Stream Drift Research Articles

Estimation of intergenerational drift dispersal distances and mortality risk for aquatic macroinvertebrates

Estimates of the total lifetime dispersal capacity of individuals comprising invertebrate drift in streams and rivers have proved very difficult to determine empirically. Here we use recent data on dispersal in the amphipod Gammarus pulex L. to illustrate a method for estimating the total distance an individual is likely to travel downstream during the period between hatching and its first reproductive episode. For the system we consider, this estimate is on the order of 1.5 km. This method may be useful in future explorations of the ecological relevance of within‐stream displacement and population persistence. Furthermore, we are able for the first time to estimate an upper bound for the mortality risk associated with entering the drift, which, for the G. pulex population under consideration, is less than 1% for individual drift events. We suggest that this risk may not be nearly as high as had previously been thought. These mortality and distance estimates may illustrate a fundamental difference between dispersal in lotic systems and those in other habitats, in that the mean dispersal distance is of much higher ecological relevance than that of rare long‐distance events.

Seasonal and diel patterns of invertebrate drift in different alpine stream types

SUMMARY1. We examined the seasonal and diel patterns of invertebrate drift in relation to seston and various habitat characteristics in two each of four different kinds of alpine streams [rhithral (snow‐fed) lake outlets, rhithral streams, kryal (glacial‐fed) lake outlets and kryal streams]. Samples were collected at four times of the day (dawn, midday, dusk and midnight) during three seasons (spring, summer and autumn).2. Habitat characteristics differed mainly between rhithral and kryal sites, with the latter having higher discharge and turbidity, lower water temperature, and higher concentrations of ammonium, and particulate and soluble reactive phosphorus. Seasonality in habitat characteristics was most pronounced for kryal streams with autumn samples being more similar to rhithral sites.3. The concentration of seston was lowest in the glacial‐influenced lake outlets and slightly higher in the stream sites; no seasonal or diel patterns were evident.4. The density of drifting invertebrates averaged less than 100 m−3 and was lowest (<10 m−3) at three of the four kryal sites. Taxon richness and diversity were lowest at rhithral lake outlets. Chironomidae dominated the drift as well as benthic communities and <30% of benthic taxa identified were found in the drift.5. Drifting invertebrates showed no consistent seasonal pattern. However, density tended to be highest in spring at rhithral sites and in autumn at kryal sites. No diel periodicity in drift density was found at any site and the lack of diel pattern may be a general feature of high altitude streams.6. Glacially influenced habitat parameters were a major factor affecting drift in these alpine streams, whereas no clear differences were observed between streams and lake outlets. Our findings indicate that invertebrate drift in alpine streams is primarily influenced by abiotic factors, and therefore, substantially differs from patterns observed at lower altitude.

Macroinvertebrate drift in Amazon streams in relation to riparian forest cover and fish fauna

Macroinvertebrate drift was studied in 12 sites in 9 lowland streams (400m), three mid-altitude streams (2100 m) and three high-altitude streams (3800 m) in Ecuador. The purposes of our study was 1) to examine levels and periodicity of drift in relation to altitude, 2) to examine levels and periodicity of drift in relation to environmental site characteristics and 3) to examine the effect of deforestation and specific fish families on levels and diel periodicity of drift in the lowland streams. At all altitudes, day drift of terrestrial invertebrates was equal to or greater than that of aquatic invertebrates, while aquatic invertebrates always dominated drift at night. Mean day drift density of aquatic invertebrates was the same at the two high altitudes, while it was lower in the lowland streams. In contrast, night drift density was higher in lowland stream compared to the two highland stream groups. In all lowland streams drift of aquatic invertebrates was higher at night than at daytime (night: day drift ratio 2-45, mean: 10), while it was largely aperiodic in both groups of highland streams. In all lowland streams, drift was dominated by insects (mean 99.6 %). No significant correlations were found between drift density, drift % or periodicity and riparian forest cover, benthic macroinvertebrate densities or any other environmental site characteristic in the lowland streams. However, path analyses indicated that riparian forest had a significantly negative indirect effect on night and day drift densities, because an increase in riparian cover was associated with a decline in the abundance of fish and benthos. Species richness, density and biomass of fish were directly correlated with invertebrate drift. These parameters had no effect on total day or night drift densities or drift %, but were positively correlated with the night: day drift ratio. The night: day drift ratios were more closely correlated with the species richness of Loricariidae catfish than to any other fish family. Loricariidae are night active, benthic herbivorous-detritivorous scrapers of biofilm, and their presumed effect on drift periodicity is due to physical disturbance of benthic invertebrates during the night, and not necessarily to a behavioural response from benthic invertebrates. Drift of the dominant invertebrate families (Leptophlebiidae, Baetidae, Tricorythidae, Hydropsychidae and Chironomidae) varied among localities, but night: day drift ratios of all families were positively correlated with species richness of fish, although this pattern was only significant for Leptophlebiidae and Baetidae.

