Field experiments on the relationship between drift and benthic densities of aquatic insects in tropical streams (Ivory Coast). III. Trichoptera BERNHARD STATZNER, JEAN-MARC ELOUARD" and CLAUDE DEJOUX? Zoologisches Institut I der Universität Karlsruhe, Federal Republic of Germany, *Laboratoire d'Hydrobiologie, O.R.S.T.O.M. , Bamako, Mali, and ?Laboratoire d'Hydrobiologíe, O.R.S.T.O.M., La Paz, Bolivia SUMMARY. 1. Based on in situ gutter trials we related the drift of caddis flies to their benthic densities and to various abiotic factors in streams in the Ivory Coast (West Africa). Members of the families Hydropsychidae, Philopotamidae , Hydroptilidae and Leptoceridae were considered in detail. 2. The drift of larvae peaked at night in both early and late larval instars. 3. Drift of a larval group (a certain instar, species or higher taxon) was more often related to the benthic density of other larval groups than to its own benthic density. 4. Self-regulation of an upper benthic density of a larval group by emigration through drift was not statistically evident. 5. There was no straightforward relationship between drift and abiotic factors. 6. Drift rates differed between taxa as well as between Xarval instars (size groups) within a taxon. Newly hatched larvae had very high drift rates, whereas the last larval instar usually had the lowest drift rate. 7. We related these results to the violently fluctuating discharge of the streams in the study area and the consequent variability of space for lotic insects. 8. Drift estimates, made at the same time as a monitoring programme on possible side-effects of insecticides (Onchocerciasis Control Programme), failed to reflect benthic densities except in the night drift of Hydropsychidae. 1. Introduction dynamics in running water has stimulated much work, which we have recently reviewed Benthic macroinvertebrates dislodged from the (Statzner, Dejoux & Elouard, 1985a). Natural stream bottom form part of the organismic drift. drift is often related to such abiotic factors as The importance of this drift for population daylight, moonlight, water temperature, dis- Correspondence: Dr B. Statzner, Zoologisches charge, current velocity, characteristics of the Institut I der Universität, Postfach 6380, 7500 bottom substratum, turbidity; and to the biotic Karlsruhe, Federal Republic of Germany. factors of benthic density of the drifting taxon, @Rg[ohij F~fidsDocumqfa ire 391 30AQ 4, .e& . 392 B. Statzner, J.-M. Elouard and C. Dejoiuc abundance of its food and predators. Past inter- &Hydrobiologie Tropicale. Another paper est has focused on the question of whether or not (Statzner, Elouard & Dejoux, 1985b) con- drift of a taxon reflects its benthic abundance or sidered Cheumatopsyche falcifera, the most even production. Drift is ‘density dependent’ if numerous caddis fly in our tests. The present emigration rates through drift change with paper deals with the remaining Trichoptera as changing benthic densities. In the classical well as with the question of how much the drift of sense, density dependence of drift is described C. falcifera was affected by the benthic density of by an exponential relationship between benthic other trichopteran taxa (section 3.3). We will density (x) and drift (y),y=aebx, where a and b focus our interest on the Hydropsychidae, since are constants. This could be the result of an this family was often represented by more than increase in the rate of drift inducing interactions one abundant species in our tests and because we within a taxon with increasing benthic density. were able to identify larvae to species and instar The relationships between drift and some of level. Other taxa sufficiently abundant to be the above factors are relatively clear. But in considered in detail were members of the many cases there remain many uncertainties Philopotamidae, Hydroptilidae and Lepto- (Statzner et al., 1985a). Although field work on ceridae. this subject is preferable (Townsend & Hildrew, Our objective was to relate drift of these cad- 1976; Goedmakers & Pinkster, 19Sl), most dis flies to light regime, application of insec- experimental studies were carried out in rather ticides in the OCP, various other abiotic factors artificial stream systems, with the attendant risk (Table 1) and, especially, to the properties of the of behavioural artefacts. We thus tried to relate benthic trichopteran community. drift of benthic stream insects to the above fac- tors under quasi-natural conditions in in situ trials with trough (gutter) systems, which were 2. Methods and Materials used from 1977 to 1981 in the Ivory Coast (West Africa). The methods we applied have been described In that area, streams exhibit extreme varia- and discussed in detail elsewhere (Statzner et al., tions in discharge. Therefore two points are of 1985a, b). special interest (Statzner, 1982; Lévêque, Dejoux & Iltis, 1983): (i) space for lotic inver- TABLE 1. Independent variables tested in drift tebrates becomes a scarce resource in stream models riffles at times of low discharge, (ii) individuals Abiotic drifting off riffles are rapidly exported to very Mean discharge through the gutter (m3 h-l): long pools of almost standing water during such DIS periods. Thus, the question is whether the den- Mean current velocity in the gutter (cm s-l): sity of riffle populations is self-regulated through VEL density dependent drift or whethe; other mecha- Mean water temperature (“C): TMP Turbidity (Secchi-transparency:cm): TRB* nisms affect population densitieh under these Moon (l=new, . 