Report of fish populations in rivers of Ivory Coast and Ghana: February, 1982

PART I: ?ART I1: PART IIl: RXPORT OF FISH POPIILATIONS IN RIVERS OF IVORY COAST AND GHANA Fish conrnrtnities of lvory Crrst rivers treated by temephos(C. L6v0que, D. Paugy & J. -l'1. Jestin) Observart-ons on fish populaiions j-n Abate-treated rivers in liorthern Ghana (E. K. Abban, C. P. Fairhurst & M. S. Cirrtis) Fish monitorinl', in West African rive'rs - problr'm and perspectivcs(8. K. Abban, C. P. l'airhurst, C. L6v€que & D. Paugy) February 1982 PART I: ?ART I1: PART IIl: RXPORT OF FISH POPIILATIONS IN RIVERS OF IVORY COAST AND GHANA Fish conrnrtnities of lvory Crrst rivers treated by temephos(C. L6v0que, D. Paugy & J. -l'1. Jestin) Observart-ons on fish populaiions j-n Abate-treated rivers in liorthern Ghana (E. K. Abban, C. P. Fairhurst & M. S. Cirrtis) Fish monitorinl', in West African rive'rs - problr'm and perspectivcs(8. K. Abban, C. P. l'airhurst, C. L6v€que & D. Paugy) February 1982 FART 1: EISH COMMUNITIES OI. IVORY COAST RMRS TREATED BY TB,IEPHOS CONTENTS Page I.NTRODUCTION 1 1. ICHTIIyOLocICAL MONITORING PROGRAI.{ME - CONDITIONS AND AIMS . I 2. CHANGES IN TITE FISH POPULATIONS . 5 2.1 Experlmental fishing by gi11 neE 2.L.L Proceeslng of experlmental fishing data . 2.7.2 Como6-L6raba Basin 2.1.3 Sassandra Basln . . 5 5 2.1.4 Bandama Basin 2.2 ELectrofishing 2.3 Conclusions 3. BIOLOGY OF TIIE },IAIN SPECIES 3.1 Reproducttve cycle . . 3.2 Coeffictent of condition . 3.3 Diet . 4 " coNclusroNs 5. BIBLIOGMPIIY L4 18 19 19 19 2l 22 25 28 ; * FART 1: EISH COMMUNITIES OI. IVORY COAST RMRS TREATED BY TB,IEPHOS CONTENTS Page I.NTRODUCTION 1 1. ICHTIIyOLocICAL MONITORING PROGRAI.{ME - CONDITIONS AND AIMS . I 2. CHANGES IN TITE FISH POPULATIONS . 5 2.1 Experlmental fishing by gi11 neE 2.L.L Proceeslng of experlmental fishing data . 2.7.2 Como6-L6raba Basin 2.1.3 Sassandra Basln . . 5 5 2.1.4 Bandama Basin 2.2 ELectrofishing 2.3 Conclusions 3. BIOLOGY OF TIIE },IAIN SPECIES 3.1 Reproducttve cycle . . 3.2 Coeffictent of condition . 3.3 Diet . 4 " coNclusroNs 5. BIBLIOGMPIIY L4 18 19 19 19 2l 22 25 28 ; * t ? - 1- INTRODUCTION An extensive Onchocerciasis Control P,i:ogranrne in West Africa was larrnched in 1974. under the aegis of lJlIO. Since th;rt rime there have been weekly applicates of insectic icles in all rivers over an area of more than two million lcnz covering llpper Volra, Tvory Coast, Ghana, the north of Benin, Niger and Mali; it is planned to continue trcatmerrt f.or 2o years Temephos, which was the only insecticide used until L98o, was adopted after a number oI tests, because of its effectiveness against the larvae of Simulium clamnosr.ro (the vector of onchocerciasis), which inhabit running water, and its apparenE slight toxicity for the non-target fauna. Nevertheless, since its prolonged and intensive use could have risks, it wa: necessary to evaluate the possible medium-term and loug-tcnn effects of applicates on the organisms of the treated watercourses. Consequently, an aquatic environmental monitoring Prograflnme was PrePared at the conrnencement of the larval eradication progrartrne, so as to be sure that the insecticide released did not excessively disrurb the functioning of the treate environments, and to provide warning to those carrying out treatment, should toxic effectsbe noted. Llhat has been done has been primarily concerned with two major categories of organisms: the invertebrates that abound in the watercourses and that are directly threatened by theinsecticide in the same way as Simulium damnosum larvae; fishes, by virtue of their economi interest for the people living along the rivers. There has also been other research of shorter duration on r^rater quality, plankton, etc. The monitoring progranmne has been jointly carried out by s,,veral teams working indifferent zones but following the same established proce,{ures (L6v6que et al., lgTg). Ihe results on ichthyology obtained in Ivory Coast by tlrr: ORSTOM Hydrobiological Laboratory at Bouak6 are set out here. 1. ICHTHYOLOGICAL MONITORING PROGRAT'IME . CONDITIONS AND AIMS Given that preliminary trials had shown that temephos apparently had no direct effect ofishes when employed at the reconrnended treatment rates, the monitorlng progra[Ere was set uplrith the aim of studying the possible medium-term and long-term effects in the naturaL environment. The terms of reference also included the introduction of simple technlques capable of being used by the teams responsible for monitoring in the various zones of theprogramne' and under envlronmental conditions that are sometimes dissimilar. In this context' al:d hav'ng regard to the difficulties of access to and work in many watercourses, as well as to the extent of the habitats for investigation, we therefore choseto give preferential treatment to a number of topics: - study of changes in the catch composition of experimental fishing carrled out at regular intervals with a standardized set of gi11 nets at selected monitoring statlon on treated and untreated rivers; - study of the developtnent of some of the biological parameters, more especially the coefficient of condition, which is a measure of the health of fiehes. This paramete may consequently be used to assess whether flshes are still abLe to, find the food thathey require ln the treated environments, and whether ecological conditions remainfavourable to them. It also reflects any possible adverse effect of Abate on the rnetabolism of fishes. Refore we undertook this prograflrne, our knowledge of the West African rivers was verylimited. In Particular, there were very few detailed studies of the biology and ecology of their fishes. It therefore very rapidly became apparent that additional research was essential if the resul-ts of the monitoring programne were to be correctly interpreted in the absence of adequate knowledge of the environment. Various studies were therefore carried out Eo provide a better knowledge of the biology of the main species: Alestes baremoze(Paugy, 1978); AlesEes nurse (paugy, 1980), Alestes imberi (paugy, 19go), A1""lgg- macrolepidoLus (Paugy, L982), Alestes longipinnis (Paugy, L982), t:r"". lgr .i Y t ? - 1- INTRODUCTION An extensive Onchocerciasis Control P,i:ogranrne in West Africa was larrnched in 1974. under the aegis of lJlIO. Since th;rt rime there have been weekly applicates of insectic icles in all rivers over an area of more than two million lcnz covering llpper Volra, Tvory Coast, Ghana, the north of Benin, Niger and Mali; it is planned to continue trcatmerrt f.or 2o years Temephos, which was the only insecticide used until L98o, was adopted after a number oI tests, because of its effectiveness against the larvae of Simulium clamnosr.ro (the vector of onchocerciasis), which inhabit running water, and its apparenE slight toxicity for the non-target fauna. Nevertheless, since its prolonged and intensive use could have risks, it wa: necessary to evaluate the possible medium-term and loug-tcnn effects of applicates on the organisms of the treated watercourses. Consequently, an aquatic environmental monitoring Prograflnme was PrePared at the conrnencement of the larval eradication progrartrne, so as to be sure that the insecticide released did not excessively disrurb the functioning of the treate environments, and to provide warning to those carrying out treatment, should toxic effectsbe noted. Llhat has been done has been primarily concerned with two major categories of organisms: the invertebrates that abound in the watercourses and that are directly threatened by theinsecticide in the same way as Simulium damnosum larvae; fishes, by virtue of their economi interest for the people living along the rivers. There has also been other research of shorter duration on r^rater quality, plankton, etc. The monitoring progranmne has been jointly carried out by s,,veral teams working indifferent zones but following the same established proce,{ures (L6v6que et al., lgTg). Ihe results on ichthyology obtained in Ivory Coast by tlrr: ORSTOM Hydrobiological Laboratory at Bouak6 are set out here. 1. ICHTHYOLOGICAL MONITORING PROGRAT'IME . CONDITIONS AND AIMS Given that preliminary trials had shown that temephos apparently had no direct effect ofishes when employed at the reconrnended treatment rates, the monitorlng progra[Ere was set uplrith the aim of studying the possible medium-term and long-term effects in the naturaL environment. The terms of reference also included the introduction of simple technlques capable of being used by the teams responsible for monitoring in the various zones of theprogramne' and under envlronmental conditions that are sometimes dissimilar. In this context' al:d hav'ng regard to the difficulties of access to and work in many watercourses, as well as to the extent of the habitats for investigation, we therefore choseto give preferential treatment to a number of topics: - study of changes in the catch composition of experimental fishing carrled out at regular intervals with a standardized set of gi11 nets at selected monitoring statlon on treated and untreated rivers; - study of the developtnent of some of the biological parameters, more especially the coefficient of condition, which is a measure of the health of fiehes. This paramete may consequently be used to assess whether flshes are still abLe to, find the food thathey require ln the treated environments, and whether ecological conditions remainfavourable to them. It also reflects any possible adverse effect of Abate on the rnetabolism of fishes. Refore we undertook this prograflrne, our knowledge of the West African rivers was verylimited. In Particular, there were very few detailed studies of the biology and ecology of their fishes. It therefore very rapidly became apparent that additional research was essential if the resul-ts of the monitoring programne were to be correctly interpreted in the absence of adequate knowledge of the environment. Various studies were therefore carried out Eo provide a better knowledge of the biology of the main species: Alestes baremoze(Paugy, 1978); AlesEes nurse (paugy, 1980), Alestes imberi (paugy, 19go), A1""lgg- macrolepidoLus (Paugy, L982), Alestes longipinnis (Paugy, L982), t:r"". lgr .i Y -2- (de Merona, 1980), SchiLbe nrvstus (L6v6que & Herblnet, 1980), Eutropius mentali9. (L6v6que & Herbinet, L9B2). An ichthyological study of the Bandama Basin (de Merona, 1980) has provided up-to-date informatlon on t.he ecology of rhe fish species and a check on the rePresentative nature of the monitoring stations. Electrofishing Las also been carried out in the rapids to give us a better understanding of the flsh populatlons of these blotopes, whlch cannot be .samrpled by gi1l net, and of their changes over time. Lastly, stornach-content studies have established that the dlet of the main species has -nJt undergone significant changes (Vidy, 1976) and fecundlty studies have demonstrated Ehat tdmephos dou" ,,oi ap,pear adversely to have affecEed reproduction (Albaret, 1982). l,lonitoring technique Observations were carried out at eight stations Located in several drainage basins(rie. 1): - Bandama Basin: the Marabadiassa and Niakaramandougou stations on the Bandama itself, the Dabakala station on the N'zi, and the Mankono station on the Maraou6. Except for l1ankono, for which records date only from 1977, all these stations have been investigated since 1974; - Cmro6 Basin: Gans6 station on the Corno6 and Pont FrontiEre station on the L6raba, record's for both of which date frour L974; - Sassandra Basin: Semien station, for which records date from 1975, treaEment starting in 1978; - Bago6 Basin (a tributary of the Niger): Kouto sEation, for which records exist from 1975, treated since 1977. In general, the monltoring statlons are located on rivers of tropical type having their hlghest water 1eve1 in September and a lengthy low-water period from January Eo June (Fig. 2). There is a special account of the hydrological and physico-chemical characteristics of these \ratercourses (I1tis & L6v6que, 1982). Experimental fishing was carried out using a set of gi11 nets 25 m long and 2 m deep, with various mesh sizes: 15, 20, 25, 30 and 40 nrn for the monitoring set. We made additional use of 12.5, 17.5 and 22.5 rmn meshes, which will be taken into consideration in the results here set out. Ihese nets, which were of multifilament nylon (12 5OO), rrere set uP at 5O%. The monitoring stations were sampled every three months at the start of the prograrme. Having regard to the slight apparent impact of the insecticides, it was decided from 1979 onwards to make only two annual observations, one at the end of the high-water period, the ot-her at the end of the low-water period. -2- (de Merona, 1980), SchiLbe nrvstus (L6v6que & Herblnet, 1980), Eutropius mentali9. (L6v6que & Herbinet, L9B2). An ichthyological study of the Bandama Basin (de Merona, 1980) has provided up-to-date informatlon on t.he ecology of rhe fish species and a check on the rePresentative nature of the monitoring stations. Electrofishing Las also been carried out in the rapids to give us a better understanding of the flsh populatlons of these blotopes, whlch cannot be .samrpled by gi1l net, and of their changes over time. Lastly, stornach-content studies have established that the dlet of the main species has -nJt undergone significant changes (Vidy, 1976) and fecundlty studies have demonstrated Ehat tdmephos dou" ,,oi ap,pear adversely to have affecEed reproduction (Albaret, 1982). l,lonitoring technique Observations were carried out at eight stations Located in several drainage basins(rie. 1): - Bandama Basin: the Marabadiassa and Niakaramandougou stations on the Bandama itself, the Dabakala station on the N'zi, and the Mankono station on the Maraou6. Except for l1ankono, for which records date only from 1977, all these stations have been investigated since 1974; - Cmro6 Basin: Gans6 station on the Corno6 and Pont FrontiEre station on the L6raba, record's for both of which date frour L974; - Sassandra Basin: Semien station, for which records date from 1975, treaEment starting in 1978; - Bago6 Basin (a tributary of the Niger): Kouto sEation, for which records exist from 1975, treated since 1977. In general, the monltoring statlons are located on rivers of tropical type having their hlghest water 1eve1 in September and a lengthy low-water period from January Eo June (Fig. 2). There is a special account of the hydrological and physico-chemical characteristics of these \ratercourses (I1tis & L6v6que, 1982). Experimental fishing was carried out using a set of gi11 nets 25 m long and 2 m deep, with various mesh sizes: 15, 20, 25, 30 and 40 nrn for the monitoring set. We made additional use of 12.5, 17.5 and 22.5 rmn meshes, which will be taken into consideration in the results here set out. Ihese nets, which were of multifilament nylon (12 5OO), rrere set uP at 5O%. The monitoring stations were sampled every three months at the start of the prograrme. Having regard to the slight apparent impact of the insecticides, it was decided from 1979 onwards to make only two annual observations, one at the end of the high-water period, the ot-her at the end of the low-water period. a( FIG. 1. Semien V,r,^ -3- LOCATION OT MONITORING STATIONS IN IVORY COAST Pont frontidre Kouto Fe rkev){.. .. Sequeke oun d ta lt u Si rasso coe Sepi kaha amandougou Abidl an [\ + F. Taou a ra Ganse a( FIG. 1. Semien V,r,^ -3- LOCATION OT MONITORING STATIONS IN IVORY COAST Pont frontidre Kouto Fe rkev){.. .. Sequeke oun d ta lt u Si rasso coe Sepi kaha amandougou Abidl an [\ + F. Taou a ra Ganse FTG. 2. AT TIIE -4- RI\IER DISCI]ARGES (MEAN MONTIILY VALUES) , 1974-7978, STATIONS OF TlrE MoNITORING PROCRAIIME (DATA SUppLIEt) BY TTIE H)DROIOGICAL SERVICE -OF IVORY COAST) 50 30 10 20 10 N'Zi-Dimbokr 1974 1975 1976 1977 80 40 20 10 1978 B andama-lvlara badias s a Sassandra- Bago6-Kouto L6raba-Pont ironti6re Maraoue-Bouafle FTG. 2. AT TIIE -4- RI\IER DISCI]ARGES (MEAN MONTIILY VALUES) , 1974-7978, STATIONS OF TlrE MoNITORING PROCRAIIME (DATA SUppLIEt) BY TTIE H)DROIOGICAL SERVICE -OF IVORY COAST) 50 30 10 20 10 N'Zi-Dimbokr 1974 1975 1976 1977 80 40 20 10 1978 B andama-lvlara badias s a Sassandra- Bago6-Kouto L6raba-Pont ironti6re Maraoue-Bouafle -5 - 2. CHANGES IN THE FISH POPULATION 2.1 Experimental fishing by gil1 net 2.1.7 Processing of experimental fishing data The experimental fishing results are expressed as the catch per unit of effort (c.p.u.e. the number or weight of flsh caught per lOO m2 of gil1 nets in the course of f.ishing for one night). In the processing of these data, we made use of correspondence analysis, the salient points of which have been amply described elsewhere (Benzecri, 1973; Hi11, L974; Lebart & Fenelon, L973). Ihis method is especially suitable for our type of data. One of its advantages is to permit simultaneous graphic represcntation of species and observations. To facllitate the proces.si.ng, the various types of nets employed lrere grouped into three categories: category 1 - nets with a mesh size of 12.5 and 15 nrr; category 2 - nets with a mesh size of 17.5/2O and 22.5 m; category 3 - nets with a mesh size of 25, 3O and 4C The mean catch per unit of effort (c.p.u.e.) was calculated for each of these categories, and for all the gilt nets ln Eerxrs of the number of fish and the weight. I'hese amalgamatlons into categorles were made after a preliminarv study had shown that there were strong affinities beEween the fish populatlons sampled by certain gill nets. The processing was carried out on raw data (c.p.u.e. expressed as numbers or as weights) and on data coded by logarithmic transformation to base 5 for numbers and to base 10 for weights. Correspondence analysi" is frequently employed in ecology for the analysis of raw data. It gives preferenEial treatment to highly variable abundant elements, and consequently often brings out Phenomena which sr,em obvious to the field ecologist. Coding, on the contrary,favours the characteristic species and observations, and here the analysis yields results somewhat dlfferent from.the previous ones, bringing out dJfferent phenornena. The extreme case of coding is presence-absence. At the present time the principle experience of many sirecial-ists in data biologists concerning the si;1ni. .gansg with species of very different sizes. sorne biologlsts to be better. of coding has only an empirical processing. It arose from the of raw data, especially data on The biological significance of basis borne out by the questions raised by numbers, when dealing weight data appears to We carried out our processing on ra!, results (numbers and weights) and also on codeddata. Only soure of the results of these analyses are given here, to the extent to which they seem to us to make a valuable ,.'ontribution to the interpretation of the results. 2.L.2 Couro6-L6raba Basin Two statlons were regr:larly sampled in this Basin: Gans6 on the Como6 and the station known as Pont FrontiEre on the L6raba. If we consider the numbers of the species caught, it may be noted that the fish populations of the two stations are qualitatively quite similar,but differ in the relative abundance of some species: S. mystus is more abundant in the I.6raba than at the station on the Como6, as are P. guentheri, H, fasc-latue,and 8.. een_eg,a1en,sito a lesser degree. Conversely, the Morrnyridae (P. bovei in particular, M. furcidens, -tl. rtrme) and L. coubie are better represented at Gans6. These characteristlcs are tllustratin Fig- 3, in which P. bovei and S. mystus make a stronE! contribution to the first axisdividing the observations at the two stations. T?rere is also a dilference in the corposition of the observations of the high-waterperiod and the low-water period at each of the two stations (Fig. 4). There are trro essential causes for this phenomenon: sampling is not carried out under precisely the same condjtions in the course of these periods, and some species migrate upstream to reproduceduring the high-water period. E. mentalis would appear to correspond to this latter case, since the abundance peaks in the caEches occurred durlng the low-water period in the Corno6 an, -5 - 2. CHANGES IN THE FISH POPULATION 2.1 Experimental fishing by gil1 net 2.1.7 Processing of experimental fishing data The experimental fishing results are expressed as the catch per unit of effort (c.p.u.e. the number or weight of flsh caught per lOO m2 of gil1 nets in the course of f.ishing for one night). In the processing of these data, we made use of correspondence analysis, the salient points of which have been amply described elsewhere (Benzecri, 1973; Hi11, L974; Lebart & Fenelon, L973). Ihis method is especially suitable for our type of data. One of its advantages is to permit simultaneous graphic represcntation of species and observations. To facllitate the proces.si.ng, the various types of nets employed lrere grouped into three categories: category 1 - nets with a mesh size of 12.5 and 15 nrr; category 2 - nets with a mesh size of 17.5/2O and 22.5 m; category 3 - nets with a mesh size of 25, 3O and 4C The mean catch per unit of effort (c.p.u.e.) was calculated for each of these categories, and for all the gilt nets ln Eerxrs of the number of fish and the weight. I'hese amalgamatlons into categorles were made after a preliminarv study had shown that there were strong affinities beEween the fish populatlons sampled by certain gill nets. The processing was carried out on raw data (c.p.u.e. expressed as numbers or as weights) and on data coded by logarithmic transformation to base 5 for numbers and to base 10 for weights. Correspondence analysi" is frequently employed in ecology for the analysis of raw data. It gives preferenEial treatment to highly variable abundant elements, and consequently often brings out Phenomena which sr,em obvious to the field ecologist. Coding, on the contrary,favours the characteristic species and observations, and here the analysis yields results somewhat dlfferent from.the previous ones, bringing out dJfferent phenornena. The extreme case of coding is presence-absence. At the present time the principle experience of many sirecial-ists in data biologists concerning the si;1ni. .gansg with species of very different sizes. sorne biologlsts to be better. of coding has only an empirical processing. It arose from the of raw data, especially data on The biological significance of basis borne out by the questions raised by numbers, when dealing weight data appears to We carried out our processing on ra!, results (numbers and weights) and also on codeddata. Only soure of the results of these analyses are given here, to the extent to which they seem to us to make a valuable ,.'ontribution to the interpretation of the results. 