a b s t r a c tSome features of galloping detonations near the detonation limits are discussed in this paper. The experimental results previously reported in , together with additional data obtained in this study for six explosive mixtures with different reaction sensitivities, each in five different diameter tubes, are analyzed in detail. It is established that galloping detonations do not occur in highly argondiluted, stable mixtures. Only in unstable mixtures, susceptible to flow fluctuations by thermo-chemical instability, when using in small tube diameters (i.e., D 6 12.7 mm), could galloping detonations be observed. For the largest tube diameter D = 50.8 mm, the detonation wave fails completely when the initial pressure approaches the detonation limit even for all unstable mixtures. In addition, the present study shows that the initial pressure range for the occurrence of galloping detonations decreases rapidly with increasing tube diameter. A collection of existing data on galloping detonations indicates that the wavelength L of one galloping cycle is on average about 350D within experimental variations.Nonetheless, there appears to be some minor decreasing trend with increasing tube diameter (i.e., D P 12.7 mm), or with increasing detonation instability of the explosive mixture. The amplitude of the galloping cycle is also investigated and the upper velocity values show more fluctuations than the lower ones. The upper and lower values of the velocity in the galloping cycle at different conditions vary from V CJ to 1.5V CJ and 0.3V CJ to 0.65V CJ , respectively. To illustrate the role of flow instability on galloping detonations, experiments are performed with a spiral inserted into the 12.7-mm diameter tube to generate perturbations artificially. Detonation waves are observed to re-initiate quickly after passing the spiral and another cycle of galloping detonation was formed in the remaining part of the tube.