120 NOTES ON LILIES 



anthers, versatile, attached on the inside by the middle of the back, 

 dehiscing outwardly along the entire margins ; ovary, sessile, cylin- 

 drical, three-celled ; ovules, numerous, horizontally arranged in each 

 cell ; style, club-shaped, more or less curved, most frequently longer 

 than the ovary; stigma, rounded, three-lobed; capsule, coriaceous, 

 obovoid, six-angled, with a three-valved loculicidal dehiscence; seeds 

 numerous, discoid, margined, and with a brown membranous skin. 



First, I will give a few notes on the characters in which the species 

 differ from one another. 



Bulbs. The bulbs in the great body of the Lilies furnish no 

 important modification in structure. So far as I am aware, they all 

 present the type of fleshy imbricated lanceolate scales which we see 

 so often drawn in botanical handbooks. 



Stem. There is nothing about this that needs to be mentioned 

 now. 



Petiole.r-TTue Petioles, sometimes as long as the leaves, are 

 present in Gordrfolium and Giganteum, and short but distinct ones 

 in Speciosiim and Aitratum ; but in the majority of species the leaves 

 are quite sessile. 



Leaves vary greatly in number, shape, texture, veining, direction, 

 and arrangement. The presence of bulblets in the axils is not, I 

 think, invariably of specific value. The large rotun date-cordate 

 leaves of Cordifolium and Giganteum are very different to the 

 lanceolate or linear leaves of all the others. One of the best 

 characters for distinguishing species is furnished by the arrangement 

 of the leaves of several of them in regular whorls ; but this character, 

 useful though it be, is not entirely trustworthy. I believe that all 

 the verticillate species are liable to have the whorls broken down, 

 sometimes partially, occasionally entirely, in exceptional cases a 

 variation that is very puzzling when a specimen of this kind comes 

 to hand by itself, without any warning of its exceptional character. 



Inflorescence. I believe that although several of the species are 

 always single-flowered in the wild state, they all may be made to 

 produce more than one flower under cultivation, and still more readily 

 may the species that commonly yield three or four flowers when wild 

 be made to show from twelve to twenty in gardens. There is always 

 a tendency, too, in what are typically thyrsoid racemes, to become 

 congested into umbels under cultivation, and of what are properly 

 scattered bracts to fall into whorls. 



Perianth. Good characters may be had from the position of the 

 individual flowers, whether erect, drooping, or intermediate in direc- 

 tion ; very good characters also from the shape of the flower, whether 

 permanently funnel-shaped, with the divisions spreading falcately 

 towards the lip only when fully expanded ; or bell-shaped, with the 

 divisions more or less spreading, or often decidedly recurved. The 

 shape of the segments of the perianth furnishes also excellent 



