Obituaries

Abstract

It has been suggested that bryophyte communities differ from those of higher plants, in that species coexistence is not limited by inter-specific competition. To test this hypothesis, bryophyte ‘lawns’ were-sampled at six locations in southern New Zealand. At each site, 625 quadrats, each ca 1 X 1cm, were arranged on a contiguous grid.

Variance in quadrat richness was significantly less than expected on a random basis at four sites, with the same trend in a fifth, i.e. quadrat richness was relatively constant. Since a patch model was used to reduce the effects of micro-environmental variation and spatial autocorrelation, this suggests that competition was restricting coexistence. The degree of restriction was similar to that found previously in grass lawn communities.

Species association was calculated on a patch basis, examining only small spatial-scale deviations from expectations. At four- sites this gave only negative associations. At the other two sites there were both positive and negative associations, but the pattern of associations gave no indication of Root/Pianka-type guilds.

Examination of the variance in guild proportions, using taxonomic (moss vs liverwort) and morphological (prostrate vs erect) guilds, gave- no indication of significant guild structure for any site (i.e. in the overall analyses for each site, index RV_gp was very close to the null-model expectation of 1.0).

Searches using the Wilson- Roxburgh method failed to reveal any significant intrinsic guild classification. That is, there was no indication of groups of species within a community that tended to be mutually exclusive because of similarity in resource use. Thus, lack of guild proportionality using a priori guilds was not due to the use of an inappropriate guild classification, but to lack of a guild structure in the communities.

It is concluded that there is community structure among bryophytes, in that species exclude each other to the same degree as higher-plants do in their communities. However, there is no evidence of structuring of these bryophyte communities into guilds; it seems that bryophyte species all form one guild.     

The effects of individual interactions and habitat preferences on spatial structure in a grassland bird community

Spatial structure and the distribution of individuals within a community might be influenced by several factors such as habitat heterogeneity and local interactions among individuals of the same and different species. We investigated the spatial distributions of eight bird species in a grassland community during the breeding season and examined whether the spatial distributions of individuals were influenced by interactions among neighboring individuals or different habitat preferences of different bird species.
In order to identify the effects of the interactions among neighboring individuals and habitat preference, we developed a randomization test in which species identifications were randomly allocated to the observed individual positions within areas with the same vegetation structure. The randomization test indicated that individuals tend to have territories near the territories of individuals of the same species or of a particular species more frequently (or less frequently) than those expected from random distributions of individuals.
Among these associations, only one case was explained by individual interactions, and 19 cases were explained by habitat preference.
The results suggest that both individual interactions and habitat preference affected the spatial distributions of individuals and possibly influence the species compositions and diversity in grassland bird communities.     

Assessing the relative contribution of functional divergence and guild aggregation to overall functional structure of species assemblages

The study of functional structure in species assemblages emphasizes the detection of significant guild aggregation patterns. Thus, protocols based on intensive resampling of empirical data have been proposed to assess guild structure. Such protocols obtain the frequency distribution of a given functional similarity metric, and identify a threshold value (often the 95th percentile) beyond which clusters in a functional dendrogram are considered as significant guilds (using one-tailed tests). An alternative approach sequentially searches for significant differences between clusters at decreasing levels of similarity in a dendrogram until one is detected, then assumes that all subsequent nodes should also be significant. Nevertheless, these protocols do not test both the significance and sign of deviations from random at all levels of functional similarity within a dendrogram. Here, we propose a new bootstrapping approach that: (1) overcomes such pitfalls by performing two-tailed tests for each node in a dendrogram of functional similarity after separately determining their respective sample distributions, and (2) enables the quantification of the relative contribution of guild aggregation and functional divergence to the overall functional structure of the entire assemblage. We exemplify this approach by using long-term data on guild dynamics in a vertebrate predator assemblage of central Chile. Finally, we illustrate how the interpretation of functional structure is improved by applying this new approach to the data set available.     

