344 MORPHOLOGY OF THE ANGIOSPERMS 



dicotyledons have a tendency to be divided, often so deeply as to sug- 

 gest four or eight. This lobing is characteristic of some Cruciferae, 

 Burseraceae, Loranthaceae, Lauraceae, Tiliaceae, some genera, Amsin- 

 ckia, Xeropetahim, Dombeya. Some species of Persoonia and Loranthus 

 have the normal two cotyledons; others, any number from three to 

 eight. That the larger numbers represent divisions of a basic two has 

 been shown for some taxa by both anatomy and ontogeny. The "four 

 cotyledons" of Cerafoplujlliiin have been shown to include the first two 

 leaves of the plumule. Polycotyledony has been reported in occasional 

 seedlings of many taxa — in as many as 2 to 4 per cent in "British 

 plants." It is characteristically prominent in some ranalian taxa. The 

 "magnolian line" in the woody Ranales has been reported to have a 

 high percentage of polycotvledonarv embryos — often as high as 87 per 

 cent with three, and 13 per cent with four, none with two cotyledons; 

 most seeds of Degeneria studied have embryos with tliree cotyledons. The 

 embryo of the Juglandaceae is remarkable in the form of its cotyledons 

 - — deeply two-lobed and the pairs united face to face ventrally. Each of 

 the apparent cotyledons consists of halves of two cotv^ledons. The fusion 

 is not complete and is evident in germinating seeds. All abnormal cotyle- 

 donary conditions in the dicotyledons — polycotyledony, syncotyledony, 

 schizocotyledony, monocotyledonv — represent only extremes of modifi- 

 cation of simple dicotyledony. All these specializations have doubtless 

 appeared more than once. The loss of a cotyledon has clearly occurred 

 many times; it is seen even within a genus — Cort/dalis, Rammctilus, 

 Clai/tonia. The presence of a single cotyledon is not alone sufficient to 

 characterize the monocotyledons, nor does it necessarily indicate that 

 all members of this taxon have been derived from the same ancestral 

 taxon. 



The absence, in some dicotyledons, of one cotyledon has been termed 

 "pseudomonocotyly," but this term has also been applied to dicotyle- 

 dons that have the two cotyledons united in a sheathing base — Podo- 

 phijJhtm, Qucrcus. Genera of the Berbcridaceae show stages in the 

 evolution of the cotyledonary tube by the connation of the lower 

 margins of the cotyledons; stages showing increasing length of the 

 tube are present in CaulophifJhim, Jeffersonia, PodopJujlIuin. The posi- 

 tion of the Nymphaeaceae as dicotyledons was, at one time, discussed 

 by morphologists; the embryo was interpreted by some as monocotyle- 

 donous. In this family also, especially in 'Nelumho, the cotyledons, 

 arising on a cotyledonary collar, form a tubular structure, which is 

 strongly lobed in some genera. Connation occurs among cotyledons, as 

 it does amonff leaves and floral orirans. 



The Epiblast. Except for the mesocotyl, the epiblast has been the most 

 puzzling part, morphologically, of the grass embryo (Fig. 124E, F). 

 The problem of its nature has been summarily disposed of several 



