Threatened freshwater pearl mussel <Emphasis Type="Italic">Margaritifera margaritifera</Emphasis> L. in NW Spain: low and very structured genetic variation in southern peripheral populations assessed using microsatellite markers

A genetic analysis of freshwater pearl mussel Margaritifera margaritifera populations from NW Spain, a peripheral area of its European distribution, was carried out using microsatellite markers. These populations were formerly reported as genetically differentiated on the basis of growth and longevity studies. Ten loci previously characterized in populations from central Europe were used to comparatively analyze the genetic variability at the southern edge of the species’ range. Iberian pearl mussel populations showed very low genetic variability and significant high genetic differentiation. Half of the total genetic diversity observed appeared to be distributed between populations, which suggested a highly structured adaptive potential in pearl mussel at the southern peripheral distribution of the species. Population distinctiveness was evidenced by assignment tests, which revealed a high accuracy of individual assignments to their population of origin. All data suggested low effective population size and major effects of genetic drift on population genetic structure. In order to avoid further loss of genetic variation in biologically distinctive populations from NW Spain, prioritization of genetic resources of this species is required for conservation and management.     

Long-term changes in the population of Eurytemora affinis (Copepoda,Calanoida) in the Gironde estuary (1978–1992)

Long-term changes in the population of Eurytemora affinis were investigated during 15 years (1978–1992) at a fixed station in the Gironde estuary (South West France). Total numbers, sex-ratio, % ovigerous females, proportion of copepodites and clutch-size were taken into account and their long-term distribution was related to temperature, river flow, salinity, suspended matter and chlorophyll a concentrations after the seasonal effect was removed for all variables.There was a great decrease of the river flow during the period of investigation, due to a general deficit in pluviosity. On the contrary, the general trend for salinity was an increase from 1978 to 1992. The same pattern was observed for chlorophyll, indicating an intrusion of marine phytoplankton. Water temperature increased significantly during the study period, following the general increase in air temperature observed in the area. For the suspended particulate matter (SPM), an increase was observed between 1978 and 1981–1982, then a very sharp decrease occurred from 1984 onwards.Numbers of E. affinis were inversely correlated with temperature and salinity, and positively correlated with the river flow. E. affinis was negatively correlated with chlorophyll concentration because of the covariation with salinity. No clear long-term trend was observed for the sex-ratio. A significant correlation was found between females carrying egg-sacs and SPM concentration, probably due to a decrease of the predation pressure in very turbid waters. The percentages of copepodites tended to decrease with time and were inversely correlated with temperature. Clutch-size significantly decreased during the 15-year period. This trend was mainly explained by temperature and salinity.It is concluded that, in the absence of strong human alteration in the Gironde estuary the long-term distribution of zooplankton can be explained by the natural environmental variability. The dominant factors are the river flow which governs the movement of the populations along the estuary and the temperature which influences the reproductive processes. Any change in climate will have consequences on the pelagic community.     

Intraspecific life history variation in the threatened Louisiana pearlshell mussel, Margaritifera hembeli

1. Margaritifera hembeli is a threatened mussel limited to twenty-two headwater streams in the Red River drainage in central Louisiana, USA. This study evaluated intraspecific variation in density, growth, size and age structure and shell morphology among several isolated populations. This study also identified the host fish and considered the role that host fish distribution played in determining mussel recruitment. 2. Mussels were aggregated in beds and average densities differed among streams. However, maximum mussel densities in beds were similar in all streams; the observed maxima were among the largest for monospecific mussel beds in North America, often exceeding 300 individuals m^–2. 3. The maximum size reached by individuals differed among streams, but all size distributions were skewed towards larger individuals. A repeated measures analysis of tagged mussels in four populations, over a 2-year period, indicated 2-fold differences in growth rates among streams, and significant variation among years. Growth rates were not affected by local population density. Maximum ages reached, determined indirectly by comparing growth rates, varied from 45 to 75 years. A canonical discriminant analysis also revealed significant differences in shell morphology across populations. 4. Half of the populations showed evidence of recent recruitment, and these sites had fish assemblages dominated by the host fish Noturus phaeus (Taylor). Host fish abundance appeared more important than adult mussel density in explaining recruitment patterns. 5. Considerable intraspecific life history variation suggests that management strategies for this species should be stream-specific, with emphasis on ensuring long-term habitat stability.     

