CRUCIFERJE. 49 



2. A fab is LINN. 



Dichotomous Key to tlie Formosan Species. 



Leaves lyrately Iol3ed AraUs morrisonemis 



Leaves nearly entire or a very sparely toothed Arabis alpina 



Arabis momsonensis HAY ATA (PI. XL) Materials for a Flora of Formosa 

 p. 29 Aralis iaraxacifdia HAY ATA Fl. Mont. Forinos. p. 49, (nou ANDERS). 

 Herb slightly liguified at the base, perennial, hirsute, hairs simple or branched 

 J mm.-l mm. long, stoloniferous, decumbent. Stem 20 cm. long, branches 

 leafv. Radical leaves clrstered radiatey arranged, long petioled 3 cm. long 

 including petioles, 6 mm. broad stellately hirsute, (hairs furcate or simple i 

 spathulate in outline, petioles 1 -J- cm. long, as long as the blades, blades 

 ovate Ivrate, lobes 4-5 011 Ijoth sides, terminal lobe obovate, obtuse. 

 Cauliue leaves simple, oblauceolate, 3| cm. long, 4 mm. broad, obtuse at 

 the apex, gradually attenuate downwards scarcely sen-ate on the margin, or 

 nearly entire. Racemes terminal or axilkry, 0-6 cm. long, pedicels 1-2 cm. 

 long, bracts 0. Sepals narrower oblong, < *2^ mm. long, 1 mm. broad, round- 

 ed and mucronate at the apex, sparingly hirsute on the back. Petals 

 ( blong-ovate. clawed, G mm. long, rounded or truncate at the apex. Stamens 

 2 mm. long. Siliques straight or slightly curved, linear, 3-4 cm. long, 1 mm 

 broad, obtuse at lx>th ends ; styles persistent, pedicels 1^ cm. long. Seeds 

 elougately oblong, 1^ mm. long, f mm. broad, rounded at both ends, com- 

 planate, somewliat scabrous. 



HAB. Mt. Monlson. 



In my paper above cited, I mention that the present plant agrees quite 

 well with the description of Arabis taraxacifdia ANDERS given in HOOK. 

 f. Fl. Biit. Ind. I. p. 136, and is also very like the European A. erenosa 

 8 r. Although I did not, at that time, see ANDERSON'S specimen, I 

 thought that the plant must be identical with A. tarcacacffoKa, While 

 studying here at Kew, I have compared it with the type of the same 

 species, and liave found that they are quite different. The Formosan plant 

 i* distinguishable from the Indian in liaviug nearly erect and stouter pods, 



