Development of Parts Not Due to Local Relations 65 
esis becomes confirmed indirectly by a whole series of 
other facts which Roux also has described and commented 
upon with his customary carefulness. 
Natural or artificial headless monsters for example, 
and in general all monsters lacking entire parts but other- 
wise normal, prove that no formative action or reaction 
is exerted by the head or by these other parts upon the 
rest of the organism. 
They speak therefore in favor of a centroepigenesis 
with independent networks of correlation in which it is 
necessary to suppose that the formative action must 
stream out entirely from a center toward the periphery. 
In all these monsters of which some part is lacking, there 
need be present only one part of the body namely the seat 
of the central zone of development. 
Roux in his researches upon the formation of half- 
embryos once observed, as an example of the disturbance 
of correlations of mass by the absence of one embryonal 
half, a lateral dislocation of the notochord and a corre- 
sponding retardation of development of the dorsal part of 
the endoblast lying near the median line as marked by the 
semi-medulla and by the ventral parts. “It is an interest- 
ing fact,’ he remarks, “that the axial parts can be laid 
down and developed with so considerable a shifting in 
relation to one another. For this indicates further that 
the development of many parts even of the main parts is 
not included in the form as such; the embryo does not 
live a formal life.” *® These facts would also confirm the 
hypothesis of independent networks of correlation, the 
displacement of whose material would not alter their 
Wilhelm Roux: Uber die kiinstliche Hervorbringung halber 
Embryonen etc Virchows Archiv. Bd. 114. October 1888. P. 132. 
Gesamm. Abhandl. Zw. Bd. P. 442. 
