New seafloor is constantly being created at the global spanning system of mid-ocean ridges (MOR). Over a wide range of spreading rates is the produced oceanic crust of surprisingly uniform thickness and composition. An exception are ridges with spreading rates slower than 15-20 mm/yr so-called ultraslow spreading MORs, at which the crustal thickness and composition drastically change. A totally different mode of seafloor spreading occurs, with discontinuous crustal accretion along-axis. Entire sections of the ridge axis lack an igneous crust and the neighbouring volcanoes receive more melt than the regional average. The Gakkel Ridge in the ice-covered Arctic Ocean and the Southwest Indian Ridge (SWIR) in the stormy Southern Ocean are the main representatives of the ultraslow spreading ridges. At present, the processes of lithosphere formation at ultraslow spreading rates are still poorly understood. The investigation of microearthquakes, with networks of ocean bottom seismometers (OBS) or hydrophones, has significantly progressed our understanding of the lithosphere accretion processes at faster spreading ridges. Until recently, the rough environmental conditions prevented the long-term deployment of OBSs at the main representatives of ultraslow ridges. This thesis makes use of the first one-year long records from two OBS networks with comparable extent that had been placed at different sections of the SWIR from 2012 to 2013. The chosen sites are characterized by contrasting crustal thickness, lithology and morphology. The Oblique Supersegment network was deployed near 13°30’E where peridotites are the dominant seafloor lithology and a typical igneous crust is absent. The Segment 8 network was deployed near 65°30’E around the volcanic center of the SWIR Segment 8 where the crust is locally thickened and the seafloor consists exclusively of basalts. I picked and located the microearthquakes for the Oblique Supersegment network and compiled an 11-month long catalogue that contains ~2000 microearthquakes. This catalogue was subsequently used to calculate so-called yield-strength envelopes of the lithosphere that provide constraints on the rheological strength and the geothermal gradient below the axial valley. I further used an existing microearthquake catalogue of the Segment 8 network for a local earthquake tomography to image the 3-dimensional structure of seismic velocities below the SWIR Segment 8. The combined seismicity catalogues from both networks showed systematic undulations in the maximum depth of faulting (brittle-ductile transition) along the SWIR axis; with deeper hypocenters below amagmatic ridge segments and shallower hypocenters below magmatic segments. The brittle-ductile transition is mainly temperature related. Thus, its position provides insight into the sub-axial thermal structure. The undulating hypocenter depths imply parallel undulations of the deeper lithosphere-asthenosphere boundary, under which molten material is constantly present. Previous studies postulated such a topography of the lithosphere-asthenosphere boundary that guides the buoyantly flowing mantle melts from amagmatic segments towards magmatic centers and thereby maintains the pattern of uneven melt supply along the SWIR. The combined microseismicity results from both OBS networks strongly support this hypothesis and provide the missing geophysical proof of this concept. An extensive aseismic region extending to 15 km depth was found in the upper lithosphere at the Oblique Supersegment, where peridotite is the dominant lithology. The aseismic behaviour is best explained by weakening of the lithosphere by serpentinization, likely focused in aseismic shear zones that constitute the rift valley bounding faults. Geochemical sampling revealed enhanced diffusive fluxes near the scarp of a bounding fault and increased methane concentrations in the valley waters that likely stem from abiotic, serpentinization-related processes. The local earthquake tomography of the SWIR Segment 8 network showed a prominent low-velocity anomaly below the segment’s volcanic center that indicates the presence of partial melts. In addition, preceding teleseismic activity and recorded microearthquake swarms with simultaneous intrusion tremor pointed to an ongoing spreading event. It turned out that this magmatic spreading episode likely lasted already over a decade and thereby vastly exceeds the duration of all previously documented magmatic spreading episodes at active MORs. In summary, this thesis provides for the first time a detailed insight into the microearthquake activity at two SWIR segments. The analysis and interpretation of the data presented in this thesis significantly contribute to a better understanding of the lithospheric structure and the seafloor accretion processes at ultraslow spreading MORs.