The divisions in transformational grammar into surface and deep structure give rise to the hypothesis that imitative-repetitive drills will never go beyond the surface structure, and that an explicit verbalization of underlying structures will result in better achievement and proficiency in second language learning. The Project described here, carried out in English classes in Gothenburg, Sweden, attempted to test this hypothesis and to find out whether the same method would work equally well at different age and intelligence levels. Three methods were used: one "Implicit," in which the pupils practiced the "do" verb construction in oral and written drills without explanations or theoretical comment; and two "Explicit," one in which explanations are given in English, and one in which explanations were given in Swedish. Apart from the explanations, the lessons in all three groups were identical. Pupils were also grouped according to IQ; in the low intelligence groups, the "Implicit" method was considered the "best." Among the more intelligent pupils, no significant differences were found. It was felt, however, that if explanations were given, they should be in Swedish. Also, one reason for the low "Explicit" scores may be that explanations were of a transformational kind. [Not available in hard copy due to marginal legibility of original document.] (AMM)