394 BOTANICAL GAZETTE [NOVEMBER 
nearly correct idea of the true state of the case may be realized 
if the reader will bear in mind the two parent species and then 
imagine every possible form of gradation and combination pro- 
ducing races capable of self-propagation. — If left to themselves 
without further contamination from their surroundings many 
would develop into good species. Var. vestita (S. Missouriensis 
Bebb) may thus be accounted for, which will be noticed later. 
The modification of characters generally proceeds along the 
whole line with more or less equal pace, though often strong and 
weak modifications of one parent will be found indiscriminately 
combined with strong or weak modifications of the other, on the 
principle stated by Darwin ‘‘of augmentation, obliteration and 
reversion of characters when well defined varieties are crossed.” 
When the departure from pure S. sericea is not too great, the 
increasing size of the stipules, the degree of discoloration i0 
drying, the basal variation from the acute towards the cordate, 
the amount of silvery pubescence on the leaves are tolerably 
fair measurements of the distance. Beyond. this the problem 
becomes more complicated. Taking our point of view from S. 
sericea, the important change first occurring is in the capsule. It 
loses its silvery pubescence, enlarges, lengthens in style aa i 
pedicel, assumes a somewhat flattish subrhomboidal base sup” 
porting a long or short beak with often a tendency to wih 
ripens a pale tawny color, and has stigmas less deeply notches, 
often remaining entire. : 
The next important character early lost is the short, silvery 
hairy coating of the under surface of the leaves. Instead, fl 
leaves become more or less glaucous beneath, or the short hat 
are transformed into the longer softer hairs seen 0M the ee 
leaves of S. cordata. Var. vestita affords an extreme meee 
of this process of change. No other character has ree pa 
demanded so much study as the one of pubescence or hairin is 
Extremely variable in amount as to leaf, bud, shoot pee 
absent altogether, variable also on scale and rachis, it 18 9° 
character of much value. 
cale 
The third important change is in the scales. 126 = 
