CLASSIFICATION OF PLANTS BY EVOLUTION. 



137 



regular flowers, especially as terminal ones, under cultivation. ' Such 

 flowers then resemble the normally regular ones Thus Salvias produce 

 regular trumpet- shaped blossoms ; members of Leguminosce, instead of 

 having the usual papilionaceous corolla, produce flowers like those of 

 the Rose family &c. Such are regarded as "reversions." 



For theoretical considerations as to how irregular flowers came about, 

 I must refer the reader to what I have said elsewhere. 



6. With regard to alterations in symmetry, I have referred to the very 

 general reduction of carpels, so that while the outer whorls may be 

 quinary, the pistil may be, and often is, reduced to two carpels, as in so 

 many of the Gamopetalee. 



But it often happens that the principle of compensation comes into 

 play, so that while some parts degenerate and vanish others alter in size 

 or form. Thus while in Mulleins (Verbascum) we find the size of the five 

 stamens decreasing from the front to the back of the flower, in most of 

 the genera of the same order (Scrophularineee), as well as of Labiatee, 

 the fifth or posterior stamen has become totally arrested, and the two 

 posterior stamens are with rare exception shorter than the two anterior. 



Again in the minute yellow disk florets of a Daisy there are both 

 stamens and pistil, but in the larger white ray florets (though the corolla 

 has really lost two petals in the change, the other three being greatly 

 enlarged) the stamens are totally arrested, those florets being female 

 only. 



7. As special instances of "arrest " it may be seen how the calyx is 

 more or less reduced or vanishes when flowers are borne in clusters. 

 Such occurs in British Rubiaceee, Dipsacece, Composites, Umbelliferee, 

 Rhododendrons, &c. On the other hand in the Incompletes it is the 

 corolla which is usually wanting ; we then generally find the stamens 

 opposite the sepals, indicating a lost intermediate whorl of petals, as in 

 Nettles, Daphne, &c. Such are believed to have been derived by pro- 

 cesses of degradation from orders having petals. Thus Nettles are allied 

 to Mallows, Daphne to Roses, &c. 



In Cornflower (Centaur ea) the ray florets have their corollas much 

 more greatly enlarged ; but this is done at the expense of the pistil as 

 well as the stamens ; hence it is neuter and sexless. A similar process 

 is seen in Guelder Rose and Hydrangeas. 



The " cohort " or group of orders known as Amentales,* from the 

 general presence of "catkins " (amentum), appears to represent primitive 

 forms. Alder is the only one which seems to have arranged its bracteoles 

 of the male flower into a whorl ; all others have nothing to indicate a 

 calyx except perhaps Poplar. Of these one is inclined to regard 

 Casuarina (the beef -woods of Australia) and our Sweet Gale (Myrica), 

 of which there are several species in South Africa, as representing the 

 earliest types, though the actual links with Gymnosperms are absolutely 

 wanting. 



All other members of the Incompletes are probably degradations from 

 plants which possessed both calyx and corolla. 



* This group includes the following orders: Betulacecs (Birch and Alder), 

 Platanece (Plane), Myricacece (Sweet Gale and Comptonia), Casuarina (Beef-wood), 

 and SalicacecB (Willow and Poplar). 



