No. 529] THE NATURE OF GRAFT-HYBRIDS 49 



explains the failure of the seeds to develop fully to the 

 fact that the fruit of the hybrid, which closely resembles 

 that of the nightshade, ripens before the seeds have had 

 time to complete their development. The tomato fruit 

 requires a very much longer time for maturing than does 

 the berry of the nightshade and a correspondingly 

 longer time is needed for the seed to be perfected; and 

 he thinks that the longer time required for the seed de- 

 velopment in S. tubingen.se is an inheritance from the 

 tomato parent, while the fruit is mainly of nightshade 

 derivation. 



All of the seedlings derived from these hybrids re- 

 verted absolutely to that parent form which the hybrid 

 more nearly resembles. Thus the seedlings of S. tubin- 

 gense and *S y . gaertueriauunt are pure 8. nigrum, those 

 of S. proteus pure tomato. This behavior also corre- 

 sponds to that of the very few cases where seedlings 

 have been secured from Cgtisus A da mi, these in all cases 

 proving to be pure Cytisus laburnum. 



Of the Cratagus-niespilus hybrids only one, Cratcego- 

 itirspihis dsineresi produced seed capable of germina- 

 ting. These seedlings were not reared to maturity but 

 so far as could be judged from the young plants, were 

 pure Cratcrgus monoggna, the parent which the hybrid 

 more nearly resembled. 3 



The third and fourth generations of the S. tubingense 

 seedlings retain perfectly the characters of 8. nigrum 

 and the same is the case when they are cross-pollinated 

 by S. nigrum. Attempts to cross 8. tubingense with the 

 tomato resulted in the formation of fruit but no seeds 

 were developed. Tt may be also recorded that crosses 

 between the two parent forms, the nightshade and the 

 tomato, were without any result. 



S. pmfrus crossed with the two parent forms produced 

 seed when crossed with the tomato to which it stands the 



