Predation rates of nemertean predators: the case of a rocky shore hoplonemertean feeding on amphipods

Estimates of the predation rates of benthic nemerteans are often based on observations of single individuals, and consequently they may not be representative for all members of a population of these predators. Herein we conducted controlled and repeatable laboratory experiments on the predation rate of the hoplonemertean Amphiporus nelsoni Sánchez 1973, which is common at exposed rocky shores along the central Chilean coast. During the austral fall (April, May 2000), nemerteans were observed in relatively high numbers crawling in the intertidal zone during early morning or late-afternoon low tides. When these nemerteans were offered living amphipods held by a forceps, they immediately attacked the amphipods and fed on them. In the laboratory experiments, nemerteans preferred the amphipod Hyale maroubrae Stebbing, 1899, which is also very common in the natural habitat of A. nelsoni. The nemerteans preyed to a higher extent on small males and non-ovigerous females than would have been expected from their abundance. We suggest that these (non-reproductive) stages of H. maroubrae are very mobile and therefore have a high likelihood of encounters with nemerteans. Predation rates reached maxima when nemerteans were provided prey densities of four or more of their preferred prey species, H. maroubrae, furthermore indicating that encounter rates with prey may affect predation rates. In long-term laboratory experiments, A. nelsoni consumed more amphipods during low tide conditions than during high tide conditions. Many nemerteans in the field prefer particular environmental conditions (e.g. nocturnal low tides), which restricts the time available for successful feeding. In the long-term experiment, predation rates of A. nelsoni never exceeded 0.5 amphipods nemertean^–1 d^–1. Maximum feeding events were 3 or 4 amphipods nemertean^–1 d^–1, but this only occurred during 10 out of a possible 2634 occasions. Nemerteans that had consumed 3 or 4 amphipods during 1 day, consumed substantially less prey during the following days. Towards the end of the long-term experiment, average predation rates decreased to 0.2 amphipods nemertean^–1 d^–1, corresponding to predation rates reported for other nemertean species (0.1–0.3 prey items nemertean^–1 d^–1). We suggest that predation rates from laboratory experiments represent maximum estimates that may not be directly transferable to field populations. Additionally, low predator–prey encounter rates with preferred prey in the field may further limit the predation impact of nemertean predators in natural habitats.     

Nemertea inhabiting the Haploops (Amphipoda) community of the northern Øresund with special reference to the biology of Nipponnemertes pulcher (Hoplonemertea)

The densities of nemerteans and associated fauna on a soft-bottom sampling station (27–30 m deep) in the Øresund were determined from 47 cores (each 135 cm² in cross-section; 20 + cm deep) collected from September to December 1989; these data were compared with 14 cores taken from the same location in December 1982. Nine species of nemerteans were identified from cores and dredge samples: Palaeonemertea — Callinera-like sp.; Heteronemertea — Cerebratulus fuscus, C. marginatus, Lineus bilineatus, Micrura fasciolata, M. purpurea; Hoplonemertea — Amphiporus bioculatus, A. dissimulans, Nipponnemertes pulcher. Mean numbers of heteronemerteans were 32 and 10 m^?2 in 1982 and 1989, respectively, and hoplonemerteans were 90 and 71 m^?2 , respectively. Only one palaeonemertean was collected during both years. Mean densities of the dominant species, N. pulcher, were similar for the two years, 74 and 68 m^?2. The dominant groups of macrofauna (n m^?2) in 1989 were ostracods (1028), amphipods (618), polychaetes (514), and ophiuroids (449). Amphipods (>90% Haploops spp.) and polychaetes (at least 30 spp.) are the major potential prey for hoplonemerteans and heteronemerteans, respectively. Laboratory feeding experiments with N. pulcher revealed that it consumed amphipods (Haploops tenuis and H. tubicola) at a rate of 2.6 worm^?1 d^?1 during the first 12 hours, but after 36 hours and beyond the rate was maintained at approximately 0.2 worm^?1 d^?1. Beyond 12 hours this nemertean showed a tendancy to only partially evacuate its prey. It was demonstrated experimentally that N. pulcher has a supply of toxin capable of killing six amphipods in approximately one hour. Limited tests showed that N. pulcher fed on the cumacean Diastylis tumida, but not on the amphipod Maera loveni or the ostracod Philomedes globosus, and that Amphiporus dissimulans readily attacked Haploops spp., but not Maera or Philomedes. Although the results of laboratory experiments are tentative, they do suggest that suctorial hoplonemerteans can exert a potentially significant effect on benthic communities. Employing seven species of polychaetes as prey for Cerebratulus fuscus and Micrura fasciolata, only the latter responded positively to one of them, Glycera alba. The hermit crab Pagurus bernhardus violently rejected N. pulcher in all feeding trials.     