5=full, . l=new): MON extreme environmental conditions. Biotic Most of the streams in the study area were Benthic density of that larval group treated in regularly sprayed with insecticides against larvae the dependent variable (ind. 0.1 m-z): OWN of Simulium damnosum s.1. in the Onchocer- Benthic density of other Trichoptera ciasis Control Programme (OCP) (Davies et al., (ind. 0.1 m-z) Aethaloptera dispar: AD1 1978). Since the OCP tries to monitor effects of Cheumatopsyche copiosa: CCO treatments on non-target fauna by means of drift Cheumatopsyche digitata: CDI net samples, we will briefly consider how far our Cheumatopsyche falcifera: CFA? results justify that practice. Macrostemum alienum: MAL We intend to publish our drift studies in a Other Hydropsychidae: OHY Chimarra petri group: CPE series of papers. The introductory one (open Orthotrichia spp. : OTT questions in stream drift, methods and experi- *Transparencylarger than depth of water: 110. mental conditions; Statzner et al., 1985a) and the ?If drift of larval instars of this species was con- concluding (synecological aspects) publication sidered, CFA represented the benthic density of all of this series did or will appear in the Revue but the instar tested in the dependent variable. Drift and benthic densities of tropical caddis flies 393 2.1. Experimental stream protocol about drift concentration (=density) are redun- Natural substrata (stones, gravel, sand, wood, dant because relatively constant discharge leaves) colonized by benthic animals and placed passed through our gutters. Thus we will present in gutters (length 1.5-3 m, width 0.09-0.2 m) results in terms of drift contributed per unit area were exposed in streams so that part of the dis- per unit time (drift transport area-l) and drift charge passed through the systems. After rate (percentage of the benthic population lost various periods of adjustment (see below) nets through drift per unit time). Data from early were fixed to the front of the gutters to prevent morning and late afternoon differed only slightly the entrance of drifting insects. These front nets from data between 10.00 and 16.00 hours GMT were shaped in a way that changes of flow char- and gave no better insight into drift phenomena acteristics in the gutters were minimized. At the of our caddis. Therefore we will concentrate on other end of the gutters a second net sampled the following: the period of peak drift (PD; animals drifting out of the system. After drift highest drift within a 24 h period), night drift samples had been collected (usually about 24 or (ND; 19.00-06.00 hours) and day drift (DD; 48 h) the abundances of the remaining benthic 1O.OC-16.00 hours). insects were evaluated from the contents of the When we seek to evaluate the relevance of gutters. The initial population density (benthic drift data for monitoring effects of insecticides, density) in a trial was calculated from the sum of data have to be gathered using techniques com- specimens which drifted plus those removed parable to those of the OCP monitoring pro- from a gutter at the end of a trial. gramme. Essentials of the drift studies in the Our data base consists of 790 drift samples, OCP monitoring programme are (Lévêque, representing a survey of more than 600 h, Odei & Pugh Thomas, 1977; Dejoux, 1983; obtained from twenty-four experiments. Elouard, 1983): (i) Drift net samples are taken Important technical details were: (i) Four from the stream for 30 min about 1.5 h before differently shaped gutters were used (see Fig. 2 (OCP day drift) and for 3 min about 1.5 h after in Statzner et al., 1985a). (ii) In the first eight (OCP night drift) sunset. (ii) Drift nets may be tests substrata from the stream bottom were put exposed either close to the water surface, close into the gutter several hours before the start of to the bottom, or close to both, depending on the the trials. Later we placed the substrata into the hydrological season. (iii) Drift is expressed as gutters 5-14 days before the tests. (iii) Eleven drift concentration (individuals m-’). (iv) OCP tests were run in stream reaches never treated night and day drift and the ratio of night drift to with insecticides or not for a long time (at least 6 day drift are expected to reflect long-term effects months), while thirteen were carried out in of the insecticide applications, if they exist, i.e.