We tested the hypothesis that functional attributes of living leaves provide a basis for predicting the decomposition rate of leaf litter. The data were obtained from standardized screening tests on 38 British herbaceous species. Graminoid monocots had physically tougher leaves with higher silicon contents than did herbaceous dicots, and this corresponded with the lesser decomposibility of the former. Total base content of living leaves was a good predictor of litter decomposition rate, but the evolutionary and ecological basis for this relationship appeared to differ between graminoid monocots and herbaceous dicots. In the monocots, litter decomposition rate was strongly predicted by leaf potassium content, which appeared to reflect other growth-related plant attributes such as seedling maximum relative growth rate, foliar nitrogen and phosphorus content, specific leaf area and short leaf lifespan. In the dicots the relationship between total leaf base content and litter decomposition rate was not unambiguously explained by growth-related leaf attributes, possibly because of the considerable calcium uptake by dicots, which varies according to calcium availability in the soil.

Chickweed plants (Stellaria media Vill.), either inoculated with cucumber mosaic virus (CMV) or not, were grown to reproductive maturity at a range of densities, both in monoculture and in mixture. CMV infection did not influence the vegetative growth of plants grown in monoculture. Similarly, the total vegetative production of a population containing infected and healthy plants was the same as that of a monoculture of healthy plants. However, infected plants were disadvantaged when in competition with healthy plants, the disadvantage increasing with density and the proportion of healthy plants. Investment in sexual reproduction was reduced in infected plants at all densities and in healthy plants at high densities. In very hard conditions, plants did not even produce seeds.

Several different fructan and sucrose hydrolysing enzyme activities were induced in roots and stubble (mainly leaf sheaths) of Lolium perenne L. plants after defoliation. Among those activities, a fructan-β-fructosidase (EC 3.2.1.80) that hydrolyses predominantly β-(2-6)-fructosyl-fructose linkages (6-FEH) was purified from the stubble. The use of the substrate 6,6-kestotetraose and high-performance anion-exchange chromatography with pulsed amperometric detection allowed linkage-specific screening and sensitive analysis of enzyme activity. A 6-FEH was extensively purified to yield one protein band as revealed by one-dimensional sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE). The 6-FEH was separated from the contaminating β-(2-1)-linkage-specific fructan-β-fructosidase and invertase activities (EC 3.1.2.26) by ammonium sulphate precipitation, lectin-affinity, anion-exchange and size-exclusion chromatography. The purified 6-FEH was a glycoprotein with an apparent molecular mass of 65000, as determined by size-exclusion chromatography, and of 69000 by SDS–PAGE. The 6-FEH had an activity optimum in the range of pH 5·1 to 5·6. Temperatures above 30°C affected the stability of the enzyme activity; however, its temperature stability was increased in the presence of 6,6-kestotetraose. The purified 6-FEH activity hydrolysed the β-(2-6)-linkages in 6,6-kestotetraose and (1&6)-kestotetraose at rates five times faster than the β-(2-1)-linkages in 1,1-kestotetraose and (1&6)-kestotetraose. Fructose up to 50 mM did not affect 6-FEH activity; conversely, sucrose substantially inhibited the enzyme activity. Other disaccharides did not affect 6-FEH. It is suggested that sucrose might modulate 6-FEH activity in vivo.

This paper reports effects of molybdenum and vanadium on Anabaena variabilis Kutz. (ATCC 29413) and Anabaena cylindrica Lemm. (PCC 7122), previously grown in NH4+-containing medium deficient in Mo and V (−Mo −V) and then transferred to media lacking combined nitrogen. Diazotrophic growth of A. variabilis was dependent on the presence of Mo or V, with little growth occurring in their absence. Neither Mo nor V was required for heterocyst differentiation in either cyanobacterium, and the highest heterocyst frequencies occurred in cultures showing the lowest growth rates. Although A. cylindrica also showed Mo-dependent diazotrophic growth and a low rate of growth in the absence of added Mo, the latter was inhibited by addition of 1 μM V. The rate of acetylene reduction by A. variabilis increased rapidly following addition of V to −Mo−V cultures, and increased ethane formation was detected in the acetylene reduction assays. Addition of Mo resulted in a smaller initial increase in the rate of acetylene reduction, but this was also accompanied by increased ethane formation, followed by a slow and much more extensive increase in acetylene reduction with a decrease in ethane formation. These observations indicate that nitrogenase polypeptides, probably those of a V-nitrogenase, are synthesized in A. variabilis in the absence of added Mo or V, that added Mo can be incorporated into the V-nitrogenase, and that synthesis of Mo-nitrogenase is dependent on the presence of Mo. By contrast, A. cylindrica appears either to synthesize only a Mo-nitrogenase, even in the absence of Mo (though 1 μM V inhibits its synthesis) or to synthesize a Mo- and V-independent nitrogenase, but only in the absence of added Mo and V. Western blots using antisera against the Mo-Fe and Fe proteins of the Mo-nitrogenase of Rhodospirillum rubrum indicated the presence of Mo-nitrogenase polypeptides in both Mo-grown and −Mo−V grown A. cylindrica, providing evidence in support of the former possibility. Physiologically, the V-nitrogenase of A. variabilis showed several distinct similarities to the Mo-nitrogenase of A. cylindrica, including responses to metal concentrations as well as the above responses to addition of V and Mo, respectively.

