M:ir iMJi.ix svuophulaiuacea:. 411) 



order. Only tho staininodes aro not drawn accurately. They are yellow- 

 in all our specimens, united with the corolla for about § of their length, 

 parallel ; the lower portion is broad, yellow-pubescent ; the free upper part 

 narrow, glabrous, much bent outwards and on itself, tip glandular, obtuse. 

 When straightened out tho tip reaches beyond tho sinus in tho lower lip. 



The measurements given in Hooker's Fl. Brit. Ind. and Cooke's Fl. B. 

 Pr. for the corollas of tho three plants in question are contradictory. As 

 a matter of fact, all have rather large flowers, up to 12mm. long, B. 

 i/randiftura perhaps the largest. 



The seeds are always rugose with small circular pits arranged in rows. 



The citation of "Wight Ic. t. 144 for B. veronicccfolia is sorawhat doubtful, 

 because of the close serraturo of the leaves in the plant figured. It may 

 be B. icptans, Spreng., a plant which has not been observed by us in the 

 tield. 



In our opinion the character of the genus should be widened in order to 

 include B. eiftamuwdioso. The genus Bonnai/a should, therefore, be distin- 

 guished from Ilysanthea by havincf no staminodes at all or entire ones, 

 whilest lhjsa7ithi'S would be characterized by bilobed staminodes. Further, 

 researches into the relationships between Vandellia , llyscmtlies and Bonnaya 

 will probably show that it is impossible to separate the three genera by 

 good characters and that they should be united into one. 



ILYSANTRES, Kaf. 



It appears that considerable confusion prevails in the nomenclature of 

 this genus as regards the two Bombay species : Ihjsanthes /ryssojnoides, 

 Benth. and /. parvi/ioia, Benth. This seems to be due to some erroneous 

 citations of Bentham's in DC. Prodr. X, 419-20. 



Under /. /lyssopioides he gives as synonyms : Gratiola hyssopioides, L. 

 (correctly) ; Koxb. Cor. PI. Ill, t. 203 (wrongly, as this plate gives Rox- 

 burgh's original Gratiola parvijiora) ; Bo7inaya hyssopioides, Wight Ic. Ill, 

 t. 857 (wrongly). 



Similarly under I. parvijiora, Benth., Gratiola parviflora, Roxb. Cor. PL 

 III., t. 204 (wrongly, it ought to be t. 203). 



These mistakes are retained by J. D. Hooker, Avho adds to the confusion 

 by expressly saying that Gratiola hyssopioides, Roxb. Cor. PI. II, t. 128, as 

 well as the corresponding description in tho Fl. Indica do not refer to 

 /. hyssopioides, Benth. Hooker refers the plate in question to Dopatrium 

 lobelioides, Benth. (Fl. Brit. Ind. IV, 274), an entiretydifferent plant, 

 correctly figured by Wight (Ic. t. 859, also cited by Hooker). 



Cooke in his Fl. B. Pr. is evidently at a loss, and excludes all references 

 except Wight, Ic. t. 857, which he, too, wrongly refers to /. hyssopioides, 

 Benth. 



These mistakes should be corrected as follows : — 



Ilysanthes hyssojnoides, Benth. in DC. Prodr. X, 419. 



Syn. — Gratiola hyssopioides, L. Mant. 174 ; Roxb. Cor. PI. II, t. 128, Fl. 

 Ind. 1, 141. Bonnaya hyssopioides, Benth. Scroph. Ind. 34. 



The plant is apparently rare, at least in the Bombay Presidency. The 

 localities given by Cooke must be regarded as very doubtful. They refer 

 in aU probability to i. pari-ijlora, Benth. W^e found the plant at Khan- 

 dala in August 1917, growing in a damp place near St. Peter's School 

 (Herb. St. Xavier's College, kos. 1527, 1528). It agrees in every respect 

 with Roxburgh's figure. 



In order to clear up doubtful points we give a description of the plant. 



Stems several from the base, very weak, but rather stout, succulent 



