268 Mr. Bentham's Review of the Order of HydrophylleEE. 



ovarium, as far as it goes, appears to be the best character that can be taken 

 for generic distinctions. In order to show how far any others may be brought 

 in aid, I shall proceed to examine them separately. 



In general habit and foliage, the distinction between HydrophiiUece and 

 Borraginece is not always so marked as appeared from the species first known. 

 Several PhaceUce and Eutocce have exactly the habit of Echittm, Cynoglossum, 

 or Anchusa ; and some of the latter genus have the leaves constantly entire, 

 although the " folia composita vel alte lobata" do run through the greater 

 number of species. Some Nemophilce may be compared to Asperugo, which 

 has frequently opposite leaves (though always entire) and the same fragile 

 trailing stem. The rough hispid hairs are the same in both tribes. As amongst 

 one anotlier, HydropliyUum, Emmenanthe, and Nemophila have each a peculiar 

 habit ; but Eutoca and Phacelia are so much blended together in this respect 

 that it would be difficult to assign any character derived from the vegetative 

 organs peculiar to either genus. 



The gyrate inflorescence of Borraginece may be very readily observed in 

 Hydrophyllum, Phacelia, Eutoca, Emmenanthe, ElVma, and in Nemophila au- 

 rita and phacelioides ; but in the remaining Nemophila; it is (as in Asperugo) 

 axillary, and can therefore only serve as a specific, not as a generic character, 

 and in the general description of the order it must be considered in the light 

 of a subsidiary, not an essential character. 



The calyx is usually the same as in the majority of Borraginece, inferior, 

 persistent, and deeply 5 -cleft, but with this particularity, that in some in- 

 stances the sinuses (as in some Campanulacece) are furnished with reflexed 

 appendages, resembling the erect divisions of the calyx in form, but smaller in 

 size. As shown by M. Alphonse De CandoUe in his Monographie des Campa- 

 nuUes, p. 11, these divisions do not indicate any organic modification in the 

 composition of the calyx, but are merely owing to the prolongation of the 

 united lateral nerves of two adjoining sepals, as is rendered evident by the 

 nervation of the calyx of Ballota, Marrubium, Leucas, and other Labiatui with 

 more than five teeth to their calyx. The character derived from this cir- 

 cumstance must consequently be inconstant, and have little or no relation to 

 general habit, as may be observed in Hydrophyllum, where it would separate 

 H. appendiculatum from its close allies H. canadense and virginicum ; and if 



