About 2,200 immigrant children in Sweden were studied by means of tests and teacher questionnaires. The children, representing 36 nationalities, were born outside Sweden and were distributed over the nine grades of the Swedish Comprehensive School. Various functional language skills in Swedish as a second language (L2), nonverbal intelligence and socio-emotional adjustment as measured by teacher opinions were studied as functions of age and length of residence. These two independent variables were found to be uncorrelated with each other and with a number of other background variables. All language variables were rather strongly related to age, and the F ratios were highly significant with one exception, Free Oral Production. The language variables were weakly or not at all related to length of residence, with the exception of Free Oral Production. The socio-emotional variables were not related to age, nor to length of residence. The pattern of nonverbal intelligence is inconsistent. As the relation of language to age is positive, i.e., L2 learning ability increases with age in the span of 7 - 17, and no optimum is found, it is suggested that the theoretical position of Penfield and Roberts, Lenneberg and others, which predict an optimum in L2 learning ability before puberty, is not tenable. (Author/AM)