2.L.2 Couro6-L6raba Basin Two statlons were regr:larly sampled in this Basin: Gans6 on the Como6 and the station known as Pont FrontiEre on the L6raba. If we consider the numbers of the species caught, it may be noted that the fish populations of the two stations are qualitatively quite similar,but differ in the relative abundance of some species: S. mystus is more abundant in the I.6raba than at the station on the Como6, as are P. guentheri, H, fasc-latue,and 8.. een_eg,a1en,sito a lesser degree. Conversely, the Morrnyridae (P. bovei in particular, M. furcidens, -tl. rtrme) and L. coubie are better represented at Gans6. These characteristlcs are tllustratin Fig- 3, in which P. bovei and S. mystus make a stronE! contribution to the first axisdividing the observations at the two stations. T?rere is also a dilference in the corposition of the observations of the high-waterperiod and the low-water period at each of the two stations (Fig. 4). There are trro essential causes for this phenomenon: sampling is not carried out under precisely the same condjtions in the course of these periods, and some species migrate upstream to reproduceduring the high-water period. E. mentalis would appear to correspond to this latter case, since the abundance peaks in the caEches occurred durlng the low-water period in the Corno6 an, FIG. 3. COMOE-I.ERABA - OT FISH CAUGHT (TOTAL FOR -6 FACTOR AMLY{;IS CARRIED IHE IHREE CA]SGORIES OF OUT ON NIiMBERS NETS, RAW DATA) a5-7s ar2-7s Lr-1"7 l\t-'u a r-78 ^ 7-77 s-7s oAro-764-76A Oz-zo dLr 7s 3_76 n A s_;s@-tt O.-"i e:_;zIo-75 671 @z-z j g,.-.,u* 8;;::,af -7s Aro-74 sltY -r-79O 8-7tf,;_zz FIG. 3. COMOE-I.ERABA - OT FISH CAUGHT (TOTAL FOR -6 FACTOR AMLY{;IS CARRIED IHE IHREE CA]SGORIES OF OUT ON NIiMBERS NETS, RAW DATA) a5-7s ar2-7s Lr-1"7 l\t-'u a r-78 ^ 7-77 s-7s oAro-764-76A Oz-zo dLr 7s 3_76 n A s_;s@-tt O.-"i e:_;zIo-75 671 @z-z j g,.-.,u* 8;;::,af -7s Aro-74 sltY -r-79O 8-7tf,;_zz FIG. 4. COMOE.LERABA - FISH CAUGHT (TOTAL FOR - 7- FACTOR Ai.IALYSIS CARRTED NUT ON THE,N1'I.{BERS TIIE THREE CAIEGORIES OF NETS, CODED DATA) L€raba Como€ high-\^tater period 1ow-waterperiod high-waterperiod 1ow-water period @ a5-7) o io-76 A A o o 2 A e-7s 610-77 sco r IO-76 r&\ A rl-7-r DRO o E-78 c3 6-7: O rr-rr t17 -77 "*o-.,, o'6'i-r, -=:-- 1 aro-74 or;:,.q -'l,-" a I-78 Ai-75 o'-'c6'o2-i7 Qz-zo L-76 rl-79 A ,J-76 -A Z-\ ,.r4, l-l; Q r-78 (J Q;-zs Qr-;: LPA TZI L4-77 llPrj I.c0 FIG. 4. COMOE.LERABA - FISH CAUGHT (TOTAL FOR - 7- FACTOR Ai.IALYSIS CARRTED NUT ON THE,N1'I.{BERS TIIE THREE CAIEGORIES OF NETS, CODED DATA) L€raba Como€ high-\^tater period 1ow-waterperiod high-waterperiod 1ow-water period @ a5-7) o io-76 A A o o 2 A e-7s 610-77 sco r IO-76 r&\ A rl-7-r DRO o E-78 c3 6-7: O rr-rr t17 -77 "*o-.,, o'6'i-r, -=:-- 1 aro-74 or;:,.q -'l,-" a I-78 Ai-75 o'-'c6'o2-i7 Qz-zo L-76 rl-79 A ,J-76 -A Z-\ ,.r4, l-l; Q r-78 (J Q;-zs Qr-;: LPA TZI L4-77 llPrj I.c0 -8- during the high-water period ln the L6raba (Fig. 5). In general, no change ascribable toinsecticidal treatment can be established in the composition of catches in the como6 between1975 and 1979' The speclfic structures of the observations for the different years are infact fairly similar (Fig. 3). on the other hand, greater dlsperslon of the observatlons isto be noted in the L€raba. one of the main causes 1s the special development of S. mystuspopulations (Fig. 5). This species, wtrich was plentiful in Late L974 and early 1975, -subsequently became scarce (although there were abundancemaxima in the catches at the end oft[e hlgh-water Period), practically dlsappeared in 1978, and reappeared in 1979. It wasalso well represented it saurPles taken in 1980 (Traore, personal comnunication). A largely 'identical phenomenon was observed for s. schal_1 €ig. 5). Although the reasons for these16ng-term develop,ments escape us, they are probably partly connected with hydrology, slncehigh-water periods in which the 1eve1 ."r.iFETIy 1ow are not conducive to the reproductionof the species. The high-water periods of L976 and 1978 hrere particularly poor in theL6raba (Iltis & L6v0que, 19g2). I'he sanre conclusions as before are reached when the observations considered areexpressed as weights rather than numbers (Fig. 6). sure species are obviously caughtpreferentially during the high-water perlod @-naiicter.i), others during the low-waterperiod G-__co"!.&, C. velifer) (fig. 5). The changes in total catches reveal a seasonal cycle with a reduction in the number offish caught during the high-water period, but it is noticeable that there was no appreciable change between 1975 and 1979 in the c.p.u.e. (Figs. 7 and 8) for the staElon on rhe Como6.The late L974 to early 1975 catches in the L6raba are larger owing to the abundance ofS. nnrstus, but they subsequently remain at a roughLy equal level, although lower than thepreceding 1eve1. This decline in the catches in 1975 would not appear to be ascribable tothe action of the insecticide, but to natural variations in S. _nqrstus populations that havealso been observed in other rivers (L6v6que & Herbinet, lggo). Nor can any aPPreclable reduction in the catch by categories of net be established overthe five-year period, especially for the fine-mesh nets (category 1), vhich would appear tolndicate that the recruitment rr,as nomal and that the insecticide had no appreciable effect onfecundity and on the survival of the young fish. 2.L.3 Sassandra Basi.n The Semien station, which may be regarded as representative of the fish populations of the middle reach of the Sassandra (Paury et al., LgTg), is the only station to have been regularly fished since 1975. As treatment hras not comtrenced until June 1978, we have datafor tbree-and-a-half years prl-or to treatment. The changes ln the c.p.u.e. (Fig. 9) showthat treatment would not appear to have had any effect on the size of the catches during theperiod under investigation. The results are of the same order of rnagnitude before and after treatmerrr, although a quite appreciable reduction is to be noted for fine-mesh nets betweenlate 1976 and late 1978. Nor does analysls of the raw data on the numbers of fish caught reveal any developmentin the popuLations before and after treabnent (Fig. 10). Ttre species Alestes baremoze,Alestes imberi and Petrocephalus bovei. make a strong contribution to the firsL t o "r..",which explain 60% ot the variance. The variations in the catches of these species (Fig. 11)in fact explain the dispersion of the representative points of the observations, in accordancetoith their relative abundance. We do not know the reasons for these non-cycllc variations. ' The processing of the coded data does not add anything to the interpretation in as far as only the species caught on several occasions make a significant contribution to the axes. D -8- during the high-water period ln the L6raba (Fig. 5). In general, no change ascribable toinsecticidal treatment can be established in the composition of catches in the como6 between1975 and 1979' The speclfic structures of the observations for the different years are infact fairly similar (Fig. 3). on the other hand, greater dlsperslon of the observatlons isto be noted in the L€raba. one of the main causes 1s the special development of S. mystuspopulations (Fig. 5). This species, wtrich was plentiful in Late L974 and early 1975, -subsequently became scarce (although there were abundancemaxima in the catches at the end oft[e hlgh-water Period), practically dlsappeared in 1978, and reappeared in 1979. It wasalso well represented it saurPles taken in 1980 (Traore, personal comnunication). A largely 'identical phenomenon was observed for s. schal_1 €ig. 5). Although the reasons for these16ng-term develop,ments escape us, they are probably partly connected with hydrology, slncehigh-water periods in which the 1eve1 ."r.iFETIy 1ow are not conducive to the reproductionof the species. The high-water periods of L976 and 1978 hrere particularly poor in theL6raba (Iltis & L6v0que, 19g2). I'he sanre conclusions as before are reached when the observations considered areexpressed as weights rather than numbers (Fig. 6). sure species are obviously caughtpreferentially during the high-water perlod @-naiicter.i), others during the low-waterperiod G-__co"!.&, C. velifer) (fig. 5). The changes in total catches reveal a seasonal cycle with a reduction in the number offish caught during the high-water period, but it is noticeable that there was no appreciable change between 1975 and 1979 in the c.p.u.e. (Figs. 7 and 8) for the staElon on rhe Como6.The late L974 to early 1975 catches in the L6raba are larger owing to the abundance ofS. nnrstus, but they subsequently remain at a roughLy equal level, although lower than thepreceding 1eve1. This decline in the catches in 1975 would not appear to be ascribable tothe action of the insecticide, but to natural variations in S. _nqrstus populations that havealso been observed in other rivers (L6v6que & Herbinet, lggo). Nor can any aPPreclable reduction in the catch by categories of net be established overthe five-year period, especially for the fine-mesh nets (category 1), vhich would appear tolndicate that the recruitment rr,as nomal and that the insecticide had no appreciable effect onfecundity and on the survival of the young fish. 2.L.3 Sassandra Basi.n The Semien station, which may be regarded as representative of the fish populations of the middle reach of the Sassandra (Paury et al., LgTg), is the only station to have been regularly fished since 1975. As treatment hras not comtrenced until June 1978, we have datafor tbree-and-a-half years prl-or to treatment. The changes ln the c.p.u.e. (Fig. 9) showthat treatment would not appear to have had any effect on the size of the catches during theperiod under investigation. The results are of the same order of rnagnitude before and after treatmerrr, although a quite appreciable reduction is to be noted for fine-mesh nets betweenlate 1976 and late 1978. Nor does analysls of the raw data on the numbers of fish caught reveal any developmentin the popuLations before and after treabnent (Fig. 10). Ttre species Alestes baremoze,Alestes imberi and Petrocephalus bovei. make a strong contribution to the firsL t o "r..",which explain 60% ot the variance. The variations in the catches of these species (Fig. 11)in fact explain the dispersion of the representative points of the observations, in accordancetoith their relative abundance. We do not know the reasons for these non-cycllc variations. ' The processing of the coded data does not add anything to the interpretation in as far as only the species caught on several occasions make a significant contribution to the axes. D -9- FrG. 5. CIIANGES IN c.p.*.e. BETITIEEN 1975 AND 1979 FoR THE MAIN SPECIES TAKEN AT THE MONITORING SIATIONS ON TI.IE COMOE AND TI{E TERABA - corroE ---- LERABA S.mystus nb ndl lCC6:/.,.it E.menlalis S.sclralI 3 2 I ::l ,L Labeo spp H.f orskatii -9- FrG. 5. CIIANGES IN c.p.*.e. BETITIEEN 1975 AND 1979 FoR THE MAIN SPECIES TAKEN AT THE MONITORING SIATIONS ON TI.IE COMOE AND TI{E TERABA - corroE ---- LERABA S.mystus nb ndl lCC6:/.,.it E.menlalis S.sclralI 3 2 I ::l ,L Labeo spp H.f orskatii FIG. 6. COMOE-LERABA - oF rIsH CAUGHT (TOTAL FOR -10- FACTOR ANALYSIS CARRIED OUT ON THN ITEIGI{T THE TtlREE CATEGORIES OF NETS, RAW DATA) t\ w ir .A-t47 -".a?\ A r?-7e 4.1,-76 a 6-76 a 2-76 A,,.-rA'o-'u At--t't a s-71 PE:: E7-r; n e-77 a1-718-75 lI ro_zo A6-79 - -- 3o-z: ^ l-tt irt 5 Ef, ro-?3 r\tr a lo-77 fi i.z-zs I l)- t tU a ?-73 r-78A a1-76 - l-76ABA LJ -tr ,-; u2-i7- Tl 3-75 D r-75-fr ,.-rt f, r-;e Aa_77 Ll f-lf Aro-74 FIG. 6. COMOE-LERABA - oF rIsH CAUGHT (TOTAL FOR -10- FACTOR ANALYSIS CARRIED OUT ON THN ITEIGI{T THE TtlREE CATEGORIES OF NETS, RAW DATA) t\ w ir .A-t47 -".a?\ A r?-7e 4.1,-76 a 6-76 a 2-76 A,,.-rA'o-'u At--t't a s-71 PE:: E7-r; n e-77 a1-718-75 lI ro_zo A6-79 - -- 3o-z: ^ l-tt irt 5 Ef, ro-?3 r\tr a lo-77 fi i.z-zs I l)- t tU a ?-73 r-78A a1-76 - l-76ABA LJ -tr ,-; u2-i7- Tl 3-75 D r-75-fr ,.-rt f, r-;e Aa_77 Ll f-lf Aro-74 - 11 - rIG. 7, TltE COMOE AT GANSE - DEVELOPMENT OF CATCHES PER UNIT OF ETFORT BETI'IEEN 1975 AND 1979. , L, 2, 3 - NETS IN CATEGORIES 1, 2 AND 3. },1 - MEAN FOR THE W}iOLE SET OF NETS c.p.u.e. k4/ 1 COm2/nrqht 197 5 197 6 1 977 1 978 1979 FIG. 8. THE LERABA AT PONT FRONTIERE . CHANGES IN CATCHES PER UNIT OF EI'roRT BETI,IEEN 1975 AND 1979. 1, 2, 3 - NETS IN CATEGORIES 1, 2 AND j. M - MEA\I FOR THE ![HOI;E SET OF NETS c.p-u.e. k g/ 1 0om2/night 197 5 t97 6 1 977 197 B 1979 - 11 - rIG. 7, TltE COMOE AT GANSE - DEVELOPMENT OF CATCHES PER UNIT OF ETFORT BETI'IEEN 1975 AND 1979. , L, 2, 3 - NETS IN CATEGORIES 1, 2 AND 3. },1 - MEAN FOR THE W}iOLE SET OF NETS c.p.u.e. k4/ 1 COm2/nrqht 197 5 197 6 1 977 1 978 1979 FIG. 8. THE LERABA AT PONT FRONTIERE . CHANGES IN CATCHES PER UNIT OF EI'roRT BETI,IEEN 1975 AND 1979. 1, 2, 3 - NETS IN CATEGORIES 1, 2 AND j. M - MEA\I FOR THE ![HOI;E SET OF NETS c.p-u.e. k g/ 1 0om2/night 197 5 t97 6 1 977 197 B 1979 -L2- FIG. 9. IIiE SASSANDRA (SEI'{IEN) - CHANGES IN MEAI{ c.p.u.e. FOR TIIE wHoLE SET OF NETS EMPLOYED (M) AND By CATEGORIES OF NETS (1, 2, 3) 1 975 1 976 1 979 I'IG. 10. SE}TIEN (SASSANDM) . CORRESPONDENCE ANALYSIS CARRIED OUT ON TIIE NUI,tsERS OF FISH CAUGTTT (TOTAL FOR THE TT1REE CATEGORIES OF NETS, RAW DATA) r 1 -7s o1-7s | 2 afltr Qz-zz CVE A e'le AMA O 2-7s O 2-76 e)11-76 @ s-zo o 11-77 e B-zz ABA o 1 2-7e o 8-76 el 7-7e 6 B-7s O PBO 5-77 os-7s o 6-78 E\liE -L2- FIG. 9. IIiE SASSANDRA (SEI'{IEN) - CHANGES IN MEAI{ c.p.u.e. FOR TIIE wHoLE SET OF NETS EMPLOYED (M) AND By CATEGORIES OF NETS (1, 2, 3) 1 975 1 976 1 979 I'IG. 10. SE}TIEN (SASSANDM) . CORRESPONDENCE ANALYSIS CARRIED OUT ON TIIE NUI,tsERS OF FISH CAUGTTT (TOTAL FOR THE TT1REE CATEGORIES OF NETS, RAW DATA) r 1 -7s o1-7s | 2 afltr Qz-zz CVE A e'le AMA O 2-7s O 2-76 e)11-76 @ s-zo o 11-77 e B-zz ABA o 1 2-7e o 8-76 el 7-7e 6 B-7s O PBO 5-77 os-7s o 6-78 E\liE FIG. 11. CHANGES SPECIES CAUGHI AT }ffiAN FOR -13- IN c.p.u.e. BETWEEN 1975 AND 1979 FOR TI]E MAIN .T'IIE SEMII]N STAT1ON (SASSANDRA) (NUI"{BER OF FISH THE WIIOLE SET OT ClLi, I.TF,TS IJMPI.OYED) c.p.u.e. nh ind./ 100m?/ni9ht A.baremoze P.bovei A.imberi FIG. 11. CHANGES SPECIES CAUGHI AT }ffiAN FOR -13- IN c.p.u.e. BETWEEN 1975 AND 1979 FOR TI]E MAIN .T'IIE SEMII]N STAT1ON (SASSANDRA) (NUI"{BER OF FISH THE WIIOLE SET OT ClLi, I.TF,TS IJMPI.OYED) c.p.u.e. nh ind./ 100m?/ni9ht A.baremoze P.bovei A.imberi - 14- 2"L.4 Bandama Basin ;'our stations were sampled in the Bandama Basini two on the main river lrom which the basin takos its name (at l*,larabadiassa and Niakaramandougorr), one on the N'Zi, a left-bank tributary (Dabakala) and one (Mankono) on the Maraou6, which enters the Bandama on its right bank. Samplirrg at Mankono rras not begun until June 1977, alnost one year before the first ir0ecticidal treatment of this river (April 1978). ' In the area investigated the N'Zi is a narrow river, the hydrological regime of which is di?ectly influenced, by local rainfa1l. Because this results in appreciable fluctuati-ons of r^rater 1evel, which may alter t.he catch results, we shall give separate consideration Lo the results obtained in this river. Apart from the Dabakala station, three other stations downstream (l,lafa, M'Bahiakro and Bocanda) were also investigated over a period of one year to test the effect of another insecticide (chlorphoxim) for blackfly control. We would recall that, whereas this new pesticide is proving to be far more toxic than Abate to the invertebrate fauna (Dejoux & Troubat, L976), no harmful eifect on the existing fish populations has been detected in the short term (L6v6que et al.,1977; I'1as1in-L6ny &M6rona, 1978). There was no significant reduction of total catches at Dabakala during the period However, it will be noted that the samplingscovered by the investigation (Fig. 12). carried out during the high-water per iod because of the difficulty of using gi11 be recalled that the high-water periods and affected the recruitment to the fish generally yield very 1o\^r results, particularly nets properly during this period. It shorrLd also in 1976 and 1977 were parti-cu1ar1y poor in the N'Zi stocks; nevertheless, despite these poor conditi-ons, a certain number of young fish did manage to survive, as i-s shown by the catches in fine-mesh nets (category 1). rIG. L2. TIiE BANDAMA (DABAKAIA AND MANKONO) - CHANGES IN MEAN c.p.u.e. FOR THE WHOLE SET OF GILL NETS EMPLOYED c p.u.e. kg / 1oonr2/night Mankono 1 976 1977 1978 --7s7s- from the previous station, and values are of roughly the same At the N'Zi-Maf f a st'ation, which j .; slightly domstream which has been regularly sampled for two year,., the c.p.u.e. order of magnitude, and an appreciable reduction of catches also to be noted on account of the poor sampling conditions. during the high-water period is The other three monitoring stations in the B;rudama Basin are representative of the area of the middle reach, which is in fact the major part of the rivcr (de M6rona, 1981), and which i6 characterized by the presence of the species Alestes baremoze, A. macrolepidotus, Eutropius mentalis and HydrocynuF forskalii. However, the influence of the Kossou 1ake, filled in 1971, gradually made itself felt during the observation period. ' Although the changes in c.ir.u.e. (Iiigs 12 anct 13) at the monitoringstationsare fairly uneven, there would not appear to be any tendency towards an appreciable reciuction in catches, since the figures obtained in 1979 at stations on the Bandama are sj-milar to those for L975. On more detailed consideration, however, it is noted that the catches r,rith fine-mesh nets at Niaka rrere Poorer between the end of L97 6 and the middle of t 'i 9. The situation existing at the cormlencment of observation would appear to have been restored at the end of L979. - 14- 2"L.4 Bandama Basin ;'our stations were sampled in the Bandama Basini two on the main river lrom which the basin takos its name (at l*,larabadiassa and Niakaramandougorr), one on the N'Zi, a left-bank tributary (Dabakala) and one (Mankono) on the Maraou6, which enters the Bandama on its right bank. Samplirrg at Mankono rras not begun until June 1977, alnost one year before the first ir0ecticidal treatment of this river (April 1978). ' In the area investigated the N'Zi is a narrow river, the hydrological regime of which is di?ectly influenced, by local rainfa1l. Because this results in appreciable fluctuati-ons of r^rater 1evel, which may alter t.he catch results, we shall give separate consideration Lo the results obtained in this river. Apart from the Dabakala station, three other stations downstream (l,lafa, M'Bahiakro and Bocanda) were also investigated over a period of one year to test the effect of another insecticide (chlorphoxim) for blackfly control. We would recall that, whereas this new pesticide is proving to be far more toxic than Abate to the invertebrate fauna (Dejoux & Troubat, L976), no harmful eifect on the existing fish populations has been detected in the short term (L6v6que et al.,1977; I'1as1in-L6ny &M6rona, 1978). There was no significant reduction of total catches at Dabakala during the period However, it will be noted that the samplingscovered by the investigation (Fig. 12). carried out during the high-water per iod because of the difficulty of using gi11 be recalled that the high-water periods and affected the recruitment to the fish generally yield very 1o\^r results, particularly nets properly during this period. It shorrLd also in 1976 and 1977 were parti-cu1ar1y poor in the N'Zi stocks; nevertheless, despite these poor conditi-ons, a certain number of young fish did manage to survive, as i-s shown by the catches in fine-mesh nets (category 1). rIG. L2. TIiE BANDAMA (DABAKAIA AND MANKONO) - CHANGES IN MEAN c.p.u.e. FOR THE WHOLE SET OF GILL NETS EMPLOYED c p.u.e. kg / 1oonr2/night Mankono 1 976 1977 1978 --7s7s- from the previous station, and values are of roughly the same At the N'Zi-Maf f a st'ation, which j .; slightly domstream which has been regularly sampled for two year,., the c.p.u.e. order of magnitude, and an appreciable reduction of catches also to be noted on account of the poor sampling conditions. during the high-water period is The other three monitoring stations in the B;rudama Basin are representative of the area of the middle reach, which is in fact the major part of the rivcr (de M6rona, 1981), and which i6 characterized by the presence of the species Alestes baremoze, A. macrolepidotus, Eutropius mentalis and HydrocynuF forskalii. However, the influence of the Kossou 1ake, filled in 1971, gradually made itself felt during the observation period. ' Although the changes in c.ir.u.e. (Iiigs 12 anct 13) at the monitoringstationsare fairly uneven, there would not appear to be any tendency towards an appreciable reciuction in catches, since the figures obtained in 1979 at stations on the Bandama are sj-milar to those for L975. On more detailed consideration, however, it is noted that the catches r,rith fine-mesh nets at Niaka rrere Poorer between the end of L97 6 and the middle of t 'i 9. The situation existing at the cormlencment of observation would appear to have been restored at the end of L979. qa rabaC lassa N- f akar ma ndougou FTG. _15- 13. I}lE BANDA(YA (MARABADIA9SA AND NIAI(A) - rOR THE WHOLE SET OF NETS EMPLOYED AND BY C'HANCIIS IN MEAN c.P.u.e. CATEGORIES OF NET Band.rma: toIal weighL for the whole of fish per lOO ml of ner per night set of ners840c: ?ooG I l I 50ccl I I 1 I I :ooc,{ I) I i1(' -/ l9;.r I97q r974 1;7 5 Marabad ias sa 10: i 1A:i 197D L917 - weight of fish catrght per per night for each of Ehe L91 8 L9i2 lon m2 of necs three caEegories ]976 1976 L977 L9i 7 t9 / b r97S 1979 ish night es of Niakaramandougotr: weight of f caught per 1OO m' of nef.s per net .= =.rJ I aTOq L.1OOO. tooJ uooJ I ,oool qa rabaC lassa N- f akar ma ndougou FTG. _15- 13. I}lE BANDA(YA (MARABADIA9SA AND NIAI(A) - rOR THE WHOLE SET OF NETS EMPLOYED AND BY C'HANCIIS IN MEAN c.P.u.e. CATEGORIES OF NET Band.rma: toIal weighL for the whole of fish per lOO ml of ner per night set of ners840c: ?ooG I l I 50ccl I I 1 I I :ooc,{ I) I i1(' -/ l9;.r I97q r974 1;7 5 Marabad ias sa 10: i 1A:i 197D L917 - weight of fish catrght per per night for each of Ehe L91 8 L9i2 lon m2 of necs three caEegories ]976 1976 L977 L9i 7 t9 / b r97S 1979 ish night es of Niakaramandougotr: weight of f caught per 1OO m' of nef.s per net .= =.rJ I aTOq L.1OOO. tooJ uooJ I ,oool - 16- IloJN L l-i* t I ln t i l.,ti\ t I l. Vt .J 'J t4 L a4 C) c, 5 a C): o a o. H (Jl!pr{ C)o FJ(aAEIEI-{ El(J El (/]P{H(a t{'zz Fl Fl 4Fl =H CJt4 :E t!HO DdH oE1fq u) O\ r,l NFIo'OJ:ts =Azt4<EH\,sdorO H[q zz Hfi3> Er E] Ea ;da . F..,1otr JT "o& .&O trl @zz F{= zgav E](J< .t <, OZ Eq .rd\t :d -lH .zC)HH Eq o oc o ! D O ,0 tn h CJ al Ct c) 0) t4 q, x{ E CJ a a -i iC t4 o - 16- IloJN L l-i* t I ln t i l.,ti\ t I l. Vt .J 'J t4 L a4 C) c, 5 a C): o a o. H (Jl!pr{ C)o FJ(aAEIEI-{ El(J El (/]P{H(a t{'zz Fl Fl 4Fl =H CJt4 :E t!HO DdH oE1fq u) O\ r,l NFIo'OJ:ts =Azt4<EH\,sdorO H[q zz Hfi3> Er E] Ea ;da . F..,1otr JT "o& .