Recent studies concerning SLF-constructions have challenged different properties of subject gaps from a theoretical point of view. For instance, Fortmann (2005) argues that under clearly defined conditions not only subjects can be elided, but also objects. Reich (2009) argues that the gap is not always structurally bound (c-commanded) by its antecedent. If these claims can be confirmed, this would have extensive consequences for the syntactic and semantic analysis of SLF-constructions. This paper investigates these hypotheses experimentally and arrives at the following conclusions: 1. Elided objects are, even under the conditions following Fortmann (2005), not rated as acceptable; 2. Reich's (2009) predictions relating to possible antecedents are not, or at least only in a limited way, confirmed. Thus, a substantial modification of the classic analysis (Höhle, 1990) does not seem indicated.