This approach was made difficult by the absence of any reliable technique for identifying platelets which, as has been known for many years (Jones, 1851; Bizzozero, 1882; Eberth and Schimmelbusch, 1886) form the major component of thrombi that arise in flowing blood. Fibrin can be observed in routine haematoxylin-and-eosin stained sections, and its nature confirmed by such methods as the picro-Mallory (Lendrum and McFarlane, 1940; Lendrum, 1949) and Mallory’s phosphotungstic acid-haematoxylin stains. At a more refined level fibrin and fibrinogen antigens have been identified by an immunofluorescence technique (Crawford and Woolf, 1960; Woolf and Crawford, 1960; Woolf, 1961). The success of these methods of demonstrating fibrin led, however, to some neglect of perhaps the more important component, platelets. Unfortunately in sections stained with haematoxylin and eosin platelets have almost the same tinctorial characteristics as fibrin. As fibrin and platelets are so often found in intimate association, it is frequently impossible to distinguish between them with this stain, unless the morphology of the section is very favourable indeed. Phosphotungstic acid-haematoxylin and the various modifications of Mallory’s trichrome are capable of demonstrating, but do not clearly distinguish between, platelets and fibrin. Phosphotungstic acid-haematoxylin stains platelets a dull blue-brown colour that is insufficiently distinct from the blue coloration of fibrin to be of great value; and the picro-Mallory method (Lendrum and McFarlane) when used in conjunction with mercuric fixatives as recommended by Lendrum usually stains platelets a colour very similar to the crimson colour of fibrin. Occasionally platelets take on a bluish tinge, but this is too inconsistent to be useful. With old lesions in which the platelet and fibrin masses have become extensively organised the special stains give little more information than the routine haematoxylin and eosin.

With these difficulties in mind I have approached the problem of platelet recognition in two ways. First, an attempt was made to modify the usual histological stains to enable platelets and fibrin to be distinguished in thrombi in which the architecture was well preserved and it could be expected that some platelet aggregates would be intact. The second approach was to use an immunological method to detect the presence of platelet antigens rather than whole platelet masses.