JUNE 14, 1912] 
such as the cytological facts as to the longi- 
tudinal splitting of the chromosomes seem 
to demand, and shows how inconsistent 
with the assumption of even the simplest 
form of determinate arrangement of ele- 
ments in the germ plasm Mendelian con- 
ceptions are. It is obvious that the coup- 
lings seen in sex-limited inheritance sug- 
gest the connection of the factors repre- 
senting them with a sex chromosome. 
That is why such characters are called sex- 
limited. But to actually imagine such di- 
verse and easily transferred characteris- 
ties of color, ete., affecting such widely 
separated parts of the adult organism as 
inhering in or connected with a single sex- 
chromosome is just as obviously opposed to 
the common interpretation of the serial 
arrangement of the chromosomes and their 
parts. Other cases of Mendelian coup- 
lings, such as that of long pollen and pur- 
ple flower color, are just as inconceivable 
on the basis of a serial corpuscular repre- 
sentation of genes in the chromosomes. To 
be sure, Bateson is now inclined to give up 
the conception of gametic coupling, but 
the idea has been widely taken up and 
seems fairly well established among Men- 
delian conceptions. It differs little, of 
course, from the older ideas of correlation 
in heredity. The factors in heredity in 
such cases are certainly more easily con- 
ceived as due to the characteristic consti- 
tution of the germ plasm, as a whole, than 
as represented by unit corpuscles. 
One of the strongest pieces of cytolog- 
ical evidence that the number of the chro- 
mosomes is due to their own specific and 
individual characters, rather than any 
mere mechanical necessities in the cell or- 
ganization, such as Fick assumes in his 
maneuver hypothesis, is found in the 
work of the Marchals on aposporous 
mosses. By regenerating the gametophyte 
directly from the sporophyte without the 
SCIENCE 
913 
intervention of the reduction division 
(apospory) they have produced diploid 
gametophytes. These in turn in fertiliza- 
tion have produced tetraploid sporophytes, 
and these by regeneration again tetraploid 
gametophytes. No stronger evidence for 
the permanence and independence of each 
chromosome could be produced. An excess 
of chromosomes can only be gotten rid of 
by a specific reduction division, that is, by 
separating them without the splitting of 
each into two. Some of the diploid gameto- 
phytes seemed vigorous, the tetraploid 
gametophytes were weak, and the physio- 
logical limit for the number of chromo- 
somes in one cell was probably reached in 
them. These cases, along with those in 
ferns, show that an excess in the amount. 
of the germ plasm doubling or trebling 
the representation of each hereditary qual- 
ity need not necessarily affect the morpho- 
logical characteristics of the organism, 
and are in strong contrast with such cases 
as that of @nothera gigas and certain 
races of bananas, in which a doubling or 
trebling of the chromosome number is as- 
sociated with marked structural changes 
in the plant. When the excess number of 
chromosomes fails to produce visible ef- 
fects, the condition is perhaps analogous 
to that of ordinary latency. The condi- 
tion of ineffectiveness of the excess chro- 
mosomes in these diploid and tetraploid 
gametophytes of the mosses, so far as their 
relation to morphogenetic processes is con- 
cerned, also seems analogous to that of the 
chromosomes of the recessive parent in the 
F, generation in cases of Mendelian domi- 
nance. Cells containing two complete sets 
of chromosomes may show complete re- 
semblance to one parent or mixed, mosaic 
and intermediate resemblance, depending 
on the degree of prepotency or latency 
represented by the conflicting sets of chro- 
mosomes. Mendelian dominance and re- 
