Vegetative Anatomy — Carlquist 
209 
Vitchia. The distribution of secretory canals in 
vegetative parts of Duhautia, Ar^roxiphmm, 
and Wilkesia follows very closely the diagram 
given by Carlquist (1957c/ 58) for F. speciosa, 
differing only in those species which do not 
characteristically have secretory canals in pith, 
cortex, or leaves. The occurrence of secretory 
canals in inflorescence structures of Dubautia, 
Ar^roxiphiim , and Wilkesia will be consid- 
ered in subsequent studies. 
Trichomes: Summary 
Although occurrence and types of tri- 
chomes have been mentioned above, the basic 
pattern underlying trichome occurrence needs 
to be stated. In the genera Duhautia, Arg^ro- 
xiphium, and Wilkesia, two types of trichomes 
are present: uniseriate nonglandular and bi- 
seriate glandular. Examination of shoot apices 
indicates that these two types may be formed 
initially, but are variously matured or de- 
generate upon maturation of organs of the 
plant. In Dubautia railliardioides, for example, 
both uniseriate hairs and biseriate glandular 
trichomes with subdivided multicellular tips 
may be seen on young leaf primordia. On 
mature portions of the plant, both types of 
hairs are absent except for retention of uni- 
seriate hairs in leaf axils. Thus, differences in 
indument of the mature plant are the result of 
elimination or retention of these two basic 
types. The leaves of D. platyphylla furnish an 
example of marked development of the bi- 
seriate hairs with subdivided pits, whereas 
leaves and stem of D. waialealae demonstrate 
conspicuously the uniseriate nonglandular 
hairs with no evidence, in the mature plant, 
of the biseriate type. 
The precise nature and distribution of tri- 
chomes in Compositae is inadequately known. 
The occurrence of the capitate glandular hairs 
mentioned above, however, appears signifi- 
cant in connection with relations of the three 
genera to certain non-Hawaiian groups. 
DISCUSSION 
The importance of the data developed lies 
in the application of anatomical characters to 
taxonomy of the genera discussed, descrip- 
tion of relationships among the three genera, 
and in determination of the relationships of 
these seemingly isolated genera. Within the 
genus Dubautia, characters of vegetative anat- 
omy seem to offer good criteria for taxonomic 
decisions, although inflorescence anatomy 
cannot be overlooked. 
The results of the present study, and one 
reported separately (Carlquist, 1958) indicate 
that Keck (1936: 24) is justified in uniting 
the species of Railliardia with Dubautia. No 
anatomical character, or combination of such 
characters, seems to support the division of 
the species into the two genera as they have 
been traditionally delineated. For example, D. 
waialealae (invariably treated as a Dubautia) 
seems closer in node and leaf anatomy to D. 
Menziesii (included under Railliardia by au- 
thors recognizing that genus) than to Du- 
bautia plantaginea (type species of the genus) 
or its close relatives (D. microcephala, D. 
railliardioides ) . 
Rather, the writer believes that a picture of 
species groups emerges, although these can- 
not yet be defined with sufficient precision to 
permit division of the genus into subgenera 
on the basis of the characters described here. 
Attention may be called, however, to certain 
alliances which seem quite certain: 
(1) D. Menziesii, D. arborea, D. Hillebrandii, 
D. Montana, and D. struthioloides agree in 
leaf, pith, and wood anatomy. 
(2) The species with narrow strap-shaped 
leaves and wide leaf bases, D. plantaginea, 
D. microcephala, D. railliardioides, D. pale- 
ata, D. magnifolia, and D. ternifolia, seem 
close in leaf anatomy, presence of secre- 
tory canals in pith, and, to a certain extent, 
in nodal anatomy. D. Knudsenii appears 
close to this assemblage on the basis of its 
foliar secretory canals and its pith structure. 
(3) A grouping of species including D. cilio- 
lata, D. lonchophylla, D. scabra, D. Sherffi- 
ana, and D. thyrsiflora seems justifiable on 
the basis of near identity in leaf structure. 
