Another method, used by the majority of more recent authors, is the relative enumeration of blood platelets in the stain dry specimen. Ehrlich found that the blood platelets were picked out by their deep red colour, corresponding to the amount of alkali they contain, in preparations treated by the iodine eosine method (see p. 46). Rabl's new method is much more complicated and in no way more serviceable, depending on a stain with iron haematoxylin recommended by E. Haidenhain for demonstration of the centrosomes. A process of Rosin's, not yet published, is more convenient. It consists in fixing the dry preparation for 20 minutes in osmic acid vapour, and staining in a concentrated watery solution of methylene blue. * * * * * With regard to the significance of the blood platelets, most authors, of whom we should before all mention Hayem, Bizzozero, Laker, assume justifiably that they are preformed in the living blood. The view opposed to this, advocated more particularly by Loewit, that these forms first arise in the blood after it has left the vessels, we may describe on the grounds of our own extensive observations as inaccurate. The blood platelets, on the grounds of their small size and complete lack of nuclear substance, are generally regarded as not analogous to real cells. Whether they represent intravital precipitation of substances of the plasma, or whether they are budded off from the cells, cannot at the present be decided with certainty, though many facts seem to support the latter assumption. That they contain glycogen (see p. 45), marks them as descendants of the blood cells. Moreover, appearances are often met with in dry preparations that arouse the suspicion that the platelets arise from the red blood corpuscles (Koeppe). Arnold has further observed processes of budding in the red blood corpuscles not only extravascularly but also intravascularly in the mesentery of young guinea-pigs, and has seen the elements that were cut off change into forms free from haemoglobin. Our knowledge too of the physiological function of the blood platelets still needs much amplification. The original view of Hayem, who regards the blood platelets as early stages of the red blood discs, and for this reason calls them "haematoblasts," is, according to the judgment of most haematologists, untenable.