Rose response to nitrogen, phosphorus, and potassium fertilization rates in plastic mulch covered beds of fine sand

Material Information

Rose response to nitrogen, phosphorus, and potassium fertilization rates in plastic mulch covered beds of fine sand
Series Title:
Bulletin University of Florida. Agricultural Experiment Station
Young, T. W ( Thomas Wilbur ), 1905-
Place of Publication:
Agricultural Experiment Stations, Institute of Food and Agricultural Sciences
Publication Date:
Physical Description:
41 p. : charts ; 23 cm.


Subjects / Keywords:
Roses -- Fertilizers ( lcsh )
Nitrogen fertilizers -- Florida ( lcsh )
Phosphatic fertilizers -- Florida ( lcsh )
Potassium fertilizers -- Florida ( lcsh )
City of Belle Glade ( local )
Fertilization ( jstor )
Soil science ( jstor )
Roses ( jstor )
bibliography ( marcgt )
non-fiction ( marcgt )


Bibliography: p. 40-41.
General Note:
Cover title.
Bulletin (University of Florida. Agricultural Experiment Station) ;
Statement of Responsibility:
T.W. Young ... [et al.].

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
023755918 ( ALEPH )
02692235 ( OCLC )
AHM1127 ( NOTIS )


This item has the following downloads:

Full Text
lletin 771 (technical) April 1975

Lose Response to Nitrogen, Phosphorus,

and Potassium Fertilization Rates


Agricultural Experiment Stations
Institute of Food and Agricultural Sciences
J. W. Sites, Dean for Research
University of Florida, Gainesville


Page 10, line 9: MXV should be NXV.
Page 26, para. 2, line 9: NO should be NO- N.
Page 28, line 15: 750 pounds should be 250 pounds.
Page 36, para. 3, line 2: Fig. 12 should be Fig. 2.

Cover photograph: the 'Christian Dior' hybrid tea rose, Plant Patent No.
1943, The Conard-Pyle Company, West Grove, Pennsylvania.


T. W. Young, G. H. Snyder, F. G. Martin and N. C. Hayslip
Dr. Young is a Professor at the IFAS Agricultural Research and Edu-
cation Center, Homestead; Dr. Snyder is an Associate Professor at the
IFAS Agricultural Research and Education Center, Belle Glade; Dr. Martin
is an Associate Professor in the IFAS Statistics Department, Gainesville;
and Mr. Hayslip is a Professor at the IFAS Agricultural Research Center,
Ft. Pierce.

This public document was promulgated at an annual cost of
$1,247.62, or .58 per copy to give rose growers in Florida,
who are using Rosa fortuniana stock on light sandy soil,
recommendations for levels and ratios of nitrogen-phosphor-
us-potassium fertilization to obtain specific results.


The rose bushes used in this work were donated by Mr. James T. Miner,
Miner's Roses, Boynton Beach, and 0. F. Nelson and Sons, Apopka. Mr.
Miner also advised on cultural and commercial aspects of the work through-
out the experiment. Dr. J. R. Iley, formerly Assistant Professor, AREC,
Belle Glade, had a major role in the design of the experiment and its opera-
tion for the first 22 months.Dr. H. W. Burdine, Professor, AREC, Belle Glade,
supervised analytical work performed by Mrs. Carrie Martin, Laboratory
Technician II, during a portion of the study. Mr. H. B. Dugan handled
the day-to-day details of operating the experiment, including harvesting
flowers and post-harvest studies. To all these, and to others aiding the work
but not mentioned here by name, the authors extend their thanks and


Materials and Methods --..........-----....--...........--------.......... 3
Results and Discussion ..............-----.......-.......------------------- 7
Flower Production --.....--...--....--...........-----...... -------- 7
Stem Length .-...--.....--...----..---.----- ------------13
Interactions .....--.----..-..-....-...---------------............ 15
Quarterly Summary ----...-.. --....... ---.-.------------19
Flower Quality ...-----...... ..-------. .-----------21
Flower Bud Size .......-..............------.........--.....--.. --21
Soil Composition .-...-.....--....-..-........... ...---------..-... 23
Leaf Composition ..----........---------------...........----------. 29
Field Ratings --...--........- ..................-.....-----------.-- 32
Root Distribution .....................---------.............----------35
Practical Considerations .......................-------.......---... ----36
Literature Cited .........------..-....---.............. --..---.......---40


Commercial rose production in Florida for the cut flower
trade reached its peak in 1963 when about 12,361,000 flowers
were produced (1). This represented the yield from approxi-
mately 450,000 bushes, practically all growing outdoors, some
in full sun, others with shade cloth covering. Greenhouse shelter
is necessary for winter production of roses only in north Florida
locations where night temperatures regularly fall below 600F.
Although the number of rose growers, as well as the number of
plants in production, has decreased in Florida during the past
several years, the potential is still good in the state for a minor
but viable cut flower industry in roses. Part of this potential is
due to a favorable winter climate, high daylight intensity
throughout the year, and nearness to eastern markets. It is also
due to the increasing widespread commercial use in Florida of
Rosa fortuniana rootstock. This stock is used almost exclusively
now for cut flower production and for the container grown bushes
that are propagated in Florida. Better yields of flowers with
longer stems and a longer period of productivity are obtained
with it than with any other stock tried under local conditions
(15, 16, 17, 18).
Most commercial rose plantings in Florida are on light sandy
soil, but a few are grown hydroponically. Under both conditions,
and regardless of rootstock, mineral nutritional problems have
been common and of major concern to growers. Most literature
on mineral nutrition of roses deals with varieties and rootstocks
for northern greenhouse production with various media in well
drained beds.
Seeley (22) found the number of salable flowers and stem
length of 'Talisman' roses growing in silty clay loam increased
with each increment of NO7- N in the soil from about 20 to 50
ppm. With 'Briarcliff' roses in the same soil, he (23) found the
number of salable flowers and stem length were optimum with
NO," N between 50 and 100 ppm in the soil, with maximum total
production at levels between 25 and 100 ppm. Usable flowers in-
creased with NOT- N, but stem lengths decreased.
With 'Better Times' roses growing in a 50:50 mixture by
volume of silty clay and river sand, Culbert and Wilde (6) re-
ported that increased KO2 fertilization alone, up to about 550
pounds per acre, did not increase yield, but increases in both N
and K20 fertilization greatly increased yield. Maximum stem
length was attained at about 270 pounds of K20 per acre. Total

growth was greatest where 385 pounds of N, 260 pounds of PO,
and 550 pounds of KO per acre were used.
Seeley (24) reported that KO fertilization sufficient to give
1.38 to 1.63% K in leaves of 'Briarcliff' and 'Hildegarde' roses in
sand culture produced longer shoots and larger flowers than those
fertilized at lower rates which had only 0.33 to 0.63% K in leaves.
Post and Fischer (20) found that 'Better Times' roses grown
in silt loam with 20% by volume of peat moss added produced
more flowers with K20 fertilization at 360 pounds per acre than
at 120 pounds, but there was no increase in number of flowers
or stem length at any level between 360 and 1920 pounds per acre
of KO.
The rose manual edited by Mastalerz and Langhans (14), es-
pecially Chapter 13 which discusses mineral nutrition of roses
and includes a key to deficiency and toxicity symptoms, is of in-
terest to the Florida rose grower.
Results of a survey by White (31) of the mineral content of
rose leaves from commercial plantings in Pennsylvania is also
helpful to the Florida grower.
The early literature on rose nutrition specifically for Florida
was aimed primarily at home gardens (4, 7, 8, 28), and recom-
mendations were not supported by results of systematic research.
Some symptoms of malnutrition of 'Pink Frill' roses on R.
fortuniana stock grown in sand culture were described (12).
Waters (30) grew 'Tropicana' roses on R. fortuniana stock on
Leon fine sand with 3 levels of N and 2 levels of KO in factorial
combinations. Increasing N fertilization increased the number
and weight of flowering stems harvested and N content of leaves.
There was no significant growth response from increased KO2
fertilization, but K content of leaves increased. The vase-life of
cut flowers was not affected by the treatments.
McFadden and Poole (19) identified a copper supplement re-
quirement for growth of young 'Baccara' roses on R. fortuniana
stock growing in light potting medium that contained no native
soil ingredient. They also found that the copper supplement regu-
lated the plant's growth responses to varied rates of P and K
fertilizers in this container culture experiment.
In response to a request from commercial growers in south
Florida for information on rose nutrition, experiments were
started at the Indian River Field Laboratory near Fort Pierce,
Florida, in November 1964. The principal part of this project
was a study with 3 levels of N, P, and K in factorial combinations
on 'Christian Dior' and 'Happiness' roses grown with plastic
mulch in the field. This work has been reported briefly (34).

The primary reason for using plastic mulch was to reduce
leaching losses, thereby minimizing this variable as much as
practicable under conditions of the experiment. It also offered an
opportunity to explore the feasibility of this sort of culture for
a long term crop as roses; it had been used on some short term
vegetable crops with considerable success. Since the influence of
the mulch on the performance of roses was not known, a supple-
mental experiment was started at the same time and location to
evaluate this influence. In this supplemental experiment both
varieties on R. fortuniana stock were grown with and without
mulch at 3 levels of N, with P and K fertilization held constant.
It was found that a heavy application of NPK fertilizer under
plastic mulch gave the same results for approximately 18 months
as an equivalent amount of fertilizer annually but applied bi-
weekly without mulch (26, 33).
This bulletin reports effects for a period of 3 years of a single
heavy application of NPK fertilizer, made under plastic mulch at
the beginning of the main experiment. Responses measured were
number of flowers, stem lengths, flower bud size, flower quality,
mineral composition of soil and leaves, and condition of plants in
the field.

