New oceanic crust is formed at mid-ocean ridges by magmatic intrusion, volcanic eruptions, tectonic extension and hydrothermal cooling. All these active processes give rise to earthquakes, which are usually too small to be recorded on land, but can be recorded locally with ocean bottom seismometers (OBS). OBS studies on slow and fast spreading ridges have greatly contributed to our understanding of active seafloor spreading. However, at ultraslow spreading rates, predictions of amagmatic spreading proved incorrect and the processes of crustal generation are still poorly understood, partly because of the remote location of the main ultraslow ridges like the Arctic ridge system. No comprehensive microearthquake studies of ultraslow-spreading ridges have yet been undertaken.During the Arctic Mid-Ocean Ridge Expedition in 2001 (AMORE2001), we made a first attempt at observing microearthquakes at Gakkel Ridge using ice floes as platforms for small seismological arrays consisting of four conventional short-period seismic stations. We successfully recorded microearthquakes including a swarm of 200 explosive sounds at the volcanic complex near 85°E which was active in 1999. We interpreted these sounds as signs of ongoing eruptive activity. The presence of a hydrothermal event plume in the water column supported this interpretation.Following this pilot study, we now started a comprehensive and comparative study of the seismicity of ultraslow-spreading ridges. As part of this project, we conducted a microseismicity survey during the Arctic Gakkel Vents Expedition 2007 (AGAVE2007). We deployed a network of three seismological arrays around the rift valley at the 85°E volcanic complex greatly improving the detection threshold and location capabilities compared to AMORE2001. The arrays had triangular shape with a central station. Each array measured about 1 km in size and was located on one ice floe. The inter-array distance was 15 km. The arrays recorded continuously at a sampling rate of 100 Hz for 16 days.We present the results of a first screening of the microearthquake data set of AGAVE2007 showing typical seismic events, their frequency and waveform characteristics and we compare the seismicity patterns to those of the pilot study AMORE2001, especially to the explosive volcanic activity of 2001 which appears to have ceased by now.