|
Instructor: Dr. P. Hildebrand
Student: M. Proenca
Differential Establishment
Characteristics
of
Floriqraze Rhizoma Peanut
Gainesville, September,1987.
3g-A Xo02
1
Introduction
"Florigraze rhizhoma peanut (Arachis glabrata Benth.) is a
warm season perennial forage legume having value as both a hay
and grazing crop" (Prine et al.,1981). Other uses of the plant
are, ornamental, and, as a cover crop. Its centre of origin,
according to Herman, is the central region of Brazil. As a crop,
the major attributes of Florigraze are; 1) high quality forage,
2)high nitrogen fixation, 3)drought resistance, 4) until now has
no pests in Florida, and 5) needs little or no fertilizer. Its
less desirable characteristics, at present, are; 1) time
necessary for, and, difficulty in, obtaining a dense homogenous
stand, 2) non-tolerance of waterlogged field conditions (i.e.
Florida Flatwoods), and, 3) slow canopy cover gives weeds a light
competition advantage. Florigraze prefers well drained soils and
a Ph range between 5.8 and 6.5 ( Prine et al.,1981).
The aim of this study is to determine several differential
establishment characteristics of A. qlabrata in a Gilchrist
County, Florida, field. This plot is approximately located
between Newberry and Trenton, two miles north of Highway 26, near
a Flatwoods area. The field has an area of about ten acres, and
was previously planted to watermelon; in one corner of the field,
an area of two acres, Florigraze has established extremely well,
whilst in the remaining acreage the peanut sparsely, if at all,
populates it.
This study will compare the physical (including texture) and
chemical properties of the soils in the two areas of different
stands; rhizome production will also be analysed. From analysis
of the formentioned characteristics, conclusions regarding
Florigraze establishment will be attempted.
Materials and Methods
A total of 11 locations were studied; of these, 6 were in
the area of better establishment, and the remaining 5 in the
lesser populated area. At each of the 11 chosen locations a 1m2
grid comprised by 25 squares was placed on the ground; 4 soil
samples (20cm deep), 4 penetrometer readings and 5 rhizome
samples (35 cm deep) were taken at each location; the same
relative pattern of sampling was maintained, in relation to
magnetic North.
Soil samples within a location will be mixed, so as to
obtain a sample that is more representative of the soil makeup;
chemical and physical analysis will follow. Rhizome samples will
be washed, weighed, dried, and re-weighed, so as to determine
differences in rhizome production between the two areas.
2
Non-randomized sampling was chosen, in the way of making
differences between both areas more discernible. From the
resulting data it may be possible to arrive at general
conclusions that will help better understand factors that
influence Florigraze establishment.
At the outset it was expected that the penetrometer readings
would show the existence of hardpans, more so in areas of poor
establishment. Should a hardpan be present, with, the amount and
distribution of rainfall in Florida, waterlogged conditions would
result, rendering Florigraze establishment less successful.
Allelopathy was eliminated as a possible causal agent of
differential establishment due to the fact that the whole field
had been previously planted to watermelon. If the field had
previously been planted to a wide variety of crops, then
allelopathic interactions would have been an inherent part of our
working hypothesis.
Rainfall was excluded from analysis, as, all the sites are
contained within a rectangular area of 80x170m2 (see map,
appendix 1). Within such a small distance rain fluctuations
between sights is insignificant.
By subjectively labelling sites M or L, depending on more
(M) or less (L) above ground growth, a correspondence between
rhizome and above ground growth would be possible.
Lab analysis results included; rhizome weight, and, soil,
texture, PH, % Organic Matter, and, concentration of Phosphorus
(P205), Potassium (K20), Magnesium (Mg), Calcium (Ca), Soluble
Salts, Zinc (Zn), Copper (Cu) and Manganese (Mn), in parts per
million (PPM).
Each of these factors was then individually related to
rhyzome weight through simple linear regression; the result is of
the type, Y(rhizome weight)= mX(one factor) + C(constant).
