This paper identifies trends in the population of students with learning disabilities (LD) in the public schools and explores reasons for these trends. It notes that there has been a 198 percent increase in the number of children served as LD between 1976-77 and 1992-93 due to a lack of consensus on the definition of LD and school-level processes designed to identify and provide services to students. It contrasts the authoritative definition of LD produced by the National Advisory Committee on Handicapped Children with various school-level processes that foster straying from these specifications. It notes the large amount of subjectivity at each stage of the process, including the teacher referral, the assessment process, and eligibility deliberations. Other reasons for the expansion of the LD concept include the common practice of ignoring exclusionary criteria and eligibility based on assessments conducted at one point in time. The paper goes on to note curricular consequences of the resulting heterogeneity of the school-identified LD population. It finds an unhealthy schism between research and practice fueled in part by the discrepancy between school-identified and research-identified students with LD. It proposes refinement of categories and multiple assessments of student progress. (DB)