Artificial neural networks for species identification by taxonomists

This paper is a study of the value of applying artificial neural networks (ANNs), specifically a multilayer perceptron (MLP), to identification of higher plants using morphological characters collected by conventional means. A practical methodology is thus demonstrated to enable botanical or zoological taxonomists to use ANNs as advisory tools for identification purposes. A comparison is made between the ability of the neural network and that of traditional methods for plant identification by means of a case study in the flowering plant genus Lithops N.E. Brown (Aizoaceae). In particular, a comparison is made with taxonomic keys generated by means of the DELTA system. The ANN is found to perform better than the DELTA key generator, for conditions where the available data is limited, and species relatively difficult to distinguish.     

Relationship between species relative abundance and plant traits for an infertile habitat

This study tested whether differences in species abundance at an infertile site could be explained by differences in the species' plant traits. Nine traits were chosen for the analysis based on results of previous studies conducted across soil fertility gradients. The traits were measured for each of seven herbaceous species whose abundance ranged from 5% to 100% of locations occupied in a ridgetop habitat. Using linear regression, significant relationships were found between species relative abundance and each of five traits. In these relationships, a trait explained between 69% and 88% of interspecific variation in abundance. Relatively abundant species had a slower growth rate, smaller shoot mass, higher root to shoot ratio, slower loss of leaf tissue to herbivores and higher infection of roots by mycorrhizal fungi than less abundant species. Using three of these five traits (i.e. shoot mass, mycorrhizal infection and loss of leaf tissue to herbivores) as independent variables in a multiple regression equation explained 99% of interspecific variation in abundance. The latter result indicates that species relative abundance can be explained for a single habitat by choosing traits found to be related to species abundance in previous gradient studies. However, not every trait chosen was significantly related to species abundance. Therefore, a large number of traits may have to be chosen initially to ensure that some subset of these traits can explain species relative abundance.     

Ecological dynamics of extinct species in empty habitat networks. 2. The role of host plant dynamics

This paper explores the relative effects of host plant dynamics and butterfly-related parameters on butterfly persistence. It considers an empty habitat network where a rare butterfly ( Cupido minimus ) became extinct in 1939 in part of its historical range in north Wales, UK. Surviving populations of the butterfly in southern Britain were visited to assess use of its host plant ( Anthyllis vulneraria ) in order to calibrate habitat suitability and carrying capacity in the empty network in north Wales. These data were used to deduce that only a portion (∼19%) of the host plant network from north Wales was likely to be highly suitable for oviposition. Nonetheless, roughly 65,460 eggs (3273 adult equivalents) could be expected to be laid in north Wales, were the empty network to be populated at the same levels as observed on comparable plants in surviving populations elsewhere. Simulated metapopulations of C. minimus in the empty network revealed that time to extinction and patch occupancy were significantly influenced by carrying capacity, butterfly mean dispersal distance and environmental stochasticity, although for most reasonable parameter values, the model system persisted. Simulation outputs differed greatly when host plant dynamics was incorporated into the modelled butterfly dynamics. Cupido minimus usually went extinct when host plant were at low densities. In these simulations host plant dynamics appeared to be the most important determinant of the butterfly's regional extirpation. Modelling the outcome of a reintroduction programme to C. minimus variation at high quality locations, revealed that 65% of systems survived at least 100 years. Given the current amount of resources of the north Wales landscape, the persistence of C. minimus under a realistic reintroduction programme has a good chance of being successful, if carried out in conjunction with a host plant management programme.     

Identifying indicator plant species of habitat quality and invasibility as a guide for peri-urban forest management

A floristic survey has been carried out in a peri-urban forest, the Sonian Forest in Brussels, to identify indicator plant species in the herbaceous layer, which could be used as an aid within the framework of a more sustainable management of the forest. Three hundred and seventy two (372) taxa have been identified, 33 of which are non-native (i.e. non-indigenous species regarding the study area, whether invasive or not). Criteria of habitat quality that have been chosen are the species richness, the commonness of the habitat, based on constitutive species, and its invasibility (vulnerability for invasion). On the basis of a comparison of the value of these criteria when each considered (potential indicator) species is present or not, 17 species have been recognised as reliable indicators of at least one of these three criteria. In particular, vegetation types containing either Anthriscus sylvestris, Galeopsis tetrahit or Senecio ovatus were found to be more susceptible to invasion than other habitats. The way to how the predictability of invasions might be effectively used as a management tool is discussed. Furthermore, we found a positive significant correlation (Bonferroni corrected probabilities) between the species richness and luminosity factor (derived from Ellenberg's indices), and the proportion of grassland and wetland species. The species richness was significantly negatively correlated with the proportion of woodland species. An increase in commonness was significantly correlated with a decrease in the proportion of geophytes. The usefulness of these results as an ecological basis for forest management is discussed.     

