144 



psrigyuous iuserfciou, iu tli8 parietal placeutas, iu the smalluess of tlie seeds, 

 in tlie deliiscence of the fruit aud in the shape of the leaves ; of the Cuuoui- 

 acese aud the Quiiuacese^^ iu exterual featm'e3 aud iu the deusely hairy 

 seeds ; of the HaTnamehdaceaB aud Dipterocarpacese"'^ in the preseuce of a riug 

 of secretory - canals in the medulla - crowu ; aud fiually, of the Rosacese aud the 

 Eucryphiacese^^ Moreover, some affiuity betweeu the Rosales aud the 

 Gerauiales is to be found iu a few aualogous features of the Pittosporacese"' 

 of the former series ou the oue side, aud of the Vochysiacese, Polygalacese, and 

 Tremandracese of the latter series ou the other (Benth.-Hooker) ; and aJso 

 in those of the HamameHdaceae and Simarubacese'^^ in the preseuce of a ring 

 of secretory-canals iu the medulla - crown. To the Myrtiflorse, its aUiauce is, 

 in some measure, to be seen iu a few similarities of the Rosacese''^ on the one 

 haud, and of the Combretacese, Myrtacese aud Thynielaeacege on the other. 

 The relation between the Rosales and the Sapindales is rather slight, as can 

 be seen from some agreement of the Saxifragacese aud the Melianthaceae'^^ and 

 from that of the Pittosporaceee^^ and Celastracese in their floral diagram. To 

 the Khamnales, the Rosales are on some degree related, as is shown iu the 

 Pittosporacese^^ aud Ehamnacese (R. Beown) ; aud also to the Umbelliflorae' 

 iu some conformities of the Pittosporacese aud Araliacese, and of the former 

 family aud the UmbeUiferae, in the pecuHar distributiou of the resiu-ducts in 

 the roots, in the formation of lateral roots aud in other anatomical characters 

 (Van Tieghem)"^^ Lastly, they are somewhat related to the Rubiales in the 

 shght agreemeut of the Saxifragaceae and Adoxaceae."^ As has been pointed 

 out, the relations of the Rosales in this poiut or that exteud from the 

 Santalales far up to the Rubiales. Their natural position is, therefore, dyuamic. 

 Next to the Rosales in his syllabus, Engler places the Pandales^^. As 

 to the relations of the latter series, we are as yet ignorant. 



1) 1. c. III.-6, p. 166. 2) 1. c. IIL— 6, p. 252. 



3) 1. c. III.— 6, p. 131. 4) 1, c. III.— 2, a, p. 108. 



5) Nnt. Pfl.-fam. m.— 5, p. 203 and HL— 6, p. 253. 



6) 1. c. IIL— 3, p. 10. 7) 1. c. III.— 5, p. 378. 



8) 1. c. III.- 2, a, p. 108. 9), 10) Nat. Pfl.-fam. IIL— 2, n, p. 108. 



11) L c. IV.- i, p. 171. 12) Engler, A. — Syllabus 1. c. p. 223. 



