On Reproduction, Animal Life Cycles. 
93 
the biophores and determinants of Weismann, the molecular units of 
the physiologists, all of these being hypothetical; then empirically 
demonstrable elements such as the centrosomes or the chromosomes. 
The amplest units are those of certain palaontalogists, such as Hyatt, 
a unit being a phylum; and the units of certain students of environ- 
mental conditions, whereby an assemblage of interdependent species 
together with the environmental influences affecting them is taken as 
a unit of study. 
With the manifold subdivisions of biology, the morphological, physi- 
ological, taxonomic and ethological, room is given for the establish- 
ment of most varied units depending upon the special problems of the 
thinker. In general we may say that two main methods are open : 
to establish the smallest possible units ; or to establish more compound 
ones. The former method has the advantage of being more analytical, 
it permits us to subdivide our ideas so far as possible and to reduce 
all phenomena to the smallest conceivable energies and quantities. 
Thereby it is commonly supposed that we have explained matters, 
for we can go no further than the conceivable. But this method of 
unit formation has a disadvantage that is very serious in any empiri- 
cal study : it introduces hypothetical quantities before there may be 
any necessity to conjure them. For when we argue from units that 
cannot be in any way experimentally studied, we have of course no 
experimental means of testing whether our conclusions are correct or 
not. 
For these reasons it would appear to be the safest working plan not 
to resort to purely hypothetical units so long as there remain phe- 
nomena that can be experimentally investigated. Biology has an end- 
less series of phenomena, large phenomena if such phrase might be 
used with propriety, many of these have been catalogued, and the re- 
lations of them should be determined as far as possible empirically 
before we resort to hypothetical aids. That is to say, it will be most 
fruitful to maintain units that are demonstrable, because at any point 
in the discussion of the problem we are able to point to an actuality. 
Now if we select an empirical unit we will soon find that we can- 
not stop at anything less than the person, the whole organization. 
For a chromosome is merely a portion of a cell, not an individual 
capable of independent existence; similarly a cell is merely a subor- 
dinated part of a tissue, as this is of an organ; then all the organs 
are mutually interdependent. The whole person acts as one, not as 
the sum of many separated parts, it is the true biological individual. 
The very development of a nervous system, directed towards co-or- 
dination of parts, is evidence of this. 
Yet the biological unit cannot stop with the person, because a single 
