ed that according to
Barker the eosinophil granulations appear to contain iron.

The key-stone of the hypothesis of the secretory nature of the granules
is the direct observation of a secretory process in the cells bearing
the granules. Naturally these researches offer extraordinary
difficulties since only the coincidence of a number of lucky
circumstances would allow the passage of dissolved granule substance
into the neighbourhood to be followed. Kanthack and Hardy have succeeded
in demonstrating the secretory nature of the eosinophil granules of the
frog. When, for example, anthrax bacilli are introduced into the dorsal
lymph sac of the frog they exert a positive chemiotaxis on the
eosinophil cells. The latter come in contact with the bacilli, and
remain for some time attached to them. During this period Kanthack and
Hardy observed a discharge of granules from these cells, which now
possess a protoplasm relatively homogeneous. Afterwards these cells move
away from the bacilli, and are succeeded by the polynuclear neutrophil
cells, as will be mentioned later. These authors were further able to
observe gradual accumulation of granules in eosinophil cells in lymph
kept under microscopic observation as a hanging drop, and thus
demonstrated that they undergo the two stages characteristic of
secretion, (1) appearance of granules within the cells, (2) discharge of
these granules externally.

The mast cells too seem suited for this purpose since their specific
substance is strongly characterised by its peculiar metachromatic
staining, and is further especially readily recognisable, since by its
great affinity for basic dyes it remains plainly stained, even in
preparations that are almost quite decolorised. In fact appearances of
the mast cells are not infrequently found, which must be referred to
excretory processes of this kind.

In the first place it is occasionally seen that the mast cell
granulation is dissolved within the cell, and diffuses in solution into
the nucleus. In place of the well-known picture of the mast cell (see
page 76) of a colourless nucleus, surrounded by a deeply stained
metachromatic granulation, a nucleus is present intensely and
homogeneously stained in the tint of the mast cell granulation,
surrounded by a protoplasm shewing but traces of granules.

Still more convincing is the presence of a peculiar halo of the mast
cells, described by various authors. Ehrlich first shortly mentioned
this ha