‘ 
NARCISSUS POETICUS AND ITS ALLIES © 19 
and pedicel within the “ng are pgs ing in some forms 
than in others, and vary also in compres nd fineness of 
striation ; the spathe show gradwtitl of fetish and thickness, 
and in one form it is inflated. The per ee differs very con- 
have examined, e cept at recent hybrids, to exhibit regularly 
one of two fore; cehioli's e termed in this paper “ unequal”’ and 
‘subequal.’ In the first of these, the anthers of the three shorter 
stamens are included in the perianth-tube and their tips only 
slightly exceed the anther-bases of the longer exserted stamens ; 
in the second, all of the sii are more or less exserted, and 
those of the three shorter stamens do not greatly fall short*of the 
others. No good distinguishing features have been observed in 
fruit, mecha is not readily produced in the garden, and was 
ignored by Haworth in his descriptions, may be developed in 
almost every sialiivated form by keeping fertilized flowers in water 
after the fading of the perianth. The shape of the capsules thus 
obtained appears to be constant in each form, allowing for the 
d t o 
every wild form, but such would appear to be probable. I have 
not succeeded in detecting any appreciable differences in the seeds. 
lt results from these conclusions that worm Poet’s Narcissi 
the bulb, foliage, scape, perianth, corona, stamens and fruit may 
all exhibit characters useful for the deter faites of species. 
Modern authors have very generally laid the greatest stress on 
the shape of the perianth-segments, which is usually visible in 
herbaria, cond as this is undoubtedly very variable in certain wild 
forms, it seems doubtful whether it is of equal importance with 
the shape of t the corona, the arrangement of the stamens and the 
form . fruit—characters which, though obscured in dried 
specimens, appear to be constant or susceptible to little variation. 
In this sounienice it may be remembered that the form of the 