&O trl @zz F{= zgav E](J< .t <, OZ Eq .rd\t :d -lH .zC)HH Eq o oc o ! D O ,0 tn h CJ al Ct c) 0) t4 q, x{ E CJ a a -i iC t4 o a- 17- It should also be mentioned that a clear reduction of the condition factor of some specic.s was noted at the Niaka station in 1976 and 1977 (s.e below). on ttrc oilrer lrand, tho c:rtclror taken with nets in category 2 incr:eased dtrling this period. Tht-r<,uAs.1 chang,c jn thc cotnposiEion of the catches, namely a reduction in Lhe weight ancl the' nrrmtrnrs <.rf A. h:rrcm,r;'etaken, offset by increases for E. mertalis (Fig. 14). From the end of 1979, howcvor, thc c.p.u.r-,. of A. harcmozc r.ec()vcr(,tj.inrl rcm,rinccl stablrin 1980 (Traore, personal communication). It seems most probable ihaL this \^ras a naturalphenomenon connecLed in pariicular with thr'1ack of water between l97b ancl 197g. porA. baremoze ' the absence of a high-water period is reflected in the main in thc catches ofthe fine-mesh nets, but mat,y adult {ish did not reach sexu;rl maturity during tlris period, which led to a Poor recruitment in the following year. Consequently, this poor recruitment cannot be ascribed to temephos pollution. In the Maraou6, a sharp decline in the numbers of fish taken is to be. note.d at Mankonoit L977 ' well before the treatment begun in mid-1978, which would not appear to have had an appreciable effect on the size of catches (Fig. L2) . The anal.ysis of the raw data for weight (Fig. 15) effects a fairly clear separationbetween the observations made at each of the three stations; this is due to a difference inthe species composition of the fish populations. A. baremoze and A. imberi make a strong contribution to axis 1, which explains 32% of t-he variance, wherea" o. a*i" 2 (16% of the variance) ttre species playing this role are L. senegalensiq, A. nurse and H. forskalii. FIG. 15. THE BANDAMA - FACTOR ANALYSIS CARRIED OU"I Oi{ THE WEIGIIT OF I'ISH cAucllr (TorAL roR rHE THREE CATEGORTES oF NETS, RAw DATA) I lv'larabadiassa O Nraka O Mankono tsE lzs ats Izs !7s 874 ..lAltU I rrui 1ru 176 .}rr I 74 :- lzo; r;l za!|^Izsrzzr.l I7e 1 i:i;-'' ' irslirl" +76 176 '"') | l', )tt7s I +r, atlla )tti .,,t a77 78O I HFo tts i 'ul * " 'u1.t" t'u I Ii .r, a- 17- It should also be mentioned that a clear reduction of the condition factor of some specic.s was noted at the Niaka station in 1976 and 1977 (s.e below). on ttrc oilrer lrand, tho c:rtclror taken with nets in category 2 incr:eased dtrling this period. Tht-r<,uAs.1 chang,c jn thc cotnposiEion of the catches, namely a reduction in Lhe weight ancl the' nrrmtrnrs <.rf A. h:rrcm,r;'etaken, offset by increases for E. mertalis (Fig. 14). From the end of 1979, howcvor, thc c.p.u.r-,. of A. harcmozc r.ec()vcr(,tj.inrl rcm,rinccl stablrin 1980 (Traore, personal communication). It seems most probable ihaL this \^ras a naturalphenomenon connecLed in pariicular with thr'1ack of water between l97b ancl 197g. porA. baremoze ' the absence of a high-water period is reflected in the main in thc catches ofthe fine-mesh nets, but mat,y adult {ish did not reach sexu;rl maturity during tlris period, which led to a Poor recruitment in the following year. Consequently, this poor recruitment cannot be ascribed to temephos pollution. In the Maraou6, a sharp decline in the numbers of fish taken is to be. note.d at Mankonoit L977 ' well before the treatment begun in mid-1978, which would not appear to have had an appreciable effect on the size of catches (Fig. L2) . The anal.ysis of the raw data for weight (Fig. 15) effects a fairly clear separationbetween the observations made at each of the three stations; this is due to a difference inthe species composition of the fish populations. A. baremoze and A. imberi make a strong contribution to axis 1, which explains 32% of t-he variance, wherea" o. a*i" 2 (16% of the variance) ttre species playing this role are L. senegalensiq, A. nurse and H. forskalii. FIG. 15. THE BANDAMA - FACTOR ANALYSIS CARRIED OU"I Oi{ THE WEIGIIT OF I'ISH cAucllr (TorAL roR rHE THREE CATEGORTES oF NETS, RAw DATA) I lv'larabadiassa O Nraka O Mankono tsE lzs ats Izs !7s 874 ..lAltU I rrui 1ru 176 .}rr I 74 :- lzo; r;l za!|^Izsrzzr.l I7e 1 i:i;-'' ' irslirl" +76 176 '"') | l', )tt7s I +r, atlla )tti .,,t a77 78O I HFo tts i 'ul * " 'u1.t" t'u I Ii .r, - 18- Relative abundance of L. senegalensis and $. imbqql is a feature of the Marabadiassa station; A. baremoze is the most plentiful at Niakaramandougou. Thc decline in the catches of this latter species Ln 1977-1978 (see above) had the effect of making the specific profile of fistring at Niakaramandougou relatively close to that of the llankono station during this period. H. fC-fr!sli.! was proportionally more plentiful at Mankono than at the other two stations, vhich explains why the points representing the observations are rclatively grouped fot this station. Analysis of coded c.p.u.e. shows that there is a contrast at Marabadiassa, as at Niakaramandougou, between the catches at the end of the high-water period and those at -the etll of the low-water period (Fig. 16), the forrnnr being characterized in particular by tfie presence of numerous A. baremoze on an anadromous spawning migration (Fig. 14). FIG. 16" I]IE BANDAMA - FACTOR ANALYSIS CARRIED OI]T ON T}IE NUMBERS OF FISH CAUGHT (TOTAL FOR THE THREE C.ATEGORIES OF NETS, CODED DATA) I Niakaramandougou: Q Niakaramandougou: !Marabadiassa: high [Marabadiassa: low high waterperrod iSYtBat"' t,vl water period vrater period c o CEB t ABATT FrCl &u"l II e t , The populations sampled by electrofishing (de M6rona et -from those sampled by gill net, and the technique is one that relatively shallow area. Roughly speaking, three groups of oo o LPA aL. , L977) are fairly different is effective for use only in a species are distinguished: I Ift II I nI I tr I lll trr: u -^-+. oo9 l ooo E 80tr o o DR9 2.2 Electrofishing The fishing, which was carried out in the main in areas of rapids where the river bed was rocky or stony, was complementary to the monitoring prog,ranrne and i-s not therefore strictly within the terms of reference. Nevertheless, it di,d ad<l some additional elements to our knowledge of the environment. and to interpretation of the effects of the insecticide. - typical running-water species: Amphilius atesuensis, Phractura clauseni, Chiloglanis occidentalis, etc. ; - 18- Relative abundance of L. senegalensis and $. imbqql is a feature of the Marabadiassa station; A. baremoze is the most plentiful at Niakaramandougou. Thc decline in the catches of this latter species Ln 1977-1978 (see above) had the effect of making the specific profile of fistring at Niakaramandougou relatively close to that of the llankono station during this period. H. fC-fr!sli.! was proportionally more plentiful at Mankono than at the other two stations, vhich explains why the points representing the observations are rclatively grouped fot this station. Analysis of coded c.p.u.e. shows that there is a contrast at Marabadiassa, as at Niakaramandougou, between the catches at the end of the high-water period and those at -the etll of the low-water period (Fig. 16), the forrnnr being characterized in particular by tfie presence of numerous A. baremoze on an anadromous spawning migration (Fig. 14). FIG. 16" I]IE BANDAMA - FACTOR ANALYSIS CARRIED OI]T ON T}IE NUMBERS OF FISH CAUGHT (TOTAL FOR THE THREE C.ATEGORIES OF NETS, CODED DATA) I Niakaramandougou: Q Niakaramandougou: !Marabadiassa: high [Marabadiassa: low high waterperrod iSYtBat"' t,vl water period vrater period c o CEB t ABATT FrCl &u"l II e t , The populations sampled by electrofishing (de M6rona et -from those sampled by gill net, and the technique is one that relatively shallow area. Roughly speaking, three groups of oo o LPA aL. , L977) are fairly different is effective for use only in a species are distinguished: I Ift II I nI I tr I lll trr: u -^-+. oo9 l ooo E 80tr o o DR9 2.2 Electrofishing The fishing, which was carried out in the main in areas of rapids where the river bed was rocky or stony, was complementary to the monitoring prog,ranrne and i-s not therefore strictly within the terms of reference. Nevertheless, it di,d ad<l some additional elements to our knowledge of the environment. and to interpretation of the effects of the insecticide. - typical running-water species: Amphilius atesuensis, Phractura clauseni, Chiloglanis occidentalis, etc. ; -19- - the,young of certain spccies, the adults ttf wtrich live in pool reaches: l:_!rro!,A. imberi, B- waldroni, C. velifer, and Tilapia spp. in relatively shallow zones btrt not in rapids ; ' - species of the rocky zonc: Labeo parvus, llastacembelus spp., Svnodontis spp. andllormyridae, etc. The species that feed in the rapids, scveral of which consume insect larvae, should,a priori, be the ones most directly affected by r,:leases of insecticicies, eithcr through theidirect acEion, or because they produce an appreciable reduction in the abundance of prey. There was no appreciable change in tl-re fish populations of the rapids on the Ifraba andon the Bandama at Mar.rir;rdiassa in 1976 an<l 1977. In the N'zi, on the other hand, profound changes were noted, but these would appear rather to have been due to the unfavourablehydrological conditions, which resulted in the drying out of the rapids for several monthsin 1977. 2.3 Conclusi.ons Catches per unit of effort (c.p.u.e.) nets in one night of fishing, are an index samp led. expressed in grams of fish caught per 1OO m2 of of the relative abundance of fish in the environmer At the stations,sampled, the mean c.p.u.e, for the whole set around 2oo} gf rcO ^Zf night of ti.ting for rhe Bandama, 15oo g forthe Sassan<lra. These result.s may be compared \./ith those obtained of sampling, in other untreated rivers in west Africa (unpubrished 198 1. - Ivory Coasr: Cavally (455-15OO); Nipoue (12OO_24OO); of nets employed here are the Como6-L6raba, 8OO g for by us, using the same mean data), and L6v6que & paugy - Senegal: Gambie (860-3480, in general between 2OOO and 3OOO) ; - Guinea: Ni 1',o1 (14oo- 17oo) ; Tornine (13oo) ; Koumba (147o) ; Kolenre (35oo) ; Irtakona(650) ; Dlani (9oo). rt therefore seems that the c.p.u.e. of the treated rivers are quite comparable to whatis observed in other rivers, ancl that temephos treatment does not have an adverse effect inthis sphere. 3. BIOLOGY OF TTIE MAIN SPECIES 3. I Reproductive cycle Many species have a reproductive cycle related to the l,ydrological regime, and they spawduring the high-water period (Table 1). rhis phenomenoniswell knorun in the sahel zor\ei wherit occurs the rivers spre-rd out over the bottom land and the floocl plains, which provideshelter and food for the young fishes. As regards the rrvers investigated, the flooded zo.erare in general of smal1 size, but nevertheless the waters of the high-water period do invadethe biotopes most suited to the development of young fishes, especially in the woodlands alonpthe bottom land. Most of the Mormyridae and some of the characidae have a reproductive period Iimited toa few months, during the high-water period, but slightly longer than that of the siluridae.Tire reproduction of the cyprinidae is protracted, with an interruption during the dry season. Some of these spocies also undertake anadromous spar.rnirrg migrations. At theNiakaramandougou station, for ^xample, we have seen large quantities of Alestes baremozeErrtropiusmenta1is-andMormyrops1ongicepscaughtby1oca1fishermeni"ffiffiiouts. whereas these species are less p1en1iful or relatively scarce in catches during the rest';;'"'the year. -19- - the,young of certain spccies, the adults ttf wtrich live in pool reaches: l:_!rro!,A. imberi, B- waldroni, C. velifer, and Tilapia spp. in relatively shallow zones btrt not in rapids ; ' - species of the rocky zonc: Labeo parvus, llastacembelus spp., Svnodontis spp. andllormyridae, etc. The species that feed in the rapids, scveral of which consume insect larvae, should,a priori, be the ones most directly affected by r,:leases of insecticicies, eithcr through theidirect acEion, or because they produce an appreciable reduction in the abundance of prey. There was no appreciable change in tl-re fish populations of the rapids on the Ifraba andon the Bandama at Mar.rir;rdiassa in 1976 an<l 1977. In the N'zi, on the other hand, profound changes were noted, but these would appear rather to have been due to the unfavourablehydrological conditions, which resulted in the drying out of the rapids for several monthsin 1977. 2.3 Conclusi.ons Catches per unit of effort (c.p.u.e.) nets in one night of fishing, are an index samp led. expressed in grams of fish caught per 1OO m2 of of the relative abundance of fish in the environmer At the stations,sampled, the mean c.p.u.e, for the whole set around 2oo} gf rcO ^Zf night of ti.ting for rhe Bandama, 15oo g forthe Sassan<lra. These result.s may be compared \./ith those obtained of sampling, in other untreated rivers in west Africa (unpubrished 198 1. - Ivory Coasr: Cavally (455-15OO); Nipoue (12OO_24OO); of nets employed here are the Como6-L6raba, 8OO g for by us, using the same mean data), and L6v6que & paugy - Senegal: Gambie (860-3480, in general between 2OOO and 3OOO) ; - Guinea: Ni 1',o1 (14oo- 17oo) ; Tornine (13oo) ; Koumba (147o) ; Kolenre (35oo) ; Irtakona(650) ; Dlani (9oo). rt therefore seems that the c.p.u.e. of the treated rivers are quite comparable to whatis observed in other rivers, ancl that temephos treatment does not have an adverse effect inthis sphere. 3. BIOLOGY OF TTIE MAIN SPECIES 3. I Reproductive cycle Many species have a reproductive cycle related to the l,ydrological regime, and they spawduring the high-water period (Table 1). rhis phenomenoniswell knorun in the sahel zor\ei wherit occurs the rivers spre-rd out over the bottom land and the floocl plains, which provideshelter and food for the young fishes. As regards the rrvers investigated, the flooded zo.erare in general of smal1 size, but nevertheless the waters of the high-water period do invadethe biotopes most suited to the development of young fishes, especially in the woodlands alonpthe bottom land. Most of the Mormyridae and some of the characidae have a reproductive period Iimited toa few months, during the high-water period, but slightly longer than that of the siluridae.Tire reproduction of the cyprinidae is protracted, with an interruption during the dry season. Some of these spocies also undertake anadromous spar.rnirrg migrations. At theNiakaramandougou station, for ^xample, we have seen large quantities of Alestes baremozeErrtropiusmenta1is-andMormyrops1ongicepscaughtby1oca1fishermeni"ffiffiiouts. whereas these species are less p1en1iful or relatively scarce in catches during the rest';;'"'the year. -20- aoHd E] Or EH z) CJ rI1H F.{ ca4oozc) AdooH>&H E]pit{ F] E{ HtrHO C] >qa&OE]&>O{H .d&& rd >{ F{ E4HZdH :)H(a4HE14z rl <f Fl5 |-l X r--trd() <n (J FAzHH L)(J<= H3 ts(atardT,A H(,hrdAz(no zTdHH< zf:l E1Fr :dF{ Ii{ .o Et H H(J =aoil A{ tdd E o\c .-l = o o(\ t{ql olx +)l o F{ Itl o Ed o =z r') .-l .{ = -t ,-l ra r-l = rn O) r{ lrd o >t ! o ltf o td o G.t{(d (l.) x U o J L,) r.! tn q, h 'oc(\ rn cc .-{ lr C.' >\ d c! q{ o d CJ == o an .-l rJ o Fl E LO tn zOH H *p F'g H a&H tq >HH OI)n,li3\-DoL)b0EI O0tr a, ,'t +l HHs2 p, o o o t-- rn ,-l o o r. .J N oo cc N N o oo c\ CD -t cc\i cf) o o O o cc -t o oo(, ca N a\ oo c{ trl a rI]H(J Idtu a rl '-.,1 U a a ...,1 CJ L] a o)r a ! l.r .-{ .-1 c) (J 'o rl LC .4 a a C] -J tr CJ uJ a j t, i, I -l r) I ol 'ti Iil :llL' I -rl-rol t,!l oUi.aftl=Zl's I <l; a) =a , ..{ c.) !{ c) ! ,,1 el oo -20- aoHd E] Or EH z) CJ rI1H F.{ ca4oozc) AdooH>&H E]pit{ F] E{ HtrHO C] >qa&OE]&>O{H .d&& rd >{ F{ E4HZdH :)H(a4HE14z rl <f Fl5 |-l X r--trd() <n (J FAzHH L)(J<= H3 ts(atardT,A H(,hrdAz(no zTdHH< zf:l E1Fr :dF{ Ii{ .o Et H H(J =aoil A{ tdd E o\c .-l = o o(\ t{ql olx +)l o F{ Itl o Ed o =z r') .-l .{ = -t ,-l ra r-l = rn O) r{ lrd o >t ! o ltf o td o G.t{(d (l.) x U o J L,) r.! tn q, h 'oc(\ rn cc .-{ lr C.' >\ d c! q{ o d CJ == o an .-l rJ o Fl E LO tn zOH H *p F'g H a&H tq >HH OI)n,li3\-DoL)b0EI O0tr a, ,'t +l HHs2 p, o o o t-- rn ,-l o o r. .J N oo cc N N o oo c\ CD -t cc\i cf) o o O o cc -t o oo(, ca N a\ oo c{ trl a rI]H(J Idtu a rl '-.,1 U a a ...,1 CJ L] a o)r a ! l.r .-{ .-1 c) (J 'o rl LC .4 a a C] -J tr CJ uJ a j t, i, I -l r) I ol 'ti Iil :llL' I -rl-rol t,!l oUi.aftl=Zl's I <l; a) =a , ..{ c.) !{ c) ! ,,1 el oo -21 - The levt:1s rcachcd dtrri-rrg the high-water perio{j appear to have a dirccl: inl lrrencr, t>n L su.cess of tl.re reproducti on of sJrecies \^ri th an annlral cyc1e. Dansoko ot- al . (197 6) lrave shown th;ft the rc<Ltction in commercial caIcl-rcs r:f Hy<lrocynus brevis and ]lyrirocl'nus_Lg_f-{g-f-l in the central delta of the Niger in 1972 ancl 1973 was a conseqrrence of inadequatc I,.ve'1s duri-n6; the higir-v'ater periods and was accourlted f or by a poor cond iL ion f,-rr-t.'r , i irrri ,- - J gro' and weak recrrtitmcnt to the f ish stocks during ttris period. The srrrvival oI 1zt'rr;.,, i j shcsis threatened in years whe,n there is insutficient water: the young fjsbes havc f,r.rer refug and are more vtrlncrablt' Lo preclators ; and there art f ewer .avai lab [e so'rrct..; ol f ooci. In our opinion, this influencc of the high-water periods cxplains thc changes in the numbers ol some species dtrring the five years ol obscrvation. We lrave sLre:rsed this, in particular, f or ljchilbe mystus , the redrrction of whost' nunbcrs in expcrimental lishi.rrg in t' Como6-L6raba Basin corrld be explained by rhe inadequalc lt'vc1:; of Ltrt'hii1[r-water pcriods in 1976 and 1978, which werc, moreover, general for all of the basins investigated here. Lt therefore seems that this phenomenon is not ascribable to temephos, as could have been thou' from a irasty interprer-ation, but lrr the action of natural phenomena. Frrrthcrmore, confirm. tion for this hypothesis is provided by f.he improvement of catches in 1979. It is plar-tsib that other species could have been affect-ed by the high-water periods, although not as spectacularly as in the previorrs example. However, some species reproduce throughout the year (Alestes uracrolepidotus_, Alestes -ilUer.! and some of the Cichlidae), and it has nol been possible to establish any especially Larvourable period. Thesr, species are apparently more sedentary and better adapted than ttr, preceding species to low-\^rater periods for the survival of their young. Albaret (1982) has studied the fecundity of many species. The fecur,dity of Cyprinida, Characidae and Schilbeidae is very high (more than 150 OOO eggs per kg o[ the female), and their eggs are of srna11 diameter. These eggs, which in general contain 1ittle yokc, are released in large numbers into the natural environment and mr:rtality is high; the eggs of Mormyridae, Claridae ahd some individual species @.i1irg, !_g!:-p.!g:yS :gfgA.r1us), arc larg, and their fecundity is in the middle range (25 OOO-9O OOO eggs per kg of tlrc, temale). Las Bagridae, Cichlidae and ccrtain species such as Mastacembeltrs nig,ronrzrrgin,ll.trx, !e]aa!S11,r endlicheri- and Papyrocramus. afer, have eggs ranging from large to vcry large, ancl tlrerr f ecundity ranges f rom ;t few thousand to 25 or JO ()( ,,r. It ir. r',,rrcra1Iy rccol;nizcd tltat tlrc species take ntc,re care with their eggs, which are l;rid in sheltered places. These differences in fecundity arrd in the survival potential of the eggs ("r" and t'ktt strrtegy) may probably play a role in the dynamics of fish populations subjer:tcd tc, releasc of insecLicides. Thc survival of the yorrng fislres is, in effect, better a priori in the case of species whose eggs are large and few in number, which are generally benthic species tl-ran in the case of other species whose pelagic young, which are sma1l and without nutrient reserves, would be more susceptible to insecticide waves. However, it has not been possib to nrake any study of this matter, particularly because of the difficrrlties of sampling. The fact that we have been unable in various observations Lo cstablish clifferences in ttre fecundity of species inhabiting treated and unt.reated rivers could indicate that tcmeph, has rro marked effect on the reproductive physiology of fishes. 3.2 Coefficient of condition 105 wthe changcs over time in the coefficient of condition (K = LSI ) , which ir. a simplifir expression of the health of fishes, are dependent on several factors. First and foremost, these are natural factors, principally the avail;rbility of food, which is in general more plentiful during the high-water period than during the low-water period, and the sexual cyc for annually reproducing species. Spawning involves a reduction of short duration in condition Iollowing the release of ripe oocytes. Because the oocytes are generally formed from reserve subsLitnces, there is no appreciable weight gain during the maturation period /Dr-rrand & Loubens, 1970). -21 - The levt:1s rcachcd dtrri-rrg the high-water perio{j appear to have a dirccl: inl lrrencr, t>n L su.cess of tl.re reproducti on of sJrecies \^ri th an annlral cyc1e. Dansoko ot- al . (197 6) lrave shown th;ft the rc<Ltction in commercial caIcl-rcs r:f Hy<lrocynus brevis and ]lyrirocl'nus_Lg_f-{g-f-l in the central delta of the Niger in 1972 ancl 1973 was a conseqrrence of inadequatc I,.ve'1s duri-n6; the higir-v'ater periods and was accourlted f or by a poor cond iL ion f,-rr-t.'r , i irrri ,- - J gro' and weak recrrtitmcnt to the f ish stocks during ttris period. The srrrvival oI 1zt'rr;.,, i j shcsis threatened in years whe,n there is insutficient water: the young fjsbes havc f,r.rer refug and are more vtrlncrablt' Lo preclators ; and there art f ewer .avai lab [e so'rrct..; ol f ooci. In our opinion, this influencc of the high-water periods cxplains thc changes in the numbers ol some species dtrring the five years ol obscrvation. We lrave sLre:rsed this, in particular, f or ljchilbe mystus , the redrrction of whost' nunbcrs in expcrimental lishi.rrg in t' Como6-L6raba Basin corrld be explained by rhe inadequalc lt'vc1:; of Ltrt'hii1[r-water pcriods in 1976 and 1978, which werc, moreover, general for all of the basins investigated here. Lt therefore seems that this phenomenon is not ascribable to temephos, as could have been thou' from a irasty interprer-ation, but lrr the action of natural phenomena. Frrrthcrmore, confirm. tion for this hypothesis is provided by f.he improvement of catches in 1979. It is plar-tsib that other species could have been affect-ed by the high-water periods, although not as spectacularly as in the previorrs example. However, some species reproduce throughout the year (Alestes uracrolepidotus_, Alestes -ilUer.! and some of the Cichlidae), and it has nol been possible to establish any especially Larvourable period. Thesr, species are apparently more sedentary and better adapted than ttr, preceding species to low-\^rater periods for the survival of their young. Albaret (1982) has studied the fecundity of many species. The fecur,dity of Cyprinida, Characidae and Schilbeidae is very high (more than 150 OOO eggs per kg o[ the female), and their eggs are of srna11 diameter. These eggs, which in general contain 1ittle yokc, are released in large numbers into the natural environment and mr:rtality is high; the eggs of Mormyridae, Claridae ahd some individual species @.i1irg, !