In this thesis, I synthesize my research on the seismicity of ultraslow spreading ridges describing the overarching theme of 10 peer-reviewed publications. At mid-ocean ridges, the lithospheric plates drift apart, magma fills the gap to form new crust. This engine splutters at very low speeds: Isolated volcanoes, capable of vigorous eruptions, pierce the seafloor of ultraslow spreading ridges; between the volcanoes, there are long stretches without volcanism. The main representatives of ultraslow spreading ridges, the Arctic ridge system and the Southwest Indian ridge, are poorly explored as they lie under the perennial sea ice cover of the Arctic Ocean and in the stormy Southern Ocean, respectively. The morphology and the mode of seafloor production at ultraslow spreading ridges differ fundamentally from all faster spreading ridges. Ultraslow spreading ridges are characterised by a cold lithosphere, a greatly reduced melt production and an uneven distribution of melt. My junior research group MOVE - Mid-Ocean Volcanoes and Earthquakes - studies the seismicity of ultraslow spreading ridges to better understand the active spreading processes that govern the generation of crust at these ridges and that lead to the focussing of melts into centres of pronounced volcanism. One of the biggest challenges of this project was to register local earthquakes in these remote survey areas. I managed to routinely record microearthquakes with seismic arrays on drifting ice floes in order to compare the local earthquake activity of magmatic and amagmatic ridge segments. I document here also our extensive experience with seismicity surveys on ice. MOVE yielded partly astonishing results: It previously appeared unlikely that a mid-ocean ridge volcano should erupt explosively at a confining pressure of 4 km of water column. Our seismometers on ice floes picked up explosive sound signals near the 85°E volcano at Gakkel ridge that we could attribute to mild Strombolian eruptions at the flank of the rift valley. We could show that this volcanic activity is part of a longer eruptive cycle that started with the largest ever-recorded teleseismic earthquake swarm at a mid-ocean ridge. Deep reaching faults play an important role as transport path for melts through the cold lithosphere of ultraslow spreading ridges and might cause high magnitude earthquakes during spreading episodes at these ridges. Even in recently active areas, local and teleseismic earthquakes occur down to a depth of 20 km below seafloor, confirming the existence of a cold and brittle lithosphere at ultraslow spreading ridges. Underneath Logachev Seamount at Knipovich ridge, microearthquake hypocentre depths yield for the first time direct evidence for an undulating lithosphere-asthenosphere boundary which has often been postulated as one possible way to channel melts towards the centres of focussed magmatism. Dike intrusions at faster spreading ridges are accompanied by small earthquakes that are below the detection threshold of global land stations. Therefore, we expected to find more earthquake swarms if the detection threshold could be lowered. However, our analysis of seismic records of the Neumayer array (Antarctica), that is sensitive to earthquakes occurring at a magmatic segment of the Southwest Indian ridge, showed that swarms consisted of significantly more smaller earthquakes, but all swarms within a 8 year period included earthquakes with M>5 and could be identified teleseismically. For ultraslow spreading ridges, teleseismic earthquake catalogues are therefore a valuable tool to study magmatic activity. We screened the teleseismic earthquake records of all ultraslow spreading ridges systematically and could make out fundamental differences to faster spreading ridges. Volcanic centres at ultraslow spreading ridges often have an increased seismicity rate relative to the background, whereas the central volcanic rises of faster spreading ridges appear as seismic gaps in along-axis profiles of teleseismic and hydroacoustic data. Increased seismicity with asymmetric accumulations of earthquakes are typically related to detachment faulting at the segment ends of slow spreading ridges like the Mid-Atlantic Ridge. In contrast, amagmatic segments of ultraslow spreading ridges are only weakly active and lack any signs of detachment faulting. Having deployed the ocean bottom seismometers for the first comprehensive local seismicity study of contrasting segments of the Southwest Indian Ridge, I continue to work on a refined model of active spreading processes at ultraslow spreading ridges.