Lipids were extracted from needles of Norway spruce plants of different provenances. The plants, seeds of which were originally collected from all over Europe, were grown at two sites in England at the same longitude but one 350 km north of the other. Significant correlations were observed between monogalactosyl diglyceride (MGDG) and digalactosyl diglyceride (DGDG) ratios on the one hand and, on the other, the lowest extreme winter temperature at the original site of seed collection (EWTS).

In the case of lipid changes, however, environmental differences due to site were more important than the inherent genotypic characteristics of the planted trees while provenances with the lowest EWTS values had more MGDG than DGDG. This implies that existing winter hardiness under mild climatic conditions does not come from changes in this ratio although DGDG formation from MGDG is known to be an important feature of winter hardening in Norway spruce elsewhere in Europe. However, closer analysis of the fatty acids in lipids from provenances at individual sites did reveal significant correlations between those provenances with low EWTS values and certain molar amounts or ratios of fatty acids in DGDG, some of which could be traced to genotypic expression. These partly involved the unusual Δ5 desaturations which are mainly confined to the fatty acids of those conifers like Norway spruce that harden to very low temperatures.

A model system was developed for the molecular analysis of plant–mycorrhizal fungus interactions using Hordeum vulgare cv. Galleon and Glomus intraradices Schenk and Smith. High levels of early-colonization stage mycorrhizas (20–35% of the total root length) were routinely obtained in the roots of 12–14-d-old barley seedlings. Detailed colonization studies showed that most colonization occurred in the first 5–10 lateral roots, and that colonization in primary roots and younger lateral roots was low. Differential screening of a cDNA library prepared from the highly colonized region of the root system and subsequent northern analyses identified four differentially expressed cDNA clones. Two clones, BMR6 and BMR78, detect RNA transcripts that accumulate to much higher levels in mycorrhizal roots than in non-mycorrhizal root tissues. DNA hybridization analysis confirmed that both of the clones were of plant rather than fungal origin. Hybridization and sequence analysis strongly suggests that BMR78 represents a partial cDNA of a low copy proton-ATPase gene. This finding is consistent with the extensive proliferation of the periarbuscular membrane and high ATPase activity associated with it. BM6 detects a low copy gene whose function is unknown.

Water contents in healthy xylem, reaction zones (column boundary layers) and decayed wood were determined in sycamore (Acer pseudoplatanus L.) naturally infected by sooty bark disease (Cryptostroma corticale (Ellis & Everh.) Gregory & Waller) or wound-inoculated with Ustulina deusta (Fr.) Petrak. or Ganoderma adspersum (Schulz) Donk. Water contents determined gravimetrically were greater in reaction zones than in healthy wood by factors of 1·1–2·1 times. Maps of imageable proton density (M0), essentially equivalent to relative water contents, were also calculated from two independent nuclear magnetic resonance (NMR) imaging (MRI) data sets, acquired from T1 or T2-weighted experiments. M0 values derived from these, corrected for T2 and T1 respectively, were in good agreement, and paralleled the results obtained when total water was determined gravimetrically, although the proportional increase in M0 in reaction zones (1·3–3·7 times) was greater than the corresponding increase in total water.

Living wood saturated with water by vacuum infiltration contained approx. 1·3 times more water, determined gravimetrically, than uninfiltrated wood. The proportional increase in M0 was greater (1·8–2·4 times). This was attributable to the alleviation of magnetic susceptibility effects by elimination of the gas–water interfaces normally present in the uninfiltrated wood. The similarities between infiltrated wood and reaction zone tissues in both total gravimetric water content and NMR-visible water suggest that the biophysical environment of the water might be similar in both. This is compatible with the hypothesis that in reaction zones the gases normally present in the lumens of empty (dead) xylem fibres are replaced by an aqueous solution of phytoalexin-like compounds. This has been confirmed by microscopy. The implications of this in relation to antimicrobial defence and xylem function are discussed.

