The Minnesota Horticulturist 23(5): 187-195 (June 1895)

*A paper read before the Mlnoesota Academy of Natural Sciences, Jan,. 1895.


The uncertainty of the regular annual fruiting of plains in the Northwest, where only the native Prunus Americana in its many varieties has been found sufficiently hardy to endure the climatic conditions, has long been a difficult problem in horticulture.

Writers on the subject of plum culture have attributed as the cause of the more or less non-productiveness "the influences of domestication and consequent high culture," "self-sterility," etc., etc. The beneficial effect of cross-fertilization has been hinted at and proposed as the remedy for all cases of infecundity. Mixed and close planting of the varieties to better ensure cross-fertilization has been suggested by nearly all of them. Reports of various horticultural societies are filled with instances of the beneficial effects of cross-fertilization, but, reading between the lines, as many or more instances of the failure of good results from cross-fertilization have been recorded. Cross-fertilization, therefore, unless it be effected in the direction of the natural affinities of the varieties, does not completely explain why certain varieties, even with the aid of cross-pollenation, may be prolific one season and the next produce no fruit at all; why one season the fruit will be large and fine, the next inferior in size and quality; why an unusually fine variety in the woods and thickets will be worthless when removed from its surroundings, even with subsequent best of care and culture.

About ten years ago I began making artificial crosses for the purpose of breeding improved varieties. My grounds contained at least two hundred trees, mostly selected from the woods and thickets along the Minnesota and Cottonwood river bottoms, together with a few horticultural varieties of P. Americana. I soon found that many of my desired crosses were difficult to obtain. I observed numerous adaptations to ensure cross-pollenation, together with differences in morphology of the stamen and pistil. Crosses between certain forms were fully fertile, while with others negative results were invariably obtained. Reciprocal crosses between varieties and between species were not equally fertile. I determined to go into the matter systematically, keeping a careful record of each cross made and noting the result, raising hundreds of seedlings and again experimenting with them.

Prunus Americana and other species of the prunus vary much in their wild state in flower, fruit, foliage, season of maturity and other botanical characteristics, so much so that an enterprising botanist might easily pick up in the thickets of almost any natural area where they abound a dozen or more varieties with characters so distinct from the type as to entitle them to the distinction of specific varieties.

The writer has no desire to inflict upon scientific botany any further division of the botanical characters of a species which is already sufficiently defined, but only offers his classification for the purposes of this paper in the interest of economic horticulture.

PLATE I. Flowers of Prunus Americana slightly enlarged.


Adopting the nomenclature used by Darwin in his "Different Forms of Flowers on Plants of the Same Species" and classifying as to morphology and function, we find the following fairly well defined forms in addition to the hermaphrodite form of botanists.


Proterogynous, (plate 1, fig. 1) on which the stigma is ready for fertilization and has passed the receptive stage before the pollen matures.

Proterandrous, (plate 1, fig 2) on which the pollen ripens and matures before the stigma is ready for fertilization.


Long-Styled, (plate 1, fig. 3) on which the pistil is nearly twice the length of the stamens.

Short-Styled, (plate 1, fig. 4) on which the stamens are nearly twice the length of the pistil.


Gynodioecious. (plate 1, fig. 5) on which the flowers in morphology consist of perfect flowers, but mostly females with anthers aborted and only a few grains of pollen, and these smaller and mostly aborted.

Andromonoecious, (plate 1, fig. 6) on which the flowers in morphology consist of perfect flowers, but mostly males with most of the pistils wanting or only present in a rudimentary form.

These divisions in morphology and function are generally well defined, but sometimes graduate into each other. The hermaphrodite form, which is the only one capable of self-fertilization, is now scarcely found in a wild state. Most of our cultivated varieties of P. Americana are transition forms of this and are somewhat difficult to classify. Some of the horticultural varieties certainly belong to the next, and a few are certainly heterostyled. Dichogamous varieties are more frequently met with in nature.

The proterogynous form is easily distinguished by most of the pistils projecting through and above the petals, which for a day or two remain incurved over the still immature stamens, thus mechanically preventing the ripening of the pollen, and fertilization, if accomplished at all, has generally been effected before the petals expand. I have also noticed in intense forms of this and the next a considerable difference in the time in which the stigma became receptive and the pollen mature after exposure to air and sunshine.

