Shelters and their use by fishes on fringing coral reefs

Coral reef fish density and species richness are often higher at sites with more structural complexity. This association may be due to greater availability of shelters, but surprisingly little is known about the size and density of shelters and their use by coral reef fishes. We quantified shelter availability and use by fishes for the first time on a Caribbean coral reef by counting all holes and overhangs with a minimum entrance diameter ≥3 cm in 30 quadrats (25 m(2)) on two fringing reefs in Barbados. Shelter size was highly variable, ranging from 42 cm(3) to over 4,000,000 cm(3), with many more small than large shelters. On average, there were 3.8 shelters m(-2), with a median volume of 1,200 cm(3) and a total volume of 52,000 cm(3) m(-2). The number of fish per occupied shelter ranged from 1 to 35 individual fishes belonging to 66 species, with a median of 1. The proportion of shelters occupied and the number of occupants increased strongly with shelter size. Shelter density and total volume increased with substrate complexity, and this relationship varied among reef zones. The density of shelter-using fish was much more strongly predicted by shelter density and median size than by substrate complexity and increased linearly with shelter density, indicating that shelter availability is a limiting resource for some coral reef fishes. The results demonstrate the importance of large shelters for fish density and support the hypothesis that structural complexity is associated with fish abundance, at least in part, due to its association with shelter availability. This information can help identify critical habitat for coral reef fishes, predict the effects of reductions in structural complexity of natural reefs and improve the design of artificial reefs.     

Observations on agonistic behaviour of Clarias gariepinus larvae and juveniles under different densities and feeding frequencies in a controlled environment

Ten-day-old African catfish larvae, Clarias gariepinus , were stocked in aquaria at 5, 13, 22, and 30 larvae per liter with fish at each density being fed three and six times per day over 35 days. Fish were fed with an artificial diet in excess of their satiation requirements at each feeding. Every week total length was determined for samples of 25 fish per tank and all fish were counted at weekly intervals. Behaviour patterns were categorized into timed behaviour (i.e. resting, swimming and browsing), and frequency events e.g. number of contacts, aggressive contacts and number of head or body attacks. Fish at the lowest density grew faster than fish at the other densities. Feeding frequency had no effect on growth. With increasing densities fish less frequently changed timed behaviour patterns while spending a greater percentage of the time swimming. Aggressive behaviour was highest at the lowest density but remained uninfluenced at densities greater than five fish per litre. The high feeding frequency provoked a higher rate of aggressive contacts than the low feeding frequency. A suggested explanation was that these fish were more often in anticipation of food and consequently more active between feedings. In summary, it was shown that density had an effect on duration behaviour patterns such as resting and swimming, while feeding frequency significantly influenced frequency events, in particular the rate of aggressive behaviour. The study contributes to an understanding of factors influencing C. gariepinus juvenile behaviour and suggestions for follow-up experiments are discussed.     

Recruitment variation in a stream galaxiid fish: multiple influences on fry dynamics in a heterogeneous environment

1. The relative importance of density‐dependent and density‐independent processes in explaining fluctuations in natural populations has been widely debated. In particular, the importance of larval supply and whether it may control the type of regulatory processes a population experiences has proved contentious. 2. Using surveys and field experiments conducted in streams in Canterbury, New Zealand, we investigated how variation in the survival of non‐migratory Galaxias vulgaris fry was affected by density‐dependent and density‐independent processes and how this variation influenced recruitment dynamics. 3. Fry populations with high settlement densities experienced a 70–80% reduction in population size from density‐related mortality during the first fourteen days after peak settlement but thereafter the influence of density‐dependent processes on fry was weak. The impact of environmental conditions on fry populations was dependent on fry size and the magnitude of the perturbation, such that flooding effects on fry survival were most severe when fry were small. 4. In streams not affected by flooding, the size and density of introduced trout (Salmo trutta and Oncorhynchus mykiss) were the most significant factors determining the abundance of eventual recruits. A field experiment manipulating brown trout access to fry populations revealed that trout as small as 110 mm may be capable of greatly reducing and possibly preventing galaxiid recruitment. 5. Overall, the results indicated density‐dependent population regulation was only possible at sites with high native fish densities because trout were likely to be suppressing the number of potential recruits at sites with low native fish numbers. Whilst density‐dependent processes had a strong effect on fry survival following the period of peak fry abundance, density‐independent processes associated with flow and predatory trout influences on fry survival largely determined recruitment variability among galaxiid populations. Focusing conservation efforts on improving habitat to increase fry retention and reducing the impacts of trout on galaxiids would ensure more native fish populations reached their potential abundance.     

