Delayed sample analysis is not a rare circumstance in clinical and laboratory practice, especially when blood samples are shipped to distant centralized laboratories, when the analysis can not be readily performed, or when retesting is appropriate. In this study we sought to evaluate the stability of conventional and new hematologic parameters in blood specimens stored for as long as 24 hours at 4 degrees C. Of the 21 hematologic parameters tested with the use of the Advia 120 hematologic analyzer (Bayer Diagnostics), means for paired samples of specimens differed significantly over the 24-hour storage period for hematocrit, main corpuscular volume, percentage of macrocytes, platelet count, main platelet volume, reticulocyte count and percentage, and reticulocyte hemoglobin content (all P < .01). We noted no significant changes in the other parameters tested or in the white blood cell differential. The overall distribution of the immature reticulocytes fractions remained substantially unchanged, though the high staining-intensity fraction showed a considerable shift from the baseline measure. Bland-Altman plots and limits-of-agreement analysis showed mean biases between -4.8% and 37.2% and relative coefficients of variations ranging from 0.4% to 32.7%. The 95% agreement interval in the set of differences was satisfactory and almost within the current analytic-quality specifications for desirable bias. The results of this investigation suggest that, within certain limitations for parameters derived or calculated from cellular volumes, blood specimens stored for as long as 24 hours at 4 degrees C may be suitable for hematologic testing.