CONVOLYULACE^E 269 



are small and the cotyledons are much folded, they appear 

 deeply bifid and bipartite after germination. The species of 

 Convolvulus and Ipomoea observed may for convenience sake be 

 divided into those which have merely emarginate, and those 

 which have bifid, deeply bifid, and bipartite cotyledons. Con- 

 volvulus Soldanella may be given as a type having the coty- 

 ledons merely emarginate. The amount of the invagination of 

 the tegrnen is small ; the cotyledons undergo little folding in the 

 seed, and they are seen to be five-nerved at the base even at a 

 very early stage. The midrib forks a little way before reaching 

 the sinus of the cotyledon, and this is what occurs in many, if 

 not in all of the species coming under my notice. There is a 

 more or less direct relation between the length of the midrib 

 and the depth of the lobing, for the longer the lobes are the 

 shorter the midrib which represents the real length of the axis 

 of the cotyledon. 



Convolvulus sepium agrees pretty closely with this type, as 

 does also C. tricolor. 



The cotyledons of Ipomcea purpurea are shallowly bifid, 

 and the amount of folding is moderate in the seed ; but the 

 infolding of the tegmen is considerable. 



The seed of Ipomoea Pes-caprae is almost two-celled owing 

 to the ingrowth of the tegmen on the ventral aspect, and 

 the invagination of the micropyle is also considerable. The 

 cotyledons are deeply bifid, but the amount of folding of 

 the lateral lobes is moderate, and the length of the midrib 

 is considerable owing to transverse folding on the dorsal 

 aspect of the seed just after leaving the tube from the micro- 

 pyle and before the lamina divides. This particular folding 

 is due to growth in length of the lamina after the apical 

 sinus comes in contact with the obstruction caused by the 

 tegmen. The advantage or economy of the latter may be to 

 prevent the lobes of the cotyledons from getting intricately 

 confused with one another, thereby rendering exit from the 

 seed during germination more difficult or even impossible. 

 I. Eoxburghii agrees pretty closely with this type. I. Qua- 

 moclit (fig. 539) also agrees with it as far as the invagination 

 is concerned and the length of the lateral lobes of the cotyle- 

 dons, but the latter are very much folded transversely so that 



