Bulletin of the Torrey Botanical Club, 59(3): 109-118 (Mar. 1932)

Control of gametophytic selection in Datura through shortening and splicing of styles1

1 This investigation was made possible by a grant from the Committee on the Effects of Radiation on Living Organisms, National Research Council.
2 University of Illinois.
3 Carnegie Institution, Cold Spring Harbor, N. Y.
4 Buchholz, J. T. and A. F. Blakeslee. Pollen-tube growth and control of gametophytic selection in Cocklebur, a 25-chromosome Datura. Bot. Gaz. 90: 366-383. 1930.

In the course of our experiments on the control of gametophytic selection, we found that restricted pollinations together with a division of the seed capsules resulted in a very marked increase in the pollen transmission of extra chromosomes in Datura.4 This increase has, for the lower part of the seed capsule, ranged from 49 to 65 per cent with pollen from the 25-chromosome plant Cocklebur. By excision of the part of the style containing the slow-growing pollen tubes in similar pollinations, we obtained a complete non-transmission of the extra Ck. chromosome. In the experiments concerned in this paper, we shall describe a method by which a part of the style containing the more rapidly-growing pollen tubes may be excised and discarded. It follows that this manipulation should result in a nearly perfect pollen transmission of the extra chromosome, permitting the use of more pollen or making unnecessary the division of seed capsules. This procedure necessitated the development of a special technique for reuniting the two parts of the style which had been severed and shortened, or the joining of two different styles in a particular place.

5 Buchholz, J. T. and A. F. Blakeslee. Radium experiments with Datura. I. The identification and transmission of lethals of pollen-tube growth in F1's from radium-treated parents. Jour. Heredity. 21: 119-129. 1930.

Recently, the discovery of many genes obtained from irradiation whose effect on pollen-tube growth renders them non-transmissible, or only slightly transmissible through the pollen,5 has stimulated renewed efforts in seeking methods for the control of gametophytic selection. These methods, therefore, fill a definite need in furthering the study of the lethals of pollen-tube growth. Aside from the splicing of styles, there are other manipulations which may ultimately prove successful where a mere shortening of the style is desired, as in the case of self- and cross-sterility.

6 Buchholz, J. T. The dissection, staining and mounting of styles in the study of pollen-tube distribution. Stain Technology, 6: 13-24. 1931.

Some interesting facts concerning pollen-tube behavior were discovered in the course of these investigations which may be advantageously described here. It happens that our technique for the study of the position of the pollen tubes and their distribution within the style at any desired time after pollination6 has enabled us to observe many phenomena of pollen tube behavior, as well as to determine the success or failure of any type or condition of pollination. Without this technique, it would be difficult or impossible for us to know the degree of success attained in attempted crosses or the cause of failure in the various types of pollinations.


We have made attempts to shorten the style by cutting off a portion and pollinating a small area of conducting tissue thus exposed on the wounded surface. In these trials, we found that, due to dessication, the pollen usually would not germinate; when protected against dessication about 20 pollen grains, at best, could be induced to germinate under the most favorable conditions. The exposed area of the conducting tissue was evidently too small. Cutting in diagonal increases this area slightly, but this gives a pointed, more exposed end.

7 East, E. M. and J. B. Park. Studies in Self sterility
II. Pollen-tube growth. Genetics. 3: 353-366. 1918.

By splitting the style through to the mid-point, placing pollen grains in this wound in direct contact with the conducting tissue, and holding these sides together by means of a straw cut from a grass culm, the pollen was found to germinate very satisfactorily. However, an examination showed that the pollen tubes were growing in both directions. Many of the pollen tubes (sometimes more than half of them) were directed toward the stigma end, where they stopped very close to the exposed style ends from which the stigma had been removed in attaching the straw. Others were growing in the normal direction toward the ovary. This peculiar behavior is of considerable interest in relation to chemotropism and pollen-tube growth. Regardless of this reversal in the direction of some of the pollen tubes, we became convinced from an examination of the dissected preparations, that, as a method of shortening the style, the pollination of the interior of a split style with suitable protection against dessication might be successful in practice, where a shortened pistil is desirable. East and Park7 reported that they obtained two seed capsules in incompatible tobacco plants by the pollination of pistils which had been mutilated.

