compare the results of two different experiments. The method can be made more reliable by setting standards for all experiments with regard to the number of sampling procedures, their frequency, the duration of the experiment and the time of day that spraying takes place (see Annex II: Drift). In our opinion, however, this will never be more than a rough method of assessing the impact of a toxic substance, since the quantity and structure of the fauna Present is not known. It is therefore not possible to compare mortality rates among organisms nor, more important, levels of sensitivity. 2. Gutters The use of gutters in situ seems to be the most precise method for the short-term studt of theET[dty of an insecticide. It gives a relatively accurate picture of the mortality of organisms because the number of individuals tested is known. Moreover, the use of multiple Elutters makes it possible to comPare the toxicity of several insecticides or several concentrations of the same substance with that of a control. Using this latter procedure, the reactions of organisms can be calculated (and expressed as LC5g or.-LC96) and they can be classified in order of "ecological susceptibilityrr. It is essential, however, to take the size or larval stage into accounE, for susceptibiLity varies widely according to the age of the larvae. OCP tests the susceptibiLiEy of S. damnosum s.t. in stages L5 and L6, stage L7 being exc1udedbecauseoffiro1ogica1andbehavioura1pecu1iaritieswhen approaching nymphosis. It would seem reesonable to use the same stages (L5 and L5) for studying other species of Simulium. We have, in fact, demonstrated that it is possibl" to s"iZEt-ne seven larval stages of S. adersi, S. tridens, S. unicornutum and S. schoutedeni by biometry. As for lotic species oEher than simulium, the susceptibility of the different stages should be ascertained before any method is recornmended- In selecting techniques the prevalence of the species in the environment should be taken into account, together with difficulties of identification inherent in Ehe stage or the species; identification of the stage of development can be carried out either by biometry or according to morphological criteria. ocP/vcu/HYBro/ 84 .2Page 9 Since the physical and chemical characteristics of the erater have a considerable effect on the efficiency of an insecticide, it would be valuableto carry out tests iq simiLar conditions as to time and place (see Annex IV:Gutters). The most favourable period in terms of both logistics and fauna would seem to be when the water leve1 is falling and at low water. Repetition of the same tests in the same ecological conditions should make it possible to monitor the susceptibility of insects during treatment peri.ods, thus removing Ehe uncertainty arising from the differences observed between the varioustests considered in this paper. Many practical problems wil-L obviously be encountered in theimplementation of this type of protocol. It would be almost impossible totest all species living in lotic ecosystems, since some of them are extremely rare. The first criterion to be adopted should therefore be density: this could be modified in Line with new knowledge gained. For example, ,or" species such as S. sshoutqdeni or the Chironomidae, wtrich were scarce or even extremel,y rare before temephos pollution, proliferated during treatmentbecause of the gradual reduction in competition. Seasonal variations in densities of species and of their preirnaginal stages depend on their breeding cycres and on ecological factors, and are therefore very wide. This constitutes a further constraint for the testsdescribed, since not all- the abundant species representative of the lotic ecosystem can be tested at the same time. The appropriate period for each species will obviously be wtren the density of the stage adopted as representative'of the species is at its maximum. However, it is out of thequestion to carry out separate experiments for each species. A compromise must be found involving groups of species, based on the development and composition of the major lotic associations. Considering the susceptibility of species in associations has the advantage of providing information on the typology of the colonies and how rhey function; this knowledge will contribute to the interpretation of variations in the structure of the entomocoenoses which appear in the long-term. 3. Sampling of Fauna on Natural Substrates comprehensive large-scale and long-term monitoring of the effects of temephos must be based on structural variations in the entomocoenoses, sincethe variations in density caused by the