CONTRIBUTIONS TO BOTANY. 
149 
easy to conceive how the ovule in its growth partakes of a similar 
kind of increment, and becomes moulded round the stationary 
fulcrum of its placentary attachment. But in Lardizabala, where 
the ovaries are multiovular, the partial protrusion of the placenta 
is not accompanied by an expansion of the walls of the pericarp. 
Something more analogous, however, occurs in the Myrsinacece, 
where the ovules often become moulded round the globular free 
placenta, and an arillus, in the shape of an inflated entire 
membranaceous sac, generated from the root of the placenta, 
encloses the seed together with its support *. 
There is, however, a close analogy in the structure of the 
seed of Lardizabala with that of the Wmleracece and Canellacece, 
if we except the circumstance of the deep cavity in the albumen, 
which does not occur in the two latter families. We And in all 
these cases a correlative, peculiar outer crustaceous shell ; in the 
next place we see the same corresponding intervening layers of 
loosely aggregated oleiferous cells ; then a similar fleshy tunic 
enclosing the cord of thp raphe, and a still inner integument 
investing the albumen, — all bearing a striking analogy to the 
structure I have described in Illicium^. This structure shows 
the very close affinity that exists between the families last men- 
tioned and the Lardizabalacem ; and in discussing the question 
of these affinities with other orders of the Polycarpicce, I have 
already suggested the place that the Lardizabalucece should 
occupy, relatively to them, in the system J. 
Ruiz and Pavon, who founded the genus Lardizabala, men- 
tion (Syst. i. 286) two species, both natives of Chile, the one 
with biternate oblong leaflets unequal at their base, the other 
with triternate ovate leaflets. DeCandolle (Syst. i. 512) am- 
plified these brief characters by fuller details founded on his 
observation of dried specimens, adding at the same time the 
feature of the peculiar petiolar leaflets, which he termed bracts, 
and which he described as being round and cordate in the former 
case, and oblong in the other ; he observed respecting the latter, 
that he found the leaves always biternate, as in the other species, 
not triternate, as mentioned by Ruiz and Pavon. L. o-ternata 
was subsequently figured in Delessert’s ‘leones,’ i. 24. tab. 91, 
where the leaves are all divided biternately, and the leaflets are 
represented sometimes inequilateral at base, or partially lobed 
and mucronate, more or less regularly oblong or ovate, and sub- 
obtuse or acute. We find, moreover, in the same specimen the 
petiolar leaflets either cordate or rounded at base, orbicular or 
ovate, entire or denticulated. I cannot, therefore, perceive any 
* I will at some futui-e time describe the observations I have made on 
this singular structure. 
t Ante, p. 141. 
X Ante, p. 126. 
