Communications in Soil Science and Plant Analysis (1979)
Effect of rootstocks on greenhouse rose flower yield and leaf nutrient levels

Nathan Gammon Jr.a & S. E. McFadden Jr.b
a Professor Soil Chemistry, Soil Science Department, and former Assistant
Professor Ornamental Horticulture, University of Florida Gainesville, FL, 32611
b Ornamental Horticulture Department University of Florida , Gainesville, FL, 32611.


'Carina' rose was grown on seven different rootstock clones. These combinations were tested for effect on flower production and mineral content of the leaves when grown in fumigated ground beds in an evaporatively cooled plastic greenhouse. Plants growing on R. odorata ISU 5710-2 produced the highest flower yields followed by R. fortuniana and R. manetti while R. multiflora ISDU 62-5 produced the lowest number of blooms. Rootstocks caused large differences in the levels of some minerals in the leaves. R. fortuniana caused a Mn accumulation that was five times the quantity developed by R. odorata but this did not seem to be related to flower production. However, R. odorata was a superior accumulator of K and under conditions of relatively low supply. R. fortuniana and R. odorata were good accumulators of N and K. These factors were related to flower yield.


Earlier experiments6 have established that Rosa fortuniana is a superior rootstock for roses grown in Florida. This improved performance was attributed to disease and nematode resistance and better adaptation to warm climate and sandy soils. The cold hardy R. multiflora is generally used for outdoor plantings in cooler climates and R. manetti seems better adapted for greenhouse use3.

A routine check in a fertilizer experiment for the mineral status of rose leaves from the same cultivar grafted to R. fortuniana and R. manetti revealed a higher manganese level in leaves from plants growing on the R. fortuniana rootstock. Since the rootstocks of other plant species have been shown to influence the mineral composition of the scion4,8, this investigation was designed to evaluate several rootstocks on flower production and foliar mineral composition of a commercial rose cultivar.


Rose cultivar Carina' was grafted to seven rootstock clones, Table 1. The experiment consisted of randomized blocks of 'Caria' on each of the seven rootstocks, each of the randomized blocks was treated with one of four fertilizer variables and there were four replications, hence 16 plants of each scion-rootstock combination were employed. Fertilizer treatments per plant consisted of: 1) Rootcontact Packet 2, prewet, containing 28 g 18-6-12, 2) Rootcontact Packet, dry, containing 28 g 18-6-12, 3) Osmocote2 18-6-12 at 363 g per 9-month interval, and 4) Liquid fertilizer 18-6-12 applied at 28 g per month. Planting in fumigated ground beds in an evaporatively cooled plastic greenhouse, located near Gainesville, Fla., was completed in January. Blocks were separated by vertical plastic barriers to prevent lateral fertilizer movement. Because of the expected delay in release of nutrients from treatments l-3, all plants received an initial liquid fertilizer application of 28 g of 18-6-12 at planting.

Samples were collected over six-week periods (a period deemed long enough to cover a flush of bloom on each plant) in July-August and October-November of the first year and in March-April, July-August, and October-November of the second year. Each sample consisted of a record of flowers produced for each single plant during the period. Total commercial flowers produced per sampling period as well as an approximate flower grade weighted as follows was used to evaluate the production of each rootstock:

Stem Weight
Less than 15 inches 1.0
15 to 21 inches 1.5
21 to 27 inches 2.0
Over 27 inches 2.5

Foliar analysis samples consisting of the first and second five-leaflet leaves immediately below each bloom were collected on a single plant basis during each sampling period. Composite samples from each plant were analyzed for N by the semi-micro Kjeldahl method, and for P, K, Ca, Mg, Mn, Fe, Cu, and Zn by atomic absorption or emission spectrophotometry of a 0.1N HCl solution prepared after dry ashing the plant tissue at 450°C. Flower data was summarized for only the last four harvest periods as many of the leaf samples taken during the first harvest period were lost in a laboratory accident.


The average number of flowers produced in a six-week period and the mean flower value index is shown in Table 1. The significant effects of rootstocks on yield during the successive sampling periods are given in Table 2. The production on R. fortuniana stock improves steadily with later sampling dates but at the last sampling R. odorata ISU 5710-2 was definitely superior in flower production. In an earlier field experiment 6 a strain of R. odorata ranked with R. fortuniana and var. 'Dr. Huey' rootstocks and much below R. fortuniana in total production and quality of bloom. The improved results with R. odorata in this experiment may be partially attributed to the ISU 5710-2 strain but it is more likely the result of the elimination of a nematode problem in this experiment as the soil was thoroughly fumigated prior to planting.

