LINNEAN SOCIETY OF LONUON. 57 



characters be mentions eitlier liad then or have since been shown 

 to be not peculiar to the Auientales ; but the point 1 wi.sb here 

 especially to criticise concerns the origin of the berniaphrodite 

 Hower. Evidently he is of the opinion that in the phylogenetic 

 sense a unisexual strobilus can revert to the bisexual state, and 

 lie cites in his support the occurrence of androgynous cones in 

 the Couiferae. These in my estimation have merely a teratological 

 significance. This may seem to be begging the question, so I 

 state my argument as follows. Until a new species of (say) Pinus 

 be discovered, which normally bears androgynous cones and which 

 is except for this peculiarity a true pine, 1 decline to attach any 

 phylogenetic importance to these freakish cones. Evidence has 

 yet to be produced to show that a species with berniaphrodite cones 

 or flowers has ever arisen from one bearing unisexual fructifi- 

 cations. Eor the converse, the evidence is overwhelming. 



Viewed in the light of our theory the origin of the hermaphro- 

 dite flower or cone presents no difficulty. The two kinds of 

 sporophylls are regarded as caught up together into one and 

 the same strobilus from the lax I'teridospermous arrangement. 

 The unisexual condition has resulted from the abortion of one 

 kind of sporophyll in the strobilus. The Englerian can, of 

 course, maintain that the Amejitifera? have been unisexual from 

 the beginning ; but then he must confess to at least a diphyletic 

 origin for Angiosperms, and must refrain from formulating any 

 direct relationshi|) between the catkin-families and hermaphrodite 

 flowering plants. But this he declines to do. Engler's system, 

 in fact, is based on a kind of general and hazy idea that naked 

 unisexual flowers of few parts are primitive, and that from these 

 have gradually emerged by a series of steps the fullv-equipped 

 hermaphrodite flower with both calyx and corolla. JVo attempt 

 is made to trace by means of examples how this evolution has 

 come about. 



Let us briefly glance at the families which may now be 

 considered to compose the Amentiferiv. The Salicales can no 

 longer be included. The group then is narrowed down to the 

 Englerian cohorts, Juglandales, and Eagales, with the addition of 

 Casuarlna. There is a general consensus of opinion that 

 Casuarina is fairly closely related to the Betulacece (12, 25). 

 There are also reasons for regarding the Juelandales as having 

 aflinities with the Eagales. Hemsley's new family, the Julianiacea) 

 (23), was looked upon for a time as a link between the Juglanda- 

 cea8 and the Anacardiacese; but probably the Jnlianiace;c should 

 more correctly be considered as merely reduced Anacardiaceoas 

 forms with no real relationship to the Juglandales (19). A 

 Eosalian origin for the Amentifera? as a whole would appear to be 

 the more |)lausible view. There maybe some element of truth in 

 Hallier's original suggestion of deriving the catkiii-families from 

 the TrocliO(lendrace;e through the Jlainamelidacea^ 



In recent years perhaps most stress has been laid on the 

 structure of the wood as pointing to the primitiveness of the 