Tube-building in two epifaunal amphipod species,Corophium insidiosum andJassa falcata

Jassa falcata andCorophium insidiosum are epifaunal tube-building marine amphipods, whose niches overlap in habitat and food requirements. Laboratory experiments were conducted to study the influence of the quality of different available particulate substrates on settlement and tube-building behaviour of these two amphipod species. Our experiments suggest thatC. insidiosum is less specialized in this respect thanJ. falcata. C. insidiosum is able to use organic materials for tube-building such as artificial (Ulva spec. powder) or mixed natural detritus as well as inorganic material (coarse sand); whereasJ. falcata utilizes organic materials, but sand only to a very limited extent.     

Food and feeding behavior of the hoplonemertean Oerstedia dorsalis

The monostiliferous nemertean Oerstedia dorsalis was collected from eelgrass (Zostera marina) beds located along the coast of New Jersey, and feeding responses to amphipods and isopods were observed in the laboratory. Tests with 46 worms showed that they fed suctorially on Ampelisca vadorum, Ampithoe longimana, Corophium acherusicum and C. tuberculata. Corophiids were preferred. Upon contact with an amphipod, the proboscis is everted and strikes the prey on the ventral side, immobilizing it in a few minutes. The worm probes the sternal region with its head and everts its proboscis one or more times during the process. The exoskeleton is eventually penetrated by the head, and the stomach is everted into the hemocoel as a flattened funnel-like structure. Peristaltic undulations of the body signify the suctorial action that removes the living contents from the exoskeleton. The actual feeding process (from head penetration to removal of the head) takes about 7 min. O. dorsalis is only the third species within the Prosorhochmidae for which the feeding behavior has been documented. The other two are terrestrial species, and are also suctorial.     

Food and feeding behavior of the nemerteanTortus tokmakovae

The food and feeding behavior of the suctorial nemerteanTortus tokmakovae Chernyshev, 1991, inhabiting the intertidal zone of Peter the Great Bay are studied. Laboratory observations show that this tortus is capable of attacking and consuming 6 species of amphipods, Mysidae gen. sp. mysids andPandalus sp. shrimps. Moreover, the tortuses suck the tissues of various dead crustaceans. Several tortuses may together attack and consume one amphipod. It is suggested thatT. tokmakovae play a significant role in intertidal communities.     

Observations on feeding in a South African suctorial hoplonemertean, Nipponnemertes sp. (Family Cratenemertidae)

Marine hoplonemerteans were collected intertidally inAlgoa Bay, Port Elizabeth, South Africa in 1983.Algoa Living worms appear to belong to the genus Nipponnemertes, perhaps N. africanus (Wheeler, 1940). The external morphology and thestylet apparatus are described and illustrated. Laboratory observations showed that this speciesattacked and consumed two species of sympatricamphipods, Elasmopus pectenicrus and Hyalegrandicornis (the only two species tested). Thefeeding behavior was similar to that documented forother suctorial nemerteans that feed on amphipods. Onecomplete feeding sequence, from the initial strike tocompletion of feeding on H. grandicornis,took approximately 12 min: the proboscis struck theventral side of the amphipod, which was immobilized in<1 min; the head eventually wedged between thesternal plates, and the internal organs were evacuatedby suctorial action. The addition of E. pectenicrus andH. grandicornis as potentialprey for suctorial hoplonemerteans brings the totalknown number of amphipod species to 25, involving atleast ten families. A summary of all species ofAmphipoda known to be potential prey for suctorialhoplonemerteans is presented.     

