172 NATURAL HISTOBY OF PLANTS. 



all Maynoliacca are more or less so, But this character, practically 

 useful though it be, is certainly of no great importance in itself; 

 and it has ceased to be absolute since the Eifjjfelra, which lack all 

 aroma, have been classed among Maf/noliacca. Nor is the direction 

 of the ovule of fundamental value in separating the two orders, 

 because a descending ovule with the micropyle exterior, as in 

 Magnolia, answers really to an ascending ovule with the micropyle 

 interior, as seen in the uni or pauci-ovulate Billcnxncpce. But here 

 again, in practice, as we as yet know no JJillcniacca with definite and 

 suspended ovules, we may assert that the ovules of Magnoliacea, 

 solitary and few in number, have the micropyle always external, 

 whether they be descending, as in the true Magnolias and Srhizan- 

 dra, or ascending, as in lUicium. In Billcniacea with pauciovulate 

 carpels the micropyle, on the contrary, looks inwards. 



Moreover we must give up the attempt to distinguish Dillcniacece 

 and Mar/iioliacca by the presence or absence of stipules, since the 

 tSchizandrca, lUicicce, and CaiicIIcce have no stipules, while certain 

 Wormias, Davillas, &c., as we have said,' possess petiolar expansions 

 which behave exactly like the organs called stipules in Magnoliacea. 

 Nor is the symmetry of the flower sufficient to separate the two 

 orders absolutely ; for if it is true that the flower of Dillemacca is 

 often on a quinary type, it is equally true that that of Magnoliacea 

 is far from being constantly composed of trimerous verticils. The 

 Billenian are almost Magnoliaccce, as no one can fail to see on an 

 exact analysis of their flowers. The quinary symmetry of the 

 perianth, the verticillate arrangement of the carpels, the spiral 

 insertion of the androceum," the stipuliform dilatations of the petioles, 

 are facts which are all met with in one or other of the types of the 

 MagnoViaceoi? These too are very near the Cafgcaiit/icce. It is true that 

 as yet we have found none of the Magiiuliaccce with a receptacle 



' Seep. 120, and Adansonia, vi. 271. tlioupli much less evidently (see Adaiisonia, vii. 



' Auinthcciweof lhe/if(i««Ho«/«<rrt', wcsliiillbo 'M'A ; viii. 12). 

 able to tiiko into Hccminl tlie develoimient of the ' We slmll not hero «poak of tlio aril, wliich is 



flowers in diHtinguihliin^ Muritiuliucea; from said to be lii^jliiy developetl in Dillenincrcr and 



JJillftiiticefF, as soon as the orj;anoj;eny of the absent in Mni/iioliiurte, lonsiderinj: that tlie aril, 



former lias bin^'n more completely studied. We as seen from our stand-iKjint, is not of the same 



may now say that in all llio Murinoluwea we conformation in the two pronps, but is rcnlly 



have as yet studied, tlie andrm-eum is develojHHl, more jjeneralized in MaijintHa than in CntuiulUa, 



not eentrifugally as in JUllminciir, but in a Ilibhertia, &c., all the supertii-ial celhi of the 



spiral order and cenlrip<tally. Tbis iMjculiarity former geims entering into its forniatiuu by their 



is very marke<l in Atoi/nolid and JJrimi/» ; it also hypertrophy (see p. l'.\2, note 7). 

 exists in Jllicium aninatum and parvijlorum, 



