In more than 30% of milk samples from clinical and subclinical bovine mastitis, bacteria fail to grow even after 48 h of conventional culture. The "no-growth" samples are problematic for mastitis laboratories, veterinarians, and dairy producers. This study provides the first investigation of the bacteriological etiology of such samples, using a real-time PCR-based commercial reagent kit. The assay targets the DNA of the 11 most common bacterial species or groups in mastitis and the staphylococcal blaZ gene (responsible for penicillin resistance) and can identify and quantify bacterial cells even if dead or growth-inhibited. A study was made of 79 mastitic milk samples with no-growth bacteria in conventional culture, originating from cows with clinical mastitis. Of the 79 samples, 34 (43%) were positive for 1 (32 samples) or 2 (2 samples) of the target bacteria. The positive findings included 11 Staphylococcus spp. (staphylococci other than Staphylococcus aureus), 10 Streptococcus uberis, 2 Streptococcus dysgalactiae, 6 Corynebacterium bovis, 3 Staph. aureus, 1 Escherichia coli, 1 Enterococcus, and 1 Arcanobacterium pyogenes. The positive samples contained as many as 10(3) to 10(7) bacterial genome copies per milliliter of milk. This study demonstrates that in nearly half of the clinical mastitis cases in which conventional culture failed to detect bacteria, mastitis pathogens were still present, often in substantial quantities. The clearly elevated N-acetyl-beta-d-glucosaminidase activity values of the milk samples, together with clinical signs of the infected cows and quarters, confirmed the diagnosis of clinical mastitis and indicated that real-time, PCR-based bacterial findings are able to reveal bacteriological etiology. We conclude that all common mastitis bacteria can occur in large quantities in clinical mastitis samples that exhibit no growth in conventional culture, and that the real-time PCR assay is a useful tool for bacteriological diagnosis of such milk samples. Low bacterial concentration is commonly speculated to explain the no-growth milk samples. This hypothesis is not supported by the results of the current study.