Proc Natl Acad Sci U S A. 1930 Jun 15; 16(6): 377-380.

Communicated May 9, 1930

Nearly twenty years ago, while attempting to produce hybrids between certain species of the genus Nicotiana, I obtained numerous plants which reproduced the maternal species. The combinations in which this phenomenon occurred, together with the haploid chromosome numbers of the species concerned, were as follows: N. paniculata (12) x N. alata, N. Langsdorffii, or N. Sanderae (9); N. paniculata (12) x N. longiflora (10); N. Tabacum (24) x N. alata (9); N. Bigelovii (24) x N. sylvestris (12); N. rustica (24) x N. Tabacum (24) and N. Tabacum (24) x N. Bigelovii (24). Since this time, similar results have been obtained so frequently in these and in other combinations that little attention has been paid to their occurrence; yet some conclusions can be drawn from the observations.

Maternals do not appear regularly in given combinations, hence environmental factors must exert an influence in their production. The species concerned do not reproduce by apogamy, and chemical stimulants, wounding, and other means of irritation do not force them into apogamous reproduction. Maternals appear both when the chromosome numbers of the species used are different and when they are the same, both when pairing occurs at reduction and when it does not occur. They appear both when the species used as female has the higher chromosome number and when it has the lower chromosome number; but they appear much more frequently in the former case. It is to be noted, however, that they appear only when the two species used are quite distinct genetically. In a single case, observed some years ago and not examined cytologically, the plant obtained was probably haploid. It was small and sterile. In all other cases the plants were probably diploid, since they exhibited complete fertility.

Such observations are not new. Many of the older hybridizers found plants of the maternal type among the seedlings produced after endeavoring to hybridize two species. Some of the records are questionable, it is true, since origin through chance self-pollination or through natural apomixis was not excluded. Nevertheless, numerous instances of "maternals" have been observed where self-pollination has been rigidly excluded and where apomixis is not the normal state of affairs.

It seems probable, then, that these "maternals" originate in one of the following ways: (1) induced development of diploid eggs; (2) induced development of adventitious embryos from the somatic tissue of the nucellus; (3) induced development of normal haploid eggs with the restoration of the diploid number of chromosomes at an early cell-division; (4) selection by the genom of the mother species of a similar genom from among the genoms of a polyploid male nucleus.

The first (1) interpretation was soon discarded as an explanation of the phenomenon observed in Nicotiana, for the following reasons. The cross-pollinations which appear to induce the production of the "maternals" are made long after the reduction divisions have occurred, hence they have no effect upon meiosis. It is indeed possible that a few diploid eggs are formed which require the stimulus of foreign germ plasm for their development. But it is difficult to accept this explanation because capsules with a full complement (no ovules missing) of hybrid seeds have been obtained when a given plant has been crossed with Species B, while the same plant crossed with Species C has produced nearly 100 "maternals" to a capsule.

Some ten years ago it occurred to me that a choice between interpretation two (development of somatic cells) and interpretation three (development of gametic cells) might be made by obtaining "maternals" through crossing a heterozygous plant Aa with a distant species. "Maternals" derived from somatic tissue should show the dominant character only. Of the "maternals" derived from gametic tissue approximately half should be dominant and half recessive. In the latter case the plants should all breed true to the characteristics exhibited. Jørgensen1 (1928) has suggested the same experiment.

I was unable to discover a satisfactory character to use in the Nicotiana experiments. In 1925, however, several "maternals" were found in populations where different Fragaria species had been mated. A red-fruited, pink-flowered variety of Fragaria vesca (n = 7) was accordingly crossed with a white-fruited, white-flowered variety. This cross resulted in plants heterozygous for two Mendelian factors known to behave regularly, color in each case being dominant to lack of color. These plants were crossed with pollen from several other species (n = 28 except F. indica in which n = 42), and a number of "maternals" were obtained in which the possibility of self-pollination is reduced to a minimum. That is to say, numerous checks lead one to believe that not more than two or three of these plants could have been fathered by stray pollen grains from the mother plant. The results are given in table 1. Red fruit is designated R and white fruit r. Pink flower is designated P and white flower p.


