134 



NATURAL HIS TORY OF PLAINTS. 



of the consistency and colour usually presented by petals. More- 

 over, the fruit, instead of having all its elements close together 



in an ovoid mass, as in 3/. 

 grandifora (fig. 16S) has a more 

 elongated axis, more or less bent 

 on itself (fig. 172), so that its 

 carpels are further apart ; some 

 of these do not attain their full 

 development.' It is however in- 

 contestible that on reviewing the 

 fruits of all the Magnolias kno\\Ti, 

 we find every possible transition 

 between the form in M. Yulan, 

 and that of ^[. grandiflora: In 

 many species, too, the number 

 of pieces in the perianth is in- 

 creased to a variable extent, 

 either normally or through cul- 

 tivation, while the form and 

 colour of these parts vary 

 greatly, without any value being 

 assignable to these characters. 

 But the flowers always termi- 

 nate the branches, and there is 

 of the stamens and that of the 



Magnolia Yulan. 



Fio. 172. 



Fruit. 



no interval between the insertion 

 carpels on the receptacle. 



Kot so in the small flowers of 



Magnolia Figo^ (f'gs. 173, 174), 



' Tlic form ami length of tlie receptacle in the 

 ripe fruit are very variable. It is sometimes so 

 short an only to bear one fertile carpel ; at other 

 timea nearly Htrai;,'ht, or slightly bent, or like a 

 book, as in fig. 172 ; or even bent twice on itself 

 into un S, like the stock of a llistort. In the 

 fniiUi of this group, some carpels oj>en along the 

 whole length of the baek, others only open lialf- 

 way down ; othern, again, arc partly iletaehed 

 from the receptacle along the inner angle, down 

 which tlie cleR extends from the back. We 

 find here, in tine, every jxisHible transition between 

 the dehiscence of .1/. i/randijiorn and that of 

 Talauma, in which the car|R>lH Ke))amte from the 

 axis, and otdy open along a variable extent of the 

 internal angle. It also happens, in certain species, 

 that several neiglibouring carpels are united 



laterally, and come off together in irregular 

 flakes. This is very well shown in T. fragrant ia- 

 sima, in plates ccix., ccx. of Hookek's Ivviifs. 



" Thus the fruit of M. CamphtUii Hook. F. 

 A- TiioMS. {Fl.Iml.,\. 77), represented in the///. 

 /'/. Himal. (t. 4), and reproducetl in the Flore 

 dm iSerres (t. 1282-12S5), has also the coniad 

 fruit of M. graiulijlora, but much elonpitwl, and 

 apjiroaching the cylindrical form foinid in Yulan 

 and the allied species, from which it further 

 differs in being nearly straight. 



* M.faxciafa Vknt., Malmaui., n. 24, not. — 

 M. fu»cala Andr., Bof. Hep., t. 229.— Zirio- 

 dendron Figo Loitk., /7. Cochinrh., ed. W., i. 

 424. — LiriopKl.ifii.ira fa Sl'ACIl, A'miV. <J Jinjf'oit ; 

 vii. 4(31. - Michelia Hanck, Ann. Sc. ^'at., 

 her. 5. V. 2li5. We have elsewhere reiuurkvil 



