IV INTRODUCTION. 



Another perplexing plant whose proper genus I found it extremely 

 difficult to determine is Disporopsis arlsanensis belonging to the Li I ia cere. 

 It is with some hesitation that I have referred the plant to the said genus, 

 whose original description is rather too imperfect to be clearly understood. 

 To begin with the perianth, the description runs " perigonium corollinum, 

 campanulatum, basi leviter saccatum, segmentis 6, subseqiialibus, crassiusculis, 

 uninerviis, tubo duplo longioribus-" So far, this agrees fairly well with our 

 plant. Then conies the corona " Corona apice tubi inserta, lobis paulo brevier, 

 ultra medium in lacinias 6, staminibus alternantes, lineares, acutas, conniventes, 

 genitalia omnino velantes, fissa." I have been very much puzzled to under- 

 stand what is meant by the " corona." As to the stamens, it is stated 

 " stamina 6, inter corona) lobos affixa." The explanation which I shall give 

 to the " corona," judging from the position of the anthers, is that it is a 

 cylindrical tube, composed of dilated filaments which are united one to another 

 and protrude far beyond the insertion of the anthers, but the points of in- 

 sertion are terminated by the anthers themselves. In other words, the corona 

 is a tube formed by the perfectly united broad filaments which are sagittately 

 2-lobed at the apex and have anthers at the sinus between the lobes. Then 

 follows " filamentis brevissimis ; antheris introrsis, brevibus, dorsifixis, secus mar- 

 gines dehiscentibus. Ovarium sessile, ovoideum, trigastrum, triloculare. Stylus 

 brevis, crassiusculus ; stigma punctiforme. Ovula in quoque loculo 4, angulo 



interno affixa, horizontalia. Bacca ? Herba rhizomatosa, caulescens, " 



If my explanation of the " corona " is correct, our plant perfectly accords 

 with HANCE'S original description in the generic characters. The Formosan 

 species has filaments with sagittate apex, partly united, but distinct nearly 

 half way down. The point of difference between HANCE'S species and mine 

 is, therefore, whether the filaments are perfectly united one another or 

 connate for half their length. 