The critical-period concept for juvenile survival and its relevance for population regulation in young sea trout, Salmo trutta

An example of density-dependent regulation is provided by a long-term investigation (1966-present) of a population of migratory trout (estuarine and sea trout), Salmo trutta L., in a Lake District stream. Evidence for the concept of a critical period for the survival of young fish is briefly reviewed and found to be rather equivocal. The concept is, however, relevant to the trout population. Loss rates were high before but low after a critical survival time ( t_c days after fry emergence) that varied between year-classes (range 33-70 days) and was inversely density-dependent on egg density. Survivor density and loss rates were strongly density-dependent on egg density before t _c, but proportionate survival with stable loss-rates occurred after t _c. Some trout established feeding territories soon after emergence and the number of fish without territories decreased from a high initial value to a negligible value at t _c. Fish size at t_c was not constant but increased as t _c increased. The range of t _c for the different year-classes was similar to that for survival times of unfed fry in the laboratory. A new stock-recruitment model, incorporating t _c, has been developed for the trout population and shown to be related to the model (Ricker curve) used in the long-term study. The critical time can also be regarded as the critical age for survival in young trout; this concept may be relevant to other fish species.     

Effects of breeding site density on competition and sexual selection in the European lobster

The availability of breeding sites has been predicted to affectthe intensity of sexual selection, including mate competition,mate choice and ultimately, variation in mating success. Wetested the hypothesis that reduced density of shelters wouldcause an increase in the intensity of sexual selection in Europeanlobsters, Homarus gammarus. However, we found little supportfor our predictions. For example, within-sex competition bymales and by females was not more intense when shelters werescarce. Indeed, females attempted to evict one another fromshelters significantly more often when shelters were common.When shelters were abundant, shelter-holding males had greatermating success than males without shelters, yet females didnot show more interest towards these males during courtshipencounters. Mate attraction was more strongly related to largemale body size when shelters were scarce. Overall, the resultssuggest that reduced shelter density does not lead to more overtwithin-sex aggression in this species. Instead, we suggest thatimpacts of breeding resource availability on sexual selectionmay depend on the range over which resources are measured, withextreme scarcity of shelters rendering overt competition uneconomical.Furthermore, females may become more selective of male traitssuch as large size, which enhance male control of breeding sitesand hence protection of females.     

Gyrodactylus salaris infecting allopatric Arctic charr Salvelinus alpinus fry: an experimental study of host survival

Artificially reared Arctic charr Salvelinus alpinus fry (post yolk‐sac stage) from Skibotnelva in northern Norway were exposed to infection with the monogenean ectoparasite Gyrodactylus salaris to determine if the parasite caused mortality in fry. About 95% of the fry became infected within the first week post‐exposure, and at the end of the experiment (77 days) all fry carried the parasite. Mortality rate was significantly higher in the groups of infected fry than in the uninfected control groups, and when the experiment was terminated c. 30% of the infected and 8% of the uninfected fish had died. Gyrodactylus salaris had a pronounced negative effect upon fry survival and caused an additive host mortality. Moreover, the concurrent temporal changes in parasite intensities and aggregation indicated that the parasite‐induced host mortality was density dependent.     

The influence of high nestbox density on the common dormouseMuscardinus avellanarius population

The response of common dormouseMuscardinus avellanarius Linnaeus, 1758 population to availability of nest sites was studied by manipulating the nestbox grid and ring-marking dormice. Abundance of adult dormice more than doubled in the 25 × 25 m nestbox grid in comparison to the 50 × 50 m grid, as a result of increased nestbox density from four to 16 boxes/ha. This effect already became apparent in the first year after additional nestboxes were made available and resulted from dormouse immigration, mostly from adjacent areas without nestboxes. In the second and third years, the number of two-year-old and older resident dormice, which had their home ranges in this plot, increased considerably. The average size of dormouse home range decreased by approximately half both in males and females in the 25 × 25 m grid compared to the 50 × 50 m grid. The proportion of breeding adult females did not differ between the two grids in spite of different adult dormouse density. Shortage of secure nest sites was a limiting factor for the common dormouse population abundance in the forest where natural tree hollows were absent, and high nestbox density increased environmental carrying capacity.     

