NATURAL HISTORY OF PLANTS. 



by a large pore.^ Noteroijliila, whose staminal pores are small, can- 

 not always be distinguished from Adsanthera. Neither can we 

 separate, except as sections, and chiefly by the habit and foliage, a 

 great many genera the flowers of which present no important diffe- 

 rences, and are distinguished only by characters of no generic value. 

 Such are : Comolia, herbs and shrubs with very variable organs of 

 vegetation, whose stamens are often but slightly unequal, and have 

 sometimes a connective less prolonged below the anther, and furnished 

 with two longer or shorter basilar projections ; Fritzcliia, from South- 

 eastern Brazil, whose receptacle and calix are glabrous and their 

 ovary four-celled; their habit and foliage are those of a Thymus;- 

 Marcetia, from the same country, with glabrous or silky calyx, whose 

 anthers have a connective somewhat thickened outwardly at the base ; 

 the leaves are small, often heathlike ; they sometimes resemble the 

 Chaincelauciem ; Chcetolepis, whose connective is articulated to the 

 summit of the filament, and its receptacle glabrous or furnished 

 with silky hairs and scales ; the habit is that of 

 certain Lahiece ; Haplodesmium, from Venezuela 

 and Columbia, which is Chcetohpis destitute of 

 accessory appendages between the sepals ; Heeria, 

 whose tetramerous flowers have eight stamens, 

 the four larger furnished with a long prolongation 

 of the connective ; they are herbaceous with 

 flowers solitary or in c^^mes, and inhabit Guate- 

 mala and Mexico ; 31icrolepiSf small shrubs of 

 southern Brazil, whose flowers are pentamerous 

 with the connective of the ten stamens prolonged anteriorly to a 

 double spur ; the leaves are larger, tomentose, and the inflorescence 

 terminal and axillary, compound and many-flowered. Desmoscelis, 

 Brazilian herbs or under-shrubs, still more hairy, with simple stem, 

 ten anthers and hairy 5 -celled ovary ; the connective is furnished 



li'jottchina {Heeria) 

 macrostachya. 



Fig. 8. Flower. 



' Several types referred to the group of 

 Microliclece appear to us not generically dis- 

 tinct from Tibouchina. Such are Tulasnea and 

 Poteranfhera, which may have the same habit 

 and flower as the true Adsanthera, hut which 

 have been artificially separated because they 

 have, it is said, a rectilinear ovoid instead of a 

 curved seed. But in Poteranthem it is often 



as much or more curved than in the Acisan- 

 thera type. This character could not, there- 

 fore, have an absolute value. For the saihe 

 reason we shall also consider Meissneria, Due- 

 quetia, Svitramia, as sections of the genus 

 Tibouchina ; the two last are scarcely distinct 

 from Fritzschia except in the form of their 

 seeds. 



