414 FOURTH GROUP. SEED-PLANTS. 



to five branched stamens. The gynaeceum is treated as a simplified transverse section 

 of the ovary, because it is thus more readily distinguished from the other parts of the 

 flower ; the smaller or larger dots inside the loculi of the ovary indicate the ovules 

 but these are given only where their position can be correctly indicated in so simple a 

 form of diagram. No indication is in any case given of the cohesion, size, or form of 

 the separate parts 1 . The construction of these diagrams is founded partly on the 

 careful investigations of Sachs, but chiefly on Payer's studies in the history of develop- 

 ment and on the descriptions of other authors (Doll, Eichler, Braun). 



Sachs distinguishes between empirical and theoretical diagrams ; the empirical gives 

 only the relative number and position of the parts, as they are found at once by 

 careful examination of the flower ; if the diagram also indicates the place where 

 members are wanting, as can be proved by the history of development and by com- 

 parison with allied plants, and if it contains indications of circumstances which are 

 suggested by purely theoretical considerations, he calls it a theoretical diagram. If 

 the comparison of a number of diagrams shows that though empirically different, they 

 yet yield the same theoretical diagram, he calls this common theoretical diagram the 

 type or typical diagram, according to which the others were formed. The determination 

 of the type facilitates the recognition of the mutual relations of the individual floral 

 structures in a group of allied plants. But it must not be forgotten that such a type 

 is purely artificial, a plan obtained by logical combination and abstraction. It has to 

 be asked in each separate case, whether, taking our stand on the theory of descent, we 

 are justified in regarding the type as a still existing or now extinct form, from which 

 the flowers with ' derivative ' diagrams have been formed by abortion or cohesion of 

 particular members 2 . Experience shows that the neglect of this consideration often 

 leads to more or less forced interpretations of the forms of flowers. 



A few instances must first be given, in which the typical diagram is actually to be 

 regarded as the original form of the flower in a group of allied plants. 



It is to be so regarded for example in the Scrophularineae. In this order the flowers 

 are typically pentamerous, as we see in Verbascum (Fig. 343 A). But the calyx does 

 not show the number five in all species ; the posterior sepal has entirely disappeared ' 

 in Veronica and Lathraea (Fig. 343 Z>, E). In some species of Pedicularis and 

 Veronica (e. g. Veronica latifolid) it is still perceptible as a minute tooth ; in 

 genera in which it has disappeared, the position of the other sepals and the structure 

 of the corolla which has often five parts (Fig. 343 E) point to the fact, that the existence 

 of a fifth sepal must be assumed in the theoretical diagram even where the early stages 

 of development no longer show any indication of it. The corolla is generally penta- 

 merous, but the two upper petals are often united into one, and it then appears to be 

 formed of four parts as in Veronica. The posterior petal shows by its greater breadth 

 that it occupies the place of two petals, but there is here no question of a cohesion. The 

 external configuration of the mature corolla is of less importance for the diagram, but there 

 are interesting points in the stamens. In some species the five stamens are all equally 

 developed and all fertile, i.e. provided with anthers having pollen-sacs (microsporangia). 

 In others the posterior stamen is sterile, rudimentary or suppressed (Fig. 343 J3, E}. In 

 Gratiola (Fig. 343 C), the posterior stamen is suppressed, but the two obliquely anterior 

 ones are at the same time sterile, being developed merely as staminodes. Lastly, in 

 Veronica (Fig. 343 D] these are entirely suppressed. Other variations, such as the 

 sterility and sometimes the entire suppression of the two obliquely posterior stamens 

 need not be enlarged upon here. The diagrams show that the structure of the ovary is 

 less liable to variation. 



The diagram of the Liliaceae is the typical diagram of many Monocotyledons. The 

 Orchideae depart from it to an antonishing degree, but even they may be referred to it 



1 A diagram showing these features is given in Fig. 343. 



2 The theory of types is also much older than that of descent. 



