226 BOTANICAL GAZETTE [skptember 



It will also be observed in the limitations of the classes of 

 Eusperms that the double fecundation of Nawaschin and Guignard is 

 accepted as a true fecundation. This position is capable of defense, 

 but even if later proved untenable the two classes of x\rchesperms and 

 Metaspernis are sufficiently distinct to justify the nomenclature. 



II. STRUCTURAL CLASSIFICATION. 



Seeds integrated with placenta. |A. Synspernis. 



(Insemin^es of Van Tiegheni/ e, g.y Loranthaceae.) 

 Seed integrated separately from placentaT B, Aposperms. 



(Seminees of Van Tieghem.) 



Under the head of structural classification should be mentioned 

 also the division of seeds into monocotyledonous and dicotyledonous 

 according to the character of the embryo, into albuminous and exalbu- 

 minous, according to the persistence of the endosperm, into ategmi- 

 nees, unitegminees, and bitegminees (Van Tieghem), according to the 

 presence and number of the integuments, etc., but these classes are 

 after all of comparatively secondary importance. They do not brmg 

 into view fundamental differences between seeds, but only incidental. 

 For example, the difference between the monocotyledonous and the 

 dicotyledenous embryo is probably by no means of fundamental 

 importance, but represents different adaptations of haustorial organs 

 during intras^minal life ; the difference between albuminous and c;xal- 

 buminous seeds arises through nutritive adaptations between embryo 

 and endosperm and is of secondary importance; the difference 

 between seeds with one or two integuments is after all a question of 

 indusial development and need not be given great weight. It is true 

 that these distinctions are of great value to the taxonomist and enable 

 him to classify plants to advantage. Yet in comparison with the dis- 



tinctions brought out below they are probably less significant in the 

 specific life. 



^The classification of flowering plants constructed by Van Tieghem principally 

 upon the basis of seed structure contains much of interest. Unfortunately it is not 

 natural, the characters taken for primitive being in many instances either vestigial or 

 derived. Thus it results in an inversion of relationships. This is well illustrated by 

 the position given to such a family as the Loranthaceae. On account of th(* high eco- 

 logical specialism of the plants herein included degenerate structures would be certain 

 to appear. It is precisely these that Van Tieghem accepts as primitive. This fact m 

 itself is perhaps a sufficient commentary on the taxonomic value of the Van Tieghem 

 sequences. 



^ 



ft 



i 