The experiment was on Oldsmar fine sand' which has a
spodic2 horizon at a depth of about 20 inches. This soil is in-
herently infertile, almost a builder's sand. With the exception of
residual P from a previous farming operation, the nutrient con-
tent was of no agronomic significance. With the combination of
light, infertile soil, plastic mulch, and a relatively shallow spodic
horizon which helped retard leaching and on which sub-irrigation
was practical, the experiment was virtually a sand culture. Thus,
results are broadly applicable.
The entire experimental area was treated with dolomitic lime-
stone at rate of 1 ton per acre shortly before the bushes were
planted. Several months later the soil at the 0 to 6 inches depth
in the beds ranged from about pH 6.1 to 6.5. Each plot (25 sq.
ft.) consisted of 4 bushes planted in a row at a spacing of 15
inches. Plots were separated by 7-foot unfertilized buffer zones
without plants in the row. The rows were spaced at 7 feet on
beds 4 feet wide at top, 5 feet at bottom, and 10 inches high. Both
varieties were grown with 300, 900 and 2700 pounds at N; 0, 200

1Formerly classified as Immokalee f.s.
2 Illuvial accumulations of organic matter, iron and aluminum.

and 400 pounds of P20s; and 300, 900 and 2700 pounds of KO
per planted acre, factorially combined, in randomized split plots
replicated 4 times, with varieties constituting the sub-units.
Thus, there were 216 plots of 4 plants each. The entire supply of
fertilizer for the duration of the experiment was applied on No-
vember 1, 1964, just before the recently grafted bushes were
planted. N was derived from ammonium nitrate, P from triple
superphosphate and K from potassium chloride. Before bedding,
100 pounds of N and 100 pounds of KO per planted acre, to-
gether with all the P and substantial but safe amounts of Mg,
Mn, Cu, Zn, B, Fe, and Mo were applied at the soil surface in an
area 8 inches wide and 5 feet long, centered along the planting
row of each plot. Plant beds were then thrown up over the fertil-
izer by means of bedding disks and a bed press. The remainder
of the N and K fertilizer was distributed evenly in a 4-inch wide
band about 20 inches from the center of the bed on each shoulder
adjacent to the planting area and covered with about 1 inch of
After the fertilizer was placed, the plot beds were covered
with 6 mil black polyethylene film, which was suitably anchored.
Four-inch diameter holes were cut in the film at the proper lo-
cations, and bushes planted in the soil at the holes. Methods of
irrigation, weed, insect, mite, and disease control have been de-
scribed elsewhere (33). As a further precaution against Zn and
Mn deficiencies, the fungicides used were alternated between
materials carrying these elements. Although deficiency symptoms
had not appeared, nutritional sprays containing Mg, Cu, Fe, B,
and Mo were made monthly, beginning in August 1966.
About 6 months after planting, and prior to recording harvest
data, all shoots with flowers were cut back to the third or fourth
full (5-leaflet) leaf from the base of the individual shoot. Flower-
ing shoots that occurred after this were disbudded to the one
central flower bud, as practiced by commercial growers. Cutting
of flowers for experimental data was according to the common
commercial practice in Florida, which in general was to about
the fourth 5-leaflet leaf from the base of the shoot. Harvest was
twice weekly on all plants. The number of flowers per plot and
their stem lengths (from cut to base of receptacle) were recorded,
beginning June 1, 1965 and continuing through August 31, 1967.
These data were grouped by periods of 3 months (summer, fall,
winter, and spring) into 9 quarters, and for the first and second
full years of harvest; quarters 1 through 4 and 5 through 8, re-
spectively. In these harvest data, all stem lengths constitute

"total" flowers, and lengths above 9 inches are "commercial" flow-
ers; those 9 inches or less in length are culls.
A grading was made of the relative quality for a given date of
2770 'Christian Dior' and 1670 'Happiness' flowers harvested
from bushes in 7 selected representative treatments from Janu-
ary through August 1967. These treatments, in terms of hun-
dreds of pounds per acre of N, P,0s and KO, respectively: 3-0-3,
3-4-27, 9-2-9, 9-4-9, 27-4-3, 27-0-27 and 27-4-27. Being on a rela-
tive basis semi-weekly, the grading automatically took into ac-
count differences in quality due to climatic factors and insects
and diseases that were not related to fertilizer treatment. The
grading scale of 1 (good) to 5 (poor) was based on conformation
and size of bud, color of petals and leaves, leaf size, length of
internodes and stockiness of stem, with approximately equal
value given each standard. High quality flowers were in grades
1 and 2.
Effects of treatments on flower bud size were examined by
measurements in 1966 and 1967 on buds from these 7 selected
treatments. Abnormally large or small buds were not used in this
study. Measurements were made of stem diameter at base of re-
ceptacle; diameter, length and weight of bud, and number of
complete and incomplete petals. Buds selected for examination
were a day or so before harvest stage with sepals starting to
open. Bud diameter and length were measured with broadfaced
calipers especially suited for measuring where soft surfaces were
involved. The 1088 'Christian Dior' and 854 'Happiness' buds
measured were well distributed among the selected treatments.
Soil samples were taken through the N-K fertilizer bands to a
depth of 6 inches 4 times during the study. Samples were air
dried, mixed and analyzed for pH and extractable N, P and K.
Nitrate was determined in 1 N NaOAc, pH 4.8, extracts by the
phenoldisulfonic acid technique (13, pp. 197-201). P was deter-
mined colorimetrically in 1 N NH4OAc, pH 4.8, extracts by the
ammonium molybdate-stannous chloride method (13 pp. 141-
144) and K was determined in the latter extracts by flame pho-
tometry (13, pp. 461-464).
Tissue samples were taken 3 times during the course of the
study and analyzed for N, P, and K. The first four 5-leaflet leaves
from below the flower bud on shoots ready for harvest were se-
lected for sampling. A preliminary sampling study indicated that
these leaves best reflected the overall nutritional status of the
rose plants. This sampling was somewhat different from that
found best for greenhouse roses (5) where the upper leaves, in-
cluding tri-leaflet leaves, were best. After oven drying at 700C,

YEAR 1964 ------- 1965------- -------- 1966----.--- -------{1967----
QUARTER 2- ---5--4-- --5-- --7-
FERTILIZATION # I #2 # 3 #2 #4
#1 -#2
Figure 1. Sequence of events during the study.

the micro Kjeldahl method was used for determining N (13, pp.
183). Portions of the samples were wet digested in a mixed
acid solution (10:4:1 mixture of HNOS:HC104,:H2S04), and an-
alyzed for P colorimetrically and for K by flame photometry.
Leaf Cl was determined on selected samples by a potentiometric
titration method (11). Figure 1 shows the soil and tissue samp-
ling dates in relation to the planting, fertilization and harvest
Visual ratings of plots in the field were made in July 1966
and 1967. Rating values were 1 (excellent) to 10 (very poor),
and were based on color, size and density of leaves and on num-
ber, length, and stockiness of stems, with approximately equal
value given each standard. 'Christian Dior' plants generally were
larger, more densely foliated and more vigorous than 'Happiness'
plants, so the two varieties were rated separately. Two inde-
pendent ratings were made on each variety each time.
Using the analysis of variance technique (25), all harvest
data were analyzed statistically by quarters for the 9 quarters
of harvest, and for the first and second full years of harvest.
Soil and leaf tissue chemical analyses data were also analyzed
statistically by this method. In order to obtain the maximum
information from these data the sums of squares associated
with the 3 fertilizer elements were partitioned, using orthogonal
polynomials, into single degree of freedom components associated
with the regression relationship between fertilization rate and
the response. These results were used to determine the degree
of the relationship and to estimate the regression equations (2).
Each equation was then examined analytically to determine the
effect of each of the 3 fertilizer elements.
These equations are the primary presentation of results. Some
of the more important relationships are shown graphically with
data calculated from the equations. The main effect means of

flower production and stem lengths are tabulated for the first
and second years of harvest, and for the 9 quarters as a matter
of general interest. The reader is cautioned against conclusions
based on superficial examination of these means. They often
would be misleading. In the next section the nature of the rela-
ships, whether linear or quadratic, and significance of differences
will be presented based upon the statistical analysis. Data on
flower grades and bud sizes require no further calculations for
proper interpretation.


Flower Production
Nitrogen had the greatest effect on flower production (Table
1). The calculated response (Table 2, Equations E and J, respec-
tively) to increased N fertilization (with P20, at 115 lbs and
KO at 900 lbs/A) by both varieties is shown in Figure 2 for the
first and second years of harvest. These graphs show that the

Table 1. Main effect means yield and stem length first and second full years.

Treat. Flowers per plant Stem lengths (in inches)
lbs/A First Year Second Year First Year Second Year
(or Var.) Total Comm. Total Comm. Total Comm. Total Comm.
300 35.2 26.2 29.2 19.2 12.1 14.6 11.05 14.04
900 48.7 38.1 40.4 28.9 12.9 15.1 11.90 14.69
2700 58.3 46.5 51.1 37.9 13.0 15.1 12.43 14.76
Q** Q** Q** Q** Q** Q** Q** Q**

0 47.7 37.6 40.7 29.2 12.9 15.1 11.92 14.54
200 45.9 35.5 38.9 27.4 12.6 14.9 11.70 14.33
400 48.5 37.7 41.1 29.4 12.5 14.8 11.76 14.41
N.S. N.S. N.S. Q* L* L* N.S. N.S.

300 48.6 36.6 39.6 26.6 12.2 14.7 11.35 14.19
900 49.1 38.6 43.2 31.4 12.7 15.0 12.02 14.60
2700 44.5 35.6 37.9 27.9 13.1 15.1 12.01 14.49
L** N.S. Q** Q** Q* Q* Q** Q**

C.D. 60.4 '46.2 50.3 35.0 11.7 14.0 11.51 14.17
Hap. 34.4 27.6 30.2 22.3 13.6 15.8 12.08 14.69
** ** ** ** ** ** **

Statistical significance and relationships:
N.S. Not significant
Significant at 5% level
** Significant at 1% level
L Linear
Q Quadratic

Table 2. Regression equations for total number of flowers per plant.

Quarter Equa.

