Flowers of Cassia fistula — -Neal 
87 
The possibility that 2^4-0 was the cause 
has been suggested. For 2 or 3' years before 
this study was made, trees A and B were ex- 
posed to 2,4-D, which was applied in solu- 
tion from a watering pot on the ground near 
their trunks in order to rid that part of the 
lawn of pennywort and other weeds. This 
treatment did not seem to affect tree A. In 
July, 1949, a year later, and without another 
application of 2,4-D, tree B again bore aber- 
rant flower forms. The effect of this hormone- 
like substance on herbs is temporary and com- 
paratively short (Zimmerman, 1943) ; but the 
effect appears later and lasts longer on woody 
plants. Applications of 2,4-D to weeds along 
a fence of living hau {Hibiscus tiliaceus L.), 
at Kailua, Oahu, caused malformed leaves on 
the hau more than 6 months later flowers, 
being absent, were not studied. According 
to Zimmerman (1943), modified organs may 
well be caused by an environment (or chemi- 
cals) which permits different potentialities of 
the protoplasm to develop, inasmuch as "the 
so-called 'normal’ characters of a plant are 
but a partial expression of the range of possi- 
bilities of which the protoplasm is capable." 
Although I had not examined the flowers 
of tree B in detail until June, 1948, for years 
before then I had noticed the general ap- 
pearance of the tree with its small, rounded 
racemes, and I believe it always has been 
different. In fact, it may be a sport. If so, this 
would also explain the comparatively small 
number of pods, which is characteristic of a 
sport and is not usually caused by 2,4-D 
(van Overbeek, 1946). Nonetheless it is possi- 
ble that fasciation found on tree B in July, 
1949, might have been caused by the appli- 
cation of 2,4-D the year before. The tree bore 
many more flowers in 1949 than in 1948, and 
among them were found four racemes with 
conspicuously flattened peduncles and ra- 
chises. No fasciation was noticed in 1948. 
The striking difference in tree B is in the 
strange flowers. These vary in the same ra- 
^Personal communication from A. J. Mangelsdorf, 
1949. 
ceme, and they follow several patterns. One 
common form of flower on tree B had four 
petals, four sepals, eight stamens (Fig. 5, ^). 
Another form had five petals, three of which 
were more or less irregular, four or five sepals, 
and eight to 10 stamens (Fig. 5, B, C, D). A 
third form had three petals, four or five sepals, 
and seven or eight stamens (Fig. 5, E, F). 
The pistils were not noticed to differ from 
those of the common form. 
The sepals were not all equal in size and 
some seemed to function as petals, one being 
large, rounded, and yellow (instead of green) 
where a petal was missing (Fig. 5, £, F). One 
or both of the innermost sepals (Fig. 2, D, b\ 
e') seemed often to be petal-like. Although 
flowers on some of the 25 other trees ex- 
amined were often found to have one or both 
of the inner sepals more or less petal-like, 
none reached the size of those on tree B. 
The petals varied greatly in size and shape, 
from round to narrow and more or less 
curved. Some petals combined with filaments 
or substituted for them and bore anthers. 
The stamens were very variable, both as to 
filaments and anthers. In extreme examples 
two anthers might combine, those studied 
combining an odd and even type anther, such 
as Nos. 3 and 4, with No. 3 at the tip (Fig. 
6, G). The presence of an anther with either 
one or two cells on the edge of a petal was 
common. The anther was always found to be 
like the anther of stamens Nos. 1, 3, 5,' 7, or 
9 (Fig. 4, E-H), that is, the outer series of 
stamens opposite the sepals; and its pollen 
grains contained starch. Nos. 1 and 3 seemed 
more commonly malformed than 5, 7, or 9. 
The calloused edge of the petal from its base 
to the anther suggested a reduced filament. 
A calloused edge might end bluntly with a 
vestigial anther or none. A filament might be 
edged or tipped with yellow petal material. 
Some of these aberrations are illustrated 
(Fig. 6, A-G ) . The stamens and petals and the 
petals and sepals were partly interfunctional. 
But the stamens and sepals were not. This 
closer relationship between petals and sta- 
