WHITE— STUDIES OF INHERITANCE IN PISUM. 535 



tendency to split up, and both these influences may be operating in 

 the seeds of the same cross. As regards shape and constitution of 

 the starch grains in the hybrid seeds, Kappert secured distinctly 

 different results depending on the round smooth parent used. Lax- 

 ton's Vorbote and Goldkonig gave starch grains approaching those 

 of the round smooth seeded parent, while Emerald Gem (round 

 smooth) and Goldkonig gave round, radially split starch grains in 

 large numbers, though the splitting was much less than in pure Gold- 

 konig starch. Further, in Fo Kappert was not able through micro- 

 scopic examination of the starch grains to separate with certainty the 

 homozygous rounds and the heterozygous rounds. Seeds of the same 

 pod (all round) gave a continuous series of seeds with clearly inter- 

 mediate starch grains to seeds with only simple starch grains. From 

 the camera-lucida drawings of F^ round seeds from two pods, one 

 from each cross as noted above, those having Laxton's Vorbote as the 

 round ancestor differed considerably in extent of split or compound 

 grains from those with Emerald Gem as the round-seeded ancestor, 

 leading the writer to believe in genetic differences between the round 

 seed varieties. Kappert himself is uncertain as to whether the con- 

 tinuous gradation in extent of splitting results from hereditary or 

 environmental differences. 



Round smooth, colored floivers X round smooth, colored flowers 

 always gives in Fj all round smooth and colored flowers and from 

 unpublished data of my own, only round smooth are present in 

 later generations. Bateson (3, p. 263), citing Tschermak (81, p. 

 30, case 9), mentions an exception to this statement. The case cited 

 is Tschermak's cross P. arv., VI. (round) X P- arv., IX. (round) ? 

 which gave distinctly dimpled seeds in F^ of seed coats (Fo of 

 cotyledons). Tschermak's description in the same paper of the 

 seeds of P. arv., IX., is " roundish, rarely few dimpled seeds," indi- 

 cating that there may be some doubt as to whether the P. arv., IX., 

 parent used was not indent instead of round. In later publications 

 (86, see formula for P. arv., IX.) he describes this variety as defi- 

 nitely round-seeded. In another place in the same paper (81) de- 

 voted to assembled results, the crossing of two smooth-seeded P. 

 arv. varieties is stated to always give smooth-seeded offspring in the 

 first seed generation, which I take to be F^ of seed coats (Fo of 



