Proc. Nat. Acad. Sci., 27: 436-440 (1941)
Communicated July 23, 1941

Mangelsdorf and Reeves have recently1 suggested that North American maize is somewhat contaminated with Tripsacum. According to their hypothesis the original Zea from South America hybridized with Tripsacum in Central America, producing a partially fertile hybrid which by further back-crossing to maize produced Euchlaena. This has since grown as a weed in Central American corn fields, serving as an intermediary for the flow of Tripsacum germplasm into Zea. A logical extension of this hypothesis (which we are calling the hypothesis of antithetical dominance) will account for many of the genetic peculiarities of Zea mays. Since critical tests which are under way will require some years for their completion, the hypothesis with some of the supporting evidence is outlined below.

Hybridization between genera as diverse as Zea and Tripsacum would bring into a single nucleus parts of two very different reaction systems. The features which distinguish these genera are so different that many of the intermediates would be monstrous and would be selected against. For example, female spikelets occur singly in Tripsacum but are paired in Zea. Obviously there is no functional intermediate between one spikelet and two and natural selection would favor those modifiers which make for the clear-cut end result of one or of two. For all reactions which are disadvantageously intermediate between Zea and Tripsacum there will be a selection of modifiers tending to make the end result of each reaction either as Zea-like or as Tripsacum-like as possible, in much the same way as the sexual balance of many organisms has been established by the selection of modifiers.2 Reaction curves which originally had the form:

will be modified into curves like:

We have named this the hypothesis of antithetical dominance which can be stated in general terms as follows: In hybrids between extremely diverse parents, natural selection will tend to encourage those modifiers favoring one parental extreme or the other and suppressing intermediates. In North American maize selection has been operating at the maize end of the maize-Tripsacum spectrum and will have tended to make Tripsacum characters recessive except in those cases where they were directly advantageous. By the sharp threshold which has been built up through selection much of the Tripsacum influence in North American maize will be quite hidden, only to be revealed when an environmental or genetic change produces a plant which does not reach the threshold. This hypothesis would explain such phenomena as inbred lines of maize which do not show tassel ear (a Tripsacum-like character) in one environment but may do so in another3 and that a single generation of inbreeding often produces plants strikingly different from their cross-pollinated parents.

Pedicel length in mm. of the upper male spikelet for 25 plants each of (from above to below) (1) Papago flour corn, first generation inbreds; (2) sweet corn inbred P39; (3) sweet corn inbred 1-45; (4) hybrid sweet corn, P-39 X 1-45; (5) Missouri dent corn.

We have made a careful study of various maize inbreds with reference to the morphological characters which distinguish Zea from its close relatives. Among the inbreds which we have examined a considerable number have tended to show a combination of characters which is characteristic of most species of Tripsacum but which is rare in North American maize and unknown in maize from Peru and Bolivia (the latter being on Mangelsdorf and Reeves hypothesis the only maize which is uncontaminated with Tripsacum). The most extreme of these inbreds are so similar to Tripsacum that we are referring to them as "Tripsacoids." They differ from the varieties out of which they were extracted by having more slender leaves, tassels with short, stiff branches and a wide-channeled rhachis, and in having the upper male spikelet sessile or sub-sessile. In all of these characters and in other less tangible ones such as leaf texture and general aspect they tend to resemble Tripsacum. Frequency diagrams for one of these characters (length of pedicel on the upper male spikelet) are shown in figure 1. It will be seen that pedicels are shorter in the inbreds than in the crossbreds or than in the open pollinated varieties. These results are to be expected on the hypothesis outlined above. An alternative explanation is parallel variation in Zea and Tripsacum but this would not explain (1) the fact that these Tripsacoid characters are recessive in corn nor (2) their absence in highland South American corn.

Through the kindness of Dr. R. A. Emerson and his colleagues we have been able to study a series of 16 crossbred F1's and the Cornell inbreds from which they are derived. Several of the inbreds in this series were quite Tripsacoid, with short tassels, stiff upright tassel branches and subsessile spikelets. In each case these characters disappeared when the inbreds were crossed. Comparisons for two characters are shown in figure 2 (pedicel length below, tassel length above). According to our theory, crosses between an inbred with a very short pedicel and one with a pedicel of normal length should have pedicels as long as those of the long-pedicelled parent, or even longer. Crosses between inbreds, both of which have well-developed pedicels (4 mm. or more), should show no such regularity of dominance. It will be seen that these predictions are verified. The data are arranged in the form of tridents, the right-hand pair of dots representing average parental values while the dot to the left shows the average for the crossbred.

It might be objected that this dominance is merely heterosis. This is refuted by the fact that some of the inbreds have extremely long pedicels and by comparisons with the tassel lengths of the same stocks. Tassel length is certainly affected in part by heterosis and if the observed dominance over very short pedicels is due, even in part, to heterosis, there should be a correlation between the results for the two characters. There is no indication of any such correlation (note, for example, 3 X 18) and we therefore conclude that the disappearance of Tripsacoid characteristics when inbreds are crossed is not an ordinary effect of hybrid vigor but is due to antithetical dominance; to the action of modifiers which render most Tripsacum characteristics recessive in North American corn.

Pedicel lengths in mm. (below) and tassel lengths in cm. (above) for 15 Cornell hybrids, each compared with its inbred parents. Further explanation in the text.

In our opinion the whole question of inbreeding effects in corn needs to be restudied in the light of this theory. Dismissing the problem with the statement that corn is highly heterozygous is a description, not an explanation. The same principle may account for some of the peculiar variability of various other crop plants, though the force of antithetical dominance would be stronger in corn where a much wider cross is suspected.

  1. Mangelsdorf, P. C., and Reeves, R. G.,The origin of Indian corn and its relatives. Bull. Texas Agr. Expt. Sta., 574, 1-315 (1939)
  2. Bridges, C. B., Sex in relation to chromosomes and genes. Amer. Nat., 59, 127-137 (1925).
  3. Huelsen, W. A., Sweet-corn inbreds and crosses. Bull. Univ. Ill. Agr. Expt. Sta., 466, 277-355 (1940).