The present research deals with the flora, phytosociology and ecology of riparian forests. The overall objective of this research is to contribute to a better knowledge of the flora, diversity and ecology of riparian forests in Benin . The specific objectives are to (a) compile a preliminary riparian forests plant species list, (b) assess plant species and ecosystem diversities, (c) investigate plant communities, (d) clarify the structural and floristic relationship of riparian forests with adjacent plant communities, and (e) assess the ecology of certain endangered tree species in riparian forests. In chapter 1 (General introduction), the research background, objectives and approach for riparian forests biodiversity assessment, and the organisation of the thesis are presented. Chapter 2 introduces the study area which covered about 70 % of Benin , from 7 ° 10' to 12 ° 20' N. Chapter 3 presents an overview of riparian forests biodiversity, their importance and the threats they face making them endangered ecosystems. A definition of riparian forests (or gallery forests) is given in the Benin context. The floristic characteristics of riparian forests in each phytogeographic district are presented. Issues related to legal protection and rehabilitation of the function and resources of riparian forests are documented: specifications and weaknesses of the forest law regarding riparian forests are presented; challenges for various stakeholders are discussed, and some improvements of the current forest law are proposed. Chapter 4 assesses plant species diversity, as well as species abundance models that best fit representative collections of plant species of riparian forests throughout the country. This study shows the richness and diversity of riparian forests in Benin , in comparison to other vegetation types in this country. They harbour about 1/3 of the estimated total number of plant species of the whole country in sample plots totalling 19 ha. This flora shares many features with riparian forests and dense forests worldwide: e.g. most abundant families, species richness/ha, trees species richness/ha, Shannon index, equitability index of Pielou, and species abundance models. Endemism is very low compared to that in rain forests, what is not surprising in the Dahomey Gap. The main conclusion is that relatively large numbers of species are still maintained in small forest fragments along waterways. These remnants with their specific plant species composition can be used for the restoration of degraded forest stands. Chapter 5 assesses the structure and ecological spectra of 19 ha of riparian forests through selected parameters (e.g. life form, geographic affinity, diameter class distribution, basal area, stem density, species dominance) that give a general picture of different vegetation types present. Figures obtained for these parameters show that riparian forests in Benin are on the one hand similar to many riparian forests in West Africa as well as in South and Central America, and on the other hand to many tropical upland forests. A brief description of the process of riparian forests degradation is also presented. Chapter 6 deals with the phytosociological assessment of representative relevés of riparian forests of Benin . Floristic ordination (DCA analysis) and classification (TWINSPAN) were derived from a comprehensive floristic inventory of a data set of 818 plant species and 180 relevés. This yielded 12 plant communities or associations, most of which had not yet been formally described: 1 - Community of Isolona thonneri and Callichilia barteri (10 relevés) along streams. This community occurs at the lowest parts of the gallery forest with frequent inundation in the centre of Pénéssoulou protected forest. 2 - Community of Motandra guineensis and Pararistolochia goldieana (24 relevés) along streams at the East and West parts of Pénéssoulou reserve forest. This community is mainly present on drained sites ( i.e . seldom inundated). 3 - Community of Chrysobalanus icaco subsp . atacoriensis and Pentadesma butyracea (22 relevés) along streams at hill feet in the Atacora mountain chain. 4 - Community of Alchornea cordifolia and Ficus trichopoda (9 relevés) along streams on regularly inundated plateaus all over the country. 5 - Community of Berlinia grandiflora and Khaya senegalensis (8 relevés) along streams on drained plateaus ( i.e . seldom inundated), mainly in the Sudanian region of the country. 6 - Community of Raphia sudanica and Oxytenanthera abyssinica (8 relevés) along streams on drained plateaus, mainly in the Sudanian region. 7 - Community of Cynometra megalophylla and Parinari congensis (31 relevés) along the Ouémé river in the Guinean region of Southern Benin . 8 - Community of Capparis thonningii and Crateva adansonii (30 relevés) along the Ouémé river in the Sudano-Guinean region of Central Benin . 9 - Community of Lepisanthes senegalensis and Drypetes floribunda (17 relevés) along the Ouémé river in the Sudano-Guinean region of Central Benin . 10 - Community of Uapaca heudelotii and Irvingia smithii (8 relevés) along the Sota river in the Sudanian region of North East Benin. 11 - Community of Garcinia livingstonei and Combretum acutum (12 relevés) along the Pendjari river in the Sudanian region of North West Benin. 