10 THE COMPARATIVE EMBRYOLOGY OF THE RUBIACEAE 
ered by an integument,” and recounts the forms in which, accord- 
ing to earlier students, the nucellus is naked. Some reference to 
this point has already been made in the introduction. In endeavor- 
ing to contribute to the answer of the question, Warming studied 
the development of the nucellus in the type above mentioned, and 
found that the epidermal cells at the apex of the nucellus, and 
therefore, those which cap the elongated cells which lie in the 
axis of the structure, have contents more granular than the re- 
maining epidermal cells, and also that the epidermal cells sur- 
rounding these capping cells divide tangentially and “il se dé- 
veloppe ainsi une assez forte couche de cellules de provenance 
épidermique qu'on doit considérer comme un tégument." His 
description with the exception of the reference to the more gran- 
ular contents of the middle epidermal cells is exact enough to 
apply to the Ga/ieae, and one cannot resist the conclusion that his 
` interpretation of this collar of tissue as a rudimentary integument 
is correct. Guignard, * seven years later, published a fuller ac- 
count of the behavior of the nucellus and embryo-sac of Thesium 
in which, however, he entirely overlooked Warming’s work and 
suggestion. Guignard here makes no reference to the tangential 
divisions seen by Warming, nor does he show any such in his 
figures. Indeed he takes the position (p. 189, 1. c.) that there is 
no integument at all in Thesium, though it must be said that in 
making the statement he is referring to the seed, rather than to 
the young ovule. 
The bottom of the micropylar canal is formed of the few epi- 
dermal cells capping the archesporium (Pl. 1, Figs. 5, 6,8). The 
walls also are formed by an epidermis continuous with and having 
its origin in the outer epidermis of the ovule. The endodermis t 
elmaier (Ueber die Keimsack einiger Compositen 
und dessen Umhiillung. Bot. Zeit 47: So : 
; me- 
diately the embryo-sac after its absorptio: 
889), comes to surround i 
5, 82 m 
n of the nucellus is complete. For this layer 
1 ion and figured being 
Hegelmaier proposed the term endodermis, the t 1 
Helianthus annuus, Bidens leucantha and 
wickelungsgeschichte der Frucht 
J Ns MAA 
Zinnia tenuiflora, 
cell-layer of the nucellus, ** bei i ö 
Nucellus," a conclusion whic 
OH CARET TUENDA 
Zi 
