These statements were based on observations and experiments with various cultivated plants. Different types of crucifers—kohl-rabi, brocoli, Brussel's sprouts, cabbage—were grown near each other; and the seed resulting from natural pollination, when grown, showed a large amount of intercrossing. The same observation was also made with different varieties of the radish, Raphanus sativus. These plants are all partially self-sterile, so that cross-fertilization is expected in somewhat greater degree than would result from random pollination. Mixing was also shown by plants which are generally self-fertile, such as tulip, hyacinth, anemone, ranunculus, strawberry, orange, rhododendron, and rhubarb. The fact that vicinism occurs when varieties of these plants are grown together is established by such observations, but this does not prove that one type of pollen is prepotent over the other. Somewhat more critical results were obtained from two other species. The monkey-flower, Mimulus luteus, was found to be highly fertile when insects were excluded. Uncastrasted flowers of a constant whitish variety were artificially pollinated by a yellowish variety; and of the 28 resulting plants all had yellowish flowers, so that the "pollen of the yellow variety completely overwhelmed that from the mother plant." Similarly a crimson variety of candytuft, Iberis umbellata, which was self-fertile, was crossed with a pink variety, the pollen being applied to uncastrated flowers as before, upon the stigmas of which he saw abundant pollen presumably from the same flowers. Out of 30 plants raised, 24 showed themselves to be crossed by the altered color of their flowers.
While indicating a selective action in favor of cross-fertilization, a number of conclusions might be drawn from these results. The crossed seeds may have germinated better and the plants grown from them survived in greater numbers. The types may not have been homozygous in their flower-color factors or the ovules may not have been receptive at the time the plant's own pollen was available, but they were when the other pollen was applied.
Darwin was familiar with so many cases of total self-sterility and was so convinced of the necessity for cross-fertilization that he was easily persuaded from these observations that a prepotency of pollen from dissimilar plants did exist, since he supposed this enabled a plant to choose between its own and foreign pollen when both were brought at the same time to the stigmas by insects or other agencies. So plausible have been the arguments in favor of such an assumption that the predominance of germ cells from individuals of somewhat different constitution, even where complete self-fertility exists, has been accepted as an established fact and incorporated in textbooks on biology.
Recent results obtained from mixed pollinations show that a directly opposite effect is obtained in several widely separated species of flowering plants. In many trials the plant's own pollen has been more efficient in accomplishing fertilization than that from other individuals which differ only in minor features. This same pollen, which is less effective when in competition with the plant's own kind of pollen, is fully able to function when not applied in mixtures. The results are noteworthy in view of the appreciable advantages which cross-fertilization gives to the immediately resulting seeds and the plants grown from them. In the experiments performed by the writer the material used consisted largely of self-fertilized strains of Zea mays which had been brought to uniformity and constancy and were considerably reduced in size and vigor. In this material, crossing increases the weight of seed within the same inflorescences as much as 50 percent in some cases. This greater amount of material is laid down in both the embryo and endosperm and is apparent in the greater size of the seeds. The crossed seeds also show a higher specific gravity and ripen earlier, as indicated by their lower water content at the close of the growing period. A large series of mixed pollinations show the ability of the cross-fertilized seeds to germinate better by an average of 16 percent. The resulting plants start to grow sooner, develop faster, mature in a shorter time, and at the end far surpass their self-pollinated sibs in size and reproductive ability. The hybrid vigor shown by maize is possibly greater than that displayed by any other plants in intraspecific crosses. In a study of a large number of crosses between inbred strains (Jones, 1918), production of grain was found to be advanced on an average of 180 percent, height of plant 27 percent, length of pistillate spike 29 percent, number of rows of spikelets on the pistillate inflorescence 15 percent, and the number of nodes on the plant 6 percent.
Notwithstanding these great immediate advantages to be gained, the plants manifest a decided preference for their own kind of pollen. This is a result which would have surprised those naturalists that were so keenly interested in methods of flower pollination in Darwin's time, but which, as I shall attempt to bring out later, is in agreement with other results from biological investigation.
In an experiment in which it was desired to compare the chemical composition of seeds of maize having different genetic constitution but produced under as nearly identical conditions as possible, advantage was taken of endosperm characters to enable proper classification of seeds produced in the same inflorescence. For example, two kinds of pollen from plants having yellow seeds and others having white seeds were mixed together and applied to a plant which normally produces uncolored seeds. The resulting yellow and white grains were distributed at random on the pistillate spikes and were developed under conditions as nearly comparable in respect to external and nutritional factors as it is possible to obtain. This same mixture was also applied to the yellow-seeded plants, and here also the two classes of seeds could be easily distinguished in most cases. On these plants the self-pollinated seeds were dark yellow while the crossed seeds were distinctly lighter in color, in most cases they had a white cap, and were as a rule readily separated.