12 - Community of Mimosa pigra and Ficus asperifolia (20 relevés) widely distributed on sandy banks along rivers. Ordination proved invaluable in the exploration of environmental characteristics of the phytosociological groups. The environmental factors (waterways, relief, topography, latitude and longitude) helped in the grouping of floristic relevés in the above mentioned 12 plant communities. The distinguished plant communities were compared with syntaxonomic data in literature. Riparian forests in Benin belong to the Mitragynetea Schmitz 1963, which is the phytosociological class of hygrophile fresh water forests of tropical Africa . Based on similarities of ecological conditions and floristic composition, the 12 plant communities can be classified into 3 orders that are Alchornetalia cordifoliae Lebrun 1947, Lanneo-Pseudospondietalia Lebrun & Gilbert 1954 and Pterygotetalia Lebrun & Gilbert 1954. Chapter 7 presents the spatial distribution and ecological factors determining the occurrence of Pentadesma butyracea (Clusiaceae), a rain forest and multipurpose species found in Benin only along certain streams. Among the 224 tree species found along waterways, Pentadesma is one of the least known, yet of great ecological and economic importance. Field survey reveals the presence of this rain forest species in four non-contiguous remnant riparian areas, some located far from its optimal ecological range. If urgent actions are not taken to protect the remaining fragmented and dispersed riparian habitats, current human-induced disturbance could result in the disappearance of this species in Benin . Chapter 8 deals with the variation of the floristic composition, structural parameters (e.g. abundance, average height, basal area, tree richness) and spatial distribution of tree species at river edges across riparian forests. Horizontal and vertical structures of tree species exhibit complex patterns at riverside. On the one hand, tree stems are characterised by an uneven distribution across riparian forests, on the other hand height and basal area variations at riverside do not show any easily interpretable patterns. The numerical analysis confirms a gradual variation in the floristic composition across riparian forests and neighbouring plant communities. These results suggest a partitioning of riparian forests in three habitats ( i.e. river front, middle and riparian forest edge). An implication for diversity assessment is that plot size, shape and layout in the terrain should take into account the river front, the middle and the edge of riparian forest. Due to the non-coverage of the whole riparian forest width and unequal chance of species and stems to be sampled, circular and square plots are not suitable for structural parameters and phytodiversity assessment in riparian forests. Instead rectangular plots with varying length and width, and covering the whole cross section of riparian forest are the most suitable sampling units under the study area conditions, and probably for savanna regions too. The present study also provides scientific guidelines for an improvement of the forest law regarding the distance to be protected at riverside, and suggests 100 m instead of 25 m. In chapter 9 the floristic composition, species richness and structure of two riparian ecosystems in West Africa (the Comoé in Ivory Coast and the Ouémé in Benin ), are compared. The overall physiognomy of the two gallery forest sites seems similar and they share the most prominent families. However, there are marked differences in terms of canopy density and height, herb layer density, number of individuals, tree richness and diversity ( H' ), and species composition. The phenomenon of single species dominance at both sites is documented from Cynometra megalophylla , an evergreen tree species, which was time and again the most frequent and dominant tree at both riversides and in the middle of the gallery forests. Only detailed comparison shows the difference and complexity of ecological processes between and within gallery forests sites. The research carried out in chapter 10 facilitates the choice between several sampling designs for the estimation of a population parameter for endangered species. This study was carried out in the Pénéssoulou forest, in Central Benin . Stratified random sampling provided the lowest variance, coefficient of variation, standard error and sampling error. This method was taken as the most precise and reliable design over simple random and systematic samplings for the density estimation of Khayasenegalensis and K. grandifoliola trees. Results have confirmed empirical knowledge about the ecology of Khaya species and shown that the selection of the most precise sampling design, with regards to estimating a given parameter, can eventually be useful for the sustainable management of forest trees in the study area. A reliable density estimate for Khaya species within the given vegetation types facilitates the selection of areas to be protected and sustainably exploited. Chapter 11 presents a general discussion on issues discussed in this thesis. Sustainable rehabilitation and restoration of riparian forests biodiversity in Benin are discussed in the general conclusion ( chapter 12 ). This study has provided detailed site-specific data on plant species that can serve for further scientific research, as well as for conservation management and planning. It fills a gap of knowledge on the flora of Benin , and can contribute to better land-use planning and conservation of riparian forests.