M3 



to inundations. It is not rare to find geminate spines grading to 

 single ones on flowering branches, thus appearing and in fact be- 

 ing alternate. On stems normally devoid of spines, it is not un- 

 common to see two or more of the prickles about the nodes some- 

 what enlarged. This condition leads to the error of taking them 

 for true geminate spines, which they are not. A little patient 

 study, however, usually suffices to clear up the difficulty in dis- 

 tinguishing the normal type 



The behaviour of the sepals during anthesis and the maturing 

 of the fruit, likewise furnish valuable indications, but not so much 

 as do their adnation. They may be persistent on the fruit when 

 fully matured ; they may be deciduous by a clear circumcision 

 at their base, or through the apex of the fruit. It is sometimes 

 difficult, not to say impossible, to distinguish the mode of adna- 

 tion in immature fruit. Sepals are either entire or lobcd ; when 

 the outer are but slightly lobed, it should be regarded as simply 

 a passing variation from the entire form. 



Beyond the mere fact that the styles may be free or connate, 

 little of practical importance has accrued from a very careful 

 study of them. While quite variable, the leaflets vary around a 

 type which is quite constant in the same species. In the more 

 glandular forms, the toothing may be compound-glandular or 

 serrulate, while in those less resinous it may be nearly or quite 

 simple. The stipules, modified leaflets, are characters of varying 

 import. In some groups and in some species they possess diag- 

 nostic value ; in others they have little significance. It is by 

 their general shape rather than by their width they merit atten- 

 tion. In drying they may become involute, revolute or remain 

 plane. Taking them all in all, they do not possess the value 

 usually allowed them. 



The petals, the habit of the bush, its in-ground ramifications, 

 the shape, size and color of the fruit, the number, shape and size 

 of the seed, although usually of secondary importance, yet fre- 

 quently furnish valuable indications. The insertion of the ovules, 

 as recently pointed out by M. Crcpin, is worthy of more than a 

 passing notice. This may be either strictly basil or basilo-parietal. 

 In the Cinnamomeae it is the latter; in the Carolina^ it is basil. It 

 is needless to say that in roses as in other plants, it is more by the 