The role of hoplonemerteans in the ecology of seagrass communities

Seagrasses of the world harbor a rich and varied fauna, but a review of the literature revealed that little has been done to evaluate the ecological importance of nemerteans in such communities. Monostiliferous hoplonemerteans are common inhabitants of some seagrasses, e.g. eelgrass (Zostera), but generally they are seldom collected or identified or are apparently absent in other species such as schoalgrass (Halodule) or turtlegrass (Thalassia). Nineteen species of hoplonemerteans (four families) have been identified from eelgrass beds around the world; they exist mainly as epifauna, and all except two species are probably suctorial feeders. Some palaeonemerteans (2 species) and heteronemerteans (4 species) are also associated with eelgrass, but mainly as infauna. Suctorial nemerteans (4 species in 3 families) from eelgrass beds located along the mid-Atlantic coast of the United States feed in the laboratory on a variety of amphipod species that inhabit eelgrass. Tubicolous species (e.g. Corophium) seem to be preferred. Zygonemertes virescens feeds on nine species of amphipods belonging to six families, and is the only species to feed on isopods (3 species). Analyses of field studies on the occurrence of hoplonemerteans in eelgrass beds in Virginia and New Jersey, along with available information on the food habits of these worms, were used as a basis for demonstrating their potential importance as predators of peracarids in seagrass systems. More careful methods for collecting and identifying worms, continued studies on food preferences and rates of predation, and emphasis on the population dynamics of worms and prey, are recommended in order to evaluate the role of suctorial hoplonemerteans in the ecology of seagrasses.     

Status of the Nemertea as predators in marine ecosystems

The ecology of nemertean predators in marine ecosystems is reviewed. Nemerteans occur in most marine environments although usually in low abundances. Some species, particularly in intertidal habitats, may reach locally high densities. During specific time periods appropriate for hunting, nemerteans roam about in search of prey. Upon receiving a stimulus (usually chemical cues), many nemertean species actively pursue their prey and follow them into their dwellings or in their tracks. Other species (many hoplonemerteans) adopt a sit-and-wait strategy, awaiting prey items in strategic locations. Nemerteans possess potent neurotoxins, killing even highly mobile prey species within a few seconds and within the activity range of its attacker. Most nemertean species prey on live marine invertebrates, but some also gather on recently dead organisms to feed on them. Heteronemerteans preferentially feed on polychaetes, while most hoplonemerteans prey on small crustaceans. The species examined to date show strong preferences for selected prey species, but will attack a variety of alternative prey organisms when deprived of their favourite species. Ontogenetic changes in prey selection appear to occur, but no further information about, e.g. size selection, is available. Feeding rates as revealed from short-term laboratory experiments range on the order of 1–5 prey items d^–1. These values apparently are overestimates, since long-term experiments report substantially lower values (0.05–0.3 prey items d^–1). Nemerteans have been reported to exert a strong impact on the population size of their prey organisms through their predation activity. Considering low predation rates, these effects may primarily be a result of indirect and additive interactions. We propose future investigations on these interactive effects in combination with other predators. Another main avenue of nemertean ecological research appears to be the examination of their role in highly structured habitats such as intertidal rocky shore and coral reef environments.     

