MAGNOLIACEJE. 137 



The Talaumas^ are Magnuliaii, in wliicli the carpels, instead of 

 dehiscing longitudinally along the back, separate by the base from 

 the common axis of the fruit, or only open for a short extent above 

 and internally, or else become woody and quite indehiscent, or 

 fleshy and pulpy, only freeing the seeds by their putrefaction. 

 This last condition is especially found in Aromadendron'; indigenous to 

 Java. The Biiergeriai which come from Japan have also a 

 similar fruit ; but the leaves of the perianth become more nume- 

 rous, as also occurs sometimes in the true Maf/noUa.s. All these 

 differences have seemed to us of little importance^ and insufficient 

 to warrant the removal of these plants from Magnolia, otherwise 

 than as sections of the genus. If we analyse the magnificent 

 flowers of M. iasigim Wall.," we see that they terminate the 

 branches, as in M. grand'ifora, and are constructed exactly the same 

 way. But on opening the carpels we find in the internal angle 

 from two to ten carpels and upwards. Hence in the fruit the 

 carpels which dehisce dorsally often set free upwards of two seeds. 

 This character is also found in four or five allied species, which 

 have been collected into a special genus under the name of j\[an- 

 glietia,^ but we do not retain this genus, for the same reasons which 

 have led us to leave Magnolia Figo with biovulate cells, and 

 Michelia Chanq^aca whose carpels are multiovulate, in the same 

 generic group.'' 



Wuss., Gen., 281. — DC, Syst., i. 460; us (toe. c;7.), in r. »Hr^«A;7/5 ; we may also see it 



Pfodr., i. 81. — Hl., Fl. Jav., xix. 29, t. 9-12. — souietimes in the fruit of Ji. filiifera (JI. pidnila 



Spach, Suit, a B'fjf., vii. 447. — Endl., Gen.,n. Axdr. ; M. CocoDC, Si/st. Veff.,i.459; Hance, 



4735. — Walp , i^e^^, i. 69; ^«»., iv. 41. — B. H., Ami. Sc. Nat., ser. 5, v. 205; — Liriodendron 



Gen., 18, n. 3. — H. Bn., Adansonia, vii. 669. — UlUfera L. ; — L. Coco LotjR., F!. Cochinch. 



Gwillimia B-OTTTu., ex HvACB, loc. cit. — Blumea 1790), 347; Givilllmia iiidica Hottl. ; — Ta- 



Nees, Flora (1825), 152. lauma pumila Bl., Fl. Jav., lot: cit., t. 12, C). 



- Bl., Bijdraj., 8 ; Fl. Jav., xix. 25, t. 7, 8. — We know that the Magnolias strictly so-called 



Spach, Suit, a Buffon, vii. 451. — Exdl., Gen., of the section Yulan present nearly all these 



n. 4736. The calyx is here sometimes tetra- variations in the way some of their carpels dehisce, 



merous, and the number of petals may be raised though in these the dorsal cleft is usually more 



to thirty. The fruit is often characterized by its marked. But we cannot see generic distinctions 



dehiscence : " carpelUs nan nisi pidredine ah axi in all this. 



secedentibus" {li. R.,loc. cit.). Seep. 134, note 1. ■'' Tentam. Fl. Nepal., t. 1; Plant. Asiat. 



3 SiEB. & Zucc, Fl. Jap. Fam. Nat., i. 78, Marior., ii. t. 182. 



t. 2.— Endl., Gen., supp. v., n. 4736 '.— B. H., « Blume, Bijdraj., 8 ; Fl. Jav., xiv. 20, t. 6. 



Gen., 183. — Miq.,^h». Mus. Lvrjd. Bat.,\\. 257. — Endl., Gen., n. 4738. — Hook. F. & TiioMS., 



■* See Adansonia, vii. 6. When the carpels Fl. Lid., i. 76. — MiQ., Fl. Jnd.-Bat., i. p. post., 



of a species like M. Plumieri {Talauma Plumieri 15. — B. \l.,Gen., 19, n. 5. — Waip., Ann., iv. 40. 



Sw.) separate in masses from the common axis, *" See Adansonia, vii. 5, where we fornnilato 



we see each carjjcl opens more or less widely into our conclusion thus ; — " The Manijlictias are to 



two lateral halves, l)eginning at the internal the Maynolias as the multiovulate Michdias arc 



angle. The same fact occurs, as observed by to the biovulate Michvlias." 