Although contrast in NMR images of lesions was predominantly attributable to changes in M0, a decrease in the spin–spin relaxation time (T2) was seen in the vicinity of some lesions. Spin–lattice relaxation times (T1) were not altered to the same extent.

Results obtained from both early and late stages of decay were much more variable than those from healthy wood or from reaction zones, and reflected both the low water content of decayed wood (and hence poor signal-to-noise ratios for NMR data) and the changing physical environment as the wood was degraded.

Three successive emergent macrophytes (Typha latifolia L., Phragmites australis (Cav.) Steudel and Glyceria maxima (Hartman) Holmbly) were each grown in substrates collected from three different zones of shoreline vegetation development (non-vegetated sediment, the interface between T. latifolia and P. australis, and degenerating P. australis). The aim of the study was to assess whether accumulation of litter changes growth conditions of P. australis, and to determine its effects on pre- and post successional plant species. The study was carried out by means of pot experiments in a glasshouse. Seedlings of the three species were cultured in fertilized and unfertilized substrates under both waterlogged and drained conditions.

In its own litter, growth of P. australis was strongly reduced, compared with the productivity of plants in substrates from preceding successional stages, and could not be compensated for by fertilization or soil drainage. The redox potential of the substrate was not strongly reduced and the sediment density was well above the critical level. Soil sterilization by gamma-irradiation did not improve growth substantially, although there was some positive effect in unfertilized substrate. Phytotoxic compounds might have caused poor growth of P. australis in its own litter. T. latifolia and G. maxima were relatively less affected by the P. australis litter. The possible importance of litter accumulation on species replacement in shoreline vegetation is discussed. It is concluded that the accumulation of organic matter should be considered as a factor affecting spatio-temporal processes in littoral vegetation owing to its specific impact on the functioning of individual dominant plant species.

Living epidermal cells of cowpea (Vigna unguiculata (L.) Walp.) were examined by contrast-enhanced video microscopy during penetration by invasion hyphae of the monokaryotic stage of the cowpea rust fungus (Uromyces vignae Barclay race 1). In resistant or susceptible host cvs, the plant nucleus migrated to the penetration site before the fungus had fully penetrated the plant wall and, at sites of unsuccessful infection, remained during the formation of a callose-containing papilla. Nuclear migration was also induced by applying hemicellulase, but not H2O2, to localized sites of wall damage. Hemicellulase-induced migration was inhibited by calcium chelators and a protein kinase inhibitor, but not catalase. In both resistant and susceptible cvs, the plant nucleus migrated away from successful infection sites at about the time that the fungal penetration peg made contact with the plant plasma membrane, and the epidermal cell showed no further cytological responses to the growth of the fungal intra-epidermal vesicle. In the susceptible cv., the nucleus migrated back to the fungus when the latter initiated tip growth. Inhibitors of transcription or translation did not affect this migration and only slightly reduced fungal growth. In the resistant cv. in which the invaded cell exhibited a hypersensitive response (HR), the plant nucleus changed its appearance before the cessation of cytoplasmic streaming and usually did not migrate to the fungus, even if the latter initiated tip growth. Nuclear DNA cleavage usually followed the subsequent cessation of cytoplasmic streaming. Treatments that delayed cell death and increased fungal growth also increased the frequency of nuclear migration to the fungus. It is argued that these and other data suggest that U. vignae negates nonspecific, penetration-induced, defence responses upon entering cells of both susceptible and resistant cultivars. The results also suggest that effects on the plant nucleus are one of the earliest signs of the HR in this system, often preceding the cessation of cytoplasmic streaming and detectable changes in plasma membrane permeability.

The biomass of Mycelis muralis (L.) Dumort plants increased with increases in growth irradiance and nitrate supply. Comparisons of the data from this study with published information indicate that biomass production and growth rate of M. muralis are comparable to that of a number of slow-growing species from resource-poor environments. Biomass allocation varied with growth irradiance and nitrate supply. At high irradiances and low nitrate supplies there was a significantly greater allocation of plant mass to roots, resulting in a lower leaf area ratio (LAR) and leaf mass ratio (LMR). In contrast to the findings of a number of other studies, light level and nitrate supply had interactive effects on leaf mass per unit area (LMA). Changes in LAR due to variations in irradiance were mainly a consequence of changes in LMR. Given the ability of M. muralis to grow in deep shade, this suggests that large changes to within-leaf allocation (LMA) are not an essential requirement for growth at low irradiances. Conversely, because M. muralis can also grow in fully exposed situations, these traits might not exclude some shade species from high irradiance habitats. However, a low growth rate, or traits which result in a low growth rate, might be advantageous in a number of resource-poor environments.

