[Vol. 8 
108 ANNALS OF THE MISSOURI BOTANICAL GARDEN 
first of these is described in the diagnoses as “straight or simply 
curved upwards" and is to be seen in such species as L. auriculata, 
L. Engelmannii, L. arctica, L. globosa, and L. Fendleri. The 
second type is that known as “recurved.” This again is a 
simple curve and differs from the first in direction and usually 
in degree: L. lasiocarpa, L. purpurea, L. recurvata, and L. argen- 
tea serve as illustrations. The third type obtains to some ex- 
tent in the great majority of species. The curve in this case is 
à compound one resembling the letter S reversed and lying on 
its side, and consequently has been termed “sigmoid” or “‘S- 
shaped.” Species characterized by the third type of pedicel 
usually have erect pods since the final or distal curve is directed 
upwards. In general, these three types are easily recognizable 
but occasionally forms of the sigmoid group are observed in 
which the final curve has become so nearly obsolete that the 
pedicel is practically recurved. However, in such cases there 
is usually at least a suggestion of the final curve in part of the 
inflorescence and this is enough to indicate the normal condition, 
since no such tendency is found in any of the species with con- 
stantly recurved pedicels. Not a great deal of stress is placed 
upon the form taken by the pedicel in indicating actual relation- 
ships. The first type may be considered primitive since it is 
characteristic of the family in general and of most of the more 
primitive species in the genus. The recurved pedicel has prob- 
ably been developed independently at least three times within 
the genus, though there is no indication that it has arisen from 
any except the first type, and from that it is, after all, but a 
minor change. There is, of course, the possibility that species 
with recurved pedicels included in sigmoid groups have passed 
through the stage characteristic of those groups and have sub- 
sequently lost the final upward curve—L. purpurea and L. 
pueblensis may have had this history. The sigmoid condition is 
surely a mark of greater specialization than either of the other 
two, although the change is relatively a simple one. It has 
probably been derived entirely from the first type and seems 
to have developed independently at least three and possibly more 
times. It is noted in such diverse species as L. Schaueriana of 
the section Enantiocarpa and L. argyraea, L. alpina, L. montana 
and L. Kingii of Eulesquerella. 