insecticide are difficult todistinguish from those due to fluctuations in ecoLogical parameters.I'lultifactor analysis (factorial correspondence analysis of correspondingfeatures, hierarchical classification in ascending order), on the other hand, makes it possible to define the typoLogies characteristic of treated and non-treated periods, without losing sight of seasonal variations. Its power of discrimination is so great that structures can be identified even where theidentification of organisms has not been achieved at the species level, but at superior taxonomic levels such as tribe, family or order. I,Ie have seen that modifications caused by temephos can be discerned in anygiven hydrological season, whatever the type of substrate sampled, that is, whatever the lotic entomocoenosis in nhich the preimaginal stages areinvolved. Thus it does not seem necessary to cover all types of substrate inthe protocol for a Large-scale monitoring progranune. Statistically speaking, the number of samples being equaI, it would be better to colLect them in the sare conditions, thus restricting the field of investigation. Sampling of saxicolous fauna is obviously indicated, in view of the number of species Present on rocks, their density, the easy application of collection techniques and the quality of the results obtained. ocP/vcu/HYBrol 84.2 Page 10 The Surber sampler appears relatively well suited to the collection of rock fauna. However, experience has prompted us to ProPose certain modifications to the protocol for this method, concerning both the size of the area sampled and the size of the sample (see Annex III: Surber sampler). The period to be monitored is also a major factor to be taken into account. On the Maraou6 and on most rivers in Ivory Coast the most favourable periods are when the water level is falling and at low water (in the case of permanent rivers). The hydrological season appears to be more important than the calendar season, to ensure hooogeneity of results between rivers whose variations in discharge are noE synchronous- Iitrile determination of the short-term effects of temephos calls for detailed studied involving the identification of the species or larval atage' long-term monitoring of the impact of the insecticide on coenotic structures should be carried out on two levels: - relatively superficial surveillance carried out over an extensive geographical area by non-speciaLised personnell - much more detailed surveiLlance, limited to one or trro reaches, which will contribute to the interpretation of results obtained from general surveillance. In the first case, counting and taxonomic identification procedures should be as sirnple as possible. Since biocoenetic modifications caused by temephos can be discerned when identification is carried out at supra-sPecies leveI, we feel that the monitoring protocol does not need changing on this point. The adoption of a logarithmic counting scale such as the one proposed in section IV would also save a 1ot of time, for Ehis method does not bias Ehe resuLts. The time saved can be used to increase the ntrmber of samplings and hence the sEatistical reliability of the results. 4. Pollution Indices 4.1 Biotic index The biocoenotic index of pollution as defined by us makes it possible to quantify structural changes. Calculation of this index, based on the density of the various taxa, is rapid and requires no special mathematical ability. It is necessary, however, to know the C coefficient and how the different taxa are to be weighted, so it cannot be used during the first year of insecticide treatment. To overcome these disadvantages it would be useful to find a method allowing rapid estimation of these parameters. The solution to the problem could be the calculation of the LC56 or LCgg of the major species. We have in fact noted that the scale of susceptibility of species in the saxicolous entomocoenoses obtained by factorial corresponding analysis of corresponding features lres similar to that obtained by means of multiple gutter tests. If this observation can be confirmed, multiple gutter studies on the short-term susceptibility of species or taxonomic groups in a given river or river basin should make it possible for the pollution index to be applied from the start of a treatment campaign, the coefficient being the values obtained, after appropriate Processing, from the LC5O or LC9g. The accuracy of the weights and coefficients will be verified only Iater, once enough post-treaEment data have been collected and processed, together with pre-treatment data, using factorial correspondence analysis of corresponding