Seasonal and spatial pattern of shelter use by badgers<Emphasis Type="Italic">Meles meles</Emphasis> in Białowieża Primeval Forest (Poland)

In 1997–2001, we investigated the use of day-time shelters by radio-collared badgersMeles meles (Linnaeus, 1758) in the Białowieża Primeval Forest, eastern Poland. Each social group of badgers utilised, on average, 9 different shelters per territory (range: 4–20). The main setts, occupied for breeding and winter sleep, were also most frequently used for day-time rest throughout the year (73% of days). Badgers living in the pristine oldgrowth stands utilised larger number of shelters and spent more days in hollow trees (mainly limeTilia cordata), compared to badgers inhabiting younger secondary tree stands. Number of shelters used by individuals varied between seasons and depended on sex and age of animals. In summer, badgers used more shelters than in spring and autumn. In winter, they stayed in their main setts only. Adult males occupied more shelters and spent fewer days in the main sett than other badgers. In spring, females rearing young used only the main setts. The average underground space used by badgers within the main sett was 128 m². It was largest in summer and smallest in winter, and also varied between males and females. We proposed that, in a low-density population, badgers used several setts and other daily shelters to reduce energy expenditure when exploring their large territories and foraging. Furthermore, setts may play a role of marking sites. Analysis of the biogeographical pattern of sett use by European badgers showed that the number of setts used by social groups increased with increasing territory size, whereas the density of setts (n setts/km²) was negatively correlated with territory size. We proposed that different factors could shape the utilisation of setts by badgers in low- and high-density populations.     

Primordial germ cells in the embryos of Medaka fish.

In the second International Microgravity Laboratory (IML-2) mission in 1994, four small Japanese killifish (Medaka, Oryzias latipes) made a space travel of 15 days aboard a space shuttle. These four adult Medaka fish successfully mated in space for the first time among vertebrate animals. Moreover, the eggs they laid developed normally, at least in their external appearance, hatching as fry (baby fish) in space. Fish mated and laid eggs every day during the first week. Near the end of the mission most of the eggs had a well-developed body with two pigmented eyes. In total, 43 eggs were laid (detected), out of which 8 fry hatched in space, as truly 'space-originated' babies. A further 30 fry hatched within 3 days after landing. This is the normal hatching rate, compared with the ground-based data. Among the 8 space-originated fry, four were killed for histological sections, and germ cells at the gonadal region were counted for each fry. Their numbers were in the range of the germ cells of the normal control fry (ground-kept samples). Thus, as embryos developed normally in their external appearance, inside the embryos the formation of primordial germ cells took place normally in space, and their migration to the genital ridges was not hindered by microgravity. The two of the remaining space-originated fry have grown up and been creating their offspring in the laboratory. This proved that the primordial germ cells formed in space were also normal from a functional point of view. The four space-travelled adult fish re-started mating and laying eggs on the 7th day after landing and continued to do so every day afterward. Fertilization rate and hatchability of these eggs were as high as the eggs laid by the laboratory-kept fish. This fact implies that in gametogenesis of adult fish, there are no specific stages of germ cells extremely susceptible to microgravity.     

Delayed density dependence and oscillatory population dynamics in overlapping-generation systems of a seed beetle Callosobruchus chinensis: matrix population model

Long-term experimental systems with overlapping generations using a seed beetle, Callosobruchus chinensis, were maintained by providing 5 g of azuki beans (Vigna angularis) in two different renewal intervals: either 7 days or 10 days. The 7-day-renewal system (system 1) showed oscillatory dynamics with a constant periodic cycle of ca. 7 weeks. More stable population dynamics were seen in the 10-day-interval system (system 2). Short-term experiments showed that survivorship of adults increased with higher adult density, and that the survival rate of adults up to the age of 7 days was much higher than up to 10 days of age. In addition, the per capita production of hatched eggs by females which had survived for 7 days increased with increasing density experienced by the females. Females aged 10 days rarely laid eggs which hatched. We constructed a matrix population model based on either 1 week for system 1 or 10 days for system 2. The model included five stages in system 1: the hatched egg, the final instar larva, the pupa, the young adult and the old adult. Four stages were incorporated in the model for system 2: the young instar larva, the pupa, the young adult, and the old adult. Logistic-difference equations were applied to formulate both overcompensatory density dependence in the hatched-egg production by adults and undercompensatory response in the larval development up to the pupa. The survivorship of young adults to the old stage and the per capita hatched-egg productivity of the old females followed a linear regression against the young adult density. Inside-bean processes were adjusted to be equivalent in the two models, irrespective of the resource renewal intervals. The model predicted that system 1 would oscillate for a long time but that system 2 would rapidly converge to the equilibrium point. Multiplicative effects of both the delayed density dependence through interstage restraint effects and the overcompensatory density dependence in hatched-egg production generated various dynamic patterns ranging from a quickly disappearing damped oscillation to stable limit cycles in system 1. The relationship between resource renewal cycles and delayed density dependence was discussed based on these simulations.     

