International Journal of Agricultural

Science and Research (IJASR)

ISSN(P): 2250-0057; ISSN(E): 2321-0087
Vol. 6, Issue 3, Jun 2016, 295 - 300
TJPRC Pvt. Ltd.

STUDY ON VARIABILITY IN FIELD EXPERIMENTS OF

COTTON CROP (BHAL AND COASTAL ZONE)
G. N. MOTAKA, V. B. DARJI, D. J. PARMAR, A. D. KALOLA & P. R. VAISHNAV
Department of Agricultural Statistics, Anand Agricultural University, Anand, Gujarat, India
ABSTRACT
The data on C.V. % for cotton crop yield along with other details of 224 field experiments conducted during
1989-90 to 2014-15 at Agriculture research station, Anand Agricultural University, Arnej and Dhandhuka center were
collected and analyzed. The frequency distribution tables were prepared for various experimental factors. The upper
fiducial limits (the yardstick) of C.V. % at 95% confidence based on non centralt distribution were worked out for
accepting the results of cotton crop experiments which emerged as 21 %.
KEYWORDS: CV%, Fiducial Limit, Plot Size, Non Central t Distribution, Experimental Variability

Received: Apr 22, 2016; Accepted: May 13, 2016; Published: May 21, 2016; Paper Id.: IJASRJUN2016036

In agricultural field experiments variation occurs due to uncontrolled factors such as soil fertility, climatic
factors etc. and controlled factors such as field layout, treatments, replications, plot size etc. Such factors play an
important role in the precision of the experimental results. C.V. % of the experiment is considered as a reliable
index of variability for accepting or rejecting experimental results. Study on variability of large number of

Original Article

INTODUCTION

experiments help to develop the yardstick of CV % for field experiments. The present investigation was carried
out with the objectives to study the experimental factors viz. disciplines, design, replications, treatments and plot
sizes responsible for variability and to develop yardstick for reliability of the experimental results of cotton
(Bhal and coastal zone) crop experiments.

MATERIALS AND METHODS

The secondary data on C.V. % cotton crop yield of 224 field experiments conducted at Agriculture
research station, Anand Agricultural University, Arnej and Dhandhuka center during 1989-90 to 2014-15 in plant
breeding, agronomy, and soil science disciplines on different cotton (Bhal and coastal zone) crop were utilized for
the variability study. Information on plot size, number of treatments, replications, experimental design and
disciplines was collected for each experiment. The data were analyzed to estimate error mean square and thereby
C.V. % of each experiment (Snedecor and Chocran, 1967). Tyagi et al and Patel et al pointed out that C.V.
obtained for the crops under study was found to be considerably higher than those reported from the uniformity
trials. They stated that the yardstick for accepting experimental results should be worked out using C.V. observed
in the experiments rather than in the uniformity trials. Bajpai and Nigam suggested a working rule for deciding the
value of W2 (weight corresponds to precision of the experiments) and developed an index to evaluate agricultural
field experiments statistically. Gomez and Gomez reported that C.V. varies greatly with the type of experiment,
the crop grown and the character measured. They pointed that the acceptable range of C.V. is 6% to 8% for
www.tjprc.org

editor@tjprc.org

296

G. N. Motaka, V. B. Darji, D.J. Parmar, A. D. Kalola & P. R. Vaishnav

varietal trial, 10 to 12 % for fertilizer trials and 13 to 15 % for insecticidal and herbicidal trials on rice. The upper fiducial
limit of C.V. % at 95% confidence level based on non centralt distribution (Johnson and Welch, 1939) was worked out on
the basis of overall average C.V. % of 224 experiments in the present study, which was considered as yardstick of C.V. %
for field experimentation on cotton (Bhal and coastal zone) crop.

RESULTS AND DISCUSSIONS

Table 1: Upper Fiducial Limit of CV % for Different Disciplines of Cotton (Bhal and Coastal Zone) Crop
Discipline

UL

No. of
Expt.

