Abstract The present paper evaluates both the variances and representativeness of the AOAC sampling procedure for bagged fertilizer (929.01a, 15th Ed., 1990) for concentrations of total nitrogen, available phosphoric acid, and soluble potash from 5% to 30% In blended, mixed fertilizers using materials In common use In 1983. The sampling variances determined did not differ significantly from sampling variances determined in earlier studies that are currently used by the Association of American Plant Food Control Officials (AAPFCO). The representativeness of the AOAC sampling procedure for bagged fertilizer was evaluated by testing the differences between AOAC bag sample means and rotary divider sample means (riffled down bags) of the blends (a) using AAPFCO investigational allowances (lAs) and (b) using t-tests. When AAPFCO lAs were used, there were only 2 bag means significantly below the rotary divider means. When t-tests were used, the bag means tested significantly lower than the rotary means for the 5, 10, 15, and 20% N blends and the 15% P205 blend. The bag means for the 5,10, and 30% P205 blends were significantly above the rotary means. There were no significant differences between bag and rotary means for K2O In any blends. Because Mests are more sensitive than AAPFCO lAs, smaller differences are judged to be significant. Based on AAPFCO lAs, the AOAC sampling procedure for bagged fertilizer probably satisfactorily represents what is In the bag. However, significantly lower N and higher P205 bag means based on Mests are cause for concern and should spur activity In determining the nature of this phenomenon, interlaboratory variances were also determined and evaluated. These tended to verify those In current use by AAPFCO. The representativeness of the AOAC stream sampling procedure (929.01b) was also evaluated and found to be an acceptable method for determining the true analysis of a blend.