Egg Size and Formation of the Geographical Range in <Emphasis Type="Italic">Thysanoessa inermis</Emphasis> (Kroyer, 1846) (Crustacea: Euphausiacea) in the Norwegian and Barents Seas

Using materials on the euphausiacean Thysanoessa inermis collected in the Norwegian and Barents seas (from 60°N to 73°N) we studied the variability of size characteristics in the eggs (diameter of embryos and egg capsules and the width of the perivitelline space) and their correlations with water temperature and salinity. The size of embryos showed almost no variability irrespective of the location of the sampling site and the water temperature and salinity; the perivitelline space performs a protective function, and its width showed a tendency to decrease with increasing water temperature and salinity. The similarities in the size of embryos in T. inermis populations thousands of kilometers away from each other might be explained, first, by the relatively small age of the populations (the age of the population from the Barents Sea did not exceed 6000–7000 years) and second, by movement of the crustaceans with currents from the Norwegian Sea to the Barents Sea.     

Salinity perception by larvae of the crab Rhithropanopeus harrisii (gould)

The mechanism of salinity perception was measured in Stage I zoeae of the estuarine crab Rhithropanopeus harrisii. The behavioral assay for perception of a salinity decrease was the loss of positive phototaxis while the reversal in geotactic sign from positive to negative was used for studying a salinity increase. A change in sodium chloride concentration is the primary environmental cue signalling a salinity change. However, neither sodium nor chloride per se is required for detection, since other salts such as lithium nitrate, can substitute for sodium chloride. The minor ions of seawater (calcium, magnesium, potassium and sulfate) have a modulating effect upon larval behavior, but play no role in salinity perception. The exact mechanism for perception of an increase and a decrease in salinity is different. Substitution experiments indicate that in the perception of a salinity decrease, only salts having a cation similar in size to sodium can mimic salinity changes whereas in the detection of a salinity increase, cation size is not an important parameter. The stimulation effectiveness of anions is correlated with neither size nor valence. Though osmolality is probably not involved in perceiving a decreased salinity, it may serve a role in the detection of a salinity increase. A charge in pH has no effect upon perception of a salinity decrease, but does affect perception of a salinity increase. A classical salt receptor appears to be involved in sensing a salinity decrease, while a different type of receptor is used for detecting an increase in salinity.     

Salinity tolerance in different life history stages of an invasive false mussel Mytilopsis sallei Recluz, 1849: implications for its restricted distribution

ABSTRACT

Although the false mussel Mytilopsis sallei Recluz, 1849 is recognised as an aggressive invasive species, its populations in several estuaries in Thailand are restricted to small areas. A salinity gradient is a major characteristic of its habitat, hence the effect of various salinity levels (0–40?ppt) on the mortality of larvae, juveniles and adults of M. sallei was investigated. Condition Indices of adults reared at different salinity levels for two months were measured. Spatial and temporal variations of salinity and false mussel abundance in a canal with a salinity gradient were also monitored. After an acute (48?h) test, survival of larvae was highest at salinity levels of 12.5 and 16.25?ppt and decreased at lower and higher levels. Juveniles survived at all salinity levels, but most adults died in the first 24?h at a salinity of 40?ppt, while condition indices were lowest at salinity levels of 30 and 35?ppt. In the field survey, highest false mussel abundance was consistently found at the middle part of the canal with mid-range salinity. The results suggested that salinity is a determinant of survival in M. sallei larvae and potentially regulates the dispersal success of false mussels. However, the importance of salinity was marginal in the later stages of its life history.     

Changes in the distribution and abundance of Dreissena polymorpha within lakes through time

Dreissena polymorpha population densities and biomass were followed in three Belarusian lakes with different trophic status over a 12-year period subsequent to initial colonization. In all three lakes zebra mussel population densities did not change once they reached a maximum. Application of the Ramcharan et al. [1992. Canadian Journal of Fisheries and Aquatic Sciences 49: 2611–2620] model for predicting population dynamics of zebra mussels was accurate for two of the three lakes studied. Population density appears to depend on the time since initial colonization, relative abundance of substrate available for colonization, lake morphometry and trophic type. Zebra mussel distribution within lakes was highly patchy, but the degree of dispersion decreased over time after initial colonization, which may be a result of saturation of suitable substrates by zebra mussels as populations increase and reach carrying capacity. In lakes where submerged macrophytes are the dominant substrate for zebra mussel attachment, populations may be less stable than in lakes with a variety of substrates, which will have a more balanced age distribution, and be less impacted by year to year variation in recruitment. Dreissena polymorpha usually reach maximum population density 7–12 years after initial introduction. However, the timing of initial introduction is often very difficult to determine. Both European and North American data suggest that zebra mussels reach maximum density in about 2–3 years after populations are large enough to be detected.     