1(summer) A 11.4825 + 3.0933N 0.2249N2 0.1898K + 3.8529V + 1.0603NV 0.0733N2V

2(fall) B 6.5271 + 2.2716N 0.1521N2 1.9840P + 0.9113P2 + 1.0245K 0.0975K2 + 1.7878V + 0.8210NV


3(winter) C 3.0038 + 1.0810N 0.0853N2 + 0.3780V + 0.0941NV

4(spring) D 6.8793 + 2.7645N 0.1702N2 0.3281K + 0.8724V + 0.3409NV

1-4(1st 4) E 29.1678 + 9.2103N 0.6325N2 0.5775K + 6.6325V + 2.4890NV 0.1465N2V

5(surmer) F 3.6947 + 2.4986N 0.1611N2 + 2.4724P + 2.8467K 0.2835K2 1.7243PK + 0.1648PK2 + 2.9645V

+ 0.1466NV

6(fall) G 8.0939 + 0.8446N 0.1123N2 + 2.4086P + 1.4267K 0.2000K2 + 0.3890NK 0.0212NK2 1.6799PK

+ 0.1606PK2 + 2.0875V + 0.2165NV

7(winter) H 1.0603 + 1.2200N 0.0846N2 0.5553K 0.0430K2 0.3420V + 0.3420PV

8(spring) I 4.4772 + 1.1802N 0.1192N2 + 1.6191P + 0.9348K 0.1081K2 1.1300PK + 0.1081PK2 + 2.5279V

+ 0.1586NV

5-8(2nd 4) J 13.0578 + 7.5180N 0.4772N2 + 5.6277P + 7.6528K 0.7940K2 4.5183PK + 0.4600PK2 + 7.7084V

+ 0.5326NV

9(sunmner) K 2.1065 + 3.4964N 0.2677N2 + 0.3398K 0.3437NK + 0.0331N2K + 0.8726V + 0.1261NV

Nitrogen Rate ; P205 K K20
N = 300 -200 300-

Note: In these equations N, P205 and K20 are expressed in pounds per acre, V = +1 for 'Christian Dior
and -1 for 'Happiness'.


- 75-
a- 50-


u- 25-




(n4 I



1200 1600 2000 2400 2800

N (pounds per acre)

Figure 2. Total flowers vs. N fertilization. First year calculated from equation
E and second year from equation J in Table 2 with P,05 at 115
Ibs/A and K20 at 900 Ibs/A.




I I l I I I I
400 800 1200 1600 2000 2400 2800
N (pounds per acre)






number of flowers of all stem lengths of both varieties increased
with increasing N up to a maximum and then decreased with ad-
ditional N fertilization. Further, the equation shows that pro-
duction on 'Christian Dior' was always greater than that on 'Hap-
piness', regardless of N level. The major terms involving the
dummy variable "V" are positive and thus in the calculation
the yield increases where V = + 1 and decreases where V=
-1, the values for 'Christian Dior' and 'Happiness, re-
spectively. The equation also indicates an MXV interaction
in which N increases yield of 'Happiness' less than that
of 'Christian Dior' throughout the range of experimental N
levels used. The percentage of difference between the two varie-
ties in number of flowers increased as N increased, even beyond
the point of maximum yield. The maximum number of flowers
of all stem lengths the first year was estimated to occur at about
2,260 pounds of N per acre on 'Christian Dior' and at about 2,075
pounds on 'Happiness'. The effect on flowers of commercial
lengths was similar, but with a wider difference between varieties
in amount of N for maximum flower production, occurring at
about 2,380 and 1,965 pounds per acre on 'Christian Dior' and
'Happiness', respectively. These maximums for N agree well with
data from chemical analysis which show that the maximum N
content of leaves for both varieties at this time occurred at a
fertilizer level of about 2,100 pounds of N per acre. Optimum
levels of N for the second year were estimated at about 2,530
and 2,195 pounds per acre for all lengths, and at 2,725 and 2,100
for commercial lengths, for 'Christian Dior' and 'Happiness', re-
spectively. This apparent increase in N level required for maxi-
mum yields the second year over that of the first year resulted
from depletion of soil N with time. As pointed out elsewhere, soil
N was considerably reduced with time, no fertilizer having been
applied after the initial treatment at the beginning of the experi-
ment. Thus, for a given N level figure, actual N in the soil was
lower the second year than the first. This resulted in artificially
high values for the estimate of optimum levels for the second
The number of flowers produced at all levels of N fertilization
was somewhat less on 'Christian Dior' the second year than in
the first year, whereas 'Happiness' yields were practically the
same for respective N levels both years. The decrease in yield
of 'Christian Dior' the second year was not due to N deficiency
because the maximum yield occurred at an N level below the
maximum rate of application. Subsequent data in this paper do
not indicate that P or K were involved in this decrease. The most

probable explanation is that it resulted from a heavy infestation
of the two-spotted mite, Tetranychus urticae Koch, in January
and February 1966. Although no systematic study was made of
the relative degree of infestation and damage on the two varie-
ties, it is commonly observed by Florida rose growers that
'Christian Dior' plants, which generally are relatively vigorous
and succulent, are more subject to attack and damage by this
mite than the less thrifty 'Happiness' plants. This hypothesis is
supported by reports on variation in severity of attack by mites
and insects with variation in vegetative vigor of other species
of plants (9, 10, 21, 29). Examination of the quarterly yield data
(Table 6) shows that for 'Christian Dior' most of the difference
in yield between the 2 years resulted from a substantial decrease
in the summer of 1966 as compared with that for the previous
summer, whereas 'Happiness' yields were practically the same
both summer quarters. This decrease correlated well with the
effects that could be expected from a heavy mite infestation.
Only the effects of various amounts of N on number of flow-
ers has been considered thus far. The effects of various levels
of N, P, and K can be calculated for a particular year by using
the equations in Table 2 and letting V = + 1 for 'Christian Dior',
V 1 for 'Happiness', and specifying the N, P20s and K20
fertilizer levels in pounds per acre. The first year P20, had no
significant effect on the total number of flowers (Table 1). The
effect of K was to linearly reduce the yield of flowers of all stem
lengths by 0.232 of a flower per plant per year for each 120
pounds of KO applied. Neither P nor K had a significant effect
on number of commercial flowers the first year. This suggests
that K20 fertilization, by increasing stem length (discussed
later), tended to reduce the number of culls.
While the general effect of N on total number of flowers was
about the same the second year as the first, the effect of K on
number of flowers of all stem lengths the second year depended
upon the amount of P applied. The nature of this interaction,
with N at 1300 pounds per acre is shown in Figure 3. From this
it is seen that at K levels giving optimum yields, increasing P
reduced yields. At lower and higher levels of K, the effect of P
in reducing yields was less. The response of flowers of commer-
cial stem lengths to P and K the second year was similar to the
response of flowers of all stem lengths. At optimum K fertiliza-
tion levels, the effect of increasing P fertilization was less for com-
mercial lengths than for all stem lengths the second year. Yields
decreased with increasing P,05 up to about 240 pounds per acre
and then increased slightly at higher P levels (Fig. 4). As with

,w 45.0
S300 -45.0 47.5
S200 42.5 50.0

100 -52.5

300 600 900 1200 1500 1800 2100 2400 2700
Figure 3. Total number of flowers per plant (average of two varieties) as
affected by PxK interaction in second year. Calculated from equa-
tion J in Table 2 with N at 1,300 Ibs/A.

N, there was a depletion of soil K with time since no fertilizer
was applied after initial treatment. Therefore, yields decreased
throughout most of the K fertilization range the second year. This
was consistent with data for the first year
Data presented in Table 1 show that there was no important
response to P fertilization at the rates used in this experiment.
It was present in the soil (about 30 lbs/A) in amounts adequate
or in excess for most crops. Except for its effect through the in-
teraction with K in the second year, P had little effect on the
number of flowers. In general, best results were obtained at low
P fertilization levels. The response obtained by McFadden to P


200- s33.75
S\ ) 31.25 /31.25
,100 -36.25
0 2875

300 600 900 1200 1500 1800 2100 2400 2700

Figure 4. Commercial flowers per plant (average of two varieties) as affected
by PxK interaction in second year. Calculated from equation J in
Table 3 with N at 900 Ibs/A.

fertilization was attributed to Cu impurities in the P fertilizer
(19). In the current study, however, Cu was maintained at an
adequate level through soil and foliar applications.

Stem Length
Average stem lengths of all flowers from both 'Christian
Dior" and 'Happiness' were affected in a parallel quadratic man-
ner by increasing N fertilization. The calculated response to N
fertilization (with P205 at 115 lbs and KO at 900 lbs/A) is
shown in Figure 5 for the first and second years. The optimum
level of N for stem length the first year was about 1,800 pounds
per acre. In the second year the effect of N on stem length was
complicated by significant interactions between N and K and
between P and K, but in general the average stem length in-
creased quadratically with increasing N to an estimated optimum

-- 15 -
"uj,_ _------------ ---

U 5--
S0o I I I I
400 800 1200 1600 2000 2400 2800
N (pounds per acre)

b 15
) 5-

400 800 1200 1600 2000 2400 2800
N (pounds per acre)
Figure 5. Stem lengths vs. N fertilization. First year calculated from equation
E and second year from equation J in Table 4 with P205 at 115
Ibs/A and K20 at 900 Ibs/A.

at 2,452 pounds per acre. This value, as with optimum N levels
for yield of flowers the second year, was probably artificially
high because of depletion of soil N by the second year.
The average increased stem length of flowers of all stem
lengths resulting from increasing N from the minimum level
(300 lbs/A) to the optimum level was about 1.4 inches for both
varieties the first year. The second year the average increase was
about 7 inches, with stems about 2.5 inches longer at the opti-
mum level than in the first year. Regardless of fertilizer levels,
'Happiness' consistently produced longer stems than 'Christian
Dior,' with the difference uniformly at about 2 inches both years.
The effect of P on stem length the first year was somewhat
different for the two varieties. With 'Happiness', average stem
length of all flowers continuously decreased with increasing P.
With 'Christian Dior' the rate of decrease became less with in-
creasing P20O until a rate of about 240 pounds per acre was
reached, and then stem lengths increased with increased P.
K affected stem lengths of all flowers in a quadratic manner
the first year, with maximum lengths at about 2040 pounds of
K20 per acre for both varieties.
The effects of an interaction between P and K on stem lengths
of all flowers the second year, with N level at 900 pounds per
acre, is shown in Figure 6. These curves indicate that P had little
effect on stem length at low K levels, but at levels of K giving
maximum stem lengths, there was also an optimum level for P205
at about 200 pounds per acre.

400 9.0 9.8 10.6 11.4 12.2 13.0 13.8 13.8



a 100
C 0 I I I
300 600 900 1200 1500 1800 2100 2400 2700

Figure 6. Stem lengths (inches--average of two varieties) as affected by
PxK interaction in second year. Calculated from equation J in Table
4 with N at 900 Ibs/A.

