TURtfERACE^J FROM RODRIGUEZ. 161 



The characters which separate Mathurina from Turnera and 

 TVorrnskioldia are so numerous and so evident that I need not 

 dwell upon them ; but its relationship to Erblichia is so close that 

 an indication of the characters on which its generic separation de- 

 pends seems advisable. In habit and mode of leafing the plants 

 are almost identical ; but the stipules in Erblichia are mar- 

 cescent, whilst in Mathurina they are very early deciduous. In 

 the sepals of Mathurina is found one of its most important 

 generic characters : each sepal is clothed at the base on the inner 

 aspect with a large bilobed emarginate gland, about one sixth of 

 the length of the sepal ; and from its apex there runs a prominent 

 broad nerve to the extremity of the sepal. This structure is 

 unrepresented in Erblichia. Coming now to the corolla, we find 

 the petaline fringes of Erblichia absent in Mathurina; nor is 

 there such a distinct claw to each petal. The stigmas of Erblichia 

 are described in the l Genera Plantarum ' as dilated and entire ; 

 but those of Mathurina are distinctly though minutely fimbriate, 

 and in the dried specimen the fringe is folded inwards into the 

 club-shaped dilated extremity of the style. Seemann (Bot. Her. 

 130, t. 27), in his description, says the stigmas of Erblichia are 

 fimbriate. I cannot say from observation whether they are so 

 or not. From the seeds we might expect to make an impor- 

 tant diagnosis ; but whilst the slightly curved foveolate seeds of 

 Mathurina, with an aril of membranous filaments extending up- 

 wards around it from a ring encircling the hilum, are very cha- 

 racteristic, the seeds of Erblichia are unfortunately very imper- 

 fectly known : in fact we only know that they are straight ; and 

 in this particular they differ from those of Mathurina. 



Although, then, our knowledge of the genus Erblichia is not 

 complete, I think the characters in which we can compare it 

 with the Eodriguez plant, especially those in the floral envelopes, 

 are, in this family, sufficient to warrant us in taking the Eodriguez 

 plant as the type of a new genus ; at least it seems to me more 

 rational to do so than to mutilate the description of an old genus 

 to embrace a new species upon imperfect data. I have therefore 

 constituted the Eodriguez plant a new genus Mathurina^ adopt- 

 ing for it that of the town, Port Mathurin, in Eodriguez. 



It is hardly within my province here to describe the relation- 

 ships of the family Turneraceae ; but a glance at them as affected 

 by the discovery of Mathurina may be not uninteresting. 



Turneraceae are alwavs regarded as very closely allied to Passi- 



o2 



