BiCKXELL : Studies in Sisyrixchiu.m 349 



I 



character in the genus and were it shown by a single specimen 

 only might reasonably be referred to an abnormal condition. It is 

 however exhibited by five perfectly healthy examples collected at 

 different times in three distinct localities. 



The origin of these basal flowers may perhaps be referred to 

 suppressed development of the usual stem. They occur either in 

 separate leaf clusters or among the leaves about the bases of 

 normal stems, but in the latter case although enclosed by the same 

 outer leaves as the stem they nevertheless belong to an inner tuft 

 of leaves which arises from a separate point of the axis. 



Some plants show no indications of these radiqal flowers and 

 Hre then with difficult}' separable from S. angustifoUuni. The 

 plant appears to need comparison with no other species than this. 

 At differs in narrower mostly smooth-edged leaves and stem^ smal- 

 ler and narrower spathes, less dev^eloped inner bract, shorter pedi- 

 cels and larger more globose or pyriform paler capsules, often 

 trans v^ersely corrugate. The thickened roots are a further note- 





worthy character. The {^\v ?^o\\cxs seen are rather smaller than 

 those of 6~'. augustifolinm, the divisions of the perianth less rounded 

 ^t the apex and more distinctly aristulate, the yellow eye also ap- 

 pears to be larger. 



In some specimens the basal and terminal capsules are of about 

 ^qual age ; others show at the same time basal capsules nearly ma- 

 ture, and upper spathcs still tx^tXy twdosmg their flower buds. 

 The successive opening of the flowers appears to be remarkably 

 prolonged or irregular for several spathes which bear nearly ma- 

 ture capsules also enclose very immature flower buds. 



^lounted on scA^eral sheets among fruiting specimens are a 

 ^^"^ plants just \x\ flower which arc larger, sXoutcr and broader 

 heaved with longer pedicels ; these are apparently not distinguish- 



tifoliun 



are 



fa 



t 



^11-pum, with smooth-edged stems, and their time of first-flowering, 

 August, is hard to reconcile with wliat is known o{ the flowering 

 period of ^. a?igiistifoHum. A k\\ of the specimens of 5. /ictiro- 

 <^ftrpum show some stems stouter than the others and somewhat 

 approaching these. Evidently the plant in all its phases cannot be 

 ^ully understood from present material and its exact relationship 

 ^o i". angustifoliitui must remain a subject for future study. 



