298 NATURAL HISTORY OF PLANTS. 



Thus we gradually arrive at a species of the same genus, called 



Matlhcea sancta^ (figs. 333-33G), in which the androceum pre- 



*r II- J- /,r.,i N , sents the greatest reduction 



Molhnedia [Matf/ifpa) sancta. ^ " 



known in the number of its 



elements. This species has 

 exactly the perianth of 

 Wilkien or Kibarojma, but 

 the sterile stamens have 

 disappeared, and we only- 

 find four stamens with 

 nearly sessile naiTow sad- 

 dle-shaped anthers, super- 

 posed to the perianth- 

 loljes." The female flower 



and fruit are altogether those of the plants described above. 



The only known species of this section is a shrub with shortly 



Mollinedia (Mad/itta) sancla. 



Fig. 333. 



Transverse section of 



male flower. 



Fig. 331. 



Long, section 



of male flower (Y)- 



Fig. 335. 

 Female flower, dehiscing. 



Fig. 336. 

 Long, auction of female flower ('j»). 



petiolate entire or serrulate leaves, and mona'cious flowers in axil- 

 lary cymes. 



In the plants we are now about to study, the fruits, instead of 



' I?i.., Mut. Lugd. jB«/.,ii. Sy.t. 10.— A. DC, 

 I'lodr., xvi., B. poHt., G(i'J. — 11. Hn., Adanson'ia, 

 ix. 118. \U. 



2 " Nonniti dehUcenlin nntherarum a Kihara 

 differre videtur." (A. 1)('., he. cit.) T\\\n do- 

 liiscence is the same in lM)th types. The two 

 ••elln, clone tf)gether hI tiie ajK-x, each dehivce l>y 



a subinHrginnl cleft, bnt later on the two clefta 

 coalesce above to form a single cnrved line with 

 its concavity downwards (tigs. 332-33 J). Those 

 Btumcns are exactly like those of Ephippiandra, 

 except that the curve tormod by the two cell* is 

 a little greater, owing to the great^jr elevation of 

 the coinicctivc. 



