222 



SCIENCE 



[N. S. Vol. XLVI. No. 1183 



plants when crossed together. They should not be 

 confused. The Nicotiana self-fertility is oom- 

 pletely dominant over self-sterility. Either of the 

 self-sterile species Nicotiana alaia or Nicotiana 

 forgetiana may be crossed with the self-fertile 

 species Nicotiana langsdorffii. The result in each 

 case is an Fi generation that is completely self- 

 fertile. The Fj plants show the usual monohybrid 

 ratio of 3 self -fertile to 1 self-sterile. Given the 

 basic factor for self-sterility in the homozygous 

 condition as in the case in Nicotiana forgetiana 

 and Nicotiana alata, two plants may be either 

 cross-fertile or cross-sterile with each other. 

 Reciprocal crosses always give the same result. 

 Thus the character behaves as if it were sporo- 

 phytie rather than gametic. In other words, the 

 constitution of the mother plants and not the con- 

 stitution of the gametes which they produce de- 

 termines whether a combination shall be fertile or 

 sterile. This fact indicates very strongly that 

 gametes have no other function than fusion with 

 their comptements, that the potential characters 

 which they carry are wholly latent until the de- 

 velopment of the zygote begins. The cross-steril- 

 ity shown is of such a nature that if plant A is 

 sterile with plants B and C, plant B must be ster- 

 ile with plant C. Generalizing upon the basis of 

 the behavior of self-sterile plants in intercrosses 

 one may say that a self-sterile population con- 

 sists of a small number of groups of plants each 

 plant being cross-sterile with all plants belongiug 

 to the same group and cross-fertile with all plants 

 of all other groups. These facts naturally lead to 

 the conclusions that the behavior of self-sterile 

 plants in inter-crosses is regulated by several 

 transmissible factors all of which are distinct 

 from the single basic factor for self-sterOity and 

 which presimiably may be carried by self-sterile 

 plants. A plant homozygous for self -sterility can 

 neither be fertilized by its own gametes nor by 

 the gametes of any other self -sterile plant of like 

 constitution as regards these regulation factors, 

 but any two plants differing in these regulatory 

 factors are cross-fertile. 



Twin hyVrids from (Enothera lamarcTciana and 

 franciscana when crossed with (Enothera pycno- 

 carpa: George F. Atkinson, head of the de- 

 partment of botany, Cornell University. 

 (Enothera lamarcTciana X (E. pycnocarpa. There 

 is a splitting in the Fi with production of twin 

 hybrids. One of the twins (pycnocarpa type) has 

 rosette leaves narrow and deeply cut over the basal 

 half as in (E. pycnocarpa, but the leaves are 



strongly crinkled as in (E. lamarcTciana. The other 

 twin (lamarcTciana type) has rosette leaves, nar- 

 row furrowed, not crinkled as in (E. pycnocarpa, 

 but with plain edge as in (E. lamarcTciana. The ro- 

 settes of the pycnocarpa type strongly resemble 

 those of CE. pycnocarpa because of narrowness and 

 cutness, while at the same time they resemble (E. 

 lamarcTciana in convexity and crLnkledness. The 

 general appearance of the rosettes of the lamarcTci- 

 ana type suggests neither parent, since the factors 

 selected represent the less striking character of 

 each. These two twin types are fixed in the first 

 generation, since they are repeated in the Fj and 

 probably in the following generations in accord 

 with the usual behavior of twin hybrids deter- 

 mined by de Vries. The progeny is remarkably 

 uniform, in that respect following the feature of 

 uniformity in the progeny of the parents, except 

 for an occasional mutant from the pycnocarpa 

 type. This mutation factor is probably inherited 

 from lamarcTciana. (Enotliera franciscana X (E. 

 pycnocarpa. There is a splitting in the Fi with 

 production of twin hybrids. One of the twin hy- 

 brids (pycnocarpa type) has rosette leaves with 

 the narrowness and cutness of (E. pycnocarpa, but 

 otherwise modified by (E. franciscana. The other 

 twin has rosettes very similar to those of (E. fran- 

 ciscana, somewhat modified by (E. pycnocarpa, 

 and showing considerable fluctuating variations, 

 parallel with those of (E. franciscana. In the F2 

 generation there is a one-sided splitting similar to 

 that which occurs in the F. of twins from CE. 

 IwoTceri X CE. lamarcTciana described by de Vries. 

 The pycnocarpa type twin has a hybrid constitu- 

 tion and in the Fi splits into two types, the 

 pycnocarpa tj^pe and the franciscana type, the lat- 

 ter presenting fluctuating variations parallel with 

 those in the parent franciscana. The other twin 

 (franciscana type) is fixed in. the Fj since it re- 

 peats itself in the Fj and probably in the succeed- 

 ing generations, but it presents the fluctuating 

 variations characteristic of the parent franciscana. 

 The franciscana tvpin probably carries the pycno- 

 carpa factors also, but in a subordinate or per- 

 manently latent condition. If so, it is a physio- 

 logical homozygote. If it is possible to introduce 

 a splitting factor into the franciscana twin by an 

 appropriate cross, and cause the pycnocarpa char- 

 acter to reappear in some of the progeny, the 

 fundamental heterozygotic constitution of the 

 franciscana twin would be demonstrated. 



Arthuk W. Goodspeed, 



Secretary 

 (To te continued) 