The influence of predatory fish on mayfly drift: extrapolating from experiments to nature

1. A knowledge of how individual behaviour affects populations in nature is needed to understand many ecologically important processes, such as the dispersal of larval insects in streams. The influence of chemical cues from drift‐feeding fish on the drift dispersal of mayflies has been documented in small experimental channels (i.e. < 3 m), but their influence on dispersal in natural systems (e.g. 30 m stream reaches) is unclear.2. Using surveys in 10 Rocky Mountain streams in Western Colorado we examined whether the effects of predatory brook trout (Salvelinus fontinalis) on mayfly drift, that were apparent in stream‐side channels, could also be detected in natural streams.3. In channel experiments, the drift of Baetis bicaudatus (Baetidae) was more responsive to variation in the concentration of chemical cues from brook trout than that of another mayfly, Epeorus deceptivus (Heptageniidae). The rate of brook trout predation on drifting mayflies of both species in a 2‐m long observation tank was higher during the day (60–75%) but still measurable at night (5–10%). Epeorus individuals released into the water column were more vulnerable to trout predation by both day and night than were Baetis larvae treated similarly.4. Drift of all mayfly taxa in five fishless streams was aperiodic, whereas their drift was nocturnal in five trout streams. The propensity of mayflies to drift was decreased during the day and increased during the night in trout streams compared with fishless streams. In contrast to the channel experiments, fish biomass and density did not alter the nocturnal nature nor magnitude of mayfly drift in natural streams.5. In combination, these results indicate that mayflies respond to subtle differences in concentration of fish cues in experimental channels. However, temporal and spatial variation in fish cues available to mayflies in natural streams may have obscured our ability to detect responses at larger scales.

Seasonal patterns in macroinvertebrate drift and seston transport in streams of an alpine glacial flood plain

1. We measured drift of macroinvertebrates and organic matter in five different channel types (upper and lower main channel, intermittent side‐channel, side‐slope tributary, and a groundwater channel) draining a glacial flood plain in the Swiss Alps. Samples were collected seasonally over 3 years (May 1997 to November 1999) to encompass major periods of floodplain expansion and contraction.2. Total wetted channel length ranged from ≤ 5 km during maximum floodplain contraction to ≥ 24 km under system expansion. Principal components analysis (PCA) of key physical and chemical measures collected over an annual cycle showed that the main and side channel sites differed spatially and seasonally from the tributary and groundwater sites, reflecting the glacial influence on the former.3. Two‐way ANOVA indicated a significant Site × Date interaction for all measured variables. Transported organic matter, caused by blooms of Hydrurus foetidus, was typically five to 10 times higher in the lower main channel and side‐channel than the other channel types, and was least variable among dates in the groundwater and tributary channels.4. The number of drifting macroinvertebrates was usually highest in the groundwater channel, although highly variable among years. Other channel types had highest drift rates in November and lowest rates in May, reflecting seasonal changes in the life cycles of particular taxa, such as Baetis alpinus and the Simuliidae. The lower main channel exhibited the greatest, and the groundwater the lowest, seasonality in drift compared with the other channels.5. The Chironomidae displayed the highest drift rates in most seasons, especially in summer when other taxa were relatively low in the drift. The number of taxa in the drift also varied among channel types and season, ranging from less than two in early spring to four to six in other seasons.6. The observed differences in drift rate and transported organic matter among channel types and seasons have important implications for ecosystem dynamics, through its effect, for instance, on dispersal among channel types or on organic matter exchange.

