386 ]\Ir. J. Mievs on the comparative Carpkal 



seeds, he is silent about the existence of integuments, raphe, 

 or chalaza, and none of his many analytical figures gives any 

 information uj^on these subjects. 



It is to be regretted that a very small amount of reliable 

 information has been recorded concerning the carpical struc- 

 ture of the family. Among the few analyses that have been 

 published, that of Gaertner is the most important : he shows 

 in his work (i. 364, tab. 76. fig. 1) that of Cordia {Sebestena) 

 Myxa.j where the seed is suspended a little below the summit, 

 with a raphe descending from that point to the l)ase, its small 

 radicle being superior, and its large fleshy cotyledons deeply 

 plicated. A very different version of this structure, in a plant 

 which he called Cordia Myxa, is given in Wight's ' Illustra- 

 tions,' pi. 169 : in the ovary the ovules are there shown to be 

 quite erect, fixed in the basal angle of each cell ; in the fruit 

 the point of the attachment of the seed is not indicated, though 

 it is drawn separately in fig. 11, without any mark of laphe or 

 chalaza. This analysis is drawn by an Indian artist, and 

 shows evident marks of inaccuracy ; for the embryo, as shown 

 in figs. 11 and 12, has a long pointed radicle, which is inferior 

 (instead of superior). I therefore place more reliance upon the 

 analysis of Gaertner, which is more conformable with my own 

 observations, as will be shown presently. Wight's ^ Icones,' 

 also drawn by Indian artists, show the ovules in the same 

 position as that indicated in the ' Illustrations,' in two other 

 species of Cordia., in pis. 1379 and 1381, while in three other 

 cases they are attached by their middle, as seen in plates 469, 

 1378, and 1380, which agrees with what I have generally 

 found in the Brazilian species of Cordia,. Prof. A. De Candolle, 

 in a note to the genus Varronia (Prodr. ix. 468), states that 

 the ovules are there laterally affixed to the internal angle of 

 the cells ; and, again, in another note (p. 471) he adds that he 

 found the ovules in C. gerascantlms attached as in Varronia^ 

 and that in C. Chamissoniana (a closely allied species) the 

 point of attachment is nearer the base \ but my observations 

 upon the same species convince me that the connexion is at 

 the middle, rather above than below it : in C. discolor he found 

 the ovules fixed as in Varronia. My examination of the uni- 

 locular nut of Cordia glabra shows that the seed, which tightly 

 fits the cell, is attached by a somewhat broad hilum to a spot 

 a little below the middle of the cell, from which point a line 

 of raphe, imbedded between the two integuments *, descends 



* The seed, as stated by Gaertner, lias two integuments : the outer 

 one, of very fi-iable texture, quite white, is composed of numerous large 

 cells rather laxly agglutinated together ; but it adheres firmly to the inner 

 integument, which is opaque, very linely reticulated, like an extremely 