Trophic Structure of One Deep-sea Benthic Fish Community in the Eastern Canadian Arctic: Application of Food, Parasites and Multivariate Analysis

Synopsis As new arctic marine fisheries develop there is need for a comprehensive ecosystem approach to long-term management. This approach recognizes the importance of community interactions such as food web structure and trophic patterns. We determined whether hierarchical clustering (guild formation) is an effective method of trophic evaluation in deep-sea Artic fish communities using stomach content and parasite data with size class, and evaluated the application of endohelminth communities (parasite species transmitted in the food) as indicators of trophic status. Cluster analysis using food group abundance with size class of fish revealed the presence of 11 guilds within the community, however the same analysis using parasite data showed little correlation between food and parasites. Redundancy analysis (RDA) within the 11 guilds also revealed no significant correlations between food group and parasite abundance suggesting that this type of ordination is not suited for environments containing mainly generalist feeders. RDA of individual taxa without a priori guild designation found that taxa in benthic deep-sea communities are defined by their ability to exploit prey species in more than one habitat zone. Benthic fish species were significantly correlated with benthic food groups and parasites that utilize benthic intermediate hosts whereas benthopelagic–pelagic species fed on a higher diversity of prey species and were infected by a larger number of non-host specific parasites. Eigenanalysis and Monte Carlo results showed that parasites and food groups are highly correlated, indicating that parasite community analysis is an effective tool for predicting feeding strategies in Arctic marine environments. It also suggests that in most cases endoparasite infections alone could be used for trophic evaluation in the absence of stomach content data.     

Substratum patch selection in the lacustrine mussels Elliptio complanata and Pyganodon grandis grandis

1. Sediment selection was investigated under controlled conditions in two common lake-dwelling species of freshwater mussels (Bivalvia: Unionidae), Elliptio complanata and Pyganodon grandis grandis .
2. Sediment choice was determined in six independent experiments under controlled conditions by distributing mussels randomly or evenly in tanks containing patches of sand and mud, and following their movement among sediment patches in experiments lasting between 30 and 45 days.
3. In all experiments, both species were found most frequently in muddy sediment patches. Movement toward muddy patches occurred rapidly: an average of nearly 80% of Pyganodon grandis grandis were found in mud after 30 days. Elliptio complanata moved rapidly to patches of mud at the start of experiments, but occupation of muddy sediments appeared to decrease after about 30 days.
4. Our results contrast with many field studies that suggest populations of lake-dwelling freshwater mussels infrequently inhabit mud and silt. We therefore postulate that large-scale mussel distribution in lakes is influenced most strongly by factors other than sediment composition.     

Spatial and temporal organisation of the pre-dispersal seed predator guild in a perennial legume, Vicia tenuifolia

Abstract.  1. The variability of species combinations and membership assembly patterns in an insect pre-dispersal seed predator guild were studied at various spatial and temporal resolutions using presence/absence and abundance data in null models. The guild consists of seven beetle species (four bruchids and three curculionids) and one moth species that live on a perennial vetch, Vicia tenuifolia Roth (Leguminosae).
2. The seed predator guild varied considerably in the number of members and species combinations in space and time, and, contrary to expectations, there was no evidence of interspecific competition among guild members, supporting the view that other processes, including chance events, could contribute to guild formation.
3. It is concluded that, apart from the possibility of stochastic co-occurrence, it is the narrow host specialisation that constrains seed predator members to participate in the guild, and small differences in habitat preference can also lead to spatial and temporal variation.     

Generating and testing predictions about community structure: Which theory is relevant and can it be tested with observational data?

Three issues are discussed relevant to the controversy over using null models and observational data on guild structure to test community-level predictions based on limiting similarity theory. First, I argue that most limiting similarity theory is not based on reasonable assumptions for plants and that the theory that is relevant does not generate any predictions about expected guild proportionality on a small spatial scale. Therefore, regardless of adequacy of the statistical methods, the predictions being tested by the body of literature using null models to test for niche limitation are unlikely to be relevant in most plant comunities. Second, assuming that the predictions are after all worth being tested, I argue that most tests using the guild approach do not provide adequate explanations of how the defined guilds could lead to greater competition within vs. between guilds. If this is not true for the particular guilds used, the predictions of guild proportionality or size constancy will not be valid. Third, I address the controversy over whether field experiments can provide more solid evidence than observational data about the role of competition in determining community structure by (1) suggesting methods of dealing with potential drawbacks of field experiments, and (2) suggesting alternative experimental approaches for directly addressing issues about community structure.     

