200 



168 Hippocastanaceae^-' : Very near to the Sapindacese and Aceracese. 



169 Sapindaceae "■' ." Close to tlie HippDcastanaceee and Aceraceae. 

 Fiirther, tliis family is related on one liand to the Mehaceae, and on the other 

 to the Anacardiacese. Its affinity with the MeHaceae is to be found in the habit 

 and anatomy, especially in the secretory cells ; to the Anacardiaceae, in the 

 habit, and in the apotropous ovules. Through the Anacardiaceae, it is connected 

 with the Burseraceae ; through the MeHacese, it is alHed to the Simarubacese 

 and Kutacese. Further, it is sometimes compared with the Malpighiaceae in 

 its winged fruit; also to the MeHanthaceae and Staphyleacefe by the same 

 fi'uit character. 



170 Sabiaceae ^^ : In the exalbuminous seeds, the curved embryos, the 

 reniform fi-uit, and the nearly ajxjcarpous ovaries, the Sabiacese are compared 

 with the Menispermaceae by Blume, Miers, and Hooker et Thomson. 

 But Bentham-Hooker put them close to the Anacardiacese and Sabiacese. 



171 Melianthaceae *^ i Closely related to the Sapindacese, as is seen in the 

 agreement of Bersarna and Meliantlms, in the zygomorphous flowers, in the one- 

 sided extra-staminal discus, in the abortion of some coroUa-lobes and stamens, 

 and in the apotropous ovules ; also to the Staphyleacefe, in the seeds with copious 

 albumen, in the straight embryo and in the absence of the continuous scle- 

 renchymatous ring. The Balsaininaceae present some relation with this family 

 in the median symmetry of their flowers, in the spur-Hke form of one of the 

 sepals and in the deposits of potassium oxalate. Kadlkofer refers to the rela- 

 tion of the Zygophyllaceae and the MeHantliacese, as he finds potassium oxalate 

 in Bersama and Meliantlms, which occurs in the same form as in the Zygophylla- 

 cese, Harvey and Baillon, however, compare the family with the Saxifi-agaceie. 



172 Balsaminaceae^^ : The only reason for this family being refen-ed to 

 the Sapindales is the position of the micropyles and raphes. On the other 

 hand, in general characters, it is closely related to the famihes of the Grera- 

 nials, 'such as the Malpighiaceae, Trigoniacese, Vochysiacese and Tropseolacese. 

 Whether it should be assigned to the Sapindales or to the Geraniales depends 



1) 1. c. in.-5, p. 275. 2) 1. c. IIL— 5, p. 298. 



3) Nat. Pfl.-fam. in.— 5, p. 369. 4) 1. c. ni.— 5, p. 378. 



5) Nat. Pfl.-fam. III.— 5, p. 388. 