Since it is probable that no one isolated factor is respons-
ible for the differential establishment, it was necessary to
conduct an analysis where, all or some, factors intervened. This
was done using SAS, where a stepwise forward/backward routine was
used. From this analysis, equations of the type, Y= Fl(factor 1)
+........+Fn(factor N)+C(constant), resulted. Each of these
equations represent models, and the stepwise procedure results
in the best model for any number of factors, or variables.
Results
Rhizome weight was determined after drying samples at 700C
for 24 hours. The weights varied between 0.35 and 1.99g.
Significant relation between observed above ground (M or L) was
established.
Weight (q)
1.99
1.91
1.48
1.29
1.18
0.90
0.81
0.52
0.48
0.47
0.35
Yield Yield
(mT/ha) (lbs/acre)
1.96
1.88
1.46
1.27
1.16
0.89
0.80
0.51
0.47
0.46
0.34
1750
1680
1300
1140
1040
0790
0710
0460
0420
0410
0310
Penetrometer Readings are graphed at the end of the report,
under appendix 2; these results show little or no relationship
between the existence of a hardpan and reduced rhizome growth.
Texture classified all samples as sands.
Note; all of the following results are tabled further along
under this tittle. The results of the simple and selected
multiple regressions follow the table.
Soil Ph varied from 6.0 to 6.8, providing a very narrow
range.
% Organic Matter (OM) varied between 0.54 and 0.93, also
providing a narrow range.
Phosphorus (P205) varied between 16 and 72 ppm.
Potassium (K20) ranged from 8 to 36 ppm.
Magnesium (Mg) ranged from 20 to 104 ppm.
Sample
Calcium (Ca) oscillated between 144 and 440 ppm.
Soluble Salts varied between 28 and 42 ppm. In 8 of the
cases the value was 42, in the remaining 3 it was 28 ppm.
Zinc (Zn) ranged between 0.32 and 2.56 ppm.
Copper (Cu) varied from 0.16 and 1.44 ppm.
Manganese (Mn) oscillated from 1.16 to 2.32 ppm
Table of obtained values;
Weight (g)
1.99
1.91
1.48
1.29
1.18
0.90
0.81
0.52
0.48
0.47
0.35
P205(PPm)
48
40
32
40
28
72
32
44
20
36
16
OM (%)
0.73
0.61
0.61
0.61
0.93
0.54
0.67
0.61
0.73
0.73
0.61
Ph
6.3
6.5
6.1
6.8
6.1
6.0
6.5
6.8
6.7
6.5
6.5
_K20 (Pm)
16
36
12
16
12
24
16
16
20
08
08
Mg(DDm)3
32
32
20
104
32
28
44
36
40
32
28
Sol.Sal.(ppm)
42
42
28
42
28
42
42
42
42
28
42
Ca(Dpm)
272
240
144
440
228
248
336
324
320
296
248
Sample
M6
M4
M5
M2
M3
Ll
Ml
L4
L3
L5
L2
Sample
M6
M4
M5
M2
M3
Ll
Ml
L4
L3
L5
L2
5
Sample Zn(Dpm) Cu(pPm) Mn(pPm)
M6 0.32 0.32 1.80
M4 2.56 0.36 1.88
M5 0.92 0.20 1.44
M2 0.72 0.60 2.32
M3 1.28 0.24 1.68
L1 2.20 1.44 1.96
Ml 0.76 0.68 1.88
L4 0.76 0.48 1.96
L3 0.40 0.16 1.92
L5 0.44 0.68 1.96
L2 0.80 0.20 1.16
Simple linear regression revealed the following Equations
and Coefficients of Correlation (R2), when each of the variables
was plotted against rhizome weight (RW). Each of the resulting
graphs is in App. 3.
OM vs RW; Y11= 0.60 +0.620M; R2=0.02
Ph vs RW; Y11= 2.87 -0.197Ph; R2=8.41 x10-4
Sol.Sal vs RW; Y11= 1.07 -8.6xl0-5SS; R2= 9.4x 10-5
P205 vs RW; Y11= 0.58 +0.012P; R2=0.10
K20 vs RW;Y11 0.46 +0.034K; R2=0.22
Mg vs RW; Y11= 1.52 -0.016Mg; R2=0.090
Ca vs RW; Y11= 1.57 -0.002Ca; R2=0.062
Zn vs RW; Y11= 0.385 +0.49Zn; R2=0.47
Cu vs RW; Y11= 1.14 -0.22Cu; R2=0.02
Mn vs RW; Y11= 0.72 +0.174Mn; R2=0.0085
6
Multiple Linear Regression, using SAS, gave 2 important
models, one for 3, and the other for 2, variables. These models
were:
2 variable; Y11= 0.113 +0.049P -1.853Cu; R2=0.566
3 variable; Y11= 0.061 +0.045P +0.230Zn -1.887Cu; R2=0.631
The analysis using SAS is included in App.4 1.