Both varieties rather consistently had an average of around
75% of flowers in commercial classes, i.e., with stems over 9
inches in length for the first seven quarters, but dropped to
around 50% in the last two quarters. The first year commercial
stem lengths responded in a quadratic manner to both increasing
N and K fertilization, and there was reduction in stem lengths
of commercial classes with increasing P. All these effects, as well
as optimum levels of N and K for stem lengths, were similar to
those for flowers of all stem lengths the first year. The PxV inter-
action that appeared for flowers of all stem lengths the first year
did not occur with flowers of commercial stem lengths.
There was a quadratic effect on commercial stem lengths from
increasing N and K fertilization again the second year, but in-
creasing P fertilization had no effect on commercial lengths.
Neither was there a PxK interaction the second year as with
flowers of all stem lengths. Estimated optimum levels of N for
commercial lengths the second year was 1,890 pounds per acre
for 'Christian Dior' and 2,320 pounds for 'Happiness'. Optimum
levels of KO were estimated to be 1,870 pounds per acre for
'Christian Dior' and 1,540 pounds for 'Happiness' for flowers
of commercial lengths the second year. This is in contrast with
the report (14, Chap. 13, p. 90) that a ratio of approximately
1 : 2 of N:K resulted in a higher percentage of longer stems on
'Colorado No. 6' rose than other ratios of N:K.

Figures 3, 4, and 6 illustrate the complexity of some of the
interactions observed. In addition, during certain quarters (Table
3, eq. F and I), and (Table 4, eq. F, G, J and K), a PxK interac-
action occurred which involved the terms PK, P2K, PK2 and
P2K2. The interaction takes the form shown in Figure 7. This
indicates an optimum value for P and K, with the influence of P
being reduced at levels lower or higher than optimum K, and the
influence of K being reduced at lower or higher levels than opti-
mum P. A somewhat similar relationship resulted from the PxK
interaction involving only the terms PK and PK2 (Table 2, eq. F,
G, I and J), and (Table 3, eq. G and J), as shown previously
(Fig. 3), except that the optimum P level appeared to be about
zero. However, in some cases the interaction was modified by a
further N interaction with P and K.
A number of other interactions occasionally occurred during
a single quarter, but which were not repeated with sufficient
frequency to make them appear to indicate a consistent trend.

Table 3. Regression equations for number of commercial class flowers per plant.

Quarter Equa.

1 (summer) A 8.2608 + 2.8872N 0.2109N2 + 3.7010V + 0.2546NV

2 (fall) B 3.7445 + 2.0193N 0.1330N2 1.7033P + 0.7830P2 + 1.2813K 0.1139K2 + 1.2421V +0.1901NV

3 (winter) C 1.8017 + 1.0391N 0.0840N2 + 0.3132V + 0.0765NV

4 (spring) D 3.9874 + 2.2454N 0.1383N2 0.2021K + 0.6301V + 0.2633NV

1-4 (1st 4) E 18.6181 + 8.1912N 0.5661N + 5.8864V + 0.7845NV
2 2
5 (summer) F 3.0474 + 2.4077N 0.1571N2 3.1943P + 2.5988P + 2.0818K 0.2083K2 + 2.2278PK 0.21301

1.8127P2K + 0.1733P2K2 + 2.2024V + 0.1307NV

6 (fall) G 4.5041 + 0.5929N 0.1157N2 + 1.1711P + 0.7725P2 + 1.5160K 0.2146K2 + 0.4459NK 0.02291

1.8810PK + 0.1798PK2 + 0.0311V + 0.1397NV + 0.5888KV 0.0483K2V

7 (winter) H 0.0244 + 1.3485N 0.0989N2 + 2.5875P 1.2938P2 + 0.3743K 0.0423K2 1.8046NP + 0.90231

0.1725N2P 0.0863N2P2 0.4151V + 0.2951PV + 0.3273KV-0.0332K2V

8 (spring) I 0.2407 + 1.2774N 0.0794N2 1.7501P + 1.3620P2 + 1.0183K 0.0983K2 + 1.2201PK 0.1166]

0.9499P2K + 0.0908P2K2 + 0.8899V + 0.1654NV

5-8 (2nd 4) J 7.8739 + 5.2238N 0.4222N2 + 5.4137P + 1.9045P2 + 5.0447K 0.5341K2 + 0.7122NK 0.05781

4.2460PK + 0.4111PK2 + 4.4011V + 0.4590NV

9 (summer) K 1.3973 + 1.2839N- 0.0688N2 + 0.5612K 0.0635K2 + 0.2197V + 0.1368NV

SNitrogen ; p= P20; K K90
300 200 300
Note: In these equations N, P205 and K20 are expressed in pounds per acre, V = +1 for 'Christian Dior'
and -1 for 'Happiness'.

P -

NK2 -

NP2 +

PK2 -

NK2 -

Table 4. Regression equations for average stem length of all flowers.

Quarter Equa.

1 (summer) A 13.1389 + 0.6519N 0.0588N2 0.3319P + 0.0765K 1.0412V

2 (fall) B 11.4986 + 0.7333N 0.0563N2 + 0.5256K 0.0378K2 2.0453V
3 (winter) C 11.7799 + 0.5585N 0.0732N 0.8613K + 0.0823K2 + 0.1988NK 0.0190NK2 0.2990V
4 (spring) D 9.5342 + 0.6706N 0.0516N 0.2518V

1-4 (1st 4) E 10.8605 + 0.6413N 0.0529N2 0.2048P + 0.3871K 0.0279K2 1.0221V 0.1521PV + 0.1479p2V

5 (summer) F 11.7484 + 3.0508N -0.2997N2 2.9610P + 1.4790P2 + 1.7382K 0.1849K2 1.7972NK + 0.1768NK2 +

0.1768N2K 0.0169N2K2 + 2.0651PK 0.1974PK2 1.0315P2K + 0.0986P2K2 0.9884V

6 (fall) G 12.5286 0.5698N + 0.0296N2 1.6350P + 1.1835P2 0.6267K + 0.0456K2 + 0.7678NK 0.0676NK2 -

0.0586N2K + 0.0056N2K2 + 1.1404PK 0.1090PK2 0.8254P2K + 0.0789P2K2 0.4713V 0.0672NV

7 (winter) H 11.6829 + 0.1950N 0.3173P + 0.5556V

8 (spring) I 5.5503 + 0.4092N 0.0261N2 + 0.4965K 0.0436K2

5-8 (2nd 4) J 5.6223 + 1.7031N 0.1679N2 6.5970P + 3.2985P2 0.0127K + 0.0253K2 + 0.3470NK 0.0276NK2 +
2 2 2 2
4.6010PK 0.4398PK 2.3005P K + 0.2199P K 1.1421V

9 (summer) K 9.4931 + 0.1582N 3.6997P + 1.7325P2 0.0293K + 0.0027K2 + 2.4166PK 1.2104P2K 0.2310PK2 +

0.1157P2K2 1.1215V + 0.4238NV 0.0317N2V 0.0630KV

N= ;Nitrogen P 205; K= K 20
300 200 300

Note: In these equations N, P205 and K20 are expressed in pounds per acre, V = +1 for 'Christian Dior' and

-1 for 'Happiness'.

Table 5. Regression equations for average stem length of commercial class flowers.


1 (summer)

2 (fall)

3 (winter)

4 (spring)

1-4 (1st 4)

5 (summer)
6 (fall)

7 (winter)

8 (spring)

5-8 (2nd 4)

9 (summer)














9.6080 + 1.2657N 0.0904N2 + 1.2511K



0.0516K 0.6171V 0.1987NV + 0.0197N2V

0.0459K2 + 0.1109NK 0.0106NK2 1.4796V

0.0357NV 0.2220KV + 0.0182K2V

0.2218K 0.0161K2 0.9037V

14.2074 +

15.3918 +

14.7472 -

12.9136 +

13.9415 +

15.8704 +

15.3818 +

14.0219 +

10.5769 +

- 0.1290NV + 0.1210KV

13.1244 + 0.0806N 0.2930P 0.6654V + 0.0684NV 0.0424KV

Nitrogen. P = P5 K =

Note: In these equations N, P205 and K20 are expressed in pounds per acre, V = +1 for 'Christian Dior'

and -1 for 'Happiness'.

0.3844N 0.0365N 0.1805P

0.4415N 0.0336N2 0.1909K


0.3828N 0.0304N2 0.1832V

0.3672N 0.0313N2 0.1333P

0.4034N 0.0298N 0.7708V

0.0706N 0.3611K+ 0.0379K2

0.0708N + 0.3101V

0.0878N + 0.3343K 0.0333K2

- 0.0300V 0.0686NV

- 0.0089V 0.0566KV

- 0.1104K2 1.0075V


5 300



0 300 600 900 1200 1500 1800 2100 2400 2700

Figure 7. Number of commercial flowers per plant (average of two varieties)
as affected by PxK interaction in fifth quarter. Calculated from
equation F in Table 3 with N at 1,300 Ibs/A.

Weather and other variables in the environment no doubt have
a strong influence on the nature of these interactions, and the
probability of the same interaction recurring often is remote.
They are more of academic than practical importance.

Quarterly Summary
Quarterly yield and stem length data (Table 6) show that sea-
sonal factors (probably temperature and perhaps total insolation,
but not moisture because of adequate irrigation) had greater in-
fluence on performance of the plants than did fertilizer treat-
ment or variety. Regardless of the fertilizer element or rate con-
sidered, in general, on both varieties, flower production decreased
drastically from summer through fall to the low level for the
year in winter, and increased in spring, reaching the year's peak
in summer. It has been thought by some growers that increasing
K fertilization in fall and winter would help maintain production
at relatively high levels during the winter season. There is no
support for this practice in these data. The only significant in;
crease in yield in winter, or at any season, from increased fer-
tilization was that from N. With time, because of nutrient
depletion, there was a general downward trend in yields in all
seasons. In general, the variations in total and commercial yields
with season were similar.
All stem lengths and commercial lengths on both varieties
followed the same general course from quarter to quarter regard-
less of fertilizer element or rate involved. Lengths increased from
the first summer into fall, decreased the following winter and

Table 6. Main effect means of total and commercial flower production per plant and
stem lengths for 9 quarter(

Lbs. N Lbs. P205
per Flowers Stem Length per Flowers Stem Length
Quar. Season acre All Com. All Com. acre All Com. All Corn.

