72 BOTANICAL GAZETTE (ory 
adjoining Venezuelan coast. Within these groups specific differentiation is 
observable, so that each island isolated by a too-fathom channel has its endemic 
species, the islands with a common coastal plain possessing little if at all differ- 
entiated forms. The almost entire absence of the genus from South America 
and the geographic grouping of species and superspecies in the West Indies indi- 
cate that Agave penetrated from the Central American mainland, where it centers, 
and overran the terrain before the disruption into islands, two or perhaps three 
parent stocks being involve 
‘The vascular plate and cotyledons of gymnosperms,” by JoHN M. CouLteR. 
—Among the various vascular structures of gymnosperms that have been used to 
si i Popeye hidsewinioceaaal changes, the vascular plate of the cotyledonary 
especially its connections with the cotyledons. 
Series of cycads, of camifers, and of other gymnosperms were shown to illustrate 
the following general tendencies: to reduce the cotyledons to two, to reduce the 
protoxylem poles of the vascular plate (and hence the root poles) to two, to elimi- 
nate certain vascular connections of the cotyledons, and to restrict the branching 
of strands within the cotyledons. 
Differentiation among chromosomes.—Crepis virens seems to afford promis- 
ing material for the solution of several difficult cytological problems. JuEL had 
already found the diploid and haploid numbers in C. tectorum to be 8 and 4. 
ROSENBERG? now finds the numbers in C. virens to be 6 and 3, the lowest numbers 
yet established for plants. The fact that the chromosomes are so few and that 
they are readily recognized as ‘‘prochromosomes” in the resting nucleus, removes 
any danger of uncertainty in counting which might be anticipated in case of large 
numbers. Not only is the number low, but the individual chromosomes are not 
alike, two being long, two rather short, and two intermediate. In the diploid 
- divisions these three kinds of chromosomes appear in pairs, the members of a 
given pair being alike. At synapsis a double thread appears, and there is 4 
fusion which is to be regarded as a fusion of whole chromosomes, reduction in 
number being brought about in this way. ROSENBERG suggests that interesting 
results might be obtained by crossing Crepis virens and C. tectorum, and he 
promises to make the attempt. 
Noting the low number of chromosomes in Crepis tectorum (8 and 4) and in 
C. virens (6 and 3), TAHARA’ examined the Japanese species, C. japonica, and 
found the numbers to be 16 and 8, just double the numbers in C. tectorum. The 
chromosomes were also found to be of different sizes and forms. If there is 4 
relation between specific characters and chromosomes, the genus Crepis would 
seem to the reviewer to be a favorable form for the investigation of this subject.— 
CHARLES J. CHAMBERLAIN. 
° ROsENBERG, QO., Zur Kenntniss von den Tetradenteilung der Compositen. 
Svensk. Bot. Tidskrift 3:64-77. pl. I. 1909. 
© TaHARA, M., Ueber die Zahl der Chromosomen yon Crepis japonica Benth. 
Bot. ars 24:23-28. pl. 2. 1910. 
