Presented is a critical inquiry about the product of the informal reading inventory (IRI) and about some of the elements used in the process of determining that product. Recent developments on this topic are briefly reviewed. Questions are raised concerning what is a suitable criterion level for word recognition. The original criterion of 95 percent correct pronunciation for word recognition is considered too high. The application of one set of performance standards uniformly across all grade levels is questioned. Neither quantitative nor qualitative uniformity across passage levels is considered appropriate in dealing with errors. It is noted that present knowledge of the IRI precludes definitive statements concerning the hierarchial relation of the independent, instructional, and frustrational reading levels. The real value of the IRI is seen as affording the possibility of evaluating reading behavior in depth and as offering potential for training prospective teachers about reading behavior. References and tables are included. (WB)