go BOTANICAL GAZETTE [AUGUST 4 
ture, originating from the hypodermal layer of cells, and developing — 
a wall of several layers, the innermost one of which is usually differ- — 
entiated as the tapetum. The only general tendencies to be observed 4 
are the gradual replacement of synangia by separate sporangia, and — 
the more rapid elimination of all evidences of an annulus (in the — 
general sense). It is noteworthy that in both these particulars the a 
cycadophytes, with their much more recent connection with th 
Cycadofilicales, are far behind the coniferophytes. 
THE OVULE , 
The origin of the ovule of gymnosperms remains in obscuri 
While the stamen and its sporangia repeat the corresponding struc 
tures of ferns, the ovules of Cycadofilicales and of Cordaitales are 
well organized, even in the modern sense, that their connection Wi 
This means a tremendous gap between the somewhat hypothetica 
Primofilices, on the one hand, and the Cycadofilicales and Cordaitales 
as we know them, on the other hand, a gap which there seems to b 
small probability of filling up with intermediate forms. 
the general changes that have occurred since. 
To select the most primitive type of ovule from among the Pal 
zoic forms that have been investigated is impossible, unless it 
assumed that those ovules which are most unlike the modern 01 
following result. The oldest ovule had a single integument entirely 
free from the nucellus; in testa-formation this integument differ 
entiated into three layers, the outer fleshy, the stony, and the inne 
fleshy; the ovule was supplied with two sets of vascular strands, t ne 
outer set traversing the outer fleshy layer, and the inner set traversi 
the peripheral region of the nucellus; and the beaked tip of 
nucellus broke down more or less completely within the firm and 
resistant epidermis to form a pollen chamber. If these are rea 
the features of the most primitive known ovules, the changes becom 
very apparent, and they represent general tendencies, for they ap 
in every phylum. | : 
