492 
L. DOXCASTER. 
side, corresponding units being opposite one another. Several 
observers (10a, 28, 29) have described chromosomes as 
twisted round each other in pairing, and Morgan suggests 
that if, when they are so twisted, the split which separates 
the chromosomes is straight, as Jannsens has maintained (26 j, 
the resulting daughter-chromosomes will not be identical 
with those which paired, but will consist of parts of each. 
When the split occurs, units which are arranged next to each 
other will usually go into the same half, but units which are 
widely separated will often go into different halves. In this 
way he accounts for the different degrees of coupling which 
are found between different characters in the same species. 
The twisting of the chromosomes round each other in synapsis 
appears undoubtedly to occur in certain cases, but until the 
splitting across the twist postulated by Morgan has been 
shown with certainty to occur, his hypothesis must be 
regarded as almost entirely speculative. 
The last class of evidence with regard to the relation of 
chromosomes to Mendelian factors depends on direct observa- 
tion. At present few cases are known which bear directly on the 
question, and only two will be mentioned at this point. First 
may be mentioned the work of Federley (20). He found that 
hybrids between moths of the genus Pygasra showed that 
certain features of one species were dominant, while in other 
respects the hybrids were intermediate. When the hybrid was 
mated with one of the parent species, in most respects the off- 
spring were all again hybrid in character ; in one or two 
features, however, segregation took place. On investigating 
the behaviour of the chromosomes, he found that in the 
spermatogenesis of the hybrid nearly all the chromosomes 
failed to pair ; the spermatocyte chromosomes were almost 
of the somatic number, and divided equationally in both 
divisions. One or two chromosomes, however, paired and 
segregated normally. When the hybrid P. curtula d' x P. 
anachoreta $ was paired back with pure anachoreta $ 
the offspring contained almost a complete triple set of 
chromosomes, for the hybrid provided almost complete 
