1887]  Metschnikof on Germ-Layers. 337 
and was enriched with many fresh facts, so that it soon became 
the password to the new road upon which embryology had 
entered. The theory received the greatest impulse from Kowal- 
evsky’s discovery of the development of Amphioxus(2). The 
embryology of this animal disclosed phenomena which linked 
together the development of Vertebrates and Invertebrates. As 
soon as Kowalevsky had discovered the two-layered ciliated 
larva of Amphioxus he began to look for analogous embryonic 
forms in other animals, and succeeded in establishing a great 
number of „very valuable facts. These investigations, having for 
their object the discovery of the most fundamental embryonic 
forms, such as might be compared with the early stages of Am- 
phioxus, were naturally followed out on animals of low grade 
with simply organized larvae. On the other hand, I turned my 
attention to the development of the higher Invertebrates, with 
the design of establishing here also the germ-layer theory. I 
first studied the embryology of Sepiola (4), found two germinal 
layers, and observed the part each played in building up the 
organism. Following up the investigation in the group of Ar- 
thropods, I failed to demonstrate satisfactorily the germinal 
layers in Insects, but found them in the higher Crustacea (5) 
(Nebalia), and particularly well in the scorpion (6). In the latter 
animal I at first (1866) found only two layers, but soon after 
(1868) discovered the third. I showed in the scorpion that the 
upper layer gives rise to the central nervous system; that the 
middle splits into two layers and forms a series of hollow seg- 
ments, by the fusion of which the body cavity arises; and that, 
finally, the under layer becomes the lining membrane of the ali- 
mentary canal. Supported by these facts, I concluded, in 1869 
{in a publication of the Educational Bureau), that the three layers 
of the scorpion embryo corresponded in all respects to the three 
Vertebrate layers. I was not deterred from this view by my 
belief at the time that the nerve-fibres were derived from the 
middle instead of the upper layer, since the peripheral nervous 
system of the Vertebrates was then generally considered to be 
mesoblastic. Thus the problems of comparative embryology 
were attacked on two sides, with the object of getting a good 
basis of facts. It was not until I had made out the main features 
in the formation of the germinal layers of the scorpion that 
Kowalevsky began to investigate the embryology of the Oligo- 
