[Vot. 8 
112 ANNALS OF THE MISSOURI BOTANICAL GARDEN 
allel lines ever produced. The septum is usually entire, but in 
a number of more highly specialized species is frequently per- 
forate. This character is so variable between species that it 
seems impossible of use in their de- 
qs limitation. No attempt has been 
made to draw phylogenetie con- 
clusions from the characters of the 
septum, since no definite progres- 
sion from one species to another 
has been discovered (fig. 2). 
The Number of Ovules.—Reduc- 
tion in the number of ovules with- 
in the ovary is so constantly coin- 
cident with specialization among 
dicotyledons in general and in the 
Cruciferae in particular that be- 
fore examining the species of a 
given genus one may confidently 
expect, if any intraspecific varia- 
tion in the number of ovules oc- 
Js 2. Characteristic M fireann n curs, to obtain evidence as to 
what are the more primitive and 
what the more recent forms. In the present genus the num- 
ber of ovules varies between two and sixteen in each cell. 
While in very few species is the number constant, it varies for a 
given species within certain limits. It must not be forgotten 
that the data obtained for the present study were from her- 
barium material only, and as this was often rather scanty the 
maxima and minima given are not necessarily exact. It is be- 
lieved, however, that the relative position of but few species in 
regard to the number of ovules would be materially changed by 
increasing the known range of variation. Of particular sig- 
nificance is it to compare pairs of species that are very closely 
related and of which one has probably been derived from the 
other. L. argyraea and L. Berlandieri furnish one interesting 
case. The former has glabrous pods and the latter stellate- 
pubescent ones. The former has eight to sixteen ovules and the 
rN 
L.Lescurii L. densiflora 
L. Fendleri L.recurvata 
