Gamma (30-80 Hz) oscillations in the brain are associated with learning and memory (Herrmann et al. 2004). In vitro, gamma oscillations are often recorded under interface conditions and oscillations lasts for hours. To study the cellular mechanisms of gamma oscillations, we established an in vitro model of kainate-induced gamma oscillations under submerged conditions. Here we characterize this model and compare it with gamma oscillations induced under interface conditions. Hippocampal slices 300 µM thick were cut from the ventral mouse hippocampus and kept in submerged conditions. Kainate (250 nM) induced gamma oscillations in the CA3 area that quickly peaked but than reduced to a steady-state level. Submerged kainate-induced gamma power and frequency was critically dependent on temperature with an optimum around 29°C, whereas 32-34°C is optimal for the interface recording. The peak power and frequency of the oscillation induced by kainate (100 nM) was 19.3±8.5 µV² (n=5, mean±SD) and 12.6±1.7 Hz at 24°C; 32.6±15 µV² and 24.5±2.5 Hz at 29°C; and 7.5±3.1 µV² and 8.5±1.3 Hz at 36°C. To achieve submerged gamma, a relatively high (5-7 ml/min) perfusion rate was required, as submerged gamma is very small (2±0.8 µV², n=4) with a perfusion rate of 2-3 ml/min commonly used in the interface condition. Optimal oxygenation was critical for kainate-induced gamma oscillations in the submerged condition, which was achieved by optimising oxygen pressure of the aCSF and by adding fine glass fibres on the bottom of the recording chamber that allow perfusion from both sides. Mapping of gamma oscillation in hippocampus indicated that, like under interface conditions, gamma oscillations were strongest in area CA3c. Peak gamma power increased with kainate concentration (25–1000 nM), but the steady state gamma power was maximal at 100 nM. The development in gamma oscillations (to reach peak) was faster in submerged recording (2 to 5 min) than in interface conditions (>30 min). The steady-state of gamma oscillations was short (20-50 min versus hours in interface conditions); the amplitude of gamma power was smaller (30.4±22.7 µV² versus 550.2±316 µV² in interface, n=5); and the washout was quicker (5-15 min versus > 30 min in interface recording). Repeated application of the same concentration of kainate (100 nM) caused 331±92% increase (n=6) in the gamma power. These observations indicate that gamma oscillations can be reliably induced under submerged conditions. The fast development and quick washout of gamma oscillations allows for using repeated application of kainate. However, submerged gamma is vulnerable to conditions that affect the metabolic state of the cells in the network, which are better under interface conditions.