66 BULLETIN 754, U. S. DEPARTMENT OF AGRICULTURE. 
some inexplicable reason they assumed that the ratio between the 
number of gametes in which the parental combinations occurred 
together to the number in which they were separated would fit a 
definite series composed of the familiar Mendelian ratios of 8 to 1, 
7 to 1, 15 to 1, etc. From this arbitrary series they have evolved an 
elaborate system of cell division to account for the gametic asso- 
clations. | 
As higher couplings were secured these authors came to the reali- 
zation that an insufficient number of cell divisions occurred between 
synapsis and the formation of the gametes to give these higher 
ratios. As a consequence they have assumed that segregation takes 
place earlier in the life history of the organism. 
Working on the same problem, Morgan came to the conclusion 
that correlations were due to the fact that the correlated characters 
were located on the same chromosome (11). At synapsis the chromo- 
somes derived from one parent pair with those from the other parent 
and presumably twist around each other. At the maturation divi- 
sion these pairs of twisted chromosomes split, resulting in the genes 
that are located close together along the chromosome falling to- 
gether more often than apart. The degree of correlation depends 
upon the distance separating the character determiners, or genes. 
This distance is determined by the percentage of gametes bearing 
the characters derived from opposite parents which are called 
“ crossovers.” Thus the adherents of this theory would explain a 
gametic ratio of 3-1-1-3 as the result of the correlated characters 
lying 25 units apart on the chromosome. 
Morgan and his coworkers, unhampered by an arbitrary gametic 
series and in fact working with material very unsuited for such an 
analysis owing to differential death rates, have amassed a wealth of 
material which has certainly served to a the linkage theory in an 
exceptionally strong position. 
These authors eae no definite gametic series, and in the’ Eas 
status of the theory such a series would be meaningless, but while 
the theory makes little provision for such a series, it does not pre- 
clude it. 
Once having established the number of units separating the cor- 
related characters, deviations too large to be ascribed to chance are 
looked upon as the mutation of the locus of one or both of the char- 
acters, the value of the theory resting upon the infrequency of such 
departures. Whether we look upon a given gametic series as the 
result of unequal cell division or whether we assume that the gametic 
ratios are the result of the correlated characters lying a certain dis- 
tance apart on the same chromosome, the question arises, Is the cor- 
relation of the same intensity between two characters for the in- 