Evaluation of potential health risks to Eastern Elliptio (Elliptio complanata) (Mollusca: Bivalvia: Unionida: Unionidae) and implications for sympatric endangered freshwater mussel species

Conservation efforts for freshwater musselspecies require identification and evaluationof potential health risks to populations. Sampling large numbers of individual mussels,however, could damage the small, fragile,extant populations of imperiled mussel species. In this study, a small sample size was designedto reduce lethal sampling, costs, andenvironmental disruption, while still allowing95% confidence in detecting health risks of25% or greater prevalence in the population. Health assessments were conducted on twentyspecimens of the Eastern Elliptio, Elliptio complanata, collected from two NorthCarolina sites as part of a survey to evaluatepotential disease threats to mussels in theregion. Bacteriological sampling of thegastrointestinal tracts yielded 18 aerobicbacterial species, of which Aeromonashydrophila (55.0%), Enterobacter spp.(40.0%), and Bacillus spp.(30.0%) were predominant. Histologicallesions of internal organs included mild tomoderate digestive gland atrophy andinflammation in one mussel, and mild tomoderate parasitism in several individuals. Adistinct difference in parasite prevalence wasevident between infections in E.complanata from the two collection sites. Thetrematode metacercaria of Homalometronarmatum and what appears to be three gillciliate species, the most abundant beingtentatively identified as the scyphidiidperitrich Mantoscyphidia sp., were foundin mussels from one site only. This studydemonstrates a comprehensive diagnosticapproach incorporating multiple modalities toassess the health status of mussel populations,while minimizing the sample size required toobtain valuable information. Furthermore, thisstudy provides baseline health data of E.complanata at two sites in south-central NorthCarolina and suggests the potential usefulnessof E. complanata as an environmentalbioindicator of health risks to sympatricthreatened freshwater mussel populations.     

Muskrat predation on a diverse unionid mussel community: impacts of prey species composition,size and shape

1. The muskrat, Ondatra zibethicus, is a predator known to affect species composition, population size and age structure of freshwater unionid mussel communities. Muskrats leave large piles of dead shells (middens) on the edges of rivers, lakes and streams. We compared the species composition and size structure of shells collected from muskrat middens to the nearby live unionid community in the lower Licking River (Kentucky, USA). 2. Like previous studies, our results suggest that muskrats are both size‐selective and species‐specific predators; however, our results indicate that mussel shape is also an important factor. 3. We generated a shape metric (‘cubocity’) sensitive to the overall shape of the mussel. Species with relatively lower cubocity values (around 0.85) are plate‐like or spike‐like, while mussels with more cuboidal shells have higher cubocity values (near 1.0). 4. Our results suggest muskrats prefer cuboidal mussels and generally avoid spike‐shaped mussels. The endangered fanshell, Cyprogenia stegaria, was the most favoured prey; the fanshell’s relative size and shape appear to make it particularly vulnerable to muskrats. 5. We believe the predictive capabilities of this shape metric will be of benefit to those who monitor and manage threatened mussel populations.     

Population dynamics of the ribbed mussel,Geukensia demissa: The costs and benefits of an aggregated distribution

Summary Although mussel beds are common in many intertidal habitats, the ecological significance of the aggregated distribution of mussels has not been examined. The ribbed mussel, Geukensia demissa, is found in dense aggregations on the seaward margin of many salt marshes in New England. Here, we examine the population structure of G. demissa in a New England salt marsh and investigate experimentally the costs and benefits of aggregation.Size, growth rate, and settlement rates of mussels decrease with increasing tidal height, whereas survivorship and longevity increase with increasing tidal height. Winter ice dislodges mussels from the substratum, resulting in mortality over all size classes, whereas crab predation results in the mortality of smaller mussels. The intensity of each of these mortality agents decreases with increasing tidal height. Effects of intraspecific competition on individual growth and mortality also decrease with increasing tidal height.At high densities, individual growth rates were reduced, with depression of growth rates most pronounced on smaller individuals. Mortality from sources other than intraspecific crowding, however, was reduced at high mussel densities, including mortality due to winter ice and crab predators. As a result, our data suggest that the mussel population at our study site would be reduced by 90% in only five years and no juveniles would survive through their second year without an aggregated distribution.Juveniles settle gregariously with or without adults present. The aggregated distribution of settlers and the postsettlement movement of smaller mussels to favorable microhabitats result in size and age class segregation within the population. This probably reduces intraspecific competition for food, while maintaining the survivorship advantages of an aggregated distribution.     