Testing for community structure: A bayesian approach

Variance deficit in richness is not reliable evidence for niche limitation. The main problem is the effect of limitation to individual plant module packing at small quadrat sizes, though environmental and historical patchiness can also confound the results unless patch models are used. More reliable approaches to community structure (assembly rules) are those that take into account the characters of the species—guild proportionality, texture convergence and limiting similarity test. A informally Bayesian approach to community structure is advocated, accepting that some hypotheses have greater prior probability, such as module-packing limitations, but using all the evidence available to estimate the likelihood of all the ecological hypotheses. The problem of environmental patchiness can be largely overcome by patch models. These models also overcome the problem of spatial autocorrelation, because they are conservative in such situations. For guild proportionality analyses, the possibility of sampling bias is a particular problem; this should be borne in mind during sampling, and checks made for the possibility in analyses. In devising Monte Carlo tests for community structure, permutation tests (i.e. randomisation tests, using sampling without replacement) are theoretically and practically preferable to bootstrap tests (i.e. using sampling with replacement). The best test is the simplest one that incorporates the intended null model and uses the intended test statistic. Experiments have intuitive appeal, but field experiments have several severe drawbacks, including the inevitability of artefacts, and insoluble problems of when to record. In practice, field experiments have told us surprisingly little about community structure.     

Limitations to species coexistence: evidence for competition from field observations,using a patch model

Abstract. Plant community structure in four dune slacks is examined for evidence of competition, expressed as local constraint on species biomass. Such constraint would result in a low variance in total quadrat biomass, compared to a null model. A method of analysis is introduced, using a patchbased null model, which is intended to discount most of the effects of environmental heterogeneity. The method is applied to data in which previous methods had failed to find such structure. There was significant biomass constraint (i.e. low variance in biomass) examining total plant-community biomass in Site 1, and there was a trend in this direction also in Site 2. When guilds defined on morphology were examined, significant biomass constraint was found for the Rhizomatous guild in Site 2. There was a trend to low variance for this guild also in Site 4, but it was significant only in a one-tailed test. After a consideration of other mechanisms, it is concluded that the most likely explanation for this biomass constraint is interspecies competition. Site 2, which showed the strongest evidence of biomass constraint in the present analyses, was also the site that demonstrated a significant deficit of variance in richness in earlier work, and was the most species-rich of the four sites and the least disturbed. The advantages of the new method, for obtaining evidence on competition, are discussed in comparison to pot experiments and field-perturbation experiments. We conclude that community structure can be seen in the pattern of species biomasses provided that a suitable method of analysis is used, i.e. examining small-scale effects and excluding the effects of environmental heterogeneity. In some cases such effects can be related to guild membership. We hypothesise that constraint on biomass is likely to be stronger in more species-rich and less disturbed communities.     

A maximum entropy analysis of protein orientations using fluorescence polarization data from multiple probes