_g!:-p.!g:yS :gfgA.r1us), arc larg, and their fecundity is in the middle range (25 OOO-9O OOO eggs per kg of tlrc, temale). Las Bagridae, Cichlidae and ccrtain species such as Mastacembeltrs nig,ronrzrrgin,ll.trx, !e]aa!S11,r endlicheri- and Papyrocramus. afer, have eggs ranging from large to vcry large, ancl tlrerr f ecundity ranges f rom ;t few thousand to 25 or JO ()( ,,r. It ir. r',,rrcra1Iy rccol;nizcd tltat tlrc species take ntc,re care with their eggs, which are l;rid in sheltered places. These differences in fecundity arrd in the survival potential of the eggs ("r" and t'ktt strrtegy) may probably play a role in the dynamics of fish populations subjer:tcd tc, releasc of insecLicides. Thc survival of the yorrng fislres is, in effect, better a priori in the case of species whose eggs are large and few in number, which are generally benthic species tl-ran in the case of other species whose pelagic young, which are sma1l and without nutrient reserves, would be more susceptible to insecticide waves. However, it has not been possib to nrake any study of this matter, particularly because of the difficrrlties of sampling. The fact that we have been unable in various observations Lo cstablish clifferences in ttre fecundity of species inhabiting treated and unt.reated rivers could indicate that tcmeph, has rro marked effect on the reproductive physiology of fishes. 3.2 Coefficient of condition 105 wthe changcs over time in the coefficient of condition (K = LSI ) , which ir. a simplifir expression of the health of fishes, are dependent on several factors. First and foremost, these are natural factors, principally the avail;rbility of food, which is in general more plentiful during the high-water period than during the low-water period, and the sexual cyc for annually reproducing species. Spawning involves a reduction of short duration in condition Iollowing the release of ripe oocytes. Because the oocytes are generally formed from reserve subsLitnces, there is no appreciable weight gain during the maturation period /Dr-rrand & Loubens, 1970). -22- Extrinsic factors may also alter the condition of fishes. By killing off the organisms on whicir fishes feed, various toxic substances reduce the amount of available food Lo thepoint of complete disappearance in extreme cases. Jhe accumulation of pesticides in thetissues (Quelennec et a1. , 1977) may also modify the metabolism of fishes. One of the difficuLties encountered in the use of the coefficient of condition is thatit is subject to variation in relation to the size of individual fish within the samesfEcies. l^Ie have demonstrated this, in particular, in schilbe mystus (L6v6que & Ilerbinet,1980). rn general, the valrres become stable on reaching rnit"ritn Consequently, the "length of the fish ,on reaching sexual maturity must be determined if one wishes to rrace ihe13ng-term development of this factor, and all values other than those observed in adult fish,or in a length ran8e preselected for each species musL be discarded, as has been done in this.study. For continuously-reproducing species, the results obtained in the various basins(rig' 17) show that values of the coefficient of condition are r:elatively random, fluctuatingaround a mean which does not seem appreciably altered at the end of fiv,-: years of monitoring.Hovever, a tendency toward reduction is to be noted in 1977 and 1g7g; this tendency affectedboth the populations of treated basins and those of the Sassandra Basin. which hras nottreated until June 1978. Annually-reProducing species were similarly affected; an annual cycle in which there isa reductior during the low-water period anrl a maximum at the ertl of the high-water period isto be noted in some instances. This is particularly the case for:1 r1ystus, in the L6raba.Nevertheless, there is generally speaking no very clear cyc1e. lloi-Is there any catastrophicreduction in the mean values to justify the conciusion that temephos has a direct or indirect e ffe ct. Going into Sreater detail, it will be noted tirat many species have a coefficient ofcondition that is in general higher in the sassandra than in the Band.ima. These differencesbetween basins are comrnon and are probably due to more favourable or less favourableecological conditions or to the presence of different genetic populations. Moreover, tirt,same phenomenon has been observed in the Bandama, where the values recorded at Marabadiassavere higher than those at Niakaramandougou for some specics (proximicy to thc Kossou lakc?).Beth,een 1976 and L977 a very significant dpcline in the condition factor of Alestes baremoze.anda1esserdec1ineinthatofEutropiusmenta1is,waSnotedintheo"..-"r.r-,.ffi, between the Kossou,and Ferke"suao.rg-""-a*i-@ffi 197g). Subsequently, the coefficient ofcondition recovered at the tl^,o stations, arthouli, it did not reach the L974-1975 values.rnquiries carried ortt in 1976 revealed that thii phenomenon was confined to this region ofthe Bandama and that it was probably connected with the sealing of the two dams, which impededmigration (Daget et a1., Lg73) andwith a very low flood. This was not entirely negative inits effect, since Labeo senegalensis, on the other hand, developed well under these newconditions 1xor.""f,ffiffi.oefficient of condition r4,as increased. 3.3 Diet A qualitative study of diet was undertaken on fishes that consume at ieast a proportionof aquatic invertebrates. Some, such as Petrocephalu1*b."".I, feed entirely on aquatici-.vertebrates. Others and EuLropius mentalis) drawextensively on exoge.,o,r" ir,@;,;U. v"a ""aA' imberi) consume both, although with a pronounced tendency toarards the inrligenous fauna.Thg analysis of stomach contents carried out in 1975 on these species in treated and untreatedrivers does not provide clear evidence of an influence exerted by the insecticide, since thediet was appreciabJ-y the same in composition whatever the provenance of the fishes (Vidy,L9Z6)' Nevertheless, going into greater detail, it appears that Ephemeroptera, which areespecially sensitive to temephos (Dejoux & Troubat, t976), are less plentiful j-n the stomachsof fi'shes from treated rivers. However, the adaptive capacity of fishes in relation tonutrition has been demonstrated repeatedly in wesl Africa (Daget, Lg52, Lg54; Blache, lg64;whitehead, L969; Reynolds, 1973; Iauzanne, lg76). consequentl.y, it does not seem .rs iftemephos has had any apPreciable influence on the fishes by lowering food stocks to a critical -22- Extrinsic factors may also alter the condition of fishes. By killing off the organisms on whicir fishes feed, various toxic substances reduce the amount of available food Lo thepoint of complete disappearance in extreme cases. Jhe accumulation of pesticides in thetissues (Quelennec et a1. , 1977) may also modify the metabolism of fishes. One of the difficuLties encountered in the use of the coefficient of condition is thatit is subject to variation in relation to the size of individual fish within the samesfEcies. l^Ie have demonstrated this, in particular, in schilbe mystus (L6v6que & Ilerbinet,1980). rn general, the valrres become stable on reaching rnit"ritn Consequently, the "length of the fish ,on reaching sexual maturity must be determined if one wishes to rrace ihe13ng-term development of this factor, and all values other than those observed in adult fish,or in a length ran8e preselected for each species musL be discarded, as has been done in this.study. For continuously-reproducing species, the results obtained in the various basins(rig' 17) show that values of the coefficient of condition are r:elatively random, fluctuatingaround a mean which does not seem appreciably altered at the end of fiv,-: years of monitoring.Hovever, a tendency toward reduction is to be noted in 1977 and 1g7g; this tendency affectedboth the populations of treated basins and those of the Sassandra Basin. which hras nottreated until June 1978. Annually-reProducing species were similarly affected; an annual cycle in which there isa reductior during the low-water period anrl a maximum at the ertl of the high-water period isto be noted in some instances. This is particularly the case for:1 r1ystus, in the L6raba.Nevertheless, there is generally speaking no very clear cyc1e. lloi-Is there any catastrophicreduction in the mean values to justify the conciusion that temephos has a direct or indirect e ffe ct. Going into Sreater detail, it will be noted tirat many species have a coefficient ofcondition that is in general higher in the sassandra than in the Band.ima. These differencesbetween basins are comrnon and are probably due to more favourable or less favourableecological conditions or to the presence of different genetic populations. Moreover, tirt,same phenomenon has been observed in the Bandama, where the values recorded at Marabadiassavere higher than those at Niakaramandougou for some specics (proximicy to thc Kossou lakc?).Beth,een 1976 and L977 a very significant dpcline in the condition factor of Alestes baremoze.anda1esserdec1ineinthatofEutropiusmenta1is,waSnotedintheo"..-"r.r-,.ffi, between the Kossou,and Ferke"suao.rg-""-a*i-@ffi 197g). Subsequently, the coefficient ofcondition recovered at the tl^,o stations, arthouli, it did not reach the L974-1975 values.rnquiries carried ortt in 1976 revealed that thii phenomenon was confined to this region ofthe Bandama and that it was probably connected with the sealing of the two dams, which impededmigration (Daget et a1., Lg73) andwith a very low flood. This was not entirely negative inits effect, since Labeo senegalensis, on the other hand, developed well under these newconditions 1xor.""f,ffiffi.oefficient of condition r4,as increased. 3.3 Diet A qualitative study of diet was undertaken on fishes that consume at ieast a proportionof aquatic invertebrates. Some, such as Petrocephalu1*b."".I, feed entirely on aquatici-.vertebrates. Others and EuLropius mentalis) drawextensively on exoge.,o,r" ir,@;,;U. v"a ""aA' imberi) consume both, although with a pronounced tendency toarards the inrligenous fauna.Thg analysis of stomach contents carried out in 1975 on these species in treated and untreatedrivers does not provide clear evidence of an influence exerted by the insecticide, since thediet was appreciabJ-y the same in composition whatever the provenance of the fishes (Vidy,L9Z6)' Nevertheless, going into greater detail, it appears that Ephemeroptera, which areespecially sensitive to temephos (Dejoux & Troubat, t976), are less plentiful j-n the stomachsof fi'shes from treated rivers. However, the adaptive capacity of fishes in relation tonutrition has been demonstrated repeatedly in wesl Africa (Daget, Lg52, Lg54; Blache, lg64;whitehead, L969; Reynolds, 1973; Iauzanne, lg76). consequentl.y, it does not seem .rs iftemephos has had any apPreciable influence on the fishes by lowering food stocks to a critical a1 Frc. 17. DEvELoP!,IEM oF rHE coErFrcrENT oF coNDrrroN (i) oF THE MArN spEcTES AT THE VARIOUS MONIIORING STATIONS. ANNUALLY-REPRODUCING SPECIES I,5 r,.r r,3 r,? I,I t.o o,9 o,B 3,o 2,9 2,8 )1 2,6 2,3 2,4 z,f 2,2 2,O l,l 7,2 I,r 1,O o,9 o,I t,6 r,5 1,4 I,3 L,2 1,r l,o ;1 A. baremozo L17 i r373 Li'i ? L<J1 7 l()78 E. menla Its - ..-.i...... l9 74 'A. nurse S. mystus 19.7 i) SJSSJatr:; a1 Frc. 17. DEvELoP!,IEM oF rHE coErFrcrENT oF coNDrrroN (i) oF THE MArN spEcTES AT THE VARIOUS MONIIORING STATIONS. ANNUALLY-REPRODUCING SPECIES I,5 r,.r r,3 r,? I,I t.o o,9 o,B 3,o 2,9 2,8 )1 2,6 2,3 2,4 z,f 2,2 2,O l,l 7,2 I,r 1,O o,9 o,I t,6 r,5 1,4 I,3 L,2 1,r l,o ;1 A. baremozo L17 i r373 Li'i ? L<J1 7 l()78 E. menla Its - ..-.i...... l9 74 'A. nurse S. mystus 19.7 i) SJSSJatr:; -24- FIG. 17 (continued) .- DEVELOPMENT OF THE COEFFICIENT OF CONDITION (K) OF TI{E MAIN SPECIES AT TI{E VARIOUS MONITORING STAIIONS. CONTINUOUSLY-REPRODUCING SPECIES t,6 1,5 I,4 1,3 2,4 ?1 2,2 2,r 2,O r,9 1,8 L,7 l,I 3,O 2,9 2,7 2,6 2,5 2,4 2,3 1974 r9 74 197 5 197 o r97 6 r9'16 L911 L97 1 t971 l<171 t97S 1976 A. imberi r.978197 5 74 1975 3 Marabadi.rssa 1979 L. senegalonsis 2,8 2,1 2,6 2,4 7,1 2,L r 197 B 19u- 9 ' : Sassandra D H. f orskalii macrolepldotus -24- FIG. 17 (continued) .- DEVELOPMENT OF THE COEFFICIENT OF CONDITION (K) OF TI{E MAIN SPECIES AT TI{E VARIOUS MONITORING STAIIONS. CONTINUOUSLY-REPRODUCING SPECIES t,6 1,5 I,4 1,3 2,4 ?1 2,2 2,r 2,O r,9 1,8 L,7 l,I 3,O 2,9 2,7 2,6 2,5 2,4 2,3 1974 r9 74 197 5 197 o r97 6 r9'16 L911 L97 1 t971 l<171 t97S 1976 A. imberi r.978197 5 74 1975 3 Marabadi.rssa 1979 L. senegalonsis 2,8 2,1 2,6 2,4 7,1 2,L r 197 B 19u- 9 ' : Sassandra D H. f orskalii macrolepldotus I_25_ threshold. Moreover, it has been noted that, overa11, insecticidal treatments with temephos result in a temporary reduction of only 3O-4O7" in the stock of aquatic insects (Dejoux et al., 1979). Nor did analyses carried out in 1976 and 1977 reveal appreciable changes in the composition of the stomach contents relative,to 1975. A remark may be made here concerning Amphilius atesuensis. HaIf the diet of this species, which is associated with running water in which the flow rate is verv high, consists of Simulium larvae (l,leronar p.'rsonol conununica- tion). Despit.e the almost total disappearance of this target species in treated r./aters, we have not noted any special effect on Amphilius populations, which remain plentiful in the rapids. CONCLUS IONS We are 1ed to conclude, trom the result.s obtained over a period of five years in the course of monitoring rivers treated !,rith insecticides in the Ivorr Coast, that releases of temephos have had no detectable effect on the fish populations. We have not been able to establish any changes ascribable to the effects of the insecticide in the composition of the populations or in the size oI the catches in test fishing. Furthermore, nor would there seem to have been any appreciable allcration in the biology of the species or in their feeding habits during the period covered by the investigation. In studics of the effecLs of DDT in the ItVictoria Nile'r, Corbet (1956) reached somewhat different conciusions, and noted a change in the diet of some fishes (Mormyridae in particular). gome speci,.s with a specialized diet, such as Mastacembelus, 957" of whose consumption consists of lichophilous insecE larvae, even suffered from malnutrition or migrated. Taken ovcral1, the results reveaL the very important role played by hydrology in the biology and dynamics of many species. I'lost of the rivers investigated have a tropical regime typified by a well-defined annual high-water period. Many species reproduce during this period, and their young enjoy conditions more favourable to their survival, finding food and shelter on the bott.om land or flood plains, where the latter exist. Flooding is also beneficial to the.adult fishes, which find that food is more plentiful at this time, principally food of terrestrial origin (insects, the fruit.s of certain trees, vegetation, ete,). When the 1eve1s attained during the high-water period are Iow, as l^as particularly the case in 1976 (Fig. 2), the survival oE the young is threatened and the recruitment is in generirl poor, with lhe consequence that fishes are less abundant in the following year. Obviotrsly, in some cases there rnay be confusion between the effect of the failure of water leveIs to rise sufficiently during the high-water period and the possible role of a pollutant. No si.mpl" sPot analysis can be made of such a situation; what are needed are long-term studies which will establish whether the effects noted are merely transitory and the result of the 1ow volume of water in Lhe high-water period, or whether they are, on the contrary, lasting and due to the insecticide. The pirenomenon is still further complicated by the existence of long-term fluctuations in some populations, the precise causes of which are not always well known. Schilbe mystus in the Llraba is a particularly clear example in these resPects. This species, which was plentiful aL the cornmencement of Ehe observations in 1974, subsequently declined and disappeared almost completety after the high-water period of 1976, in wlrich the rise in level was slight, only to reappear in abundance at the end of 1979. This exampre ir; a good illustration of the difficulty of interpreting results, of the need to be cautious and to avoid any hasty interpretation concerning the possible effect of insecticides, and of the value of long-term studies in ichthyological research. Hydrology also has other consequences for sampling. Thus, some fishes make spawning migrations to the upper reach of rivers at the start of the high-nater period and may disappea from some stations at this time. On the other hand, sampling by gill net in rivers is not carried out under the same condltlons in the low-water and high-water periods. In the first case, the nets are set up ln pools of still water, and the catches provlde a Beneral idea of the populat.ton. During the high-water period, Ehe flow rate and the debris (br'rnches, dead leaves) transported by the wator restrict the sampling sites to a few calmer localities and the catches are less plentiful, equally due to the fact that the same number of fish are occupying a far larger volume of uater. I_25_ threshold. Moreover, it has been noted that, overa11, insecticidal treatments with temephos result in a temporary reduction of only 3O-4O7" in the stock of aquatic insects (Dejoux et al., 1979). Nor did analyses carried out in 1976 and 1977 reveal appreciable changes in the composition of the stomach contents relative,to 1975. A remark may be made here concerning Amphilius atesuensis. HaIf the diet of this species, which is associated with running water in which the flow rate is verv high, consists of Simulium larvae (l,leronar p.'rsonol conununica- tion). Despit.e the almost total disappearance of this target species in treated r./aters, we have not noted any special effect on Amphilius populations, which remain plentiful in the rapids. CONCLUS IONS We are 1ed to conclude, trom the result.s obtained over a period of five years in the course of monitoring rivers treated !,rith insecticides in the Ivorr Coast, that releases of temephos have had no detectable effect on the fish populations. We have not been able to establish any changes ascribable to the effects of the insecticide in the composition of the populations or in the size oI the catches in test fishing. Furthermore, nor would there seem to have been any appreciable allcration in the biology of the species or in their feeding habits during the period covered by the investigation. In studics of the effecLs of DDT in the ItVictoria Nile'r, Corbet (1956) reached somewhat different conciusions, and noted a change in the diet of some fishes (Mormyridae in particular). gome speci,.s with a specialized diet, such as Mastacembelus, 957" of whose consumption consists of lichophilous insecE larvae, even suffered from malnutrition or migrated. Taken ovcral1, the results reveaL the very important role played by hydrology in the biology and dynamics of many species. I'lost of the rivers investigated have a tropical regime typified by a well-defined annual high-water period. Many species reproduce during this period, and their young enjoy conditions more favourable to their survival, finding food and shelter on the bott.om land or flood plains, where the latter exist. Flooding is also beneficial to the.adult fishes, which find that food is more plentiful at this time, principally food of terrestrial origin (insects, the fruit.s of certain trees, vegetation, ete,). When the 1eve1s attained during the high-water period are Iow, as l^as particularly the case in 1976 (Fig. 2), the survival oE the young is threatened and the recruitment is in generirl poor, with lhe consequence that fishes are less abundant in the following year. Obviotrsly, in some cases there rnay be confusion between the effect of the failure of water leveIs to rise sufficiently during the high-water period and the possible role of a pollutant. No si.mpl" sPot analysis can be made of such a situation; what are needed are long-term studies which will establish whether the effects noted are merely transitory and the result of the 1ow volume of water in Lhe high-water period, or whether they are, on the contrary, lasting and due to the insecticide. The pirenomenon is still further complicated by the existence of long-term fluctuations in some populations, the precise causes of which are not always well known. Schilbe mystus in the Llraba is a particularly clear example in these resPects. This species, which was plentiful aL the cornmencement of Ehe observations in 1974, subsequently declined and disappeared almost completety after the high-water period of 1976, in wlrich the rise in level was slight, only to reappear in abundance at the end of 1979. This exampre ir; a good illustration of the difficulty of interpreting results, of the need to be cautious and to avoid any hasty interpretation concerning the possible effect of insecticides, and of the value of long-term studies in ichthyological research. Hydrology also has other consequences for sampling. Thus, some fishes make spawning migrations to the upper reach of rivers at the start of the high-nater period and may disappea from some stations at this time. On the other hand, sampling by gill net in rivers is not carried out under the same condltlons in the low-water and high-water periods. In the first case, the nets are set up ln pools of still water, and the catches provlde a Beneral idea of the populat.ton. During the high-water period, Ehe flow rate and the debris (br'rnches, dead leaves) transported by the wator restrict the sampling sites to a few calmer localities and the catches are less plentiful, equally due to the fact that the same number of fish are occupying a far larger volume of uater. -26_ Lle are able to arrive at sonn' conclusions on the methodology employed from our experience of these five years of fielrl work, and to make some proposals for future fish monitoring Programmes in West Africa. First and ftrremost, we shal1 stress the value oflong-term sttrdies. lnsecticides may be selected before their field use in relation to theireffectivener,s against the target fauna and to their low inrnediate toxicity against the non_target fauna. Such was the case for temephos. The point of monitoring is then to verify rShether rePeated releases over a long period may or may not have effects on the fish fauna notdetectable when tests are carried out (accumulaEion in tissues and physiological consequences,fot example). In other cases, when insecticides are known to be not without danger, it canLe seen frorn long-term studies whether the consequences of use i1 the natural environment are compatible with the overall ecological balance. InterPretaLion of monitoring results also requires a good knowledge of the biology and ecology of the species. Even if we now have adequate results in some basins, they must nevertheless be verified on proceeding to examine another river network, and it is most often necessary to suPplement them. Consequently, a special programme should be put in hand atthe start of observations. 