The material used for these mixed pollinations consisted largely of inbred strains which had been reduced to uniformity and constancy by continued self-fertilization, so that the genetic differences between the two kinds of seeds sharply differentiated them, much more than in ordinary cross-pollinated varieties of this plant in which the yellow color is usually variable, owing to more than one hereditary factor for this color and various modifying conditions such as the consistency of the endosperm in respect to corneous and floury starch.
Many pairs of yellow- and white-seeded plants were treated in this way by mixing their pollen and applying to both types. After harvesting, it was realized that here was an excellent method of determining whether or not any selective action was shown by the plant's own pollen as compared to that from a plant of somewhat different type. If one member of the pair of plants which furnished the pollen for the mixture is designated A and the other B, the two kinds of seeds grown on A plants are A x A (self-fertilized) and A x B (cross-fertilized); on the B plants there are B x A seeds (cross-fertilized) and B x B (self-fertilized). Since the same pollen mixture is applied to both, the ratio of the seeds resulting from A pollen to the seeds resulting from B pollen on A plants should be the same as the ratio of the seeds resulting from A and from B pollen on B plants. In other words, the numbers form a proportion which, irrespective of the relative amounts of functioning A and B pollen in the mixture, should be a perfect proportion within the limits of random sampling if fertilization takes place equally. The end terms of the proportion comprise the self-pollinated seeds; and the middle terms, the reciprocally cross-pollinated seeds. If a true proportion is obtained, the products of the end terms should naturally equal the products of the middle terms. If they do not, the deviation is either in an excess of cross-fertilized or of self-fertilized seeds, indicating a selective action in one or the other direction.
The advantages of this method of testing are readily apparent. It is difficult to make up a mixture of large amounts of pollen in which the proportion of each kind is known. Measuring or weighing the pollen is not satisfactory, as maize pollen takes up moisture from the air rapidly and when any quantity is brought together in a humid atmosphere it becomes aggregated in a flocculent mass. Moreover, this pollen loses its viability rapidly so that even in case equal numbers of pollen grains could be had there would be no way of knowing the ratio of functional pollen grains in the mixture.
By mixing the pollen and applying it to both types of plants from which it is obtained, all of these difficulties are automatically overcome, and the experiment is as simple as could be devised. In all mixtures an attempt was made to have as nearly equal quantities of pollen as possible by measuring out the two kinds roughly. But in many cases the results showed that one kind of pollen was far more effective than the other. This, however, does not lessen the value of the figures, as it is the relative efficiency of each type of pollen, when it is applied to its own and to other stigmas, that we want to know.
Eighteen pairs of yellow- and white-seeded plants produced seed as the results of the application of mixed pollen. When these were counted and when the deviations of the proportions found from the closest perfect proportions were obtained, there were 12 pairs which showed a deviation in favor of the self-fertilized seeds and 6 in the opposite direction. The results as a whole showed a tendency to favor the plant's own pollen.
The inbred material used was so reduced in growth that the number of seeds produced on one plant was not large enough to give the results much weight. Additional mixtures were made, using a number of plants of two different self-fertilized strains to supply the pollen and applying this mixture to all of the plants of two different types. Most of the strains used had been self-pollinated for 6 generations or more, some as many as 10, so that the plants within one strain were practically identical in hereditary constitution. Pollen was collected from about the same number of plants as it was applied to. The two lots of pollen were put together into a paper sack and thoroughly mixed by shaking. This mixture was then applied to the plants of the two strains which supplied the pollen.
From 10 to 15 plants in each of the paired strains were pollinated, and from many mixtures 1,000-2,000 seeds were obtained on each lot of plants. The usual precautions were taken to prevent the entrance of extraneous pollen. The ear shoots were bagged before any silks appeared. The tassels were covered several days before pollen was used, so that any foreign pollen that may have lodged on the tassels had lost its viability, as tests (Kiesselbach, 1922) have shown that pollen rapidly loses its ability to function after 48 hours. The tassels that were not bagged were removed before pollinations were made, so that there was very little pollen in the air at the time the mixtures were applied. In spite of all precautions, out-crossed seeds regularly occur but in small numbers compared to the legitimate pollinations. The error from this source was detected when colors or other characters differing from either of the strains used were brought in by the undesired pollen. In a total of over 63,000 seeds in the first experiments only 30 visibly out-crossed seeds were found. A somewhat larger number of contaminations, however, undoubtedly occurred which could not be detected. But, giving a reasonable allowance to this source of unreliability, the results could not be seriously affected, as the error of this kind is not all in one direction.