We investigated the effect of soil compaction and phosphorus (P) application on morphological characteristics of mycorrhizal colonization and growth responses, to determine the reasons for reduced responses observed in our previous work with compacted soil. Growth, phosphorus (P) uptake and intensity of vesicular–arbuscular (VA) mycorrhizal colonization were studied in clover plants (Trifolium subterraneum L.) with and without VA mycorrhizal colonization at two P applications and three levels of soil compaction. Phosphorus was supplied either at constant mass concentration (mg P kg−1 soil) or at constant volume concentration (mg P dm−3 soil). Increasing bulk density of the soil from 1·1 to 1·6 Mg m−3 significantly decreased root length and shoot d. wt, but increased the diameter of both main axes and first order lateral roots regardless of P application. Total P uptake and shoot d. wt of clover plants colonized by Glomus intraradices (Schenck & Smith) were significantly greater than those of non-mycorrhizal plants at all levels of soil compaction and both P applications. However, soil compaction to a bulk density of 1·6 Mg m−3 (penetrometer resistance = 3·5 MPa at a matric potential of −33 kPa) significantly decreased mycorrhizal growth response. There was no evidence that the increased volume concentration of P at high bulk densities was responsible for the reduced responses. Soil compaction had no significant effect on the fraction of root length containing arbuscules and vesicles, but total root length colonized by arbuscules, vesicles or by any combination of arbuscules, vesicles and intra-radical hyphae significantly decreased as soil compaction was increased. The air-filled porosity of highly compacted soil, which varied from 0·07 to 0·11 over the range of matric potentials encountered (−33 and −100 kPa), had no significant effect on the intensity of internal colonization.

The presence of copper-binding proteins produced in response to added copper was examined in isolates of Laccaria laccata (Scop. ex Fr.) Cooke and Paxillus involutus (Batsch ex Fr.) Fr. taken from copper-contaminated and uncontaminated sites, and in a single isolate of Scleroderma citrinum Pers. from a contaminated site. Two isolates of Laccaria (GLac4 and ELac1) grew better in 1·5 mM and 2·5 mM copper than a third (Lac3G) and were considered to be more tolerant. Amongst five isolates of P. involutus, three (WJPax2R, GPaxRSp2 and Pax4) were capable of growth in media containing 4·0 mM copper and were regarded as tolerant. All isolates of both Laccaria and Paxillus were capable of some growth in 2·5 mM copper, but S. citrinum was much more copper-sensitive and the concentration had to be reduced at least 10-fold before any growth occurred. Tolerance of isolates was not related to whether they were taken from copper-contaminated or uncontaminated sites. Copper-binding proteins were detected in response to copper in the culture media in the two tolerant isolates of Laccaria (GLac4 and ELac1) but not in the least tolerant isolate. In Paxillus, similar proteins were found in two tolerant isolates (GPaxRSp2 and Pax4) but not in WJPax2R, which was also regarded as tolerant, nor in any of the less tolerant isolates. Copper-binding proteins were not detected in S. citrinum. The copper-binding protein purified from the Laccaria isolate ELac1 appeared as a single band in modified SDS–PAGE electrophoresis. Its molecular mass and spectral characteristics were consistent with it being a metallothionein.

The lower part (460–650 cm) of a lacustrine sequence from Lac de Creno, Corsica (1310 m) is analysed on the basis of 68 pollen spectra and with the support of 10 14C dates (including nine A.M.S. dates). This sequence, which extends from the end of the Würm to the beginning of the Postglacial, reveals a complete late-glacial. The absence of forest dynamics during the late glacial Interstadial in Corsica is a real mystery. Pollen data clearly suggest that Pinus laricio ( = Pinus nigra Arnold ssp. laricio Maire) and perhaps other tree species did not exist in Corsica at that time.

Mycorrhizal mycelium was allowed to grow into a compartment from which roots were excluded. Ammonium or nitrate was supplied exclusively to the fungus and the respiration rate before and after nitrogen addition was measured. Mycelial respiration in the fungal compartment represented 11–29% of total fungal and root respiration. The mycelium in the fungal compartment received 20–29% of shoot net assimilation, and 43–64% of the carbon allocated to the mycelium was respired. The nitrogen-induced respiration increase ranged, in the fungal compartment, from 54 to 180%.