Aspects of the biology of Pantinonemertes californiensis,a high intertidal nemertean

I studied the distribution, feeding biology, and reproductive biology of Pantinonemertes californiensis, described as a semi-terrestrial nemertean, along the central California coast. At the sites used in this study, maximal tidal height is about 2 m, and P. californiensis typically occurred under boulders between 1.3 and 1.7 m tidal height. Worms fed primarily on the semi-terrestrial amphipod Traskorchestia traskiana. Distribution of nemerteans was similar to that of the prey, although prey extended higher on the beach than did the worms. Nemerteans were largest and most abundant at the site with highest abundance of T. traskiana and smallest and least abundant at the lowest prey abundance site. In laboratory feeding trials, nemerteans from the site with lowest prey abundance fed most readily. Non-reproductive nemerteans lived for at least a week when submerged in sea water; some prey died within a week of being submerged. Nemerteans only lived minutes when submerged in fresh water; 50% of prey lived 4.5 h. Eggs are approximately 90–100 μm in diameter and hundreds to thousands are shed per female. Larvae are planktonic and apparently planktotrophic, and are morphologically similar to other marine hoplonemertean larvae. At the sites studied life history characteristics of P. californiensis provided little evidence of adaptations to terrestrial life in these worms and were not helpful in elucidating the role of semi-terrestrial nemerteans in the evolution of terrestrial nemerteans.     

Exploitation of sediment bacterial carbon by juveniles of the amphipod Monoporeia affinis

1. Carbon budget parameters were measured for young-of-the-year Monoporeia affinis in a combined field and laboratory (microcosm) study, designed to quantify the role of sediment bacteria as a carbon source for juvenile amphipods. Special emphasis was placed on the stimulative effects of amphipod activity (foraging, feeding, bioturbation) on sediment bacterial production and abundance by including the carbon thus generated in carbon budget calculations. 2. Amphipod production was clearly higher at lower densities, suggesting strong intraspecific interactions. Negative production was recorded at amphipod densities of 10000 and 20000 ind. m^?2. Negative production was not accompanied by a decrease in amphipod total lipid content, however, probably due to the lack of easily mobilized lipids in juvenile amphipods. Amphipod respiration rate was 0.45 μg O₂ ind.^?1 h^?1, or O.15 μg C ind.^?1 h^?1. Sediment bacterial carbon content averaged 1.31 and 0.90 mg g^?1 DW under field and laboratory conditions, respectively. 3. Bacterial carbon was not quantitatively important for Monoporeia. Due to higher bacterial abundance and production in natural, stratified sediment, assimilation of bacterial carbon was highest for the field population, providing 6.3% of the amphipods' carbon requirement. In microcosm populations, bacterial carbon accounted for between 1.7 and 5.2% of overall amphipod carbon demand, increasing with amphipod density and bioturbation. 4. Ingestion rate, rather than the quantity of bacterial carbon in the sediment, was found to limit absorption of bacterial carbon from the sediment.     

Prey selection by an intertidal beetle: field test of an optimal diet model

Summary Adult beetles Thinopinus pictus LeConte (Staphlyinidae) live on sand beaches in temporary burrows from which they emerge at night to prey on amphipods Orchestoidea calforniana (Brandt). I constructed models of amphipod size selection by beetles, using the size distributions of amphipods measured on the beach, and the results of laboratory experiments on capture success, reaction distance and feeding rates. Capture success decreased and the probability that an amphipod was detected increased with increasing amphipod size. Beetles observed during beach searches selected larger sizes of amphipods than predicted from availability and vulnerability of different sizes. To apply an optimal foraging model, I estimated the profitability of different sizes of amphipods from the number of amphipods of a given size required to satiate a beetle in the laboratory. Profitability was highest for large amphipods and lowest for small amphipods and isopods. However, amphipod abundance on the beach was always below the threshold at which specialization on larger sizes was predicted to occur.     

