510 NATURAL EISTOEY OF PLANTS. 



estimate the number of species at about three huhdreS. They are 

 plants of the cold and especially temperate regions of the northern 

 hemisphere, particularly 6f the old world, less abundant relatively in 

 N. America. S. America, chiefly in the western and Andean regions, 

 is rich in Valeriana, Plectritis a^d Astrephia. There are few Valeri- 

 anacece in the eastern part of S. America and the Antilles, or in tropical 

 Asia. The genera Nardostachys and Patrinia belong to the central 

 temperate regions of Asia and the extreme East. The Valerianacece 

 observed in southern Africa are introduced Valerianellas and Valeriana, 

 capensis, of the indigenous character of which doubts have been 

 expressed. Not a plant of this family, it is said, is known to be 

 spontaneous in Australia. Europe possesses only the three genera 

 Valeriana, Valerianella, and Centranthus. 



Affinities. — The Valerianacece have naturally close affinities with 

 the Dipsaceoe, since they have been ranged in the same family. They 

 are nearly always distinguished by their tricarpellar gynsecium aud by 

 the absence in the Dipsacece of the rudimentary cells with or without 

 aborted ovules. The albumen in the seed of the Dipsacece is com- 

 pletely wanting or nearly so in the Valerianacece ; the latter, though 

 their flowers may be accompanied with bracts morfe or less united or 

 accrescent, have not the true involucel which surrounds the flowers of 

 the' Dipsacece. A character is also not unreasonably derived from the 

 odour, ordinarily foetid and easily recognizable, possessed by the 

 Valerianacece. As the corolla of the latter is almost always irregular,' 

 and as the stamens are always inferior in number to the divisions of 

 the corolla,^ we shall find the types of the Buhiacece most analogous 

 ,to the Valerianacece not among the first twelve series of that family, 



> On the disposition of the parts, see Pater, covering. In CenirantMis, the tube which en- 



OrQanog. t. 130. — EiCHii. Stuthendiagr, i. 275. closes the stamen and style corresponds, not to 



This disposition is always easily derived from this lobe, but to lobe 1 of the quincunx, that of 



that of the flower of Nardostachys, as given in the two posterior which covers the other. The 



fig. 398 (p. 505). If ormally the anterior divi- fertile cell of the gynsecium is never in the plane 



sion of the corolla, corresponding to the spur of the symmetry which would pass through the 



when it exists, is covered by the two adjacent middle of the spur, but is lateral and normally 



lobes, and these again by the two posterior. situate on the side on which the stamens are 



There are frequent anomalies ; but the anterior most numerous. 



lobe ofFedia Cormwopiie, corresponding to the ' The androecium is also derived from that of 



gland of the tube, is normally the same though Jfardostaehys, where, with five lobes to the 



EicHLER [loc. cit. fig. E) has represented it as corolla, there are four stamens. That which 



