Michotte et al.(1964) studied the tunnel effect in detail. When amoving object was occluded by an opaque still object for a while， the movingobject's identity was maintained. Even when the first object was white and thesecond object was red， the object identity was maintained. The moving object'sidentity is very strong.  Komatsu (2000) reported the situation that it was easy to destroy the movingobject's identity. In the condition that as soon as the whole of moving object wasoccluded by a black still object， a moving gray object appeared in front of the stillobject， some observers perceived that the moving object before occlusion wasdifferent from the moving object after occlusion. But in there， all objects were thesame size， and when the occluded moving object appeared in front of the stillobject， it always appeared at the same place and then went straight. If the stillobject is bigger than the moving object， nobody can expect where the occludedmoving object appears again， and then which direction it goes.  In this study， first， I examined the effect of the size of the still object on themoving object's identity. I found that when the still object was bigger than amoving object， more subjects perceived that the moving object before occlusionwas different from the moving object after occlusion.  Next， I examined the effect of the difference of the moving object's direction onthe moving object's identity. In the condition that after the whole of moving objectoccluded by the still object， it appeared at g places and then went to 4 directions，even the moving object went straight， more subjects perceived that the movingobject before occlusion was different from the moving object after occlusion.