The efficiency of detection of single base substitutions by single-stranded conformation polymorphism (SSCP) analysis was tested on 86 randomly distributed point mutations in a 193-bp-long DNA fragment of the mouse β-globin gene. Multiple parameters were varied, including electrophoresis temperature, buffer concentration, gel concentration, acrylamide-to-bis-acrylamide ratio, and/or addition of different compounds to the gel matrix. Gels with a higher concentration of acrylamide and lower crosslinking gave optimal separation, and all 86 mutations can be clearly distinguished from the wild type on a 5% or 7.5% (2.6% C) acrylamide gel at 4°C. Most of the mutations are also resolvable from wild type on gels with 5% urea or formamide, or 10% dimethylsulfoxide or sucrose. The relative position of the purine and pyrimidine-rich single strands were followed by an asymmetric PCR–SSCP technique. We found that most of the informativity comes from the purine rich strand, which appears to be much more sensitive to changes in the gel. The position or type of mutation showed no correlation with its ability to be detected. However, the neighboring base sequence around the mutation appears to have an effect on mobility. For example, A→G substitutions in GC-rich regions significantly increase the mobility shift of the purine-rich strand, while most G→A changes decrease it. We conclude that SSCP is a very efficient method for the detection of point mutations, if the parameters that effect the separation are optimized for a particular DNA fragment. © 1993 Wiley-Liss, Inc.