1889.] On Inheritance in Evolution. 1069 
hypotheses present the difficulty that we must conceive of each 
particle or " gemmule " derived from a j^iven part of the organism 
finding its way through the circulation to its exact place in the 
growing embryo; or otherwise, of transmitting its peculiar mode 
of motion to the correct molecules of the embryo, without error 
as to locality. The difficulties to be encountered in accomplish- 
ing such a feat seem to be insuperable. The transmission of a 
mode of motion organized in a central nervous system, is less in- 
conceivable. This central system is the seat of a composition of 
incoming stimuli and of outgoing energies, the resultant of both 
combined constituting the active agency in the production of 
automatic adaptive or intelligent adaptive movements of any and 
all of the organs. It appears to me that we can more readily 
conceive of the transmission of a resultant form of energy of this 
kind to the germ-plasma than of material particles or gemmules. 
Such a theory is sustained by the known cases of the influence 
of maternal impressions on the growing fa.tus. Going into greater 
detail we may compare the building of the embryo to the un- 
folding of a record or memory, which is stored in the central 
nervous organism of the parent, and impressed in greater or less 
part on the germ-plasma in the order in which it was stored. 
The basis of memory is reasonably supposed to be a molecular 
(or atomic) arrangement from which can issue only a definite cor- 
responding mode of motion. That such an arrangement exists 
in the central nervous organism is demonstrated by automatic and 
reflex movements. That the entire record is not repeated in 
automatic and reflex acts, but only that part of it which was last 
acquired, may be regarded as due to the muscular and other 
systems concerned in it having performed it most recently, and 
having for a longer or shorter period omitted to perform the 
older movement, because the latest structures of the organs would 
render the performance of the old movements impossible. In 
other words, the physiological division of labor extends to mem- 
ory at the basis. In the case of the germ-plasma no specialisation 
exists, so that the entire record may be repeated stage after stage, 
thus producing the succession of type-structures which embryol- 
ogy has made familiar to us. In the process of embryonic growth, 