1t is also essential to ask for observations at reference stations two years before theforeseeable coflmencement of treatment. This is done for several purposes: to have time toverifv the representative nature of these stations, and to acquire p."li*i.r".y data on thebiology of the species and the seasonal dynamics of the popurations. We therefore Propose the following methodology for subsequent monitorlng progranmes. (a) Sampling Lechniques The trse of fishing gear in the context of a fish monitori.g certain requirements: particularly a good representation of thequantification of the data, s;l-andardization, and ease of use andhydrological and morphological conditions. programme should sati_sfy indigenous f ish popuiations installation under various We do not crrrrently have any fishing gear that satisfies all the conditions. A11 suclr that is capable of exhaustive sampling and gear catches only certain species and can beused only in certain tyPes of environment. Thus, the drag seine, which is perhaps the leastselective fishing gear, can be used only on uniform bottoms (g.nerally sandy) and inrelatively calm water. rts use h/as coinpleteJ-y ruled out in the lvory coJst, wSere the rivershave rocky and stony bottoms cluttered with the trunks of dead trees. I"Ie have usedreconrnend the use of a set of gill nets with different types of mesh. Although thiswhich is often found i-n west Africa, is selective, it can be used in various types ofenvironment, except for rather shallow rivers and very fast-running waters. Complemenlary information may be obtained using different fishing gear. Iheis the easiest to use, and it yields good rer,u1ts, although these are difficult toElectrofishing is also a promising technique for r;ampling sharlow zones of runninginhabited by certain, generally smal1 species, but its use is limited and difficultthe quantification of the results is uncertain. (b) Site selection and gear, ca st- ne t quantify. I^Iater and The monitoring sites selected must be accessible in alltb main roads. These sites must also be representative oftion, and every operation should begin witlr a zonal researchor not the fish populations of the selected station arl wellthe river. It is less worthwhile to take a reach inhabited a reference. seasons, and as close as possible the environments under investiga- progranme to establish whether represented in a large part of by a partictrlar population as -26_ Lle are able to arrive at sonn' conclusions on the methodology employed from our experience of these five years of fielrl work, and to make some proposals for future fish monitoring Programmes in West Africa. First and ftrremost, we shal1 stress the value oflong-term sttrdies. lnsecticides may be selected before their field use in relation to theireffectivener,s against the target fauna and to their low inrnediate toxicity against the non_target fauna. Such was the case for temephos. The point of monitoring is then to verify rShether rePeated releases over a long period may or may not have effects on the fish fauna notdetectable when tests are carried out (accumulaEion in tissues and physiological consequences,fot example). In other cases, when insecticides are known to be not without danger, it canLe seen frorn long-term studies whether the consequences of use i1 the natural environment are compatible with the overall ecological balance. InterPretaLion of monitoring results also requires a good knowledge of the biology and ecology of the species. Even if we now have adequate results in some basins, they must nevertheless be verified on proceeding to examine another river network, and it is most often necessary to suPplement them. Consequently, a special programme should be put in hand atthe start of observations. 1t is also essential to ask for observations at reference stations two years before theforeseeable coflmencement of treatment. This is done for several purposes: to have time toverifv the representative nature of these stations, and to acquire p."li*i.r".y data on thebiology of the species and the seasonal dynamics of the popurations. We therefore Propose the following methodology for subsequent monitorlng progranmes. (a) Sampling Lechniques The trse of fishing gear in the context of a fish monitori.g certain requirements: particularly a good representation of thequantification of the data, s;l-andardization, and ease of use andhydrological and morphological conditions. programme should sati_sfy indigenous f ish popuiations installation under various We do not crrrrently have any fishing gear that satisfies all the conditions. A11 suclr that is capable of exhaustive sampling and gear catches only certain species and can beused only in certain tyPes of environment. Thus, the drag seine, which is perhaps the leastselective fishing gear, can be used only on uniform bottoms (g.nerally sandy) and inrelatively calm water. rts use h/as coinpleteJ-y ruled out in the lvory coJst, wSere the rivershave rocky and stony bottoms cluttered with the trunks of dead trees. I"Ie have usedreconrnend the use of a set of gill nets with different types of mesh. Although thiswhich is often found i-n west Africa, is selective, it can be used in various types ofenvironment, except for rather shallow rivers and very fast-running waters. Complemenlary information may be obtained using different fishing gear. Iheis the easiest to use, and it yields good rer,u1ts, although these are difficult toElectrofishing is also a promising technique for r;ampling sharlow zones of runninginhabited by certain, generally smal1 species, but its use is limited and difficultthe quantification of the results is uncertain. (b) Site selection and gear, ca st- ne t quantify. I^Iater and The monitoring sites selected must be accessible in alltb main roads. These sites must also be representative oftion, and every operation should begin witlr a zonal researchor not the fish populations of the selected station arl wellthe river. It is less worthwhile to take a reach inhabited a reference. seasons, and as close as possible the environments under investiga- progranme to establish whether represented in a large part of by a partictrlar population as _27 - (c) Frequency of observations We would recal1 that we regard a preliminary two-yeo.5l,rdy as essenEial before the comnencement of treatment. Having regard to the considerable life span of fishes and to the existence of an annual cycle, sampling should be carried out e\/ery two or three months during the initial years of monitoring, with the object of arrivirrg aE a better understatlding of possible changes in catches as a function of the seasonal cyc1e, and of achieving ctarity on the biology of t.he main species. Subsequently, we reconrnend sampling aL the end of the low-water period, which provides a status report on the populations after the dry season, and sampling at the end of the high-water period. The latter ls the most interesting, because of the possibility of sampling the young of annually-reproducing species and of verifying whether the recruitment is proceeding normal1y. Our experience leads us to propose that several other stations should also be sampled once a year, at the end of the high-water period, from the cormnencement of progranunes. These stations should preferably be located on either side of reference stations in the same cat(l)ment, and could be used to verify, in long-term research, that the results obtained at these reference stations were well representative of the river as a who1e. They would also have the advantage of enabling the reference station to be replaced, should it be altered as a consequence of road works or hydraulic engineeri_ng works. (d) Complementarv research It would be worthwhile making a deeper study of the effect of pollutants on the physiology of fishes. In 1'61.1i.,r1ar, it would be of interest to be able to recognize the concentration rates and the organs in which the insecticide is liable E.o accumulate. In conclusion, if an ichthyological monitoring prograrme is to be carried out seriously it ca1ls for the establishmenE of a pennanent team for a period of several years. Proper interPretation of the results also requires a deep knowledge of the environment and of the species involved, and is unobtainable from occasi.onal missions. Consequently, this type of programme cannot be tackled in a superficial manner, and concentration of effort seems far preferable to g,eographical dispersion. The need for in- depth studies has also been emphasized as regards invertebrates. _27 - (c) Frequency of observations We would recal1 that we regard a preliminary two-yeo.5l,rdy as essenEial before the comnencement of treatment. Having regard to the considerable life span of fishes and to the existence of an annual cycle, sampling should be carried out e\/ery two or three months during the initial years of monitoring, with the object of arrivirrg aE a better understatlding of possible changes in catches as a function of the seasonal cyc1e, and of achieving ctarity on the biology of t.he main species. Subsequently, we reconrnend sampling aL the end of the low-water period, which provides a status report on the populations after the dry season, and sampling at the end of the high-water period. The latter ls the most interesting, because of the possibility of sampling the young of annually-reproducing species and of verifying whether the recruitment is proceeding normal1y. Our experience leads us to propose that several other stations should also be sampled once a year, at the end of the high-water period, from the cormnencement of progranunes. These stations should preferably be located on either side of reference stations in the same cat(l)ment, and could be used to verify, in long-term research, that the results obtained at these reference stations were well representative of the river as a who1e. They would also have the advantage of enabling the reference station to be replaced, should it be altered as a consequence of road works or hydraulic engineeri_ng works. (d) Complementarv research It would be worthwhile making a deeper study of the effect of pollutants on the physiology of fishes. In 1'61.1i.,r1ar, it would be of interest to be able to recognize the concentration rates and the organs in which the insecticide is liable E.o accumulate. In conclusion, if an ichthyological monitoring prograrme is to be carried out seriously it ca1ls for the establishmenE of a pennanent team for a period of several years. Proper interPretation of the results also requires a deep knowledge of the environment and of the species involved, and is unobtainable from occasi.onal missions. Consequently, this type of programme cannot be tackled in a superficial manner, and concentration of effort seems far preferable to g,eographical dispersion. The need for in- depth studies has also been emphasized as regards invertebrates. -?"8- 5. BIBLIOGRA?HY ALBAREI, J' J' (1982) Reproduction et- f6condit6 des poissons d'eau douce de Cote d'IvoireRev. Hvdrobio. trop., 15 fl.BARET, J. J., MERONA, B. de, BIGORNE, R. & HERBINET,icirtyologique de la Maraou6 avant son traitement P. (1978) Observations sur 1a taune A 1'Abate, Rapp. ORSTOM, Bouak€, no 20, 118 p. multigr. o BENZECRI, J. P. (1C73) L'analyse des donndes. I. correspondances - Dunod, paris, 63I et 619 p. La taxinomie: II. L'Analyse des BLACHE, J' (1964) Les poissons du bassin du Tchad et du bassin adjacent du Mayo-Kebbi,M6m. ORSTOM, paris, 4: 493 coRBET, P' s. (1956) Some effects of simulium control by insecticirle on the feeding habitsof insectivorous fishes. Second symposium on African hydrobi.ololxr and inland fisheries,Brazzaville, 1956, CCTA: g1 DAGET' J' (1952) M6moires sur la biologie des poissons du Niger moyen, I: biologie de croissance des espEces du genre Alestesl Bul1. Inst. Fr. Afr. Noire, A, !4,1: lgl-225 DAGET, J. (1954) Les poissons tlu Niger supErieur, DAGET, J., PLANQUETTE, N. & PLANQUETTE, p. (1973) des peuplements de poissons du Bandama (C0te 3Bme s6r. 151, 7: L?9-L4Z M6m. lnst. Er. Afr. Noire, i6: 391 P. Premilres donn6es sur la dynamique d'Ivoire), Bul-1. Mus. Nat. Hist. Nat., DAGET, J, & ILTIS, A. (1905) poissons de Cdte d.Ivoire (eaux douces et saum6tres), M6m. IFAN, 74, 385 p. DANSOKO, D., BREHMAN, M. & DAGET, J. (1976) Influence de la s6cheresse sur les populationsd'Hydrocynus dans Ie Delta Central du Niger, Cah. ORSTOT'I, s6r. Hydrobiol., 10 (2): 7L-76 DEJoux, c., ELOUARD, J. M.,IEVEQIIE, c. & TROUBAT, J. J. (Lg7g) La lurre contre simuliumdamnosg en Afrique de 1'ouest et la protection du milieu aquatique; c. n]El-Eoo=e.e.de Marseille. s,rr la lutte contre les lnsectes en milieu tropical, II, 873-83 DEJoUx' c' & TRoUMT' J' J. (L976) Toxicit6 cornparEe de tleux insecticides organophosphor6s sur la faune aquatique non-cible en milieu tropical, Rapp. ORSTOM, Bouak6, f: i: p.multigr. DURAND' J. R. & LOUBENS, G. (1970) observations sur 1a sexualit6 et 1a reproduction des Ales tes *3#"r. du Bas-Charl et du lac Tchad, Cah. ORSTOI"I, s6r. Hydrobiol.,4 (2): IIILL, M. O. 340-354 62-81 (1,974) Correspondence analysis: a neglectecl multivariate method, App. Stat., 3: LLTIS, A. & LEVEQW, C. (1982) Caracr6risriques physico-chimlques des riviBres 15de Cdtre d'Ivoire, Rev. Hydroblol. Trop., IIAUZANNE, L. (1976) Regimes alimentalres et relations trophi.ques des poissons du lac Tchad, Cah. OR.STOM. s€r. Hvdrobiol., 10 (4): 26l-3LO -?"8- 5. BIBLIOGRA?HY ALBAREI, J' J' (1982) Reproduction et- f6condit6 des poissons d'eau douce de Cote d'IvoireRev. Hvdrobio. trop., 15 fl.BARET, J. J., MERONA, B. de, BIGORNE, R. & HERBINET,icirtyologique de la Maraou6 avant son traitement P. (1978) Observations sur 1a taune A 1'Abate, Rapp. ORSTOM, Bouak€, no 20, 118 p. multigr. o BENZECRI, J. P. (1C73) L'analyse des donndes. I. correspondances - Dunod, paris, 63I et 619 p. La taxinomie: II. L'Analyse des BLACHE, J' (1964) Les poissons du bassin du Tchad et du bassin adjacent du Mayo-Kebbi,M6m. ORSTOM, paris, 4: 493 coRBET, P' s. (1956) Some effects of simulium control by insecticirle on the feeding habitsof insectivorous fishes. Second symposium on African hydrobi.ololxr and inland fisheries,Brazzaville, 1956, CCTA: g1 DAGET' J' (1952) M6moires sur la biologie des poissons du Niger moyen, I: biologie de croissance des espEces du genre Alestesl Bul1. Inst. Fr. Afr. Noire, A, !4,1: lgl-225 DAGET, J. (1954) Les poissons tlu Niger supErieur, DAGET, J., PLANQUETTE, N. & PLANQUETTE, p. (1973) des peuplements de poissons du Bandama (C0te 3Bme s6r. 151, 7: L?9-L4Z M6m. lnst. Er. Afr. Noire, i6: 391 P. Premilres donn6es sur la dynamique d'Ivoire), Bul-1. Mus. Nat. Hist. Nat., DAGET, J, & ILTIS, A. (1905) poissons de Cdte d.Ivoire (eaux douces et saum6tres), M6m. IFAN, 74, 385 p. DANSOKO, D., BREHMAN, M. & DAGET, J. (1976) Influence de la s6cheresse sur les populationsd'Hydrocynus dans Ie Delta Central du Niger, Cah. ORSTOT'I, s6r. Hydrobiol., 10 (2): 7L-76 DEJoux, c., ELOUARD, J. M.,IEVEQIIE, c. & TROUBAT, J. J. (Lg7g) La lurre contre simuliumdamnosg en Afrique de 1'ouest et la protection du milieu aquatique; c. n]El-Eoo=e.e.de Marseille. s,rr la lutte contre les lnsectes en milieu tropical, II, 873-83 DEJoUx' c' & TRoUMT' J' J. (L976) Toxicit6 cornparEe de tleux insecticides organophosphor6s sur la faune aquatique non-cible en milieu tropical, Rapp. ORSTOM, Bouak6, f: i: p.multigr. DURAND' J. R. & LOUBENS, G. (1970) observations sur 1a sexualit6 et 1a reproduction des Ales tes *3#"r. du Bas-Charl et du lac Tchad, Cah. ORSTOI"I, s6r. Hydrobiol.,4 (2): IIILL, M. O. 340-354 62-81 (1,974) Correspondence analysis: a neglectecl multivariate method, App. Stat., 3: LLTIS, A. & LEVEQW, C. (1982) Caracr6risriques physico-chimlques des riviBres 15de Cdtre d'Ivoire, Rev. Hydroblol. Trop., IIAUZANNE, L. (1976) Regimes alimentalres et relations trophi.ques des poissons du lac Tchad, Cah. OR.STOM. s€r. Hvdrobiol., 10 (4): 26l-3LO _29- LEMRD' L. & FENELON, J. P. (1973) Statistique et informatique appliqu€es (2i:me €dition) Dunod, Paris, 457 p. LEVEQIIE, C., MERONA, B. de & PAUGY, D. (L977) Les peuplements ichryoLogiques du N'Zi avanr son traitement au chlorphoxlrm, Rapp. oRSToM. Bouak6, No. g, 143 p. multigr. LEWQITE, C., ODEI, M. & PUGH THOI'IAS, M. (1979) Ttre Onchocerciasis Control Prograrmne and the monitoring of its effects on the riverine blo1ory of the Volta River Basin. InI'Ecologtcal Effects of pesticidestr, F. H. Perrlng and MeLlanby 6d., Linnean Society Symposium series, no 5: 133-143 LEVEQUE, C. & PAUGY, D. (1,981) Prospection Hydrobiologique de 1a r6gion S6n6gambienne dans 1e cadre de 1'extension du programme de lutte contre 1'Onchocercose. II. Ichryofaune. Rapp. URSTCI'I Bouak6, no 40, 37 p. multigr. LEVEQIIE' C. & HERBIMT, P. (1980) Caract6ristiques m6ri-stiques et biologie des qqbi&g__g5.!_qf(Pisces, Schilbeidae) en Cdre d'Ivoire, Cah. ORSTOM, s6r. Hydrobiol., f: 1I-aJ: fOT:1ZO LEWQIIE, C. & HERBINET, P. (1982) Caract6ristiques m6ristiques et biologie des Eutro-pius mentalis (Pisces, Schilbeidae) en COte d,Ivoire, Bev. Zool-. afr. MASLIN-LENY, Y. & MERONA, B. de (1978) Evolution des peuplements ichtyologiques du N'Zidepuis son traitement au Ctrlorphoxym, Rapp. OR^STOM, Bouak6, No. 22, 103 p. multigr. MASLIN-LEI'ry, Y., ALBARET, J.J., BIGORM, R., HERBINET, P., LEVEQUE, C., MERQNA, B. de & PAUGY, D. (1978) Evolution des peuplements ichtyologiques du Como€ depuis son traitementA l-'Abate, Rapp. OR^STOI"[. Bouak€, no 23t 65 p. multigr. llERoNA, B. de (1979) rutfo..uph"1r" burr..orno.r,rit n. ssp. (Poisson: Mormyridae) du bassin du Como6 (C0te d'lvoire). Donn6es morphologiques. Cybium. 3Eme s6r., 7: 45-5L MERONA, B. de (198o) Ecologle et biologie de Petro-cephalus borrei (Polssons, Mormyridae) dans lesriviEres de Cdte d'Tvoire. Cah. ORSTOM. s6r. HydrobioL., L3 (3-4): LL7-LZ7 MERONA, B. de (1981) Zonation ichtyologique du bassin du Bandarua, Rev. Hydrob. Trop., 14 (1): 63-75 MERONA, B. de, ABBAN, E. K., HERBINET, P, & SAPE, E. R. (L979) Peuplements ichtyologiques des biotopes d'eaux peu profondes des riviEres du nord du Ghana, Rapp. ORSTOM. Bouak6, no 31, 23 p. multigr. MERONA, B. de & ALBARET, J. J. (1978) R6partition spatiale des poissons dans Ies radiers des riviEres de cOt.e d'rvoire, Rapp, oRSToM, tsouak€, no 17, 7g p. multigr. lmRONA, B. de, LEVEQLTE, C. & HERBIMT, P. (L977) obscrvations preliminaires sur les peuplements ichtyologiques des radiers. Rdsultats des p6ches 6lectriques affectu6es dans les stations du prograrmte de sunreillance de 1'environnement aquatique, Rapp. ORSTOM, Bouak6, no. 9, 51 p. multigr. MERoNA, B. de, LEVEQIIE, C., PAUGY, D., ALBARET, J. J., BIGoRNE, R. & I{ERBINET, P. (197S) Observations sur la faune lchtyologique du Bandama en amont du barrage de Kossou(COte d'Ivoire), Rapp. ORSTOM, Bouak6, no 2L, 193 p. multigr. PAUGY, D. (1978) EcoLogie et- bioLogie des Alestes baremoze (Pisces, Characidae) des riviEres de Cate d'Ivolre, Cah. oR^STOM. s€r. Hydrobiol., 12, 3-4: 245-275 PAUGY' D. (1980) Ecologie et bir:logie des Alestes nurse (!1sces, @-i{3g) des riviEresde COte d'Ivoire, Cah. ORSTOM. s6r. Hvdrobiol., L3 (3-4) 143-159 _29- LEMRD' L. & FENELON, J. P. (1973) Statistique et informatique appliqu€es (2i:me €dition) Dunod, Paris, 457 p. LEVEQIIE, C., MERONA, B. de & PAUGY, D. (L977) Les peuplements ichryoLogiques du N'Zi avanr son traitement au chlorphoxlrm, Rapp. oRSToM. Bouak6, No. g, 143 p. multigr. LEWQITE, C., ODEI, M. & PUGH THOI'IAS, M. (1979) Ttre Onchocerciasis Control Prograrmne and the monitoring of its effects on the riverine blo1ory of the Volta River Basin. InI'Ecologtcal Effects of pesticidestr, F. H. Perrlng and MeLlanby 6d., Linnean Society Symposium series, no 5: 133-143 LEVEQUE, C. & PAUGY, D. (1,981) Prospection Hydrobiologique de 1a r6gion S6n6gambienne dans 1e cadre de 1'extension du programme de lutte contre 1'Onchocercose. II. Ichryofaune. Rapp. URSTCI'I Bouak6, no 40, 37 p. multigr. LEVEQIIE' C. & HERBIMT, P. (1980) Caract6ristiques m6ri-stiques et biologie des qqbi&g__g5.!_qf(Pisces, Schilbeidae) en Cdre d'Ivoire, Cah. ORSTOM, s6r. Hydrobiol., f: 1I-aJ: fOT:1ZO LEWQIIE, C. & HERBINET, P. (1982) Caract6ristiques m6ristiques et biologie des Eutro-pius mentalis (Pisces, Schilbeidae) en COte d,Ivoire, Bev. Zool-. afr. MASLIN-LENY, Y. & MERONA, B. de (1978) Evolution des peuplements ichtyologiques du N'Zidepuis son traitement au Ctrlorphoxym, Rapp. OR^STOM, Bouak6, No. 22, 103 p. multigr. MASLIN-LEI'ry, Y., ALBARET, J.J., BIGORM, R., HERBINET, P., LEVEQUE, C., MERQNA, B. de & PAUGY, D. (1978) Evolution des peuplements ichtyologiques du Como€ depuis son traitementA l-'Abate, Rapp. OR^STOI"[. Bouak€, no 23t 65 p. multigr. llERoNA, B. de (1979) rutfo..uph"1r" burr..orno.r,rit n. ssp. (Poisson: Mormyridae) du bassin du Como6 (C0te d'lvoire). Donn6es morphologiques. Cybium. 3Eme s6r., 7: 45-5L MERONA, B. de (198o) Ecologle et biologie de Petro-cephalus borrei (Polssons, Mormyridae) dans lesriviEres de Cdte d'Tvoire. Cah. ORSTOM. s6r. HydrobioL., L3 (3-4): LL7-LZ7 MERONA, B. de (1981) Zonation ichtyologique du bassin du Bandarua, Rev. Hydrob. Trop., 14 (1): 63-75 MERONA, B. de, ABBAN, E. K., HERBINET, P, & SAPE, E. R. (L979) Peuplements ichtyologiques des biotopes d'eaux peu profondes des riviEres du nord du Ghana, Rapp. ORSTOM. Bouak6, no 31, 23 p. multigr. MERONA, B. de & ALBARET, J. J. (1978) R6partition spatiale des poissons dans Ies radiers des riviEres de cOt.e d'rvoire, Rapp, oRSToM, tsouak€, no 17, 7g p. multigr. lmRONA, B. de, LEVEQLTE, C. & HERBIMT, P. (L977) obscrvations preliminaires sur les peuplements ichtyologiques des radiers. Rdsultats des p6ches 6lectriques affectu6es dans les stations du prograrmte de sunreillance de 1'environnement aquatique, Rapp. ORSTOM, Bouak6, no. 9, 51 p. multigr. MERoNA, B. de, LEVEQIIE, C., PAUGY, D., ALBARET, J. J., BIGoRNE, R. & I{ERBINET, P. (197S) Observations sur la faune lchtyologique du Bandama en amont du barrage de Kossou(COte d'Ivoire), Rapp. ORSTOM, Bouak6, no 2L, 193 p. multigr. PAUGY, D. (1978) EcoLogie et- bioLogie des Alestes baremoze (Pisces, Characidae) des riviEres de Cate d'Ivolre, Cah. oR^STOM. s€r. Hydrobiol., 12, 3-4: 245-275 PAUGY' D. (1980) Ecologie et bir:logie des Alestes nurse (!1sces, @-i{3g) des riviEresde COte d'Ivoire, Cah. ORSTOM. s6r. Hvdrobiol., L3 (3-4) 143-159 -30 PAUGY, D. (1980) Ecologie et biologie de Cdte d'Ivoire. Comparaison Hydrobiol, 13 (3-4) | L29-L4L des Alestes imberl. (Pisces, qlgr""ida".) des riviEres --.