Particular attention was given to the accuracy with which the seeds were classified. Strains were selected to be used which gave sharp differences between self-fertilized and cross-fertilized seeds, and in most cases separation was made very satisfactorily. In a few mixtures there was some doubt, and in two experiments the seeds on the yellow-endosperm plants could not be distinguished. In these two cases the seeds were planted, and classification was made with the mature plants. Also in all the other mixtures involving yellow and white endosperms a sample was taken and grown to determine the approximate percentage of error in separating the seeds. Since the self-fertilized seeds gave small inbred plants, pure for yellow or white color, while the cross-fertilized seeds produced large vigorous hybrids segregating into yellow and white seeds, classification of the mature plants was made without the least doubt. However, it should be noted that the better germination and greater vigor give the advantage to the cross-fertilized classes in every case if there is any difference. About 120 plants in each lot were grown, and the percentage of error obtained was used to calculate the total amount of misplaced seeds. Since the numbers of seeds were so large, it was impossible to grow all of them. In 25 out of 45 lots no faulty separations were discovered. In all of the other cases except one, the number of wrongly classified individuals was not more than 3 percent. These errors tend to balance each other, so that this source of doubt can be placed aside. Calculating the results with regard to the errors of classification gives practically the same result as when they are ignored.
In other mixed pollinations other characters were used which permitted even more positive classification than the single character difference of yellow and white endosperms. Plants with yellow sweet seeds were paired with white starchy-seeded plants. Each variety possessed one dominant endosperm character. Plants of the pop or Zea mays everta type were used in most of the mixtures for the white starchy seeds because the clear corneous endosperm differentiates very clearly the yellow and white seed color. In the reciprocal cross the smooth opaque kernels are clearly contrasted with the wrinkled translucent seeds of the sweet-corn type. In a similar manner purple sweet and white starchy types and sweet and shrunken endosperm were paired. In a few cases the purple-crossed seeds were not as distinct as could be desired, but the error is certainly small.
A few seeds were found in many inflorescences which had failed to reach a stage of developmment so that they could be classified. Seeds at the tips of the spikes and where the seeds were closely crowded were abortive. This introduces another source of unreliability, that of selective elimination of zygotes. Since crossing gives to seeds of maize an enormous advantage in development, it can be expected that as a rule more of the self-pollinated seeds will be found among the abortions than cross-pollinated. We are attempting to find out the relative fertilizing efficiency of pollen grains from different plants. But this can only be arrived at by counting the zygotes some time after fertilization has taken place. In the meantime a differential destruction of zygotes may have taken place. This effect must be considered in any organisms employed; but because of the short time that elapses between fertilization and the maturation of the seeds, and from the fact that they develop in an exceedingly favorable and uniform environment, maize is the very best material the writer can think of in which this problem can be attacked, especially considering the large numbers which can be obtained. All animals and those plants which do not show xenia in the seeds have the objection that a comparatively long time intervenes between fertilization and sufficient development to permit classifi-cation. In plants many cases are known in which there is a selective elimination of certain classes of individuals owing to a lesser germination and unequal ability to grow. In Drosophila (Hyde, 1914) cross-fertilization does not influence the number of eggs laid as would be expected, but markedly regulates the percentage that hatch. Therefore, the error from this source always tends to show an apparent deviation in favor of cross-fertilization. In plants such as maize, where the seed progenies can be classified, selective elimination is at a minimum and probably is not sufficient to affect the numbers appreciably; but, since the tendency is in the opposite direction to the results which have been obtained, the data are even more convincing.
FIG. 6.—White starchy and yellow sweet types of maize resulting from
the application
of the same mixture of pollen from the two kinds of plants that produced them.
The figures obtained from the preliminary experiments in which pairs of single plants only were used need not be given here because the numbers are too low to give the results much weight. It is sufficient to state that the data taken together indicate a slight selective action favoring the plant's own pollen.
Twenty-three pollen mixtures have been made in which pollen was applied to several plants of the same strain in each case. Some of these results have been published (Jones, 1920). A summary of all the pollen mixtures including additional, unpublished results is given here in Table V. The material used in these pollen mixtures is described as follows, the pedigree numbers corresponding to those used in the table. Several different strains from the same original source were used.
No. 1. Several distinct strains from a yellow
dent variety originally obtained in Illinois and known as Chester's Leaming;
self-fertilized 10 or more generations.
No. 10. A strain with white floury seeds with no traces of
corneous starch; self-fertilized 9 generations.
No. 14. Two distinct strains from a yellow dent variety from
Connecticut known as Stadtmueller's Leaming, selected for high protein content
during 6 generations of self-fertilization.
No. 20. Two distinct strains from a white dent variety originally
selected for high-protein content at the Illinois Agricultural Experiment
Station and further selected during 4 generations of self-fertilization.