F. glauca 1      
F. indica 1      
F. virginiana   1    
F. virginiana 1      
F. virginiana   1    
F. platypetala 1      
F. sq. (Friday Harbor) 2     1
F. chiloensis 6 5 3 1
Total 12 7 3 2

There were 12 plants having red fruits and pink flowers, 7 plants having red fruits and white flowers, 3 plants having white fruits and pink flowers and 2 plants having white fruits and white flowers. In addition to the plants recorded above, however, there were four "maternals" in which only flower color could be recorded. One, coming from a cross in which F. virginiana pollen was used, had white flowers. Of the other three, from a cross in which F. chiloensis pollen was used, one had pink flowers and two had white flowers. The segregation in fruit color, therefore, was 19 red to 5 white; the segregation in flower color was 16 pink to 12 white.

CybeRose note: D = dominant, R = recessive

Now self-pollination would give 3D:1R ratios, development of adventitious embryos from somatic tissue would give dominants only, while induced development of gametic cells would give 1D:1R ratios. The results obtained obviously do not exclude the occasional development of adventitious embryos from sporophytic cells. It is just as obvious that self-pollination of the mother plants with stray pollen grains is not always avoidable. I can only state my conviction, formed from a considerable experience in checking results with Fragarias, that no more than three plants from the above records are likely to be the result of self-pollination. It seems safe to conclude, therefore, that these "maternals" ordinarily arise through the development of gametic cells. The diploid number of chromosomes is presumably restored through the omission of the dividing wall after the first nuclear division. A more rigorous proof of this conclusion might be made by determining whether the "maternals" thus obtained breed true; but, owing to certain technical difficulties, it is doubtful whether this can be done satisfactorily. Perhaps a geneticist working upon more favorable material will make the experiments required.

A single plant affords a basis for the conclusion that "maternals" may also arise through the selection by the genom of the mother plant of a similar genom from among those offered by the male nucleus of a polyploid species.

Millardet2 (1894) reported a cross in which the female was F. vesca variety alba, a white-fruited type which we now know has 7 chromosomes (n). The male parent was a red-fruited variety of F. chiloensis, which has been found to have 28 chromosomes (n). Millardet obtained one plant which was "maternal" in every respect except that it had red fruits. Naturally the geneticists of today have been somewhat skeptical of this case, so different from any other known result of hybridization, and for which no obvious explanation appeared. Fortunately a similar instance has appeared among our cultures. And since the cytological control is complete, it is an observation having considerable theoretical importance.

A plant of F. vesca with white fruit, and known to carry 7 (n) chromosomes, was crossed with the pollen from a plant of F. virginiana with red fruit and known to carry 28 chromosomes. Similar crosses had given, as did this cross, numerous hybrids having 35 (2n) chromosomes which resembled the male parent, owing to the dominant effect of the preponderance of chromosomes from the male. In addition to these normal hybrids a plant was obtained resembling the female parent (F. vesca) in every respect except that it had red fruits. This plant had 14 chromosomes. It was completely fertile.

Here, then, is a diploid species (14 chromosomes) from Hawaii which is able to select from an octoploid (56 chromosomes) species native to the United States a 7-chromosome genom which is sufficiently like its own genom to reproduce a "maternal" type plant that is completely fertile.

This case affords cogent evidence in support of the polyploid theory as a factor in plant evolution. It also offers additional, though perhaps redundant, evidence that fertility depends upon the similarity of the two genoms uniting to form a sporophyte without reference to their immediate origin. Of course the evidence now available does not exclude mutation as the explanation of this case; but the white variety is an old one which has exhibited no red mutations under other conditions. Moreover, seven pairs of chromosomes are known to mate in the meiosis of the thirty-five chromosome hybrids ordinarily produced in this cross. Finally, the plant exhibits several other characteristics of F. virginiana.

  1. Jørgensen, C. A., "The Experimental Formation of Heteroploid Plants in the Genus Solanum," J. Genet., 19, 133-210 (1928).
  2. Millardet, A., "Note sur l'hybridation sans croisement au faux hybridation." Mém. Soc. Sci. phys. nat. Bordeaux, 4 Sér.. 4, 349-372 (1894).

See also, Mangelsdorf & East: Millardet's "Faux" Fragaria hybrids.