Prey: predator ratio dependence in the functional response of a freshwater amphipod

1. First known for their shredding activity, freshwater amphipods also behave as active predators with consequences for prey population regulation and amphipod coexistence in the context of biological invasions. 2. A way to quantify predation is to determine the average consumption rate per predator, also known as its functional response (FR). 3. Although amphipods are gregarious and can display social interactions that can alter per capita consumption rates, previous studies using the FR approach to investigate amphipod predation ignored such potential mutual interference because they did not consider variations in predator density. 4. We investigated the FR of Echinogammarus berilloni feeding on dipteran larvae with joint variations in prey and predator densities. This bivariate experimental design allowed us to estimate interference and to compare the fits of the three main classes of theoretical FR models, in which the predation rate is a function of prey density alone (prey‐dependent models), of both prey and predator densities (predator‐dependent models) or of the prey‐to‐predator ratio (ratio‐dependent models). 5. The Arditi–Ginzburg ratio‐dependent FR model provided the best representation of the FR of E. berilloni, whose predation rate showed a decelerating rise to a horizontal asymptote as prey abundance increased. 6. Ratio dependence means that mutual interference between amphipods leads to prey sharing. Mutual interference is likely to vary between amphipod species, depending on their level of aggressiveness.     

Multiple predator effects in an intertidal food web

1. We examined the effects of multiple predators from an intertidal boulder food web to test whether and how three different predator species affected the survival of a small amphipod species. 2. Predators were chosen because they differ in their foraging mode, two feeding at the bottom and in benthic refuges (nemertean and shrimp) and one in the water-column (juveniles of a fish). 3. Mortality of amphipods was not affected by nemerteans, but was high in the presence of shrimp or fish. Highest mortalities were observed in predator-combinations that contained both shrimp and fish. Amphipods responded to shrimp by escaping into the water column, while they avoided fish by remaining in the refuge. We conclude that predator-specific defence causes conflicts for prey when both shrimp and fish are present. 4. Comparing observed effects of multiple predators with expected effects revealed risk enhancement for the shrimp + fish combination. A comparison of different predictive models revealed that the multiplicative model was most appropriate, although additive models may work well under certain conditions. 5. Based on known consumption-ranges of the predators used, we conclude that nemerteans were saturated with prey while fish were far from their saturation point. A predator's functional response curve (prey consumption in relation to prey abundance) determines its impact on prey populations. This knowledge appears essential in order to predict whether prey organisms face risk enhancement, risk reduction or additive effects of multiple predators.     

Amphipod (Gammarus minus) responses to predators and predator impact on amphipod density

Recent theoretical work suggests that predator impact on local prey density will be the result of interactions between prey emigration responses to predators and predator consumption of prey. Whether prey increase or decrease their movement rates in response to predators will greatly influence the impact that predators have on prey density. In stream systems the type of predator, benthic versus water-column, is expected to influence whether prey increase or decrease their movement rates. Experiments were conducted to examine the response of amphipods (Gammarus minus) to benthic and water-column predators and to examine the interplay between amphipod response to predators and predator consumption of prey in determining prey density. Amphipods did not respond to nor were they consumed by the benthic predator. Thus, this predator had no impact on amphipod density. In contrast, amphipods did respond to two species of water-column predators (the predatory fish bluegills, Lepomis macrochirus, and striped shiners, Luxilus chrysocephalus) by decreasing their activity rates. This response led to similar positive effects on amphipod density at night by both species of predatory fish. However, striped shiners did not consume many amphipods, suggesting their impact on the whole amphipod “population” was zero. In contrast, bluegills consumed a significant number of amphipods, and thus had a negative impact on the amphipod “population”. These results lend support to theoretical work which suggests that prey behavioral responses to predators can mask the true impact that predators have on prey populations when experiments are conducted at small scales. Received: 21 March 1997 / Accepted: 15 December 1997     

Ecological studies of the nemertean fauna in an estuarine system of the northwestern Gulf of Mexico

The distribution and abundance of nemerteans in the brackish-water lakes of Sea Rim State Park, Texas, near the Louisiana border, were studied and compared with other macrobenthos during one year. Six of 93 macrobenthic species collected were nemerteans (0.9% of the total number of specimens). Only one species of nemertean, Carinoma sp., was consistently present. This species is the most ubiquitous and, probably, the most abundant nemertean in the estuarine systems of the Texas coast. Carinoma sp. was collected at Sea Rim from a salinity range of 0–21 ppt and at other Texas estuaries from 2–26 ppt. Preliminary experiments with Carinoma sp. as predator and as prey indicated that it feeds on polychaete worms and in turn is fed upon by white (Penaeus setiferus) and brown (P. aztecus) shrimp.     