Re-examining the drift paradox

Why do the upper reaches of streams remain populated by aquatic insects, other invertebrates and algae if many thousands of these organisms are swept downstream by the current every day? Invertebrate drift – basically, any invertebrate animal big enough to see with the naked eye that lives on the streambed but which is found entrained in the water column of a stream or river – provides the most observable example of this ‘drift paradox’.The idea of the drift paradox itself has been the subject of ecological angst since it was first noticed by Karl Muller nearly half a century ago. Surprisingly, of the hundreds of papers concerning invertebrate drift produced over the past five decades, only two have tried to formalize the problem and address it using theory, and both of these papers were published in the last six years 1xDensity dependence resolves the stream drift paradox. Anholt, B.R. Ecology. 1995; 76: 2235–2239Crossref | Scopus (74)See all References, 2xExact compensation of stream drift as an evolutionarily stable strategy. Kopp, M. et al. Oikos. 2001; 92: 522–530CrossrefSee all References. In a recent paper 3xPopulation persistence in rivers and estuaries. Speirs, D. and Gurney, W.S.C. Ecology. 2001; 85: 1219–1237CrossrefSee all References, Douglas Speirs and William Gurney present a treatment of this problem that uses equations describing advection and diffusion in molecular systems to approximate movement of animals entrained in the water column, as well as randomly directed movements of animals travelling along the streambed by their own volition. Their results explain the persistence of several real animals in real systems and suggest that these are cases where either advection is minimized or other forms of movement are very prevalent. These conclusions themselves are a fascinating take on an old problem that deserves much more investigation.However, one aspect of this work that is not elaborated on in their paper could, if true, have profound effects on our preconceptions regarding the management of streams and rivers. Speirs and Gurney suggest that population persistence is a delicate balance between advection by water flow and other movements acting to counter this downstream drift. However, a speculative implication of their conclusions is that small reductions in stream or river flow, not large enough to affect oxygen exchange, food availability, or other niche parameters for a given species, might actually increase population size and stability through an increase in the probability of population persistence.However, this speculation rests on two assumptions. The first, and more obvious, is that the species involved is not so specifically adapted that a small change in mean water velocity will make the habitat unsuitable. However, the key assumption is whether we believe that displacement owing to drift is adaptively compensated for and, therefore, whether reduced water velocities will result in reduced compensatory movements.

Exact compensation of stream drift as an evolutionarily stable strategy

The colonization cycle hypothesis predicts that adults of stream‐dwelling insects preferentially disperse in the upstream direction in order to compensate for larval drift. Upstream biased dispersal has indeed been shown in many, albeit not all, natural populations. Based on a recently published analysis, we develop a simple stochastic model for the competition of genotypes with different dispersal strategies in a stream habitat. By means of an invasion analysis, we show that exact compensation of larval drift by upstream biased adult dispersal is an evolutionarily stable strategy. Exact compensation means that, on average, the net movement of individuals from birth to the time of reproduction is zero. At the population level, we show that, in general, upstream biased dispersal is not necessary for persistence, unless the reproductive rate is very low. Under all conditions, however, populations of exact compensators attain highest sizes or persistence times, respectively. Although selection pressure towards exact compensation is arguably very general in populations subject to stream drift, trade‐offs or constraints might change the outcome of selection. Therefore, the analysis presented in this paper has to be viewed as a null model for optimal dispersal behavior in stream habitats.

