102 Mr. B. Clarke on the Structure and Affinities of 



same catkin ; and in one instance nearly forty were all attached 

 to one pedicel-like torus, as though the filaments had become 

 monadelphous ; when numerous, some of them were very im- 

 perfect, but the greater part of them had the usual double ap- 

 pearance from their deeply divided connective. It is not intended 

 to allude to each cell as constituting one stamen, however distinct 

 they may be ; and even in Comjdonia, which has been described 

 as having six stamens united in pairs, I believe there can be no 

 doubt that there are only three, the connective being rather more 

 deeply divided than in Myrica, but each lobe being strictly one- 

 celled, as in that genus*. 



But in other parts of their structure jNIyricacese so nearly 

 agree with Urticacese and Cannabinete as to form a connecting 

 link between the Amental and Urtical families, approaching the 

 Urticacepe in their erect ovule and occasionally succulent calyx, 

 there being also in Urticacese a tendency to unite the carpels^ 

 when two are present, by their margins, as may be seen in the 

 barren ovaries of Dorstenia ; and showing some degiee of affinity 

 with Cannabincte in the calyx being either free, as it is in Hu- 

 mulus, or partially adherent, like that of Cannabis (PI. VI. fig. 8)t, 

 and in the inflorescence being frequently covered with resinous 

 glands having a cellular appearance like those of Humulus. 



Plataxe.e. 



This is also one of the Amental families, the ovary of which 

 is understood as consisting of a smgle carpel, which has doubt- 

 less occasioned the structure of the flower to be overlooked, so 



* The anther of Myrica has been described as either two- or four-celled; 

 but the four-celled anther I should suppose must refer to a deep furrow at 

 the line of dehiscence, which makes the half-anther appear two-celled. 



t That the same membranous cup which is present in the ftniit of Hu- 

 mulus, loosely surrounding it, exists also in Cannabis, but adherent, is 

 shown by the fruit of a variety of Hemp formerly in cultivation in the 

 Botanic Gardens of Chelsea under the name of Cannabis indica, in which 

 the u]iper margin of this membranous cup or calyx is frequently free, 

 forming a shallow, semitransparent ring, as occasionally hap])ens in the 

 fruits of Cupuliferae, where it is not usually developed, as in the common 

 Nut. In an early stage of the flower it is discernible in the common Hemp 

 (fig. 8), and it may sometimes be easily separated as an entire, colourless, 

 semitransparent, membranous cup, agreeing v.ith that of Humulus in being 

 higher on the postei'ior side, — a character which occurs also in Urticacea;. 

 In some few flowers it may be found loose, being nearly detached, and in 

 a somewhat monstrous flower with three stigmas this calyx was free and of 

 a greenish colour, having two or three green ribs, the bractlet enclosing it 

 as usual. In this view of the structure of Cannabiuca;, the sepal or peri- 

 gone as usually understood becomes a bractlet, of which there is then one 

 enclosing each flower, and which in Humulus forms the large, permanent, 

 scale-like envelo))e of the fruit. 



