392 
AMERICAN JOURNAL OF BOTANY 
[Vol. 8 
become consolidated except in their distal parts. Stronger evidence for 
the same conclusion is described for U. americana on page 390. 
The vascular supply to the flower duplicates that of U.fulva (PI. XVII, 
figs. I, 2) and of U. scabra (PI. XVII, figs. 8-10). Commonly one and some- 
times two perianth lobes have the stamen suppressed even in the vascular 
supply. 
Ulmus scabra Mill, is described as having the perianth lobes and stamens 
ranging from 6 to 5 ; a study of many flowers reveals that 6 perianth lobes 
and 6 stamens appear in the majority of cases. The vascular supply to the 
perianth and to the stamens arises separately (PI. XVII, fig. 8, b) as it does 
in U. Julva and in U. campestris. The strands to the perianth lobes (PI. 
XVII, fig. 9, w . . . m^) are well out in the "cortical" regions when the 
supply to the stamens is just passing out of the stele (fig. 2, n . . . n^). 
U. scabra presents a feature not found in any of the other species. 
Alternating with the traces to the perianth lobes and arising with them are 
bundles that apparently lag behind (PI. XVII, figs. 9, 10, x, x\ x^; fig. 8, 
X, x^; PI. XVIII, figs. I, 2, X, x\ x^). These are perianth bundles but 
they always continue inside the strands to the perianth lobes. They do not 
have the number of lignified cells, nor the size of cells that the strands to the 
perianth lobes possess. They weaken rapidly and vanish on a level with 
the origin of the lobes of the perianth. Such bundles were not found in the 
posterior part of the flower (PI. XVII, fig. 9). The anterior part of the 
flower of Ulmus is clearly the conservative part of the flower, since in this 
part the stamens opposite the perianth lobes are always present. Sup- 
pressed stamens occur in the posterior part of the flower, or here the stamen 
is present and the perianth lobe is suppressed. Also, the abortive bundles 
described in the above named species usually occur in the anterior part of 
the flower except in U. fulva where they make one complete whorl (PI. 
XVII, fig. 2, d), but the abortive stamen, s^, in the same figure, is in the 
posterior part of the flower. Thus, as described above, U. scabra presents 
an additional feature in the anterior part of the flower, namely, the weak 
bundles to the perianth lobes. These bundles, x, x\ x^, may be vestigial 
parts of suppressed perianth parts which alternated with the present 
perianth lobes. The origin, the position, and the appearance of these weak 
bundles offer no other disposition except that of a vascular supply to corolla 
parts which have been reduced and consolidated in the gamophyllous 
perianth. 
Alternating with the strands of the perianth, except with the posterior 
strand, are organized tissue regions as in the other species suggesting bundles, 
but these show no lignification (PI. XVII, fig. 9, d). The same condition 
has been fully discussed above in the other species. Again, there seems to 
be no other alternative here than to look upon these as abortive bundles to 
a suppressed outer whorl of stamens. 