Ecological assessment of plant communities by reference to species traits and habitat preferences

Where vegetation is managed for nature conservation, results should be assessed against criteria linked to the objectives of management. An assessment method is presented in which the goal vegetation was defined as a specific biotope limited by a set of biophysical conditions. Vegetation was sampled from field boundary strips intended for the conservation of arable plant communities. Species suited to the specified conditions (suited species) were defined by applying rule sets to a matrix of species with their traits and habitat preferences, compiled from a range of data sources. The proportional contribution of suited species to the total vegetation was calculated for each set of conditions. Scores for each condition (suited species scores) were combined to provide overall comparative site values representing the extent to which the goal vegetation was established. There was close correlation with results of a previous assessment by expert opinion poll. Variation between sites in soil type and cultivation frequency was reflected by differences in individual suited species scores. Ideally, suited species selection would be by traits alone but currently there are insufficient data available. The method is readily applicable to other vegetation types.     

Local plant species richness increases with regional habitat commonness across a gradient of forest productivity

We tested the prediction that forest habitat types with relatively high productivity are not only relatively low in species richness but are also regionally uncommon. This relationship was supported by an analysis of data from 146 forest communities in southern Ontario, Canada. Potential forest habitat productivity was determined based on a classification scheme developed for the Canadian Land Inventory (CLI) project. Vascular plant species richness approximated a unimodal distribution across forest productivity classes with the lowest mean species richness recorded for the two most productive classes. The contemporary regional commonness of forest habitat productivity classes were also displayed as a unimodal frequency distribution. Hence, mean species richness per CLI class was positively correlated with the regional area of land encompassing each of these productivity classes and this relationship was increasingly significant at increasingly larger spatial scales of regional CLI class land areas. These results are consistent with the species pool hypothesis, which postulates that species richness is relatively low in highly productive habitats because such habitats have been relatively uncommon in both space and time and hence, have had relatively little historical opportunity for the origination of adapted species.     

Consistency in the habitat degree of invasion for three invasive plant species across Mediterranean islands

Habitats are known to vary in their vulnerability to invasion by alien plants and different species often colonise distinct habitats. To assess the consistency in the degree of invasion of particular habitats, this study examined the frequency of occurrence and local abundance of three invasive plant taxa: Ailanthus altissima, Carpobrotus spp., and Oxalis pes-caprae across different habitat types on four representative Mediterranean islands. We conducted systematic field surveys recording the presence-absence and cover of these taxa on the islands of Mallorca, Corsica, Sardinia and Crete. Drawing on the results of 5,285 sample points, the frequency of occurrence of the three invaders tends to be higher than expected in urban, ruderal and roadside habitats. In contrast, scrub habitats rarely contain any of the three invaders, indicating that they may be more resistant to invasion. The degree of invasion, determined by the local abundance of an invasive plant in any one habitat, varies according to the identity of the invader and the island. However, based on average abundance, Oxalis pes-caprae exhibits the highest degree of invasion, and Carpobrotus spp. the least. There is no indication that any one of the four islands is more prone to either higher frequencies or abundances of the three invaders. These patterns suggest that anthropogenic changes in Mediterranean islands will increase the vulnerability of certain habitats to invasion and increase the distribution of these three invasive taxa at any of the four islands.     

Applying DNA barcodes for identification of plant species in the family Araliaceae

An effective DNA marker in authentication of the family Araliaceae was screened out of the five DNA regions (matK, rbcL, ITS2, psbA-trnH and ycf5). In the present study, 1113 sequences of 276 species from 23 genera (Araliaceae) were collected from DNA sequencing and GenBank, in which 16 specimens were from 5 provinces in China and Japan. All of the sequences were assessed in the success rates of PCR amplifications, intra- and inter-specific divergence, DNA barcoding gaps and efficiency of identification. Compared with other markers, ITS2 showed superiority in species discrimination with an accurate identification of 85.23% and 97.29% at the species and genus levels, respectively, in plant samples from the 589 sequences derived from Araliaceae. Consequently, as one of the most popular phylogenetic markers, our study indicated that ITS2 was a powerful barcode for Araliaceae identification.     

