192 



UMriEIJ.IFERyE 



central axis or carpiiplmrc, which is offen V-shapcd, and from 

 which, as they ripen, they separate from below upward. The 

 fruit is often flattened, either dorsallv, from front to back, that is, 

 or in a direction at right angles to the commissure, so that this 

 lateral plane is broad, or latiTallv, from side to side, or in a 

 direction parallel to the commissure, which is, therefore, narrow. 

 Each carpel or ?/tcricarf> is marl<ed by vertical riifi^rs, normally 9 

 in number, 5 prima>-\\ usually the more i)rominent, and 4 sccniuiayY 

 alternating \\itli them Of tire 5 primary ridges, the 2 nearest 

 the commissure are termed latently and are often expanded into 

 wings ; the midrib down the back of the carpel is the carinal or 

 dorsal ri,lge : and the two to the right and left of it are called 

 the uUennediak ruii:,cs.. I'lie sjiaces bt^tween the primary ridges 



Diaqram Section of Fruit o^ Uniljellilei". c, commi.^sure ; </r, dorsal or carina! ridge ; 

 /r, intermediate ridge ; h\ lateral ri<.ige ; f'r, printary ridge ; jr, seci.indary ridge : /; furrow ; 

 7', vitta ; JT', stjiitary \dtta ; ( r', commissural vitta ; _/.v, Hat seed ; ^-j-, groo\-ed Seed. 



are called furrows. In the wall of the fruit there are often 

 long narrow sacs called fv/A/' or s/ripcs containing coloured 

 essential oil. They generally occur betwi-en the ridges, rarely in 

 them. Normally there are 6 m each carpel, 4 between the 

 primary ridges and 2 in the comniissijral face ; but instead of 

 being thus solitary in each furrow, they may be j, 3, or more 

 together, or they may be fewer in number. There is one pendulous 

 seed in each carpel, containing horny ailn'inien and it is import- 

 ant to notice whether it is flat or groo*ved along the side facing 

 the commissure. The ridges, vittK, and seeds are best examined 

 by a lens on a cross seclii^n of a fruit ; the so-called ("arawav-seed, 

 for instance, which is, accurately sjieaking, a mcricarp or carpel. 

 The order being, as has been said, a verv natural one, its genera 

 agree in almost all their more obvious characters, and can only he 

 accurately discriminated by minute differences, chielly in the ripe 



