STRUCTURE.] CARUNCLES. 51 



whole lower half of the surface of the testa. These excres- 

 cences I have many reasons to think are analogous to those 

 of Polygala, but not having myself seen the seeds of these 

 two genera,, J cannot be certain on this point. 



Between the excrescences we have just examined and those 

 found on the seeds of the Lasiopetalese and of several 

 Bytneriads, the connection is very striking. There is the 

 same position next the hilum, the same variations in form, 

 the size and toothings, and the same fleshy consistence. In 

 addition to these external resemblances, we may add, that in 

 the latter plants, as in Polygala, the radicle is constantly next 

 the point of insertion of the caruncle, and that this is inde- 

 pendent of the umbilical cord. And lastly, to make the 

 identity perfect, I may state that I have seen in two species 

 of Commersonia, the micropyle scarcely visible and placed on 

 the caruncle itself. I have not been able to repeat the same 

 observations on the seeds of Seringia, Thomasia, Lasiopetalum, 

 and of others having caruncles ; but if we may judge from 

 the figures of the latter given by M. Gay, there exist such 

 analogies in form and especially of position between these 

 excrescences and those of Commersonia, that I should hardly 

 hesitate in regarding them as expansions of the exostome. 



Let us remark before proceeding further, that the produc- 

 tions of the micropyle have hitherto appeared frequent in 

 families not distantly related to one another. Now that, for 

 example, many botanists have placed Spurgeworts among the 

 Polypetalous orders, we cannot but see the near affinities 

 which unite this order to Lasiopetalese, Butneriaceae, and in 

 short to the whole of those groups which compose Jussieu's 

 order of Malvaceae. In all these families, as in Euphorbiaceae, 

 the presence of a caruncle is combined in a constant manner 

 with that of a layer of parenchyma which covers one or more 

 crustaceous plates of the testa. But because this coincidence 

 is always found, we must not therefore conclude that the 

 layer of parenchyma cannot exist without a caruncle. 



In the preceding examples the false aril has been as it 

 were rudimentary ; we shall now proceed to some of those 

 cases in which it is found extending much more over the 

 seed. 



E 2 