Techniques have recently become available to label protein subunits with fluorescent probes at predetermined orientation relative to the protein coordinates. The known local orientation enables quantitative interpretation of fluorescence polarization experiments in terms of orientation and motions of the protein within a larger macromolecular assembly. Combining data obtained from probes placed at several distinct orientations relative to the protein structure reveals functionally relevant information about the axial and azimuthal orientation of the labeled protein segment relative to its surroundings. Here we present an analytical method to determine the protein orientational distribution from such data. The method produces the broadest distribution compatible with the data by maximizing its informational entropy. The key advantages of this approach are that no a priori assumptions are required about the shape of the distribution and that a unique, exact fit to the data is obtained. The relative orientations of the probes used for the experiments have great influence on information content of the maximum entropy distribution. Therefore, the choice of probe orientations is crucial. In particular, the probes must access independent aspects of the protein orientation, and two-fold rotational symmetries must be avoided. For a set of probes, a "figure of merit" is proposed, based on the independence among the probe orientations. With simulated fluorescence polarization data, we tested the capacity of maximum entropy analysis to recover specific protein orientational distributions and found that it is capable of recovering orientational distributions with one and two peaks. The similarity between the maximum entropy distribution and the test distribution improves gradually as the number of independent probe orientations increases. As a practical example, ME distributions were determined with experimental data from muscle fibers labeled with bifunctional rhodamine at known orientations with respect to the myosin regulatory light chain (RLC). These distributions show a complex relationship between the axial orientation of the RLC relative to the fiber axis and the azimuthal orientation of the RLC about its own axis. Maximum entropy analysis reveals limitations in available experimental data and supports the design of further probe angles to resolve details of the orientational distribution.     

Analysis of comparative data using generalized estimating equations

It is widely acknowledged that the analysis of comparative data from related species should be performed taking into account their phylogenetic relationships. We introduce a new method, based on the use of generalized estimating equations (GEE), for the analysis of comparative data. The principle is to incorporate, in the modelling process, a correlation matrix that specifies the dependence among observations. This matrix is obtained from the phylogenetic tree of the studied species. Using this approach, a variety of distributions (discrete or continuous) can be analysed using a generalized linear modelling framework, phylogenies with multichotomies can be analysed, and there is no need to estimate ancestral character state. A simulation study showed that the proposed approach has good statistical properties with a type-I error rate close to the nominal 5%, and statistical power to detect correlated evolution between two characters which increases with the strength of the correlation. The proposed approach performs well for the analysis of discrete characters. We illustrate our approach with some data on macro-ecological correlates in birds. Some extensions of the use of GEE are discussed.     

An investigation of dominance in phytoplankton using the PROTECH model

1. Using a model (PROTECH-C) that simulates the simultaneous daily growth of eight phytoplankton species, the following hypotheses were tested: (i) for each given set of simulated conditions, the species with the most appropriate trait, as predicted by a functional group classification, should dominate the community; (ii) with removal of this dominant species, the next best-adapted species should dominate and should be from the same, or a close, functional group where available; (iii) a reduction in the inoculum size of the initially dominant species will not prevent its eventual dominance of the community.
2. For clearer insight into the mechanisms underlying these community processes, a functional group classification based upon species morphology has been used to produce a matrix analogous to Grime's CSR (C, competitor; S, stress tolerator; R, ruderal) paradigm. The effects upon this phytoplankton community of temperature, grazing, limiting light and nutrients over a simulated year were recorded.
3. The results supported all three hypotheses. It was found that, for a given selective constraint, functional traits provided excellent predictors of the dominant types. Also, under conditions of resource competition, the number of functional groups represented decreased. Competition was greatest within functional groups where niche overlap was high, but one species was always clearly the strongest competitor, i.e. its superiority over its nearest functional competitor was regularly expressed even when the difference in inoculum size was great (1000-fold). These conclusions emphasized the power that trait selection can have in the shaping of communities.     

Exploring co‐occurrence of closely‐related guild members in a fragmented landscape subject to rapid transformation