Zur Nahrungsaufnahme vonTetrastemma melanocephalum (Nemertini)

The proboscis of hoplonemerteans is armed with a stylet apparatus, which is used for capturing prey.Tetrastemma melanocephalum (Johnston) is a common hoplonemertean in the littoral of the Baltic Sea and the North Sea. It lives in the phytal and on sand- and mudflats and occurs in large numbers in theCorophium belts in the waddenseas.T. melanocephalum feeds on arthropods, mainly on copepods and amphipods. The prey is caught with the proboscis, penetrated by the stylet and grows weak (reduction of movements) within a few seconds; it is then sucked out. During the summer, a 15–35 mm long nemertean captures about 3 specimens ofCorophium volutator of 3.5–6 mm length per day. In theCorophium belt at Sahlenburg, German Bight, North Sea, 29 nemerteans/0.25m² were found; they feed mainly onC. volutator, so that on average more than 10 000 specimens ofC. volutator/m²/month are sucked out by these nemerteans. Hence,T. melanocephalum is an important consumer ofCorophium.     

Distribution of circular single‐stranded DNA viruses associated with benthic amphipods of genus Diporeia in the Laurentian Great Lakes

相似文献   

INTENSE GRAZING AND PREY‐DEPENDENT GROWTH OF PFIESTERIA PISCICIDA (DINOPHYCEAE)1

Grazing and growth of Pfiesteria piscicida (Pfiest) were investigated using batch and cyclostat cultures with Rhodomonas sp. (Rhod) as prey. Observed maximum growth rates (1.4 d^?1) and population densities (2 × 10⁵ cells·mL^?1) corresponded to values predicted by Monod functions (1.76 d^?1; 1.4 × 10⁵ cells·mL^?1). In batch cultures under a range of prey‐to‐predator ratios (0.1:1 to 180:1) and prey concentrations (1000–71,000 cells·mL^?1), Rhodomonas sp. was always depleted rapidly and P. piscicida concentrations increased briefly. The rate of Rhodomonas sp. depletion and the magnitude of P. piscicida population maxima depended on the prey‐to‐predator ratio and prey concentration. Starvation resulted in cell cycle arrest at G1 and G2+M and ultimately the demise of both P. piscicida and Rhodomonas sp. populations, demonstrating the dependence of P. piscicida on the supply of appropriate prey. The depletion of Rhodomonas sp. populations could be attributed directly to grazing, because P. piscicida did not exert detectable inhibitory effects on the growth of Rhodomonas sp. but grazed intensely, with maximum grazing rates>10 Rhod·Pfiest^?1·d^?1 and with no apparent threshold prey abundance for grazing. The results suggest that 1) the abundance of appropriate prey may be an important factor regulating P. piscicida abundance in nature, 2) P. piscicida may control prey population, and 3) high growth and grazing potentials of P. piscicida along with cell cycle arrest may confer survival advantages.     

FATTY ACIDS IN PHOTOTROPHIC AND MIXOTROPHIC GYRODINIUM GALATHE‐ANUM (DINOPHYCEAE)

