158 REV. PROF. G. HENSLOW ON THE ORIGIN OF 
S arrangement. Thus, taking Rumer, we have, supplying the corolla, 3+3 . [34 3]. 
3--3.8—21; while Fagopyrum, like Polygonum, shows the attempt to combine threes 
with fives. Thus, P. Persicaria has 5 . [5] . 3+5 . 3 * —21 with 8 circles. This may 
be regarded as partially descending from 38; to $ and 4. Flowers of P. Hydropiper 
examined were always 5.0.5.3. 
In some cases a supernumerary whorl of stamens has been introduced. Thus Rheum 
has 3+3.{8+3].3+8+3.3. But the number of stamens may vary from 3 and 
6 to 9. 
I have elsewhere f described how the stamens arise in Rheum by tracing their fibro- 
vascular cords :—“ Three sepaline cords give rise to the three staminal (7. e. in front of 
the sepals), while six others arise in pairs. The remainder (ż. e. three in number) are 
formed from the sepaline which have not given rise to the cords of stamens; these 
supply the carpellary cords.” Hence it will be seen that the “ pairs ” arise from the axis 
where the petals ought to be, the cords of which have supplied those of the pairs of 
stamens instead. 
One genus of this order carries us back to what was possibly a more primitive 
condition of a long spiral. Symmeria, as given by Bentham and Hooker, has (2) 3+3 . 
[3+3] . 20-50 . 0; (2) 323-3 .[3--3]. 0. 3. Here the corolla only is supplied, while 
the 20-50 may mean whorls of fives; or jf more, perhaps they might accurately be 21 
to 55; i. e. cycles of 38; and $$. 
Lythrarieze supply some interesting corroborations. In the genera of this order the 
calyx is gamosepalous $. This necessitates the two whorls appearing as one, composed 
of six parts. The corolla follows suit, but the stamens reveal themselves as multiples of 
three. Thus the formula of Lagerstremia is, 6 .6 . c0 . 3, that of Lythrum being, 6; 
6. 3+3, 3+3, 384-3 . 2. 
In some species of Lythrum symmetrical reduction has taken place, and we find 4; 
4.4.44-4.2. The first number represents the teeth-like appendages. 
The central flower of the 3-flowered cyme of L. Salicaria may be 7- or 6-merous ; when 
this occurs, the two lateral flowers are 6- or 5-merous, respectively. 
Peplis, 6; 6.6.3 -3.2, or else, but rarely, 5; 5.5.5.2. 
The angular distance between any two consecutive parts in the $ arrangement is 144, 
and between any two in a whorl of three is 120°. This, as observed, offers no difficulty 
in nature; such changes are frequent. Thus, on vertical shoots of the common Laurel, 
the leaves are arranged on the 2 plan; but on the lateral horizontally growing branches 
the leaves are}. It is the same in the Ivy, the free-growing leaves are $, but those on 
* The position of the three extra stamens is somewhat anomalous, The sepals are quincuncial, or 2. The petals, 
which should alternate with the sepals, are wanting; then the apparently outer whorl of five stamens is also alternate 
with the sepals; but the three inner stamens are opposite sepals. I would suggest as an explanation that these three 
are the representatives of the outer whorl; but, as in Geranium, they have become dislocated in consequence of the 
presence of glands, so that they may either represent a whorl of five, or else be a single cycle of j divergence. M 
the latter, then the whole number would make a single cycle of 5. 
+ “The Vascular Systems of Floral Organs,” Journ. Linn. Soe., Bot. vol. xxviii. (1889) p. 151. 
t The tube is probably of a receptacular nature. 