No. 21. One strain from the same source as above but selected
for low protein in field-pollinated cultures and during 5 generations of self-fertilization.
No. 65. A small-, white-, round-seeded strain from a variety
of pop corn, Zea mays everta; self-fertilized 10 times and characterized
by clear corneous starch.
No. 75. An inbred strain of sweet maize of miscellaneous
origin having reddish leaves and husks and white, translucent, and very much
wrinkled kernels.
No. 76. Two similar strains from a sweet variety of latent
flint type having purple aleurone and known as Black Mexican; self-fertilized
4 years.
No. 77. An inbred strain from a sweet variety of latent dent
type with deeply wrinkled, white seeds known as Evergreen; self-fertilized
4 generations.
No. 117. A variety of pop corn with sharp-pointed seeds having
clear corneous endosperm; self-fertilized 3 times.
No. 126. A small, early-maturing, sweet variety of variety
latent flint type with dark yellow kernels, known as Golden Bantam; self-fertilized
3 generations.
No. 142. A variety of pop corn known as California rice pop
having the smallest seeds known in maize. The plants are medium in size, rather
late in maturing, and produce several very small ears. This material has been
self-fertilized until it is very uniform.
No. 146. Another variety of Golden Bantam sweet corn from
a different source than No. 126 and somewhat different in type; self-fertilized
2 years.
No. 220. Another variety of Golden Bantam sweet corn similar
to 126 and 146; self-fertilized 3 generations.
No. 285. A variety of mixed origin having white, shrunken
seeds.
No. 375. A variety of yellow sweet maize, known as Golden
Giant, having larger plants and ears than Golden Bantam.
No. 405. A white, shrunken endosperm type that originated
from a large dent variety known as Pride of Saline.
The pollen mixtures Nos.1-9, inclusive, comprise various inbred strains differentiated only by yellow and white endosperm color. All show marked heterosis in the crossed seeds and in the resulting first-generation hybrid plants. A sample of all the different lots of seed secured from these mixtures was grown to test the accuracy of classification. Mixtures Nos. 10, 11, and 12 were made with first-generation hybrids, one having all yellow seeds, the other all white. The second crossing gave still more increase in vigor, although not as great as the stimulus following the first cross. The plants used were vigorous and productive, and a large amount of seed was obtained from a few plants. It was desired to know whether the same selective action would be shown by vigorous plants with segregating gametes as contrasted with non-vigorous plants whose gametes were all alike. The seeds were easily classified, and the amount of error when tested was found to be quite low. Mixtures Nos. 13 and 14 involved yellow, wrinkled, sweet seeds in one member of the pair and white, smooth, starchy endosperm in the other. Each contributed one dominant factor, so that the differentiation was perfectly distinct in the reciprocal applications. In mixtures Nos. 13 and 16 it was intended to use the same characters as in Nos. 13 and 14, but the yellow sweet seeds, when planted, turned out to be all crossed with a white-seeded sweet strain. The seed had undoubtedly been wrongly labeled. Self-pollinated, these plants gave ears that were all sweet but were segregating for yellow and white color.
This material would not ordinarily have been used, but the selective action is very high in these mixtures and, since similar results have been obtained later when uncrossed seed was used, results are given for comparison with the other material. The effects of the starchy-carrying pollen showed up all right among the all-sweet seeds, but the reciprocal cross-pollination differentiated only the yellow cross-fertilized seeds. The white cross-fertilized seeds could not be distinguished from the self-fertilized seeds. But since half of the pollen grains carried yellow and half of them white, the number of yellow seeds can be doubled to give the total number of cross-pollinated seeds on the starchy ears, and an equal number of white, cross-fertilized seeds subtracted from the white seeds. This increases the probable error of random sampling somewhat; but, since the number of yellow seeds is very low in coin-parison with the white in these mixtures, the data are reliable in view of the great selective action shown. That the yellow color in this material was a unit-factor difference and that there were equal number of pollen grains carrying yellow and white, is proven by the self-fertilized ears produced by the hybrid which gave a monohybrid ratio (actual count, 318 yellow and 7 white seeds). The starchy crossed seeds on the hybrid plants were of two kinds, yellow and white, and were produced in equal numbers. Since the ovules of the heterozygous plant were segregating equally and the self-fertilized seeds gave the monohybrid ratio, the pollen grains must have carried the two colors in equal amounts.
In the next four mixtures the characters, purple sweet and white starchy, were used. In Nos. 17 and 18 the plants were not productive and the numbers of seeds are low. Also the separation of purple starchy cross-fertilized seeds and white starchy self-fertilized seeds was not as sure as in the other mixtures. In mixtures Nos. 19 and 20, satisfactory numbers were obtained and the differentiation was clear-cut on both sides.