Microtubule arrays were studied in Spiranthes sinensis (Pers.) Ames protocorms cultured in vitro either asymbiotically or symbiotically with the fungus Ceratobasidium cornigerum (Bourdot) Rogers by localizing β-tubulin in hand sections observed with laser scanning confocal microscopy. Cortical microtubules present in uncolonized cells disappeared when cells became colonized by the fungus. However, microtubules were observed between the hyphae forming hyphal coils (pelotons) in colonized cells. In these cells a close relationship between pelotons, microtubules and nuclei was often observed, and microtubules associated with hyphae formed a net-like structure through the pelotons. During senescence and condensation of the pelotons, microtubules were observed surrounding the pelotons, between collapsing hyphae and as circular profiles within the collapsed hyphal masses. In cells which contained completely collapsed hyphal masses and which were not recolonized, cortical microtubules reappeared and continued from the cell periphery through the hyphal remains to the collapsed hyphal masses. Microtubules associated with hyphae seemed to be involved in the condensation of hyphal masses in the host cells. Microtubules were also observed within hyphae, except in those which were collapsing or had collapsed.

The early development of indigenous and introduced Scots pine (Pinus sylvestris L.) ectomycorrhiza in natural forest humus was examined using molecular fingerprinting techniques. Non-mycorrhizal or mycorrhizal (Suillus bovinus (L. ex Fr.) O. Kuntze or Paxillus involutus (Batsch ex Fr.) Fr.) seedlings were placed in transparent two-dimensional microcosms on a thin layer of sieved humus to allow observation of ectomycorrhizal development and fungal growth. Twelve ectomycorrhizal morphotypes, mainly based on colour, gross morphology and outer mantle structure, were identified over a 3-month period. No successional trends in mycorrhiza formation were observed. For further characterization, individual ectomycorrhiza representing the different morphotypes were subjected to analyses of esterase (EST) isozyme profiles and RFLPs of the internal transcribed spacer (ITS) from fungal rRNA genes following amplification using PCR. All 10 morphotypes analysed displayed different esterase isozyme profiles, and the characteristic S. bovinus species diagnostic band (S. b. EST8) was detected in two white morphotypes. Ten Pink and Black ectomycorrhiza were all separated into one and three groups, respectively. A fast-running plant-specific polymorphic locus (Z) was also confirmed in most ectomycorrhizal morphotypes. Diagnostic species-specific (S. bovinus and Pink) EST bands were detected in intact external mycelium colonizing soil. Successful amplification of the ITS from individual ectomycorrhiza, of eight different morphotypes, was found to be mainly influenced by the DNA template concentration. Different RFLPs (Hinf I, Mbo I and Hha I) of the ITS placed the white morphotypes into two groups, one corresponding to S. bovinus, in agreement with the esterase fingerprinting. By contrast, Pink and Beige morphotypes displayed different EST profiles but nearly identical RFLP fingerprints. The ITS amplification success rate of standardized DNA template concentrations from 10 individual Pink and Black ectomycorrhiza was 90 and 50%, respectively. In all successful amplifications, the black morphotype yielded the smallest ITS fragment, similar to earlier reports for Cenococcum geophilum, that gave identical RFLPs. The ecological significance of the observed mycorrhizal diversity and combined application of these two identification methods is discussed.

Solution culture experiments were made using two subspecies of Cerastium fontanum and Cochlearia pyrenaica ssp. alpina (Bab.) Dalby and Festuca rubra L. of ultramafic and non-ultramafic origin to test: (1) whether Mg and Ni toxicities are likely to exclude non-ultramafic races of the three species; (2) whether tolerance to Mg and Ni toxicities by the ultramafic plants differed markedly among the species; and (3) to use (2) to generate a hypothesis about controlling factors of species distribution on the Meikle Kilrannoch (MK1) ultramafics. The solution cultures were based on soil solution extracts with experimentally varied concentrations of Mg, Ni and Na. The solution with high (7·5 mM) Mg and high (0·7 mg l−1) Ni concentrations simulated the ultramafic soil solution. Care was taken to use a Fe source which did not react with Ni and microelements. Of the non-ultramafic plants, the growth of Cerastium was reduced by high concentrations of Mg but not by Ni, whereas both the Cochlearia and the Festuca suffered severe reduction in biomass production in the presence of high Mg or Ni or both. Of the ultramafic plants, Festuca was tolerant of high Mg and Ni concentrations; Cerastium was less tolerant of high Mg; and Cochlearia had low shoot production in high Ni (although the effect was ameliorated by high Mg). The different interspecific responses of the ultramafic plants to Mg and Ni supported field observations that the three species might occupy microhabitats which differ in Mg and Ni.