CV %

55
161
8
224

15.60
16.16
20.48
16.18

Agronomy
Pl. Breeding
Soil Science
Average

Range

(0.05)

(0.10)

(0.05)

(0.10)

20.40
20.51
24.70
20.64

19.16
19.42
23.69
19.51

4.80
4.36
4.22

3.56
3.26
3.21

CV % > 16.18
No. of
Proportion
Expt.
10
0.18
32
0.19
4
0.50

The results presented in Table 1 revealed that mean C.V. % (16.18) and upper fiducial limit (20.64) of average
value of all the discipline except for soil science discipline, were below the mean C.V. %. The experiments on soil science
discipline showed large variation (average CV = 20.48 %) showed poor precision may be because of experimental
requirements such as sample size, natural population of pests and diseases. Use of proper statistical tools may help to
improve the precision of the results. More than 50% experiments of this disciplines had more than 20.64 %.
The results presented in Table 2 indicated that most of the experiments were carried out in RBD and about 20 %
of them had C.V. % higher than the fiducial limit worked out. This proportion was about 30 % in other designs (Except
RBD and SPT)
Table 2: Upper Fiducial Limit of CV % for Different Design Cotton (Bhal and Coastal Zone) Crop
Design
FRBD
RBD
Split plot
Average

No. of
Expt.
10
199
15
224

UL
CV %
18.02
16.19
14.77
16.18

(0.05)
22.06
20.80
17.51
20.64

Range
(0.10)

21.08
19.63
16.86
19.51

(0.05)
4.04
4.61
2.74

(0.10)
3.06
3.44
2.09

CV % > 16.18
No. of
Proportion
Expt.
3
0.30
40
0.20
3
0.20

Influence of number of treatments was also examined and results are given in (Table 3) according to different
treatments group. 11-15, 16-20, 21-25, 26-30 and > 30 treatments in an experiment showed higher C.V. % than the overall
average (16.18 %).
Generally increased number of treatments in the experiments increases blocks (replication) size which increases
error variance, affecting the precision of the results. Therefore, it is advisable to use such experimental designs
(when treatment exceeds 10) which can help in controlling within block variation. Other means such as optimum plot size,
more number of replications, proper site of the experiment etc. need to be considered. About 31to 100 per cent experiments
showed higher C.V. % compared to the overall average C.V. % except the treatment groups < 6, 6-10 and 11-15.

Impact Factor (JCC): 4.7987

NAAS Rating: 3.53

Study on Variability in Field Experiments of Cotton Crop (Bhal and Coastal Zone)

297

Table 3: Upper Fiducial Limit of CV % for Different Treatments of Cotton (Bhal and Coastal Zone) Crop
Treatments
<6
6-10
11-15
16-20
21-25
26-30
>30
Average

UL

No. of
Expt.

CV %

15
93
79
32
2
2
1
224

15.66
14.82
16.28
18.02
16.63
40.58
33.16
16.18

Range

(0.05)

(0.10)

(0.05)

(0.10)

21.17
19.67
20.32
21.71
19.57
48.45
38.99
20.64

19.73
18.41
19.39
20.82
18.89
46.55
37.60
19.51

5.51
4.85
4.04
3.69
2.95
7.87
5.83

4.07
3.59
3.04
2.80
2.27
5.97
4.44

CV % > 16.18
No. of
Proportion
Expt.
2
0.13
17
0.18
14
0.17
10
0.31
0
00
2
1.00
1
1.00

The results presented in (Table 4) indicated that the average C.V.% for different plot size experiments were below
the average C.V.% (16.18) in plot size 6-9, 12-15, 15-18, 18-21, 24-31 and > 31 sq.mt. In 3-6 and 9-12 sq.mt. Plot size the
proportion of C.V. % having higher C.V.% from 0.50 to 0.28 respectively. Plot size of 6-9 and 12 to 21 sq.mt. seems to be
an optimum plot size for cotton crop. Therefore, this needs to be confirmed by plot technique study for different locations.
Table 4: Upper Fiducial Limit of CV % for Different Plot Size of Cotton (Bhal and Coastal Zone) Crop
UL

Plot size
(m2)

No. of
Experiment

CV %

<3
3-6
6-9
9-12
12-15
15-18
18-21
21-24
24-31
>31
Average

14
4
32
28
12
116
0
12
6
224

21.29
14.29
18.55
15.57
15.24
15.68
14.42
10.58
16.18

Range

(0.05)

(0.10)

(0.05)

(0.10)

27.81
18.67
23.34
19.82
19.01
19.95
19.12
14.02
20.64

26.12
17.55
22.15
18.75
18.06
18.87
17.92
13.13
19.51

6.53
4.38
4.79
4.25
3.77
4.27
4.70
3.44

4.83
3.27
3.60
3.19
2.82
3.19
3.50
2.55

CV % > 16.18
No. of
Proportion
Expt.
7
0.50
0
00.0
9
0.28
5
0.17
3
0.25
21
0.18
0
00.0
1
0.08
0
0.00

As far as replications are concerned, the experiments conducted with 3 and 7 replications showed large variation
26 and 25 per cent experiments showed C.V. > 20.64 respectively (Table 5). Analysis showed that majority experiments
were conducted with 3 and 4 replications in field experiments on cotton (Bhal and coastal zone) in which about 26 percent
were having C.V. % > 20.64.
Table 5: Upper Fiducial Limit of CV % for Different Replications of Cotton (Bhal and Coastal Zone) Crop
UL
Replication

No. of Expt.

