Following the preceding paper, the author now discuss the effects due to the depth of water to the damages of cylindrical shells, because it seems to him not being well discussed yet on the quantitative measurement of water pressures due to effects of expansion of successive gas globes and damage amounts occurred by explosion in the deep sea. The cylindrical shells used were made of high strength steel and stiffened by inner frames with different span lengths. After hanging the cylindrical shells by wire ropes at the depth of about 30, 150 or 300 meters under the water on the about 2, 000 meters depth sea, the explosion tests were pursued by blasting TNT spherical explosives from the bottom, yielding to the small damages on shells. This paper is mainly concerned with small local damages occurred between frames, and hence the large ones in case of buckling of frames were disregarded. Comparing both results of experiments and calculations on water pressures and displacements, the semi-empirical formulas of time lag among peak pressures due to gas globes and the amounts of maximum concavities of damages on the cylindrical shells were determined. The formula of maximum concavities was further compared with the experimental formula which was developed quite recently in USA in case of variable depth of water, showing good coincidences between both results up to the so-called safe depth of water. The formula of maximum concavities which is available to the mild steel material was also determined and compared with the experimental results of the cylindrical models carried out by changing the depth of water, showing reasonable coincidences between both calculated and experimental results.