264 MR. A, S. HOENE— A COXTEIBUTIOxV TO 



Coynparison uitli the CaprifoliacecB and Hamamelidacea. 



Kuntli * influenced by the resemblance between Corniis and Vthurnimi, placed the 

 Corneae, typified by Corniis, in the Caprifoliacese.' But the Corneoe were soon afterwards 



separated from the latter to form a new family, the Cornete ( De CandoUe \, Endlicher J) 

 or Oornacea3 (Lindley §J. Lindley considered that their habit and general characters, 

 and especially their polypetaly, justified this removal. Benthani and Hooker note that 



polypetalous condition is tlie only real d 



It is, in the writer's 



opinion, a valuable one, for the Caprifoliaceae are definitely gamopetalous, notwithstanding 

 tlie fact that the corolla is highly variable in different genera. 



In variance with Kunth, Lindley and Oliver j| are of opinion that the Cornese more 

 nearly approach the Hamamelidaceae. Oliver considers even that the latter family may 

 not improbably be regarded as a section of the former. Oliver is guided by general 

 morphological resemblances, such as arrangement of the leaves, the form of inflorescence 

 (capitate in Benthamia and N'l/ssa capitata), disposition of the stamens, the " inferior, 

 bicarpellary and bilocular ovary," the ovules solitary and pendulous, etc. But the facts 

 regarding the embryology of the Hamamelidaccte revealed in Shoemaker's ^\ paper on 

 Ilamamclis , and those deduced from the study of the ovule in the Cornacese, are against 

 the approximation of these families : in the Hamamelidaceas, the nucellus is bulky, it is 

 invested by two integuments and an extensive parietal tissue is formed, as in the 

 Saxifragaceae ; whereas, in Cornus, the nucellus is small, without or with very little 

 parietal tissue and invested by a single integument. 



The writer finds a close correspondence in the floral organization between Cornus 

 and the Sambuceoe, and Cornus and Samhucus may be profitably compared. The whorls 

 in >S'. canadensis are pentamerous, those in Cornus tetramerous M'ith the exception of 

 the ovary. In both genera, the loculi are uniovulate, the ovules terminal, axially bor 

 and tangentially orientated; the nucellus is very similar and invested by a sing 



• 



ith regard to the vascular structure, a single series of bundle 



pr 



m the ovary, dorsal branches pass into the style, and the vascular connections of the 

 ovule are similar. Kow the variability in the number of units contributing to the ovary 

 in Samhucus is accompanied by a corresponding variability in the number of loculi, axial 

 bundles, and styles, and the writer has expressed the opinion that the dorsal portions of 

 the carpels have become definitely incorporated as an integral and invariable portion 

 of the ovary-wall. Hence the ovary is not strictly hi-, tri-, or quadri-carpellary in hi-, tri-, 

 or quadrilocular ovaries. Passing to Cornus, the presence of an additional unit, as 

 shown by Clarke, is asserted by the appearance of an extra loculus, a perii3heral ovular 

 bundle and two additional lobes to the stylar canal, the structure of the ovary-wall 

 remaining unchanged, as in Samhucus \ so that it does not seem unreasonable to inter- 

 pret the organization In terms of that of Samhucus, nor to consider the ovary bicarpellary 



Kunth, Kov. Gen. et Sp. iii. 430. f a. de Candolle, Prodr. iv. 271. 



t S. Endlicher, Gen. Pi. (1S3G) 798. § J. Lindley, Yeg. King. 2iid ed. 



il D. Oliver in Trans. Linn. Soc, Bot. xxlii. (1 860) S3. f D. N. Shoemaker in Pot. Goz. : 



(1905) 



