The paper is part of an on-going study of English equivalents of the 10 most frequent Czech prepositions. In this case it is limited to the most frequent of them v/ve (in). The general aim is to determine whether and to what extent the choice of translate equivalent is influenced by the fact that the sequence including v/ve is an, open grammatical structure or a prefabricated lexical string. The first task is to identify which of the sequences including v/ve are grammatical and which are lexical, to find their representation and compare them with their translation equivalents. The assumption was that SL prefabricated strings tend to be translated by means other than a grammatical sequence (i.e. a prepositional equivalent) and that accordingly Czech lexical sequences (prefabricated, formulaic, etc.) with v/ve will have a higher proportion of non-prepositional equivalents than found in the whole sample. The results show however that the concept of prefabricated lexical strings and consequently their identification by purely statistical methods is not without problems. Still, the tests of correlation (using T-score,between presumably grammatical sequences and lexical sequences, and their corresponding translations, give some support to the hypothesis.