Discussion
Factors that had little or no influence on the establishment
of florigraze, in this study, were; soil texture, existence of
hardpans, Ph, OM, soluble salts and manganese.
Rhizome weight varied considerably, from 0.34 to 1.96mT/ha
(9310-17501bs/acre, respectively).
As to what regards simple linear regression (see app. 3),
the equation with highest R2 was Zn (0.47), followed by K20
(0.22) and P205 (0.1). All the other variable models explained
less than 10%.
Considering the multiple linear regression, using SAS (app.
4), in the first part, Forward Selection, only P, K and Cu met
the 0.500 significance level; Backward Elimination, the second
part, eliminated all the variables exept P and Cu, considered
relevant at the 0.100 significance level. In the last part, Max-
R2, usually the most important, two models are outstanding; these
are, the best, 2, and, 3, variable models found, with, P and Cu,
and P, Cu and Zn, respectively. These 2 models can be seen on
pages 17 and 18 of appendix 4.
Conclusion
There appears to be a positive relation between the above
ground growth and rhizome weight; in this case the presence of
hardpans did not significantly affect rhizome production; this
may be due to the fact that the texture of the top 20cm of soil
was, in all samples, sand. Soil Ph in the 6.0-6.8 range does not
1 Three variables were excluded from the multiple
linear regression ; 1) Ph for having such a narrow range, 2)
Solution Salts for having such a small R2 value under simple
linear regression analysis, and 3) OM for having such a low
% and range.
7
show significant correlation to rhizome production (the
recommended Ph range is 5.8 to 6.5). OM, Ph, soluble salts and
Mn, have, in their respective ranges present in this field,
little or no effect on Florigraze establishment.
This field appears to have been limed with Dolomitic
limestone, as there is a strong positive correlation between Ca
and Mg.
During the course of the statistical analysis, the nutrients
that appeared to be most critical, in the sampled concentration
ranges, are, Zn, K, P, Mg, Ca and Cu. Further immediate
research, should, in my opinion, study the effects of, soil
texture, hardpans, weed competition, P, K, Cu, OM and Ph, on
Florigraze establishment.
Appendix 1
Field Map
*M4
*M3
"M5
-M2
*M1
.tree
-M6
.L5
- L4
.L2
Scale 1:1 000
Appendix 2
Penetrometer Results; Graphs
70
60
50
40
30
20
10
0
70
60
50
40
30
20
.-10
O
0
70
60
50
40
30
20
10
0
70
60
50
40
30
20
10
M2 0
Small Pt. kgs/cm2
10 20 30 40 50 6
Small Pt. kgs/cm2
I I _
/ \ '\ '
I
10 20 30 40 50 6
Small Pt.kgs/cm2
70
60
50
40
30
20
10
LI 0
Small Pt. kgs/cm2
-- -- I
_____I
___ _I
____________________________________________________ ________________________
10 20
40 50 60
Small Pt. kgs/cm2
10 20
50 60
Small Pt.kgs/cm2
=^^S^=
/___ I _II_
10 20 30 40 50
10 20 30 40 50 60
70
60
50
40
30
20
10
M3 o
- -- ~
_ ii
________________________________________________ ________________________________________________ ________________________________________________ I
/
Small Pt.kgs/cm2
I p
70
60
50
40
30
20
10
0
10 20 30 40 50 60
Small Pt. kgs/cm2
H' i 3
10 20 50 40 50 6_
Small Pt. kgs/cm2
70
60
50
40
30
20
10
0
Small Pt. kqs/cm2
70
60
50
40
30
20
IO
10
L5 0
70
60
50
40
30
20
10
0
Appendix 3
Simple Linear Regression
Organic Matter vs Rhizome Weight
Yll=o.60+0.62o.m. : R2=0.02
2.00,
1.85.
