39 
NATURE 
ee 
[Vov. 9, 1882 
north by anyone else. 
E. longitude. 
Quitta is situated 5° N. latitude, and 1° 
WALTER HIGGINSON 
B. MANNING 
Quitta, West Coast Africa, September 25 
Two Kinds of Stamens with Different Functions in the 
same Flower 
Ir may be worth mentioning that cases strongly analogous to 
those described in NATURE (vol. xxiv. p. 307, and vol. xxvi. 
p- 386, are also to be observed among the Monocotyledons in 
the family ef Commelynacez, ani that these cases offer some 
graduations. 
In Zyradescantia virginica, L., the flowers, as is generally 
known, are turned upwards and quite regular, the leafy organs 
of each whorl (3 sepals, 3 petals, 3 outer, 3 inner stamens, 3 united 
carpels) being alike and equal in size, As Delpino has clearly 
shown (U/teriori osservazioni, parte il. fascic. 2, p. 297) these 
flowers are adapted to Apidz, which in order to collect pollen 
take hold of the articulated hairs of the filaments, In some 
other species here to be considered the adaptation to pollen- 
collecting bees has remained, but the flowers have turned late- 
rally, and thus not only has their form become irregular (bi-laterally 
symmetrical or zygomorphous), but also the function of the 
stamens has gradually changed. 
In ZYinnantia undata, Schlecht. (Fig. 1), sepals and petals 
are still almost unaltered in form and size, only stamens and 
pistil have become markedly irregular. The broad roundi-h 
petals, which are light purple, spread in a perpendicular plane. 
The 3 upper stamens, with shorter filaments projectiig from 
Fic. 1. 
Fic. 2. 
Fic. 1.—Front view of the flower of Timnantia undata, Schlecht. Fic. 2.— 
Front view of the androeceum and gynecium of Commelyna celestis, 
Willd. s,s, s’, sepals; ~, Z, 2 petals; a, a, a’, outer whorl of anthers; 
a, a, a’, inner whorl of anthers, or ovary ; g7, style (“‘Gr-fiel’’); s¢, 
stigma. 
the middle of the flower, are highly conspicuous by a diverging 
tuft of bright yellow articulated hairs, which on the last third of 
the light-purple filaments surround the golden yellow anthers 
like a cone of golden rays. At the tips of these filaments golden 
yellow pollen-grains are presented by the whole front side of 
the three upper anthers. 
The three lower stamens are much longer, directed obliquely 
downwards and forwards, with only their tips bending upwards, 
a little overtopped by the pistil, which has the same direction 
and incurvation. These parts, like the same parts in the de- 
scribed Melastomacez, will hardly be perceived by anadvancing in- 
sect, ‘‘ owing to their projection against the broad-petalled corolla 
of the same colour in the background,” fornot only the style and the 
filaments, but also the hairs on the base on the two Jateral lower 
filaments are of the same purple colour as the petals, and even 
the bluish lower anthers with their yellowish pollen are but 
feebly conspicuous. Any one of the Apidz or Syrphidz of 
suitable size, however, when making for the upper yellow 
stamens in order to collect their pollen (I have only once ob- 
served the honey-bee doing so), will involuntarily repose on the 
projecting part-, and at first bring the stigma and then the two 
lateral of the lower anthers into contact with the under-side of 
its abdomen, and thus regularly effect cross fertilisation. 
Here, then, as in Heeria, &c., the anthers have differentiated 
into upper ones, which attract insects and afford food to them, 
and lower ones which attach their pollen to the visitors, and 
cause it to be transported by them to the stigma of the next 
visited flowers. Also differentiation in the pollen of the two 
kinds of anthers in our Tinnantia has begun to take place, but 
contrary to Melastoma, the pollen-grains of the short stamens 
here are smaller than those of the longer ones. I measured 
numerous pollen-grains of two individuals in a moistened state 
(where they are of elliptical form), and found in the one stem 
the pollen- grains of the short stamens (in I-1000 m.m.) 62-75 
long, 31-38 broad, those of the longer ones 68-94 long, 38-44 
broad ; in the other stem, those of the short stamens 53-69 
long, 28-37 broad ; those of the longer ones 59-78 long, 31-40 
broad. Both kinds of pollen proved to be quite fertile. 
