140 



an Formen einer anderen Reihe gestatten, dann sine! dieselben axis der ersteren zu 

 entfernen itnrl an die andere anznischlieBen. 



In my opinion, however, such is not the case. As I have discussed series 

 after series, the several groups bear a close or loose relation to other groups, 

 some of which lie near the group under consideration and some far from it, 

 even judging from our present limited knowledge. Some species assignable 

 to a series in one respect are just as clearly referable to another series in 

 another respect. It is absolutely impossible to make up more natural groups, 

 by taking one form from its old place to a new one. Natural relations should, 

 in any case, appear in the dynamic view. 



Now, we should examine the Polyp3talous groups. Let us begin with 

 the Centrospermse. This series is related to the Batidales, Polygonales, 

 Urticales, and Santalales, as has been discussed before. It is also allied to the 

 Parietales, as is indicated by the agreement in the position of stamens in the 

 Portulacaceae and Loasacefe n . On the other hand, the resemblance is to be 

 found in this series and the Rhoeadales, the floral structure being somewhat 

 similar in the Phytolaccacepe and Tovariace?e 2) (BENTH.-HoOKER). Further, it 

 bears some relation to the Geraniales, in the matter of the incontestable re- 

 semblance in the Caryophyllacere-Silenoidese and the Liuace?e s) . Moreover, a 

 striking affinity of the Centrospermso to the Opuutiales is suggested in 

 Mescmbrianthemwn of the Aizoacete and in Opuntia of the Cactacese in the 

 floral structure . The Centrospermae are, therefore, related to the Batidales, 

 Polygonales, Urticales, Santalales, Parietales, Ranales, Rhceadales, and Geraniales, 

 in this point or that. Their natural position should be a very dynamic 

 one. 



Next come the Ranales r . As this series implies many different things, 

 its relations extend to many other series. Some families of the Ranales are 

 closely allied to the Monocotyledones and some are, if not very evidently, even to 

 the Metachlamydere. The relations of the Ranales to the Urticales, Helobife 

 and Aristolocliiales, we have treated above. Their affinity to the Ebenales can not 



1) Nat. Pfl.-fam. III. 6, a, p. 106. 2) Nat. Pfl.-fam. III. -2, p. 207. 



3) Nat. Pfl.-fam. III. 4, p. 30. 4) WETTSTEIN, R. R 1. c. p. 533. 



5) ENGLEB, A. Erliiuterungen, 1. c. p, 364. 