1.70_
1.55-
1.40
1,25
1.10-
0.95
0.80
(g)
0.65
0.50
0.35-
1M6
M4
,M5
M2
y
M13
YL1
. L4
'L5
L2
ORGANIC
0.70
MATTER (%)
0
0.
50
0.60
0.b0
0. )
--
Ph vs rhizome weight
Y =2.87 0.197Ph
11
R2=8.41x104
M6 '
* M4
Y1
SM5
M2
2.00-
1.85-
1.70-
1.55-
1.40-
1.25
1.10-
0.95 -
0.80-
0.65-
0.50-
0.35-
' LI
. L5
* L3
*L2
___________, -- - - -- - - -- -- -- - -
6.0
6.1
6.1
6.2
6.3
6.4
6.5
6.6
15
5.8
6.7
P h
* M3
Phosphorus vs Rhizome Weight
2
Y = -0.128+0.035P :R2=0.34
10
2
Y =0.58+0.012P :R =0.10
M6
M4
Y10
R
II
I
Z
0
M .
E
W
E
I
G
H
T
(g)
2.00 -
1.85 -
1.70
1.55
1.40
1.25
1.10
0.95
0.80 -
0.65
0,50
0,35
.L4
.L3 L5L4
,L2
16 3 5 7
P205 (ppm)
M3 .
. M5
- ", "--i "~
2.000
1.85
1.70
1.55
1.40
1.25-
1.10-
0.95-
0.80-
0.65-
0.50-
0.35-
0 28
SOLUBLE SALTS
. L5
. M6
, M2
SY11
. LI
* L3
SL2
' '2' - = _" C -,-e -
,L "
: -~.n;rO-'
(ppm)
Potassium vs Rhizome Weight
YL=0.46+0.034K : R2=0.22
2.00-
1.85-
1.70-
1.55-
1.40-
1.25-
(;-)
1.10-
0.95-
0.80-
0.65-
0.50-
0.35
0
M6.
114
M5 .
M2
M3 .
L1.
Ml "
Y11
I4 .
L5
L3.
L2.
,_- -- -i
K20 (ppm)
Magnesium vs Rhizome Weight
Y =1.52-0.016Mg :R2=0.090
2.00 *M6
.M4
1.85 -
R
S 1.70 -
I
1.55
'0 M5
1.40
E
M2
1.25 -
W I .M3
S 1.10
I
G 0.95
H L
T 0.80 m
(g)
0.65 -
0.50- L5 L4 3
0.35- L2' 11
0 I I F---
20 40 60 80 100
Mg (ppm)
Calcium vs Rhizome Weipht
Y11=1.57-0.002Ca : R2=0.062
M6
rIM4.
M5
M3
Y
L1
id'
ul-.
L5. 1J3.
L2 .
20
240
Ca(ppm)
340
2.00-
1.85-
1.70-
S1.55-
1,c-
S1.40-
1.25-
. 1.10-
(g)
0.95-
0.80-
0.65-
0.50-
0.35-
O
140
S/ ()O
Zinc vs Rhizome Weight
2
Y9= 0.32+0.65Zn : R =0.61
2
S10=0.45+0.45Zn :R =0.42
2
Yll=0.385+0.49Zn :R =0.47
2.00_ -M6 /9
M4
1.85
R
il 1.70
I
1.55
S. M5
1.40
1.25 *M2
M3
J; 1.10
J
G, 0.95
i* L1
0.80 M11
(Cg)
0.65
0.50 L3 .L4
'L5
0.35 .L2
0 1
0.30 1.10 1.90 2.70
Zn (ppm)
2.00
1.85
1,70
1.55-
1.40 -
1.25 -
1.10 -
0.95 -
0.80 -
0.65 -
0.50-
0.35 -
0.15
1.55
0.85
Cu (ppm)
Copper vs Rhizome Weight
2
Y =1.14-0.22Cu :R =0.02
11
M6
M4
SM5
M2
M3
SL1
Y11
.L3 .L4 L5
L2
L2
e
2.00-
1.85-
1.70-
1.55-
1.40-
1.25-
1.10-
0.95-
0.80-
0.65-
0.50-
0.35-
0
1.35
1.55
1.75
1.95
2.15
2.35
Mn (ppm)
Manganese ;vs Rhizome Weight
2
Y =0.72+0.174Mn :R =0.0085
11
M6 .
