Carcasses are potentially a valuable source of DNA for genetic studies of Pacific salmon (Oncorhynchus sp.). They can be collected at spawning grounds across a wide geographic area and include populations in streams in which it is logistically difficult or too intrusive to sample spawning fish. However, the quality and quantity of DNA from salmon carcasses is highly variable. Using a standardized set of microsatellite loci, we investigated the relationship of amplification success and genotyping errors to time since death and locus size in Chinook salmon (O. tshawytscha) carcasses. Amplification success declined rapidly from death, smaller loci had greater amplification success than larger loci, and genotyping errors were present in 5% of scored samples. Salmon carcasses can be a valuable source of genetic information; however, the level of effort needed to produce accurate and reliable data with microsatellite loci is considerable. Ideally, field collections should be frequent and focus on fresh carcasses. Data replication should also be incorporated into analyses to reduce amplification failures and genotyping errors. New techniques and markers should improve future cost-effectiveness when analyzing degraded tissue samples. However, we recommend that an evaluation of cost and time be performed at the onset of any population genetic study using salmon carcass tissues, regardless of the technique used.