Effects of Environmental Factors on Byssal Thread Formation in Some Members of the Family Mytilidae from the Sea of Japan

The ability to attach repeatedly to a substrate (glass, boulders, sand) in three common mussel species of the upper sublittoral zone of the Sea of Japan, Grayan's mussel Crenomytilus grayanus, the Korean mussel Mytilus coruscus, and northern horse mussel Modiolus modiolus, was studied under experimental conditions. It was found that during 120 h of the experiment C. grayanus and M. modiolus produced more byssal threads than M. coruscus. A decrease in the water temperature from 20 to 0°C slowed the rate of production of byssal threads down to full passivity in some experimental mollusks. This was more typical of M. coruscus and less typical of C. grayanus. Renewed threads differed in their length, thick, size of the adhesive plate, and strength. M. coruscus formed the shortest, thickest, and strongest threads with rather a large adhesive disk. The observed differences are discussed from the position of morphophysiological adaptations of species for colonization of different natural substrata under contrasting conditions of the upper sublittoral zone.     

Influence of the White Sea sponge <Emphasis Type="Italic">Halichondria panicea</Emphasis> (Pallas, 1766) on physiological state of the blue mussel <Emphasis Type="Italic">Mytilus edulis</Emphasis> (Linnaeus, 1758), as evaluated by heart rate characteristics

During laboratory testing it was shown that the presence of the White Sea sponge Halichondria panicea on the shell of the mussel Mytilus edulis negatively affects the functional state of the mollusk as expressed in the increased heart rate recovery time after functional test loads (a 50% water salinity drop for 1 hour). It was found that the degree of this negative influence rises with an increase both in the projective cover of mussel shells by the sponge and in the size/age parameters of mussels.     

Population genetic structure of mussels from the Baltic Sea

In a macrogeographic survey, the population genetic structure of mussels from various regions of the Baltic Sea, a large semi-enclosed brackish-water basin, was examined with reference toMytilus edulis andM. galloprovincialis samples from the North Sea, Irish coast and southern Portugal. Electrophoretically detectable variation was analysed at 6 polymorphic enzyme loci (Ap, Est-D, Lap-2, Odh, Pgi andPgm). Evidence was provided of a remarkably large amount of biochemical genetic differentiation among ecologically and morphologically divergent mussel populations in the Baltic. Patterns of allele frequencies in low-salinity populations from the area of the Baltic Proper were demonstrated to be widely homogeneous but contrast strongly with those of the western Baltic, the latter resembling populations from marine habitats of the North Sea. Associated with a pronounced salinity gradient, the spatial heterogeneity in gene-pool structure is indicated by steep clines of allele frequency changes in the area of the eastern Danish isles. The adaptive significance of the observed allozymic variation is suggested. From genetic distance estimates, the subdivision of population structure is discussed in relation to the significant amount of differentiation detected withinMytilus populations to date and to the evolutionary time required for the divergence of Baltic mussel populations. The allozymic data provide evidence for the genetic distinctiveness of mussels from the low-salinity areas of the Baltic. Their position at the specific or subspecific level of classification requires further consideration.     

秦王川盐沼湿地角果碱蓬种群聚集分布格局与特征

植物种群的多尺度集聚与聚块特征变化是植物对环境协同适应的结果。运用成对相关函数与零模型方法,依据微地形的土壤盐分分布规律设置4个样地,在2 m×2 m的样方内设置400个小格子记录植物株数并取土样,分析了兰州秦王川盐沼湿地土壤盐分梯度下角果碱蓬种群(Suaeda corniculata)种群空间格局的集聚分布内在特征。结果表明:角果碱蓬种群集聚分布范围内呈现2个关键尺度的集聚现象,小尺寸聚块的集聚或叠加形成复合大尺度聚集,整体表现为嵌套双聚块分布;随着土壤盐分递减,种群集群分布中大聚块尺寸趋于增大,小聚块尺寸差异不明显;大尺寸聚块的数量明显减少,小聚块数量随种群植株数量变化,整个梯度呈现降低趋势;小聚块中植株个体平均数逐渐减少,复合大聚块中包含小尺度聚块平均数量呈降低趋势。内陆盐沼湿地土壤环境异质性背景下角果碱蓬种群集群分布中聚块特征的梯度变化,是植株个体形态与构成的调整适应,种群正负向生态关系、庇护与自疏效应梯度转换的结果。     