1 summer 300 54.1 43.7 12.7 14.6 0 68.4 58.3 13.6 15.0
2 fall 38.7 29.3 13.3 15.7 53.0 42.2 14.2 16.2
3 winter 16.0 11.0 11.3 14.5 21.2 15.6 12.0 14.7
4 spring 32.2 20.9 10.2 13.3 48.3 34.3 11.0 13.8
5 summer 39.1 32.0 14.3 16.2 54.6 46.4 15.0 16.8
6 fall 39.9 28.2 11.9 14.9 52.6 40.0 12.7 15.3
7 winter 9.5 6.9 11.3 14.0 15.2 12.3 12.6 14.6
8 spring 28.2 9.6 6.8 11.0 40.6 18.0 7.4 11.5
9 summer 21.8 12.5 9.4 12.9 32.5 20.9 10.2 13.5

1 summer 900 71.7 60.1 13.6 15.1 200 67.7 55.5 13.0 14.7
2 fall 52.0 41.2 14.3 16.3 48.7 38.6 14.2 16.3
3 winter 21.9 16.7 12.4 14.9 19.5 14.6 12.2 14.7
4 spring 48.9 34.4 11.1 13.8 47.9 33.4 10.8 13.6
5 summer 53.9 46.2 15.1 16.8 51.1 43.6 14.9 16.6
6 fall 51.5 39.2 12.8 15.3 50.4 37.0 12.5 15.2
7 winter 16.6 13.0 12.2 14.4 15.9 12.7 12.1 14.1
8 spring 39.6 17.2 7.4 11.4 38.3 16.3 7.3 11.4
9 summer 33.9 20.5 9.6 13.0 34.9 21.5 9.8 13.1

1 summer 2700 81.1 68.7 13.2 14.8 400 70.9 58.7 12.9 14.7
2 fall 62.7 51.4 14.7 16.5 51.7 41.1 13.9 16.1
3 winter 23.2 17.4 12.2 14.9 20.6 14.9 11.8 14.8
4 spring 66.2 48.5 11.4 13.9 51.2 36.1 10.8 13.5
5 summer 67.5 58.8 15.6 17.1 54.8 47.1 15.1 16.8
6 fall 64.4 49.8 13.2 15.6 52.8 40.2 12.7 15.3
7 winter 21.5 18.0 13.0 14.6 16.5 12.9 12.0 14.3
8 spring 51.2 25.0 7.9 11.8 40.1 17.6 7.3 11.3
9 summer 46.3 31.5 10.6 13.6 34.8 22.2 9.7 12.9

Lbs. K20 Variety
per Flowers Stem Length Flowers Stem Length
Onar. Season acre All Com. All Com. Var. All Com. All Com.

1 summer 300 70.3 57.3 12.8 14.6 C.D. 93.9 76.7 12.1 13.9
2 fall 49.0 36.9 13.5 15.9 65.6 48.9 12.1 14.7
3 winter 20.6 14.3 11.6 14.7 23.5 17.6 11.7 14.1
4 spring 54.6 37.9 10.7 13.5 58.5 41.7 10.6 13.2
5 summer 52.4 44.0 14.7 16.5 67.9 56.7 14.0 15.9
6 fall 48.2 33.9 11.9 14.9 64.1 45.9 11.9 14.9
7 winter 16.3 12.6 11.8 14.1 16.6 13.8 12.8 14.6
8 spring 41.3 16.0 7.0 11.3 52.5 23.7 7.4 11.1
9 summer 35.0 21.5 9.7 13.1 39.7 24.8 9.4 12.6

1 summer 900 71.6 60.1 13.2 14.8 Hap. 44.1 38.3 14.2 15.7
2 fall 54.0 43.5 14.2 16.3 36.6 32.4 16.1 17.7
3 winter 21.4 16.1 12.1 14.7 17.3 12.5 12.3 15.4
4 spring 49.3 34.9 10.9 13.7 39.7 27.5 11.1 14.1
5 summer 57.6 50.1 15.2 16.8 39.1 34.6 16.0 17.5
6 fall 55.5 42.1 12.9 15.5 39.8 32.1 13.4 15.6
7 winter 17.7 14.2 12.4 14.4 15.2 11.5 11.7 14.0
8 spring 41.9 19.2 7.6 11.7 26.8 10.9 7.3 11.6
9 summer 36.1 23.9 10.2 13.3 28.4 18.3 10.4 13.7

1 summer 2700 65.0 55.1 13.5 15.0
2 fall 50.0 41.5 14.6 16.4
3 winter 19.2 14.7 12.2 14.8
4 spring 43.3 31.0 11.0 13.7
5 summer 50.4 42.9 15.1 16.9
6 fall 52.1 41.1 13.1 15.4
7 winter 15.8 12.6 12.2 14.3
8 spring 39.7 17.3 7.4 11.4
9 summer 34.0 21.5 9.9 13.2

spring, increased the second summer, decreased the following
fall, winter and spring, and increased the third summer. As
with the number of flowers, stem lengths decreased gradually
with time because of nutrient depletion.
Results from this experiment are in good agreement with
those of Waters (30) within the N fertilizer range common to the
two tests. As N fertilization was increased on 'Tropicana' roses
on R. fortuniana stock from about 320 to about 960 pounds per
acre per year the number of flowering stems increased linearly,
without evidence that maximum production was reached at the
high rate. Increasing K20 fertilization (with KNO, and K2S,4)
from about 300 to about 600 pounds per acre per year had no
measurable effect on the growth responses of 'Tropicana' roses.
In the present study, KO fertilization was from 300 to 2700
pounds per acre, all derived from KC1. The possibility that the
decrease in number of flowers with increased K fertilization re-
sulted from excess chlorides is recognized, but is discounted, and
will be discussed later.

Flower Quality
The post-harvest grading of flower quality showed a highly
significant effect due to N for both varieties. A slight, but not
clear cut, effect due to P was detected for 'Happiness' but not
for 'Christian Dior', while K apparently had no effect on quality
of either variety. The nature of the fertilizer effects on flower
quality for each level of N, P and K is shown in Table 7. As the
amounts of N increased, the percentage of high quality flowers
(Grades 1 and 2) increased.

Flower Bud Size
No significant difference in flower bud size due to treatments
was detected for any of the 'Happiness' responses (Table 8). For
'Christian Dior', only bud diameter and bud length were signifi-
cantly affected by the treatments. Buds of greatest diameter and
length were obtained from the 27-0-27 (N, P2,0 and K,O, respec-
tively, in 100s Ibs/A) treatment, and those of least diameter and
length from the 3-4-27 treatment. Although there was a trend
towards increase in bud diameter, bud length and bud weight
with increased N for both 'Christian Dior' and 'Happiness', the
relationship between rate of any particular fertilizer element or
combination of elements and increase in size was not consistent.
The differences were small and probably of little practical im-

Table 7. Percentage distribution of flowers by grades.

Christian Dior Happiness
Good to Poor Good to Poor
Treatment- 1 2 3 4 5 1 2 3 4 5

N P205 K20

N 9

P205 2

K20 9









































1/ In 100s lbs/A.

Table 8. Treatment means Flower bud size.

Stem Bud
Treatment-/ Diameter Diameter Length Weight No. of petals
Variety N P205 K20 (in mm) (in mm) (in mm) (in gms) Complete Incomplete

Christian 3 0 3 4.120a2/ 24.59a 37.81 bc 7.930a 33.82a 3.432a
Dior 3 4 27 3.785a 22.74 b 36.78 c 7.268a 34.10a 2.908a
9 2 9 4.158a 24.44a 38.36abc 8.325a 34.98a 3.352a
9 4 27 4.118a 24.73a 39.17ab 8.362a 33.98a 3.188a
27 0 27 4.240a 25.00a 39.93a 8.685a 36.19a 3.165a
27 4 3 3.965a 23.87ab 39.25ab 8.115a 33.99a 3.092a
27 4 27 4.212a 24.43a 39.22ab 8.452a 35.06a 3.175a
Happiness 3 0 3 3.878a 22.21a 35.03a 7.108a 31.44a 3.695a
3 4 27 3.952a 21.11a 35.81a 6.800a 29.37a 2.592a
9 2 9 3.908a 22.30a 36.22a 6.738a 28.19a 3.218a
9 4 27 3.812a 21.82a 37.22a 6.692a 29.50a 3.465a
27 0 27 3.998a 22.26a 37.59a 7.178a 29.42a 3.988a
27 4 3 4.030a 21.98a 36.26a 6.705a 30.74a 2.960a
27 4 27 3.992a 22.07a 37.24a 7.287a 28.62a 3.940a

/ In 100 lbs/A.

SAny means within a column followed by the same letter are not significantly different
at the 5% level (Duncan's MRT).

Soil Composition
The effects of N, PO, and KO fertilization on soil pH, NH1 N,
NOT-N, P and K at the four sampling dates are summarized in
the regression equations presented in Table 9. Using these equa-
tions, the value of soil pH, NHT-N, NO-N, P or K at any level
of N, PAO, and K20 fertilization in any sampling can be calcu-
lated. Variables not significant at the 95% confidence level do not
appear in the equations. Of the many points that could be dis-
cussed from inspection of these equations, only those more con-
sistent and important will be considered here.
From an examination of the equations for soil pH, (Table
9, eq. A, B, C, and D), it is clear that only N and K significantly
affected pH, as P does not appear in any of the four equations.
N and K independently affected pH in the first two samplings
(Table 9, eq. A and B). However, the general trend was for in-
creasing N fertilization to decrease soil pH and for increasing K
fertilization to increase pH. A calculated example of these trends
is given in Table 10 for equation C. The effect of N has been
widely observed and is explained on the basis of acidity produced
during nitrification of the NH4-N in the fertilizer. Increasing
K probably increased soil pH by mass action displacement of this
acidity from the fertilizer band area. In the latter two samplings
(Table 9, eq. C and D), N was less effective in decreasing soil pH.

Table 9. Regression equations for the soil analysis data.