Applying DNA barcodes for identification of plant species in the family Araliaceae

Genetic variants of tPA (PLAT) and PAI-1 genes have been suggested to be the risk factors for stroke. In the present case-control study we investigated the association of − 7351 C/T polymorphism (rs2020918) and I/D polymorphism of tPA gene and Insertion/deletion polymorphism (4 G/5 G) of PAI-1 gene with genetic predisposition to ischemic stroke. 516 stroke patients and 513, sex and age matched healthy controls were involved in the study. We did not find a significant association of tPA − 7351 C/T polymorphism and PAI-1 4 G/5 G polymorphism with stroke. However, in case of I/D polymorphism significant difference was observed in the genotypic distribution and allelic frequency between the stroke patients and healthy controls. DD genotype and D allele associated significantly with stroke (p = 0.002 and < 0.001 respectively). We also found significant association of I/D polymorphism with intracranial large artery atherosclerosis and stroke of undetermined etiology. Exploring the association between gene-gene interaction (26 combinations including the three variants) and stroke, we found that individuals with CC + 4G4G + DD, CC + 5G5G + ID, CT + 4G5G + ID, CT + 5G5G + II, CT + 5G5G + ID and TT + 4G5G + II had a significantly higher risk of stroke. The results of this study suggest that − 7351 C/T polymorphism of tPA and 4 G/5 G polymorphism of PAI-1 are not associated with stroke, while as DD genotype and D allele of tPA gene are important risk factors for ischemic stroke. Further we found that the subjects with different tPA and PAI genotype combinations displayed a significantly high risk for overall ischemic stroke suggesting that gene-gene interaction involving more variants may change the susceptibility of particular subjects to the disease.     

The nested assembly of plant facilitation networks prevents species extinctions

Facilitation is a positive interaction assembling ecological communities and preserving global biodiversity. Although communities acquire emerging properties when many species interact, most of our knowledge about facilitation is based on studies between pairs of species. To understand how plant facilitation preserves biodiversity in complex ecological communities, we propose to move from the study of pairwise interactions to the network approach. We show that facilitation networks behave as mutualistic networks do, characterized by a nonrandom, nested structure of plant-plant interactions in which a few generalist nurses facilitate a large number of species while the rest of the nurses facilitate only a subset of them. Consequently, generalist nurses shape a dense and highly connected network. Interestingly, such generalist nurses are the most abundant species in the community, making facilitation-shaped communities strongly resistant to extinction, as revealed by coextinction simulations. The nested structure of facilitative networks explains why facilitation, by preventing extinction, preserves biodiversity.     

Phylogenetically poor plant communities receive more alien species, which more easily coexist with natives

Alien species can be a major threat to ecological communities, but we do not know why some community types allow the entry of many more alien species than do others. Here, for the first time, we suggest that evolutionary diversity inherent to the constituent species of a community may determine its present receptiveness to alien species. Using recent large databases from observational studies, we find robust evidence that assemblage of plant community types from few phylogenetic lineages (in plots without aliens) corresponds to higher receptiveness to aliens. Establishment of aliens in phylogenetically poor communities corresponds to increased phylogenetic dispersion of recipient communities and to coexistence with rather than replacement of natives. This coexistence between natives and distantly related aliens in recipient communities of low phylogenetic dispersion may reflect patterns of trait assembly. In communities without aliens, low phylogenetic dispersion corresponds to increased dispersion of most traits, and establishment of aliens corresponds to increased trait concentration. We conclude that if quantified across the tree of life, high biodiversity correlates with decreasing receptiveness to aliens. Low phylogenetic biodiversity, in contrast, facilitates coexistence between natives and aliens even if they share similar trait states.     