* "Different Forms of Flowers
on Plants of the Same Species

The proterandrous form is also easily distinguished from the fact that the pistil, before the petals expand, is found curved within the corolla, the stigma being inverted and partly within the calyx tube. About twenty-four to thirty-six hours after the petals expand and, usually, after the pollen on the anthers has all dehisced, the pistil gradually assumes an upright position and becomes receptive, and as the stigma does not become receptive until after an exposure to the influence of light and air for at least an hour or two, self-fertilization is prevented. The peculiarities of the two dichogamous forms not only in a measure mechanically prevent self-fertilization but certainly prevent the maturation of the pollen grains and stigmatic secretion at the same time, and this alone in the more intense forms is sufficient to prevent self-fertilization. With respect to heterostyled plants Darwin* says: "Unless it be proved that one form is fully fertile only when it is fertilized with pollen from another form, we have not complete evidence that the species is heterostyled. But when pistils and stamens differ in length in two or three set of individuals, and this is accompanied by a difference in the size of the pollen grains or in the state of the stigma, we may infer with much safety that the species is heterostyled." If the above test is correct, then a few individuals of our species are heterostyled, and many more are tending in that direction. Bisexuality is clearly defined in many individuals in nature, the male form being more numerous. I have frequently met forms entirely devoid of pistils. Now and then, I have found forms which do not appear to come under any division of the foregoing classification. For instance, I have a tree which for three years has produced flowers, each of which had two and, in a few instances, three apparently perfectly developed pistils. So far no fruit has set, although I made last spring a number of hand crosses to determine its affinity. These freak forms are the exception, and with them this paper has nothing to do.


The fruit buds of P. Americana are developed on the spurs and spur-like branches of the current season's growth. The following spring, on approach of steady warm weather in May, the buds swell and expose from one to five flowers in a simple umbel-like cluster. The period of bloom and the time when pollination may be effected generally extends over two or three days, and in cool and cloudy weather it may extend over a week. Pollination is effected by the aid of wind or insects. Within from two to twenty-four hours after the blossom has fully expanded, or in the dichogamous forms after the pistil and stigma have been exposed to light and warmth, the stigma becomes receptive, as may be plainly seen with a glass of moderate diameters by the glistening secretion on the stigma. Pollen ripens during clear, warm weather in about the same time, varying slightly in the different varieties. Within three or four days after fertilization has been effected, the petals drop off and the calyx tube is parted over the now slowly swelling ovary and drops off. When pollination has not been effected, the blossom continues fresh for several days, although the stigma may have become covered with duet and withered and become non-receptive, and it finally drops off, the peduncle remaining for a day or so longer. The peduncle lengthens to nearly its full length from the time the blossom bursts from the bud until fertilization is complete and, when legitimately fertilized, enlarges in diameter. When fertilization has been illegitimately effected the peduncle does not enlarge in diameter as much, and the slightly enlarged ovary usually falls, together with the peduncle, within from three to twenty days after fertilization. The season of full bloom ranges in different varieties over a period of about ten days. The past season, 1894, my earliest blooming varieties were in full bloom May 2d and the latest May 10th.

The actual time in the life of a blossom during which fertilization may be effected scarcely exceeds two hours and is not, as many suppose, during the whole life of the expanded flower.


From the many artificial crosses that I have made and recorded, I long ago became convinced that fertilization might be effected in different degrees and that many plants had the power of throwing off such ovaries as were fertilized by pollen lacking in sexual affinity and that this was especially true in P. Americana. It should be borne in mind that the production of seed is the chief end of the act of fertilization and the vivification of the ovule is the primary object of pollination. By systematic crossing and hybridizing, I determined that the union of the reproductive elements of two trees possessing the proper selective affinity for each other readily produced a stronger developement of the ovary; a union of this kind I shall call legitimate.

It is well known that by crossing distinct species fertilization is effected with more or less difficulty; that reciprocal crosses of the same two species vary in the intensity of fertilization. As to the union of the reproductive elements of varieties lacking in sexual affinity for each other or in which the reproductive elements have become too greatly differentiated and the development of the ovary either fails entirely or is below the normal, I shall use the term illegitimate, and in the same sense as used by Darwin in his "Different Forms of Flowers on Plants of the Same Species." The simplest test to determine the sexual affinity of any variety, and one which I have never known to fail when done under proper conditions, is to take several sets of flower clusters and pollinate each individual stigma with pollen of a different form. The union of such crosses as possess the proper degree of affinity will prove fertile, while the union of those lacking in affinity will prove sterile. No matter how many of the flowers of each cluster are pollinated legitimately or illegitimately, the result will be as above. If all of the flowers of a cluster are pollinated legitimately, they will all set fruit, barring accident, of course. This experiment may be modified by many different combinations.


In the diagramatic chart following (plate 2), I have attempted to give in the simplest and most condensed manner the results of several thousand artificial hand pollinations made by me during a period of five successive years, together with some earlier data and observations, all made in a very careful manner and carefully recorded, the details of which would be too long and burdensome for this paper. Most of my data for pollinations made with pollen of the andromoncecious form on the long-styled heterostyled form were accidentally lost, and at the time of compilation consisted of but three records. I am unwilling at this time to state positively that pollination in the above direction would produce legitimate fertilization; I have, however, on the basis of incomplete data and from theoretical conclusions, indicated the probable rule by a dotted line. The central column represents the form of pollen used; the lines between the different forms indicate the directions in which fertilization is positive or negative, or, in probably more expressive terms, legitimate or illegitimate. Of the forty-nine possible combinations, or directions. of pollinations but one form, the hermaphrodite, is fully fertile with its own pollen. Including the hermaphrodite form, cross-fertilization is legitimate in only thirteen directions. Thus it will be seen that among the seven forms of P. Americana pollination is possible in forty-nine directions, thirty-six of them giving negative or illegitimate results, and that there are only thirteen directions in which cross-fertilization is possible.


