THE EVOLUTION OF STRUCTURES 151 



In Cycas circinalis the broadening of the midrib 

 region has progressed farther, so that only the tips of 

 the leaflets remain free (Fig. 86). The number of ovules 

 is about the same as in tJie preceding case. 



Cycas media shows several features in the reduction. 

 The midrib region has broadened until the leaflets appear 

 only as spines, giving the upper part of the sporophyll 

 the appearance of a serrate leaf; the number of ovules 

 may be as high as eight, but usually not more than six, 

 often four, and occasionally only two (Fig. 87). This 

 latter feature is important, since all the other genera 

 have two ovules. 



In Dioon edule there remains scarcely any external 

 evidence of lateral leaflets, but the broadened portion 

 of the sporophyll is rather thin and tapering and still 

 preserves something of the contour of the foliage leaf 

 (Fig. 88). The basal part has become very thick, and 

 it may not be clear to the layman that the two ovules 

 have the same morphological position as in Cycas. 



Dioon spinulosum has a shorter and thicker sporo- 

 phyll and furnishes an easy transition from the preceding 

 examples, in which the contour of the leaf blade is more 

 or less preserved to the rest of the series, in which such 

 a contour has nearly or entirely disappeared (Fig. 89). 



In Macrozamia the basal thickening noted in Dioon 

 edule has become extreme, but the midrib is prolonged 

 as a spine and thus leaves the sporophyll this vanishing 

 trace of its leaf character. In some species of Macro- 

 zamia the spine is very long, reaching a length of four 

 inches in Macrozamia Fraseri; but in some other species 

 the spine has become so short that the sporophyll looks 

 like that of Encephalartos (Fig. 90). 