Equation Sampling

A 1 pH = 7.1172 0.03753N + 0.06130N2 + 0.03114K

B 2 pH = 7.0053 0.03657N + 0.06111N2 + 0.04988K

C 3 pH = 7.4513 0.03359N + 0.05134K + 0.0 692NK

D 4 pH = 7.6663 0.03500N 0.03717K + 0.06245K2 + 0.06601NK 0.09188NK2
E 1 NH4-N = 4.9628 + 0.0730N 0.0213K

F 2 NH -N = 1.1943 0.03490N + 0.06559N2 + 0.03117K 0.06477NK

G 3 NH4-N = 1.1093 0.03282K

H 4 NHt-N = 2.4614 + 0.03275N 0.03352K

I 1 N05-N = 34.2556 + 0.1503N

J 2 NO5-N = 27.3612 + 0.0682N + 0.03652K 0.04158NK + 0.5587P 0.02276P2

K 3 NO3-N = 0.9640 + 0.02443N 0.02189K + 0.05114K2 0.05147NK 0.0739P +

0.03417P2 + 0.04108NP + 0.03192PK 0.07785PK2 0.05117P2K 0.0944P2K2

L 4 NO3-N = 0.0508 + 0.0306N + 0.02375K 0.04104NK

M 1 P = 13.3333 0.02418N + 0.05109N2 + 0.02445K 0.05145K2

N 2 P = 12.8058 0.02127N + 0.03787K

0 3 P = 18.7771 0.02225N + 0.02100K

p 4 P = 16.8605 0.02284N 0.03197K + 0.06889NK + 0.0112P

Table 9. (continued)

Equation Sampling

Q 1 K = -20.3397 + 0.8169K

R 2 K = 199.3077 0.3247N + 0.04869N2 + 0.3489K

S 3 K = -37.2743 0.1153N + 0.04733N2 + 0.6509K 0.03158K2 0.03367NK

+ 0.06121NK2

T 4 K = -47.3941 0.0926N + 0.05630N2 + 0.5627K 0.03146K2- 0.03339NK


(a) The units on the left hand side of the equation are in ppm, and on the right hand side are in pounds per
acre of the element.

(b) The superscripts above the zeros indicate the number of zeros following the decimal point, i.e.
0.03519 = 0.000519.

Table 10. Calculated effect of N and K 0 fertilization
on soil pH (3rd sampling).-/

N 20pH

300 300 7.35
1500 300 6.94
2700 300 6.52

300 300 7.35
300 1500 7.37
300 2700 7.39

Calculated from equation C, Table 9, with K20 con-
verted to its elemental equivalent.

Equations for soil NH-N (Table 9, eq. E, F, G, and H) readily
show that only N and K affected soil NH+N. Generally, increas-
ing N fertilization increased soil NH+-N, as should be expected, with
the increase per 100 pounds of N fertilizer per acre being 7.3 and
0.028 ppm for the first and last sampling dates (Table 9, eq. E
and H) respectively. For three of four sampling dates, increasing
K fertilization decreased soil NH+-N linearly, probably due to
mass action displacement.
At the first sampling date (Table 9, eq. I) approximately 8
months after fertilization, only N fertilization affected soil NOT-N.
A 15.0 ppm increase was observed for each 100 pounds of N applied.
But in later samplings (Table 9, eq. J, K and L) all three factors
affected soil NO;-N and the relationship became quite complex.
Generally, soil NO`-N increased with increasing N fertilization
and decreased with increasing K fertilization. An example of
these trends, using equation J, Table 9, is presented in Figure 8.
It was not clear why soil NOT was statistically linked with P in
the second and third samplings. It should be remembered that all
fertilizer P was placed in a band below the plant row, whereas
the bulk of the N and K was placed on the bed surface over 22
inches away from the P band. The soil samples came from the

N-K band. Since P probably moved little in the soil, it should
have had little or no direct influence on the soil N and K in the
N-K band. This conclusion is further supported by the observa-
tion that the variable P only occurs in 3 of the 20 equations pre-
sented in Table 9. It is probably for this reason that P was not
a significant factor in affecting soil P in the region of the N-K
fertilizer band, at least until the last soil sampling, taken 3 years
after the fertilization. Thus, the soil P being measured in these
samples was the residual P in this soil, and not that applied ex-
perimentally. The extractable soil P averaged about 15 ppm of P
(about 70 pounds PO,/A), which would be considered an adequate
to high level for most crops. This probably accounts for the gen-
eral lack of plant response to P as mentioned elsewhere in this
bulletin. An example of the effect of N and K on soil P, using
equation N in Table 9, is presented in Table 11. The general effect
of N to decrease soil P and for K to increase soil P in the region
of N-K band is consistent with the effects of N and K on pH, as
discussed above. Decreasing pH would favor P movement from
the sampling region, and increasing pH would favor its immo-


S120 200 400
S100- P205
80 K
40 I I I I
0 500 1000 1500 2000 2500 3000
N, P205 or K20 (POUNDS PER ACRE)

Figure 8. Calculated effect of N, P.,O and K,O fertilization on soil NO, (second
sampling). When N is the variable, P.,O5 and K.,O are set at 115 and
300 Ibs/A, respectively. When PsO5 is the variable, N and K20 are
set at 2000 and 1200 Ibs/A, respectively. When K20 is the vari-
able, N and P0s5 are set at 250 and 115 Ibs/A, respectively.

Table 11. Calculated effect of N and K20 fertilization
on soil P (2nd sampling).-/

(lb./acre) (ppm)

300 300 12.62
1500 300 11.10
2700 300 9.58

300 300 12.62
300 1500 13.41
300 2700 14.20

SCalculated from equation N, Table 9, with K20 con-
verted to its elemental equivalent.

Soil K (Table 9, eq. Q, R, S, and T) was affected by N and K
fertilization in a manner analogous to the way soil N was af-
fected by N and K fertilization. That is, increasing K fertiliza-
tion generally increased soil K and increasing N generally de-
creased soil K (Fig. 9). However, only K fertilization significant-
ly affected soil K in the first sampling (Table 9, eq. Q). In the
second sampling (Table 9, eq. R) soil K increased linearly at a
rate of 33 ppm, for each 100 pound per acre of K fertilization,
and N fertilization, decreased soil K up to an N rate of 1,868
pound per acre. In the final two samplings (Table 9, eq. S and T)
the relationship was more complex, but the same general trends
persisted, though they were affected by interactions. At low K
fertilization it took less N fertilization to minimize soil K than
was required at higher K fertilization. For example, at a K fer-
tilization rate of 750 pounds per acre, soil K was minimized at an
N fertilization rate of 1358 pounds. At a K rate of 750 pounds,
soil K was minimized at an N rate of 2193, and at a K rate of
2250, soil K was minimized at an N rate of 2247. The reduction
in soil K with increased N fertilization was probably largely due
to mass action displacement by NH+ and by H+generated by nitri-
fication. Naturally, as K fertilization was increased, a higher N
fertilization was required to minimize soil K.

- 500 -
E 300-
I I I I I I-
0 500 1000 1500 2000 2500 3000

Figure 9. Calculated effect of N and K,O fertilization on soil K (second
sampling). When N is the variable, KO is set at 2,100 Ibs/-A. When
KO is the variable, N is set at 300 Ibs /A.

Leaf Composition
In general, the trends observed for soil N, P, and K with
varying N, P, and K fertilization were also reflected in the leaf
tissue analyses. The main effect means for the three samplings
(Table 12) are summarized in the regression equations in Table
13, where the dummy variable "V" is set at +1.0 for the variety
'Christian Dior' and at -1.0 for the 'Happpiness' variety. These
equations (Table 13, eq. U, V and W) show that tissue N was
consistently affected by N fertilization. At the time of the first
sampling (Table 13, eq. U) approximately 17 months after initi-
ation of the experiment, the maximum N content of the leaves
of either variety occurred at an N fertilization rate of about
2,100 pound per acre. It was at about this same N rate that the
maximum blossom production occurred at this point in the ex-
periment, as discussed earlier. In the latter two tissue samplings
(Table 13, eq. V and W) the N content increased linearly with N
fertilization, increasing by 0.0099% and 0.0060% per 100 pound
of N fertilization per acre, respectively, in the second and third
samplings. Naturally, at the time of these latter samplings the
actual soil N was considerably reduced, since the only fertiliza-
tion occurred at the start of the experiment. Tissue N was linear-
ly reduced by K fertilization in the last sampling (Table 14) and
was similarly affected by K fertilization at the 90% confidence
level in the first sampling (not shown in Table 13, eq. U). This

Table 12. Main effect means leaf tissue analysis for 3 samplings.

% dry weight
Treat. Nitrogen Potassium Phosphorus
Ibs/A April June Oct. April June Oct. April June Oct.
(or Var.) 1966 1967 1967 1966 1967 1967 1966 1967 1967
300 2.592 1.695 1.812 1.412 1.205 1.313 0.270 0.279 0.372
900 2.778 1.819 1.818 1.268 1.086 1.191 0.238 0.212 0.282
2700 2.893 1.951 1.947 1.144 0.946 1.049 0.211 0.177 0.222
Q** L** L** Q** Q* Q* Q* Q** Q**

0 2.725 1.810 1.834 1.269 1.076 1.186 0.240 0.217 0.281
200 2.770 1.850 1.862 1.269 1.048 1.166 0.232 0.217 0.286
400 2.769 1.805 1.880 1.286 1.113 1.202 0.247 0.244 0.310
N.S. N.S. N.S. N.S. Q* N.S. N.S. L** L**

300 2.802 1.823 1.880 1.027 0.897 0.996 0.237 0.213 0.277
900 2.745 1.856 1.910 1.287 1.090 1.212 0.247 0.229 0.298
2700 2.716 1.786 1.786 1.511 1.250 1.346 0.236 0.235 0.302
N.S. N.S. L** Q** Q** Q** N.S. L* L*
C.D. 2.706 1.852 1.816 1.310 1.075 1.169 0.233 0.232 0.304
Hap. 2.802 1.791 1.902 1.239 1.083 1.200 0.247 0.220 0.281
** N.S. ** ** N.S. **

Statistical significance and relationships:
N.S. Not significant
Significant at 5% level
** Significant at 1% level
L Linear
Q Quadratic

could be expected in light of decreased soil N brought about by
heavy K fertilization, as discussed above. The 'Christian Dior'
variety, which produced more blossoms than 'Happiness', also
had a higher N content in its leaves.
The P content of the leaves was generally reduced by N fer-
tilization, at least up to N rates of about 2,200 pounds per acre,
and was increased by K fertilization (Table 13, eq. X, Y and Z).
This correlated with soil P, as discussed above, which was also
reduced by N and increased by K fertilization, and these soil
levels of P no doubt influenced the P uptake by the roses. At the
time of the first sampling (Table 13, eq. X) P fertilization had no
effect on tissue P, but in later samplings (Table 13, eq. Y and Z),
perhaps after residual soil P levels were reduced, the tissue P
reflected the P fertilization. The N, P, and K effects on tissue P
are depicted in Fig. 10 for the second sampling (Table 13, eq.
Y). The 'Happiness' variety had significantly higher tissue P
than the 'Christian Dior'.