Effects of the population density on growth of SD-JCL rats (author's transl)]

The effects of the population density on growth and life span of 21 to 26 months in SD-JCL-rats were investigated. The mean life spans of males and females were 23 and 24 months, respectively. The population density had no definite influence on increase in body weight of females, whereas a most accelerated growth of males was seen with 2 rats per cage. The total amounts of food intake of each animal throughout its life were about 15 and 11 kg for males and females, respectively, and those of water intake were about twice the food. No correlation was observed between the population density and the weights of the liver, kidney, heart and femur, though mals reared at 5 animals per cage had lightest adrenals among all the animals.     

Shadow competition in wild juvenile sea-trout

Shadow competition occurs in a group of sit-and-wait predators when those closer to a source of mobile prey reduce the feeding success of those further from the prey source. It was examined in territorial juvenile sea trout Salmo trutta in a small stream. The fry formed groups of two to six fish with adjacent territories and a social hierarchy within each group. It was hypothesized that: (i) as group size increased, the mean number of prey eaten per fish within a group decreased and the variability in prey consumption between fish increased; (ii) prey consumption by individual fish decreased with increasing distance from the food source; (iii) group size increased as the mean water velocity immediately upstream from a group, and hence potential drifting food, increased. Five groups of fry were fed on small shrimps released upstream from each group at a rate of one every 15 s over a 10 min period, this procedure being repeated over 5 days to provide five replicates per group. Experiments were performed three times in 1967, 1969 and 1974 to provide information on 45 groups of fry. The first and third hypotheses were supported, but the second was only partially supported. In 1967 and 1969, territory size and shrimp consumption by individual fry decreased with increasing distance from the food source. This also occurred in 1974, except during a critical period for survival when fry density was exceptionally high with large numbers of sea trout lacking territories. This resulted in sea trout fry with the largest territories eating fewer shrimps than those with medium-sized territories because they spent more time defending their territories against sea trout lacking territories. This study is the first to demonstrate shadow competition in a vertebrate species, but has also shown that territorial defence may modify the consequences of shadow competition when densities are high and there is strong competition for the acquisition of a territory.     

Effect of Population Density,Size of Habitat and Oviposition Substrate on the Breeding System of Pupfish (Cyprinodon pecosensis)

Size of habitat, availability of oviposition substrate, and population density were manipulated to determine their effects on the structure and dynamics of the pupfish breeding system. The fish established a territorial breeding system in large tanks, irrespective of population density or oviposition substrate. A dominance hierarchy, in which one male controlled most of the oviposition substrate and spawned with most females, was established in small tanks at low densities. Both population density and oviposition substrate affected the number of males defending territories. Male spawning success was inversely related to availability of oviposition substrate, but was not affected either by density or tank size. Location of a territory and its size also influenced spawning success. Males with bigger territories had higher spawning success. A “carryover” effect of neighbors was observed; males with reproductively active neighbors spawned more often than those with less successful neighbors. Interference with spawning was affected both by tank size and density. Agonistic behavior of breeding males was not affected by any manipulation, but it was positively associated with male spawning success at high but not at low densities. In all treatments the mortality of breeding males was approximately 8 times as high as that of females, suggesting that, under these experimental conditions, the cost of reproducing is substantially higher for males. Comparisons between breeding pupfish in these experiments and in natural environments provide a basis for understanding the selective pressures that have favored the maintenance of these extremely plastic breeding systems.     

