As a sample of semen contains spermatozoa of various types depending on their normality, stage of maturity and degree of cytolysis, a quantitative study of the structure of the sample must consider the frequency of the different morphological types in a representative sub-sample. The effects of treatments may be determined by comparing different sub-samples. Since the methods that have been described for preparing spermatozoa for electron microscopy (Birch-Andersen & Blom, 1963; Saacke & Almquist, 1964; Healey, 1969; Quinn, White & Cleland, 1969) have disadvantages which make them unsuitable for use in such a study, a number of different methods have been examined. The procedures described below were developed for speed and convenience. They have given consistent results when used to prepare samples of spermatozoa from various sources (e.g. epididymis, semen, diluted and stored semen) and from a variety of mammals including the domestic ungulates, carnivores and rodents. For electron microscopy,