From a review of literature on bovine mastitis published between 1833 and 1973 it is concluded that although herd control has caused a shift of emphasis from the clinical to the subclinical forms, diagnostic criteria have remained basically unaltered. Concurrent with this shift, improved bacteriological and cytological techniques are more sensitive than those available at the turn of this century. The sensitivity of these methods apparently has led to great confidence in their use in the diagnosis of mastitis. This culminated recently in a definition of mastitis as well as standardization of diagnostic criteria by the International Dairy Federation (IDF) which stressed the significance of potentially pathogenic bacteria and eleva1ed somatic cell counts (threshold 5 x 10⁵ cells/ml) in aseptically sampled foremilk. The validity of these widely accepted diagnostic criteria, particularly in the case of subclinical mastitis, is based on the assumption that an increase in the sensitivity of the methods should lead to an increase in diagnostic accuracy. Consequently, one would expect to find conclusive evidence of the diagnostic significance of the criteria used by the IDF. In contrast, this review attempts to reveal a conspicuous lack of adequately controlled experimental evidence supporting the diagnostic significance of criteria laid down by the IDF. They are of particular importance when dealing with subclinical mastitis, since the demarcation between physiological and pathological changes is considerably less distinct than in acute clinical mastitis. Thus no reference to pathological alteration of the udder epithelium, the primary symptom of mastitis, could be found in the literature assessed in terms of examination of milk. Similarly, the misleading influence of teat canal infections or lesions on the diagnostic accuracy of bacteriological and/or cytological examination of milk samples, obtained by "aseptic'' withdrawal via the teat canal, is disregarded. The physiological significance of Selye’s General Adaptation Syndrome or Schalm's Leucocytic Udder Barrier as non-inflammatory causes of elevated epithelial or leucocytes counts in milk appear to have been overlooked by workers who regard the presence of somatic cells in general or leucocytes in particular as synonymous to pus in milk. Due to absence of appropriately controlled experimental data, there is also no conclusive evidence to support the diagnostic significance of the IDF criteria referring to latent udder infections. Likewise it is not possible to distinguish between clinical and subclinical mastitis in udders with chronic indurative tissue changes in the absence of clinical alterations of the milk. The above considerations do not preclude the control of certain types of mastitis by application of IDF standards. However, because such criteria may result in some 43,13 ± 20,8% false positive diagnoses, it is conceivable that losses resulting from procedures based upon inaccurate diagnoses may far outweigh those caused directly by mastitis. The criteria of the IDF could, however, be used to great advantage for the diagnosis and control of mastitis if augmented by a test capable of establishing the physiological and pathological slate of the mammary epithelium. This is possible by means of a radial immunodiffusion test based on diffusion of bovine serum albumin (BSA), present in milk. The combination of BSA and the IDF criteria permits accurate assessment of udder health and teat canal infections.