ANGIOSPERMS. 



525 



stamens are branched, this is indicated by the signs being grouped, as in Fig. 377, where 

 the five groups correspond to five branched staminal leaves. The gyuceceum is treated 

 as a simpHfied transverse section of the ovary, since it is thus most easily distinguished 

 from the other parts ; the marks within the loculi of the ovary indicate the ovules, 

 which however are only represented in those cases where their actual position can be 

 expressed by so simple a plan. The size, form, and cohesion of the separate parts are 

 not taken into account at all. The construction of these diagrams is based partly on 

 careful investigations of my own, but chiefly on the studies of Payer in the history of 

 development (Organogtnie de la fleur), as well as on the descriptions of other authors 

 (Doll, Eichler, and Braun). 



I draw a distinction between empirical and theoretical diagrams. The empirical dia- 

 gram only represents the relative number and position of the parts, just as a careful 

 observation shows them in the flower; but if the diagram also indicates the places 

 where members are suppressed— which can only be determined by the history of de- 

 velopment and by comparison with allied species, especially if it points out relationships 

 which are entirely the result of theoretical considerations— I call it a theoretical diagram. 

 If the comparison of a number of diagrams shows that, although empirically different, 

 they nevertheless yield the same theoretical diagram, this common theoretical diagram 

 may be termed the type or typical diagram according to which they are all constructed. 

 I consider the careful determination of such types an important problem, the solution of 

 which may be extremely useful in the classification of Angiosperms. When the type has 

 once been ascertained, the theoretical diagrams which correspond to it may be treated 



FlC. -?7S.— Diatjrnni of the flower of a Grass ; /I Darnbiisa ; n of liiost Grasses; C of Nardus (from Doll, Flora von Baden, 



vol. I, pp. 105, 133). 



as derivative forms from which particular members have disappeared, or where they have 

 been replaced by a number of members. From the stand-point of the theory of 

 descent the type corresponds to a form still in existence or that has already disap- 

 peared, from which the species to which the derivative diagrams belong have arisen by 

 degeneration (/. e by abortion^) or by multiplication of the parts. 



A f<;\v examples will explain this. The flower of Grasses which is seated among the 

 pales may be deduced, as is shown in Fig. 378, on the theory of the abortion of certain 

 parts from the typical flower represented in Fig. 375? which is itself the typical diagram 

 of Liliacea3. A is the diagram of Bambusa, which only deviates from the type in 

 the absence of the outer perianth-whorl which is indicated by dots. But in most other 

 Grasses {E) the posterior leaf of the inner perianth-whorl (this whorl appearing generally 

 only in the form of small colourless scales), the whole of the inner whorl of stamens, and 

 the anterior carpel, are also wanting. In Nardus again (C), the anterior carpel only is 

 present (as far as the pistil is concerned); all the absent parts are represented by dots, 



' The construction of the diagram itself shows that the theory of abortion is justified even 

 where the earliest state of the flower-bud gives no indication of the absent member if the 

 number and position of the parts present point distinctly to such a hypothesis. If the idea of abor- 

 tion in this sense is not admitted, neither can the increase in number of individual parts, or their 

 replacement by several, be allowed. It is only the theory of descent that gives a rational ex- 

 planation of either fact, and that a very clear one. 



