COMPARATIVE MORPHOLOGY OF GYMXOSPERMS 159 



cient Cordaitales the most obvious explanation of the ovule sec- 

 tion is the presence of two distinct integuments, and the testa 

 of the Bennettitales indicates the same fact in that group. In 

 Cjcads, Ginkgo, Cephalotaxus, Podocarpus, etc., the outer 

 fleshy and inner bony layers may be regarded as representing 

 two integuments structurally, which have become connate; 

 while in Taxus, Dacrydium, Microcachrys, etc., the outer flesh- 

 forming integument has remained or has become distinct as the 

 so-called " aril." In the Pinaceae the most remarkable modifi- 

 cation is displayed, the outer integument having entered into the 

 structure of the " ovuliferous scale." Among the Gnetales the 

 inner integument is produced into the remarkably long and tubu- 

 lar micropyle, while the outer one has been usually designated a 

 " perianth." 



It is evident that between distinct bracts on the one hand 

 and a genuine integument on the other there may be all sorts of 

 intergradations, and that a hard-and-fast morphological line 

 probably does not exist. We can recognize in all Gymnosperms 

 a single undoubted integument, but whether the next outer 

 structure should be called an outer integument, an aril, a peri- 

 anth, an ovuliferous scale, or what not, seems impossible and 

 perhaps unimportant to decide. 



The nucellus of the group presents some features which seem 

 to be important. In the whole group there is a remarkable de- 

 velopment of sterile tissue above the single or several sporoge- 

 nous cells, which thus appear imbedded in the chalazal region of 

 the ovule. Among the more primitive Gymnosperms (Cor- 

 daitales, Bennettitales, Cycadales, and Ginkgoales) this sterile 

 mass differentiates into a firm persistent beak at the apex, and a 

 loose tissue between the beak and the embryo sac. In the same 

 groups a deep pollen chamber is developed within the beak, the 

 loose tissue becomes broken dovra by rhizoidlike pollen tubes, 

 and the beak settles down upon the embryo sac. This uniform 

 structure in all the more primitive forms certainly has to do with 

 the more primitive phases of siphonogamy, which will be dis- 

 cussed more fully under fertilization. Among the Coniferales 

 and Gnetales the beak is not developed, but the outer micropylar 

 layers of the nucellus are of firmer tissue than those beneath, 

 and occasionally the tip of the nucellus breaks down into a cup- 

 like depression which holds the pollen grains, as has been ob- 