Mean number of commercial 'Carina' flowers per plant produced in a six-week
sampling period and the corresponding flower value index.

No. Rootstock Identification Flowers†
per plant
value index
1. R. fortuniana 18.04 27.63
2. R. multiflora ISU 62-5 (thornless) 10.14 14.23
3. R. multiflora (Cincinnati grower source) 17.02 23.80
4. R. multiflora Brooks 56-6 (thornless 13.85 20.35
5. R. odorata ISU 5710-2 20.56 31.74
6. R. manetti (locally grafted) 17.59 26.08
7. R. manetti (Carlton Nursery plants) 16.82 24.78
†Average of 5 selected 6-week sampling periods.
‡Sum of flowers in each grade x weighted number assigned to that grade (stem length).

Relative rootstock yield responses as measured by Duncan's Multiple Range Test.†

†Any rootstock numbers underlined by same line are not significantly different at the .05 level.
‡FVI = flower value index; FLWS = number of blooms.
§See Table 1 for rootstock identification.

The mean foliar nutrient contents of 'Caria' cultivar for the March-April sampling period of the second year are given in Table 3. This period was selected as it represented a very vigorous spring growth and because in the two later samplings it became obvious that the two Rootcontact Packet treatments, which were not refertilized, were no longer supplying adequate levels of nutrients. The N and K levels in the Rootcontact Packet treatments were already depressed in this sample. When the average values for all rootstocks on these two fertilizer treatments are considered, only K and Mg levels are below the minimum normal range for greenhouse roses as proposed by Boodley and White 3. For the last two harvest periods, average N and K levels were below the low normal ranges of 3 and 1.8%, respectively.

Mineral content of 'Carina' cultivar leaves sampled during the March-April
period of the second year, dry weight basis.

Rootstock N K Ca Mg P Mn Fe Cu Zn
  - - - - - - - - - - - - - - - - - % - - - - - - - - - - - - - -- - - - - - - - - - - - - ppm - - - - - - - - - - - -
1. 3.3 1.4 1.4 0.22 0.27 664 67 4.8 45
2. 3.3 1.5 1.3 0.24 0.27 131 61 6.1 40
3. 3.7 1.5 1.5 0.24 0.32 251 65 6.4 45
4. 3.6 1.5 1.4 0.23 0.31 198 66 4.7 44
5. 3.6 1.7 1.5 0.26 0.29 130 63 6.4 38
6. 3.6 1.4 1.3 0.20 0.27 276 60 5.8 38
7. 3.3 1.3 1.2 0.20 0.26 251 69 5.5 38
1. 3.0b ¶ 1.2 1.4 0.24a 0.26 180 61b 6.4 36
2. 3.1b 1.0 1.4 0.25a 0.26 242 59b 5.6 39
3. 3.8a 1.7 1.4 0.21b 0.28 318 67ab 4.7 41
4. 3.8a 2.0 1.3 0.20b 0.33 345 72a 5.9 50
3.0 1.8 1.0 0.25 0.2 30 50 5 15
5.0 3.0 1.5 0.35 0.3 250 150 15 50
† See Table 1 for rootstock identification.
‡ 1. = Rootcontact Packet, prewet; 2. = Rootcontact Packet, dry; 3. = Osmocoat; 4. = Liquid Fertilizer, 18-6-12.
§ L = Low and H = High normal range, Boodley and White3.
¶ Numbers in same column followed by same letter are not significantly different at the .05 level by DNMRT.

While the effect of individual rootstocks on the mineral content of 'Carina' leaves is shown in Table 3, a statistical evaluation of those differences for each of four sampling periods is given in Table 4. The earlier observations of the increased Mn content of roses grown on R. fortuniana rootstock is confirmed. The significant fertilizer interaction during two sampling periods did not alter the significantly high Mn contents on R. fortuniana or the consistently low values for R. fortuniana ISU 62-5 and R. odorata ISU 5710-2, Likewise, Zn uptake by R. fortuniana was high and R. odorata ISU 5710-2 low. The reverse trend was true for Cu and may explain the positive responses to Cu sprays noted when roses are grown on R. fortuniana stocks.

There was a very strong interaction between K fertilizer supply and the rootstocks (data not shown). In general, R. odorata and R. fortuniana supplied relatively more and R. manetti the least K when little K was supplied from the fertilizer. However, R. manetti was a strong K accumulator when the amount of K supplied in the fertilizer was high.