Table 13. Regression Equations for the Tissue Analysis Data.

Equation Sampling

U 1 %N = 2.4716 + 0.03433N 0.06102N2 0.0481V

V 2 %N = 1.6930 + 0.04989N

W 3 %N = 1.8413 + 0.04556N 0.04554K 0.0427V

X 1 %P = 0.2909 0.04733N + 0.07161N2 0.02694V

Y 2 %P = 0.2930 0.03133N + 0.07301N2 + 0.03152P + 0.05944K + 0.0158V 0.05725NV

Z 3 %P = 0.4051 0.03209N + 0.07489N2 0.03170P + 0.04101K + 0.0264V 0.04114NV

AA 1 %K = 1.1419 0.0356N + 0.07713N2 + 0.03661K 0.06187K2+ 0.07367NK

+ 0.0191V 0.04328NV 0.02172PV + 0.05909P2V + 0.03210KV 0.07710K2V

BB 2 %K = 0.9726 0.03258N + 0.07502N2 0.03837P + 0.05601P2 + 0.03530K

CC 3 %K = 1.0598 0.03267N + 0.07524N2 + 0.03607K 0.06173K2 0.0153V -0.06141K2

Notes: (a) Units are in pounds per acre of the element.

(b) The superscripts above the zeros indicate the number of zeros following the decimal point,

i. e., 0.03519=0.000519.

(c) V=+I for the 'Christian Dior' variety and -1 for the 'Happiness' variety.

Table 14. Calculated effect of N and K20 fertilization
on 'Christian Dior' leaf tissue (3rd sampling). 1

N K0 %N
300 300 1.80
1500 300 1.85
2700 300 1.95

1000 300 1.84
1000 1500 1.79
1000 2700 1.73

SCalculated from equation W, Table 12, with K20 con-
verted to its elemental equivalent.

There was also an apparent correlation between the effect of
N and K on tissue K (Table 13, eq. AA, BB, and CC) and the
effects of N and K on soil K (Table 9, eq. Q, R, S and T). Nitro-
gen fertilization reduced tissue K up to about 2,200 to 2,700
pounds of N (the exact rate depending on certain NxK or NxV
interactions). Part of this was probably due to reduced levels
of soil K, and part may be attributable to a dilution effect since
N increased stem lengths and perhaps leaf size. K fertilization
naturally increased tissue K, with maximum percentages occur-
ring at K fertilization rates of over 1,600 pounds per acre (2,000
lb. K0z). These effects are shown in Figure 11 for the third sam-
pling (Table 13, eq. CC). 'Christian Dior' tissue had greater K
content in the first sampling, 'Happiness' had more K in the last
sampling, and variety was not a significant factor in the second

Field Ratings
The results of the field ratings are summarized in the regres-
sion equations presented in Table 15. There was a very simple
quadratic relationship for 'Happiness' variety, with N being the
only factor significantly affecting the field ratings (Table 15, eq.
FF and GG). For 1966 and 1967, the amounts of N that would
have minimized the rating values (i.e., have given the best appear-

0 500 1000 1500 2000 2500 3000
N, P205 or K20 (POUNDS PER ACRE)
Figure 10. Calculated effect of N, P205 and K20 fertilization on 'Christian Dior'
rose leaf tissue P (second sampling). When N is the variable, P205
and K20 are set at 115 and 300 Ibs/A, respectively. When P205
is the variable, N and K20 are set at 300 Ibs/A each. When K20 is
the variable, N and P205 are set at 300 and 115 Ibs/A, respectively.

ing plants) were 2,250 and 2,072 pounds per acre, respectively.
These amounts are in the general range mentioned previously as
giving the maximum blossom production and the maximum leaf
N content. The relationship for 'Christian Dior' was more com-
plex. During the second summer, 1966, the rating depended both
on N and K (Table 15, eq. DD). As with 'Happiness', the rating



v 1.20
LL 1.10

S1.00 -

0.80 I I I t
0 500 1000 1500 2000 2500 3000

Figure 11. Calculated effect of N and K20 fertilization on 'Christian Dior' rose
leaf tissue K (third sampling). When N is the variable, KO is set at
300 Ibs/A. When K20 is the variable, N is set at 300 Ibs/A.

Table 15. Field Ratings for 'Christian Dior' and 'Happiness' Varieties.

'C.D.' rating

+ 0.05278NK -

'C.D.' rating

- 0.06249NK +

'Hap.' rating

'Hap.' rating

= 10.1888 0.02786N + 0.05200N2 0.03944K


= 8.3937 0.02191N + 0.06424N2 0.02270K + 0.05124K2

0.02173P 0.04492P2 + 0.04224PK 0.08859PK2

= 9.0832 0.03451N + 0.05100N2

= 8.7337 0.02358N + 0.06865N2

Notes: (a) One is the best possible and 10 is the poorest possible.

(b) The superscripts above the zeros indicate the number of zeros following the decimal

point, i.e., 0.03519=0.000519.








was minimized at a particular N fertilization rate, but this rate
was dependent on the K fertilization. For instance, at a K rate
of 750 pounds per acre, an N rate of 2,166 pounds would have
produced the best appearing plants, whereas at K rates greater
than 1,460 pounds it would have taken N rates greater than 2,700
pounds to produce the best appearing plants. Interestingly, based
on the regression equation DD, Table 15, a K rate of 0 and an N
rate of 1,968 pounds would have produced better appearing plants
than a K rate of 1,460 pounds and an N rate of 2,700 pounds (the
optimum N rate for this K rate), the ratings being 2.46 and 3.68,
respectively. Nevertheless, at any given N fertilization rate, the
appearance of the plants improved linearly as K fertilization in-
creased. In the third year, 1967, P fertilization also affected the
ratings (Table 15, eq. EE), resulting in a rather complex rela-
tionship involving NxK and PxK interactions. The effect of N,
however, was similar to that in 1966, being influenced by K in
much the same way.
Since KCI was used as the K source and high K fertilization
rates were used in some treatments, the possibility of Cl injury
should be considered in interpreting the K data presented above.
However, no visual evidence of salt damage and of Cl toxicity
was observed. The ratings themselves either show no effect of K
or a beneficial effect of K fertilization, and naturally the Cl fer-
tilization paralleled the K fertilization. Chloride determinations
on leaf tissue from the first sampling averaged 0.28% and 0.38%
for bushes fertilized at 300 and 2,700 pounds KO per acre, re-
spectively, when both received no P and N was at 900 pounds
per acre. Both of these Cl values are well under the 1.0% Cl level
which has been suggested as being indicative of Cl injury (3).
Also, lengths increased linearly with increasing K (Cl) fertiliza-
tion. For these reasons, it is concluded that Cl toxicity was not a
factor in this study.
Root Distribution
At the termination of the experiment, an examination was
made of root distribution by both varieties in the soil of represen-
tative treatments. Root distribution was rather uniform in the
top 18 inches of soil in the bed. Apparently there had been no
tendency for roots to concentrate in the vicinity of the fertilizer
bands. This indicated rather effective movement of fertilizer
salts in the soil, primarily by water movement in the irrigation
cycle. There was practically no root penetration into the light
gray sand of the B soil horizon, which was found at a depth of
about 18 inches. It was not determined whether this limitation in

depth of rooting was because of low fertility or because of excess
soil moisture at this depth. No difference between the two vari-
eties was observed in root distribution.

Practical Considerations
There are some practical aspects of these data that deserve
brief consideration. Generally speaking, the number of flowers
produced was dependent upon the inherent vegetative character-
istics of the variety. Normally, the greater the number of shoots,
the more flowers produced. While extra N may have forced de-
velopment to flowering stage of more shoots on either variety by
stimulating vegetative activity, yield of flowers on 'Happiness'
probably could, never, on the average, be made to equal that on
'Christian Dior' with equal spacing, regardless of fertilizer levels
used, because of genetic differences.
From inspection of the yield vs. N fertilization curves (Fig.
12), it is obvious that as N fertilization increases up to that pro-
ducing maximum yields, there is a decreasing gain in yield per
increment of additional monetary return which, at some point,
might not cover the added fertilizer cost. However, with favor-
able fertilizer and flower prices, it appears that generally it
would be economically justifiable to fertilize at an N rate very
near that producing the maximum number of flowers.
The decrease in number of flowers with increasing K fertiliza-
tion perhaps was caused by decreasing N absorption with the
increase in K. Results of analysis (Tables 9 and 13, resp.) show
that N in both soil and leaves decreased as K fertilization in-
creased. This suggests that applying N and K each in separate
bands would reduce N leaching and thereby increase N.uptake.
The increase in stem length with increasing K fertilization
may have resulted indirectly from the reduced N uptake due to
K fertilization. With reduced N absorption, vegetative activity
was probably reduced and fewer shoots developed. Thus, the sup-
ply of elaborated materials and minerals available to individual
developing shoots might be greater than where there were more
shoots developing, and longer growth occurred on the reduced
number of shoots.
The favorable effect of increased K fertilization on stem
length, but adverse effect on number of flowers, raises the ques-
tion as to whether or not it would be economically advantageous
to use the optimum level of K for stem growth. The relationship
between total number of flowers and stem length and K fertilizer
level, with N and P at optimum levels for stem lengths, is shown
for the first year in Figure 12. This shows that increasing K fer-


S70 16
-60- / 15 -

W 4

40- -'HAPPINESS' 13 5

30 1 I 12
0 600 1200 1800 2400 3000

Figure 12. Relationship between total number of flowers, stem length and level
of K20 fertilizer for 'Christian Dior' and 'Happiness' roses for the
first year. Calculated from equation E in Table 2 for number of
flowers and equation E in Table 4 for stem length, with N and P205
at 1818 and 115 Ibs/A, respectively.