Spatial and temporal patterns of invertebrate drift in streams draining a Neotropical landscape

1. Invertebrate drift in streams draining a tropical landscape in Costa Rica was studied to assess differences in assemblage composition above and below a major gradient break in geomorphic landform and to assess temporal patterns of drift in lowland reaches below the gradient break. The gradient break (∼50 m a.s.l.) is the point at which the foothills of the Costa Rican Cordillera Central (piedmont) merge with the Caribbean coastal plain (lowlands). 2. Spatial patterns were assessed along two streams by sampling drift over 24 h once a month for 3 months in both the piedmont (90 m a.s.l.) and lowlands (30 m a.s.l.). Temporal patterns of drift were assessed through monthly diel sampling of three lowland sites over 8–10 months, encompassing both ‘dry’ (<400 mm precipitation per month, November to May) and wet (July to October) seasons. 3. Drift composition was insect dominated in piedmont sites and larval shrimp dominated in the lowlands. Percent similarity of assemblages between piedmont and lowland sites was low (range 26–43%) because of high larval shrimp densities in lowland versus piedmont sites. 4. Drift densities were higher during night than day, with peaks at sunset on all dates and at all sites. Diel patterns in drift agree with previous observations for the study area and support the ‘risk of predation’ hypothesis. 5. Analysis of monthly patterns in lowland sites showed high variability in drift densities; however, all major taxa were found every month. Overall, there was a trend for high invertebrate densities during the ‘dry’ season, but these trends were not significant. 6. Observed changes in drift composition support the concept of river zonation, which predicts a change in community composition along the stream continuum due to geomorphic features. Drift at lowland sites below the gradient break was dominated by shrimps, which are linked to marine environments via their migratory behaviour.

Spatial and temporal patterns of invertebrate drift in streams draining a Neotropical landscape

Spatial and temporal patterns of invertebrate drift in streams draining a Neotropical landscape

Selective foraging on terrestrial invertebrates by rainbow trout in a forested headwater stream in northern Japan

The important contribution of terrestrial invertebrates to the energy budget of drift‐foraging fishes has been well documented in many forested headwater streams. However, relatively little attention has been focused on the behavioral mechanisms behind such intensive exploitation. We tested for the hypothesis that active prey selection by fishes would be an important determinant of terrestrial invertebrates contribution to fish diets in a forested headwater stream in northern Japan. Rainbow trout,Oncorhynchus mykiss, were estimated to consume 57.12 mg m−2day−1(dry mass) terrestrial invertebrates, 77% of their total input (73.89 mg m−2day−1), there being high selectivity for the former from stream drift. Both the falling input and drift of terrestrial invertebrates peaked at around dusk, decreasing dramatically toward midnight. In contrast, both aquatic insect adults and benthic invertebrates showed pronounced nocturnal drift. Because the prey consumption rates of rainbow trout were high at dawn and dusk, decreasing around midnight, the greater contribution of terrestrial invertebrates to trout diet was regarded as being partly influenced by the difference in diel periodicity of availability among prey categories. In addition, selectivity also depended upon differences in individual prey size among aquatic insect adults, and benthic and terrestrial invertebrates, the last category being largest in both the stream drift and the trout diets. We concluded that differences in both the timing of supplies and prey size among the three prey categories were the primary factors behind the selective foraging on terrestrial invertebrates by rainbow trout.

Life history of Kaniwhaniwhanus chapmani Boothroyd (Chironomidae: Orthocladiinae): Population dynamics, emergence, and drift

The life history of Kaniwhaniwhanus chapmani Boothroyd is described through examination of larval population dynamics, larval drift, and adult emergence. Development of larvae occurred all year but peak larval densities occurred in February, July, and November and indicated several generations with longest generation times in the autumn. Emergence occurred all year and a bimodal diel emergence pattern was apparent at certain times. K. chapmani larvae were present in the stream drift at all times but not in high densities, and generally only during daylight hours with an early morning/late afternoon drift pattern. K. chapmani exhibits a generally non‐seasonal life history but rapid growth tends to occur when conditions of temperature and food quality are likely to be optimal resulting in shorter generation times and peaks in larval populations.