features. ocP/vcu/HYBto/ 84 .2 Page 11 4.2 Shannon index Without wishing to appear dogmatic as to the use of biotic and biocoenotic indices, we should Like to stress once again that the Shannon index is not appropriate for assessing the pollution caused by temephos, at least when it is calculated globally for an entomocoenosis and not a taxocoenosis. This assertion does not call into question the heuristic and intrinsic value of the index. Studies on the short-term effects of temephos carried out in gutters have shown that the index does not vary when the impact of the insecticide is slight or nonexistentl that it falls along with the quantity of fauna present when the concentration of temephos is rnoderate; and that urith high concentrations the index faLls initially and then rises. These contradictory variations are due first to the reduction in the quantity of fauna and then to a reduction in specific density. In the long-term, Shannon index values calculated for saxicolous fauna do not make it possible to distinguish betlreen samples which have been subjected to temephos treatment and Ehose which have not. This lack of variation in the index can be attributed to the fact that while some species have been eLiminated or have spffered a reduction in numbers, others have appeared orproliferated, compensating arithmetically for Ehe lossses caused by temephos. In studies on the distribution of fauna on floating substrates, on the other hand, the Shannon index makes it possible to monitor the depletion of conununities as the rate of flow increases. However, the same study carried out when the river was treated with temephos shows that species diversity resembles that obtained in the first study only in moderate currents; it is higher where the current is rapid and lower wtere it is slow. Ttrese variations depend largely on the density of a dominant species. Ttre sum of the results obtained with the Shannon index gives rise to the following observations : - The fact that differing values are obtained with the Shannon index shows that one or more parameters have modified the structure of the entomocoenoses; and temephos is not necessarily the active parameter. - Temephos pollution may cause a rise or a fall in species diversity indifferent cases. These variations often depend on the susceptibiLity of the numerically dominant species - Lastly, the fact that the Shannon index gives identical values before and after temephos treatment does not prove that there has been no change in the structures of the entomocoenoses. It must therefore be concluded that the index cannot be used forquantification of the pollution caused by temephos in the short-, medium- or long-term. 4.3 Drift index ratios This paper has not mentioned monitoring of the long-term effects of temephos using the drift, estimated by the day drift index (Opt) ana the nightdrift index (NDr). This is because these parameters have revealed no effect of temephos, even after processing the data by factorial correspondence analysis of corresponding features, principal components or hierarchical classification in ascending order. This failure seems to be due to the Large number of biotic ocP/vcu/HYBro/ 84.2 Page L2 and abiotic factors that influence the drift. The value of studying the drift as a means of assessing the effect of temephos could be ascertained by systematic collection of day and night drift on one or two stretches the day before treatment and on the treatment day, according to the protocol used in the monitoring progranrne. If, as we have shown, there is no significant different between the two sets of data, there is little chance that the action of temephos can be revealed by the drift. We have not, however, applied multifactor analysis to the devel,opment of the relationship between the day and night drift indexes (DDI and I{DI). It is possible that the impact of temephos could be quantified in this way. If temephos produces morbidity in surviving populations, and if it teduces numbers and modifies the nyctohemeral pattern of drift, it is probable that an abnormal increase in the day drift index would occur, thus modifying the DDI/NDI ratio. As far as the monitoring programne is concerned, these calculations should be made by WIIO through OCP. It is essential that sampling of the drift in a long-term monitoring programme and study of variations in the DDI/NDI ratio should take account of the following conditions (see Annex II: Drift ) : - Collection should be carried out a fixed number of days after the last spraying, which rras not the case in the monitoring progratrune and which calls for coordination between research teams and vector control teams. In the case of weekly treatment, a lapse of 5-5 days after spraying is desirable, to eliminate all extraneous factors such as delayed morbidity and the effects of rraves of insecticide arriving from the treatment of breeding sites upstream. - Calculation of the DDI/I{DI ratio should be carried out taxon by taxon and not on the total fauna collected. Selection of classes according to size or larval stage is also recormnended, to buffer non-significant variations in density due to maas eoergence or eclosion. 