For a few years observations have indicated that teosinte has more tolerance to heat and drought and possibly more resistance to certain diseases and insect damage than corn. Efforts to improve inbred lines of corn by modifying them with teosinte characters have progressed far enough to give a suggestion of the results to be expected. Various Texas lines were crossed with Florida teosinte, backcrossed to corn from once to three times, and selfed each generation afterwards. In the development of the modified lines, no effort was made to select by observation among the segregates available for use. Plants were selfed at random, and only those plants or ears that were seriously affected with such abnormalities as disease, insect damage, and sterility were later discarded.

Tests of the desirability of the modified lines as compared to the original (unmodified) corn lines were of two kinds: (1) Tests of the lines themselves to compare their tolerance to artificially applied heat; (2) Yield tests of the various lines crossed to a common tester, conducted under field conditions.

1. Heat-tolerance tests. The procedures followed in making tests for tolerance to heat were based on those used for several years at the Kansas Agricultural Experiment Station, although in some respects there are considerable differences between the Kansas methods and mine. Inbred plants of Tx4R-3 ana of eight modifications of it were grown and given artificial heat treatments in an oven in six replications, each replication being grown and treated at a different time. Glazed pots with top inside measurement of four inches were used. The pots were selected for uniformity. The soil used for the first five replications was a thorough mixture of sandy loam and compost. That used for the sixth was relatively homogeneous Houston Black Clay.

In each replication, five pots of each line were planted, and an effort was made to have a final stand of two plants to the pot. This procedure usually resulted in 10 plants of each line for each replication.

The plants were given the artificial heat treatment when 13 to 15 days old. The oven used was electric, automatically controlled, with forced ventilation. It was designed for other purposes, and the fluctuation in the temperatures obtained led to some difficulties. However, after a few replications had been treated for practice, the method was found to be usable.

Prior to each application of heat, the soil in the pots was well-saturated with water. The pots were randomized in the oven and kept under heat treatment for eight hours at 55� C. After the treatment was complete, the plants were kept in the greenhouse for 5 to 30 days without water while the readings of the results were taken. It was found most practicable to take the first reading about 24 hours after treatment, because the extent of the damage to the plants was more readily determined after this lapse of time. The best method found of recording the results was to tabulate the number of days that each plant lived after treatment. In most of the replications no plants were living 10 days after treatment, and those which did live this long or longer were considered not to have been killed by the treatment.

For the purpose of analyzing and studying the results, it was found desirable to assemble all the data for each entry into a single score. In order to accomplish this objective, the combined number of days that all the plants of an entry lived after treatment was adopted as the score. Thus, in the fifth replication of modified line No. 1, the 10 plants lived the following numbers of days: 3, 6, 3, 20, 3, 17, 3, 3, 5, 15. But, since a plant is not considered to have been killed by the heat treatment when it lived 10 days or longer, all numbers above 10 were reduced to 10, and therefore the numbers actually added in order to get the score of this entry were 3, 6, 3, 10, 3, 10, 3, 3, 5, and 10. The score of this entry, therefore, is 56. The highest possible score is 100, and the lowest is zero. The score of each entry is shown in table I, the various lines being listed in descending order of their observed tolerance to heat:

Table I.

	Replications
Lines	1	2	3	4	5	6	Average
3	45	38	20	85	100	62	58.3
9	32	26	30	100	96	60	57.3
5	47	22	12	90	96	30	49.5
6	28	22	16	71	100	59	49.3
4	26	22	16	60	94	61	46.5
Tx4R-3	36	18	10	50	87	42	40.5
1	22	32	26	77	56	26	39.8
2	18	14	20	30	93	33	34.7
7	36	14	16	34	69	30	33.2

For significance, .05 = 14.6

Since the difference necessary for significance on the .05 level is 14.6, the indication is that two of the lines modified with teosinte characters are more tolerant to heat than the original line Tx4R-3. Whether tolerance to heat and to drought are related phenomena, as reported by some investigators, has not been determined in this study. However, the yield tests, to be discussed in the following paragraphs, were conducted with that possibility in mind.

2. Yield tests. One yield test was conducted each year from 1943 to 1946 on hybrids involving the group of Tx4R-3 lines tested for heat tolerance, and several tests were conducted on certain other groups. In all the yield tests, the uriiform tester was a single cross, commonly one with which the original inbred is combined when put into agricultural use. One or more checks were always included. Except where the contrary is indicated, one check was the original inbred crossed with the uniform tester, and various hybrids whose usual performance was known were often used as supplements.

The most satisfactory results of yield tests were obtained with groups of lines other than Tx4R-3 and its modifications. Although results of the heat tests indicate that additional tolerance has been introduced into Tx4R-3 by crossing it with teosinte, no field test has shown convincingly that the yielding ability of any of the modified Tx4R-3 lines should be adjudged superior to that of the original. Tests conducted during 1945 and 1946 showed only that some of the modified lines were in the same class with the original Tx4R-3 and that others were inferior. As would be expected, one or more modified lines gave actual yields greater than the original Tx4R-3 in each test conducted, but in none of these instances was the difference significant. It should be pointed out, however, that tolerance to drought did not have a fair chance to manifest itself in terms of yield in any test conducted on the Tx4R-3 group. In 1943 and 1944 the yield tests were a failure, principally because of poor stands and accidental damage. In 1945 and 1946 there was no appreciable drought during the critical part of the season.

More interesting results of yield tests were obtained with a group of modified Tx127C lines. A small portion of the results of the two tests conducted in 1945 and 1946 is shown in table II.

The 1945 test of the Tx127C lines contained 36 entries and the 1946 test contained 25 entries. Since the two tests did not contain the same entries, but had only certain ones in common, it is impracticable here to combine all the results briefly in one table. However, the following does include the highest-yielding entry and one check in each test. The lowest-yielding entry tabulated here from the 1945 test stood 14th among the 36 in the test, and the lowest shown from the 1946 test stood 16th among the 25 in the test. A blank indicates the omission of the entry from the test.

Table II.

	Average yield bu. per acre
Pedigree*	1945	1946
42116-21-2	44.8	59.5
42116-25-3	42.6
Tx. Hybrid No. 18 (Ck.)	40.8
42116-15-2	39.4	65.7
42116-27-1	38.2	49.3
42116-28-5	37.0	55.4
42116-28-4		45.6
Tx127C (Ck.)		44.0
Difference for significance, .05	7.26	9.75
Difference for significance, .01	9.63	12.06
*The tester in each instance was Tx173D x Tx203

It may be observed from these results that some of the Tx127C modified lines, such as 42116-21-2 and 42116-15-2, show considerable promise. It is interesting that some of them gave improved yields during a seasan when there was no serious drought or other hazard to which teosinte is known to be especially tolerant. Of course there are possible explanations. It seems fairly probable that the introduction of teosinte germ plasm into Tx127C resulted in modified lines with more remote relationship to the tester. Remoteness of relationship between the two parents of a cross is often believed to be an important factor affecting hybrid vigor. Another possible explanation is simply that additional "yield genes" have been acquired from teosinte.

A few teosinte-modified lines of Tx132A and Tx1O2A have been developed and tested, but the results to the present do not indicate appreciable improvement in them.

R. G. Reeves