410 TRELEASE— THE DESERT GROUP XOLIXE^. [April 21. 



escaped my search, but the process may be inferred with some 

 probabiHty from the character of the fruit. In all. the ripened fniit, 

 with its enclosed or attached seed or seeds and the persistent but 

 unenlarged perianth, falls by disarticulation of the pedicel, — close 

 to the fruit in Dasylirion and Calibaiuis, somewhat further from it 

 in Bcaiicarnca, and usually at a still greater distance in Nol'ina. No 

 provision for dissemination other than through rolling or being 

 blown over the ground appears in the round fruit of Calibanus. 

 The winged fruits of Dasylirion and Bcancarnca are as evidently 

 wind-scattered as the similarly disarticulating and equally small 

 fruits of Rtiiiicx, — though in the latter the wings are not carpellary 

 but consist of the enlarged persistent sepals. The very different 

 fruits of Nolina are likewise evidently wind-disseminated, their 

 more or less inflated carpels giving them a character intermediate 

 between winged and balloon fruits. 



Germination, of which no published records have been found, is 

 of Klebs' Asphodelus-Tradcscantia type.^"--'^' the seed — freed from 

 the remnants of the fruit in NoVuia or still contained in them in the 

 other genera — remaining in the ground with the arched haustorium 

 elongating with the cotyledonary sheath so as to reach a length of 

 even 10 mm. InKoVuia longifolia and in specimens of N .parviflora 

 preserved by Dr. Rose, the haustorium is apical, though a slight 

 elbowing is sometimes seen near the the top of the arch ; and it 

 sometimes straightens and lifts the seed from the ground. Seed- 

 lings of Bcaiicanica and Calibaints preserved by Dr. Rose show that 

 in these genera the sheath is produced above the arch in form of a 

 pointed ventral ligule, as is true in such species of Dasylirion as I 

 have observed. In these cases the haustorium appears to be dis- 

 tinctly dorsal on the sheath, along which it is often sharply re- 

 fracted (pi. i^-ij). Initial growth is evidently at the princii)al 

 expense of the granular protoplasm, oil and " reserve-cellulose " of 

 the endosperm. In Calibanus and Bcancarnca, as is shown in excel- 

 lent specimens in the National Herbarium prepared by Dr. Rose, the 

 formation of the thick trunk follows germination quickly. 



