The importance of vertebrate animals as seed dispersers (zoochory) has received increasing attention from researchers over the past 20 years, yet one category in particular, diploendozoochory, remains understudied. As the term implies, this is a two-phase seed dispersal system whereby a secondary seed predator (carnivorous vertebrate) consumes a primary seed predator or granivore (rodent and bird) with undamaged seeds in their digestive tract (mouth, cheek pouch, crop, stomach, or other organ), which are subsequently eliminated with feces. Surprisingly, although snakes are among the most abundant predators of granivorous vertebrates, they are the least studied group insofar as our knowledge of seed rescue and secondary dispersal in a diploendozoochorous system. Here, using live snake subjects of the Sonoran Desert (one viperid and two colubrid species) and seeds of the Foothill Palo Verde (Parkinsonia microphylla), a dominant tree of the same region, we experimentally tested germination frequency and rate, and seedling viability. Specifically, to mimic rodents with seed-laden cheek pouches, we tested whether wild-collected P. microphylla seeds placed in the abdomen of thawed laboratory mice and ingested by the snakes would retain their germination viability. Second, we examined whether seeds exposed to gut transit germinated at a greater frequency and rate than the controls. While we found strong statistical support for our first hypothesis, both aspects of the second one were not significant. Accordingly, we provide an explanation for these results based on specific life-history traits (dormant and non-dormant seeds) of P. microphylla. Our study provides support for the role of snakes as important agents of seed rescue and dispersal in nature, their potential as ecosystem engineers, and crucial evidence for the investment of field-based studies on diploendozoochorous systems in deserts and other ecosystems.