1887] Metschnikoff on Germ-Layers. 431 
ued as before to function as phagocytes. This condition is found 
in many Ccelenterates (where, however, in certain cases, addi- 
tional cells migrate from the ectoderm, as in Corals, according 
to Kowalevsky and Marion), in Echinoderms, many Worms, etc. 
In some forms the whole mesoderm, in others only a part, per- 
manently retained the original phagocytoblast condition. With 
many of the latter it came to pass that the mesoderm was formed 
by means of special sacs, which were constricted off from the 
endoderm. The cells of which these sacs were built ceased 
more or less completely to play the part of phagocytes. Such 
a condition appears in the Ambulacraria and in Vertebrates. 
In other animals with a simpler development,—for instance, 
Worms, Molluscs, and many Arthropods,—such mesoderm sacs 
have been looked for in vain; so that in these animals it is pos- 
sible that the entire mesoderm has been derived from the origi- 
nal phagocytoblast without the aid of special endodermal sacs, 
though probably with some help from the ectoderm. In suc 
cases the mesoderm has been able gradually to differentiate itself 
into a somatopleure and a splanchnoplcure, without running 
through a sac-like stage. In Arthropods with a large amount 
of food-yelk the mesoderm is functional at a very early date, 
it being the habit of the mesoderm cells to devour the yelk- 
globules. Under such circumstances it can be understood how © 
the mesoderm, even within the most recent time, has often been 
taken for the endoderm. In its origin dependent on the pha- 
gocytoblast, in part also on the kynoblast, the mesoderm sooner 
(Ctenophora) or later acquires its freedom and appears as a special 
germ-layer, which plays an important part in the development of 
the embryo. In opposition to the authors who think the meso- 
derm had its origin in sexual organs or muscles, I believe it was 
originally a part of the phagocytoblast, and as such took part 
in the inception or absorption of food. The importance of the 
mesodermal phagocytes in physiological and pathological pro- 
cesses, as well as the morphology of the mesoderm, have been 
discussed by me in other papers (7, 9, 10, II, 12, 13), to which I 
may refer, and thus avoid repetition. 
In conclusion, I wish to say that not until we are enlightened 
to the utmost as regards the primitive condition of the Metazoa 
will it be possible to place comparative morphology on a safe 
basis. But so long as the question of the germinal layers is 
