62 Rennert : Seeds and Seedlings of Arisaema 



sion of a plumule, of course, prevents the destruction of young 

 seedlings by grazing animals ; but the well- developed plumule 

 contains enough raphide cells to protect it from attacks of this 

 kind and the advantage which the plants enjoy in this respect is 

 largely over-balanced by positive disadvantages. This variation, 

 therefore, can scarcely have been perpetuated because of the ad- 

 vantage as a means of defense which plumuleless seedlings pos- 

 sess. As far as can be discovered at present this variation is of 

 no benefit to the plant but, on the contrary, there are great disad- 

 vantages in this method of germination, which tend to throw the 

 plant entirely upon the food supply of the parent form for another 

 season. 



The germination of A. Draconthmi consists essentially in the 

 conversion of the hypocotyl of the embryo into the corm of the 

 seedling by the transfer of the food material contained in the 

 endosperm. In some aroids,* this resorption of the endosperm 

 and conversion of the base of the hypocotyl into a corm takes 

 place before the embryo leaves the seed ; that is to say, the de- 

 velopment of the seed is not completed until a bulbiform embryo 

 with true fibro- vascular bundles and no trace of cotyledon has been 

 formed. An example of this type is Spatliycma foctida. When 

 the seed of this plant germinates the stem-bud pushes out of the 

 micropyle and breaks through the ground first ; later adventitious 

 roots spring from the nodes. No primary root is produced. It 

 will be seen that the only difference between the germination of 

 this seed and the sprouting of the first year's corm of A. Dracon- 

 tium is the fact that in the former case the seed coats which sur- 

 round the bulbiform embryo must be penetrated by the stem-bud. 

 If the corm formation of the seedling of A. Dracontiiun took 

 place within the seed coats (and this could be brought about 

 simply by the arrest of the elongation of the cotyledon) the two 

 cases would be exactly similar. In fact the cotyledon of A. 

 Dracontuiiii is varying in just this direction, for it often displays a 

 tendency to be shorter than the cotyledon of A. tripliyUiiui as 

 has been pointed out above. The delay in germination at first, 

 moreover, may be another indication of a tendency to carry on 

 the entire development of the corm within the seed. 



* Engler. Monographiae Phanerogamarum, Araceae, 1 1 : 34. 