Adjusting parental investment to changing environmental conditions: the effect of food ration on parental behaviour of the convict cichlid, Cichlasoma nigrofasciatum

Animals that care for their offspring may vary the amount of care provided for a particular brood in relation to environmental conditions. Food availablity is one factor that may affect the costs and benefits associated with parental investment. The convict cichlid, Cichlasoma nigrofasciatum, is a small, substrate-spawning cichlid from Central America. Both male and female provide parental care for eggs and fry. Paris were kept at one of three ration levels, high, medium or low. Time spent in parental fanning by females was positively related to ration. Males spent less time fanning than females and their parental behaviour varied considerably between individuals. Males on the high ration spent slightly more time fanning than those on the lower rations. The number of eggs produced per spawning and the post-spawning weight of both males and females were significantly and positively related to ration. Foraging frequency of both males and females was inversely related to ration. There was no significant effect of ration on the frequency of mouthing eggs and young or on intra-pair aggression. Eggs of low-ration fish hatched earlier than those of medium- and high-ration fish but there was no significant difference in the number of days that the young survived. These results suggest that the allocation of time and effort between parental and maintenance activities differs in relation to food supply. Parents may provide more care for the large brood produced when food is plentiful but place more emphasis on their own survival when food is short and broods are small.     

Effect of vertebrate predation on the spatio-temporal distribution of cladocerans in a temperature eutrophic lake

We analysed the spatio-temporal distribution of zooplankton along a profile of 10 stations from the shore to the pelagic zone from April to September 1988, the period when the larvae and juveniles Rutilus rutilus, the most abundant species in the Lake, are in the littoral zone. The digestive tracts of the young roach were analysed. They fed essentially on rotifers and on cladocerans. For comparison, zooplankton was also analysed at one littoral area without fish fry. There was an increase of cladoceran density from the vegetated nearshore zone to the offshore zone. Considering the density of Bosmina longirostris, Daphnia longispina, Chydorus sphaericus and Ceriodaphnia quadrangula, we observed a different distribution pattern in the course of the year. In the nearshore zone, the relative abundance of small species, Bosmina and Chydorus, was much higher than that of the larger Daphnia. From April to September, predation pressure mainly affected the smallest species: in contrast to the inshore station without fish fry, the density of Bosmina decreased in May in the littoral with fish. Chydorus was concentrated in the littoral between February and April, then grew into the pelagic zone, where predation pressure obviously was low during the warm season. The number of Daphnia, which was eaten by the fish fry at any time, remained low in the nearshore zone, which suggests that the presence of fish may cause Daphnia to avoid this zone. Ceriodaphnia which was not affected by this predation, was scarce in the nearshore zone during mid-summer. The low density of the cladocerans in the nearshore zone is likely associated with vertebrate predation by roach fry and juveniles, the result of such a process being either a depletion in density of the prey, or an avoidance behaviour.     

Breeding success of four male life history types of spring Chinook Salmon spawning in an artificial stream

In 1997 the Cle Elum Supplementation Research Facility was established to enhance spring Chinook salmon returning to the upper Yakima River, Washington State. This effort increased spring Chinook abundance, yet conditions at the hatchery also significantly elevated the occurrence of jacks and yearling precocious males. The potential genetic effect that a large influx of early maturing males might have on the upper Yakima River spring Chinook population was examined in an artificial stream. Seven independent groups of fish were placed into the stream from 2001 through 2005. Males with four different life history strategies, large anadromous, jacks, yearling precocious, and sub-yearling precocious were used. Their breeding success or ability to produce offspring was estimated by performing DNA-based pedigree assessments. Large anadromous males spawned with the most females and produced the greatest number of offspring per mate. Jacks and yearling precocious males spawned with more females than sub-yearling precocious males. However, jacks, yearling and sub-yearling precocious males obtained similar numbers of fry per mate. In the test groups, large anadromous males produced 89%, jacks 3%, yearling precocious 7%, and sub-yearling precocious 1% of the fry. These percentages remained stable even though the proportion of large anadromous males in the test groups ranged from 48% to 88% and tertiary sex ratios varied from 1.4 to 2.4 males per female. Our data suggest that large anadromous males generate most of the fry in natural settings when half or more of the males present on a spawning ground use this life history strategy.     

Effects of dietary lipid source on reproductive performance and tissue lipid levels of Nile tilapia Oreochromis niloticus (Linnaeus) broodstock

Nile tilapia were fed diets supplemented with one of the following lipid sources at 5% level: cod liver oil, corn oil, soybean oil, a coconut oil-based cooking oil or a combination of cod liver oil and corn oil (1 : 1). The control diet had no lipid supplement and tad fish meal as a sole protein source. A diet with soybean meal as a protein source was also tested. The number of females that spawned, spawning frequency, number of fry per spawning, and total fry production were increased at varying degrees by the supplemental lipid sources except for the cod liver oil. Fish fed the soybean oil diet tad the best overall reproductive performance over a 24-week period. Fish fed the cod liver oil diet had the highest weight gain but the poorest reproductive performance. The suplemental lipids significantly increased crude fat levels in the liver and ovaries. Both males and females Ld the cod liver oil diet had the highest levels of fat in the liver and muscle. The ratio of total n-6/n-3 fatty acid in the liver, ovaries and testes was influenced by the supplemental lipid sources. It was highest in fish fed either the soybean oil diet, the corn oil diet, or the soybean meal diet and lowest in fish fed the control diet or the cod liver oil diet.     