I know of no group of plants more favorable than the genus Prunus for the study of the order of evolution from the hermaphrodite stage to the higher stage of bisexuality. Their organs of reproduction, as I have shown, present a number of peculiarities of morphology and function, unusually interesting and significant and at the same time unusually intelligible, nor are these peculiarities exhibited to the same degree by any other group of plants. A study of these numerous adaptations to insure cross-fertilization must necessarily end in the conclusion that our species is gradually approaching a state of dioeciousness, and, fortunately for our discussion, there appears to be no missing link in the chain.

With these numerous adaptations and structures to prevent self-fertilization and to insure cross-fertilization in view, we are prepared to understand why, in the several cases, self and cross-fertilization is possible, and why impossible; why cross-fertilization is possible in a certain direction, while the reciprocal cross may be sterile, and, finally, by what means our species is gradually becoming dioecious. The wedge of variation, having gained a hold of our hermaphrodite form, still in existence and capable of self-fertilization and forced by long continued self and occasional cross-fertilization, produced the earliest types of our dichogamous group which the better insured cross fertilization. In the proterogynous form, the pistil, protruding through the still undeveloped petals and stamens and receiving the advantage of sunlight, air and warmth, was encouraged to greater development, the stamens being correspondingly retarded. In the proterandrous form, the stamens received the benefit and the pistil was retarded. Through successive generations, the influence of the law of balancement has been at work, the evolution towards a separation of the sexual organs has fairly started, and we have the foundation in the proterandrous and proterogynous forms for the pistillate and staminate forms of a future dioecious species. Simultaneously with the development of the pistil and retardation of the stamens and vice versa, came the further adaptation of difference in time of maturity of the reproductive elements, with an additional protection against self-fertilization.

PLATE 2. The direction of legitimate and illegitimate fertilization.

The development of the reproductive organs, aided by the law of balancement, continued, and we have developed the heterostyled group. What was before accomplished to quite an extent by purely mechanical adaptations, is now accomplished by a differentiation in the reproductive elements. So great a differentiation in the reproductive organs surely caused a differentiation in the sexual elements.

Our species has now become divided against itself. The differentiation of the reproductive elements was followed by still further development and retardation of the reproductive organs—and we have nearly reached the bisexual stage, not only in morphology but in function. We now have only a step further to the complete separation of the sexual organs.

In the earlier stages of the dichogamous group, self-fertilization was possible but mainly prevented by mechanical adaptations to insure cross-fertilization. In the heterostyled group self-fertilization is prevented by the great differentiation in the reproductive elements, and the sexual affinity destroyed. There can be no doubt that the differentiation into species was accomplished by variation of form, foliage, cell structure, etc., without a corresponding differentiation of the sexual elements. We know this from the fact that distinct species will sometimes cross, but not freely, in a reciprocal direction. I have myself, within the past ten years, produced hybrids between P. angustifolia and P. Americana; between P. domestica and P. Americana; between P. Besseyi, Bailey (P. pumila Lin.) and P. hortulana, B.; between Cerasus avium var. and P. Besseyi, B. I made several hundred crosses to produce hybrids between our sand cherry (P. Besseyi) and horticultural varieties of Cerasus avium. Pollen of C. avium var. on P. Besseyi invariably proved sterile; reciprocal crosses set fruit, but they failed to germinate, the seed containing only a trace of the aborted ovule. When I finally used the pollen of a proterandrous form of P. Besseyi on a short-styled form of C. avium fertilization was effected and developed a normal fruit, the seed of which germinated and produced an undoubted hybrid. The reciprocal crosses of the same varieties failed to fertilize a single ovule out of over fifty crosses made. I had applied the same principle in the production of hybrids between P. hortulana and P. Besseyi with fair success. The successful crosses just mentioned were made with pollen which had not been too greatly differentiated, on a pistil which, in accordance with the theory advanced for the evolution of the different forms, had been retarded. The unsuccessful crosses were made with differentiated pollen on a pistil not sufficiently differentiated. Finally, we must conclude that the means by which the bisexual forms have been produced, though gradually and necessarily very slow, are identical with the forces that produced the different species. ln the crossing of the different species, we find that by applying the same rule for cross-fertilization we can trace the genealogy back to the forms wherein the differentiation of the sexual elements had not destroyed their affinity. From these experiments we deduce the following:


Self-sterility of Prunus Americana in the heterostyled and bisexual forms is caused by the great differentiations of the sexual elements. Pollination by wind and insects cannot be controlled to any extent.

Mixed planting, therefore, unless it be done with respect to the natural affinities of the varieties, may produce the most disastrous result for the horticulturist.

Other species of Prunus observed show these characteristics of P. Americana, and it may be possible to bring them together and under a similar classification.

Finally we have gained some knowledge in the summary production of hybrids.

This study is necessarily preliminary to further investigation touching horticultural varieties of P. Americana, etc.

Sexual Systems in Prunus