Changes in polypeptide contents following inoculation of roots of Pisum sativum L. (wild type cv. Frisson (myc+, nod+)) with the arbuscular mycorrhizal fungus Glomus mosseae were analysed by two-dimensional gel electrophoresis (2D-PAGE). A different polypeptide pattern was obtained in a mycorrhiza-resistant pea genotype P2 (myc-, nod-) inoculated with G. mosseae. In order further to characterize the polypeptide modifications detected and to provide evidence for some possible symbiosis-related (SR) proteins, a time course experiment, from appressoria formation to fully developed symbiosis, was carried out on two genotypes allowing fungal colonization: the wild type and its isogenic mutant P56 (myc+, nod-). The same experiment was done with the mycorrhiza-resistant pea genotype (myc-, nod-). After G. mosseae inoculation, we characterized 12 additional polypeptides in the two mycorrhiza-compatible pea genotypes which were never observed in root extracts from the mycorrhiza-resistant mutant. Five polypeptides were first detected in the early stage of the symbiosis (5 d of inoculation) while others were observed later (8 d of inoculation). The induction and accumulation of these polypeptides seem to be more correlated to the establishment of the functional symbiosis than to the recognition stages and appressorium formation. Furthermore, none of the additional polypeptides were detected in the mycorrhiza-resistant pea genotype. This mutant was more characterized by a great repression of polypeptides. In addition, up-regulated and down-regulated polypeptides from the mycorrhiza-compatible genotypes were different from those of the mycorrhiza-resistant genotype.

Previous studies demonstrated the role of the O-antigenic side chain of the outer membrane lipopolysaccharides of Pseudomonas fluorescens strain WCS417r in its ability to induce resistance against fusarium wilt disease. The present study compared the colonization of tomato roots (Lycopersicon esculentum Mill.) by P. fluorescens strain WCS417r with that of a mutant of this strain, lacking the O-antigenic side chain of its outer membrane lipopolysaccharides (OA−). Special attention was paid to colonization within root tissues and changes of the root ultrastructure induced by the bacterial strains. Experiments were performed with tomato plants rooted in rock wool and tomato seedlings grown on agar. Numeration of colony-forming units showed that wild-type strain WCS417r colonized the root interior of tomato grown in rock wool to a higher extent than its OA− mutant B4, whereas colonization of the rhizoplane did not differ between the two. Light microscopy observations of cross sections of roots grown on agar showed that (i) colonization of the root interior by both wild-type and OA− mutant was inter- and intracellular, (ii) colonization of the internal root tissues mainly occurred if the Pseudomonas strains were inoculated at sites of lateral root emergence, (iii) the wild type colonized the root interior more frequently than did its OA− mutant. Thickening of the cortical cell walls, in reaction to the bacterial colonization of epidermal or hypodermal cells, or cortical intercellular spaces, suggested that local defence reactions of the plant occurred. Altogether, these results indicate the involvement of the O-antigenic side chain of the outer membrane lipopolysaccharides in the endophytic colonization of the tomato root by P. fluorescens WCS417r, and lead us to hypothesize that induction of disease resistance by strain WCS417r might be related to the extent of colonization of the internal root tissues.

Little is known about non-mycorrhizal endophytic fungi in tropical orchids; still less is known about how endophytes vary within and between individual orchid plants. Fungal endophytes were isolated from roots and leaves of epiphytic and lithophytic orchids in the genus Lepanthes; seven species, from rainforests in Puerto Rico, were sampled. The endophytes observed most frequently were Xylaria species and Rhizoctonia-like fungi, found in 29% of roots and 19% of leaves, and 45% of roots and 31% of leaves, respectively. Five deuteromycete genera were also isolated, occurring in 19% of roots and 43% of leaves (combined). At least nine species of Xylaria were found, with several species sometimes occurring in a single plant. Differences between roots and leaves in frequency of Xylaria and Rhizoctonia isolates were not significant, although differences among orchid species in number and types of endophytes were. Heterogeneity of endophytes in single plants and plant organs was greater than differences between species. Many Lepanthes species are very restricted in distribution, and knowledge of their interactions with endophytes might be useful in species management.