Factors controlling the colonial structure of <Emphasis Type="Italic">Pediastrum tetras</Emphasis> (Chlorophyceae)

Accounting for morphological plasticity in phytoplankton populations is relevant for taxonomy, systematic/evolutionary, and ecological studies. In this work, the green alga Pediastrum tetras (Ehrenberg) Ralfs was used to describe the variation in population size structure over its growth cycle and to analyze responses to changes in biotic and abiotic factors. Pediastrum cultures reached a final stable concentration in approximately 10 days. This density (8 × 10⁵ cells ml⁻¹) remained stable for at least another 13 days and the intrinsic growth rate was 0.24 ± 0.01 day⁻¹. In the exponential phase, the relative number of single cells and the proportion of large cells (with vesicles inside) within colonies increased. When density peaked, a relative increase of single cells as well as small cells in new colonies took place. Finally, during the stationary phase, the trend reversed: fewer single cells and a larger cell size (without vesicles) were observed. Results indicated that nutrient supply could affect population structure, diminishing the proportion of eight-cell colonies. Daphnia magna Straus significantly reduced the Pediastrum population density due to predation, and this led to a significant decrease in the density of the largest colonies. In addition, info-chemicals induced a slight increase in the density of the largest colonies compared to the control treatment. Our study suggests a possible trade-off in P. tetras colonial size in natural environments: during the stationary growth period in a lake, Pediastrum populations tend to increase in size for efficient use of nutrients, while they decrease in size in the presence of herbivores. Handling editor: J. Padisak     

Susceptibility of larval dragonflies to zebra mussel colonization and its effect on larval movement and survivorship

Colonization by the zebra mussel, Dreissena polymorpha, was quantified for five dragonfly species that differed in size and larval habits in a Michigan lake. Both larger size and a non-burrowing habit independently increased susceptibility to colonization. In 2005, over 50% of the final instars of the sprawlers Didymops transversa and Hagenius brevistylus were colonized, as well as younger instars. Rarely colonized were Progomphus obscurus and Dromogomphus spinosus, whose larvae burrow under sand, and the sprawler Epitheca princeps, whose final instars were lightly covered with sand. Hagenius larvae that had been preyed upon carried more mussels than those dying of other causes. More generally, mussel attachment decreased the probability that sprawlers left the water to emerge, the distance that some species traveled before emerging, and the ability of an overturned sprawler to right itself. On average, final instars of Didymops and Hagenius remaining in the water carried three times as many mussels as individuals known to emerge. Compared to uncolonized individuals, Epitheca and Progomphus with mussels emerged closer to the water line. Among colonized Didymops, the distance traveled on land before emerging decreased with increasing mussel load. Of the colonized Didymops that could right themselves, righting time increased with mussel load. Because the two common species of sprawlers were disproportionately colonized, and mussel attachment decreased their chances of emerging, our results suggest that D. polymorpha has the potential to affect the community structure of this guild of aquatic and terrestrial predators. Handling editor: R. Bailey     

Allozyme Diversity among Different Size Cohorts of Spat of Mytilus trossulus (Bivalvia,Mytilidae) Dwelling together in Amursky Bay,Sea of Japan

Allozyme variability in 16 polymorphic loci in two size cohorts of large (30.0 ± 6.3 mm) and small (9.8 ± 2.3 mm) spat of the mussel Mytilus trossulus (Gould) coinhabiting the intertidal zone in Brazhnikov Bay (Amursky Bay, Sea of Japan) was studied 1988–1989. It was determined that the size cohorts compared were highly similar genetically; Nei's minimum distances were 0.014 and 0.016 in 1988 and in 1989. However, statistically significant differences in allele frequencies were found at 2 loci, both in 1988 and in 1989. The loci differing by allele frequencies in 1988 did not coincide with those in 1989. Significant differences by heterozygosity between the cohorts were found only at the Idg-1 and Est-3 loci, 1988; and at the Aat-1, 6-Fgd, Est-3 loci, 1989; but the average heterozygosity was not significantly different in the cohorts for both years. A deficiency of heterozygotes (Dg) was observed almost at all loci in both cohorts. The mean Dg values were significantly higher in the cohort of small size individuals. Since the environmental conditions for both cohorts were similar, the allozyme differences between them were not a result of natural selection. A possible cause of the allozyme and size differences among the mussel cohorts is the gap in the settling time of larvae, which entails the formation of two size cohorts of mussels. Such a gap was caused by protraction of spawning period of M. trossulus and by the formation of segregate larval pools from various mussel populations differing in allele frequencies.     