An update on the quantification of stream drift

The reliable quantification of stream drift is of vital importance for all drift studies. Although variation between drift samples is smaller than between benthic samples, some six to seven replicates have been considered necessary by ALLAN & RUSSEK (1985) to obtain a precision of 95 %-confidence limits within 50 % of the mean. While a review of drift studies published during the last twelve years showed that many researchers have probably used too few replicates to allow a reliable quantification of stream drift, it may be often unnecessary to use as many as six or seven. In the present study conducted in a Swiss river, we analysed data from 24 drift samples taken on six different occasions during three studies conducted in winter, spring, and summer. In each season, drift was sampled twice. The 95 %-confidence limits obtained with just four replicates were within the 50 %-boundary on all six sampling occasions for the total number of invertebrates and the five relatively most common taxa in spring and summer. In the two winter collections, taxon-specific confidence limits were outside 50% of the mean in five cases out of ten. However, overall drift densities in winter were low, and all but one of the taxa with high variation between replicates were quite rare in actual numbers. A similar or even better statistical precision may be possible in other rivers, because the higher number of replicates recommended by ALLAN & RUSSEK (1985) may have been partly due to collecting some of their replicate samples during periods of low overall drift. When studying the drift of rare taxa, however, actually more than six to seven replicates may be required to generate reasonable estimates of their drift densities. We suggest that researchers conduct more pilot studies in their individual study systems to determine the appropriate number of replicates needed, since this has been rarely done in the past.

Inter‐stage survival of wild juvenile Atlantic salmon, Salmo salar L.

A biological model was developed to calculate annual survival between life stages of juvenile Atlantic salmon, Salmo salar L., in Catamaran Brook, a small stream basin (52 km2) in the Miramichi River catchment in New Brunswick, Canada. Seven years’ data (1990–1996) were used in the model. Input variables included: daily fish counts and measurements of parr (3–4 age classes), smolts, and adult salmon at a fish‐counting fence near the stream mouth; biennial quantification of all habitat types along the watercourse; fish density estimated by electric fishing at 30 sites; and estimates of young‐of‐the‐year emigration via stream drift. Continuous recording of stream discharge provided data to assist in interpretation of survival estimates. Annual survival for juvenile salmon in their first 3 years of life in the stream averaged between 31% and 34%. The greatest annual variation (CV = 0.699) occurred at the egg to 0+ (summer) stage with a low of 9.2% survival recorded for a winter with an atypical midwinter flood event; parr and pre‐smolt survival were similarly affected. Survival from egg deposition (after correction for losses caused by predation and retention/non‐fertilization) to smolt emigration was between 0.16% and 0.52%, which is low relative to estimates from many other studies. Survival of smolts to returning 1‐sea‐winter adults (grilse) averaged 8.5%. Potential errors in the computation of the model are discussed, e.g. inaccurate counts of spawning adults during high autumn stream flow. A possible explanation for the low egg to smolt survival was the environmental conditions experienced during various winters. Mean egg survival was 1.3 times higher (39.3%) and egg to smolt survival increased to 1.03% when the two winters characterized by extremely low discharge or midwinter freshets were excluded from the calculation. Density‐dependent factors related to a beaver dam, which limited spawning distribution, may also have contributed to poor survival and increased fry emigration in one year. Environmental factors, particularly winter conditions, in streams such as Catamaran Brook may act as bottlenecks to natural production of Atlantic salmon.