In No. 21 two very distinct types are tested. Pollen from a yellow sweet type with large seeds is mixed with that from plants having the smallest seeds and ears known in maize. In foliage, tassel type, and pistillate inflorescence it is quite different from all other kinds of maize. Its aleurone color composition is AA CC RR P1P1 II, having all complementary factors for purple aleurone; but this is prevented from showing by the dominant inhibitor.
Still another set of character differences is used in mixtures Nos. 22 and 23. When the dominant allelomorphs of sweet and shrunken endosperm are brought together, the seeds are normal starchy. Being smooth and plump, they are quite distinct from either recessive type. Plants having shrunken endosperm from two very different sources were used. In No. 22 the plants carrying shrunken had previously been crossed with several common varieties and the recessive character extracted. In mixture No. 23 the shrunken material originated from a large, white-seeded dent variety in which shrunken kernels had been found on one ear. When tested with the other character of the same appearance, the shrunken factors were found to be the same.
The number of seeds obtained from each mixture, the numbers in the four groups, and the deviations in percent from the closest perfect proportion are given in Table V. If the deviation is in an excess of self-fertilized seeds, it is plus; and if in the opposite direction, minus. The deviations can range from +50.0 to -50.0 percent. These extreme values would result if there were complete non-functioning of each kind of pollen on one set of plants and exclusive functioning on the other.
The total number of seeds obtained from these 23 pollen mixtures is 76,620. Large numbers are necessary in any test of differential fertilization. Mixtures Nos. 6, 7, 13, 17, and 18 are less reliable than the others, because they have rather small numbers from either the A or the B plants—less than 200 in each case. The value of each set of figures is limited by the numbers in the less populous half of the proportion.
Over 1,000 individuals in each of the four groups were obtained in some of the mixtures. Since the experimental error is low and not all in one direction, as has been shown, and the selective elimination of zygotes tends to obscure the result which has been obtained, these figures are convincing. Of the 23 mixed pollinations, 20 show a deviation indicating a selective action in favor of the plant's own pollen, while 3 of the mixtures show the opposite result. These 3 are all low in numbers on one or the other side of the proportion. Mixtures Nos. 6 and 7 could not be classified by the seeds on the yellow-seeded plants, so consequently the progenies were grown and classified at maturity. This brings in other sources of error—differential germination and competition between plants which are weak with those that are vigorous—which certainly tend to result to the apparent advantage of cross-fertilization. Some of the deviations in favor of self-fertilization are too small to be significant; but, disregarding these, so many of the pollen mixtures show such a marked deviation from random assortment that the conclusion is inescapable that in maize the plant's own pollen is more effective in consummating fertilization than pollen from plants of only slightly different construction. This selective action is shown even though the foreign pollen is perfectly capable of fertilizing the plants when not acting in competition with the plant's own pollen, as has been definitely proven.
Mixtures Nos. 15, 16, and 17 show 47 percent out of a possible 50 percent deviation—almost complete non-functioning of the dissimilar pollen. Nos. 15 and 16 include the first-generation hybrids in which allowance has to be made for segregation, but the results can be discounted but very little. Mixture No. 17 is low in numbers, having only 104 seeds on the B plants, of which are cross-fertilized. But when this same mixture is applied to the A Plants, only 3 cross-fertilized seeds are to be found among 1,303 self-fertilized seeds. Surely there is some powerful action working to hold back the pollen from the dissimilar plants. Mixtures Nos. 3, 4, 8, 10, 19, 20, 21, and 23 are the most convincing, as in these the numbers are large, the differentiation of the seeds in both groups is precise, and the deviations clearly show the superiority of self-pollination.
Taking all the data in consideration, magnifying the actual experimental error to its fullest extent, and giving due allowance for variations inherent in an experiment of this kind, the conclusion can be no other than that these plants manifest a definite receptiveness to their own pollen, discriminating against foreign pollen even though it comes from plants only slightly differentiated from them, both of which might easily be descended from the same individual at no very distant period back. This selective action is shown by plants of weak growth or full vigor, whether each strain descended from a line of similar ancestors or whether its immediate parents were diverse and, finally, irrespective of the plants being homozygous or heterozygous. The one significant feature in common in all these experiments is the fact that the cytoplasm is the same as the medium in which the pollen fulfils its function.
Mixed pollinations were also made with the garden tomato, Lycopersicum esculentum Mill. Advantage was taken of plant characters such that the seedlings could be distinguished in both reciprocal applications. A variety with unserrated leaves with a tall habit of growth was paired with a dwarf variety having serrated leaves as shown in Figure 7. Tall stature and serrated leaves are dominant characters so that the cross-fertilized and self-fertilized seedlings from one variety are differentiated in the early seedling stage by the leaf formation and in the other variety by the length of the stems. Dwarf plants are characteristically shorter and more compact in stems and leaves, which gives them a distinct appearance. Two mixtures of pollen were made, and the numbers of plants obtained were 340 and 272. Separation was easily made between the serrate and non-serrate leaved seedlings, but the tall and dwarf plants could not be distinguished clearly in those seedlings that were somewhat stunted in growth. These doubtful plants were set in the field, but even at the end of the season classification was not made with certainty in every case.
