Mem. Hort. Soc. New York 1: 93-101.
International Conference on Plant Hardiness and Acclimatization - 1902

Improvement of Roses by Bud Selection, or
Blind vs. Flowering Wood for Rose Cuttings

L. C. Corbett
, Department of Agriculture, Washington, D. C.

The results which are recorded in the article prepared under the above title are based on a series of tests with rose cuttings made from "blind" and "flowering" wood, and cover a period of five years.

The work was undertaken to settle a point in dispute among commercial growers of roses as to the relative value of plants grown from blind and flowering wood for flower productions.


1) Do cuttings tend to perpetuate the individual peculiarities of the parent branch from which they are taken?

2) Can accumulative results be obtained from a continuous use of cuttings from wood with like habits? i.e., can the flowering habit of plants be increased by the continuous use of flowering wood selected through successive years from plants which have themselves been produced from flowering wood?

B—Flowering Wood A—Blind Wood

For many years discussion has been rife among commercial growers of roses is regard to the flower or bloom-producing power of plants grown from what is known as "blind wood," and those grown from "flowering wood." These terms are familiar to all accustomed to the propagation and cultivation of the rose under glass; but in order that none may hold a misconception of the point in question, I present a figure representing the two types of wood spoken of. The branch at (a) in the engraving represents a characteristic shoot of the so-called "blind wood." This shoot, it will be noticed, is of slender growth, somewhat willowy in character, and is terminated by a leaf instead of a flower bud. As a rule, "flowering shoots" are larger and more vigorous, but are also softer and contain a larger percentage of pith in proportion to the woody tissue than "blind wood" shoots. In general, however, the "flowering wood" shoots are longer than the "blind wood" branches, and if the flowers are cut with short stems, then there is a considerable length of wood suitable for purposes of propagation at the base of each flowering shoot. This wood is harder and more mature than the wood close to the base of the bud, and for that reason is better suited for making cuttings. With plants which are blooming profusely it not uncommonly happens that the whole length of the flowering shoot is needed to satisfy the market demand for long-stemmed roses, and the supply of suitable wood from which to propagate the next season's stock of plants is greatly lessened, or it may be entirely cut off. But there is always a greater or less supply of "blind wood." Consequently, why not use it for purposes of propagation? If the plants grown from blind wood do not perpetuate the tendency of the parent shoots (which are non-flowering), then there can be no objection to its use for the purpose of propagating the next season's stock. But if it does perpetuate the tendency of the parent stem, then there is danger.

In order that more definite statements may be made on this point, the writer has for five years carried on a test with rose plants from the two types of wood above mentioned. Before stating the results of this experiment, however, I wish to call to mind a few of the experiments which have been made which throw light upon the point in question.

Do cuttings tend to perpetuate the peculiarities of the parent branch from which they are taken? To answer this, I need hardly do more than call attention to the fact that propagation by cuttings is employed almost exclusively for the perpetuation of cultural varieties of all fruit and ornamental plants which are capable of being grown from cuttings. Many annual plants, however, come true from seed and varietal differences, sufficiently close for all commercial purposes. If we were to go a step farther, we might be justified in considering the various processes of budding and grafting as identical in their results with that of propagation by cuttings.

Budding and grafting are in reality processes of division, the same as is the growing of plants from cuttings. In all three of these modes of reproduction the results are so constant that we never stop to question the fact; yet we constantly commit the blunder of ignoring qualities quite as important as the varietal peculiarity itself. In fruit growing, nurserymen propagate from a Baldwin tree, whether it has ever borne fruit or not, simply because they know it to be a Baldwin. Yet in the face of this we are being taught by our advance agents that each tree has an individuality, and, in fact, that each branch and bud is in its peculiar way different from every other branch or bud, even upon the same tree. If we accept these statements as true, and we have no good reason to doubt them, then the peculiar tendencies of the plant, or of a branch of a plant, may be expected to play a more or less important part in determining the behavior of the plant or plants propagated from it. Orchardists have observed these differences, and are slowly coming to exercise greater care in the selection of cions. This precaution not only influences fruit production, but it has been clearly pointed out by Smith, Fairchild and others the health of plants from which buds and cions are taken measures to a very marked degree the health and longevity of the resulting tree.

