ANALOGIES AND AFFINITIES 43 



portance — something merely incidental to tlic genus or the 

 family. It will be curious at least — i)er]iaps more than a 

 mere gratification of curiosity — to note instances in Avhich 

 the organographical type of a genus or family breaks (lo\vn 

 on all hands, changing into something else, and bidding 

 defiance to the formulated character at many a point, while 

 the common property alone remains as the one absolute 

 peculiarity of the group, pervading every species in it, and 

 being conterminous with it. Quite such a family as this we 

 have before us in that extremely natural order, the Cruciferjio. 



The morphological type of this order is most characteristic 

 and distinct from all others. The six tetradynamous stamens, 

 along with the peculiar bilocular and bivalvular fruit concur 

 nowhere else in nature. If silique and silicle had been con- 

 stant throughout the family, its character as a natural order 

 could have been drawn up in one or two brief sentences. 

 That framed by Jussieu (Gen. 237) a century ago, com- 

 prised in a few short lines of large type, would have been 

 abundantly sufficient. But that of Bentham (Benth. & Hook. 

 Gen. 57), some seventy years later, occupies, the formm 

 ahnormes included, a very large page of small type, yet falls 

 considerably short of embracing all the odd forms of pericarp 

 now known within the order. 



The true silique, happily the prevailing type, retains all 

 the essentials of its structure even while disguising them more 

 or less effectually throughout a long series of shortened or 

 siliculose forms, cylindrical, globose and double-globose, 

 compressed and obcomin-essed, immarginate and winged, 

 cordate and obcordate, triangular and fiddle-shaped. Then 

 again in certain genera, its valves are sealed and it becomes 

 aa altogether indehiscent fruit, siliquiform in outline only 

 {Raphanns) ; or the whole breaks transversely into indehis- 

 cent one-seeded joiuts (certain species in several genera) ; or 

 the joints are two only and very dissimilar (Crambe and 

 li^ajnstrnm). In many a genus we have fruits altogether 

 nut-like in their conformation, that is to say, one-celled, one- 

 seeded, indehiscent, with pericarps of leathery, cartilaginous, 



