44 THE AE1L IS AN OVULARY LEAF. [BOOK i. 



that the proper and accessory envelopes of the seeds, like 

 other parts of a plant, are confounded the one with the other 

 by insensible degrees ; this I have observed in the seeds of 

 the very anomalous Cytinus hypocistis. 



"We have hitherto seen in the aril an expansion of the 

 umbilical cord, which, the ovule being considered a bud, is 

 its external leaf. Depending on the retrograde manner in 

 which the integuments of the ovule form, the aril appears 

 very late on the outside of the other envelopes of the ovule, 

 and is the last appendage sent out by the exhausted axis. 

 This feeble production, inclosing no vessels even in its highest 

 state of development, is found immediately below and on the 

 outside of the primine, the most perfect leaf of the ovule, 

 just as the pieces of the disc, the most rudimentary append- 

 ages of the floral axis, are found immediately surrounding 

 the carpels, the most vigorous and perfect organs of the 

 whole flower. But M. Aug. de St. Hilaire, the learned and 

 ingenious author of the idea that an aril was a leaf of the 

 ovule, knew too much of vegetable organography not to see 

 that this part, compared in a series of species, insensibly lost 

 its characters of an appendage, and at last became confounded 

 with the simple layer often found at the summit of the umbi- 

 lical cord ; just as the pieces of the disc, which are seldom 

 petaloid, and are more generally represented by little scales, 

 are in many cases nothing but simple projections of the 

 receptacle, which at last disappear altogether. I shall show 

 that the aril passes in allied plants through this series of 

 gradual alterations, and I shall trace it from its usual state 

 until it will be impossible to mark the limit between it and 

 the thickened extremity of the umbilical cord. 



On the seeds of many Soapworts (Sapindacese) we find the 

 organ in question ; whilst on others no trace of it is discover- 

 able. In an undetermined species of Cupania, for example, the 

 great short funicle supporting each seed expands around the 

 hilum into a circular membranous edge, which is a true cup- 

 shaped aril leaving but a small portion of the testa uncovered. 

 The seeds of Paullinia and Schmidelia are only half covered 

 by an analogous cup, the free portion of which is very 

 narrow and inserted around a large hilum, without the limit 