Diversity and habitat relationships of hypogeous fungi. III. Factors influencing the occurrence of fire-adapted species

Among the huge array of hypogeous ectomycorrhizal fungi so far documented from Australia, six genera and more than 30 species occur within the family Mesophelliaceae, all of which show various adaptations for surviving in fire-prone landscapes. These mostly endemic fungi are critical to postfire reestablishment of regenerating vegetation, and their fruit-bodies provide essential food resources for diverse ground-dwelling fauna. We developed habitat models for five common representatives of the Mesophelliaceae based on repeat collections of their fruit-bodies from 136 study plots situated along a series of environmental gradients across the south-eastern mainland of Australia. At a meso- or landscape scale, temperature influenced the occurrence of Castoreum radicatum, Mesophellia clelandii and Nothocastoreum cretaceum, with the type of response varying. Below a threshold, C. radicatum preferred sites with cooler mean annual temperatures. In contrast, M. clelandii and N. cretaceum had optimal ranges of temperature, above and below which the probability of detecting them dropped. Also at a landscape scale, C. radicatum was more likely to be detected at sites with lower levels of precipitation during the driest quarter of the year. At a micro-site scale, M. clelandii and N. cretaceum were more likely to occur in stands with an intermediate number of host eucalypt stems, likely relating to successional age of the stand. Sites with a higher number of large fallen trees were more likely to have N. cretaceum, while sites with intermediate litter depths were more likely to have C. radicatum and M. clelandii. Mesophellia glauca and M. trabalis showed no consistent patterns. They are apparently the most broadly adaptable in terms of the independent variables tested. Although fire has been previously suggested to be heavily implicated in the life cycle of several members of the Mesophelliaceae, we found no relationship between time since disturbance by fire and other factors and likelihood of occurrence. Instead, other habitat attributes appeared to be more important in explaining their distribution. The complex and differing responses of the species of Mesophelliaceae studied here, to features of their environment, reinforce the need to manage multiple-use forest landscapes across the region for a diversity of attributes.     

Linking local ecological knowledge and habitat modelling to predict absolute species abundance on large scales

Assessing the spatial structure of abundance of a species is a basic requirement to carry out adequate conservation strategies. However, existing attempts to predict species abundance, particularly in absolute units and on large scales, are scarce and have led to weak results. In this work we present a scheme to obtain, in an affordable way, a predictive model of absolute animal abundance on large scales based on the modelling of data obtained from local ecological knowledge (LEK) and its calibration. To exemplify this scheme, we build and validate a predictive absolute abundance model of the endangered terrestrial tortoise Testudo graeca in Southeast Iberian Peninsula. For that purpose, we collected distribution and relative abundance data of T. graeca using a low cost methodology, such as LEK, by means of interviewing shepherds. The information from LEK was employed to build a predictive habitat-based model of relative abundance. The relative abundance model was transformed into an absolute abundance model by means of calibration with a classical absolute abundance sampling method such as distance sampling. The obtained absolute abundance model predicted the observed absolute abundances values well in independent locations when compared with other works (R ² = 36%) and thus can offer a cost-effective predictive ability. Our results show that reliable habitat-based predictive maps of absolute species abundance on regional scales can be obtained starting from low cost sampling methods of relative abundance, such as LEK, and its calibration.     

Interactions between alien plant species traits and habitat characteristics in agricultural landscapes in Finland

The survival and success of alien plant species is determined by species traits (i.e., invasiveness) and the characteristics of the habitats in the region of introduction (i.e., invasibility). However, little is known about species traits as related to habitat characteristics. We assessed the characteristics of successful invaders and the interaction of environmental factors and life-history traits for alien plant species. The vascular plants were recorded from 52 agricultural landscapes in Finland. We compared the traits of native and alien plant species with Fisher’s exact test and used a three table ordination analysis, RLQ analysis, to relate species traits to environmental conditions. Species were clustered according to their position on the RLQ axes, and the clusters were tested for phylogenetic independence. The successful alien plant species were associated with life form and preferences for moisture and nitrogen, but the trait composition varied according to the habitat type. Two RLQ axes explained 80.5% of the variation, and the species traits were significantly associated with environmental variables. The clustering showed that the occurrence of alien plant species in agricultural habitats was driven by invasion history, traits related to dispersal (dispersal type, seed mass) and habitat preferences, as well as environmental features, such as geographical location, temperature and the quality and disturbance regime of the habitats. All clusters were phylogenetically non-independent. Thus, the clusters of alien species comprised species of diverse taxonomic affinities, although, they shared the traits explaining their occurrence in particular habitats. This information is useful for understanding the link between species traits and the environmental conditions of the habitats, and complexity of the invasion process.     

