EVOLUTION OF SEED-PLANTS 107 



of a Cycad stem; In some details, especially 

 the structure of the vascular strands passing 

 out to the leaves, the agreement is remarkably 

 exact. The leaf-stalk, however, only received 

 a single strand, or at most a pair, which branched 

 over and over again to supply all the complex 

 subdivisions of the frond. The general structure 

 of the frond internally was like that found in 

 many of the Ferns. Besides the spines already 

 mentioned, both stems and fronds were studded 

 with spherical glands, usually stalked; we do 

 not know whether they secreted honey or some 

 other substance, but they have proved important 

 as a means of identification. 



Up to the year 1903 nothing certain was 

 known of the reproduction of Lyginodendron, 

 and some botanists thought that it was probably 

 a true Fern in this respect, though so peculiar 

 in structure. In that year, however, Professor 

 F. W. Oliver was able to show that a previously 

 unassigned seed, then known by a name of its 

 own, belonged in reality to Lyginodendron, a 

 result quite unexpected before. The seed, a 

 small barrel-shaped body, about a quarter of 

 an inch long, is enclosed in an envelope or cupule, 

 like a hazel-nut in its husk (see fig. 12, a). This 

 cupule bears glands, exactly like those on the 

 frond and stem of Lyginodendron, with which 

 it occurs in association; it was by the help of 



