g 66 Davis. — A Comp avis on of the Reduction Divisions of 
to be simply features of a continued chromatic condensation and contraction 
which transforms the system of thickened threads into the spireme. 
Gregoire (’ 10 , pp. 325 and 387) has expressed the opinion that the 
‘ strepsinema 5 stage will be found in Oenothera exhibiting the characteristic 
longitudinal division or doubling of the spireme, which is so important 
a feature of his views on a side-by-side association of chromosomes (para- 
synapsis) leading up to diakinesis. He believes that the chromosomes 
will be found to arise side by side in pairs as a result of this longitudinal 
doubling of the spireme. Gregoire, then, suggests that the arrangement of 
the chromosomes in Oenothera in a chain is a later development due to 
the separation at one end of the chromosomes in each pair during dia¬ 
kinesis and the secondary union of the free ends of the pairs to form a chain 
not of 2, x independent sporophytic chromosomes but of x pairs of chromo¬ 
somes. This ingenious interpretation is intended to bring certain pecu¬ 
liarities of chromosome formation and arrangement in Oenothera into line 
with Gregoire’s well-known theory, which he believes to be of general 
application throughout the animal and plant world. 
The difficulties confronting Gregoire’s interpretation described above 
seem to the writer insurmountable for the following reasons :— 
First. With respect to the segmentation of the spireme, this is one 
of the most evident and direct processes in the entire history of reduc¬ 
tion phenomena in Oenothera. It is illustrated in PI. LXXI, Figs. 11-16 
for biennis (Davis, ’ 10 ), Figs. 13-17 for Lamarckiana, and PL LXXII, 
Figs. 57-59 for gigas. The spireme is not complexly coiled, the stages 
of segmentation are easily found, the form of the chromosomes is so simple 
that no especial difficulty is offered to the study of their arrangement, and 
the usual diploid (sporophytic) number, fourteen (twenty-eight in gigas), is 
so small that they may be counted with accuracy. It is perfectly clear 
that the spireme segments into the full set of 2 x (diploid) chromosomes 
arranged end to end and not into .r pairs of chromosomes arranged side 
by side. There is to the writer no doubt of the facts of the case or 
of the interpretation that these constitute the full set of sporophytic 
chromosomes formed end to end by the segmentation of a univalent and 
not a double spireme. 
Second . There is no general pairing of the chromosomes in biennis , 
Lamarckiana , rubrinervis , or gigas. The occasional association in these 
species of chromosomes to form ring-shaped pairs, whether united at one or 
both ends, comes about through the approximation of chromosome seg¬ 
ments forming the two sides of loops in the spireme. The exceptional 
conditions in certain material of grandiflora (Davis, ’ 09 , pp. 558 and 559), 
where seven ring-shaped pairs of chromosomes are formed, result from the 
presence of a spireme composed of complexly coiled loops which segment in 
such a manner that seven loops (later rings) are generally differentiated, 