The habitat and conduit functions of roads in the spread of three invasive plant species

Nonnative plant species commonly occur along roadsides, and populations are often assumed to invade by spread along the road axis. To distinguish between the function of roadsides as movement corridors and as habitat, nonnative plant species were surveyed along roads in deciduous forest sites in southeastern Ohio, USA. The importance of road proximity was tested by comparing nonnative species abundance in 100 m transects along roads with transects in undisturbed forest. Nonnative species were most abundant and most frequently observed in roadside sites in valleys. Three common species were chosen for closer scrutiny. In a seed sowing experiment roads and open sites proved to be better locations for the germination and growth of Microstegium vimineum than non-roadside and closed-canopy sites. Tussilago farfara and Rosa multiflora occurred in a small number of disjunct patches suggesting infrequent arrival in the sampled transects. Both species were strongly clustered at scales consistent with diffusive spread by vegetative growth and short-range seed dispersal. Comparisons of distributions parallel and perpendicular to roads showed no evidence for enhanced dispersal along the road axis. Microstegium distributions were correlated with local light availability implying site saturation. Microstegium micro-distributions suggested that spread along the road axis was facilitated by movement of dormant seeds in road maintenance. Thus, roadsides appear to function as both habitat and a conduit for population expansion, with the rate of spread dependent on the life history of the individual species. These results suggest a hierarchical process of regional invasion, with different dispersal mechanisms functioning at different spatial scales.     

Host plant use by the Heath fritillary butterfly, Melitaea athalia: plant habitat, species and chemistry

We present a study of habitat use, oviposition plant choice, and food plant suitability for the checkerspot butterfly Melitaea athalia Rottemburg (Lepidoptera: Nymphalidae) in Åland, Finland. We found that in Åland, unlike in the mainland of Finland and many parts of its range, M. athalia flies mainly in open meadows. When offered an array of plants in a large (32 × 26 m) field cage, they predominately oviposited upon Veronica chamaedrys L., V. spicata L. and Plantago lanceolata L. (Plantaginaceae), which grow in open meadows. The relative abundance of the butterfly in Åland, and its habitat and host plant use there, may reflect local adaptation to land use practices and geology that maintain clusters of small open meadows with little successional change. At the scale of a plant patch, preferred species were used as frequently in mixed species patches as in mono-specific patches, and more oviposition occurred in open than in grassy patches. All of the host plants used by M. athalia are defended by iridoid glycosides (IGs). However, oviposition choice among species and among individual plants within species was largely independent of IG concentration. This contrast with the more discerning congener, M. cinxia, supports the idea that host discrimination decreases with increasing host range. Finally, although the adult butterflies chose specific plant species for oviposition, as larvae they performed well on twelve out of thirteen species of plants, including both known hosts and related novel plants that occur in Åland, indicating a much wider range of larval food plant species than adult oviposition species.     

Ecological dynamics of extinct species in empty habitat networks. 1. The role of habitat pattern and quantity, stochasticity and dispersal

We examined a remnant host plant ( Primula veris L.) habitat network that was last inhabited by the rare butterfly Hamearis lucina L. in north Wales in 1943, to assess the relative contribution of several spatial parameters to its regional extinction. We first examined relationships between P. veris characteristics and H. lucina eggs in surviving H. lucina populations, and used these to predict the suitability and potential carrying capacity of the habitat network in north Wales. This resulted in an estimate of roughly 4500 eggs (ca 227 adults). We developed a discrete space, discrete time metapopulation model to evaluate the relative contribution of dispersal distance, habitat and environmental stochasticity as possible causes of extinction. We simulated the potential persistence of the butterfly in the current network as well as in three artificial (historical and present) habitat networks that differed in the quantity (current and X3) and fragmentation of the habitat (current and aggregated). We identified that reduced habitat quantity and increased isolation would have increased the probability of regional extinction, in conjunction with environmental stochasticity and H. lucina 's dispersal distance. This general trend did not change in a qualitative manner when we modified the ability of dispersing females to stay in, and find suitable habitats (by changing the size of the grid cells used in the model). Contrary to most metapopulation model predictions, system persistence declined with increasing migration rate, suggesting that the mortality of migrating individuals in fragmented landscapes may pose significant risks to system-wide persistence. Based on model predictions for the present landscape we argue that a major programme of habitat restoration would be required for a re-established metapopulation to persist for >100 years.