.M4
M5'
M2 .
M3
LY
M1.
L4
L3. L5
L5
2
'L
1.15
Appendix 4
Multiple Linear Regression
SAS
Forward Selection Procedure for Dependent Variable WEIGHT
Stec 1 Variable K Entered
R-square = 0.22179172
Sum of Squares
Regression
Error
Total
Variable
INTERCEP
Parameter
Estimate
0.45839450
0.03444381
0.75237800
2.63989472
3.39227273
Standard
Error
0.39506886
0.02150627
Mean Square
0.75237800
0.29332164
Type II
Sum of Squares
0.39489105
0.75237800
C(p) = 3.91616268
F Prob>F
2.57 0.1437
F Prob>F
1.35 0.2758
2.57 0.1437
Bounds on condition number:
Step 2 Variable CU Entered
1.0000,
1.0000
R-square = 0.28639614
Sum of Squares
Regression
Error
Total
Variable
INTERCEP
K
CU
Parameter
Estimate
0.58620673
0.03875314
-0.41023361
0.97153383
2.42073890
3.39227273
Standard
Error
0.42844814
0.02242273
0.48204070
Mean Square
0.48576692
0.30259236
Type II
Sum of Squares
0.56645189
0.90384717
0.21915583
C(p) = 5.00993689
F Prob>F
1.61 0.2593
F Prob>F
1.87
2.99
0.72
0.2084
0.1222
0.4195
Bounds on condition number:
Step 3
Variable P Entered
R-square = 0.61873943
Sum of Squares
Regression
Error
Total
Parameter
Estimate
2.09893291
1.29333982
3.39227273
Standard
Error
Mean Square
0.69964430
0.18476283
Type II
Sum of Squares
C(p) = 2.34805715
F Prob>F
3.79 0.0667
1.0537,
4.2150
F Prob>F
Variable
SAS
INTERCEP
p
-0.01277585
0.04220428
0.01895545
-1.71392095
0.41338220
0.01708539
0.01926735
0.64839690
0.00017648
1.12739908
0.17882937
1.29096456
Bounds on condition number: 3.5942, 23.9724
No other variables met the 0.5000 significance level for entry
into the model.
0.00
6.10
0.97
6.99
0.9762
0.0428
0.3580
0.0333
SAS
Summary of Forward Selection Procedure for Dependent Variable WEIGHT
Variable
Entered
Number Partial
In R**2
Model
R**2
c(p)
F Prob>F
1 0.2218
2 0.0646
3 0.3323
Step
0.2218
0.2864
0.6187
3.9162
5.0099
2.3481
2.5650
0.7243
6.1019
0.1437
0.4195
0.0428
SAS
Backward Elimination Procedure for Dependent Variable WEIGHT
Step 0
All Variables Entered
R-square = 0.78613136
Sum of Squares
Regression
Error
Total
Variable
INTERCEP
P
K
ZN
CU
CA
MG
MN
2.66677198
0.72550075
3.39227273
Parameter
Estimate
1.21778179
0.03302447
0.03381020
-0.21387934
-1.15640713
-0.00739374
0.02261940
0.00424156
Standard
Error
1.57707191
0.02424273
0.04835383
0.58274200
1.08180619
0.00659108
0.01595667
1.19288740
Mean Square
0.38096743
0.24183358
Type II
Sum of Squares
0.14419579
0.44877154
0.11823611
0.03257629
0.27633711
0.30432058
0.48595269
0.00000306
C(p) = 8.00000000
F Prob>F
1.58 0.3841
F Prob>F
0.60
1.86
0.49
0.13
1.14
1.26
2.01
0.00
0.4963
0.2664
0.5347
0.7380
0.3635
0.3436
0.2513
0.9974
Bounds on condition number:
Step 1 Variable MN Removed
10.3142,
329.4
R-square = 0.78613046
Sum of Squares
Regression
Error
Total
Variable
INTERCEP
P
K
ZN
CU
CA
MG
Parameter