Quantifying biotic responses to landscape transformation is a major research focus. Most past studies have explored co‐occurrence of entire communities of a given group (e.g. birds) within largely intact ecosystems or over a limited time‐frame. By contrast, here we use data from a 15 yr experimental study, to explore intra‐guild co‐occurrence of six closely‐related and functionally‐similar sets of birds within 55 woodland fragments. Areas surrounding these remnants are undergoing transformation from grazed paddocks to Pinus radiata plantations, leading to a novel assemblage of forest and woodland birds. We sought to determine if the occurrence of a given species in a guild influenced the occurrence of other closely‐related species in that guild, and through this relationship whether there was evidence of co‐occurrence between species. After controlling for environmental and habitat variables which can affect species occurrence like time since commencement of landscape transformation, patch size and vegetation type, we found the occurrence of a given species was influenced by the occurrence of a closely‐related species in the same guild. Co‐occurrence varied among bird guilds and included: 1) positive co‐occurrence in which occurrence of one species within fragments positively affected the occurrence of another closely‐related guild member (e.g. eastern and crimson rosellas); and 2) negative co‐occurrence in which the occurrence of one species was negatively associated with the occurrence of another within the same guild (e.g. willie wagtail and grey fantail). We also identified interactions between patch size and species recording frequency within members of two guilds. For example, modelling of conditional recording frequency revealed the eastern rosella increased with increasing recordings of the crimson rosella in large patches, but decreased with increasing recordings of the crimson rosella in small patches. Our results provide empirical evidence of co‐occurrence among guild members and underscore the complexity of biotic responses to landscape transformation.     

The effects of fire on communities,guilds and species of breeding birds in burnt and control pinewoods in central Italy

Breeding bird communities in burnt and unburnt residual pinewoods were studied over 3 years by line-transect method, following a catastrophic fire event in Castelfusano (Rome, Central Italy; July 2000). We applied bootstrap procedures to evaluate whether the observed data were true or just produced by chance, and then examined the emerging patterns at three levels: community, guild and species levels. At the community level, fire acted on breeding bird communities by altering especially the total abundance patterns: the species abundance decreased in the burnt pinewood compared to the residual one, but other parameters were not significantly affected by fire. As a consequence of fire, the destruction and structural simplification of the canopy and shrubby component, as well as the increase of edge habitat and patchiness at landscape scale, induced a turnover in species between pinewoods. Species turnover was higher at the burnt than at the residual pinewoods, during all the 3 years of study. At the guild level, the forest species decreased strongly in terms of richness and abundance in the burnt pinewoods, contrary to the edge and open habitat species which increased in terms of richness, abundance and evenness. Edge species showed the highest turnover in burnt pinewood during the whole period of study. At species level, after an a priori subdivision (based on bibliographic search) of the various species in two ecological guilds (forest versus edge species), it was found that an a posteriori statistical analysis confirmed the expected trend, i.e. that the species which decreased significantly in burnt pinewood were essentially the forest species, whereas the species which increased were essentially the edge/open habitat ones. Overall, in order to investigate the effects of fire catastrophes on birds, the guild approach seems more exhaustive than the taxonomic community approach, where intrinsic confounding trends are present.     

中国远古人类的进化

本文从中国人类化石的年代顺序、共同形态特征、渐进变化、形态的异样性、镶嵌性、与其他地区的基因交流和古文化证据等方面论证了中国人类进化以连续性为主,还与世界其他地区之间有渐增的基因交流。本文还附带讨论了近年云南发现的古猿化石和湖北发现的人科头骨化石。     

An Ant Mosaic Revisited: Dominant Ant Species Disassemble Arboreal Ant Communities but Co-Occur Randomly

The spatial distributions of many tropical arboreal ant species are often arranged in a mosaic such that dominant species have mutually exclusive distributions among trees. These dominant species can also mediate the structure of the rest of the arboreal ant community. Little attention has been paid to how diet might shape the effects of dominant species on one another and the rest of the ant community. Here, we take advantage of new information on the diets of many tropical arboreal ant species to examine the intra- and inter-guild effects of dominant species on the spatial distribution of one another and the rest of the tropical arboreal ant community in a cocoa farm in Bahia, Brazil. Using null model analyses, we found that all ant species, regardless of dominance status or guild membership, co-occur much less than expected by chance. Surprisingly, the suite of five dominant species showed random co-occurrence patterns, suggesting that interspecific competition did not shape their distribution among cocoa trees. Across all species, there was no evidence that competition shaped co-occurrence patterns within guilds. Co-occurrence patterns of subordinant species were random on trees with dominant species, but highly nonrandom on trees without dominant species, suggesting that dominant species disassemble tropical arboreal ant communities. Taken together, our results highlight the often complex nature of interactions that structure species-rich tropical arboreal ant assemblages.     