The reversal of pollen tubes calls to mind another observation first made nearly ten years ago. At that time, flowers were pollinated, kept in a cool, moist place for several days, and then re-pollinated with the idea of observing the amount of change in the growth-rate of pollen tubes in the partially depleted style of an old flower. A very pronounced decrease in the rate of growth was obtained in the pollen tubes from the second pollination, but a much more interesting phenomenon was observed, namely; that some of the pollen tubes from the first pollination were found to be growing in the reversed direction. They appeared to be emerging from the ovary and growing toward the stigma. An examination of the styles of 30 or more flowers two- and three-days after pollination (self and open pollinated flowers in the field) showed that in more than half of these preparations anywhere from 1-11 pollen tubes were growing in the direction of the stigma. In flowers only a day old, there were very few reversed pollen tubes.

Possibly we should revise some of the commonly accepted ideas concerning chemotropism and pollen-tube growth in the light of these observations. There is certainly a question as to whether chemotropism, if present, exerts its influence through the length of the style. If a pollen tube happens to become reversed in its orientation at any time, it will continue its growth toward the stigma. On the stigma there is only one direction in which a pollen tube may grow and find nutriment when the pollen grain germinates. All cells of the stigma and conducting tissue are so oriented that the path of least resistance (the long axis of the nutritive cells) leads them in the direction of the ovary. Once the pollen tubes are started in this direction there is usually nothing to reverse their orientation, but within the ovary, we find that the path followed becomes very devious and irregular. (We have successfully dissected the entire conducting tissue from the pistil, removing the ovules which remained attached to the dissection by their pollen tubes.) If the ovules in the path of a particular pollen tube have been previously fertilized, and it happens to find its way back to the conducting tissue through which it entered, it may emerge from the ovary and continue its growth toward the stigma. Of course, these observations do not necessarily deny the existence of chemotropism at the time a pollen tube enters the micropyle.


Our experiments, in which we spliced styles on cut flowers which were kept in the laboratory, revealed some very interesting features. We found that when the parts of two different styles were joined perfectly in square joints, properly protected, and held in this position, many pollen tubes could pass across the wounded area. When the styles were not perfectly joined, the ends of the pollen tubes would enlarge considerably and extend themselves slightly beyond the conducting tissue as they were stopped. Figure 1 is taken from one of our laboratory tests, and shows the behavior of pollen tubes at the joint between two styles. This figure is from a camera drawing of a selected part of this region and the conducting tissue is only schematically represented. The pollen tubes shown at the left had little difficulty in passing through the wounded area; those shown at the right were held in check by the gap at this point between the two neighboring conducting tissues. Some of the pollen tubes were attempting to cross the barrier and the figure shows how they may crowd up to and extend themselves beyond the extreme end of the upper cut surface. With the first tubes as support, others may extend themselves farther and farther. Thus it could be observed in many instances that when crowded, a few of the pollen tubes may sometimes actually find their way across a considerable space to the nourishing tissue beyond the gap and continue their growth to the ovary.

These observations demonstrated the importance of obtaining a perfect union of the two strands of conducting tissue in the styles which are spliced. Though some pollen tubes are held in check at the joint, we counted hundreds of pollen tubes in our test slides which successfully passed over into the second style. We were, therefore, encouraged to make attempts to obtain seed capsules from such pollinations.


The conditions under which it may be desirable to splice a style are illustrated by the pollen tube distributions shown in figures 2 and 3. Here we find the pollen tube population, growing under favorable conditions, and resolved into two groups. Only the ends of the pollen tubes were counted and plotted in their appropriate interval as they were growing from the stigma at the left toward the ovary at the right. Burst pollen tubes are plotted below the datum line and any ungerminated pollen grains are shown by the vertical bar at the left. Figure 2 shows a test using a moderate amount of the pollen of an F2 plant derived from irradiation experiments. The bi-modal distribution is due to a gene which is recessive in the sporophyte carrier. Genes of this type are usually transmitted through half of the eggs, giving very satisfactory 1:1 ratios. Half of the male gametophytes also receive the gene which retards their growth. The group of long pollen tubes are growing at the same rate as those obtained from the pollen of normal plants. Their arrival in the ovary several hours before the second group, may enable them to monopolize the ovules and leave none to be fertilized by the second group of pollen tubes which, therefore, may appear not to be transmitted through the pollen in ordinary pollinations. Since the pollen tubes in the longer group could only contribute the dominant allelomorph of this gene upon fertilization it is desirable to eliminate this group entirely and obtain a result similar to a back cross to the recessive, if this plant is self ed.

