Mr. J. Miers on the Winteracese. 35 



racter is wanting in the Magnoliacece ; or, if sometimes present, 

 these dots are exceedingly minute and faint. In the latter family 

 the several parts of the flower spring from a broad and highly 

 conical torus, whereas this is extremely small in the Winteracece. 

 In the latter group the ovaries are generally few, and always in 

 a single whorl, sometimes reduced to two in number, or even 

 solitary ; in Magnoliacece, on the contrary, they are constantly 

 very numei'ous, being arranged imbricately in many series upon 

 a conical or almost cylindrical torus. The structure of the fruit 

 in this last-mentioned family affords a very characteristic fea- 

 ture, generally consisting of a large cone or ball of many-seried 

 aggregated capsules, more or less free, but sometimes forming a 

 solid syncarpium ; these capsules generally open by two valves, 

 each exhibiting one or two tolerably large seeds (covered by a 

 brilliant scarlet fleshy tunic), which fall out and remain sus- 

 pended each by a long elastic thread : in the Winteracece the 

 fruit is small, consisting of a few radiating carpels generally 

 distinct; in Illicium, somewhat two-valvular; but in Drimys 

 and Tasmannia, baccate, enclosing a few small, shining, black, 

 cochleate seeds, of a structure different from those of Magno- 

 liacece, and remarkably similar to those of the Canellacece : the 

 hard crustaceous tunic, hitherto mistaken for the testa, so con- 

 spicuous in the latter family and the M^interacece, presents a 

 striking contrast to the scarlet soft tunic, suspended by a long 

 thread, in the Magnoliacece : in the latter order the raphe is found 

 in this external fleshy coating, while the tunic beneath it is thick, 

 hard, and bony ; but in the IVinteracece and Canellacece the outer 

 coating is hard, brittle, and void of vessels of any kind ; the 

 raphe is seen in the second tunic, which is thick, soft and 

 spongy (analogous to the outer tunic of Magnolia), while the 

 coating next beneath it is thin and membranaceous. The cm- 

 biyo, in the former order, is situated in the axis of the albumen, 

 at the extremity farthest removed from the hiluni, and beneath 

 the apical chalaza : in the two last-mentioned families, the em- 

 bryo is excentrically placed near the rostrated summit of the 

 albumen, at no great distance from the hilum, and at a more con- 

 siderable interval from the lateral chalaza. There can be little 

 doubt, therefore, that the Winteracecu have far less afiinity with 

 the Magnoliacece than they have with the Canellacece; and it 

 appears to me that the reasons here given fully justify their 

 removal from the station hitherto assigned to them, and their 

 approximation to the last-mentioned family, in the manner I 

 now proceed to indicate. 



The circumstances that seem to connect the Winteracece with 

 the SchizandracecB should not be lost sight of in this investiga- 

 tion. There are certainly many points of affinity between them ; 



3* 



