338 BOTANICAL GAZETTE [way 
interpretation is correct, then we have here, somewhat as in 
ROSENBERG’S account for Drosera, a lateral pairing of long thread- 
like chromosomes, which afterward by contraction become short 
and thick. 
It has sometimes been urged as an argument for the parasynap- 
tic method of reduction, that the chromosomes are in pairs side- 
by-side in the somatic mitoses. When, as is very generally the 
case in somatic mitoses, the long axis of the chromosomes is several 
times their short axis, then for mechanical reasons, if they are 
to continue to keep together in pairs, they will naturally lie side- 
by-side in the crowded somatic nuclear prophase and metaphase. 
The same is true of Oenothera (as already referred to in this paper), 
in which the somatic chromosomes are from three to six times 
longer than broad. They are paired side-by-side in the equato- 
rial plate of somatic divisions. It is not yet known whether in 
prophase they lie paired side-by-side or end-to-end on a spirem, 
though the latter is probably the case. But in any case, they are 
clearly arranged end-to-end in the stages of reduction preceding 
diakinesis, though after the looped and folded chain segments, they 
frequently come to lie paired side-by-side. Such pairs, as I have 
pointed out in previous papers, are almost invariably joined by a 
linin connection at one end, showing clearly their manner of origin. 
This all goes to prove that while both end-to-end and side-by-side 
pairings of chromosomes occur, yet no great significance attaches 
to the difference. In the case of the heteromorphic chromosomes 
figured by C. MULteR (16) in root tips of Yucca, the long chromo- 
somes lie in pairs side-by-side, but in the practically globul 
chromosomes both axes are of the same length, and the distinction 
between a lateral or endwise pairing breaks down. : ; 
Miss Stevens (20) has shown that in one of the mosquitoes 
(Culex) “parasynapsis of homologous chromosomes often change 
to telosynapsis in the metaphase of the first spermatocyte.” She 
says (p. 216): “It is of especial interest to see in Culex a perfectly 
clear case of parasynapsis change in some cases to an eque y 
clear case of telosynapsis before metakinesis, while intermediate 
Ting stages and cases of overlapping ends also occur.” 