The effects of population density on fecundity and fertility in the guppy, Poecilia reticulata (Peters)

Experiments were performed to examine the influence of population density on the fecundity and fertility in the guppy, Poecilia reticulata . The frequency of females with only ova or with different combinations of reproductive units and developmental stages in the ovary was significantly different between different densities but not between females with one and two gestations from the same population density. The ovarian width of females with one gestation from the lowest density was significantly larger than in the highest population density. Gonad length and height and ovarian volume of females with two gestation periods were larger in the lowest compared with the highest density. After one gestation, females at the lowest population level had greater ovarian length, height, gonad weight, width and volume compared with those from the higher density. Ovarian length, height and volume and gonad weight were also significantly greater among females of two gestation periods in the lowest than in the second lowest density. The average gonadosomatic ratio was highest in the lowest population density. A significant correlation was found between gonad volume and total body weight and between gonad volume and ovarian weight of females at every population level. This was also the case when ovarian weight vs. total body weight was examined in each population density. Only at the lowest population level was there a significant correlation between fecundity and total body length and between fertility and body length. A significant relationship was also found between gonad volume and total body length and between gonad weight and body length at the same population level. No significant difference of fecundity and ovum size was found between females of one and two gestation periods from the same density. Fecundity and fertility were significantly higher among females of one and two gestation periods from the lowest than from the highest density.     

Larval-rearing and growout of the red porgy (Pagrus pagrus) in the Riopesca hatchery (Greece)

Juveniles of red porgy (Pagrus pagrus 1758 L.) were experimentallyproduced in RIOPESCA hatchery during Spring 1992–95. Wild broodstock,weighing from 500 to 1000 g, was caught and brought to the hatchery insummer 1991. They were placed in a 30 m³ rectangular outdoorbroodstock tank with a sea-water flow of 5 m³ per hour.Salinity ranged from 3.9 to 4.1%, temperature from 13 to25°C , the stocking density was 4 kg m^–3,and the sex ratio was 1:1. The females spawned spontaneously in captivity producing approximately 200 000 eggs per kg of body weight. The percent offertilization was between 85–95%. Hatching was completed 85 hours later at 18 °C with a hatching percent of80–90%. The larvae were introduced into a 15 m³tank using green water (Nannochloropsis gaditana & Isochrysis galbana)and fed rotifers (Brachionus rotundiformis), Artemia salina nauplii andmetanauplii, and artificial food. After the end of the weaning period (day80 posthatch) the juveniles reached a weight of 1.4 g with a mean percentof survival of 10%. The fry were transferred to a cage site in which the growout proceeded without particular difficulties. The onlyproblem during growout was the skin coloration, that was darker than thatof wild porgies. Red porgies growth in our installations is satisfactory,with fry reaching 360 ± 12 g within 19 months, with a foodconversion rate of 1.8:1 , and a 6% mortality.     

An experimental study on the natural removal of dead trout fry in a Lake District stream

The chief objective was to discover why few moribund or dead Salmo trutta fry were observed on the stream bed when mortality within the stream was known to be high ( c . 13 000 dead fry year ⁻¹ for whole stream). Newly dead fry were placed in 20 boxes embedded in the stream bed (20 fry of known total weight per box) and arranged in pairs with one box open and one closed. One pair was removed every 2 days, the fish remains being weighed and the invertebrates in the open box being identified and counted. The experiments were performed from late April to early May in 1967, 1968, 1969 and the results were similar in each year. Both wet and dry weights of fry decreased exponentially but the rate of decrease was much higher in the open boxes; detectable fish remains were about 55% of initial weights after 20 days (end of experiment) in closed boxes but zero after 16 days in open boxes. Invertebrate scavengers were responsible for the higher loss rates in the open boxes and showed a definite succession with caddis larvae and carnivorous stonefly larvae dominant at first, but then being replaced by detritivorous stonefly larvae and freshwater shrimps. These experiments show clearly why dead fry disappear rapidly from the stream bed.