84 
PACIFIC SCIENCE, Vol. V, January, 1951 
Fig. 3. Receptacle of common form of flower of 
Cassia fistula. A, Diagram, front view, showing points 
of attachment of stamens (1-10) in relation to petals 
(a-e) and pistil. B, Side view, showing also points of 
attachment of sepals a', h', e'. 
genus Fistula., to which C. fistula belongs, has 
the three lower stamens (my Nos. 5, 7, 9) 
with long, bent filaments, and with anthers 
opening by longitudinal clefts; the seven up- 
per stamens with short filaments and with 
anthers opening by basal pores ; the anthers of 
the one to three uppermost stamens (my Nos. 
1, 2, 3) sometimes abortive. This description 
of the stamens of C. fistula or the subgenus 
Fistula is the description commonly found, 
the majority of opinions being that the three 
smallest stamens (my Nos. 1, 2, 3) were alike 
and without pollen (and, according to some 
authorities, indehiscent) ; that the arrange- 
ment was seven (or four) anthers with basal 
pores and the three on long filaments with 
longitudinal clefts. 
Examples of descriptions of the stamens of 
C. fistula follow: “Bentham and Hooker 
(1865: 572): "Stamina perfecta 10, filamentis 
3 inferioribus elongatis, antheris saepius ob- 
longis breviter 2-rimosis, 7 superioribus ab- 
breviatis antheris minute 2 -porosis." Trimen 
(1894: 103, 104): "3 lowest stam. with very 
long doubly curved exserted fil. and oblong 
anth. dehiscing longitudinally, 4 lat. ones 
with short straight fil. and versatile anth. 
opening by pores at the base, 3 uppermost 
ones much smaller, erect, with indehiscent 
abortive anth." Brandis (1907: 253): "The 
three lower stamens have long curved fila- 
ments and anthers, which open by longi- 
tudinal slits, four stamens have short filaments 
and anthers opening by basal pores; the rest 
have minute anthers without pollen." Rock 
(1920: 75): "Three stamens with long fila- 
ments, 7 stamens with short filaments, anthers 
of three basifixed opening in longitudinal 
slits, anthers of seven dorsifixed opening in 
basilar pores." 
As seen by my diagrams and descriptions, 
the arrangement is of five stamens of each of 
two kinds in two series, one series opposite 
the petals, the other opposite the sepals. I 
have examined the pollen of the "abortive" 
stamens (Nos. 1, 2, 3) and found apparently 
normal grains, some slightly smaller than 
some of those of No. 8, for example, which, 
however, were turgid with starch. Some of the 
grains of Nos. 1, 2, 3, also, were slightly 
smaller than those of Nos. 5, 7, 9, even after 
the starch contents of Nos. 5, 7, 9 had evi- 
dently been converted in mature flowers from 
starch by diastase (Tischler, 1910: 241). 
Grains of Nos. 1, 2, 3, from flowers at dif- 
ferent stages of development, were tested for 
starch and dextrin (with iodine dissolved in 
potassium iodide) and were found to contain 
varying amounts of both. Thus it seems 
probable that the small stamens do function — 
Nos. 1 and 3 like Nos. 5, 7, 9, No. 2 like 
Nos. 4, 6, 8, 10. 
Tischler (1910: 220) agreed with Knuth 
(1904: 375-377) in Knuth’s description of 
the stamens of C. fistula, which is translated 
as follows (Knuth had studied the flower 
ecology of C. fistula in Buitenzorg) : 
Cassia fistula has three abortive stamens; 
these have anthers to 2 mm. long, which are 
