The study exploits access to parallel corpus data in order to explore the type and distribution of English equivalents of two of the ten most frequent prepositions in Czech, spatiotemporal v/ve and po. Based on 600 parallel concordance lines for each preposition (4 contemporary fiction texts), the analysis identifies prepositional and non-prepositional equivalents and correlates their distribution with syntactic analysis and the semantics of the Czech prepositions. The equivalence patterns of the two prepositions exhibit clear differences (in the proportion of prepositional/non-prepositional equivalents, the presence of a dominant equivalent and other features) and specific idiosyncrasies. The results are of theoretical and practical relevance, offering a substantial improvement on the lexicographic description in Czech-English dictionaries.