1900 | LIFE HISTORY OF SILPHIUM 1 | 
such a stamen shows that its anther is already fusing with those 
of the other stamens in the flower. In each corner of the anther 
a single large hypodermal cell is found (jig. 46), which soon 
divides into an outer and an inner cell, which may be called, 
respectively, the primary tapetal and primary sporogenous cells. 
If we bear in mind that Goebel applies the term archesporium to 
the row which we have called primary sporogenous, his state- 
ment ( Outlines of Classification, etc., p. 363) that in the Compo- 
Sitae the archesporium consists of a single row of cells finds 
here an easy confirmation (fig. 47). 
Periclinal walls soon appear in the primary tapetal row, by 
which it becomes a cell layer. This soon divides into two layers 
(fig. 48), of which the inner is to become the external portion 
of the tapetum, while the outer divides again to form the 
‘“endothecium ” and the “ middle layer” (fig. g9). The tapetal 
layer is completed by the addition of cells lying between the 
Sporogenous mass and the connective, the whole layer finally 
Consisting of large and usually binucleate cells with densely 
Stanular contents. The layer of cells just behind this addition 
to the tapetum grows to resemble the middle layer, thus com- 
pleting a second nutritive layer (fig. 50). It will be seen that 
the foregoing account of the origin of the sporogenous cells and 
their surrounding layers agrees with the early work of Warming. 
Meanwhile, longitudinal divisions have occurred in the pri- 
mary sporogenous row, giving finally a mass of four or five cells 
M cross section. The appearance of the various layers in 
longitudinal section is also shown, fig. 5Z corresponding quite 
Closely to fig. 47, fig. 52 to fig. 49, and Jig. 53 showing a stage 
just younger than fig. 50. A comparison of figs. 52 and 53 
shows that the rule that when the tapetum is differentiated the 
Pollen-mother-cell stage has been reached is not of such uni- 
versal application as Guignard claims for it (18). Exception to 
this rule has already been noted by Miss Lyon in Euphorbia 
(26). That the sister layer to the tapetum divides to produce 
the endothecial and middle layers is shown at 4, figs. 49, 52. 
This division May not occur at all, and in such cases the original 