Ecological regime shifts in salinised wetland systems. I. Salinity thresholds for the loss of submerged macrophytes

This paper is the first in a pair investigating potential mechanisms for ecological regime change in salinising wetlands. In this first paper, we investigated some of the factors responsible for the formation and maintenance of a macrophyte-dominated ecological regime by studying the germination and flowering of four submerged macrophyte species common in saline Western Australian wetlands; Ruppia polycarpa R. Mason, Ruppia megacarpa R. Mason, Lamprothamnium macropogon (A. Braun) Ophel and L. cf. succinctum (A. Braun in Ascherson) Wood, and by following the survival of adult R. polycarpa as salinities were increased to a range of endpoints (6, 15, 45, 70 and 100 ppt). Increased salinity led to a decrease in the number of germinating plants, an increase in the time to emergence of germinating plants and a decrease in the number of inflorescences (or fertile plants) produced by R. polycarpa, L. macropogon and L. cf. succinctum. Germination of R. megacarpa was low, providing limited information with regard to salinity response. The survival of adult plants also decreased as salinity increased and was negatively affected by faster rates of salinity increase. The upper salinity limits for germination, within the 70 day experiment, were 40–50 ppt for R. polycarpa and L. cf. succinctum, and 30–40 ppt for L. macropogon. Survival of adult R. polycarpa also declined markedly at above 45 ppt.     

Extinction,recolonization, and the genetic structure of tidepool copepod populations

Summary Extinction and recolonization in metapopulations may either increase or decrease genetic differentiation among populations, but recent genetic models predict increased differentiation under most circumstances of recolonization. I examine this prediction empirically using tidepool populations of the marine tidepool copepodTigriopus californicus. The probability of extinction of tidepool populations was sufficiently high to invoke the model's predictions, but varied among populations. Nearly 75% of colonizing groups consisted of 10 or fewer individuals. The genetic effective size of colonizing groups might be as high as 18, depending on assumptions, but colonists probably originated from a subset of local populations. In contrast to my predictions, genetic differentiation was smaller among younger tidepool populations than among older populations on each of three rock outcrops, suggesting that genetic differentiation was reduced by metapopulation dynamics. The discrepancy between the prediction and the results may be explained by the unmet assumptions of classical metapopulation structure underlying the genetic models.     

Intraspecific variation in group size in the blackbuck antelope: the roles of habitat structure and forage at different spatial scales

The main ecological factors that are hypothesized to explain the striking variation in the size of social groups among large herbivores are habitat structure, predation, and forage abundance and distribution; however, their relative roles in wild populations are not well understood. I combined analyses of ecological correlates of spatial variation in group size with analyses of individual behaviour in groups of different sizes to investigate factors maintaining variation in group size in an Indian antelope, the blackbuck Antilope cervicapra. I measured group size, habitat structure, forage, and the occurrence of predators in ten blackbuck populations, and, at a smaller spatial scale, within an intensively studied population. To examine the processes by which these ecological factors influence group size, I used behavioural observations and an experiment to estimate the shape of the relationship between group size and potential costs and benefits to individuals. Group size varied extensively both among and within populations. Analyses of spatial variation in group size suggested that both forage and habitat structure influence group size: large-scale, among-population variation in group size was primarily related to habitat structure, while small-scale, within-population variation was most closely related to forage abundance. Analyses of individual behaviour suggested that larger groups incur greater travel costs while foraging. However, individuals in larger groups appeared to experience greater benefits, namely the earlier detection of a “predator”, a reduction in vigilance, and an increase in the time spent feeding. Overall, these findings suggest that individuals in groups experience a trade-off between predation-related benefits and costs arising from feeding competition. Habitat structure and forage likely influence the nature of this trade-off; thus, variation in these ecological factors may maintain variation in group size. The role of predation pressure and other factors in explaining the remaining variation needs further exploration. Electronic supplementary material The online version of this article (doi:) contains supplementary material, which is available to authorized users.