CONTRIBUTIONS TO BOTANY. 
211 
the margins of the cotyledons. In Berchemia the cotyledons 
are incumbent, both with regard to the ovary and to the dorsal 
raphe. 
In this investigation, I have noticed only the proper integu- 
ments of the ovule, leaving out of consideration that coating 
which, after impregnation, frequently grows over the primine 
and produces in the seed either an incomplete or an entire tunic, 
called an arillus. Schleiden and others are of opinion that the 
true aril must necessarily be pervious at its extremity, and he 
concludes (/. c. p. 431) that wherever an actually closed structm’c 
surrounds the seed, it is undoubtedly a layer of the seed-coats : 
but this is mere opinion unsupported by proof ; for it is equally 
probable that the aril may become closed just as much as the 
proper tunics of the ovule. It has also been thought that it must 
necessarily be fleshy ; but I have shown* that it is often perfectly 
entire, and frequently hard and testaceous (as in Canellacea, 
Winteracece, &c.), in which case the true testa, or development 
of the primine, is generally either fleshy or membranaceous, 
contrary to its usually hardened condition. 
Where, on the other hand (as in Magnolia, Clusia, &c.), the 
inner tunic of the ovule becomes hardened by osseous deposits, 
the primine, as in the former case, remains fleshy, and assumes 
the appearance of a complete arillus, for which reason I proposed 
to call it an arilline t instead of testa, to which name the hard- 
ened tunic is more entitled. It has been contended by Dr. Asa 
Gray that the osseous tunic and the aril-like covei’ing which 
contains the cord of the raphe are both developed from the pri- 
mine, the former resulting from hardened deposits upon its 
inner layers of cells, while the outer cells remain soft and 
fleshy. I have argued that, if such a deposition took place 
in the manner stated, these two dissimilar textures must be en- 
closed by a single epiderm and one endoderm ; but we And, on 
the contrary, each of the tunics provided with its respective ex- 
ternal and internal epidermis, showing that the two formations 
are independent in their origin. It is also clear, that if these 
two coatings, so very dissimilar in their nature, were produced 
from a single ovular tunic, then the cord of the raphe ought 
rather to be found in the nut, as that would correspond with 
the inner layers of mesodermic tissue in which the raphe exists 
in the primine; whereas that cord of vessels is really imbedded 
in the fleshy tunic, while the nut is free from vessels of any kind. 
This is further shown in the general structure of the seed of 
Magnolia, where the raphe passes in its usual course through 
* Ann. Nat. Hist. 3 ser. ii. 39; Contributions to Botany, i. 128. 
t Trans. Linn. Soc. xxii. p. 89. 
2 E 2 