--.--....--.. m6ristique avec A. nigricauda, Cah. ORSTOM. sdr. PAUGY, D. (1982) Synonymie d'Alestes rr.,jibe!.. B1gr. 1-9L6 avec A:g@.tus r (Valenciennes, 1849). Iliologie r:t variabllit6 morphologique, Itev. zoo1. Afr. (sous presse) ?Aucv, D., Brc,oRNE,'R., ALBARET, J. J., IIERBTNET, p., MASLTN-LE}.IY, y., LEvEeuE, c. & l{ERONA, B. de. (1979) Observatlons sur Ia faune ichtyologique du Sassandra (Cdte d'Ivoire) avant son traitenent a l'Abate, Rapp. ORSTOM, Bouak6, no 30, 106 p. multlgr. PAUCff, D. & LEVEQUE, C. (1977) Observations sur la faune ichtyologique du Bandama en aval du barrage de Kossou (Cdte d'Ivoire), Rapp. ORSTOM. Bouak6, no 12, 63 p. multigr. QUELENNEC, G., MILES, J. W., DEJOUX, C. & I"IERONA, B. de (1977) Chemical monitoring for temephos in mud oysters and fish from a ri.ver r.rithin the onchocerciasis control progranme in the Volta river basin area. WtOfvrc/ll.OAlz 6p REYNOLDS, J. D. (1973) Biolory and Characidae) in the new Volta fisheries potential of four Lake, Ghana, Rev. ZooL. Bot. species of Afr., 87, VIDY, G. (L976) Etude du r€gime alimentair:e de quelques poissons insectivores dans les riviEres de Cdte d'Ivoire. Recherche de f infLuence des traitements insecti-cides effectu€es di-rns le cadre de la lutte contre L'Onchocercose, Rapp. ORSTOM. Bouaki, No. 2, 36 p. multigr. WIiITEHEAD' V. (1969) Investigations into the food habits of some juvenile fish in the Volta Lake during the period October 1967 to March 1969, with scme notes on distribution and abundance. Volta Lak. Re : 16 p. muLtlgr Alestes (Pisces 2: 298-3Lo -30 PAUGY, D. (1980) Ecologie et biologie de Cdte d'Ivoire. Comparaison Hydrobiol, 13 (3-4) | L29-L4L des Alestes imberl. (Pisces, qlgr""ida".) des riviEres --.--.--....--.. m6ristique avec A. nigricauda, Cah. ORSTOM. sdr. PAUGY, D. (1982) Synonymie d'Alestes rr.,jibe!.. B1gr. 1-9L6 avec A:g@.tus r (Valenciennes, 1849). Iliologie r:t variabllit6 morphologique, Itev. zoo1. Afr. (sous presse) ?Aucv, D., Brc,oRNE,'R., ALBARET, J. J., IIERBTNET, p., MASLTN-LE}.IY, y., LEvEeuE, c. & l{ERONA, B. de. (1979) Observatlons sur Ia faune ichtyologique du Sassandra (Cdte d'Ivoire) avant son traitenent a l'Abate, Rapp. ORSTOM, Bouak6, no 30, 106 p. multlgr. PAUCff, D. & LEVEQUE, C. (1977) Observations sur la faune ichtyologique du Bandama en aval du barrage de Kossou (Cdte d'Ivoire), Rapp. ORSTOM. Bouak6, no 12, 63 p. multigr. QUELENNEC, G., MILES, J. W., DEJOUX, C. & I"IERONA, B. de (1977) Chemical monitoring for temephos in mud oysters and fish from a ri.ver r.rithin the onchocerciasis control progranme in the Volta river basin area. WtOfvrc/ll.OAlz 6p REYNOLDS, J. D. (1973) Biolory and Characidae) in the new Volta fisheries potential of four Lake, Ghana, Rev. ZooL. Bot. species of Afr., 87, VIDY, G. (L976) Etude du r€gime alimentair:e de quelques poissons insectivores dans les riviEres de Cdte d'Ivoire. Recherche de f infLuence des traitements insecti-cides effectu€es di-rns le cadre de la lutte contre L'Onchocercose, Rapp. ORSTOM. Bouaki, No. 2, 36 p. multigr. WIiITEHEAD' V. (1969) Investigations into the food habits of some juvenile fish in the Volta Lake during the period October 1967 to March 1969, with scme notes on distribution and abundance. Volta Lak. Re : 16 p. muLtlgr Alestes (Pisces 2: 298-3Lo 31 I'ART I I : 1. INTRODUCTION OBSI]RVA'IIONS ON }'1SH RIVERS IN NORTHERN POPUT,A'I'IONS IN ABATE TRF]A'TI']) CHANA CONTENTS 2. METHODS Page 32 32 JJ 33 34 34 34 35 35 36 i6 3. DATA ANALYSIS 4. 5. RESULTS AND DISCUSSIONS CATCH PER UNIT EFFORT OF (r) Schilbe mystus FISHING (b) Entropius niloticus and E. mentalis (c) Al es Les species (d) lJggd"rlr-g- species 6. 7. CONCLUSIONS FLUCTUATIONS IN COEFFICIF]NT OI' CONDITION 31 I'ART I I : 1. INTRODUCTION OBSI]RVA'IIONS ON }'1SH RIVERS IN NORTHERN POPUT,A'I'IONS IN ABATE TRF]A'TI']) CHANA CONTENTS 2. METHODS Page 32 32 JJ 33 34 34 34 35 35 36 i6 3. DATA ANALYSIS 4. 5. RESULTS AND DISCUSSIONS CATCH PER UNIT EFFORT OF (r) Schilbe mystus FISHING (b) Entropius niloticus and E. mentalis (c) Al es Les species (d) lJggd"rlr-g- species 6. 7. CONCLUSIONS FLUCTUATIONS IN COEFFICIF]NT OI' CONDITION 32 INTRODUCTION the Onchocerciasis Control programme (ocr,), t." *ri.i, used a larvicideAbate' to atrack rhe aquatic developmenrri'ra.go; oi tr-," try. TheocP has other sections including epidemiology Ioa redevelopmenr(Davies S. al , lgTg ) . . The weekly use of chemicals over 1g ooo kms of rivr:r make itimperative thitr some srrrveillance must be kept on the non-target fauna.particrrlarly the invertebrates and fish. rn rhis reporr, rre presentresults of the fish surveillance work in Ghana Monit.ring started in Augus t 1976 on the rivers oti and white Volta,and January 7977 on the Black Volta. The rast two rivers were arreadybeing treated with insecticide and the oti forlowed in March 1977.These rivers are wide savanna systems in northern Ghana and flow intothe Volta Lakc. Their ecology ias been studied hy Hopson (1964) andAbban (1976). The sites, .rs indicated on rig.1t were selecred becauseof (a) the existence of Simulium breeding sites, (b) access arall tirnes of year, and (") t,"l.,g the nearest riverine portions of themain rivers to tlre volta Lake. rt was anticipated that crata collecredat these points would serve as indications of oossible effects ofinsecticide on the lake itself, resulrins f;;r-;;;;;;r* appricationof Abate. The fact that these sites were also situated near villageswith large conrnercial fisheries has 1d to certain problems and maygo a long way to explain the low catches as compared to the stationsmonitored in the rvory coast. More recently, stations have been establishedon the Pru (September 197g) and on the Wawa (December 19gO) to gaininformation prior to the extension of the .cp control area. The world Healrh organization is crrrrently carrying out a progremmeof controlling populations of the biacktly, simulium damnosum. vector ofonchoeerciasis or river blindness. The n..f * ,- i-r, l ffa ly designated METHODS T:u::: et.a1 (1q78) standardized the protocol of fish monitoring :l:n^:1.. dgr ;."., rn"." ;;;;-;;;;-;;"';".;";;":hilr;':used. First Iy,the Catch Per Uni r rif fort (CPUE) was estimated regrrl ar.l y by the use ofa battery of gill nets of five different mesh sizes (15, 20, 25, 30an! ,aO *r. ): Th2 results were exnressecl as catch in term of nrrmbers an6 ,ergnE per: Io0 m surface area of net r)er night. I'or the resul ts inGhana' it must be mentioned that variaLions in protocrrr occ,rred at flu.a:I.rof sampling. Unril March 1971, a hali ser of nets was used(r 'e' )um surface area of each mesh size) and the recorded cpuE obtainedby extrapolatiorr- From then until october 1979, the half set was usedfor two nights. since then, a fulr set has been.sed for one or t'ronights. Secondly, a condition factqr was es{imated by the formula coefficientof condirion (K) = (weight x 1O')/Lengthr. An atiempr was also madeto estimate the fish fry population fronr catches made by a pair of "fish-drift, nets. These were 3m rong with a mesh size of 1.5 and arectangular opening at one end of a 40 x 7O cm frame. 32 INTRODUCTION the Onchocerciasis Control programme (ocr,), t." *ri.i, used a larvicideAbate' to atrack rhe aquatic developmenrri'ra.go; oi tr-," try. TheocP has other sections including epidemiology Ioa redevelopmenr(Davies S. al , lgTg ) . . The weekly use of chemicals over 1g ooo kms of rivr:r make itimperative thitr some srrrveillance must be kept on the non-target fauna.particrrlarly the invertebrates and fish. rn rhis reporr, rre presentresults of the fish surveillance work in Ghana Monit.ring started in Augus t 1976 on the rivers oti and white Volta,and January 7977 on the Black Volta. The rast two rivers were arreadybeing treated with insecticide and the oti forlowed in March 1977.These rivers are wide savanna systems in northern Ghana and flow intothe Volta Lakc. Their ecology ias been studied hy Hopson (1964) andAbban (1976). The sites, .rs indicated on rig.1t were selecred becauseof (a) the existence of Simulium breeding sites, (b) access arall tirnes of year, and (") t,"l.,g the nearest riverine portions of themain rivers to tlre volta Lake. rt was anticipated that crata collecredat these points would serve as indications of oossible effects ofinsecticide on the lake itself, resulrins f;;r-;;;;;;r* appricationof Abate. The fact that these sites were also situated near villageswith large conrnercial fisheries has 1d to certain problems and maygo a long way to explain the low catches as compared to the stationsmonitored in the rvory coast. More recently, stations have been establishedon the Pru (September 197g) and on the Wawa (December 19gO) to gaininformation prior to the extension of the .cp control area. The world Healrh organization is crrrrently carrying out a progremmeof controlling populations of the biacktly, simulium damnosum. vector ofonchoeerciasis or river blindness. The n..f * ,- i-r, l ffa ly designated METHODS T:u::: et.a1 (1q78) standardized the protocol of fish monitoring :l:n^:1.. dgr ;."., rn"." ;;;;-;;;;-;;"';".;";;":hilr;':used. First Iy,the Catch Per Uni r rif fort (CPUE) was estimated regrrl ar.l y by the use ofa battery of gill nets of five different mesh sizes (15, 20, 25, 30an! ,aO *r. ): Th2 results were exnressecl as catch in term of nrrmbers an6 ,ergnE per: Io0 m surface area of net r)er night. I'or the resul ts inGhana' it must be mentioned that variaLions in protocrrr occ,rred at flu.a:I.rof sampling. Unril March 1971, a hali ser of nets was used(r 'e' )um surface area of each mesh size) and the recorded cpuE obtainedby extrapolatiorr- From then until october 1979, the half set was usedfor two nights. since then, a fulr set has been.sed for one or t'ronights. Secondly, a condition factqr was es{imated by the formula coefficientof condirion (K) = (weight x 1O')/Lengthr. An atiempr was also madeto estimate the fish fry population fronr catches made by a pair of "fish-drift, nets. These were 3m rong with a mesh size of 1.5 and arectangular opening at one end of a 40 x 7O cm frame. 33 In addition to the routine been undertaken concerning the Abate on tire fish of the rivers study to e,:timate the immecliate carried ouL on the Oti with the and Samman 1981). monitoring ! s()me speci fi c s tudics havc estimation of the possihle impact of of Ghana. For example, a 48-hour elfect o[ Abate on fish catch,,s was first treatment of the river (Abban I The effect of the flood regime o1 rhc Ori and trlhite Volta on fistr catches was reported by Abban (1979). A further strrdy of the effect of Abate treatment on the fish catches, especially during the dry season is given in Abban .'nd Samman (in press). l,.rl.,)rarory studies oI the reaction of fish to Abate was reported ro the Ocl,in 1981 (Abban and Samman, unpublished).and the aspects of the biology of a ma.]or specio ELttropius mentalis was discussed by Abban (1978). An ichthyologist was appointed with the responsibility of OCp monitoring in March 1977. The introduction of this expertise is clear in the analysis of resultsr and in some of the multivariate statistics it is necessary to exclude data prior to this time. 3. DATA ANALYSIS Standardized coding forms werc used for monitoring throughotrr rhe OcP area, and facilitated the transfer of information for daL:r handling using computers, Apart from the analvses carried orrt by the monitoring teams, the Salford University data handling team was contracted ro assist in the handling and interpretaLion ot the information. The tvpe of statistics used, has been the subject of a series of reports to OCP (19/9, 1980, 1981). In vierv of tl.re variability of rhe resulrs, the main analyses used have involved graphical presentations for ease of visual interpretation. I'lre Slrannon-Weincr diversrLy indc.r ancl its component pat ts, species richness and evenness (Odtrm 1974) aide<l estimation of community change as did the technique of correspon(ionce analysis (a1so known as reciprocal averaging) (Renzecri, L973). 'l'his latter technique, whilst not being the most statistically sophisticated method and not including overall abrrndance, has a higir visrr,rl information content and sirnultaneous portrayal of site/months ancl species a11ows indic.;rtor species to be recognized. 4. RESULTS AND DISCUSSION Surveys oI the norrhern Ghana rivers have identified some 107 species of fish with the Pru containing 69 species alone. Each river was considered rich in fish fauna and mosr specic,s were widely distrihrrred. These observations were artributed ro rhe facr thar all the rivers were directly connected to the Volta lake wlrich'has many species. Also the lake contains many fishes which are legacies of many smal1 rivers which became partly or completely absorbed in the fornr;rLion of the lake. since the lake is relatively young (aborrt 18 years) most of the fishes have not become cornpletciy adapted to the lacustrine condition of tlre lake. Thus many species migrate up into the rivers under observation, to snawn dtrring the wet season. Lastly. being tropical rivers, the Volta basin rivers have a wide ranlie of habitars within a relatively short distance and thus there is provision for various ecological groups of fishes to be widely riistributeci. futropius mentalis, E. nilgti cI:, Alestes leuciscus, !S!:_1U" !ys!u9, Synodonrl. sorex and S.gambit.nrir r." pu.f,ro" tt,. n stEililnr ""a "laefydistributed specit,:; with S.ocellifer also connnon in the Oti. 33 In addition to the routine been undertaken concerning the Abate on tire fish of the rivers study to e,:timate the immecliate carried ouL on the Oti with the and Samman 1981). monitoring ! s()me speci fi c s tudics havc estimation of the possihle impact of of Ghana. For example, a 48-hour elfect o[ Abate on fish catch,,s was first treatment of the river (Abban I The effect of the flood regime o1 rhc Ori and trlhite Volta on fistr catches was reported by Abban (1979). A further strrdy of the effect of Abate treatment on the fish catches, especially during the dry season is given in Abban .'nd Samman (in press). l,.rl.,)rarory studies oI the reaction of fish to Abate was reported ro the Ocl,in 1981 (Abban and Samman, unpublished).and the aspects of the biology of a ma.]or specio ELttropius mentalis was discussed by Abban (1978). An ichthyologist was appointed with the responsibility of OCp monitoring in March 1977. The introduction of this expertise is clear in the analysis of resultsr and in some of the multivariate statistics it is necessary to exclude data prior to this time. 3. DATA ANALYSIS Standardized coding forms werc used for monitoring throughotrr rhe OcP area, and facilitated the transfer of information for daL:r handling using computers, Apart from the analvses carried orrt by the monitoring teams, the Salford University data handling team was contracted ro assist in the handling and interpretaLion ot the information. The tvpe of statistics used, has been the subject of a series of reports to OCP (19/9, 1980, 1981). In vierv of tl.re variability of rhe resulrs, the main analyses used have involved graphical presentations for ease of visual interpretation. I'lre Slrannon-Weincr diversrLy indc.r ancl its component pat ts, species richness and evenness (Odtrm 1974) aide<l estimation of community change as did the technique of correspon(ionce analysis (a1so known as reciprocal averaging) (Renzecri, L973). 'l'his latter technique, whilst not being the most statistically sophisticated method and not including overall abrrndance, has a higir visrr,rl information content and sirnultaneous portrayal of site/months ancl species a11ows indic.;rtor species to be recognized. 4. RESULTS AND DISCUSSION Surveys oI the norrhern Ghana rivers have identified some 107 species of fish with the Pru containing 69 species alone. Each river was considered rich in fish fauna and mosr specic,s were widely distrihrrred. These observations were artributed ro rhe facr thar all the rivers were directly connected to the Volta lake wlrich'has many species. Also the lake contains many fishes which are legacies of many smal1 rivers which became partly or completely absorbed in the fornr;rLion of the lake. since the lake is relatively young (aborrt 18 years) most of the fishes have not become cornpletciy adapted to the lacustrine condition of tlre lake. Thus many species migrate up into the rivers under observation, to snawn dtrring the wet season. Lastly. being tropical rivers, the Volta basin rivers have a wide ranlie of habitars within a relatively short distance and thus there is provision for various ecological groups of fishes to be widely riistributeci. futropius mentalis, E. nilgti cI:, Alestes leuciscus, !S!:_1U" !ys!u9, Synodonrl. sorex and S.gambit.nrir r." pu.f,ro" tt,. n stEililnr ""a "laefydistributed specit,:; with S.ocellifer also connnon in the Oti. 34 a a sampling of Lhese main rivers was carried orrt a;'proximately every month on tlre oti and every two months 'on the i./hire Volta, Bl ack volta and Pru. Nrrmbers of fish caught during sa,npling were low in c,,mparisonto the lvory Coast records and the intensive local fishing and differencesin the character of rivers in the trro countries must be borne in mind. . The correspondence analysis in !'ig.3 demonstrares somc separationint. weL and dry sc.ason on tlre white Volta wirlr tlre lale start of the 1a_ins in 111.7 and the dominancr: of Ales res baremose in Novembe r 197g,Delrg notable. A decrease in rhe n,mber of speii-eJ with inereasingrlver freigrrr . is noticeable with various lJlgqgrj]: spp. beingcnaracrerrstrc of the flood periods. For the Oia-(fig.4) there is aclear separation into wet and dry seasons, 1977 is again abnormal with a very high catch of Eutropius mentalis in July and Alestes macrolepidotus in Novembe.. 0*.,ges i.E*ru.,ity .ir,-,"triEli?T"u, reflect the srart of the frdod. There is a cleai rise in numbers of species recorded, corresponding to changes in personnel and thereforepre-trcatment coll.ections canlr,)t be easily compared. Tlre Bl ack Volta at Ramboi again shows a drop in numbers ;,.rd species at times of maximum f1ood, and the otlrer seasonal iuflrrencesare c1ear. 5. CATCH PER UNI'I' EIFORT OF !.ISIIING Tt is important to conment on (see tigs.5.6 and 7) catchesrecorded for the more dominant species in the basin: (") f-l'i-]1"'rr:r'r-: Tlri s f ish f rom I;ig.5 appears ro br. in rhc riversat) rlrrough the year. lr"wever, the cpuE was related to the riverlreight. G.ncral1y the cpuri of rlre specier increasecr immediately af tertl)e start of the wet season with rains in May/June and also during thereceding period of frood usually about Novernber/December. Theini ri al increase in cprE was attributed to the mi grarion of thespeci cs into the wi rh trre sta.r orrl::;:'"d:"1::";# l::::;::0.i1;:*,:;: ;i":;.i:.,"0was only due to the faet that gi11 nets can not be as r,ffectivelydeployed in high curre.t rivers. The apparently abnormar shift ofpeaks during 1979/_Bo was.mainly due to a corresponding shif t in rheflooding period of the rivers irig.Z). (b) Eutropius niloticus and E.mentalis: Tht'se two species generally appear and disapoear from the riversat Lhe same periods in relation to the flood regime of the rivers(except for r977 when F-mentalis dominated), ttre relationship beingsimilar ro thar or schir6Elil..rr.. rr is also apparent that E.nilotic,sconmences it" ,igrriT6i-T?o.--tri.-trtr. prior to E,mentar is;. n"s"a- nn--Lhe cpuE recorded for rire rwo species,'E. m.ntrffi. ;; ,;;;";""r"1argcrsclroo1sthanE.ni1oti"u".a"iuffio."",",abundanceof adul ts and -iuvenires-orE-iEnTrlis in the Llhi," ,oii" showed rhat arthe peak of the floods "n*eitte,,arlts are apparently not availabl,r(attributed to influence of flood on effective deployment of gi11 ners)the f ry r,r'ere most abundant. 34 a a sampling of Lhese main rivers was carried orrt a;'proximately every month on tlre oti and every two months 'on the i./hire Volta, Bl ack volta and Pru. Nrrmbers of fish caught during sa,npling were low in c,,mparisonto the lvory Coast records and the intensive local fishing and differencesin the character of rivers in the trro countries must be borne in mind. . The correspondence analysis in !'ig.3 demonstrares somc separationint. weL and dry sc.ason on tlre white Volta wirlr tlre lale start of the 1a_ins in 111.7 and the dominancr: of Ales res baremose in Novembe r 197g,Delrg notable. A decrease in rhe n,mber of speii-eJ with inereasingrlver freigrrr . is noticeable with various lJlgqgrj]: spp. beingcnaracrerrstrc of the flood periods. For the Oia-(fig.4) there is aclear separation into wet and dry seasons, 1977 is again abnormal with a very high catch of Eutropius mentalis in July and Alestes macrolepidotus in Novembe.. 0*.,ges i.E*ru.,ity .ir,-,"triEli?T"u, reflect the srart of the frdod. There is a cleai rise in numbers of species recorded, corresponding to changes in personnel and thereforepre-trcatment coll.ections canlr,)t be easily compared. Tlre Bl ack Volta at Ramboi again shows a drop in numbers ;,.rd species at times of maximum f1ood, and the otlrer seasonal iuflrrencesare c1ear. 5. CATCH PER UNI'I' EIFORT OF !.ISIIING Tt is important to conment on (see tigs.5.6 and 7) catchesrecorded for the more dominant species in the basin: (") f-l'i-]1"'rr:r'r-: Tlri s f ish f rom I;ig.5 appears ro br. in rhc riversat) rlrrough the year. lr"wever, the cpuE was related to the riverlreight. G.ncral1y the cpuri of rlre specier increasecr immediately af tertl)e start of the wet season with rains in May/June and also during thereceding period of frood usually about Novernber/December. Theini ri al increase in cprE was attributed to the mi grarion of thespeci cs into the wi rh trre sta.r orrl::;:'"d:"1::";# l::::;::0.i1;:*,:;: ;i":;.i:.,"0was only due to the faet that gi11 nets can not be as r,ffectivelydeployed in high curre.t rivers. The apparently abnormar shift ofpeaks during 1979/_Bo was.mainly due to a corresponding shif t in rheflooding period of the rivers irig.Z). (b) Eutropius niloticus and E.mentalis: Tht'se two species generally appear and disapoear from the riversat Lhe same periods in relation to the flood regime of the rivers(except for r977 when F-mentalis dominated), ttre relationship beingsimilar ro thar or schir6Elil..rr.. rr is also apparent that E.nilotic,sconmences it" ,igrriT6i-T?o.--tri.-trtr. prior to E,mentar is;. n"s"a- nn--Lhe cpuE recorded for rire rwo species,'E. m.ntrffi. ;; ,;;;";""r"1argcrsclroo1sthanE.ni1oti"u".a"iuffio."",",abundanceof adul ts and -iuvenires-orE-iEnTrlis in the Llhi," ,oii" showed rhat arthe peak of the floods "n*eitte,,arlts are apparently not availabl,r(attributed to influence of flood on effective deployment of gi11 ners)the f ry r,r'ere most abundant. 