I propose a reading of Berkeley's Essay towards a New Theory of Vision in which Molyneux-type questions as thought experiments instead of arguments. First, I present the general argumentative strategy in the NTV, and provide grounds for the traditional reading. Second, I consider some roles of thought experiments, and classify Molyneux-type questions in the NTV as constructive conjectural thought experiments. Third, I argue that (i) there is no distinction between Weak and Strong Heterogeneity theses in the NTV; (ii) that Strong Heterogeneity is the basis of Berkeley's theory; and (iii) that Molyneux-type questions act as illustrations of Strong Heterogeneity.