Estimate
1.22083657
0.03305489
0.03387594
-0.21433819
-1.15611831
-0.00738561
0.02263055
2.66676892
0.72550380
3.39227273
Standard
Error
1.14534183
0.01964406
0.03869290
0.49214032
0.93422896
0.00535330
0.01354943
Mean Square
0.44446149
0.18137595
Type II
Sum of Squares
0.20607463
0.51355700
0.13902723
0.03440337
0.27776506
0.34523058
0.50597356
C(p) = 6.00001264
F Prob>F
2.45 0.2025
F Prob>F
1.14
2.83
0.77
0.19
1.53
1.90
2.79
0.3465
0.1677
0.4307
0.6857
0.2836
0.2398
0.1702
Bounds on condition number:
9.0720, 227.9
SAS
Step 2
Variable ZN Removed
R-square = 0.77598877
Sum of Squares
Regression
Error
Total
Variable
INTERCEP
P
K
CU
CA
MG
Parameter
Estimate
0.87116579
0.03685361
0.01922476
-1.43130483
-0.00564311
0.01954768
2.63236556
0.75990717
3.39227273
Standard
Error
0.74770891
0.01611220
0.01749949
0.62992070
0.00325579
0.01057585
Mean Square
0.52647311
0.15198143
Type II
Sum of Squares
0.20631322
0.79513564
0.18342646
0.78466278
0.45657844
0.51921860
C(p) = 4.14227314
F Prob>F
3.46 0.0995
F Prob>F
1.36
5.23
1.21
5.16
3.00
3.42
0.2965
0.0709
0.3220
0.0722
0.1436
0.1238
Bounds on condition number:
Step 3 Variable K Removed
4.0046,
82.3697
R-square = 0.72191692
Sum of Squares
Regression
Error
Total
Variable
INTERCEP
P
CU
CA
MG
Parameter
Estimate
0.96867437
0.04414494
-1.58383071
-0.00554457
0.01955320
2.44893909
0.94333363
3.39227273
Standard
Error
0.75511429
0.01493272
0.62493397
0.00331019
0.01075665
Mean Square
0.61223477
0.15722227
Type II
Sum of Squares
0.25872852
1.37403496
1.00986500
0.44110704
0.51951195
C(p) = 2.90075533
F Prob>F
3.89 0.0681
F Prob>F
1.65
8.74
6.42
2.81
3.30
0.2469
0.0254
0.0444
0.1450
0.1190
Bounds on condition number:
Step 4
Variable CA Removed
R-square = 0.59188403
Sum of Squares
Regression
Error
Total
2.00783205
1.38444067
3.39227273
Mean Square
0.66927735
0.19777724
C(p) = 2.72476602
F Prob>F
3.38 0.0834
4.0016,
57.4387
SAS
Variable
INTERCEPT
P
CU
MG
Parameter
Estimate
-0.06574469
0.05051373
-1.92383303
0.00421807
Bounds on condition number:
Step 5 Variable MG Removed
Step 5 Variable MG Removed
Standard
Error
0.48735687
0.01619625
0.66291013
0.00633332
3.0490,
Type II
Sum of Squares
0.00359918
1.92382852
1.66571907
0.08772852
F Prob>F
0.02
9.73
8.42
0.44
0.8965
0.0169
0.0229
0.5267
21.2768
R-square = 0.56602275
Sum of Squares
Regression
Error
Total
Variable
INTERCEP
P
CU
Parameter
Estimate
0.11311872
0.04919623
-1.85380026
1.92010354
1.47216919
3.39227273
Standard
Error
0.39228725
0.01550588
0.63134508
Mean Square
0.96005177
0.18402115
Type II
Sum of Squares
0.01530130
1.85241687
1.58657282
C(p) = 1.08753001
F Prob>F
5.22 0.0355
F Prob>F
0.08
10.07
8.62
0.7804
0.0131
0.0188
Bounds on condition number:
2.9723,
11.8891
All variables in the model are significant at the 0.1000 level.
