Pollen Development in Lactuca . 373 
the interpretation that they have anything to do with a possible conjugation 
of parallel threads. While it is difficult to affirm from direct observation 
that no such parasynaptic conjugation of parallel threads takes place during 
synizesis, yet we have found no evidence whatever which would bear such 
an interpretation, and we believe the nature of the later stages excludes the 
possibility of its occurrence. The thread might of course become double at 
this time through a split, as has been described in various forms, but in 
Lactuca we have found no indication of a split during this period. The 
thread remains persistently single as it becomes progressively shorter and 
thicker. 
The segmentation of the spireme into separate loops is a progressive 
affair, different loops becoming gradually detached. Figs. 23-25 show 
a stage in which a few loops are free while the others are still connected in 
a continuous spireme. At this time the characteristic twisted appearance 
of the loops becomes yet more marked. The detached loop in the lower 
part of Fig. 23 shows this very clearly. That certain chromosome pairs 
frequently become detached from the spireme and undergo precocious con- 
densation was shown in the case of Oenothera (Gates, 1908 , PI. II, Figs. 20-24), 
but in that genus the chromosomes are so short and thick that there is 
never any twisting of the members of a pair about each other during the 
condensation period. Twisting of the chromosomes about each other, with 
subsequent breaking at the nodes, is therefore apparently excluded. Its 
absence in Oenothera may perhaps be genetically significant, since it is 
known that in that genus the characters in crosses usually have a strong 
tendency to remain together in groups instead of showing either free 
assortment or ‘ crossing over ’. 
Fig. 30 represents a later stage in lettuce, where all the loops have 
become free and the twisting has reached its climax. The loops clearly 
correspond in number to the ;tr number of chromosomes, and one is 
consequently led to the conclusion that each loop represents a pair of 
homologous chromosomes. The later fate of the loops, which will be 
described below, completely confirms this view. The only satisfactory 
interpretation of such an arrangement would appear to be that the spireme 
is made up of the full somatic number of chromosomes, which are arranged 
end to end in pairs, the members of which are alternately of maternal and 
paternal origin. The two members of each pair then bend round and form 
a loop, so that the parental chromosomes of each pair come eventually to 
lie side by side but still connected at one end. This is very similar to the 
course of events in Oenothera (Gates, 1908 , PI. II, Figs. 20-26). 
The sides of the loops in lettuce then twist around each other. Usually 
this results in two or three or more turns (Fig. 30), but sometimes the 
overlapping is only at one point giving a figure 8 (Fig. 31). Such figures 
of course are familiar enough in the cytological literature, having been 
