Omission of prepositions (null prepositions) in wh-questions (and relative clauses) by L2 learners of English was taken to indicate in Klein (1993) that UG in L2 acquisition might not be fully functioning. However, in two more recent analyses, the phenomenon is argued to fall within the limits of UG, reflecting either a general process in L2 acquisition whereby interrogatives might be represented as bound construals in the IL grammar (Dekydspotter, Sprouse, and Anderson, 1998), or resulting from the P-stranding property of L2 English and the learners' attempt to reconcile this property with their L1 grammar (Klein, 2001). In this paper I test these claims against empirical findings from L2 Hebrew, a non-P-stranding language with overt A'-movement. Based on this I argue that null prepositions in L2 acquisition might have two independent sources: (i) incorporation of the null preposition into the verb and subsequent A'-movement of the relevant operator; or (ii) misanalysis of small locative Ps (e.g. in) as a null D-head in the nominal structure of locative PPs. The former is necessarily triggered by P-stranding in L2, and therefore not attested in L2 Hebrew, while the latter is not limited to L2 acquisition at all and therefore may be attested in any L2. The proposed analysis supports the relevance of UG in L2 acquisition, showing that null prepositions in L2 acquisition reflect an interaction between UG-based syntactic processes such as Case-checking, or licensing of empty categories and their language specific manifestations.