A critical study of the cytology of Crepis virens. 
117 
da durch die Zellwand hindurch und in die benachbarte Zelle hinein- 
gepreßt wird” (p. 68). 
There is a typical extrusion of bodies in Crepis taraxacifolia, and 
these pass into the adjacent cells as in Galtonia candicans (8), a phenom- 
enon which in Galtonia cannot be ascribed to faulty fixation. 
To eliminate misunderstanding, it seems advisable to define the pre- 
cise meaning of the terms ‘univalent and bivalent’, which will constantly 
be used in the succeeding description of the meiotic phase. A ’univalent’ 
Segment of spireme signifies that which will eventually constitute a ‘uni- 
valent’ chromosome; a ‘univalent’ chromosome being characteristic both 
of somatic and of homotype divisions. A ‘bivalent’ Segment of spireme 
is of a duplex character, and represents the association of two ‘univalent’ 
lengths of spireme, in preparation for the formation of the ‘bivalent’ chro- 
mosome ; the bivalent chromosome, characteristic of the heterotype division 
alone, represents a combination of two ‘univalent’ chromosomes. Further, 
the signtficance of the two forms of fission must be borne m mind. Fission 
in the substance of the ‘univalent’ spireme, or chromosome, divides it 
into two daughtcr halves preparatory for the next vegetative or homo- 
type mitosis ; whilst the so-called ‘fission’ in the substance of the bivalent 
spireme disjoins the two univalent Segments or chromosomes, in prepara- 
tion for their dissociation on the equatorial plate of the heterotype division. 
It is proposed to confine the term fission exclusively to the split in the 
substance of the univalent spireme and of the univalent chromosome, and 
to use other descriptive words for the disjunction of the univalent Seg- 
ments of the bivalent combination. 
Passing out of synapsis, 
Whcreas during the presynaptic phases the split univalent Segments 
concentrate for the evolution of the future univalent chromosomes, so 
during synapsis, as revealed by the course of subsequent events, there is 
a further sorting out, and also an approximation in pairs, of these uni- 
valent Segments preparatory for the association of two univalent chromo- 
somes for the heterotype combination. (In Primula this association 
takes place during the hoUow sph-eme stage.) The Strands and loops, 
as they come out of synapsis, are therefore of the bivalent Order, though 
they Show all stages in the Separation of their component univalent Seg- 
ments. This, combined with their viscous character, gives them a varied 
appearance. The clear line in the substance of the bivalent spireme 
indicates the Separation of the univalent spiremes, and must not be con- 
fused with true fission which for the time is invisible, the individual 
