354 
The Ohio Naturalist. 
[Vol. V, No. 8, 
The subject has proved very interesting because of the large 
number of closely related plants which have already been studied 
and the accumulated literature which was easy of access either 
in the original publications or through brief reviews and abstracts. 
The writer became thoroughly familiar with her own material 
before making any comparisons in order to avoid having pre- 
conceived ideas of what ought to be expected. 
Material for study was collected at Licking Reservoir in 1901 
by Professor J. H. Schaffner and at Sandusky Bay, Lake Erie in 
the summers of 1902-1903 by the writer. The usual methods of 
killing, imbedding, sectioning and staining were employed. 
Thickness of the sections varied from 8-20 microns the older 
material being cut thickest. 
The development of the carpel is almost identical with that 
of Ranunculus abort! vus as described by E. A. Bessey (1). The 
rounded pyramid of the receptacle first appears from which 
numerous conelike projections arise (Fig. 1). Those nearest the 
base develop into the stamens. Near the summit of the receptacle 
the arrangement of parts is spiral but approaches the cyclic 
among the outer stamens. The number of stamens found by 
actual count varied from 17-21 while the number of carpels was 
approximateh^ half as great. A lamina or flap develops from the 
distal side of the young carpel enveloping the inner portion 
which begins to grow away from the receptacle (Fig. 2). This 
lamina thins out as it meets the axillary placenta and traces of 
the integuments can be seen (Fig. 3). As the nucellus develops 
it describes an angle of 180° and when the gynoecium is mature 
the tip of the nucellus is directed downward while the opening of 
the micropyle is towards the receptacle. (Fig. 4). Only a single 
integument develops. The outer cells of the integument nearest 
the placenta are large and glandular and seem to function in 
conducting the pollen tube to the micropyle (Fig. ")). After the 
closing of the carpel an elongated style develops having finger 
like, glandular cells on the stigma which afford a lodging place 
for the pollen. 
The microsporangium develops a plate of four or five hypo- 
dermal archesporial cells which divide by periclinal walls to 
form primary wall and primary sporogenous cells (Fig. 6). The 
primary wall cells then divide and the inner cells develop into 
the tapetal layer (Fig. 7). The outer cells may divide once or 
twice forming two or three distinct layers between the epidermis 
and the tapetum (Fig. 8). The layer next to the epidermis forms 
the endothecium with thickenings in the angles of the walls, 
exactly as were found in the endothecium of R. delphinifolius. 
Further divisions by anticlinal walls occur in both tapetal and 
wall layers and later the tapetum becomes binucleate by kary- 
okinesis, without forming walls, instead of by fragmentation of 
the nucleus (Fig. 10-11). 
