An Aspergillus mold was isolated that had contaminated an unpreserved semisoft baked cookie with visible mycelial growth. Due to the xerophilic nature of the contaminating mold, the Aspergillus isolate went undetected using conventional plating media and incubation times. Spread plating onto rose bengal osmophilic potato dextrose (rose-osmo) agar, a reduced water activity (aw) medium, was used for isolation and enumeration of the isolate from contaminated cookies and in microbiological assessments of the production facility. Rose-osmo plates were sealed in plastic and incubated at 25°C for up to 6 weeks to allow the isolate to sporulate. The minimum aw that permitted growth was determined by inoculating both MY70GF agar plates and production cookies with fresh spores of the Aspergillus isolate. Inoculated samples were incubated for up to 6 weeks in humidity chambers adjusted to various moisture levels. The minimum aw required for growth of the Aspergillus isolate was found to be between 0.60 and 0.65 on MY70GF agar and approximately 0.65 on the cookie. To evaluate the potential delayed growth of the isolate on reformulated cookies of a lowered aw, long-term growth studies were conducted on cookies adjusted to various aw levels and monitored for up to 280 days. In these studies, the minimum aw test range found to support visible mycelial growth of the Aspergillus isolate was 0.67. Microbiological assessment of the production facility identified potential sources of xerophilic molds and the Aspergillus isolate (teleomorph Eurotium chevalieri) and identified the manner in which the cookies became contaminated. These methods and results proved helpful in both production and process risk assessments and in successful reformulation of the cookie at an aw low enough to prevent spoilage and provide an organoleptically acceptable product.

Twenty laboratories participated in a collaborative study to validate a 2-day hydrophobic grid membrane filter method using YM-11 agar for enumeration of yeast and mold in foods. Six naturally contaminated food products were included in the study: garlic powder, raw ground beef, walnuts, flour/meal, orange juice, and yogurt. The test method produced significantly higher results than the 5-day pour plate reference method for orange juice and significantly lower, though numerically similar, results for walnuts and yogurt. Differences between the test and reference methods were not significant for garlic powder, raw ground beef, or flour/meal. Repeatability and reproducibility were similar for both the test and reference methods in all cases. The hydrophobic grid membrane filter method for enumeration of yeast and mold in foods has been adopted by AOAC INTERNATIONAL.