FIG. 7.—Two types of tomatoes differing in serration of leaf and in stature
give hybrids which can be distinguished from both parental varieties.
The results, such as they are, agree with those from maize. There is a deviation favoring the plant's own pollen of +2.06 and +6.84 percent in the two mixtures. In both mixtures the differences may be due to random sampling; but in view of the probability that differences in the germination of the seeds and viability of the plants tend to decrease the proportion of self-fertilized individuals, the figures are somewhat significant.
Heribert-Nilsson (1920) found that Oenothera Lamarckiana. was pollinated more readily by its own pollen than by a mutant form gigas which originated from it. How the two types compared in the reverse pollination is not stated. Here a chromosome reduplication is involved so that these results may not be comparable to those obtained from other plants. The gigas plants develop more slowly and flower later; and the growth of the pollen tubes may be similarly reduced, whether traveling in tissues of their own or different plants.
In 1911, Balls reported the results of mixed pollinations with cotton. Two quite different types of cotton were used: the Egyptian, commonly grown in North Africa, and the Upland, which includes most of the varieties of cotton grown in North America. Equal quantities of pollen were put on the stigmas of both types. It was found that the seed produced by the Egyptian flowers to which had been applied both the plant's own pollen and pollen from Upland plants gave 10 hybrids out of 330 plants—less than 3 percent—in spite of the fact that in this cross the hybrid plants are notably more vigorous and more likely to survive. A reciprocal application was made to the Upland type, and about the same proportion of hybrids was obtained.
Both the Egyptian and Upland types were also pollinated with a mixture of the plants' own pollen and pollen from the first generation hybrid of these two types. In both cases the F1 pollen was more effective than the other pure type, being represented in 20 percent of the seedlings from Egyptian plants and in 28 percent of the seedlings from Upland plants, but was not as effective as the plant's own pollen.
More recently, extensive investigation with these same types of cotton have been made by Kearney (1923) and Kearney and Harrison (1924). Owing to the nature of cotton pollen, the mixed pollinations were made by brushing the stigmas of the emasculated flowers with the clustered anthers of the staminate flowers. Pollen from one type was applied in this way, followed immediately by the pollen from the other type. To half of the flowers one type of pollen was applied first and to the remaining half the other type was put on first. This is not as satisfactory as mixing the pollen where possible. In every case the pollen which was applied first gave relatively more fertilizations; but even where the unlike pollen was applied first, it never fertilized 50 percent of the ovules. Large enough numbers were employed to make the results significant in view of the marked selective action obtained.
In two experiments a total of 2,349 plants from Egyptian seed from mixed pollen gave 25.8 ± 0.61 percent hybrids, and 1,419 plants from Upland seed produced 27.2 ± 0.80 percent hybrids. While these figures show considerably more fertilizations by the unlike pollen than Balls found, a marked selective action is apparent favoring the plant's own pollen.
These are the results obtained by representatives of the Gramineae, Solanaceae, Malvaceae and Onagraceae, widely separated orders in the angiosperms. What is the cause of this differential fertilization? The plant's own pollen may prevent or retard the germination of the foreign pollen. The pollen tubes may grow faster in the plants of the same germinal constitution than in other plants, or, when the generative nuclei are brought to the embryo sac, fertilization may take place more readily between like nuclei than unlike.
 |
| FIG. 8.—PistiIlate inflorescence of maize at the time of pollination, with leaf sheaths removed. (About one-half natural size.) |
Five pollen mixtures were made to test this. The plants used were white-seeded, pointed pop corn of the variety known as Squirrel Tooth, and a yellow-seeded, sweet type known as Golden Bantam. First-generation hybrid plants of two strains, self-fertilized several generations, were used in both types. The plants were vigorous and productive and very uniform. In previous pollen mixtures (Nos. 15 and 16 in Table V) this material had given the greatest selective action. Pollen was collected from bagged plants, thoroughly mixed by shaking in a closed paper sack, and applied to about 10 plants of each type in the several mixtures. Just before the pollinations were made, the pistils were cut-off evenly at a short distance beyond the tip of the spike, and the pollen was applied to the cut ends of the filaments. The distance that the pollen tubes had to travel to reach the ovules differed considerably in the case of the seeds in the upper half as compared to the lower half. The mature pistillate inflorescences ranged from 10 to 20 cm. in length. At the time fertilization took place they were considerably shorter than this. It is estimated that the pollen tubes traveled through a distance which varied from 5 cm. to 15 cm. If the plant's own pollen tubes grow faster than the foreign tubes, we would expect fewer cross-fertilized seeds at the base of the spike than at the tip. The selective action exhibited in these mixtures, together with the number of seeds, is shown in Table VI. The white, smooth-seeded pop corn plants are designated A in this tabulation, and the yellow wrinkled sweet corn as B. The deviation ranges from 12 to 41 percent out of a maximum of 50 percent and shows a marked selective action in favor of the plant's own pollen.
