THE FLORA OF SOUTH AMERICA. 367 
Gen. Plant, tom. 2, tab. 69—73), it is not now necessary to 
determine, it being sufficient for our present purpose to indi- 
cate the fact, and to add that they differ from the whole of 
that family, in the form and position of their embryo, as well 
as in habit and inflorescence. With Convolvulacee, as Prof. 
Lindley remarks, many analogies exist, but they differ in 
their simple, not imbricate calyx, in their distinct ovaria and 
the important character of their embryo. I shall presently 
attempt to show that the hitherto anomalous genus Gra- 
bowskya, is referrible to a position between Borraginee and 
Nolanee, and Dichondree will then form an excellent con- 
necting link between Nolanee and Convolvulee, to the former 
of which this small group has a very close affinity, on account 
of the almost gynobasic origin of its nearly distinct carpels, 
and also because its embryo is really cyclical, notwithstanding 
. thatthe cotyledons, at their extremity, are bent back in a 
. Sigmoid form, after having completed more than an entire 
. helix, somewhat in the manner of the embryo of Convolvulus, 
. butit is to be remarked, that although the cotyledons of 
— Dichondra are broader than the radicle, and more foliaceous 
_ than those of Nolanee, they are simply parallel, and have not 
_ their margins crumpled and conduplicated, as in the true 
= Convolvulee. Doctor Lindley observes (Nat. Syst. 230) that 
“if we attend to the embryo, they will stand among Convol- 
vulaceæ, if to the carpels, among Nolanacee : upon the whole 
thelatter must be accounted of the most importance, and 
Consequently it is with Nolanacee that I arrange them." I 
find on examining the seed of Dichondra repens, that the 
— Utricle falls away from its receptacular attachment, showing 
à distinctly round perforate aperture at base, and on the 
Teceptacle are to be seen opposite the opening, two distinct 
. Prominences, corresponding to what, by analogy, may be 
. TOnsidered as of a similar nature to the areolar processes 
Observed in Nolanee and Grabowskya. The reasons above 
Offered will therefore probably justify the position I have 
 Ventured to assign for Nolanacee in the general system, so 
