OF THE RAPHE IN SEEDS, 107 



hilum, near the bottom of the ventral margin of the nut, and at last arriving at the basal 

 point, it terminates abruptly by penetrating the integuments : between this point and the 

 hilum a very short space intervenes, which is obsoletely angular, and immediately at this 

 angle, close to the margin, the small embryo is found imbedded in a large albumen, the 

 radicle pointing outwards, and the two short fleshy cotyledons, greatly divaricated, being 

 directed towards the vertex of the seed. Here then, as far as regards its direction with 

 the hilum, the embryo is anatropal, having become inverted half a turn on its centre from 

 its normal position, while the integuments have performed an entu-e gyration, by which 

 the embryo is deprived of its anatropal characteristic. AU the species of Ilollinedia 

 which I have examined have regular anatropal seeds, with one exception, where the 

 raphe is cyclotropal, as in Citrosma. "We have thus in Citrosma a verification of the 

 fact observed in Stemomirus, the raphe terminating at a point exactly contiguous and 

 corresponding to the radicle of the embryo, a fact quite incompatible with the laws of 

 structure so beautifully expounded by Mirbel : it is a development that has hitherto been 

 considered impossible, but I will proceed to show that these are not the only instances 

 of the same occurrence we find in the structure of seeds. 



In the development of the seed of the Cucurbitacece, we meet with a structure that 

 offers much resemblance to the anomalies observed in Stemonurus ; we find also other 

 features, not only difficult of explanation, but contrary to all our ideas of the nature and 

 fimction of the raphe. To aid us in the solution of these phsenomena, we will first recur 

 to the detailed account of the development and growth of the ovule of Ciicumis Angaria, 

 L., from the observations of Mirbel (M6m. Acad. Sci. Paris, ix. 621. tab. 1). The 

 growth of the anatropal ovule is here shown in all its earlier stages, and we see that by 

 the time of the fall of the flower, the micropyle of the primine and secundine becomes 

 quite closed over the tercine ; soon after which, a fine thread is observed to grow out of 

 the micropyle, which was first noticed and described by Brongniart, in the ovule of Pepo 

 macrocarptis and Momordica Elaterium. As the grovrth of the ovule proceeds, this 

 tubular thi-ead becomes covered by an envelope formed of cellxilar tissue : according to 

 the drawing of Mirbel {Joe. cit. tab. 1. fig. 10), it is a prolongation of the mouth of the 

 embryo-sac; but Griffith says (Posth. Notes, 153), that although he had observed the 

 same fact in several instances in Cucurbitacece and Asclepiadacece, he was never able to 

 trace any connexion with it and the embryo-sac, and was rather disposed to consider this 

 thread as formed by the boyaux or pollen-tubes, an opinion confirmed by the observations 

 of Schleiden, and which appears the more probable conclusion, as no remainder of this 

 thread is visible in the ripe seed. At the period above mentioned, the bottom of the 

 embryo-sac breaks away from the gangylode, becoming thus suspended by its apex, and 

 the embryo is formed in its summit : there is no appearance in the raphe at this time to 

 indicate any other change than the ordinary anatropal development. The same course of 

 metamorphosis is said by Mirbel to occur generally among the Cucurbitacew. We will 

 now compare these changes with the development found in the ripe seed, and I select my 

 analysis of Citrullus Colocynthis as a fan* example of the structure of the seed in this 

 family. 



Here the seed is obovate, much compressed, its hilum being an open slit along its 



p2 



