<p>Behavior shown by severely mentally retarded childreren may be classified into two categories, i. e. Self-Stimulative Behavior (SSB) and Outer-Directed Behavior (ODB). The former is the kind of behavior directed to their own body like waving hands before eye or body rocking and often observed in the institutionalized severely mentally retarded, and the latter is the kind of behavior directed toward outer world surrounding them. SSB seems to be the behavior whose stimulus and response are completely enclosed within their body. Engaging in this kind of behavior results in the loss of the chance which advances their development of ODB. The purpose of this paper is to clarify the effect of situations in a ward of residentialinstitution for severely mentally retarded children on SSB and ODB, and to specify the type of SSB according to the degree of the effect. Three situations were studied. 1. Free-time in a dayroom where all children of the ward spent with a few maniplatable objects under supervision of one or two attendants. 2. Training-time in which children were intentionally given a kind of edcational care in small group with many sorts of materials fitted to their sbilities. 3. Meal-time. 36 ambulatory, severely and profoundly retarded children were observed employing a check-list by time-sampling method. Results are followings: 1. On the average, the highest occurence of SSB and the lowest occurence of ODB were observed at the free-time in day-room. At the training-time, however, decremment of SSB and increment of ODB were shown. Occurence of SSB and ODB decreased at meal-time. 2. Large individual differences of children in the degree of the effect of these situations on SSB and ODB were found, so 5 groups of children were divided by a cluster analysis. Group 1: The children showed the highest occurenrences of SSB and the lowest occurence of ODB at all three kinds of situations. The SSB in this group were charaterized by especially patterned form including quick movements such as complex hand movements. In the ward, these children retired to the corner and appeared to be withdrawn from the outer world. Group 2: Being dissimilar to group 1, the occurence of SSB decreased at meal-time. Most children of this group tended to be hypoactive and their SSB consisted of self-contact with their own body or gross movements. ODB in this group was scarecely observed but these children reacted to stimulation from other person passively. Group 3: higher occurence of SSB and ODB are observed regardless of the kinds of the situation. Group 4: In this group, transition of the children from free-time to training-time caused their SSB to decrease remarkably in accordance with so much increment of ODB. Group 5: The children in this group showed the lowest occurence of SSB, the occurence of ODB, however, was changed according to the kind of situation. These reults suggest that two types of SSB may be considered as to severely mentally retarded children. The first type of SSB, which was observed in the children belonging to group 4 and 5, appeares only in the situation where manipulatable objects or attendants are absent and rapidly disappeares in the situation where are the proper stimulus from the environment. This resembles the one which appear at an infant phase in the normal development. The second type of SSB, which was observed in the group 1 and 2, occurs with high frequency in every situation. This type of SSB consists of two distinctive forms of behaviors, the one is idiosyncratically characterized pattern including fine quick movements, the another includes body contact to himself or gross movements. Although this type of SSB predominates over the ODB which is scarecely observed in these children, all of them showing the second type do not always seem to be more retarded than those showing the first type. Induldging themselves in this type of SSB may lower the responsiveness to the e</p><p>(View PDF for the rest of the abstract.)</p>