Likewise, in figure 3 from the pollen of one of the 23-chromosome plants where an extra chromosome is carried in the slow-growing half of the pollen tube population, we may wish to eliminate the longer pollen tubes which do not carry the extra chromosome. Since an extra chromosome behaves as a dominant character in heredity we may expect to approach 100 per cent transmission of this condition, if the fast growing pollen tubes are eliminated.


The splicing of styles on plants growing out of doors was most successful when done at night. The method which was finally elaborated is illustrated by steps in A-J, in figure 4.

Healthy, cut flowers from normal plants were pollinated under controlled conditions. Pollen was used which was known to give a bi-modal distribution of tubes. After sufficient growth had been attained to separate the two groups (8-12 hours at 18-20 degrees C.), one or more styles were killed, dissected, and stained as rapidly as possible (at least 30 minutes required). When properly stained, the condition of the tubes and the position of the two modes were determined by actual measurements. The distance from the stigma to the low point X between the two modes was measured and marked with parallel lines on a wooden gauge. This gauge served as a guide in cutting the styles in the field. While the test slide and gauge were in process of preparation, the styles were removed from the remaining flowers and inserted into previously selected, closely-fitting grass straws C, as shown at A (the living culms of Digitaria sanguinalis were used). These were immediately set in vials containing a small quantity of water (fig. 4B) and transported to the field in this condition.

All the paraphernalia required for the field work must be prepared in advance, for speed and accuracy may determine the difference between success and failure. The following equipment is needed:

  1. Suitable tags showing pollination legends.
  2. A hunter's headlamp for each operator. These free both hands for work and give a narrow beam of intense light which is necessary for the rather delicate manipulations.
  3. A sharp razor for each operator.
  4. A wooden gauge as described above, marked in such a way as to enable one of the operators to cut the style quickly, squarely and at the desired point corresponding to the low place (X) between the two modes shown in figs. 2 and 3.
  5. A small portable stool or table for instruments and material.

On the plant which is to serve as maternal parent a nearly mature, unopened bud is selected, one whose pollen has not been shed, such as is shown at D. At the point k, the fluted inner surface of the corolla is found to be constricted as shown in figure 4 at E and F. This constriction furnishes a natural clamp for holding the straw splint in place, as shown in position at H. For this reason the point k is selected and cut by one operator. At the same time the other operator places a style on his gauge and makes a clean square cut at the predetermined desired distance from the stigma and then discards the lower part of the style together with its contained tubes. The stigmatic portion together with its slow-growing tubes is retained and its newly cut surface drawn up into the body of the splint as seen at I. The straw is then carefully lowered over the newly cut end of the lower style which remains on the plant, as shown at I, H and J. The instant that the two cut surfaces come in contact further lowering of the straw thrusts upward the part bearing the stigma. Thus one may know when contact is made and the process is complete. If this operation is carefully carried out on field plants and then left undisturbed for a period of 10-12 hours, a sufficient number of fertilizations may be accomplished to cause a fruit to set. It is usually a small one, especially if the fast-growing tubes have been discarded, but contains sufficient seed to give as large a culture as one usually wishes to plant.


For the greatest success, certain precautions should be borne in mind. (1) General growth conditions and fruit setting should be good. (2) A rather high degree of atmospheric humidity is also necessary because otherwise the cut surfaces dry quickly and stop the passage of tubes. For this reason, we worked at night after the air approached the dew point. In damp weather, the operations can be performed in the evening thus taking advantage of twilight. (3) As will be seen from the cross section of the style G, fig. 4, the strand of conducting tissue is not circular, but elongated in cross section. If the two abutting strands are properly orientated, a perfect union is made, otherwise many tubes will be stopped because they find no continuity of conducting tissue at this point. In practice, this structural difficulty can be only partly overcome. (4) The straw splints, C should be obtained from living plants, and carefully selected to fit the style very closely. Diagonally cut ends will facilitate their insertion.