The highest deviation of 41 percent agrees fairly well with the results obtained in mixtures Nos. 15 and 16, where similar material was used. But why there should be such a range of deviations from 12 to 41 percent in the same plants is not understood. Mixtures Nos. 24 and were made in 1920, and mixtures Nos. 26, 27, and 28 in 1921; but the plants were all grown from the same two lots of seed. Some of the mixtures were made shortly after the silks appeared; while in others, pollination was delayed several days after fertilization normally takes place. It was thought that the plants might not put such a restriction on foreign pollen when fertilization was delayed.
TABLE VI
The amount of selective action in favor of the plant's own pollen
shown by plants of the same strains of maize in fibe pollen mixtures.
| Pollen |
No. of seeds in the combination |
Total no. |
Deviation
from perfect proportion (percentage) |
|||
| A x
A |
A x
B |
B x
A |
B x
B |
|||
| 24 |
811 |
11 |
381 |
2,006 |
3,209 |
41.35 |
25 |
4,222 |
27 |
466 |
1,404 |
6,119 |
37.22 |
| 26 |
1,568 |
2 |
319 |
224 |
2,113 |
20.56 |
| 27 |
1,930 |
20 |
73 |
309 |
2,341 |
39.71 |
| 28 |
4,084 |
6 |
963 |
290 |
5,343 |
11.50 |
This was tested by making a mixture of pollen, using the same material and applying part of the mixed pollen to plants of both types in which the pistils were fresh and others in which they were old. In the fresh lot the pistils did not extend more than 5-8 cm. beyond the tips of the inflorescences and had been receptive less than 24 hours, In the old lot the pistils had grown to a length of 10 cm. or more and probably had been receptive to pollen for more than 24 hours. The fresh lot gave a deviation of 26.0 percent of self‑fertilized seeds, while the old lot gave 26.9 percent. Evidently some other explanation than age of pistils must be sought for the varying results from the same material.
TABLE VII
The ratio of crossed seeds in the upper and lower halves of ears resulting
from mixed pollinations.
| POLLEN MIXTURE NO. |
FLINT TYPE | TOTAL NO. OF SEEDS |
NO. OF CROSSED SEEDS PER 100 |
RATIO OF CROSSED SEEDS Top:Bottom |
|
| Top | Bottom | ||||
| 24 | White starchy | 822 |
1.99 |
.54 |
3.7:1 |
| 24 | Yellow sweet | 2,387 |
17.23 |
14.64 |
1.11 |
| 25 | White starchy | 4,249 |
1.22 |
.05 |
24.4:1 |
| 25 | Yellow sweet | 1,870 |
26.66 |
23.01 |
1.2:1 |
| 26 | White starchy | 1,570 |
.25 |
.00 |
.25:0 |
| 26 | Yellow sweet | 543 |
60.07 |
57.14 |
1.1:1 |
| 27 | White starchy | 1,959 |
2.54 |
.32 |
7.9:1 |
| 27 | Yellow sweet | 382 |
24.73 |
13.78 |
1.8:1 |
| 28 | White starchy | 4,090 |
.10 |
20 |
.5:2 |
| 28 | Yellow sweet | 1,253 |
78.43 |
75.35 |
1.1:1 |
| Total and average |
White starchy Yellow sweet |
12,690 6,435 |
1.22 41.42 |
.22 36.78 |
5.6:1 1.1:1 |
The ears of the five pollen mixtures given in Table VI were divided arbitrarily into top and bottom halves, and the numbers in these two parts are arranged in Table VII. In every case but one there are more crossed seeds in the upper halves of the ears. In the one exception only 6 crossed seeds out of 4,090 were obtained, so that their position on the ears is hardly significant. In all the other cases there are relatively less cross‑fertilized seeds in those seeds which resulted from the pollen tubes which had to grow the longer distances.
Although all but one lot show more crossed seeds in the upper halves, it will be noticed that the differences are much greater on the pop corn ears than on the sweet type. Most of the pronounced selective action in the mixtures of these two types of maize is shown by the one type, and undoubtedly pollen-tube factors described in the preceding chapter cause most of the difference in fertilizing ability.