If we find these differences among plants grown from buds and cions, quite as marked peculiarities may be anticipated in plants grown from cuttings. Upon this point recorded observations are exceedingly meagre, but some light can be gathered from the work published by myself in the Ninth Annual Report of the West Virginia Agricultural Experiment Station. From these studies it is evident that varieties are perpetuated true to type by cuttings through many generations. A single exception in the case of the tomato entering to break the constancy of the results. Not only are general varietal differences retained, but acquired characters also, temporarily, at least, as is shown in the cuttings of grape, poplars, currants, etc., grown in Northern and Southern latitudes. For a specific example of this, nothing could be more conclusive than the results shown by New York grown potatoes, given one season's outing in Maine, for, when brought back to New York the next year, they retained their Maine tendency towards increased vigor and yield.

Trees propagated from fruitful trees are themselves more fruitful. Potatoes grown a year in the North become more prolific. Plants grown from cuttings, taken from Northern and Southern grown parents of the same species, retain the characteristics of their parents. What, then, should be expected from plants grown from blind and flowering wood? To anticipate the results of the test I will state that they accord with and justify the logical conclusion which would follow from the particular instances above set forth. In this test rose plants propagated from "flowering wood" gave on the average 29 4/9 blooms per plant for the season, while the "blind wood" plants produced 11 1/2 flowers per plant.


In the Spring of 1897, when the time for making rose cuttings had arrived, cuttings were made from both flowering and blind wood of each of the varieties of rose then in the house. The cuttings were all made on the same day, placed in the same cutting frame in contiguous rows. In all respects the conditions for the several cuttings were as nearly the same as it is possible to obtain in a greenhouse.

On February 16 the cutting plants were examined and potted, with the following results:

Bride, flower15 9 6 .
Bride, blind20 9 9 2
Bridesmaid, flower9 5 4 .
Bridesmaid, blind23 15 8 .
Perle des Jardins, flower15 2 11 2
Perle des Jardins, blind17 7 7 3
Mme. Hoste, flower13 5 8 .
Mme. Hoste, blind9 1 8 .
Meteor, flower18 11 3 4
Meteor, blind5 4 1 .

From this it would appear that there is little difference in the tendency to form roots between the cuttings made from flowering and from blind wood.

From this time until the plants were set in their permanent places upon the greenhouse benches (August 19, 1897), they were given like treatment. They were grown in pots in a sunny greenhouse, and all received the same number of shifts, and like attention in regard to soil, water and food supply. As the plants were planted upon the benches August 19, the following note was made: "At this time there is little difference in the size and general vigor of the plants from the blind and flowering wood cuttings." As soon as the plants became established in the soil on the bench growth on all was vigorous, and while there was a marked difference in the growth and general behavior of different varieties, there was no notable difference in vigor between the blind wood and flowering wood of the same variety.

The bloom record of the several lots is quite different, however, as is shown in the accompanying table, which gives the total number of flowers produced between December 1, 1897 and July 1, 1898.

Bride, flower13 242 18.62
Bride, blind18 32 1.78
Bridesmaid, flower8 192 24.00
Bridesmaid, blind3 56 18.67
Perle des Jardins, flower8 141 17.63
Perle des Jardins, blind14 224 16.00
Mme. Hoste, flower7 204 29.14
Mme. Hoste, blind9 176 19.56
Total No. blind 44 488 11.09
Total No. flower35 779 22.26

As is seen from the table, the product of the flowering wood plants in this one year's test was more than double that of the blind wood plants, amounting in this particular instance to 156 per cent. gain in favor of the flowering wood plants.

