Cox, Friedman, and Sadiraj (2008) develop a model of reciprocity, „Revealed Altruism‟, which posits that a „more generous than‟ (MGT) offer elicits a „more altruistic than‟ (MAT) response. A MGT ordering is defined by two conditions. Condition a) states that MGT is ordered by the maximum potential increase in income of the recipient, or that the more you stand to receive from an offer, the more generous it is to you. Condition b) states that the increase in maximum potential income of the recipient cannot be less than the maximum potential increase in income of the proposers. In other words, Condition b) states that an offer cannot be self-serving, but it is not specified in Cox, Friedman, and Sadiraj (2008) precisely how b) affects the MGT ordering. I propose that a violation of b) is considered self-serving and is less MGT than when b) is not violated. I then experimentally study the empirical relevance of b) using two designs that hold a) constant, comparing MGT differences implied by responses. The first design is a variant of the Lost Wallet Game (Dufwenberg & Gneezy, 2000) with a negative outside option, and the second design is a modified Investment Game (Berg, Dickhaut, & McCabe, 1995) with elements of the Dictator Game implemented by Andreoni and Miller (2002). I find no empirical support that b) affects the MGT ordering.