602 REVISION OF AEAUCAEIOXYLON KNOWLTON. 



synonymy is also iu some cases wrong. The following paper embraces 

 fifty species, the additional ones having been mostly detected since the 

 publication of Kraus's paper, although a few are evident omissions 

 fiom it. 



As above mentioned, it was originally intended to present in this 

 paper only the descriptions and principal references of the well-known 

 forms, and bring it as nearly as possible up to date, by including the 

 species that had been described since the publication of Kraus's paper. 

 But early iu this work it became evident that the investigation of Pale- 

 ozoic woods by Grand 'Eury and Eenault, and later by Morgenroth, 

 Vater, and Felix, have made certain changes necessary in the genus 

 Araucarioxylon as defined by Kraus. These investigations show very 

 conclusively that it must be divided into two parts, and the later 

 studies of Felix make it equally clear that the present state of our 

 knowledge is sufficient to amply justify its division into three parts. 



As the evidence which had led to the separation of Cordaites from 

 Araucarioxylon is of considerable interest, it is presented somewhat in 

 extenso. 



CORDAITES Unger. 



The name Cordaites was first employed by Unger in 1850.* It was 

 founded principally upon the leaves, and Unger admitted but a single 

 species (C. borassifolius), which had previously been described by Stern- 

 berg t (1820-'25), first as Flabellaria borassifolia and later as Cycadites 

 palmatus.% Sternberg, it will be observed, had regarded the first as a 

 palm, the latter as a cycad. 



The first to investigate the internal structure was Corda,§ who 

 in 1845 created the Flabellariacese for the species of [Sternberg. He 

 devoted a large plate to the elucidation of the internal structure, but 

 as his specimens had not been well preserved he fell into error in his 

 interpretation of the histological elements. He compared them in 

 habit to Aletris and Dracaena, but regarded the internal structure as 

 similar to Lomatojioyos of the Lycopodiaceae, and Unger, in his work 

 above cited, placed Cordaites iu the Lycopodiacese on this account. 



In 1848Germar|| obtained a second species from Wettin iu Germany, 

 which he referred to Flabellaria observing that it was very closely 

 allied to F. borassifolia of Sternberg. Unger, however, regarded it as 

 a true palm, and retained it iu the genus Flabellaria, although subse- 

 quent investigation has shown that Germar was correct iu regarding 

 it as allied to Cordaites. In the following year Bronguiart published 

 his celebrated " Tableau des generes de Vegetaux fossiles,^ in which 



* Gen. et sp. Plant, foss., p. 277. 



t Versuch., I, fuse. 2, p. 27, PI. xviii ; fuse. 4, p. 34. 



\ Op. tit., I, fuse. 4, p. 32, PI. xl. 



§ Beitriig. Fl. d. Vorweldt, p. 44, Pis. xxiv, xxv. 



|| Verstein. d. Steiukohlgeb. Wettin u. Lobjun, p. 56, PL xxm. 



IT In Dictionnaire univ. d'Hist. Nat., Vol. xnr, pp. 113-115. 