During the forcing season of 1898 and 1899, extending over the same period as in 1897 and 1898, e.g., from December 1 to May 31, inclusive, the two varieties retained in the experiment produced bloom as follows:

Bride, blind13 113 8.69
Bride, flower15 253 16.87
Bridesmaid, blind8 119 14.88
Bridesmaid, flower12 210 17.50

The results here recorded show the superiority of the flowering wood over the blind wood plants. But strange as it may appear, the difference between the flower-producing power of the plants grown from flowering wood and from blind wood is less than in the first generation, and this, despite the fact that the plants used in this test were grown from cuttings selected from plants used in the previous years' tests.


The flowering wood plants used in the tests of 1898 and 1899 were grown from cuttings selected from the base of flowering shoots produced by the plants previously grown from flowering wood. This course was followed in order to test the effect of constant selection in one direction, the idea being to secure plants with the maximum blooming capacity. The experiment was to test a theory which may be stated as follows: If flowering wood from commercial plants is capable of producing plants able to throw more than double the number of blooms produced from similar plants grown from blind wood, is it not possible by selecting flowering wood from plants grown from flowering wood to increase the flower producing tendency in a distinct strain of forcing plants? Beginning with blind and flowering wood from good commercial plants of Bride and Bridesmaid in the Spring of 1897, plants were grown and flowered in the Winter of 1897 and 1898; from these, as above stated, cuttings were taken and flowered during the Winter of 1898-1899; and from these, in turn, came the flowering plants for 1899 and 1900, and so on for five years; the flowering plants being from flowering plants of the previous season in each and every case. The blind wood plants were treated in exactly the same way, blind wood cuttings from blind wood plants produced the blind wood plants for the succeeding season, and so on for the five years over which the experiment has now extended. The following records for each of the years 1899-1900, 1900-1901, and 1901-1902 serve to show how constantly the flowering wood plants retain their supremacy over those grown from blind wood:

Record from December 1 to May 31, Inclusive
1899-1900 No.
Bride, blind15 190 12.67
Bride, flower18 192 10.67
Bridesmaid, blind18 137 7.61
Bridesmaid, flower18 220 12.22
1900-1901 No.
Bride, blind17593.47
Bride, flower411027.50
Bridesmaid, blind16674.19
Bridesmaid, flower14805.71
1901-1902 No.
Bride, blind14805.71
Bride, flower161247.75
Bridesmaid, blind16986.13
Bridesmaid, flower1213611.33
* * *
Record of Plants Second Year on Bench
1900-1901 No.
Bride, blind1419313.79
Bride, flower822327.88
Bridesmaid, blind1013813.80
Bridesmaid, flower1218315.25

But, strange as it may seem, there is no apparent cumulative effect from the selection of cuttings from flowering wood plants. Neither is there any marked degradation from the continuous use of blind wood. True, there is a constant falling off in the average number of flowers produced per plant from the first season to the close of the experiment, but the result is not more marked in the case of the blind wood plants than with the flowering wood plants.

The following table, which summarizes the results for the five years, is of interest in that it shows but a single departure from the rule that flowering wood produces plants which are more floriferous than those grown from blind wood:

Average No, Flowers per Plant for Season—Dec. 1 to May 31, Incl.
1897-981.88 19.84 19.00 28.62
1898-998.66 16.86 14.87 17.50
1899-0012.66 10.66 7.62 12.22
1900-0112.21 27.87 13.80 15.25
1901-025.71 7.75 6.12 11.33
Average for 5 years8.26 16.59 12.29 16.98

This is seen in the case of the blind wood plants of Bride grown during the forcing season of 1899 and 1900, in which year this particular lot of plants produced an average of two blooms per plant more than did the flowering plants of the same variety. The table is of interest also in showing the ratio of the flowering wood plants to the blind wood plants based on the average number of flowers produced per plant during the season. In the case of Bride the blind plants averaged 8.26 blooms per plant, while the flowering plants produced 16.59 blooms per plant, or a little more than twice as many. In the case of Bridesmaid, the difference is very decided, but not so great as with Bride. Bridesmaid blind produces an average of 12.29 blooms per plant, while the "flowering wood" plants of the same variety produced 16.98, or one and one-third times as many as the blind wood plants.


