AmphicarpiBa iiwnoica. 25 



(a) Subterranean Legume Resulting from Buried Purple Flower. 

 The surface of the legume with its raised stomata, its numer- 

 ous hairs, including the form with multicellular base, was not to 

 be distinguished from that of any typical subterranean legume. 

 No indurated elements were to be found in the walls. The seed 

 filled the entire cavity ; its coat showed but the one indurated 

 layer typical of the subterranean seed ; not the slightest indi- 

 cation of the great supporting hypodermal layer of the aerial 

 seed could be discovered. The size of the cotyledonary cells, 

 also the size and shape of the starch granules, agreed with 

 those previously described for the normal subterranean seed. 



(b) Legume Resulting from Burying a Flat but Well-Developed 



Legume Produced from a Green Aerial Flower. 



The external appearance of the legume was strikingly 

 subterranean in color and fullness, owing to the swelling 

 of the seeds (two in number), which quite filled the space 

 within. The sutures were as distinct as upon any aerial 

 legume. Upon opening the pod, the entire inner epidermis 

 had separated as a sheet of tissue from the remaining portion 

 of the walls. Exactly what agencies caused this is question- 

 able. It may have been the combined action of the new con- 

 ditions, or perhaps of moisture alone. The surface of the 

 legume was typically aerial in character, possessing no stomatic 

 papillae, and but a scanty growth of hairs. The short rods in 

 the supporting layer were well developed. Chlorophyll could 

 not be detected, but a pinkish-purple coloring was irregularly 

 diffused through the etiolated cells. The seeds, however, 

 were subterranean in all details of structure and in contents. 



(c) Legume Resulting from Biirying a More Mature Legume 



Produced by Green Aerial Flower. 

 This aerial legume having already assumed its character- 

 istic form, showed but little change after experiment. An 

 inclination to turn brown v.'as obsen,-cd, but this would have 