Fatty acids were measured in G. galatheanum grown either phototrophically, or mixotrophically with Storeatula major (Cryptophyceae) as prey. G. galatheanum, like many photosynthetic dinoflagellates, contains high amounts of n‐3 long‐chain‐polyunsaturated fatty acids (LC‐PUFA) such as docosahexaenoic acid (DHA, 22:6n‐3) and the hemolytic toxic fatty acid 18:5n‐3. We hypothesize that a benefit of phagotrophy in G. galatheanum is the acquisition of precursor linolenic acid (18:3n‐3) that fuels LC‐PUFA synthesis. Phototrophs grew at 0.37 d^?1, while mixotrophs grew at 0.40 d^?1 with a feeding rate of 0.62 d^?1. Photosynthesis was lower in mixotrophs (3.7 pg C cell^?1 h^?1) than phototrophs (4.9 pg C cell^?1 h^?1). DHA levels were higher in mixotrophs [3.7 (+/? 0.11) pg cell^?1] than phototrophs [3.0 (+/? 0.16) pg cell^?1] and prey [0.4 (+/? 0.01) pg cell^?1]. 18:5n–3 levels [1.7 (+/? 0.03) pg cell^?1] were similar in phototrophs and mixotrophs. An intermediate in n‐3 LC‐PUFA synthesis, 20:4n‐3, accumulated in mixotrophs [0.6 (+/? 0.27) pg cell^?1] relative to phototrophs (not detected) and prey [0.03 (+/? 0.002) pg cell^?1]. Low ratios of linolenic acid to DHA in phototrophic G. galatheanum (0.14) relative to mixotrophic G. galatheanum (0.29) and prey (2.14) are consistent with substrate limitation of LC‐PUFA synthesis in phototrophs. Accumulation of 20:4n‐3 suggests incomplete conversion of linolenic acid to DHA, possibly due to conditions in batch culture. We conclude that precursors for n‐3 LC‐PUFA biosynthesis in G. galatheanum may be acquired through ingestion of S. major, and may partially control feeding/photosynthesis in mixotrophic populations.     

The effect on Orchestia hurleyi (Amphipoda: Talitridae) of a whitey disease caused by Bacillus subtilis

Abstract

Field observations made over 10 years suggested that a bacterial disease of adults of the terrestrial amphipod Orchestia hurleyi Duncan, caused by Bacillus subtilis, is progressing southwards down the eastern side of New Zealand's South Island. As the disease spread, amphipod density appeared to decline and population age structure became truncated. In the vicinity of Dunedin and further south the amphipods are still disease-free. Signs of the disease are a progressive weakening and wasting. The animal cannot jump, and its speed of walking is reduced. Its body becomes opaque white instead of the normal translucent reddish-brown. Diseased females do not brood. There is no evidence that diseased animals moult. Death is caused by general wasting or by predators. The disease-causing organism was isolated, and healthy amphipods were re-infected from the isolate. Signs of the disease were apparent within 7 days of inoculation. The presence of the disease-causing organism in the haemocoel causes host defences to be mobilised, as shown by elevated haemocyte counts (4512 mm^?3, cf. 300 mm^?3 in healthy, disease-free adults), but as the disease progresses the animal's defences are overcome, and haemocyte counts fall to an average of 784 mm^?3 during the later stages of disease. The blood of terminally diseased amphipods is thick and creamy white, packed with motile bacterial cells, and few (if any) haemocytes are present in the circulation. Two populations were studied, one disease-free (at Dunedin) and the other heavily diseased (at Christchurch). The incidence of disease (as measured by a performance test) was about 30%r in Christchurch adults. The disease-causing strain of B. subtilis was found on the body surface of almost all adults in the diseased population. It is possible that the bacterium gains entry to the haemocoel through wounds suffered during ecdysis, conflict, or predator attack. The main differences shown by the diseased population relative to the disease-free population were: lower average density (992 m^?2, cf. 1677 m^?2); lower maximum density (3104 m^?2, cf. 9971 m^?2); smaller average size, with fewer adult instars; a smaller proportion of females brooding in each instar; and much lower egg production. The brood size/mother age relationship was the same for both populations — number of eggs in brood = -4.9 + 0.64(instar number of mother)—because in the diseased population only healthy females breed. Lower egg production in the diseased population reflects the smaller proportion of healthy females, and the number of broods per female is lower since life expectancy is much less. A computer model based on Leslie matrices was used to simulate the ecological effects of the disease. It gave predictions which conformed with the observed population features with respect to age structure and density.