GYMXOSF'EKM.K. 



branch which is studded with floral-leaves (Figs. 263, 264); an ex- 

 it-mill integument is developed on all sides and surrounds the seed 

 as a scarlet aril. According to this conception the aril corresponds 

 in an eternal integument, and the Taxoideae thus possess a partly 

 dichlainvdeons ovule. Only Ginkgo and Cephalotaxus appear to 

 deviate from this, as in these there is only one integument (unless 

 the small outgrowth indicated by ar, in Fig. 260 D, really is a 

 rudimentary, external integument) ; in CYCAPEJE, to which Gink go 

 is most closely related, there is likewise only one integument. But 

 in these genera the testa is differentiated into two layers, and the 

 seed resembles a drupe ; like the Cycadese there is an external 

 fleshy covering and an internal hard one, and these two layers 

 mav probably be considered homologous with the two integuments. 

 This theory is also borne out by the arrangement of the vascular 

 bundles in Ceplilot<t,fns and Podocarpus, which present the xylem 

 in the fleshy external layer to the outside of the testa, which is 

 therefore the upper side of the integument (Celakovsky). 



The coalescence of the integuments into one is only slight in 

 Torreya, more pronounced in Podocarpus and strongest in Cephalo- 

 taxus and Ginkyo. Celakovsky terms these ovules " holochlamy- 

 deous." 



If we pass from these to the order PIXOIDE.E, we find the female 

 flowers collected into catkin-like cones, which have been considered 

 from various points of view to be sometimes single flowers, at 

 t>t her times compound inflorescences. The structure in ABIETACE.E 

 is as follows : a number of spirally arranged, scale-like leaves, 

 cover-scales (Figs. 267, 268), are situated on a long axis. In the 

 axil of each cover- scale a larger leaf-like projection, the ovuliferous 

 scale, is borne, which turns the upper side towards its cover-scale 

 (which is shown by the fact that the wood of its vascular 

 bundles is turned downwards and towards the wood in the bundles 

 of ihe cover-scale : Fig. 269). Two ovules, with micropyles turned 

 towards the central axis, and with apparently only one integument 

 ( Fig. 268), are situated on the dorsal side of each ovuliferous 

 scale, i.e. the side turned away from the cover-scale. The ovuli- 

 trnms scales grow after fertilisation, into the woody or leathery 

 "cone-scales," which are usually much larger than the cover-scales. 

 Tin.-, ovuliferous scale with its axis may, according to Celakovsky, 

 ! considered as a dwarf-branch which is situated in the axil of 

 tin- cover-scale, and bears two ovules (in the same way as in 

 Oinlcgo, one long-stalked flower, reduced to two ovules, is situated 



