364 INTRODUCTION TO EMBRYOLOGY OF ANGIOSPERMS 



megaspores; (4) long synergids; and (5) embryo sac constricted in 

 the middle portion. On the other hand, if we turn to the Passi- 

 floraceae, we find a number of striking differences, for here (1) the 

 nucellar epidermis does not undergo any periclinal divisions; (2) it 

 is usually the uppermost megaspore which functions, not the cha- 

 lazal; (3) the ovule is completely anatropous with a perfectly sym- 

 metrical nucellus ; and (4) the outer integument grows up to a higher 

 level than the inner and takes part in the formation of the micro- 

 pyle (see Schnarf, 1931). 



Garryaceae. The systematic position of the Garryaceae has 

 been disputed for a long time. Engler and Gilg (1924) placed it 

 among the most primitive families of the dicotyledons, and the 

 same course has been adopted by Engler and Diels (1936) in the 

 latest edition of the "Syllabus der Pflanzenfamilien." Others, like 

 Bentham and Hooker (1880) and Wangerin (1910), have taken a 

 different view and assigned it to the highest of the Archichlamydeae 

 placing it close to the Cornaceae. 



Hallock's (1930) work on the morphology and embryology of 

 Garry a elliptica necessitates a fresh appraisal of the situation. She 

 reports as follows: 



1. The staminate flower, although apparently monochlamy- 

 deous, invariably shows the primordia of the sepals in the earlier 

 stages of its development. 



2. The pistillate flower is not naked. The two whorls of alter- 

 nately arranged "bractlets" or "folioles" seen on the top of the 

 ovary represent the reduced perianth lobes (the members of the 

 outer whorl are thick and green, and those of the inner are more 

 delicate and petal-like). The presence of a vascular supply in these 

 structures and the definite position which they occupy with respect 

 to each other and to the pistil supports this view, and from this fact 

 it further follows that the ovary is inferior and the flower parts 

 epigynous and not hypogynous. 



3. The integument, described as "complete or incomplete" 

 (Engler and Gilg, 1924), is a thick and massive structure. It is only 

 in later stages that it begins to be consumed by the embryo and 

 therefore appears to be "incomplete." 



4. The nucellus is thin and ephemeral, disappearing completely 

 at the sides of the embryo sac. 



5. There is a single archesporial cell which divides to form a wall 

 cell and the megaspore mother cell. 



