194 



dimerous ovary, in tlie nnmerous seeds, in the free styles and in tlie anatro- 

 pous ovules. But in otlier respects, it is comparable with the Caryoph^-llacesB. 



115 Hydrostachyaceae : Related to the Podostemonacese. 



116 Crassulaceae ^^ : This family is somewhat related to the Elatinaceas 

 the isomerous flowers and in the habit, as can be seen in IMlcea aqucdica L. 

 in which is at present referred to the Crassulacese, but is closely related to 

 Elcdine. In general, it is connected with the Saxifragacese more closely than 

 with any other family. 



117 CeplialotaceaB '^ ." Near the Saxifragacese and Ci-assulacese. 



118 Saxifragaceae ^^ .* Certainly quite close to the Crassulaceae and Cuno- 

 niacese, and then, to the HamameHdacese and Kosacese. But, the family has 

 some relation to the Trochodendracese in the separate carpels, in the perigy- 

 nous insertion and in the small embryo ; then, to the Kceberliniacese in the 

 placenta with numerous seeds in many series. Further, the resemblance of 

 this family to the Begoniacese can not be denied, as there are agreements in 

 both families, in the perigynous insertion, in the parietal placenta, in the 

 small seeds, in the dehiscence of the fruit, and in the shape of the leaves. 



119 Pittosporaceae ^^ : R. Brown compares tliis family with the Celastra- 

 ceae and Rhamnacese. De Candolle places it between the Polygalaceae and the 

 Caryophyllacese, wliile Richakd and Schnitzlein insist upon the existence of 

 a relation between it and the Caryophyllaceae. Endlichee and Eichler con- 

 sider it better to bring it back again near the Celastracese ; but Bentham and 

 HooKEE class it with the Polygalaceae, Tremandracese and Vochysiacese in the 

 Polygalinese. Van Tieghem compares it with the Umbelliferae and AraHaceae 

 in the peculiar distribution of resin-ducts in the root and in the formation of 

 lateral roots. Baillon says that there is no absolute difference between the 

 floral structure of the Pittosporaceae and that of the Saxifragacese. Some re- 

 lations seem to exist between the Pittosporacese and the Hamamehdacese in 

 the presence of resin-ducts. 



120 Brunelliaceae^^ : Somewhat near the Cephalotaceae, this family is 



1) Nat. Pfl.-fam. 111—2, a, p. 28. 2) 1. c. III.— 2, a, p. 40. 



3) 1. c. III.— 2, a, p. 45. 4) Nat. Pfl.-fam. III.— 2, a, p. 108. 



5) Nat. Pfl.-fam. Nacht. p. 184. 