It should be noted that in mixtures Nos. 26 and 28 where the deviations are low, showing less restriction on the unlike pollen, the ratio of crossed seeds in the top and bottom halves is no more in the white starchy type than in the yellow sweet-in fact, slightly less, in marked contrast to the other mixtures. For some reason these plants had lost some of their ability to erect a barrier to the pollen from dissimilar plants. The fact that all of the mixtures of some i different varieties representing most of the distinct types of maize exhibit an unequal rate of fertilization in pollen mixtures shows how numerous and well distributed these pollen-tube factors are. Some of them are much more pronounced in their effect than others.
That the selective action is due to something else than the factors which determine the endosperm differences in color and texture which are used to separate the seeds is proved in the following manner. Plants were grown from the crossed seed of pollen mixture No. 15 in Table V which gave the greatest difference in favor of the plant's own pollen. One of these plants was self-fertilized; and from the resulting seeds which were segregating for yellow and white color and smooth and wrinkled texture, white smooth and yellow wrinkled seeds were selected and the plants grown from them again self-fertilized. This was repeated until after four generations the two original combinations were regained, each breeding true for the same endosperm color and texture factors used in the original mixture. Three pollen mixtures were made with this material. They gave deviations of +0.98, -0.17, and -2.65. No such selective action, as was originally obtained from this combination, +47.75, was shown after crossing and extracting the same endosperm characters.
It would be possible to obtain the same combination of endosperm characters and pollen-tube factors, and, if that had been the case, there would have been no way of separating the differential fertilizing effect from the visible different factors themselves; but since this is not the result that was obtained, it is clear that whatever causes the difference is not derived from the particular endosperm characters themselves.
How these factors operate to reduce the fertilizing ability of the unlike pollen when two kinds of pollen are in competition is not known. The evidence is conflicting. There is the possibility of a difference in the time of germination of the pollen grains and in their rate of growth. Pollen may germinate and grow equally well when not acting in competition; but when more than one kind is present at the same time, one lot of pollen may inhibit the germination or growth of the others.
Miller (1919) has observed that in maize many pollen tubes may start to grow down the style, but in about 100 examinations only 1 tube was seen to reach the ovary cavity in every case. Vanilla has species with flowers having long columns and others with short columns. McClelland (1919) found that pollen of the long type fertilized the short type flowers more readily than the plant's own pollen. In contrast to this are the results that Tokugawa (1914) obtained from long-styled and short-styled species of Lilium, whereby the pollen tubes of either species grew more rapidly in the pistils of the same species than the pollen tubes of the other type.
The flower of Oenothera Lamarckiana. has a comparatively long style. By cutting this off just above the ovary at intervals alter pollination, Heribert-Nilsson, in the experiment previously referred to, found that Lamarckianna pollen tubes reached the ovary in a shorter time than those from the pollen of a mutant form gigas. Different flowers on the same plant were used. The rate of pollen-tube growth varied with the temperature, but under equal conditions the plant's own pollen in one comparison accomplished fertilization in 20 hours while the other pollen required 21 hours.
A similar experiment was carried out by Kearney with cotton, but with different results. Some flowers of the Pima variety of Egyptian type were pollinated with their own pollen; and others, with Upland pollen. The pistils were then cut off at 8, 10, and 12 hours after pollination. No seeds were obtained in 8 hours, but after 10 and 12 hours the bolls developed from 2 to 18 percent of a normal amount of seed, as shown in Table VIII. No differences in favor of the like pollen are indicated by these results in the rate of pollen-tube growth, as somewhat more seed resulted from the off-pollination than in the self-pollination during an equal period of pollen-tube growth. Also in cotton when two different kinds of pollen are present together, the plant's own pollen tubes grow no faster than the others apparently, as shown in an experiment performed by Kearney and Harrison in which the bolls resulting from a pollen mixture from Egyptian and Upland plants applied to the same types were divided into upper and lower halves and the two lots of seed grown separately. In cotton, with its single pistil, the faster-growing tubes reach the ovules in the upper part of the ovary and fertilize them first, leaving the slower-growing tubes to pass on down to the ovules in the lower part. More hybrids would therefore be expected in the lower halves of the cotton seed bolls, directly opposite to the situation found in maize, if there were the same differences in pollen-tube growth. But no significant differences were obtained by Kearney and Harrison; so apparently the tubes travel at the same rate. An examination of the two kinds of pollen in sugar solutions showed them to be equally viable. Also the pollen mixtures yielded approximately equal percentages of hybrids when applied to each type, indicating equal viability.
TABLE. VIII
The amount
of seed produced after stated intervals of pollen-tube growth
of two different types of cotton compared on the flowers of one type.
(Data from Kearney and Harrison in the Journal of Agricultural Research)
