I9IS.] PURE SPECIES OF CENOTHERA. 233 



species show hardly more than a trace of pollen abortion, and 

 Janiesii from Texas only a small amount of sterile pollen. I have 

 this winter tested the seed fertility of some of these species by ger- 

 minating the seeds in Petri dishes after the method recently de- 

 scribed (Davis, '15&). The Dutch biennis gave a germination of 

 about 96 per cent., the Dutch muricata about 72 per cent., grandi- 

 flora B about 95 per cent., franciscana about 61 per cent., venusta 

 about 87 per cent., and Janiesii about 91 per cent. 



It is interesting to note in the above list that the Dutch biennis 

 with its very high percentage of fertile seeds (96 per cent.) has 

 extensive pollen abortion and the Dutch muricata with seed ger- 

 mination of about 72 per cent, has an even lower degree of pollen 

 sterility. On the other hand there are species of CEnothera with 

 both high seed and pollen fertility as illustrated by some races of 

 grandiflora, venusta and Jamesii. I was especially interested in 

 the conditions shown by my race grandiflora B with its almost per- 

 fect fertility both as to pollen and seeds. This race isolated from a 

 collection of mixed seeds gathered by Tracy in 1907 at Dixie Land- 

 ing, Alabama, has always seemed to me to present a type of unusual 

 purity. The line was started in 1908 by a cross of two similar 

 plants (Davis, '11, p. 203) representing the broader-leaved forms 

 of grandiflora that were present at Dixie Landing and I have grown 

 in small cultures several generations of the plant without noting 

 departures from the type. I cannot accept the criticism of De 

 Vries ('14, p. 348) that my race grandiflora B is impure because 

 from the same collection of mixed seeds of Tracy's he obtained a 

 diversified culture as I also reported (Davis, '11, p. 203) when the 

 line was first isolated, and because De Vries and Bartlett found 

 the Dixie Landing station " desolate " five years after the visit of 

 Tracy. This type may prove to be nearer to the desired pure spe- 

 cies than the Dutch biennis. 



Jeffrey in recent papers ('14a, '14&, '15) has taken the position 

 "that in good species the spores or pollen is invariably perfect 

 morphologically " and from this standpoint refuses to consider La- 

 marckiana and other Oenotheras as suitable material on which to base 

 experimental studies on mutations. To him the mere presence of 



PROC. AMER. PHIL. SOC, LIV. 2l8 P, PRINTED AUG. 9, I915. 