5. Special Surveill,ance Long-term surveillance cannot be carried out without more detailed studies, for the correct interpretation of results ca11s for some knowledge of the bioecology of the entomic species, and any replacement of species within a taxonomic group must be recognized. We therefore recorrnend that once or twice a year, on a fixed data, a qualitative survey be made of the species present on the principal substrates colonized by the lotic entomocoenosis. It is also essential Lhat on one or trdo reaches of the river net\rork under surveillance a quantitative study of saxicolous entomocoenoses be undertaken down to the species leve1, including preirnaginal stages. On those reaches detailed taxonomic identification should be accompanied by an assessment of the infLuence of abiotic parameters on variations in the faunal and structural composition of the entomocoenoses. Special aEtention should be paid to the action of the current. Since it is not possible to modify the distribution of rocky substrates in l-otic reaches and in view of the difficulties involved in collecting samples from them, we recortrnend the use of artificiaL substrates for this work. A yearly study of insect distribution would be desirable, carried out at a fixed hydrological season. Artificial substrates of the brush type should be used, ri.r"" they are easy to handle and suitable for the collection of the majority ocP/vcu/HYBro/ 84.2Page 13 of lotic fauna. V. CONCLUSIONS The massive use of insecticides, whether for health reasons (control ofdiseases spread by arthropods) or for plant protection (control of crop pests)is widespread throughout the world and is generally effective. The practice can, hovever, prove very dangerous in the long-term because of the app""."a""of resistance in some target species, the often unknown ecological fate ofthese substances and their degraded residues, and the breaks caused in foodchains. The use of pesticides over long periods is therefore open tocriticism from the ecological point of view. Before endorsing this opinion it would be useful to review the results recorded for non-target fauna in the basic prografine. [rle recall that the aimof temephos spraying operations is to prevent the spread of onchocerciasis. rn the absence of other methods of controlling this endemic disease(prophylaxis or chemotherapy), the treatment of rivers with larvicide is the only practicable means now available for carrying out a mass campaign. According to the results obtained, temephos can be described as onlyr:moderately" or tttolerablyt'toxic for the insect fauna of lotic ecosystems end non-toxic for fish. It is thus difficult to weigh the rarefaction oi ,or" aquatic species in the balance against the health of millions of people,particularly in view of the fact that in the OCP area most small tributary streams are left untreated, thus constituting a reservoir for the threatened species and creating the possibility of recolonization of the major watercoursee treated with temephos once apraying operations have ceased. Moreover, there hae been a considerable reduction in treatment over thelast fen years in the central part of the Ocp area. systematically weekly treatment has been replaced by ilopportunistictr treatment carried out only where S. damnosum s.1. are present. Ttris lightening of insecticide pressure ontheffimostfavourab1etothemainienanceofnon.targetfauna. In view of its lack of toxicity for marmnals and its low level of residue, Ire are far from condemning the use of temephosl rre recortrnend it as long as noinsecticide equally effective against S. damrosum s.1. but less harmful Eo non-target fauna has become operationaF-- At present, Bacillus thufingiensis serotype H-14 seems the most likelyinsecticidetor@ecol.ogica1effectsofB.t.H.14wi1l. probably be very different from those brought out in this st@ since in thelotic environment it is toxic only for all Simuliidae. Other shifts in the coenotic baLance are likely to appear; we hope thet the monitoring rnethods reco"rnended here will permit their early