as distinguished from the Symplocacese. 135 



is very slightly imbedded in the torus at an early stage, while 

 in other cases it is generally free ; but even in these two in- 

 stances, in a short time, by the upward growth of the ovary, 

 this minute portion emerges, and the fruit is quite free from the 

 calyx. In Btyrax and Strigilia, the greatest increment of the 

 ovary occurs in its upper portion, and the persistent calyx in- 

 creases very little in size; but in Halesia there is no growth 

 whatever of the upper portion, which remains unchanged, per- 

 sistent, and crowning the fruit, the whole amount of increment 

 being confined to the lower moiety, and with it a corresponding 

 growth of the tube of the adnate calyx, as I will presently 

 show. 



2. Since I have been able to examine Halesia and Pterostyrax, 

 I am satisfied that in these cases the aestivation of the corolla is 

 decidedly imbricated, as shown by Dr. Gray ; but it is certainly 

 valvate, as I stated, in every other instance I had met with, 

 especially in Strigilia, Pamphiliaj and Cyrta, However import- 

 ant this feature of aestivation may be in some cases, as an acces- 

 sory character, it is not considered of any ordinal importance in 

 many families ; in proof of which I need only refer to Rubiacece 

 as furnishing numerous examples of both kinds. 



3. Although it be quite true, in a general sense, as I affirmed, 

 that the stamens are uniserial in Styracece, and pluriserial in 

 SymplocacecBj I admit that they are uniserial in Barberina, and 

 that in Halesia the stamens are often four times the number of 

 the petals; but in the latter case the filaments, though more 

 numerous than usual, still only constitute a single whorl, being 

 slenderly agglutinated by their adhesion to the base of the co- 

 rolla in a single series. In Symplocos, however, where the fila- 

 ments are broad at the base, and the stamens frequently as 

 numerous as thirty or forty, they are arranged in three or four 

 imbricated series of different heights, and are all agglutinated by 

 their base to the corolla for more than half their length: in 

 Barberina the stamens do not exceed fifteen or twenty, and they 

 are quite free from each other and from the corolla ; but as the 

 filaments are there very narrow at the base, they may still be 

 normally three- or four-seried, although forced by pressure to 

 assume a uniserial position : in Sympleura, however, where the 

 number of stamens is sometimes reduced to five, they are of 

 course uniserial, as in Styrax ; but this does not affect the ge- 

 neral rule. This consideration, after all, is of very trivial im- 

 portance where other and far more essential points of differential 

 structure are manifest. The point in question therefore remains 

 valid as a common rule of distinction, especially when connected 

 with the following consideration. 



4. My definition, that the Styracece are distinguished by linear 