Invertebrate drift and longitudinal transport processes in streams

On a deloge des macro-invertebres benthiques des substrats de cours d'eau, et on les a attrape a l'aide de filets derivants a diverses distances en aval. On a effectue quatre experiences, avec plusieurs debits, dans quatre cours d'eau a caracteristiques de transport hydraulique tres differentes. On a ajuste dans une fonction exponentielle negative les valeurs obtenues a chaque distance pour tous les invertebres et les Plecopteres dans la colonne d'eau. Le chenal ayant la vitesse moyenne la plus elevee donnait le taux de capture (β) le plus faible. Dans tous les chenaux, β diminuait en raison inverse de la vitesse; ce rapport obeissait approximativement a une fonction de puissance et la pente propre a chaque cours d'eau correspondait aux caracteristiques hydrauliques du chenal. Le cours d'eau ayant la plus forte proportion de zones d'eaux mortes avait une forte valeur de β variant peu avec la vitesse du courant, alors que ceux ayant moins de zones mortes avaient une plus faible valeur de β, diminuant en rapport inverse avec la vitesse du courant. A une vitesse particuliere, la profondeur et la turbulence propres au chenal peuvent expliquer les valeurs differentes de β dans les cours d'eau comportant des zones mortes semblables. Les taux de capture pour deux familles de perles presentaient une variation relative, peut-etre a cause de changements de comportement propres a l'espece en reponse au courant. Dans un cours d'eau lent, on notait des valeurs plus elevees de β chez les Leuctrides que chez les Nemurides, et des taux semblables dans un cours d'eau plus rapide. Nos resultats sont coherents avec l'opinion selon laquelle l'habitat physique influence la distance de derive des invertebres deloges.

Density Dependence Resolves the Stream Drift Paradox

The downstream displacement of individuals by drifting in the current is a prominent feature in the population biology of stream invertebrates. To compensate for the loss of individuals it has been proposed that adults of aquatic insects preferentially fly upstream to oviposit and thereby maintain populations. The observation of adult flight biased in the upstream direction has been advanced as confirmation of the hypothesis. I argue here that upstream—biased dispersal is not sufficient for population persistence nor is it necessary. Population persistence can only be explained on the basis of density dependence at some point in the life cycle. Computer simulations showed that density dependence by itself is not sufficient for population persistence. Infrequent dispersal by randomly flying adults coupled with density dependence does allow population persistence. Upstream—biased dispersal can be explained by increased individual fitness. If upstream reaches are depopulated by drift, these reaches provide more rapid growth to successful colonists. Computer simulations of competition for space by two genotypes that differ only in the directionality of dispersal showed that genotypes with upstream—biased dispersal always drove random dispersers to extinction.

The relative importance of refugia in determining the drift and habitat selection of predaceous stoneflies in a sandy-bottomed stream

Patch structure in sandy, compared to rocky streams, is characterized by isolated snags that can only be colonized by drifting. By measuring drift from patches (snags) of various quality we determined the factors that influence habitat selection and drift of the predaceous stoneflies Acroneuria abnormis and Paragnetina fumosa. The presence of refugia (loose bark and leaf packs) was more important than hunger level and modified the effects of increased predator densities and aggressive interactions. Stoneflies concentrated to 8x natural densities with access to refugia remained longer on snags than a single stonefly without access to refugia. During periods of activity, refugia were defended with larger stoneflies always displacing smaller nymphs. During long periods of inactivity, two and sometimes three nymphs would rest side-by-side sharing the same refuge. Hunger level (starved versus satiated stoneflies), an indirect measure of a predator's response to prey availability, had no significant effect on drift or habitat selection regardless of the presence of refugia. Stonefly predators had a uniform distribution while their prey were clumped. Drift was deliberate and almost always delayed until night, usually at dusk or dawn. An examination of previous research plus the results of this study suggest that non-predatory intra- and interspecific interactions can be an important mechanism causing drift in streams.

Long‐term patterns of invertebrate stream drift in an Australian temperate stream

SUMMARY1. Invertebrate stream drift was sampled bimonthly in the Acheron River, Victoria, Australia, over a period of 18 months. Replicated hourly samples were collected over a 25‐h period on each sampling date. A total of 194 taxa were identified in the drift. However, total drift density was dominated by few taxa.2. Some evidence was obtained for a seasonal pattern in drift: this was most pronounced in relative abundances of individual taxa and the composition of the drift, rather than in total drift densities. Most of the commonly collected taxa reflected the seasonal pattern of total drift. However, some of the common taxa did not.3. A small number of taxa showed behavioral drift, with a nocturnal increase in drift densities. One species of ‘Baetis’ drifted in high densities over short periods of time around dusk and dawn. It did not drift in higher densities during the night than during the day. The results emphasize the need for drift studies to be more rigorously designed than is typically the case.