Patterns of morphological traits shaping the feeding guilds in the intertidal mudflat fishes of the Indian Sundarbans

Broad-scale patterns of resource utilization and the corresponding morphological evolution is a result of an integral relationship among form and function. In addition, there is also an inherent role of the latter in determining species co-interaction and assemblage pattern that forms an integral aspect of ecological research. The present study aimed to evaluate the ecomorphological relationship among 37 fish species inhabiting the intertidal mudflats of the Indian Sundarbans by outlining the following objectives: (i) identifying and characterizing feeding guilds/groups and (ii) understanding the inter-relationship between morphometry with (a) the established feeding guild classifications and (b) observed prey taxa (that characterizes these feeding groups) for determining the role of morphometry in prey acquisition followed by (iii) the evaluation of their potential phylogenetic convergence among the species. For the first objective, two approaches for feeding guild classification were made (3-Guild and 8-Guild) for assessing the prediction accuracy of morphological characters in identifying the different guilds. While the former was based on troph values, the latter classification mode relied on the similarities in diet composition among the different fish species. For addressing the second objective, we employed two different models namely, linear discriminant (LDA) and redundancy analysis (RDA). While the LDA model tested the prediction accuracy of morphological traits in classifying the different feeding guilds, RDA was applied to model the correlation between the morphological traits and the prey categories. In the LDA model, morphological characters showed higher accuracy (78.4%) in classifying three feeding groups rather than eight feeding groups (73%). Following this, the RDA model (explaining 79.78% of constrained variance) showed gill raker intensity, protrusion length, head depth, caudal peduncle, eye diameter and inter-orbital distance to be highly associated with selection of specific prey types by species, thereby characterizing a particular feeding guild. However, generalized linear models testing for correlation between troph value and feeding groups showed substantial variation (90.35%) in the dietary index being explained by the 8-Guild classification. Hence, our study maintains the assumption that broad morphological differentiation acts as one of the underlying processes resulting in dietary variations that results in the varying modes of resource utilization by the coexisting species, thereby determining the structure of a trophic guild. Furthermore, it also suggests that in terms of prey abundance or selectivity, the 8-Guild model is much more conducive in representing the feeding habits of the species while the morphological traits reflected a relatively broader scheme of classification, (i.e., 3-Guild model) with certain traits being phylogenetically conserved within these groups.     

Associations between avian functional guild response and regional landscape properties for conservation planning

This project facilitates a regional approach to conservation planning in Pennsylvania based on avian breeding habitat selection. The objectives were to: (1) determine the sensitivity of spatial pattern in avian diversity to changing thresholds of intra-guild species richness and (2) relate change of spatial pattern in avian diversity with landscape characteristics of bird Atlas blocks. Two state-wide spatial data layers, based on Landsat satellite data were constructed for this study. These regional landscape data were compared to Breeding Bird Atlas data from 1983 to 1989 using a geographic information system. Breeding bird data were recorded from 4928 blocks that form a grid covering Pennsylvania. Correlation analysis reduced landscape variables to 12 originally derived from forest, urban, roads, streams, and topographic data.Avian functional response guilds were used to analyze associations between breeding bird data and landscape variables. Functional response guilds were created by grouping organisms based on shared habitat preferences or behavioral characteristics. Most of the 18 avian guilds identified for this study were based on shared structural resource characteristics of preferred breeding habitat. Preferred structural resources frequently included the amount and type of forest. For this study, guilds separate resource characteristics by: (1) primary habitat (i.e. forest interior, forest edge), (2) area sensitivity (i.e. forest and grassland), (3) migratory status (i.e. resident, temperate, and tropical), and (4) nest placement (i.e. canopy nester, forest ground nester). Wetland obligate species were treated as a separate guild. Breeding Bird Atlas blocks were tabulated with respect to the number of species present from each guild. For a given guild, the number of its species in a block is termed guild-specific species richness. Sample blocks having high species richness for a given guild often occur adjacent or in close proximity forming spatial clusters in the landscape. Spatial coherence (adjacency/proximity) among the blocks forming these islands is shared guild-specific richness. Spatial clustered blocks of each guild represent areas that presumably possess required resources for members of that guild. Blocks having high intra-guild richness were evaluated with a group of block-level continuous variables using multiple logistic regressions. Logistic regression results indicate that a convincing connection exists between landscape properties of Breeding Bird Atlas blocks and habitat selection characteristics of guild members. Percent of forest cover and mean elevation were the most important habitat characteristics influencing intra-guild richness for most of the guilds tested. Concordance values from logistic regression were used to determine the strength of each guild model. Concordance, the proportion that represents the percent of correct guild richness predictions versus incorrect predictions, suggests a relationship between guild-rich clusters and habitat resources required by each guild. The highest concordance was for the exotics guild at 76.3% and the next highest was 74.8% for the grassland area sensitive guild. This signifies a 75% certainty that landscape variables could predict occurrence of a guild-rich block. Eight more guilds had concordance values greater than 65%.By using a guild approach, this study goes beyond total diversity to the more informative structural and functional diversity of guilds. Spatially clustered blocks of high species richness for a particular guild are more indicative of habitat availability and quality than would be the case for overall species richness. Clusters of blocks having high intra-guild species richness become candidate areas for conservation efforts.     