In the course of these observations a cultural problem of some moment presented itself, and as it could be brought under observation without deranging the observations on the production of bloom from blind and flowering wood plants, the experiments were planned to admit of retaining a number of plants upon the benches a second season in order to compare their flower-producing power with plants grown from cuttings and placed on the benches in July for the succeeding Winter's flower supply. Accordingly 14 blind and 8 flowering wood Bride, with 10 blind and 12 flowering wood Bridesmaid plants were retained on the middle bench of the house used continuously in this test. These plants were severely pruned in August, after having been kept quite dry and inactive during July. After pruning at least one-half of the soil of the bench was replaced by fresh compost. The earth was removed from on top the roots and between the plants and replaced with fresh earth. After this treatment they were slowly started into growth and the record of flower production began in October. The following table will serve to show the bloom record of these plants:

MonthBlind Flower Blind Flower
October57 49 51 49
November25 29 15 20
December39 36 33 33
January14 18 9 19
February28 26 18 29
March46 47 26 41
April39 47 34 33
May27 49 48 28
Totals275 301 204 252
Average per plant19.6 37.6 20.4 21.0

In the case of plants grown from cuttings struck March 11, and planted on the benches in July, the record was as follows:

October4 10 7 5
November6 5 1 10
December. 2 5 2
January4 8 4 7
February7 12 7 7
March8 11 12 16
April19 40 22 29
May21 37 18 19
Totals69 125 75 95
Average No. per plant4.05 31.25 4.68 6.70

From the comparison of the average number of flowers produced per plant from October 1 to May 31, inclusive, in each of the two sets, it is evident that there is little difference between the two, but upon comparing the monthly flower product of the two it at once becomes apparent that the two-year-old plants produced their crop in the Fall and early Winter, while the cutting plants produced the heaviest bloom later in the season. One must, therefore, be guided by the demands of the market. If a heavy crop before the holidays is the end to be achieved, then year-old plants are desirable, but if the rose market is more profitable from the first of February to the first of June, the cutting plants will give best return. While these results are interesting in showing the value of strong year-old plants, we do not consider that the record of the one season, during which this comparison was made, sufficient to be taken as a basis for extensive commercial undertakings. With the tests of blind and flowering wood plants, however, the case is different, and the results may be considered conclusive.


It is clear from the results of these tests that the tendencies manifested in a branch are perpetuated from generation to generation in plants propagated by asexual processes.

It is equally demonstrated that cumulative results are not to be expected by selecting parts showing like tendencies through successive generations. The flowering habit of plants which themselves had been produced from flowering wood was not increased, even in the fifth generation, over what it was in the first. On the other hand, plants repeatedly propagated from blind wood through five successive generations were not markedly less floriferous in the fifth than in the first generation.

In both plants propagated from blind and from flowering wood, there was a slight tendency towards lessened flower production. This may be accounted for in that the stock from which the plants were propagated each season had been grown and forced under artificial conditions, and no attention was given to selecting cuttings from the strongest plants. The commercial side of this experiment is, of course, the most important one from the standpoint of the practical grower. It is clearly more economical for the florist to produce his plants each season from blind wood, and since there is no cumulative effect from such a procedure, the plants so produced are not necessarily less floriferous than the parent stock. But where bloom rather than stock plants is the end sought the tests above recorded are emphatic in declaring the superiority in this respect of plants grown from flowering wood. A rose grower can well afford to send short-stemmed roses to market during the months of January and February, if by so doing he can secure sufficient flowering wood for propagating purposes to insure a stand of flowering wood plants for the production of the succeeding crop.

Hubbell: Causes of blind wood in roses (1934)
Cooper: Causes of blind wood (1970)