THE EXPERIMENTAL STUDY OF GENETIC RELATIONSHIPS I4I 
are frequently characterized by different chromosome numbers; and, 
exceptionally, within a species, well-characterized nominal varieties 
may differ in this respect. Lists of genera in which there is variation 
in chromosome number have been published by Gates and by Strass- 
burger. Among the genera included in these lists, or concerning which 
information has been more recently published, we find the flowering 
plants Alchemilla, Antennaria, Crepis, Dahlia, Daphne, Drosera, 
Funckia, Hieracium, Humulus, Houttuynia, Musa, Oenothera, Pri- 
mula, Rosa, Rumex, Spiranthes, Taraxacum, Thalictrum, and Wik- 
stroemia, the ferns Athyrium, Lastrea and Nephrodium, the mosses 
Amblystegium, Bryum, Mnium and Phascum. The numbers range 
from 3 (x) and 6 (2x) in Canna to 48 (x) and 96 (2x) in Castalea. The 
processes of species formation must have been frequently attended by 
changes in the number of chromosomes. We cannot escape this con- 
clusion when we take into consideration the two striking facts that 
within a species the number is constant, but that from species to 
species and from genus to genus it shows the greatest diversity. 
Recent discoveries are making it very clear that mutative changes 
in the chromosome number occur frequently, and that such changes are 
always associated with a modification in the morphological characters 
of the plant. In other words, certain mutations are probably de- 
pendent upon, or, at any rate, closely associated with, visible changes 
in the nuclear mechanism. We have every reason to believe, therefore, 
that the different chromosome numbers of different species were ac- 
quired simultaneously with the acquisition of other specific characters. 
Among the mutations of Oenothera Lamarckiana there is just one 
which de Vries regards as unquestionably progressive. It is the re- 
markable Oenothera gigas, which differs from its parent not only in 
numerous external characters but also in having twice as many chro- 
mosomes. The gametophytic and sporophytic numbers are 14 and 
28 in Oe. gigas, 7 and 14 in Oe. Lamarckiana. Gates has shown that 
the double chromosome number of the former is correlated with a 
larger cell size in all corresponding tissues and that many of the gross 
characters of the plant are in turn dependent upon the difference in 
the cells. Hybrids between Oe. gigas and Oe. Lamarckiana show no 
semblance of Mendelian inheritance. In general they are very sterile, 
but in one case a fertile strain was obtained which remained constant 
through five generations. The hybrids are intermediate whichever 
way the cross is made, and, if back crossed with either parent, the 
