MK. G. UENTHAM ON MYllTACE^. 121 



frviit, we can generally find a perfect seed to examine ; but in 

 Leptospermese it is generally only one or two of the uppermost 

 seeds of each cell that come to perfection ; and those fall out the 

 moment the capsule bursts, so that, in herbarium specimens appa- 

 rently loaded with fruit, nothing is to be found but open capsules, 

 either empty or full of abortive seeds of various sizes, presenting 

 nothing but a hard mass of homogeneous matter. 



The modifications in the form of the embryo are, however, so 

 constant in species and, often, in large groups of the order, 

 that the principal objection to taking them as absolute artificial 

 characters is the practical one above-mentioned, of the frequent 

 impossibility of observing them; and they are far from being 

 always natural. The embryo is perhaps nowhere so diversified 

 in a large natural family as in Myrtacese. Large cotyledons 

 connected by a slight radicular protuberance, or all radicle * 

 with the cotyledons minute or absolutely imperceptible before 

 germination — nearly globular or long and slender embryos, either 

 quite straight or variously bent, folded or spirally twisted — thick 

 and fleshy, or broad and flat, or contortuplicate and leafy cotyledons 

 are found in genera nearly allied in other respects, or even 

 scarcely otherwise distinguishable. Still there are some pecu- 

 liarities which, as far as observed, are so constant as to justify, 

 in some degree, the high opinion entertained of their value, and 

 requiring therefore some special notice. 



In the few species of Euchama^lauciea? where I have been able 

 to observe it, the embryo, under a thin testa, fills the cavity of 

 tlie fruit, taking the general form of the calyx-tube- It presents 

 an obovoid mass, the top nearly flat, upon which lies a slender 

 appendage or neck, which I at first took to be the radicle till I 

 obiserved in one case a minute notch at the point, showing that it 

 is, on the contrary, the cotyledonar end, analogous to that of 

 ^(echea. In the Calythricese, always included among Chamac- 

 lauciese, the embryo is quite straiglit, usually linear and terete, 

 with short cotyledons at the upper end, althougli in two species 

 it becomes broader, almost obovoid, with the short cotyledons at 

 the upper broad end, not at the lower small end as in Bceckea, 

 In the T[iryptomeneae, connecting Chama^laucieae with Ba^ckeeae, 

 the embryo is only known in two or three species, where it is 

 similar to that oi BcecJcea, 



m 



* I use the word radicle in the ordinary sense, designating the solid pai-t of 

 the embryo below the cotyledons, quite independently of the question of how 

 mueh of it forms part of the descending axis. 