tilization from 250 to 1,726 pounds per acre (optimum for stem
length) reduced total yields from 83 to 79 flowers per plant per
year on 'Christian Dior' and from 39 to 36 flowers on 'Happiness',
but increased average stem lengths from about 12.25 to 13.25
inches on 'Christian Dior' and from 14.25 to 15.25 on 'Happiness'.
Increasing K fertilization had no significant effect on the number
of commercial flowers. Evidently increased stem lengths added to
the number of flowers falling into commercial classes sufficiently
to compensate for the decrease in yield caused by increasing K.
For flowers at the borderline between culls and commercial
lengths, increasing stem lengths only slightly, by increased K
fertilization, might move them from cull class, with probably no
value, to a commercial class, which for the past few years have
had an average wholesale price per flower of around 12t or more.
Furthermore, with a favorable market, commercial growers
place a value of 10 on each 1 inch in length of stems 9 inches or
over. The average commercial yield the first year on 'Christian
Dior' was 58.1 and on 'Happiness' 36.8 flowers per plant at all

levels of K fertilization from 250 to 1,726 pounds per acre (with
N at 1818 lbs/A and P at 50 lbs/A). With this increase in K fer-
tilization, average commercial stem lengths on 'Christian Dior'
increased from 14.24 to 14.80 inches and on 'Happiness' from
16.05 to 16.61 inches, or an increase of 0.56 inch on each vari-
ety. Thus, on 'Christian Dior' there was a total commercial stem
length increase of about 32 inches and on 'Happiness' about 21
inches per plant per year. The cost of the extra K fertilizer would
be about 10 per plant per year. It seems that this minor added
expense would be justified in view of the possible returns through
increased stem length alone, regardless of effect on number of
Although there were some slight differences between 'Chris-
tian Dior' and 'Happiness' varieties in optimum levels of N and
KO for both flower production and stem lengths, a single fer-
tilizer mixture with a ratio of N:P05 :KO2 of 22:1:18 (11:1:9 or
10:1:9) would give results reasonably close to optimum on either
variety. If a level nearer optimum of either N or KO was de-
sired for increased flower production or stem lengths on either
variety, the level could be adjusted by supplementing the mixture
with a carrier of N or KO. With plastic mulch, as in this experi-
ment, about 5 tons per planted acre per year would be required
to maximize flower production and stem lengths. Without the
mulch or other protection against leaching, the annual fertilizer
requirement would be substantially greater under Florida con-
For some years, south Florida vegetable growers have been
using plastic mulch to reduce fertilizer leaching losses which are
potentially great due to the high fertilization rates commonly
used, and the low exchange capacity and high infiltration rates
of the sand soils lying along the coastal areas. The region is also
characterized by intense rainfall which contributes to the prob-
lem. The plastic mulch also aids in weed control and fumigation
and reduces damage from blowing sand. Although it would be
expected that the befiefits of this cultural method on vegetables,
short season crops, would be magnified on roses which would be
in the field for more than a year, a test was necessary to estab-
lish the point. Experience gained from this experiment, from the
supplemental experiment (26, 33) in which fertilization under
plastic mulch was compared to bi-weekly fertilization without
mulch, and from a third experiment (27) indicates that rose
production using infrequent, heavy fertilization in conjunction
with black plastic mulch is both feasible and advantageous in
these soils. However, in sand soils which exhibit little capillarity

and therefore little lateral water movement, proper irrigation
must be provided. Naturally, overhead irrigation would not be
sutiable for plastic mulched plots, and furrow irrigation gener-
ally would not be practical in sandy soils both because of the
rapid infiltration rate which may make it difficult to get water to
run the length of a furrow and because of the short lateral water
movement, which makes it difficult to get water very far under
the plastic mulch. In this experiment, a relatively impermeable
spodic horizon occurred at 18 to 24 inches depth, so that a perched
water table was created by furrow irrigation, and lateral water
movement was enhanced. In certain areas of south Florida a
natural water table occurs near the soil surface or a water table
can be so created by suitable ditching and pumping. In areas of
deep sands where it is not feasible to maintain a high water table
by conventional means, a moisture barrier might be installed at
a depth of 18 to 24 inches to facilitate furrow irrigation. Such a
barrier was installed experimentally, using six mil clear plastic,
and was used for several years with no indication of deterioration.
Methods for installing asphalt barriers (32) are being developed
commercially. When the proper barrier conditions are present in
combination with plastic mulch, a weekly irrigation has been
found suitable during dry periods, with less frequent irrigations
being used during times of adequate rainfall. Where finer tex-
tured soils are utilized, the moisture barrier would probably be
unnecessary, although some of the advantages of the plastic
mulch would be less pronounced.

1. Cut Flowers-Production and sales 1962 and 1963. June, 1964. USDA
SpCR 6-1 (64).

2. Bancroft, T. A. 1968. Topics in intermediate statistical methods, Vol. 1,
Iowa State Univ. Press, Ames.

3. Bernstein, L. 1964. Salinity and roses. Amer. Rose Annual.

4. Blackmon, G. H., J. V. Watkins and W. L. Floyd. 1939. Roses in Flori-
da. Fla. Ag. Ext. Serv. Bul. 102.

5. Carlson, W. H., and E. L. Bergman. 1966. Tissue analysis of green-
house roses (Rosa hybrida) and correlation with flower yield. Proc.
Amer. Soc. Hort. Sci. 88:671-677.

6. Culbert, John R., and E. I. Wilde. 1948. The effect of various amounts
of potassium on the production and growth of 'Better Times' roses
under glass. Proc. Amer. Soc. Hort. Sci. 52:528-536.

7. Floyd, W. L., and J. V. Watkins. 1930. Rose growing. Fla. Agr. Ext.
Serv. Bul. 59.

8. 1934. Rose growing in Florida. Fla. Agr.
Ext. Serv. Bul. 78.
9. Garman, Philip, and B. H. Kennedy. 1949. Effect of soil fertilization
on the rate of reproduction of the two-spotted spider mite. J. Econ.
Ent. 42(1) :157.
10. Griffiths, J. T. 1951. Possibilities for better citrus insect control through
the study of the ecological effects of spray programs. J. Econ. Ent.
44(4) :464-468.

11. Horwitz, W. 1970. Official Methods of Analysis of the Association of
Official Analytical Chemists. Wash., D. C.

12. Iley, J. R., and T. W. Young, 1964. Some symptoms of malnutrition in
roses. Proc. Fla. State Hort. Soc. 77:505-510.
13. Jackson, M. L: 1958. Soil Chemical Analysis. Prentice-Hall, Inc. Engle-
wood Cliffs, N. J.
14. Mastalerz, John W., and Robert W. Langhans (editors). 1969. Roses-
a manual on the culture, diseases, insects, economics and breeding
of greenhouse roses. Pennsylvania Flower Growers, N. Y. State
Flower Growers Assoc. Inc., Roses, Inc. 331 pp.
15. McFadden, S. E., Jr. 1956. Comparison of 'Happiness' rose produc-
tion on four rootstocks. Proc. Fla. State Hort. Soc. 69:368-370.
16. 1959. Production of tree roses on Rosa
fortuniana stock, Proc. Fla. State Hort. Soc. 72:413-417.
17. 1962. Rosa fortuniana compared with other
rose rootstocks grown in Florida. Proc. Fla. State Hort. Soc. 75:
18. H. N. Miller, L. C. Kuitert, and A. M. Wil-
son. 1963. Roses in Florida. Fla. Agr. Ext. Serv. Bul. 180.

19. and R. T. Poole. 1967. Copper and other
nutrient requirements of 'Baccara' rose in lightweight potting
medium. Proc. Fla. State Hort. Soc. 80:480-486.
20. Post, Kenneth, and C. W. Fischer, Jr. 1951. The potassium and cal-
cium nutrition of greehouse roses. Proc. Amer. Soc. Hort. Sci.
21. Schoene, W. J. 1941. Plant food and mealybug injury. J. Econ. Entomol.
22. Seeley, John G. 1943. The effects of various nitrate levels on growth
and production of greenhouse roses. Proc. Amer. Soc. Hort. Sci.
23. .1948. Soil nitrate levels for roses. Proc.
Amer. Soc. Hort. Sci. 51:613-617.
24. 1950. Potassium deficiency of greenhouse
roses. Proc. Amer. Soc. Hort. Sci. 56:466-476.
25. Snedecor, G. W. 1956. Statistical Methods, 5th edition. Iowa State
Univ. Press, Ames.
26. Snyder, G. H., T. W. Young, and N. C. Hayslip. 1969. A three year
comparsion of infrequent heavy fertilization under plastic mulch
vs. frequent light applications without mulch for rose production
on Immokalee fine sand. Soil and Crop Sci. Soc. of Fla. Proc.
27. Snyder, G. H., and T. W. Young. 1974. Fertilizer packets, fertilizer
tablets, and a plastic mulch method for flower production of field
grown roses. J. Amer. Soc. Hort. Sci. 99:276-278.
28. Stoutamire, Ralph. 1932. The growing of roses in Florida. Fla. State
Dept. Agr. Bul. 59.
29. Thompson, W. L. 1939. Cultural practices and their influences upon
citrus pests. Jour. Econ. Ent. 32(6)782-789.
30. Waters, W. E. 1967. Influence of nutrition on yield, quality and
chemical composition of 'Tropicana' roses on Rosa fortuniana
rootstock. Proc. Fla. State Hort. Soc. 80:396-400.
31. White, J. W. 1966. Plant analysis for flower crops. Penn. Flower
Grower Bul. 187.( Presented at the 1966 Penn. State Florists
32. Woods, Chuck. 1969. Paving the way to higher yields (Asphalt goes
underground). Sunshine State Agr. Res. Rept. 14(1)7-11.
33.' Young, T. W., J. R. Iley, and N. C. Hayslip. 1966. Response of roses
grown with and without plastic mulch at three nitrogen levels.
Proc. Fla. State Hort. Soc. 79:438-443.
34. Young, T. W., G. H. Snyder, F. G. Martin, and N. C. Hayslip. 1973.
Effects of nitrogen, prosphorus and potassium fertilization of roses
on Oldsmar fine sand. J. Amer. Soc. Hort. Sci. 98(1)109-112.


Smeing Aankin4
1875 1975