MONOCOTYLEDONES. 277 



may be traced to suppression, very rarely to a splitting of certain 

 members, the typical relative positions not being changed. Thus, 

 the Iridaceas, the Cyperacece, most of the Grramineas and some 

 Juncaceas deviate in having only 3 stamens (Fig. 279), the inner 

 whorl (indicated by *) not becoming developed. The Musacete 

 differ in the posterior stamen not being developed ; Zingiberacere 

 (Fig. 314), Marantacea?, and Cannacecv, in the fact that only 1 of 

 all the stamens bears an anther, and the others are either sup- 

 pressed or developed into petaloid staminodes, with some perhaps 

 cleft in addition. The Orchiclea? deviate in having, generally, 

 only the anterior stamen, of all the 6 developed (Fig. 280). In 

 this, as in other instances, the suppression of certain parts of 

 the flower is often connected with zygomorphy (i.e. symmetry in 

 one plane), chiefly in the inner perianth- whorl, but also in the 

 other whorls. In the Orchids, the perianth-leaf (the labellum, 

 Fig. 280 /) which is directly opposite the fertile stamen, is larger 

 and altogether different from the others. The perianth-leaves 

 may also be suppressed; see, for example, the two diagrams of the 

 Cyperaceaa (Fig. 284). In some orders the suppression of these 

 leaves, which form the basis of the diagram, is so complete that it 

 is hard to reduce the actual structure of the flower to the theo- 

 retical type, e.g. the Grasses (Fig. 290) and Lemna (Fig. 303). In 

 the first family, which especially comprises water-plants, a some- 

 what different structure is found ; thus Fig. 282 differs somewhat 

 from the ordinary type, and other flowers much more so ; but the 

 floral diagrams which occur in this family may perhaps be con- 

 sidered as the most probable representatives of an older type, from 

 which the ordinary pentacyclic forms have taken their origin. In 

 favour of this theory we have the larger number of whorls, the 

 spiral arrangement of some of these in the flower, with a large and 

 indefinite number of stamens and carpels, the perfectly apocarpous 

 gynceceum which sometimes occurs, etc., etc. 



The Monocotyledons are divided into 7 Families : 



1. HELOBIFJE. This family forms a group complete in itself. It commences 

 with hypogyiious, perfect flowers, whose gyucecium is apocarpous arid ter- 

 minates in epigynous and more or less i educed forms. 



2. GLUJUFLORJE. These have as a starting point the same diagram as the 

 following families, but otherwise develops independently. 



3. SPADICIFLOR^E. Also an independent branch, or perhaps two different 

 ones which terminate in much reduced forms. 



4. ENANTIOBLAST.E. These ought perhaps to be amalgamated with the follow- 

 ing family. 