SAS
Summary of Backward Elimination Procedure for Dependent Variable WEIGHT
Number Partial
In R**2
6 0.0000
5 0.0101
4 0.0541
3 0.1300
2 0.0259
Model
R**2
0.7861
0.7760
0.7219
0.5919
0.5660
C(p)
6.0000
4.1423
2.9008
2.7248
1.0875
0.0000
0.1897
1.2069
2.8056
0.4436
F Prcb>F
0.9974
0.6857
0.3220
0.1450
0.5267
Stec
Variable
Removed
SAS
Maximum R-square Improvement for Dependent Variable WEIGHT
Step 1 Variable K Entered
R-square = 0.22179172
Sun of Squares
Regression
Error
Total
Variable
INTERCEP
K
Parameter
Estimate
0.45839450
0.03444381
0.75237800
2.63989472
3.39227273
Standard
Error
0.39506886
0.02150627
Mean Square
0.75237800
0.29332164
Type II
Sum of Squares
0.39489105
0.75237800
C(p) = 3.9161626S
F Prob>F
2.57 0.1437
F Prob>F
1.35 0.2758
2.57 0.1437
Bounds on condition number:
1.0000,
1.0000
The above model is the best 1 variables model found.
Step 2 Variable CU Entered
R-square = 0.28639614
Sum of Squares
Regression
Error
Total
Variable
INTERCEP
K
CU
Parameter
Estimate
0.58620673
0.03875314
-0.41023361
0.97153383
2.42073890
3.39227273
Standard
Error
0.428448i4
0.02242273
0.48204070
Mean Square
0.48576692
0.30259236
Type II
Sum of Squares
-0.56645189
0.90384717
0.21915583
C(p) = 5.00993689
F Prob>F
1.61 0.2593
F Prob>F
1.87
2.99
0.72
0.2084
0.1222
0.4195
Bounds on condition number:
Step 3 Variable K Removed
Variable P Entered
1.0537,
4.2150
R-square = 0.56602275
Sum of Squares
Regression 2
Error 8
Total 10
1.92010354
1.47216919
3.39227273
Mean Square
0.96005177
0.18402115
C(p) = 1.08753001
F Prob>F
5.22 0.0355
SAS
Variable
INTERCEP
Parameter
Estimate
0.11311872
0.04919623
-1.85380026
Standard
Error
0.39228725
0.01550588
0.63134508
Type II
Sum of Squares
0.01530130
1.85241687
1.58657282
F Prob>F
0.08
10.07
8.62
0.7804
0.0131
0.0188
Bounds on condition number:
2.9723,
11.8891
The above model is the best 2 variables model found.
Step 4 Variable ZN Entered
R-square = 0.63128687
Sum of Squares
Regression
Error
Total
Variable
INTERCEP
P
ZN
CU
Parameter
Estimate
0.06134074
0.04474381
0.22998727
-1.88747747
2.14149724
1.25077549
3.39227273
Standard
Error
0.38934342
0.01579419
0.20661499
0.62285442
Mean Square
0.71383241
0.17868221
Type II
Sum of Squares
0.00443520
1.43401170
0.22139370
1.64086090
C(p) = 2.17205044
F Prob>F
3.99 0.0598
F Prob>F
0.02
8.03
1.24
9.18
0.8793
0.0253
0.3024
0.0191
Bounds on condition number:
3.1760,
22.2830
The above model is the best 3 variables model found.
