It is becoming increasingly recognized that asthma is a heterogeneous disease, whether based on clinical factors, including the patient's age at diagnosis, symptom spectrum and treatment response, triggering factors, or the level and type of inflammation. Attempts to analyze the importance of these characteristics to the clinical presentation of asthma have led to the appreciation of numerous separate and overlapping asthma phenotypes. However, these approaches are 'biased' and based on the clinician/scientist's own experience. Recently, unbiased approaches have also been attempted using both molecular and statistical tools. Early results from these approaches have supported and expanded on the clinician's concepts. However, until specific biologic markers are identified for any of these proposed phenotypes, the definitive nature of any phenotype will remain speculative.