The research described studies delayed failures in software using high volume automated testing (HVAT) and investigates the effectiveness of different HVAT techniques; such techniques include genetic algorithms, model-based testing, penetration testing, robustness testing, and random (stochastic) testing. A delayed failure is a failure that occurs some time after the conditions that lead to the failure are applied. There appear to be no studies of delayed failures of software in the literature and no comparative studies of the effectiveness of different HVAT techniques; therefore research in this area can make an important contribution. Delayed failures in software are unlikely to be revealed by conventional testing techniques; a HVAT technique that systematically reveals delayed failures could lead to improved reliability of software and reduced costs. Experimental work is in progress using the MySQL database server as the software under test.