This thesis investigates restored and naturally developed multiple-channel sections in German mountainous areas. Without human interference most streams in these areas would show this hydromorphological pattern. Unfortunately only short multiple-channels sections can be found or have been built in restoration projects. The quantity of hydromorphological differences compared to anthropogenically straightened single-channel sections and effects on macroinvertebrate communities are largely unknown. In a paired site study seven multiple- and seven single-channel sections were investigated and compared. The hypothesis was tested that multiple-channel sections have a larger area and higher habitat diversity, which also influences the biota. A standardised transect-point-protocol was applied in the summer of 2004 and 2005. Width of channel features (main and secondary channels, sidearms, standing water bodies, bars, banks, floodplains and embankments) and their habitat characteristics (aquatic, terrestrial or transient) were recorded along 20 equidistant transects across the bankfull width. Along these transects, aquatic habitats (substrate type, depth, current velocity) were recorded at 400 points per stream section. Section-wise analyses of both years built the elaborate basis against which macroinvertebrate data were tested. Altogether 199 substrate-specific macroinvertebrate samples were taken and processed separately. Finally, the relationships between hydromorphological characteristics and each section’s community were evaluated. Twelve hydromorphological metrics were calculated from the transect and point data. Metrics at three scales differentiated well between single- and multiple-channel sections: the macro-scale metrics mean channel width and shore length increased by factors of 2.1 and 2.4, respectively; the meso-scale metric mean number of channel features increased from two to ten per section; several micro-scale metrics, such as current velocity and depth variance, were significantly different between single-channel sections and most multiple-channel sections. Furthermore, substrate composition was more diverse in the multiple-channel sections. Correlation analyses of hydromorphological parameters suggest that metrics of different scales should be combined in order to evaluate hydromorphological diversity. The sites were revisited in 2005 and transect-point recordings were repeated to quantify annual changes within the sections. The amount of changes between two years differed between single- and multiple-channel sections. Water depths, aquatic widths and cross-section, riparian areas, and the waterside margins of the embankments have been modified by the streams in the course of the year. Multiple-channel sections showed more annual changes than the single-channel sections. Within the multiple-channel sections, natural and restored sites showed different extents of change with the latter showing higher and more varied changes in the multiple-channel sections. 199 substrate-specific macroinvertebrate samples were analysed for alpha- and beta-diversity and nestedness patterns. Taxa number, abundance, and evenness of communities found on multiple-channel substrates did not significantly differ from their single-channel section counterparts. Nestedness was neither different from a section's perspective, i.e. single- and multiple-channel sections were equally nested, nor from a substrate's perspective, i.e. substrates from either section were equally nested. The findings imply that substrates at single- and multiple-channel sections have similar macroinvertebrate communities, thus alpha-diversity is not changed. Different substrates host different communities, so beta-diversity might be influenced. Hence, stream restoration projects aiming at a re-development of macroinvertebrate diversity should focus on the generation and availability of high quality habitats, such as large wood. These results are further supported by occurrence of ten Coleoptera and seven Trichoptera taxa present exclusively on multiple-channel substrates, with the highest differences found for the fine mineral substrates such as loam or sand and for organic substrates such as living parts of terrestrial plants (LPTP), coarse particulate organic matter (CPOM) and large wood. Macroinvertebrate data from 140 substrate specific samples were then combined to representative communities for each section. Community data were subject to similarity and cluster analyses. 35 metrics were calculated with the taxalists, including taxa number, abundance, feeding type, habitat and current preferences. Biological metrics were correlated to hydromorphological parameters. Micro-scale morphological metrics, such as certain substrates areas and substrate spatial diversity, showed consistent correlations with biological metrics, as did several meso- and macro-scale metrics like aquatic width and shore length. The biological metrics percentage of shredders, number of taxa, percentage of littoral preference and the percentage of current preference (rheo- to limnophil and rheobiont) showed the most consistent correlations with hydromorphological parameters. Bray-Curtis similarity was very high (69−77%) between communities of single- and multiple-channel sections. The hydromorphological diversity of the multiple-channel sections is improved towards the reference condition. Habitat diversity is increased and a higher sediment dynamic becomes manifest in cross-section changes. However, effects on the macroinvertebrate community of substrates or the overall stream sections are not detectable and at best display a general tendency towards improvement. This lack of diversification of the macroinvertebrate community is attributed to several causes and a mixture of abiotic and biotic constraints likely apply. Reasons for high similarity of macroinvertebrate communities from single- or multiple-channel sections are discussed, including influence of large scale catchment pressures, length of restored sections and lack of potential re-colonisers.