CV %

3
4
5
6
7
Mean

30
108
74
4
8
224

17.87
15.98
16.01
11.33
16.41
16.18

www.tjprc.org

Range

(0.05)

(0.10)

(0.05)

(0.10)

24.46
20.05
20.11
14.94
21.97
20.64

22.69
19.04
19.09
14.00
20.54
19.51

6.60
4.06
4.10
3.61
5.57

4.83
3.05
3.08
2.67
4.13

CV % > 16.18
No. of
Proportio
Expt.
n
8
0.26
20
0.18
16
0.21
0
00.0
2
0.25

editor@tjprc.org

298

G. N. Motaka, V. B. Darji, D.J. Parmar, A. D. Kalola & P. R. Vaishnav

YARDSTICK OF C.V. % FOR MEDICINAL AND AROMATICCROP EXPERIMENTS

The C.V. % data of 224 field experiments were used to fit non centralt distribution and to work out upper
confidence limit of C.V. at 0.05 level of probability. According the upper fiducial limit of C.V. % at 95 percent confidence
level of C.V. % was worked out to be 20.64 per cent. Thus the results suggested that about 21 percent C.V. % should be
considered as a yard stick for cotton (Bhal and coastal zone) crop field experiments. These having C.V. % > 21 should be
rejected for drawing scientific conclusion.
Table 6: The Average Upper Fiducial Limit and Yardstick for CV % for the
Experiments of Cotton (Bhal and Coastal Zone) Crop
Name of Crop
Cotton (Bhal and
Coastal Zone) Crop

No. of
Experiments

Mean
CV %

224

16.18

Upper Fiducial Limit

of CV %
0.95
0.90
20.64

Overall Yardstick
of CV%

19.51

21

Table 7: Power of F-Test as Influence by CV%

Classes
CV%
1.0-11.0
11.0-21.0
21.0-31.0
31.0-41.0
41.0-51.0
>75
Total

No. of Experiments
48
131
37
7
1
0
224

Significant
34
71
7
3
0
0
115

F-Test
Non-Significant
14
60
30
4
1
0
109

Ratio
0.41
0.85
4.29
1.33
0.95

CONCLUSIONS
The power of F test was examined with the non significant/ significant ratio of experiments (Table 7). The results
revealed that the ratio consistently increased with the increase in C.V. of the experiments. It also indicated that the
efficiency (of detecting difference in treatment means) of F-test decreased with the increase in C.V. of experiments.
The average ratio was observed to be 0.95. The ratio for the class 11.0-21.0 % was almost equal to the average ratio which
included 16.18 %, the mean C.V. of all experiments results clearly showed that when the coefficient of variation in cotton
(Bhal and coastal zone) crop field experiments exceeds 21 %, the experimental finding should not be considered for
scientific purpose.

AKNOWLEDGEMENTS
Research scientist, Agriculture research station, AAU, Arnej center and Assistant Research Scientist, Agricultural
Research Station, Anand Agricultural University, Dhandhuka for providing necessary data.
REFERENCES
1.

Bajpai, S. N. and Nigam, A. K. 1980. Statistical evaluation of Agricultural field experiments. J. Indian Soc. Agric. Stat., 32 (2):
41-45.

2.

Gomez, K. A. and Gomez, A. A. 1984. Statistical procedure for Agricultural workers, John Wiley and Sons, New York, 2nd ed.

3.

Johnson, N. L. and Welch B. L. 1939. Application of noncentral t-distribution. Biometrika, 31: 362-389.

Impact Factor (JCC): 4.7987

NAAS Rating: 3.53

Study on Variability in Field Experiments of Cotton Crop (Bhal and Coastal Zone)
4.

299

Patel, N. M., Prajapati, M.R. and Prajapati, K. H. 1978. Paper presented at XVII Annual conference of Gujarat Statistical
Association, Anand, November, 4-5.

5.

Snedecor, G. W. and Cochran W.G. 1967. Statistical Methods. (Sixth Edition), Oxford and IBH Pub. Co., New Delhi.

6.

Tyagi, B. N., Kathuria, O.P., Sahni, M. L. and Kulharni, G.1973. A.A study of coefficient of variation associated with plots and
blocks of different sizes for some important field crops. J. Indian Soc. Agric. Stat., XXV (II), 37-47.

www.tjprc.org

editor@tjprc.org