The concept of 'reference intervals' proposed by NCCLS DOCUMENT C28-P has been introduced in Japan, and it has been widely accepted among laboratory scientists and clinicians. However, the terms 'reference values' and 'reference intervals' appear to be misunderstood by some. Normal ranges which have been used conventionally should not be simply renamed 'reference intervals'. The term 'reference intervals' should be strictly reserved for the ones which are determined based on the NCCLS recommendation. However, it is not an easy work, almost impractical for most clinical laboratories to follow strictly the NCCLS document C28-P. Therefore, the author suggests [correction of suggest] a multi-institutional cooperative activity to establish the reference intervals; the reference intervals for 13 serum proteins among Japanese adults have been successfully determined with the cooperative work of 7 clinical laboratories. 'Reference values' in a strict sense has been defined by NCCLS to be the observed values among reference individuals. However, 'reference values' in a broad sense should be understood to be any consensus value set by a professional group(s) for the purpose of clinical decision making.