Factors influencing invertebrate drift in small forest streams, south-eastern Queensland

Aspects of invertebrate drift were investigated in four small rainforest streams in south-eastern Queensland. Hourly sampling in two streams revealed low drift rates and no distinct diel patterns. This may have been a consequence of the low flows experienced during the study period. The drift was dominated by Simuliidae, Chironomidae and to a lesser extent Tasmanocoenis sp. (Caenidae) and Cheumatopsyche sp. 6 (Hydropsychidae). Additional studies were undertaken to examine why some individuals of particular taxa drift while others (of the same taxa) remain in or on the substratum. Analysis of the relative gut fullness of simuliid and chironomid larvae showed that a higher proportion of non-drifting individuals had full or nearly full guts compared with those in the drift. Hungrier individuals of these two taxa may enter the drift as a food searching mechanism. An analysis of limb loss of hexapod nymphs indicated that a disproportionate number of individuals in the drift had one or more limbs missing compared with those that remained in the benthos. These data support the hypothesis that invertebrates present in the drift are less fit than non-drifters.

Macroinvertebrate drift in streams of the Nepalese Himalaya

SUMMARY1. Macroinvertebrate drift was investigated in seven small headwater streams along an altitudinal gradient (600–3350 m) in two adjacent river systems in east‐central Nepal. Study streams in the Likhu Khola were at the lowest altitude and flowed through terraced agriculture. At higher altitudes, Langtang streams flowed through catchments of forest and alpine scrub. Samples were collected every 3h over a 24 h period in each stream.2. Terrestrial macroinvertebrate drift was greatest in streams with semi‐natural catchments, but was scarce in streams where trees were absent.3. At lower altitudes drift was dominated by Hydropsychidae and Baetidae, which were most abundant in night samples, and Hydracarina, which tended to be day active. Baetidae also dominated drift at high altitudes, but here drift was increasingly aperiodic or diurnal. The composition of the drift reflected a significant correlation between the percentage abundance of taxa in the drift and benthos.4. As in other mountain regions of the world, drift was aperiodic in fishless streams (high altitude), but strongly nocturnal in streams where insectivorous fish were present (lower altitude). However, a wide array of potentially important variables along the altitudinal gradient, such as temperature, climate and community structure, might be important influences on these patterns.

Effects of trout on the diel periodicity of drifting in baetid mayflies.

Some benthic invertebrates in streams make frequent, short journeys downstream in the water column (=drifting). In most streams there are larger numbers of invertebrates in the drift at night than during the day. We tested the hypothesis that nocturnal drifting is a response to avoid predation from fish that feed in the water column during the day. We surveyed diel patterns of drifting by nymphs of the mayfly Baetis coelestis in several streams containing (n=5) and lacking (n=7) populations of rainbow trout, Oncorhynchus mykiss. Drifting was more nocturnal in the presence of trout (85% of daily drift occurred at night) than in their absence (50% of daily drift occurred at night). This shift in periodicity is due to reduced daytime drifting in streams with trout, because at a given nighttime drift density, the daytime drift density of B. coelestis was lower in streams occupied by trout than in troutless streams. Large size classes of B. coelestis were underrepresented in the daytime drift in trout streams compared to nighttime drift in trout streams, and to both day and night drift in troutless streams. Differences in daytime drift density between streams with and without trout were the result of differences in mayfly drift behaviour among streams because predation rates by trout were too low to significantly reduce densities of drifting B. coelestis. We tested for rapid (over 3 days) phenotypic responses to trout presence by adding trout in cages to three of the troutless streams. Nighttime drifting was unaffected by the addition of trout, but daytime drift densities were reduced by 28% below cages containing trout relative to control cages (lacking trout) placed upstream. Drift responses were measured 15 m downstream of the cages suggesting that mayflies detected trout using chemical cues. Overall, these data support the hypothesis that infrequent daytime drifting is an avoidance response to fish that feed in the water column during the day. Avoidance is more pronounced in large individuals and is, at least partially, a phenotypic response mediated by chemical cues.