35 t (c) Alestes species: A.nrtrse and A.macrolepidotus: The species o.cur most in catches durinf tltelecedi, -le ,rerioa-"f tfte f 1ood, when river height is below 5 ftr'ditlr tlre p(,ak catcites of A.macrolcpjrlotrts following that of A.ntrrse in time. The data ,lso-su![est-ttii-ttrgt catches of A.nurse n,[ UJ-o.,"urring in alternate years. A. lettciscrts and A.luterts: f'hese are Ltrro sma1.l and almost identical rpu. i "r Ai-a .i" "r.,gtl t u1r*i-"xc1us ive I y by the sma 1 Les t mesh of thebattery of gi11 nets (i.e. 15nrn.). The peak catches of these occur between .Ianrrary and March (dry season). The high catches recorded during 0ctober ar.rd November 1979 was because the duration of the flood during that ye;rr was so short that by Novomber/December the rivers were in their dry : ':ason state. (d) .Synodonris species Svnodontis sorex, S.oceilifer and S.gambiensis all migrate from the I,rkJ-into tt* ri""rr to rp*.r after-tbe-Schilbeidae and their peak catches occur when the rivers are high (above 5.5 ft). The ability of the gi11 nets to catclr the Synodontis when most species were aPDarently unavailable lras been attribuLed in part to the spiny nature of the species River Pnr Irr rhe Pru, which is yet to be treated, Alestes macr()lepidotrrs and sclrilb_t'Ty-lrrs lrave been co.siclerecl th.,ffi- in t.rms of regularity in eatch and numbers of individuals caught. The i)ru has a sligh L 1.,7 rlelayed f looding period compared to Lhe rivers in northern Gl.rana. The reriods of high catctres of various fishes correspond with similar hydrological srates in ttre otlrer rivers. During 1979 when rhe flooding of Pru lasted longer than in 1978 and 198O, the peak cat<'hes of S.mystus occurred in January and February of 1980. Fig.8 again shows separ.rtitrn of wet and dry seasons althougl-r quantification of discharge was impossible, due to the at,sence of a water gauge. s.mystrrs is seen to be characteristic of high water 1evels (November, December) and Alestes macrole_pidotus of 1c'w water (February). An examp1eortne@nceanalysis,isinNovember1978 when only Sync,Llontis gambienlis was found. In conclusion, it could be said fluctuations., which can not be fu1ly of the rivers was the most important relation to changes in CPUE. Since flexible in time, rlight changes in of years must be expected. that in addition to narural explained, tire flooding regime variable to be considered in the flooding regime of rivers is peak catch periods over a number 35 t (c) Alestes species: A.nrtrse and A.macrolepidotus: The species o.cur most in catches durinf tltelecedi, -le ,rerioa-"f tfte f 1ood, when river height is below 5 ftr'ditlr tlre p(,ak catcites of A.macrolcpjrlotrts following that of A.ntrrse in time. The data ,lso-su![est-ttii-ttrgt catches of A.nurse n,[ UJ-o.,"urring in alternate years. A. lettciscrts and A.luterts: f'hese are Ltrro sma1.l and almost identical rpu. i "r Ai-a .i" "r.,gtl t u1r*i-"xc1us ive I y by the sma 1 Les t mesh of thebattery of gi11 nets (i.e. 15nrn.). The peak catches of these occur between .Ianrrary and March (dry season). The high catches recorded during 0ctober ar.rd November 1979 was because the duration of the flood during that ye;rr was so short that by Novomber/December the rivers were in their dry : ':ason state. (d) .Synodonris species Svnodontis sorex, S.oceilifer and S.gambiensis all migrate from the I,rkJ-into tt* ri""rr to rp*.r after-tbe-Schilbeidae and their peak catches occur when the rivers are high (above 5.5 ft). The ability of the gi11 nets to catclr the Synodontis when most species were aPDarently unavailable lras been attribuLed in part to the spiny nature of the species River Pnr Irr rhe Pru, which is yet to be treated, Alestes macr()lepidotrrs and sclrilb_t'Ty-lrrs lrave been co.siclerecl th.,ffi- in t.rms of regularity in eatch and numbers of individuals caught. The i)ru has a sligh L 1.,7 rlelayed f looding period compared to Lhe rivers in northern Gl.rana. The reriods of high catctres of various fishes correspond with similar hydrological srates in ttre otlrer rivers. During 1979 when rhe flooding of Pru lasted longer than in 1978 and 198O, the peak cat<'hes of S.mystus occurred in January and February of 1980. Fig.8 again shows separ.rtitrn of wet and dry seasons althougl-r quantification of discharge was impossible, due to the at,sence of a water gauge. s.mystrrs is seen to be characteristic of high water 1evels (November, December) and Alestes macrole_pidotus of 1c'w water (February). An examp1eortne@nceanalysis,isinNovember1978 when only Sync,Llontis gambienlis was found. In conclusion, it could be said fluctuations., which can not be fu1ly of the rivers was the most important relation to changes in CPUE. Since flexible in time, rlight changes in of years must be expected. that in addition to narural explained, tire flooding regime variable to be considered in the flooding regime of rivers is peak catch periods over a number 36 FLUCTUATIONS IN COEFFICIDNT OT CONDITION The coefficient of condition of fish (K), generally thought of as a constant for a species is used to express the state of health of a species in numerical terms. Some variations in the condition of afish is, howcver, expecteri in connexion with the rcproductive cyclc offish. This is because the process of gon:d developrnent, especialiy inthe females, is usually associated with the accumulation of fats inthe mr-rscle and around the viscera. A1so, periods.f abundance or other- wise of preferred food materials coul<I contribute to slight fluctrrationsin the condition of fish. Tn the present work the condition of fisheshas been estimated to give indications of deviation or evol,tion, whi ch might be attributed to the use of Abare in r'lre rivers . Thefluctuations in K which appear in the acoompanying diagrarns (rig.9) were attributed to natural parameters. There appeared to be 1itt1e variation i, K with length of fisn except where numbers were Iow. Thetype of analysis used, involved a logarithmic transformation of the data,to stabilize the variance, enabling a greater size range to treincluded (Salford report lgTg) . Examples of fluctuations <;f K in differentfish groups as shown in Fig.9 is explained as follows: Synodontis c.ambiensis on the I{hite Volta (9) shows the effect of sp"roning on tt "-"n"altTon of Lhe speiias. This f ish spawns cluringthe middle of the flood (i.e. July to mid-Augrrst) and has lower coefficient from ahout September unti1 about l-ebr-uary the fo1lowing year, when the fish starts developing for the next spawning season. Aleitqs racrolepidotus which spawns d.ring the later part of the r-1oorlperiod shows an increasing condition rrom about the r,,_rk of the flood(Ju1y/August) ti11 spawning period (November,/oecemhei) and the coefficientis lower until July/August again. Ifutropius mentalis spawns about themidd1eofthef]oo<lseason-asshoffiles(,.,t:.rlna] abundance of the adults and juveniles in the trhite Volta - (Abban, rAB report 1978). Thus the condition of the fish apparently increasesduring the early part of the flood period when fat is being accumulatedfor spawning. CONCLUSlONS The effect of any chemical used as Abate is appiied in the rivers of the Volta basin could affect fish and fish popr.riations in one of, or a combination of, the following lrays, Firstly, the cheuricals couldaffect the survival of both adults andjuvenire risnes. secondly, thedistribution of fish populations could be affectecl, which in thl longterm also alters the fish population st.rcture of entire rivers (Abban and Samman, in press). ThircJly, Lhe corrr inuorrs application .f a chemical in a river system could cause changes in the coefficient of condition of the fishes. This could happen if the chemical adversely af fects fish foorl materials in the water or puts the fish in a state which will divert the atEention of fish frc,nr food. Lastly,the possibility of the accumulation of the chemical or its breakdownproducts in various tissues could affecr the generar physiology aodin parEicular the reproductive activities of tlu fi.f,. 36 FLUCTUATIONS IN COEFFICIDNT OT CONDITION The coefficient of condition of fish (K), generally thought of as a constant for a species is used to express the state of health of a species in numerical terms. Some variations in the condition of afish is, howcver, expecteri in connexion with the rcproductive cyclc offish. This is because the process of gon:d developrnent, especialiy inthe females, is usually associated with the accumulation of fats inthe mr-rscle and around the viscera. A1so, periods.f abundance or other- wise of preferred food materials coul<I contribute to slight fluctrrationsin the condition of fish. Tn the present work the condition of fisheshas been estimated to give indications of deviation or evol,tion, whi ch might be attributed to the use of Abare in r'lre rivers . Thefluctuations in K which appear in the acoompanying diagrarns (rig.9) were attributed to natural parameters. There appeared to be 1itt1e variation i, K with length of fisn except where numbers were Iow. Thetype of analysis used, involved a logarithmic transformation of the data,to stabilize the variance, enabling a greater size range to treincluded (Salford report lgTg) . Examples of fluctuations <;f K in differentfish groups as shown in Fig.9 is explained as follows: Synodontis c.ambiensis on the I{hite Volta (9) shows the effect of sp"roning on tt "-"n"altTon of Lhe speiias. This f ish spawns cluringthe middle of the flood (i.e. July to mid-Augrrst) and has lower coefficient from ahout September unti1 about l-ebr-uary the fo1lowing year, when the fish starts developing for the next spawning season. Aleitqs racrolepidotus which spawns d.ring the later part of the r-1oorlperiod shows an increasing condition rrom about the r,,_rk of the flood(Ju1y/August) ti11 spawning period (November,/oecemhei) and the coefficientis lower until July/August again. Ifutropius mentalis spawns about themidd1eofthef]oo<lseason-asshoffiles(,.,t:.rlna] abundance of the adults and juveniles in the trhite Volta - (Abban, rAB report 1978). Thus the condition of the fish apparently increasesduring the early part of the flood period when fat is being accumulatedfor spawning. CONCLUSlONS The effect of any chemical used as Abate is appiied in the rivers of the Volta basin could affect fish and fish popr.riations in one of, or a combination of, the following lrays, Firstly, the cheuricals couldaffect the survival of both adults andjuvenire risnes. secondly, thedistribution of fish populations could be affectecl, which in thl longterm also alters the fish population st.rcture of entire rivers (Abban and Samman, in press). ThircJly, Lhe corrr inuorrs application .f a chemical in a river system could cause changes in the coefficient of condition of the fishes. This could happen if the chemical adversely af fects fish foorl materials in the water or puts the fish in a state which will divert the atEention of fish frc,nr food. Lastly,the possibility of the accumulation of the chemical or its breakdownproducts in various tissues could affecr the generar physiology aodin parEicular the reproductive activities of tlu fi.f,. 37 I In the routine monitoring attempts were made to collect samples at given intervals from treatme'nt days. How,,ver, in practice treatment schedules varied and some samptes were taken on treatmenr tlays. This could explain some of the variations in Catch Per llnit Effort although no clear pattern emerges because the hydrological conclirions were varied. Secondly, no 1o'rg term changes in the condition of fish emerges from the interpretation of data. lndeed, if there rr,as a trend it would be ditficutr to separate the role of Abate from continued evolution resulting from the development ot the Volfa lake and its numerous influences on the fisheries of the tributary rivers. The temporary changes in the catch composition at the sampling sites can not be said to have been due to the apolicarion of Abare, and no species could be said to have been eliminated or otherwise. However, a regular study on representative rivers to monitor fish distribution may have to be undertaken - perhaps once every two or three years. The dominant patterl'r is seasonal, dependant on the hydrological regime and the presence of many species migrating from the lake. Differences between year may often be great although explanations can often be found in annual variations in river depth and the timing of the collections. 37 I In the routine monitoring attempts were made to collect samples at given intervals from treatme'nt days. How,,ver, in practice treatment schedules varied and some samptes were taken on treatmenr tlays. This could explain some of the variations in Catch Per llnit Effort although no clear pattern emerges because the hydrological conclirions were varied. Secondly, no 1o'rg term changes in the condition of fish emerges from the interpretation of data. lndeed, if there rr,as a trend it would be ditficutr to separate the role of Abate from continued evolution resulting from the development ot the Volfa lake and its numerous influences on the fisheries of the tributary rivers. The temporary changes in the catch composition at the sampling sites can not be said to have been due to the apolicarion of Abare, and no species could be said to have been eliminated or otherwise. However, a regular study on representative rivers to monitor fish distribution may have to be undertaken - perhaps once every two or three years. The dominant patterl'r is seasonal, dependant on the hydrological regime and the presence of many species migrating from the lake. Differences between year may often be great although explanations can often be found in annual variations in river depth and the timing of the collections. -38 -D'--T- - - 17t- DAI]OYA 1\ raulu sa\anr I t T'BENDE FI(;.I - I,OCATION OF MONIl'()RINC STATIONS IN GHANA a a t o 0,J ! t I I t \ t t t t \ I I t I t I\ t I t ( I I\ ) I I I I t 1 I () 5 ) o I , t t I) I I I I I \ I I -38 -D'--T- - - 17t- DAI]OYA 1\ raulu sa\anr I t T'BENDE FI(;.I - I,OCATION OF MONIl'()RINC STATIONS IN GHANA a a t o 0,J ! t I I t \ t t t t \ I I t I t I\ t I t ( I I\ ) I I I I t 1 I () 5 ) o I , t t I) I I I I I \ I I o @ c -d F-c oz FJ o t!F =t! F z It Hc XF JF,O ,r.lL!Ii + :l L] H& I N (i l& a G] co@ocaNN -; !t\ f! o @ c -d F-c oz FJ o t!F =t! F z It Hc XF JF,O ,r.lL!Ii + :l L] H& I N (i l& a G] co@ocaNN -; !t\ f! _40 t rr. ct x rs (RFE -i .tE <r ol, }{ !ae0$ 0/ ql -CtJ!5LU3t .a@!c Iar\ paSr'F t<' <r tf fie ,Eq A -'F<; "/a, c) a a J LL L) 7akl< oo c./') r! lll o(F&t< O:EU3 lttt EOJ I -] .J & rlJ F43 Jd ;T c1 t! J mK &*5 + ev tI* <E {t lD iE,o H noa{ €xF J' s( _40 t rr. ct x rs (RFE -i .tE <r ol, }{ !ae0$ 0/ ql -CtJ!5LU3t .a@!c Iar\ paSr'F t<' <r tf fie ,Eq A -'F<; "/a, c) a a J LL L) 7akl< oo c./') r! lll o(F&t< O:EU3 lttt EOJ I -] .J & rlJ F43 Jd ;T c1 t! J mK &*5 + ev tI* <E {t lD iE,o H noa{ €xF J' s( xA.n t a .a o r.3 CORNF,SPONDENCE ANAI,YSIS ON OTi A Al?l ,.1r5r arln A" I - urcu ulrrr A - r.ou t.tlrr.n .'r.r tr r.b e .?, 'a. rltA Paor cg -l(Tl'' 'rpr,? a aatt .^_ T rr .toa 4l,&,tl4ll afii An art r5* r?rr I' 5rr $JDA ,o, x '5 rOtl *U *A tS.V xA.n t a .a o r.3 CORNF,SPONDENCE ANAI,YSIS ON OTi A Al?l ,.1r5r arln A" I - urcu ulrrr A - r.ou t.tlrr.n .'r.r tr r.b e .?, 'a. rltA Paor cg -l(Tl'' 'rpr,? a aatt .^_ T rr .toa 4l,&,tl4ll afii An art r5* r?rr I' 5rr $JDA ,o, x '5 rOtl *U *A tS.V 42 Schi lbe mystus - .a 6(x) 400 200 200 1 000 500 I 500 2 500 5()o 1978 1979 19Bo 1981 t976 1977 t978 1979 C.P.U.E. (wt-g) OIr MAJOR FISHES ON THE WHtTLt Vor.TA 200 4()0 200 1916 1977 r-IG.5 - o : 42 Schi lbe mystus - .a 6(x) 400 200 200 1 000 500 I 500 2 500 5()o 1978 1979 19Bo 1981 t976 1977 t978 1979 C.P.U.E. (wt-g) OIr MAJOR FISHES ON THE WHtTLt Vor.TA 200 4()0 200 1916 1977 r-IG.5 - o : 41 t97 6 FIG.6 - ON 1977 C.P.ll .E.p r97 8 OF AI,ESTES I,EUCISCUS WHITE VOI,TA o'r I 1980 41 t97 6 FIG.6 - ON 1977 C.P.ll .E.p r97 8 OF AI,ESTES I,EUCISCUS WHITE VOI,TA o'r I 1980 -44 -a .a pd o FJ c tI I F z a A ot 0 o 6 0 o o t E' a o c a -44 -a .a pd o FJ c tI I F z a A ot 0 o 6 0 o o t E' a o c a _ tl' & ra * *(.f 'J6 7- o 6 a >.1 ,&l(J z Hnzo!"5ndhl q.x/,a()z(J< I I(J J\ F s<f :€ t<l! *"t <<3i ,tl rd x & -f c B(( ot$*tJx I HJo _ tl' & ra * *(.f 'J6 7- o 6 a >.1 ,&l(J z Hnzo!"5ndhl q.x/,a()z(J< I I(J J\ F s<f :€ t<l! *"t <<3i ,tl rd x & -f c B(( ot$*tJx I HJo Alestes mcrolepidotut flucruArroNs nr corirrdfrxf oF co[DrrroN Svaodon t i genbi enr i I 1 1 'l t \ l,\r\l\rlt A ,\t i I I t I I I ,\ ,\ Eutrop i us mental i s Alestes mcrolepidotut flucruArroNs nr corirrdfrxf oF co[DrrroN Svaodon t i genbi enr i I 1 1 'l t \ l,\r\l\rlt A ,\t i I I t I I I ,\ ,\ Eutrop i us mental i s 47 'a t a a : 1.1 1.2 1.3 1.4 ') 2.L ?? 2.3 3. PART III : The effects of Problems in the insecticides on int e rpre t at ion fish in West of results FISII }'.)NITORING tN WIST AFRICA}'I R1VERS PROBLEMS AND PERSPECTIVES CONTENTS Page Africa 48 49 49 50 50 50 50 51 51 RECOMMNDA] IONS Monitoring sites Cont inui ty Collaborat ion Methodo logy Calendar of CONCLUS IONS FOR FUTURE PROGMMMES Operat i ons 47 'a t a a : 1.1 1.2 1.3 1.4 ') 2.L ?? 2.3 3. PART III : The effects of Problems in the insecticides on int e rpre t at ion fish in West of results FISII }'.)NITORING tN WIST AFRICA}'I R1VERS PROBLEMS AND PERSPECTIVES CONTENTS Page Africa 48 49 49 50 50 50 50 51 51 RECOMMNDA] IONS Monitoring sites Cont inui ty Collaborat ion Methodo logy Calendar of CONCLUS IONS FOR FUTURE PROGMMMES Operat i ons 48 For reasons of economic importance, fish populations h,"ve beeo the ob,jecr of particular study within the framework of aquatie srrrveillance, lt is also inporrant for psychological reasons to show the villagers mainly occupied in fishing, tbat care is being taken about the risks of pollurion. 'l 't During work carried out since 1974 by t,,,rms based in Ivory Coast and Ghana, the conclusion is that.Temephos as an anti-Simulium larvicide. has no rletecrable effect. in the long-terrlr , on the composition and stocki-iF-fis_t, This is, of corrrse. based on the methods used and does not include chemical analysis. Furthermore, effects have not been seen in srudies on the biology of principal species, notably fecundity and reproduction. It is obvious that dietary regimes will be affected to some extent. but the insectivorous species apPear to have not been affected by the destruction of part of the entonological populations. The scientific results obtained in the course of monitoring fisir oopr:lations are in reports from the dilferenr teams, presented The experience acquired and the knowledge we now have of rhe problems of acquisition and treatment of the data, enable us to present cc,nstructive recorrlnen<iations for futureprogrames of fish monitoring. These could be continuation of existing progranrnes. extension of OCP areas or where new insecticides are introduced. 1.1 The effects of insecricides on fish in West Africa Insecticides mav have rapid, medium or long-term effecrs. therefore, must be adapted to the different possibilities. 'Ihe nroni tgring methods, lrr the case of a rapid effect of insecticide on the fish, a drop in catch per rrrrit effort will be noticed in a short space of time. Such observations may be confirmed by fish mortality following treatment. This eventuality ought not to occur if tests are carried outin vitro followed by field testing prior to utilization in a progranrne. Such preliminary experiences are very importantr Lrere carried out in the case of Abate, and should be carried out for new insecticides For medium- and long-term effects, the impact of an insecricide may be low and indirecr. Accumulation of roxic products may modify the metabolism of the fish (respirarionr assimilation nutrition and reprodrrction). In the last case dimunition of fecuncliry will be detecred bv regular estimation during periods of reproduction, However, it is important to executein vitro tests on 1oca1 species before regular field use, in order to know which organs are affected and whar quantity of insecticide accumulates. Furthermore. it is necessary to makethe same observation on natural populations at the start of the campaiBn to see if thereis anv difference from laboratory resr.rlts. Ir must be underlined that different species will react in different ways, according to their behaviour. Such studies have only been partially carried out for Abate. The toxicity of the insecticide will be most pronounced in juveniles and pc'1agic species with smal1 and abunrianr eggs. This effect will be mosr clear in the catches of juveniles in the nets of the smallest mesh size and at the enri of the flood season when most of the populations repr,.luce. The insecticide may not affect the fish rlirectly, but destroy an important part of 'tfreir diet. Some species are able to modify their food choice if an alternative is available. -Irl contrasr, more specialized fish may srrffgr malnutrition. The coefficient of condition will agt as an indicator of this. rn conclusi.n, we report the necessity of key preliminarlr strrriiesThis will lead to clarification of the direction of research and of the in laboratory and field. survey methodoloEly, 48 For reasons of economic importance, fish populations h,"ve beeo the ob,jecr of particular study within the framework of aquatie srrrveillance, lt is also inporrant for psychological reasons to show the villagers mainly occupied in fishing, tbat care is being taken about the risks of pollurion. 'l 't During work carried out since 1974 by t,,,rms based in Ivory Coast and Ghana, the conclusion is that.Temephos as an anti-Simulium larvicide. has no rletecrable effect. in the long-terrlr , on the composition and stocki-iF-fis_t, This is, of corrrse. based on the methods used and does not include chemical analysis. Furthermore, effects have not been seen in srudies on the biology of principal species, notably fecundity and reproduction. It is obvious that dietary regimes will be affected to some extent. but the insectivorous species apPear to have not been affected by the destruction of part of the entonological populations. The scientific results obtained in the course of monitoring fisir oopr:lations are in reports from the dilferenr teams, presented The experience acquired and the knowledge we now have of rhe problems of acquisition and treatment of the data, enable us to present cc,nstructive recorrlnen<iations for futureprogrames of fish monitoring. These could be continuation of existing progranrnes. extension of OCP areas or where new insecticides are introduced. 1.1 The effects of insecricides on fish in West Africa Insecticides mav have rapid, medium or long-term effecrs. therefore, must be adapted to the different possibilities. 