170 THE CORONA OF NARCISSUS. 
That these appendages are sometimes present in all the floral organs 
of plants seems to me clear from Figs. 1, 2, 3, and 4, Plate XLVIII. The 
leaf-stipules of Trifolium incarnatum (Fig. 1) are almost repeated, with 
the exception of colour, in the petals of Silene maritima (Fig. 2) ; there is 
such a close resemblance in form and position in both objects, that it 
is impossible to doubt their being identical in character. A slightly 
modified form of stipule exists in the stamen of Ornithogalum nutans 
(Fig. 3) ; and they are very distinct and most characteristic in the 
stigma of Iris Pseudacorus (Vig. 4). If. reference is made to Fig. 2, 
and if the whole of the petal-stipules of the complete flower are ima- 
gined to be connate, we have a corona precisely resembling Narcissus. 
The true explanation of the corona in the small section of the Order, 
I believe, consists in the recognition of a series of confluent petal- 
stipules, leaving the normal 6 stamens and 6 petals as in the rest of 
the Amaryllidacee. That there is nothing improbable in confluent 
stipules, I give examples of them in all the floral organs; Fig. 5 is 
an example of confluent leaf-stipules in Greffia calyculata, figured in 
Seemann’s ‘ Flora Vitiensis, plate vi. (a somewhat analogous growth 
may be seen in many of the Euphorbiacee). Fig. 6 is one-half a 
flower of Narcissus Pseudo-narcissus, shown with half the corona, i. e. 
half the series of confluent petal-stipules. If this figure be compared 
with Fig. 2, where the stipules are disconnected, it will be better 
understood. Fig. 7 shows the stamens of Lobelia Dortmanni, confluent 
near the anthers, but free below; if we imagine the two appendages 
of Fig. 3 to be connate, we would have such a growth as is here 
represented, with the filaments disconnected below. Fig. 8 gives an 
example of confluent pistil-stipules in Sarracenia purpurea, and may 
be compared with the stigma of Jris in Fig. 4, where, if we imagine 
the appendages to be connected, we have an object similar to that 
represented in Fig. 8. 
In Paneratium lilyricum (Fig. 10 4) the confluent stipules are at- 
tached to the filaments, instead of to the corresponding parts of the 
perianth segments, as in Ornithogalum nutans, only that they are connate. 
Dr. Masters, in Journ. Bot. Vol. III. p.107, endeavours to show that 
the corona probably “ consists intrinsically of two rows of stamens.” 
He says, “in the species with lobed cups three of the lobes are opposite 
to the sepals and alternate with the petals (a. A. Fig. 9 a), and these 
three in wstivation decidedly overlap the three inner lobes which are 