detection. ocP/vcu/HYBrol 84 .2 Page L4 ANMX I: ARTIFICIAL SUBS?RATES A new type of artificial floating substrate, knolm ag a trbroomtt, has been developed to meet the ecological requirements of preimaginal stages of s. damnosum s.1. rt consists of a bunch of 15 plastic fibresr T cm long and;iEhffiavy, attached at one end by a rubber iand. Each fibre is about l runin diameter. The bunch of fibres is hooked on to a very flexible wire whichis in turn attached to a cable fixed 30 or 40 cm above rreter level, at right angles to the direction of the current. These substrates float on natural floating objects such asperifolia or the leaves and AII{S Irlithout temephos the surface of the lrater in the as leaves from trees, the aerial stalks of Kahania and Chloris. same lray as roots of Ficus - Study of the distribution of organisms on floating substrates - Couposition of colonies - Study of rheophilous valences and preferred rates of flow - Study of species competition. With temephos - Assessment of quantitative variations brought ebout by ternephos - Detection of variations in rheological niches - Observation of structural changes. TYPE OF SUBSTMTE: Broonrtype artificiaL substrate. NT'I.{BER OF SUBSTRATES: Tvo or three per unit of flow rate measured in 10 m/s segments The speed of the current must therefore be measured. It is advisableto install more than the reconunended number in case of accident. PERIOD OF COLONIZATION: Fifteen days. If the experiment takes place during a treatment period, collect the substrates the day before spraying. IDEMIFICATION OF ORGANISMS: Larval stage for the more plentiful organisms; identification at the species level for others. ocP lvcu|HYBT}/84.2Page 15 ANMX II: DRIFT It has long been known and demonstrated that flowing streams aLways contain a quantity of drifting organisms which have either become detached from their substrates or are actively migrating. The quantification of this phenomenon has made it possible to deEermine nyctohemeraL and seasons patterns and to detect isolated cases of pollution. Moreover, hydrobiologists are stil1 nursing the hope that a mathematicaL relationship, complex or otherwise, will be found between rates of drift andthe quantity of fauna present. This hope is strengthened by the fact that itis relatively easy to take samples of the drift; interpretation of theresults, however, is rmrch more tricky than appears at first. Drifting invertebrates are collected by means of one or more nets mounted on metal frames. Ttre mesh size of the sieves is about 250u. Each net isfitted with a collector with a removable filter top. At high water the nets are attached to floats and weights to keep them facing vertically into the current. I. SHORT-TERM STUDIES AIMS: - Detection of short-Eerm toxicity of an insecticide - Study of kinetics of detachraent. }IETS: One net 2 m long with a Z0 x 20 cm aperture. SAI.{PLII{G PERIOD: Three minures. HYDROLOGICAL PARAMETERS TO BE TAKEN IMTO ACCOUM: - Rate of flow - Hydrological season - Variations in discharge of the river and rainfall. PROTOCOL OF EXPERIMEM: - During the 2 hours preceding spraying, one sample should 6e taken everyhalf-hour, i.e. 4 samples. - Spraying around noon, to avoid the intensive morning and evening drift associated with behavioural petterns. - During the 4 hours following spraying, one sample should be taken everyhalf-hour, i.e. 8 samples. LEVEL oF TAxoNoMrc rDENTrFrcATroN: species (and 1arval stages). EXPRESSION OF RESULTS: - Drift index - Increase-in-drift ratio. ocP/vcu/HYBro I 84 .2Page 16 II. LONG.TERM STUDIES AIMS: - Detection of any modifications in patterns. - Numerical changes: variations in the day and night drift indices(DOt and DNI). During periods of treatment with temephos, an increase in the day drift index can be expected owing to morbidity in some parts of the populations weakened by the insecticide and to a reduction in the quantity of fauna in situ. - Ethological changes: these can be reveaLed by modified patterns, i.e.: a shift in drift acrophases lack of variation (DDI = NDI?) - Modifications in the structure of colonies. NETS: One triple net, each unit 2 m l-ong with an aperture of 20 x 20 cm. SAMPLING PERIODS: - Two 15 minute sampling periods, L ll2 hours before sunset, for the day drift (i.e., 6 samples). - Two 3,-minute sampling periods, t hour after sunset, for the night drift (i.e., 6 samples). PARAMETERS TO BE TAIGN INTO ACCOUM: - Speed of current - Ilydrologica1 season - Previous fluctuations in the discharge of the river - Recent rainfall - Discharge measured on a river gauge. LEVEL OF TAXONOMIC IDEMITICATION: - For general surveillance: taxonomic