Guilds and assembly rules in lawn communities

Abstract. The community structure of 11 lawn sites in New Zealand and Fiji was examined in terms of guilds, seeking assembly rules based on guild proportionality. First, associations were analysed, using a new patch model which examined the mean of associations within patches of about 4 cm x 4 cm. As expected from the previously-demonstrated existence of niche limitation at this scale in lawns, the majority of associations between individual species were negative. Even in a lawn only four months old there were significant associations (mostly negative, and one positive). At some sites those species with the most negative associations were those whose morphology might be expected to restrict co-existence with other species. Much ecological theory is based on the assumption that there are limitations to coexistence, related to the resource usage of the species, such that species that are too similar in resource use are less likely to coexist. This theory was tested by looking for evidence of guild proportionality. We defined two guilds, graminoids and forbs. Using these guilds, three of the sites showed significant evenness in proportional representation from the two guilds in quadrats containing four species, i.e. there was less variation in the graminoid:forb ratio than expected on a random basis. Inclusion of species-poor quadrats in a joint analysis over all richness categories overwhelmed this effect, though there was still a strong trend for two of the New Zealand sites: the two species-rich grass lawns. Preliminary analyses of alternative guild delimitations gave no indication that other guild delimitations would have revealed stronger community structure. There was a strong and significant tendency in most sites for the graminoid guild to be more strongly represented in species-poor quadrats, an effect caused by species frequencies.     

Silica and diatom growth in Lough Neagh: the importance of internal recycling

1. Lough Neagh is a large eutrophic lake with a high dissolved silica (SiO2) concentration arising from the basaltic rock in parts of the catchment and the normal winter maximum concentration is over 8 mg L−1. Based on frequent observations between 1974 and 1997, the annual cycles of SiO2 input, uptake and release are explored.
2. Large spring blooms of the planktonic diatoms Aulacoseira subarctica Haworth and Stephanodiscus astraea occur every year and are terminated by SiO2 depletion. Although there are periods when one diatom species has dominated the other, over the period as a whole, the two diatoms appear to be in a stable coexistence.
3. A model of the effect of river inputs on lake concentration shows that without SiO2 release from the sediment, the maximum spring dissolved SiO2 concentration (and hence diatom crop) would be only about one-third of that actually observed.
4. It is concluded that within-lake processes play a large role and are potentially more variable than catchment processes in determining the available SiO2 in Lough Neagh.
5. The role of benthic animals, especially Chironomus anthracinus, in the sediment SiO2 release process is discussed. Field data suggest the SiO2 release rate is highly sensitive to temperature, but this could be partly caused by an interaction between temperature and animal activity.