In further pursuit of the byssinosis problem, viable counts for total, gram-negative, and presumptive coliform bacteria were made on fiber from across the U.S. Cotton Belt in the crop of 1985. As in earlier work, the 1985-crop samples were highly variable in counts even within a single grade from a single location. No clear relations of counts to grade nor to fiber pH were seen. Many samples had a total population of more than 10 million bacteria per gram of fiber. Many had more than a million gram-negatives. Presumptive coliforms were always delectably present but at very low levels. Average levels for samples for five crop years are shown here for comparison, the data in total representing more than 800 samples. The results show that values for samples from the San Joaquin Valley of California have been regularly lower than the Beltwide average, a result attributed to the dry weather which occurs in the Valley between boll cracking and harvest. No other major areas exhibiting regular occurrence of counts either distinctly below or above the Beltwide average have been detected to date. Implications of the data in respect to the byssinosis problem are discussed.