30 AKT. I.— B. HAYATA : 



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

 a^'recs quite well witii the description of Arabis taraxacifolia 

 Andeks. given in Hook. f. Fl. Brit. Ind. I. p. 136, and also very 

 like the Em^opean A. arenosa Scop. Although I did not, at that 

 time, see Anderson's specimen, I thought that the plant might 

 be identical with A. taraxacifolia. While studying here at Kew, 

 I have compared it with the type of the some species and 

 have found that they are quite different. The Formosan plant 

 is distir^guishable from the Indian in having nearly erect and 

 stouter pods, larger seeds, and smaller leaves and also in 

 bearing very long stolons. The Indian plant appears to be of 

 more tender habit, witli curved pods, narrower, smaller seeds, and 

 larger, thinner leaves. Moreover, the present plant differs from 

 A. arenosa Scop, in having longer pods, smaller flowers, and 

 leaves with more rounded lobes. The description given above is 

 drawn up from a specimen numbered 680, which is of a rather 

 elongate form and prostrate habit. 



Cardamine Linn. 



Cardamine hirsuta Linn. var. formosana Hayata n, v. Caulis 

 glaber, ramosus, ascendens, 15 cm. altus, foliosus. Folia pinnati- 

 secta, obovata in circumscriptione, cum petiolis 5 cm. longa, 

 2J cm. lata, petiohs 2 cm. longis complanatis vel alatis, segmentis 

 lateralibus obovatis apice rotundatis, 3-5 lobulatis, basi subito 

 attenuatis 2J cm. longis. liacemi 4 cm. longi, pedicelhs 1 mm. 

 longis, bracteis obtusis | mm. longis. Sepala 4, oblonga IJ mm. 

 longa I mm. lata. Petala 4, obovata, spathulata IJ mm. longa, 

 apice rotundata vel emarginata, basi angustata. Stamina IJ mm. 

 longa, filamentis complanatis. Ovarium cylindraceum IJ mm. 

 longum. Siliquœ lineares rectge, 17 mm. longae 1 mm. latœ, 



