MR. D. OLIVER ON SYCOPSIS, A NEW GENUS OF HAMAMELIDE^. 85 



anthers of Sycopsis I have not been fortunate enough to find subsequent to dehiscence ; 

 they are very similar before bursting to those of Distylium and the African genus 

 Trichoclaclus. 



In connexion with this notice it is scarcely within my province to inquire at length 

 into the general relations of the Samamelidecs. Excepting a somewhat anomalous mode 

 of dehiscence in the anthers of a few species (at least of Mistigma, Tetrathyria, and Hama- 

 melis chinensis), and the singular stipules of BnchlancUa, the order does not offer any 

 very salient peculiarity, neither does it afford any well-marked type of structure pervading 

 its members. Mr. Griffith, in his 'Observations on Cantor's Plants*,' and Tulasne in 

 " Fragmenta Ploree Madagascariensis f," remark the rather complex character of the affi- 

 nities of the order, and the variety in the structure and disposition of some of the floral 

 whorls presented by its various genera. The more characteristic features of the Hama- 

 melide^B may be said to reside in their arborescent or frutescent growth ; the prosenchyma 

 bearing discoid markings from the presence of minute intercellular lenticular cavities ; 

 leaves nearly invariably alternate and stipulate; perigynous or epigynous and definite 

 stamens in nearly all of the hermaphrodite species ; the more or less inferior bicarpel- 

 lary and bilocular ovary (in Distyliimi quite superior) with distinct styles ; the oviiles 

 solitary and pendulous, or in the pluriovulate genera, with the exception of one or two 

 in each loculament, mostly abortive. The order presents in these characters much in 

 common with Cornacem (including ^to«^iert? and Nyssa), near which Dr. Lindley disposes 

 it. Of this order it may not improbably be regarded as a section, especially as a Tri- 

 chocladus of South Africa and a Bicoryphe described by Tulasne, from Madagascar, have 

 opposite leaves ; and as, further, in some genera we find the flowers closely aggregated 

 into a dense capitulum (as in Benthamia or in Nyssa %), and sometimes also provided with 

 conspicuous involucral bracts analogous to those of some species of Cormis (as in Bhodo- 

 leia and Parrottia) : the frequently tetramerous symmetry of the flower and the tendency 

 to a valvate sestivation of the petals favour this view of their affinity. With Cimoniacece 

 and the Saxifmgal alliance (as observed by Drs. Hooker and Thomson §) an affinity is 

 indicated by the curioiis stipules of Biicklandia, as well as by other of the characters 

 above enumerated. To Bruniacece the order is allied in some important particulars, as 

 noted by Eobert Brown and Prof. Lindley. 



tive, p. 375), that it might not improbably be considered generically distinct from H. virginica, on the ground chiefl_v 

 of the different structure of the anthers ; and Asa Gray remarks (' Botanical Jlemoirs ' [from Mem. Am. Acad, of Arts 

 and Sciences, new ser. vol. vi.], p. .390), that H. japonica (a species which I have not myself examined) is close to 

 H. virginica, " its only other strict congener " — implying, consequently, his sense of the importance of this distinction. 

 n. chinensis ought, perhaps, to be separated from the American species, under the name of Loropetalmn, suggested by 

 R. Brown (I. c. p. 375). 



The question maybe open whether Tetrathyria, with the same stamens, but having the petals represented by didy- 

 mous scales, ought to be reduced to the same genus with this plant. As Hamamelis stands at present, including both 

 the American and Chinese species, this feature in Tetrathyria must be held of less than generic value, estimated by 

 the range permitted in the yet more important particular of the structure of the anthers. 



* p. 24. t Annales des Sciences Naturelles, 4"'° ser. vol. viii. p. 142. 



X The elongate revolute laterally-stigmatose style of Nyssa recalls that of some HamamelidecB. 



§ Proc. Limi. Soc. vol. ii. p. 54. 