This method should not be called grafting. The tissues of a style are mature and do not unite by meristematic growth. They are merely brought closely together and held in place. If the splicing has been successful we estimate that the pollen tubes usually cross the joint within a 5-9 hour period following the operation. The pollen tubes actually serve as a bond to hold the united parts together, but the upper part of the joined style frequently dries out long before the lower portion, showing that no actual organic union of the stylar tissues occurs.


Our first attempts in splicing the styles on garden plants were made with normal pollen. We were naturally interested in demonstrating the possibility of obtaining successful fertilization by this method regardless of pollen tube distributions. In some cases, we cut off only a portion of the longest pollen tubes or made our incision in front of all of them. The seed capsules from 15 successful trials contained from 5-175 seeds, an average of 72. In another set of trials, we obtained from our successful attempts from 50-275 seeds, an average of 120. In a set of experiments using the pollen of (2n+1) Globes, we obtained 7 capsules ranging from 35-175 seeds with an average of 100.

We have now made a total of 50 successful pollinations with spliced styles. Many of these are from crosses in which the desired phenotypes would not appear in ordinary pollinations. The latter have not been planted as this is written. From the last series, the ovaries which dropped without setting fruits were collected and dissected. These yielded some valuable data concerning the minimum number of fertilizations required in an ovary to ensure a reasonable probability of a set. We found that 36 had no ovules fertilized, 15 had between 1 and 10 fertilizations, 15 had 11-20 enlarged ovules and 2 others between 21 and 30. Among the successful sets there were no capsules with less than 10 seeds, 3 with 11-20, 2 with 21-30, 5 with 31-50 and 13 with from 51-140 seeds each. Thus it appears that a minimum of about 20-30 ovules should be fertilized in order to ensure the setting of seed capsules in field-grown plants at times when the conditions of fruit setting are favorable.

Some of the crosses Normal X (2n+1) Globe were planted in order to test the Globe-transmission through the pollen by this method. Though the highest had only 31 per cent Globes, which is fewer than we might expect to obtain, this is about a 15-fold increase over Globe-transmission in ordinary pollinations, and demonstrated that the 13-chromosome gametophyte of this mutant could be successfully transmitted by the method of splicing styles. Evidently we had placed our incision too far forward, thus including a majority of 12-chromosome pollen tubes. In our later experiments, in which the place of incision was determined with greater accuracy, so that more of the longer pollen tubes were discarded, the proportion of the progeny especially desired was much higher, in spite of the fact that the average number of seeds obtained was lower.



* Style was cut too far forward, including longest group of pollen tubes.

In order to test the possibilities of this technique with distinct bimodal distributions of pollen tubes, we made some double pollinations using pollen which was known to give only a single mode of distribution from each separate pollination. The two pollinations were separated by an interval of four hours thus giving two modes. By using the pollen of the dominant gene for purple flowers and seedlings in one pollination, and the pollen from white plants in the other, and splicing these onto the flowers of white plants, we could tell from the color of the seedlings whether the desired group of longer pollen tubes had been excised. Table 1 gives the results of these tests together with similarly pollinated control-flowers which were not spliced.

The first four pollinations given in table I were some of our controls, and were divided into pedigrees representing seeds from the upper and lower half of the seed capsules. It is obvious that the pollen tubes of the second pollination (which were largely excised in treated styles) were almost completely excluded from the upper half of the seed capsules, but fertilized ovules in the lower portion to a much greater extent. The average for the entire capsules in the controls was about 15 per cent.

In our splicing tests the proportions obtained from the second pollination ranged between 62.5 and 94 per cent, with an average of 83.2 per cent. This is nearly a six-fold increase over the controls. However, one case (Planting #31752 listed at the bottom of this table) gave practically the same results as the unspliced controls. In this case, there can be little doubt that the incision was made too far forward, thus including all or nearly all of the longest group of pollen tubes. This serves to demonstrate the importance of knowing accurately the point X (see figures 2, 3 and 4A) at which the incision is to be made in order to discard the longest group of pollen tubes. On the whole, these genetic tests indicate that the method of splicing styles is practical, and may be expected to give greatly increased yields of zygotes coming from slow-growing pollen tubes.