Mr. Bentham's Review of the Order q/TIydrophyllese. 269 



that be really the only distinction between Nemophila and Ellisia, it proves 

 the expediency of uniting these two genera, more especially as in N.parvijlora 

 and pedunculata the appendages are sometimes scarcely perceptible, or even 

 entirely wanting, and at other times nearly half as long as the calyx. 



Ellisia Nyctelea has, indeed, another character, derived from the same organ, 

 which distinguishes it from Nemophila, that is, the large size and expansion 

 which it acquires after the fall of the corolla ; but this, again, is but a charac- 

 ter of degree, in respect of which the Ellisia chrysanthemifoUa would stand 

 intermediate between the two genera, and is not, therefore, available to sepa- 

 rate them. 



The corolla of Hydrophylleae varies in general form from campanulate to 

 rotate, showing but rarely {Eutoca phacelioides) an approach to the infundi- 

 buliform corolla of most Borraginece. But these variations are slight, difficult 

 to characterize, and so little in relation to general habit as to be unavailable 

 for generic distinction. 



I have never observed in Hydrophyllece any trace of those coroUine appen- 

 dages, or squamae, which may be termed laminal, from their position at the 

 base of the limb of many infundibiiliform corollse, whether monopetalous (as 

 Borraginece), or polypetalous (as Cai-yophyllece) ; but the unguicular or tubal 

 squamce are often remarkably prominent. These squamae are analogous to 

 those placed at the base of the tube of many Borraginece (having at the same 

 time ligiilar squamae at the mouth of the tube), and to the annulus of hairs or 

 scales in the tube of many Lahiatae. 



The laminal squamae are evidently of no organic importance, but mere 

 excrescences of the petals, showing an approach to a transformation analo- 

 gous to that by which the ligulae are converted into anthers, but for a different 

 purpose, whatever that purpose may be. But many circumstances might in- 

 duce a supposition that the case may be different with regard to the tubal 

 squamae. Their origin is always below that of the stamina, and in some 

 cases (as in Emmenanfhe and some Eutocce) tliey appear to be connected with 

 the filaments of the stamina in a manner analogous to the abortive filaments 

 in pentandrous Caryophyllece, Paronychiece, and Ainaranthacew. On the other 

 hand, like the ligular squamae, although constant in form in the same species, 

 they vary much in the most natural genera, such as Ajuga, Salvia and Stachys 



2 N 2 



