Canines trained to alert to traces of flammable liquids at a fire scene are useful to identify locations to collect samples for laboratory analysis. In some instances, no samples are collected or laboratory testing of samples collected following a scent alert fails to identify a residual flammable liquid and potential accelerant. Then, an attempt may be made, through testimony of the dog's handler, to introduce at trial information regarding the canine alert to indicate the presence of an ignitable liquid at the scene. Canine handlers contend that the dog has greater sensitivity to typical accelerants than laboratory techniques but scientists counter that the specificity of canine detection is unknown. Unverified canine indications have been used in a number of cases and challenges to several of these have reached the appellate level. Examinations of court decisions on admissibility of canine alerts shows that they have been as varied as the arguments pro and con. References, note