'Ihe nroni tgring methods, lrr the case of a rapid effect of insecticide on the fish, a drop in catch per rrrrit effort will be noticed in a short space of time. Such observations may be confirmed by fish mortality following treatment. This eventuality ought not to occur if tests are carried outin vitro followed by field testing prior to utilization in a progranrne. Such preliminary experiences are very importantr Lrere carried out in the case of Abate, and should be carried out for new insecticides For medium- and long-term effects, the impact of an insecricide may be low and indirecr. Accumulation of roxic products may modify the metabolism of the fish (respirarionr assimilation nutrition and reprodrrction). In the last case dimunition of fecuncliry will be detecred bv regular estimation during periods of reproduction, However, it is important to executein vitro tests on 1oca1 species before regular field use, in order to know which organs are affected and whar quantity of insecticide accumulates. Furthermore. it is necessary to makethe same observation on natural populations at the start of the campaiBn to see if thereis anv difference from laboratory resr.rlts. Ir must be underlined that different species will react in different ways, according to their behaviour. Such studies have only been partially carried out for Abate. The toxicity of the insecticide will be most pronounced in juveniles and pc'1agic species with smal1 and abunrianr eggs. This effect will be mosr clear in the catches of juveniles in the nets of the smallest mesh size and at the enri of the flood season when most of the populations repr,.luce. The insecticide may not affect the fish rlirectly, but destroy an important part of 'tfreir diet. Some species are able to modify their food choice if an alternative is available. -Irl contrasr, more specialized fish may srrffgr malnutrition. The coefficient of condition will agt as an indicator of this. rn conclusi.n, we report the necessity of key preliminarlr strrriiesThis will lead to clarification of the direction of research and of the in laboratory and field. survey methodoloEly, 49 1.2 Problems in the interpretation of results It is important to realize the length of,the biological cycle of the fish. The majority of species reproduce each year and liye for se.veral years. Therefore, an insecticide having a medium- or lonp;-term effect, would necessitate seyeral years of observations on thepopulation structure and abundance. The interpretarion is cornplicated b., natural variationsin population abundance. Some of these are related to I'i,drologicaI phenomena. Strch a sitrration could be exemplified by low srrrvival abiliry of young fish duringperiods of low flood or short flood duration. There are also other unmeasurable natural phenomena which must be taken into account to avoid hasty interpretarions. An example worth considering is given by catches of Schilbe mystus, which became very low in the Leraba jrrst after treatment started and in the nag",ra UET6iJ-t."utment. In hoth rivers, however, the catches of S.mystus increased considerably in 1979, by which time both rivers were under treatment. Consideration should also be given to variations due to spawning migrations in theinterpretation of results (consequently the need for biological srudies in the progranme area).Finally' to compare results obtained over long periods, it is necessary to ensllre uniformityin methodology and experience of scientists and technicians in the fie1d, This makes ir necessary to have qualified and experienced staff capable of replacing existing personnel. fn conclusion, many years of observation are required to confirm that an insecticide is adversely or otherwise affecting fish populations or communities becarrse of the many parametersthat have to be considered before a satisfactory conclusion can be reached. The recording and sLorage of field data in its raw form is important for accgrateinterpretation by the field teams, the coordinating hydrobiologist and/or outside bodies andprovides a full historical record in case of staff changes. In view of the variability of the data and the importance of biotogical knowledge inthe interpretation, the analyses need to aid the construction of hypotheses, retaining theidenrity of each sample and preferably involve graphical represeotation. Techniques such as correspondence analysis can detect major trends and distrrrbances andmay identify indicator species. statisticai irp..,r"rents may be made, but the naruralpopulation variability at any one p1ace, renders precise ,.,niy""" uonecessary. Field workers should be able to interpret the diagrams, but not necessarily to performthe statistical task. However, most analyses could be carried out by the coordinatingtrydrobiologist using a pre-progranrned micro-computer. copies of the data shorrld be held inHQ, Geneva, and external statistical work is only n."""sary in case of problems, developmentand independent assessment. 1.3 Continuity If conclusions are to be made from data collected over many years, then the infornrationmust be gathered in the same way. obviously, the method of collection must be standard.changes in protocol are only permissible if some conversion factor may be employed withorrtloss of staristical va1ue. rt is evident from the present work that the abundance, speciescomposirion and condition factors vary with the hydrological season. collections made evenat monthly intervals show considerable variation. The idea is ro sample at the same hydro-logical season of each year. This is still not ideal because of differences in rrater1evel between years, but would be better than a catendar date. other connnitments on the parrof monitoring personnel wou1d, lrotrever, cause complications. rt is also essential ro be aware,from sector offices, or hydrological stations, of the development of the water regime and notesshouLd be taken, and included on the forms, concerning possible influences from facrors sucl.ras fishing activity, algal blooms, and other evidence of disturbance including the date ofthe last treatment. such corEnelrts are invaluable to the interpretation of data, e.g, the 49 1.2 Problems in the interpretation of results It is important to realize the length of,the biological cycle of the fish. The majority of species reproduce each year and liye for se.veral years. Therefore, an insecticide having a medium- or lonp;-term effect, would necessitate seyeral years of observations on thepopulation structure and abundance. The interpretarion is cornplicated b., natural variationsin population abundance. Some of these are related to I'i,drologicaI phenomena. Strch a sitrration could be exemplified by low srrrvival abiliry of young fish duringperiods of low flood or short flood duration. There are also other unmeasurable natural phenomena which must be taken into account to avoid hasty interpretarions. An example worth considering is given by catches of Schilbe mystus, which became very low in the Leraba jrrst after treatment started and in the nag",ra UET6iJ-t."utment. In hoth rivers, however, the catches of S.mystus increased considerably in 1979, by which time both rivers were under treatment. Consideration should also be given to variations due to spawning migrations in theinterpretation of results (consequently the need for biological srudies in the progranme area).Finally' to compare results obtained over long periods, it is necessary to ensllre uniformityin methodology and experience of scientists and technicians in the fie1d, This makes ir necessary to have qualified and experienced staff capable of replacing existing personnel. fn conclusion, many years of observation are required to confirm that an insecticide is adversely or otherwise affecting fish populations or communities becarrse of the many parametersthat have to be considered before a satisfactory conclusion can be reached. The recording and sLorage of field data in its raw form is important for accgrateinterpretation by the field teams, the coordinating hydrobiologist and/or outside bodies andprovides a full historical record in case of staff changes. In view of the variability of the data and the importance of biotogical knowledge inthe interpretation, the analyses need to aid the construction of hypotheses, retaining theidenrity of each sample and preferably involve graphical represeotation. Techniques such as correspondence analysis can detect major trends and distrrrbances andmay identify indicator species. statisticai irp..,r"rents may be made, but the naruralpopulation variability at any one p1ace, renders precise ,.,niy""" uonecessary. Field workers should be able to interpret the diagrams, but not necessarily to performthe statistical task. However, most analyses could be carried out by the coordinatingtrydrobiologist using a pre-progranrned micro-computer. copies of the data shorrld be held inHQ, Geneva, and external statistical work is only n."""sary in case of problems, developmentand independent assessment. 1.3 Continuity If conclusions are to be made from data collected over many years, then the infornrationmust be gathered in the same way. obviously, the method of collection must be standard.changes in protocol are only permissible if some conversion factor may be employed withorrtloss of staristical va1ue. rt is evident from the present work that the abundance, speciescomposirion and condition factors vary with the hydrological season. collections made evenat monthly intervals show considerable variation. The idea is ro sample at the same hydro-logical season of each year. This is still not ideal because of differences in rrater1evel between years, but would be better than a catendar date. other connnitments on the parrof monitoring personnel wou1d, lrotrever, cause complications. rt is also essential ro be aware,from sector offices, or hydrological stations, of the development of the water regime and notesshouLd be taken, and included on the forms, concerning possible influences from facrors sucl.ras fishing activity, algal blooms, and other evidence of disturbance including the date ofthe last treatment. such corEnelrts are invaluable to the interpretation of data, e.g, the 50 water flow may stoD at certain times of the flood in alrnrrst a1l rhe rivers and nels rnay not bein the same place for a series of colleetions It is also clear that changes in personnel anrl expertise inevitahly affecr tht, resrrlMinor differences in scienrific methodology and raxonomy knowledge will be reflected indifficulties in interpretation. A long-term connnitment on the part.f sraff is reqtrired. ts. :ljt.::11:borafive field meetings.as well as tabte discussions. unfortuoately, yearty L&: ttris ainr.wt tl Long-term planning is not a short-term task. If training of new personnel important with respect fo staff staff changes occur, sufficient to ensure compatibility and cont . The deve I oDmcnf of expert i se i s time must be allocated to rhe inuity of dara. i.4 Collaboration In'any large-scale monitoring exercise with several teams, a coorclinating biologist isLrr DEvu(d Ledtils, C ofOlnatlng D1Oessential' His office must be considered as headquarters and informarion bank. FLrrLs ru DL us url ro o n rters r r ari. n k. ield staffwill be required to report proElress and any specific corrnents or prohlems. aod to answerqueries and carry out specific exercisur r" required. The hydrobiologist in return would beresponsible for the continued standardization of methodoiogy and for the suoply of informationon e'g' hydrology and treatment dates, if not 1oca11y available. 1t ir also important rhatthe hydrobiologist visit the field teams to discuss prog.""" and view the sampling siresduring a study period, particularly when new personnel are involved. Visits of experts to the field teams monitoring personnel can recognize areas also be able to inform all the teams of subsequent reports are distributed. is good for the requiring outside rhe ob-l ectives of progr€uTme in general and all heip. The hydrrrbiologist woul d such visits and ensure thar relevanr ) 2.1 sites s tudy RECOMMENDATIONS FOR FUTURE PROGRAMMES Monitoring sites The choice of sampling station is firstlv a also need to be representative of irnportantis indispensible for the selection. of access througborrt the year. The the treated river and a previous function parts of An expert with experience of the sarnpling problems and, preferably of rhe region. isneeded for counselling the 1oca1 team. rt is not reconrnended to choose stations in the inrnediate proximity of villages and camps having substantial fisheries. This is not conducive to optimum srudy because of pollution,overfishing and disturbance to both parties. rt is also important to know the 1eve1 offishing activity in the river basin, if such knowledge is available, and the hydrological regime. It also is saving time at thr'starl of the programe if a systematic inventory iscFrlied out by an expert. -! 2 .- 2' Me thodo l ogy .t Until the present time the monitoring protocol assessed thebatterv of 5 gi11 nets with different mesh sizes. The experiencepropose certain modifications, namely the adciition of three morebattery should contain the following mesh sizes: 12.5, 15, 17.5, variation in CPUE and K by a of five years lead is to mesh sizes. The standard 20, 22.5, 25, 3O and 40 nrn. 50 water flow may stoD at certain times of the flood in alrnrrst a1l rhe rivers and nels rnay not bein the same place for a series of colleetions It is also clear that changes in personnel anrl expertise inevitahly affecr tht, resrrlMinor differences in scienrific methodology and raxonomy knowledge will be reflected indifficulties in interpretation. A long-term connnitment on the part.f sraff is reqtrired. ts. :ljt.::11:borafive field meetings.as well as tabte discussions. unfortuoately, yearty L&: ttris ainr.wt tl Long-term planning is not a short-term task. If training of new personnel important with respect fo staff staff changes occur, sufficient to ensure compatibility and cont . The deve I oDmcnf of expert i se i s time must be allocated to rhe inuity of dara. i.4 Collaboration In'any large-scale monitoring exercise with several teams, a coorclinating biologist isLrr DEvu(d Ledtils, C ofOlnatlng D1Oessential' His office must be considered as headquarters and informarion bank. FLrrLs ru DL us url ro o n rters r r ari. n k. ield staffwill be required to report proElress and any specific corrnents or prohlems. aod to answerqueries and carry out specific exercisur r" required. The hydrobiologist in return would beresponsible for the continued standardization of methodoiogy and for the suoply of informationon e'g' hydrology and treatment dates, if not 1oca11y available. 1t ir also important rhatthe hydrobiologist visit the field teams to discuss prog.""" and view the sampling siresduring a study period, particularly when new personnel are involved. Visits of experts to the field teams monitoring personnel can recognize areas also be able to inform all the teams of subsequent reports are distributed. is good for the requiring outside rhe ob-l ectives of progr€uTme in general and all heip. The hydrrrbiologist woul d such visits and ensure thar relevanr ) 2.1 sites s tudy RECOMMENDATIONS FOR FUTURE PROGRAMMES Monitoring sites The choice of sampling station is firstlv a also need to be representative of irnportantis indispensible for the selection. of access througborrt the year. The the treated river and a previous function parts of An expert with experience of the sarnpling problems and, preferably of rhe region. isneeded for counselling the 1oca1 team. rt is not reconrnended to choose stations in the inrnediate proximity of villages and camps having substantial fisheries. This is not conducive to optimum srudy because of pollution,overfishing and disturbance to both parties. rt is also important to know the 1eve1 offishing activity in the river basin, if such knowledge is available, and the hydrological regime. It also is saving time at thr'starl of the programe if a systematic inventory iscFrlied out by an expert. -! 2 .- 2' Me thodo l ogy .t Until the present time the monitoring protocol assessed thebatterv of 5 gi11 nets with different mesh sizes. The experiencepropose certain modifications, namely the adciition of three morebattery should contain the following mesh sizes: 12.5, 15, 17.5, variation in CPUE and K by a of five years lead is to mesh sizes. The standard 20, 22.5, 25, 3O and 40 nrn. 51 , ,'t 'f t length of one side. The 12.5 and 15 nnn allows better estimation r:f the annual recruitment, fish. Moreover, a good sample of .j uvcnile'-s and sma11 spacies is beneficial as tlre'sr,,rrl apriori more susceptible to insecticide. We also recormnend a wider sLlrvey, cach year, in some river basins at the end of the flood. These st-rrdies will permit an improvemenr of thc knowledge of the fish comnunities. For this it is necessary to use other fishing te<:lrnique. such as cast nets and electric fishing and fisheries investigation. It is also recornrnended that for the pre-treatment and 2-years post-treatment periorJ, an investigation of the diet and fecundity be carricci out for the main soecies. After rhisperiod, these investigations could be carried ou6 evcry two years. Chemical monitoring for the accumulation of insecticides can be carried out at the same time. 2.3 Calendar of operations We propose the time schedule in the accompanying diagi'am. After the installation of tl team and choice of station, an intensive survei llanr:e sirorrld be carried out tr{ro ye,ars bef trrt and two years after the start of treatment. I'he frequency of visits shotrld be every three months to the principal station. After this period the station shoutd be sampled twice a year, before and after the flood. If there is any doubt about the action of the insecticidt this frequency is maintained. On the other hand, if an effect is not detected after two years, the surveillance may be reduced to once per year after the flood. Fina1ly, if a new insecticide is used the intensi"e monitoring is re-established CONCLUSIO}s Becartse of the economic.rnd ecological importance of fish popualtions, it is necessary to carry out surveillance in areas where insecticides are being used to ascertain any important effects on non-target fauna. Any meaningfui prograrnrne must take accounf of delayed effects, the longevity of the species and long-term natural variations. In the design and execution of such a monitoring scheme it is important to create a balance between scientific desirability and practical necessity. A minimum progranrne of observat i rn irre lrr,les studies on population size, strlrcture and key work on the biology of the impc,rtant species. The logistic problems are Ereat but r,ne consider that the enclosed reconrnendations represent this minimum cffort requirecl to assess any changes. 3. la ., aI 51 , ,'t 'f t length of one side. The 12.5 and 15 nnn allows better estimation r:f the annual recruitment, fish. Moreover, a good sample of .j uvcnile'-s and sma11 spacies is beneficial as tlre'sr,,rrl apriori more susceptible to insecticide. We also recormnend a wider sLlrvey, cach year, in some river basins at the end of the flood. These st-rrdies will permit an improvemenr of thc knowledge of the fish comnunities. For this it is necessary to use other fishing te<:lrnique. such as cast nets and electric fishing and fisheries investigation. It is also recornrnended that for the pre-treatment and 2-years post-treatment periorJ, an investigation of the diet and fecundity be carricci out for the main soecies. After rhisperiod, these investigations could be carried ou6 evcry two years. Chemical monitoring for the accumulation of insecticides can be carried out at the same time. 2.3 Calendar of operations We propose the time schedule in the accompanying diagi'am. After the installation of tl team and choice of station, an intensive survei llanr:e sirorrld be carried out tr{ro ye,ars bef trrt and two years after the start of treatment. I'he frequency of visits shotrld be every three months to the principal station. After this period the station shoutd be sampled twice a year, before and after the flood. If there is any doubt about the action of the insecticidt this frequency is maintained. On the other hand, if an effect is not detected after two years, the surveillance may be reduced to once per year after the flood. Fina1ly, if a new insecticide is used the intensi"e monitoring is re-established CONCLUSIO}s Becartse of the economic.rnd ecological importance of fish popualtions, it is necessary to carry out surveillance in areas where insecticides are being used to ascertain any important effects on non-target fauna. Any meaningfui prograrnrne must take accounf of delayed effects, the longevity of the species and long-term natural variations. In the design and execution of such a monitoring scheme it is important to create a balance between scientific desirability and practical necessity. A minimum progranrne of observat i rn irre lrr,les studies on population size, strlrcture and key work on the biology of the impc,rtant species. The logistic problems are Ereat but r,ne consider that the enclosed reconrnendations represent this minimum cffort requirecl to assess any changes. 3. la ., aI -\2 a t{ d o, )i ) I I I l I \o rn \..- t tla I t "l _t I -l I l -l I -1 qJ o ! o OJ6 3 q,) o lr L-i F{ o Hz td Fi rrl&H IHt/)o Ar ts a Hz Td EFi rI]&H If:l&A I I I x I Il6IrrIHI r,:I r,1(JorZaH&FJ Ftr{HHZ(Joz &El Ar HZ ac)z rI]MH&zoHE Il>rtklIV I I I I I I I I tr(n oh r+ ap r! c, >.6t il !(d c) rl !$ q, \, ! OJ eEo '.1rr}' .a q)6>!li vtAlq c4 [-o---T---1 I j E.i I!; 6383 rr..A(,.c la t t;l olool -/o i / . .,t or-CLq-ll.o-Jq-qr - J -\2 a t{ d o, )i ) I I I l I \o rn \..- t tla I t "l _t I -l I l -l I -1 qJ o ! o OJ6 3 q,) o lr L-i F{ o Hz td Fi rrl&H IHt/)o Ar ts a Hz Td EFi rI]&H If:l&A I I I x I Il6IrrIHI r,:I r,1(JorZaH&FJ Ftr{HHZ(Joz &El Ar HZ ac)z rI]MH&zoHE Il>rtklIV I I I I I I I I tr(n oh r+ ap r! c, >.6t il !(d c) rl !$ q, \, ! OJ eEo '.1rr}' .a q)6>!li vtAlq c4 [-o---T---1 I j E.i I!; 6383 rr..A(,.c la t t;l olool -/o i / . .,t or-CLq-ll.o-Jq-qr - J I T ri 'f I PART I: PART II: PART III: REPORT OF FISH POPUTATIONS IN RIVERS OF IVORY COAST AND GITANA Fish conmunities of Ivory Coast rivers treated by temephos(C. Ldvaque, D. Paugy & J.-M. Jestin) Observations on fish populations in Abate-treated rivers in Northern Ghana (E. K. Abban, C. P. Fairhurst & M. S. Curtis) Fish monitoring in West African rivers - problem and perspectives(E. K. Ablran, C. P. Fairhurst, C. L6vGque & D. paugy) 11 -! t'! t! Februarv 1982 I T ri 'f I PART I: PART II: PART III: REPORT OF FISH POPUTATIONS IN RIVERS OF IVORY COAST AND GITANA Fish conmunities of Ivory Coast rivers treated by temephos(C. Ldvaque, D. Paugy & J.-M. Jestin) Observations on fish populations in Abate-treated rivers in Northern Ghana (E. K. Abban, C. P. Fairhurst & M. S. Curtis) Fish monitoring in West African rivers - problem and perspectives(E. K. Ablran, C. P. Fairhurst, C. L6vGque & D. paugy) 11 -! t'! t! Februarv 1982