Commelyna coelestis, Willd. (Fig. 2) possesses in general the 
same contrivances for cross-fertilisation, but has gone a step 
further in differentiation. Its upper sepal is plainly smaller, its 
lower petal plainly larger than the two other ones; its upper 
anthers (a, a’ a) have differentiated -in themselves ; two small 
lateral portions of each of them (go) produce a little pollen and 
four cross-like diverging flaps (j7), which are much larger, 
actract insects by their bright yellow colour strikingly contrasting 
with the azure corolla, and perhaps at the same time serve as 
food to the visitors. The articulated hairs of the filaments thus 
having lost not only their original function (which they have in 
all stamens of Tradescantia) as supports for the feet of pollen- 
collecting bees, but also their secondary function (which they 
have in the upper stamens of Tinnantia) of attracting insects, have 
disappeared altogether. The middlemost of the lower anthers, 
which in Tinnantia is nearly useless from its position behind the 
style here, has erected and become much larger than the two 
lateral ones, so as to be eminently useful. 
The pollen-production of the upper anthers appears to be 
vanishing, not only from the diminution of the quantity of pro- 
duced pollen, but also from the great variability of the size of 
the pollen grains. For whilst the pollen grains of the two lateral 
lower anthers only differ in length from 75 to 90, in breadth 
from 45 to 68, and those of the middlemost lower anther in 
length from 56 to 82, in breadth from 37 to 56, those of the 
three upper anthers fluctuate from 50 to 87 length, and from 
31 to 56 breadth. 
In Cummelyna communis, differentiation has gone still further ; 
the upper sepal and the lower petal are relatively very small ; 
the upper filaments, like the upper petals, are blue-coloured ; 
the lower filaments, like the pistil and the lower petal, are 
colourless. The upper anthers, as far as I have seen (without 
microscope) produce no nore pollen 
The examinati »n of other species aad genera of Commelynacezx 
probably would show a longer scale of gradations. 
Lippstadt, October 25 HERMANN MULLER 
A Curious Halo 
THERE appeared in NATURE, vol. xxvi. pp. 268, 293, two 
articles headed ‘‘ A Curious Halo,” which reminded me of an 
analogous and still more curious phenomenon occurring some- 
times here in China, during the hot season. I beg to hand you 
a few lines on that subject, from the Monthly Bulletin of 
the Zi-ka-wei Observatory for August, 1877 :— 
“*A phenomenon to which I wish to call the attention of 
meteorologists was observed many times duriog that month 
(August), as also in July. It does not seem to take place in 
Europe, and I am inclined to think that it cannot occur except 
with an atmosphere over-charged with aqueous vapour, as it 
is the case with us in July and August. In the evening, just 
after sunset, or in the morning even long before sunrise, no 
matter what the direction of the wind and the barometric 
pressure may be, provided the day or night were very warm, 
éands of a tint varying from the faintest 1o the deepest d/ue are 
seen to appear upon the whitish or roseate vault of heaven. 
They usually are first seen in the east at evening and in the west 
at morning time, seemingly radiating from a common centre 
diametiically opposite the sun’s position. At other times they 
emerge from the very position of the sun, or from both points at 
_once, the interval being either free from bands or compleiely 
encircled by them. 
“Last year (1876), on the morning of September 4, I en- 
joyed a most interesting sight. It was about 5 a.m., the moon, 
then on her nineteenth day, was above the western horizon, and 
ju-t being partially eclipsed ; now from ber bright disc, as from 
a radiating ceutre, shot out a number of those bands or blue 
beams; they traversed the whole expanse of the sky, and 
seemed to converge towards a point whose situation in the east 
