258 MORPHOLOGY OF THE ANGIOSPERMS 



in Magnolia and species of Acacia. As applied to the angiosperm ovule, 

 aril is obviously merely a descriptive term. The interpretation of the aril 

 as a third integument is unfortunate, especially where it is clearly an 

 elaboration of one, or of part of one, of the typical integuments. The 

 term caruncle is often applied to smaller, Heshy parts of the ovule, espe- 

 cially to parts of the integument, funicle, and raphe. Use of the terms 

 aril and caruncle is inconsistent; fleshy outgrowths of the integument 

 are called caruncles, if small; arils, if large. The term "third integument" 

 has also been applied to the inner layer of the inner integument when it 

 is sclerified or consists largely of tracheids. Four integuments have been 

 described for the ovules of some of the Annonaceae. Two of these are 

 the usual integuments; one is an aril — an elaborated part of the outer 

 integument; the fourth is described as a new structure, which is 

 formed by proliferation of chalazal tissues, a "middle integument," 

 lying between the inner and outer integuments. In the seed, this "middle 

 integument" takes over the function of the outer integument. This struc- 

 ture, found in only three genera of the Annonaceae, seems hardly worthy 

 of the rank of integument. Arils are associated chiefly with tropical 

 fruits, are rare in temperate-climate plants. The aril has been considered 

 characteristic of the primitive angiosperm ovule, but this seems most 

 unlikely, since it occurs in taxa scattered throughout the angiosperm 

 system. Like the caruncle, it seems to represent specialization in ovule 

 structure related to dissemination by animals. 



Ovules without integuments have been described in several families. 

 These ovules are doubtless all reduction types, with integuments lost in 

 simplification of the flowers. Probably only in the Santalales and Balano- 

 phorales are truly naked ovules present. Reports of naked ovules in the 

 Rubiaceae — Houstonia, Coffea — Olacaceae, Apocynaceae, and Ipomoea 

 have been shown to be in error. In the Amaryllidaceae, Crinum is re- 

 ported to have naked ovules, but this genus needs further study. In 

 greatly reduced ovules, difficulties of interpretation are great; the 

 nucellus may consist of a uniseriate layer of very few cells, even only one 

 or two, and these may be ephemeral, as in Houstonia and the Olacaceae. 

 The nucellus has been interpreted as the integument and the integument 

 as the nucellus in the Apocynaceae. 



Enlargements of the carpel wall adjacent to the ovule, of the chalazal 

 region, and of the top of the outer integument that overtop and block, 

 or appear to block, the micropyle to some extent are called obturators 

 (Fig. 103). The term is best restricted to protuberances of the carpel. 

 The term obturator, like caruncle and aril, has no morphological value. 

 Good examples of obturators are seen in many of the Rosaceae — Pijrus, 

 Prunus, Spiraea. In this family, the obturator consists of loosely ar- 

 ranged parenchyma and may serve as an aid in pollen-tube conduction 