level as defined in OCP monitoring prografitre. - At two fixed points: species level, but only certain species will be monitored, selected according to criteria of density and susceptibility to the insecticide. Moreover, only certain stages or sizes will be counted, in order to avoid numerical variations due to mass eclosion or emergence. IIYDROLOGICAL SEASON: Months when the rilater level- is falling and at low water. EXPRESSION OF RESI'LTS: - Ntrmeration according to a logarithmic scale would be preferable. - Day and night drift indices (DDI and I'IDI). - Drift index ratio = NDI/DDI. &PlVCUIEYBI.O/84.2Page L7 ANMX III: SURBER SAI'{PLER This device, which nay be modified in various lrays, is.the one most often used in hydrobiology for sanpling river fauna. It is designed for the study of rock (saxieolous) fauna, and can collect in its net the individuals present on a surface area of 15 x 15 cm. Ttre organiems are detached from the eubstrate by means of a brush and pase into a collector of the usual type with a funnel and a sieve. The advantages and disadvantages of the method need nofurther discussion. In interpreting the results, however, the following points should be bornein mind: - Ttris method of sampling is 4ot practicable at high lrater wtren the rocky substratee are inaccessible, being eubmerged under large maaaea of water. - The device can onl.y be used on relatively flat rocks, and is not suitable for sanpling fauna on laterite substretes. - Becauge of the aggregetive distribution of aquatic invertebratee, five sanpling oPerations are required to obtain a representative aample of the fauna in the rrrockg in running waterrt biotope. This condition ie not always uet ot ing to the difficulty of finding enough rocke in sinilar rheological eituationa. AIHS: - Deteroination of aseociations among saxicolous ineecte. - Lon5term uonitoring. - Demon8tration of structural. nodificatione in saxicolous entonocoenoaea. SIZE Of SITRFACE SAI'iPLES: Ttre surface area of 15 x 15 cm (usedlarge. We recomend an area of 10 x cm. by 15 OCP) is too or10x10 SIZE OF SAMPLE: 10 sampling procedures carried out in homogeneous conditions. PARAUETERS TO BB TAIGN INIO ACCOI]M: - Speed of current - Extent to which rocks are covered with rristicha trifarii. LEVEL OF TAXONOI.GC IDENTIFICATION.. General surveillance: taxa (family, order, tribe) as monitoring programe. Special gitea: two points with identification ar the and larval stage. defined in the oCP level of epeciee ocP/vcu/EYBro/u.2 Page 18 ANNEX IVr GUTTERS The multiple gutter device recomrended for testing different insecticidee or different formulations consists of five plastic (pVC) gutter units mounted on a suPPort of adjustable height. The upstream end of the gutter consists of a funnel fitted with a nylon filter, mesh size 250,/1, which allows water toflow into the gutter wtrile retaining the organisuJ drifting in the river. The downstream part of the trough is fitted with a collecting net ending in a funnel with a filter which also has a 25lp mesh. fire body of each trough is 1.5 m long and 9 cm in diameter. Ttre isolation of the organisms inside the trough unit gives a knownquantity of fauna, so that the natural drift rates and those due to pesticides can be calculated. Wtrere several concentrations of pesticides are teated, the LC5g and LC96 can be calculated. AIMS: - Short-term impact. - Kinetics of detachment. - Determination of LC5g or tCgg of species or preimaginal stages of lotic insects, - Establishment of a scale of susceptibility of the various species so that they can be classified in order of vulnerability. - Monitoring of developoents in susceptibility of the species at different seaaons and sites, and detection of any resistance phenomena. - Couparison of several ineecticides. EQUIPUEM: Multiple gutters. NTUBER OF CONCENTRATIONS: At leest 5 + 1 control. DUMTION OF SPRAYING: 10 minutes. PERIOD: I{hen water level is falling and/or at its lowest. NTMBER OF DAYS IN PLACE BEFORE EXPERII'IENTI ?-5 days. DURATION OF EXPERIMEM: - 2 hours before spraying - 4 hours thereafter. FREQUENCY OF READINGS: Every 15 minutes. SPRAYING TIME: Around noon, to avoid drifr due to morning and evening activity. LEVEL OF TAXONOMIC IDENTIFICATION: Species and larval stages. PROTOCOL OF EXPERIMEM: t - 3 h:, placing of filter in upstream end of gutter. t - 2 L/Zhz inecallation of net at downstream end of gutter. t - 2hr collection of first pretreatment drift. t = 12h: treatment of gutter. t + 4 h3 end of experiment. PARAMETERS TO BE TAKEN IMO ACCOUNT: - Turbidity of the rrater (measured with a Secchi disc). - Rete of flow in the gutters.