Step 5 Variable MG Entered
R-square = 0.68425513
Sum of Squares
Regression
Error
Total
Variable
INTERCEP
P
ZN
CU
Parameter
Estimate
-0.21436925
0.04567834
0.28214939
-1.99846993
2.32118000
1.07109273
3.39227273
Standard
Error
0.47641262
0.01581427
0.21296273
0.63231706
Mean Square
0.58029500
0.17851545
Type II
Sum of Squares
0.03614384
1.48935438
0.31334795
1.78320492
C(p) = 3.42904876
F Prob>F
3.25 0.0961
F Prob>F
0.20
8.34
1.76
9.99
0.6685
0.0278
0.2334
0.0196
SAS
0.00622503
0.00620477
0.17968276
1.01 0.3545
Bounds on condition number:
Step 6 Variable ZN Removed
Variable CA Entered
3.187C,
34.8179
R-square = 0.72191692
Sum of Squares
Regression
Error
Total
Variable
INTERCEPT
P
CU
CA
MG
Parameter
Estimate
0.96867437
0.04414494
-1.58383071
-0.00554457
0.01955320
2.44893909
0.94333363
3.39227273
Standard
Error
0.75511429
0.01493272
0.62493397
0.00331019
0.01075665
Mean Square
0.61223477
0.15722227
Type II
Sum of Squares
0.25872852
1.37403496
1.00986500
0.44110704
0.51951195
C(p) = 2.90075533
F Prob>F
3.89 0.0681
F Prob>F
1.65
8.74
6.42
2.81
3.30
0.2469
0.0254
0.0444
0.1450
0.1190
Bounds on condition number:
4.0016,
57.4387
The above model is the best 4 variables model found.
Step 7 Variable K Entered
Regression
Error
Total
Variable
INTERCEP
P
K
CU
CA
MG
5 -
5
10
R-square = 0.77598877
Sum of Squares
-2.63236556
0.75990717
3.39227273
Parameter
Estimate
0.87116579
0.03685361
0.01922476
-1.43130483
-0.00564311
0.01954768
Standard
Error
0.74770891
0.01611220
0.01749949
0.62992070
0.00325579
0.01057585
Mean Square
0.52647311
0.15198143
Type II
Sum of Squares
0.20631322
0.79513564
0.18342646
0.78466278
0.45657844
0.51921860
C(p) = 4.14227314
F Prob>F
3.46 0.0995
F Prob>F
1.36
5.23
1.21
5.16
3.00
3.42
0.2965
0.0709
0.3220
0.0722
0.1436
0.1238
Bounds on condition number:
4.0046, 82.3697
SAS
The above model is the best 5 variables model found.
Step 8 Variable ZN Entered
R-square = 0.78613046
Sum of Squares
Regression
Error
Total
Variable
INTERCEP
P
K
ZN
CU
CA
MG
Parameter
Estimate
1.22083657
0.03305489
0.03387594
-0.21433819
-1.15611831
-0.00738561
0.02263055
2.66676892
0.72550380
3.39227273
Standard
Error
1.14534183
0.01964406
0.03869290
0.49214032
0.93422896
0.00535330
0.01354943
Mean Square
0.44446149
0.18137595
Type II
Sum of Squares
0.20607463
0.51355700
0.13902723
0.03440337
0.27776506
0.34523058
0.50597356
C(p) = 6.00001264
F Prob>F
2.45 0.2025
F Prob>F
1.14
2.83
0.77
0.19
1.53
1.90
2.79
0.3465
0.1677
0.4307
0.6857
0.2836
0.2398
0.1702
Bounds on condition number:
9.0720,
227.9
The above model is the best 6 variables model found.
Step 9
Variable MN Entered
R-square = 0.78613136
Sum of Squares
Regression
Error
Total
Variable
INTERCEP
P
K
ZN
CU
CA
MG
MN
Parameter
Estimate
1.21778179
0.03302447
0.03381020
-0.21387934
-1.15640713
-0.00739374
0.02261940
0.00424156
2.66677198
0.72550075
3.39227273
Standard
Error
1.57707191
0.02424273
0.04835383
0.58274200
1.08180619
0.00659108
0.01595667
1.19288740
Mean Square
0.38096743
0.24183358
Type II
Sum of Squares
0.14419579
0.44877154
0.11823611
0.03257629
0.27633711
0.30432058
0.48595269
0.00000306
C(p) = 8.00000000
F Prob>F
1.58 0.3841
F Prob>F
0.60
1.86
0.49
0.13
1.14
1.26
2.01
0.00
0.4963
0.2664
0.5347
0.7380
0.3635
0.3436
0.2513
0.9974
Bounds on condition number:
10.3142, 329.4
SAS
'Ie .above model is the best 7 variables model found.
No further improvement in R-square is possible.
|
RSS
HIDE
