Jantjaet 15, 1909] 



SCIENCE 



99 



roses {Primulales) , phloxes {Polemowiales) 

 to figworts (Personales) and mints (Lami- 

 ales). Another line was sketched from 

 roses (Eosales) to umbelworts {Umbell- 

 ales), madders {Bubiales) and sunflowers 

 (Asterales). Certain details of that ar- 

 rangement, as the disposition of Celastrales 

 and Sapindales, and the retention of the 

 Choripetalae and Gamopetalae as valid 

 gronps, were subsequently found to be 

 erroneous, and were corrected, but in the 

 main the system as then outlined has been 

 sustained by subsequent careful studies of 

 the families. 



Two years later^" this general arrange- 

 ment was expanded so as to include brief 

 characterizations of the orders, suborders 

 and families, and in it the Celastrales and 

 Sapindales were brought into the phyletie 

 line extending from Bosales to Asterales, 

 but the Choripetalae and Gamopetalae were 

 still recognized as valid groups. 



A year later, in an elementary text- 

 book^^ the Choripetalae and Gamopetalae 

 were abandoned as definite groups of angio- 

 spermous orders, since it is evident that 

 gamopetaly has been attained independ- 

 ently in at least two phyletie lines. 



In my presidential address^^ before the 

 Botanical Society of America in 1897, the 

 dicotyledons were arranged in "two some- 

 what diverging genetic lines or phyla, each 

 beginning with apocarpous, hypogynous, 

 ehoripetalous plants, and both attaining 

 syncarpy and gamopetaly, one remaining 

 hypogjTious, the other becoming epigy- 

 nous." A little later it is explained that 

 "since gamopetaly has evidently been at- 

 tained at more than one point, it is no 



" In Johnson's " Universal Cyclopedia," Vol. 

 VIII. (1895). 



" " The Essentials of Botany," sixth edition, 

 1896, p. 322. 



" " The Phylogeny and Taxonomy of Angio- 

 sperms," August 17, 1897. Published in Botanical 

 Gazette, Vol. XXIV., p. 145. 



longer desirable to retain the Gamopetalae 

 as a distinct group." 



The latest restatement of this arrange- 

 ment of the flowering plants was published 

 at the beginning of the present year in my 

 "Synopsis of Plant Phyla, "^^ already re- 

 ferred to earlier in this address. In it 

 many minor corrections were made, and 

 some suggestions were hazarded as to the 

 point of origin of angiosperms, and con- 

 ifers. These suggestions followed the line 

 sketched by Arber and Parkin in their 

 paper on the "Origin of Angiosperms"^* 

 published a few months earlier. Basing 

 their argimient upon the discoveries of 

 Wieland^^ these authors concluded that 

 angiosperms were derived from Cycadean 

 ancestors of the Bennettitean type, with an 

 open flower-like strobilus ("pro-anthostro- 

 bilus") of megasporophylls, microsporo- 

 phylls ("amphisporangiate") and asporo- 

 phylls ("perianth"). As a consequence 

 they arrive at the conclusion that primitive 

 angiosperms were necessarily polypetalous, 

 hypogynous and apocarpous, precisely the 

 conclusion reached by me on theoretical 

 grounds more than fifteen years ago, and 

 since then persistently held in the face of 

 the increasing popularity of Bugler's sys- 

 tem. It would now appear probable that 

 there must soon be another rearrangement 

 of the fiowering plants. "We have recently 

 witnessed the almost complete inversion of 

 the sequence of the families of flowering 

 plants in our systematic manuals, and it 

 appears now that we shall barely have time 

 to become accustomed to the new order 

 before we shall have to learn still another. 

 It seems now inevitable that such orders as 



^^ " A Synopsis of Plant Phyla," in University 

 of Nebraska Studies, October, 1907 (issued Feb- 

 ruary, 1908). 



" " The Origin of Angiosperms," by P. A. Newell 

 Arber and John Parkin, in Linnean Society's 

 Journal— Botany, Vol. XXXVIII., July, 1907. 



'° " American Fossil Cycads," 1906. 