Although there were no significant differences, R. fortuniana and R. odorata appear consistently near the top as suppliers of Fe. The Mn levels produced by R. fortuniana were often five times the levels produced by R. odorata but this difference was evidently not enough to cause a reduction in Fe uptake sometimes observed for high levels of Mn5,7,9.

Relative rootstock effects on mineral contents of 'Carina' cultivar leaves as measured by
Duncan's Multiple Range Test† for four harvest periods.

†Any rootstock numbers underlined by the same line are not significantly different at the .05 level.
‡See Table 1 for rootstock identification.
§Absence of underline indicates significant interaction with fertilizer treatments

Total flowers and flower value index were highly correlated, hence the correlation coefficients between flower value index and mineral contents of the leaf tissue for four sampling periods as presented in Table 5 could represent either. Although these r values account for 22% or less of the variation, some of the significant effects are worthy of note. The increasing significance of the N and K correlations is attributed to the fact that the leaf contents of many plants were below the 3 and 1.8% levels, respectively, during the later sampling periods. The increasing negative P correlation is accounted for by the high flower yields and relatively low P concentration in the leaves of R. fortuniana and R. odorata which were .30 and .29%, respectively, during the last October-November harvest while the other rootstocks ranged from .29 to .37% P. The high Mn and Fe correlations probably are the result of secondary responses to the higher total salts resulting from the Osmocote and liquid fertilizer treatments during the second year. Increased salt concentrations would result in higher localized soil acidity and hence increased Mn and Fe solubility. In addition to favoring increased Mn and Fe availability, these treatments supplied more N and K which favored a higher flower production.

Correlation coefficients† for Flower Value Index and mineral levels in the leaves of 'Carina' cultivar.

  Sample period
  1st year 2nd year
Element Oct.-Nov. March-April July-Aug. Oct.-Nov.
  - - - - - - - - - - - - - - - r - - - - - - - - - - - - - - -
N .001 .240 .246 .474
K .192 .095 .155 .320
Ca .381 .232 -.084 -.026
Mg .195 .083 -.024 -.125
P .204 -.101 -.192 -.364
Mn -.103 .135 .129 .335
Fe -.045 .250 .108 .313
Cu -.180 -.025 .118 .079
Zn -.359 -.095 .087 .141
†Coefficient must exceed in absolute value 0.187 to
be significant at the .05 level and 0.244 at the .01 level.

These data show that rootstocks may strongly influence not only the mineral levels in the leaves of cultivars grown on them but also the flower yield of the cultivar. The importance of mineral accumulation induced by the different rootstocks cannot be clearly evaluated. Certainly rootstocks that tend to increase K and N levels, especially when the soil reserves are relatively low, would be expected to stimulate flower yields. However, the variations produced in some elements, such as Mn, may be large with no apparent effect on flower production. Although the effects of rootstocks on mineral accumulation may be of considerable significance in flower production, it is likely that biochemical and other biological factors (including possible soil microorganisms) are important in determining the flower production stimulated by a given rootstock.

  1. Florida Agricultural Experiment Station Journal Series Paper No. 1386.
  2. Use of trade names in this manuscript does not constitute endorsement of the product.
  3. Boodley, J. W., and J. W. White. 1969. Fertilization. p. 78-92. In J. W. Mastaierz and R. W. Langhans (eds.) Roses, A manual on the culture, management, diseases, insects, economics, and breeding of greenhouse roses. Pennsylvania Flower Growers, Assoc., Inc., and Roses Incorporated Pub.
  4. Kleese, R. A. 1967. Relative importance of stem and root in determining genotypic differences in Sr-89 and Ca-45 accumulation in soybeans (Glycine max L.). Crop Sci. 7:53-55.
  5. Lingle, J. C., L. O. Tiffin, and J. C. Brown. 1963. Iron uptake-transport of soybeans as influenced by other cations. Plant Physiol. 38:71-76.
  6. McFadden, S. E. 1956. Comparison of 'Happiness' rose production on four rootstocks. Proc. Fla. State Hort. Soc. 69:368-370.
  7. Moraghan, J. T., and T. J. Freeman. 1978. Influence of FeEDDHA on growth and manganese accumulation in flax. Soil Sci. Soc. Am. J. 42:455-460.
  8. Polson, D. E., and L. J. Smith. 1972. Nature of scion control of mineral accumulation in soybeans. Agron. J. 64:381-384.
  9. Tiffin, L. O. 1967. Translocation of manganese, iron, cobalt, and zinc in tomato. Plant Physiol. 42:1427-1432.