PREFACE. Vll 



serial position assigned to the Gymnospermce. in 1861, which has 

 persisted in those Colonial Floras published since that year. 



The acceptance of the modern view as to the relationships of the 

 Gymnosperfnce no doubt justifies the action of those who would 

 modify that serial position. But what may be lawful is not neces- 

 sarily expedient. Considerations of practical convenience render 

 it desirable to adhere in this work to the serial arrangrement adopted 

 in 1861. The maintenance of that arrangement derives considerable 

 support from the action taken by Bentham and Hooker in the 

 " Genera Plantarum " in 1880. These authors, while presenting 

 the Gymnospermce in such a fashion as to indicate that this gioup 

 is not to be regarded as similar in status to the Dicoiyledone.s and 

 the Monocotyledones, have nevertheless accorded it a serial position 

 between the two classes of Angiospermcp. In following that example 

 here, an adequate safeguard against any possible misunderstanding 

 is provided by the subjoined synoptic statement, prepared by Dr. 

 Stapf, wherein the natural position and the divisions, according to 

 modern views, of the Gymnospermce that occur in tropical Africa, 

 are clearly indicated : — 



Division PHANEROGAM/E (Siphonogamge). 



Subdivision A. ANGIOSPERMM. 



Class I. Dicotyledones. 



Class II. Monocotyledones. 



Subdivision B. GYMNOSPERM.^. 



Class I. Gnetales. 



Order i. Gnetace^. 



Class II. Coniferales. 



Order i. Pinace^. 



Order ii. Taxace^. 



Class III. Cycadales. 



Order i. Cycadace^. 



The admission of the Gym,nospermce to the rank of a pliylum 

 distinct from that to which the rest of the Phaneroqamce belong, 

 creates another difficulty which is harder to settle than the one 

 disposed of above. This difficulty is concerned with the appli- 

 cability to the Gymnospermce of the conventional terminology applied 

 to the " flower." Attempts to maintain the old uniformity of 

 terms have been frequent ; these attempts have proved far fro)n 

 successful. Those who lay most stress upon function have found 



