230 S.D. Koutroubas et al. European Journal of Agronomy 11 1999 227–237
of the first raceme and the others and the thousand and r
=0.899, p0.01 at Loudias but not in the other years. The maturity of the first raceme was
seed weight were calculated on a dry-seed basis. The oil content was measured by the Soxhlet
completed from the beginning to the middle of August in all genotypes in all three years 49–
method American Oil Chemists’ Society, 1983 in 1995 and 1996 and by nuclear magnetic resonance
62 days after the beginning of the flowering on the University farm and 48–64 days at Loudias, fol-
NMR spectrometry in 1997. NMR estimates of oil
content were
calibrated against
Soxhlet lowed by a period of about 1.5 months with
favourable environmental conditions for the matu- estimates.
A statistical analysis was performed according rity of secondary racemes. On the University farm,
122–137 days were needed for the maturity of the to Steel and Torrie 1980. The homogeneity of
the variances was checked, and all measured and reproductive racemes and at Loudias more time,
121–151 days, because the mean over genotypes derived data were subjected to analysis of variance
combined over locations separately for each year. number of secondary racemes was higher. The
corresponding temperature sums base temper- A combined analysis of variance over locations
and years was also performed for the data concern- ature 10
°C varied from 1421 to 1792 d°C on the University farm and 1377 to 1895 d
°C at Loudias, ing the genotypes used for more than one year.
LSD values were calculated and used to compare depending mainly on the genotype rather than on
the years of experimentation. treatment means. The relative contribution of the
seed yield and the seed oil concentration to the sum of squares of the oil yield was determined by
linearizing the multiplicative relationships by 3.2. Morphological characteristics
taking logs according to the method of Moll et al. 1982. According to this analysis, the sum of
The plant height and the height of insertion of the first raceme differed between locations and
cross products of each component trait by the resultant trait
∑x i
y i
divided by the sums of among genotypes in all years. The locations
× genotypes interaction was significant in 1995 and
squares of the resultant trait ∑y
2 i
gives the relative contribution of each component variable to the
1996 for the plant height and only in 1997 for the height of insertion of the first raceme Table 2.
resultant variable. The over genotypes mean values of both character-
istics were higher at Loudias compared to those on the University farm, but the differences between
3. Results
the locations were greater for the plant height than for the height of insertion of the first raceme. The
3.1. Stages of development plant height on the University farm varied from
79 cm Negus in 1995 to 230 cm H530 in 1996 Seedlings emerged after 19–26 days in 1995, 18–
26 days in 1997 and 11–14 days in 1996. The and at Loudias from 117 cm HD912 in 1997 to
278 cm Polluce in 1995, and the height of inser- flowering of the first raceme began 2–19 days
earlier at Loudias compared to the University tion of the first raceme from 23 cm HD912 in
1997 to 111 cm Castore in 1997 and from 34 cm farm, depending on the genotype and the year
Table 1. The time from sowing to the beginning Riscio in 1995 to 128 cm Castore in 1997, on
the University farm and Loudias, respectively. The of female flowering on the first raceme varied from
52 days Pronto in 1996 to 73 days B9 in 1997 plant height of most genotypes was less than
150 cm on the University farm and more than on the University farm and 42 days 125 in 1996
to 68 days B9 in 1995 at Loudias. The time of 150 cm at Loudias, and the height of insertion of
the first raceme less than 70 cm in both locations. the beginning of the female flowering on the first
raceme was positively correlated with the time of Both the plant height and the height of insertion
of the first raceme were affected by the year, but maturity of the first raceme in both locations in
1997 r =0.704, p0.05 on the University farm
the variation of the height of insertion of the first
231 S.D. Koutroubas et al. European Journal of Agronomy 11 1999 227–237
Table 2 Height of insertion of the first raceme, plant height and number of secondary racemes of 19 castor plant genotypes
a Genotype
Height of insertion of the first raceme cm Plant height cm Number of secondary racemes
1995 1996
1997 LSD
b 1995 1996
1997 LSD
b 1995 1996
1997 LSD
b UF
L UF
L UF L
UF L UF L
UF L UF L
UF L UF L
Negus 26
36 25
38 26
37 6
79 176 106 182 93 125 16
2.3 4.1 1.9 3.8
3.3 3.3 0.6 Pronto
38 51
31 50
24 44
115 206 132 171 80 131
3.5 4.1 2.2 3.1
3.4 2.0 H530
102 104
55 101
9 230 245 176 222 23
1.7 2.1 2.2 2.6 0.5
H526 47
63 34
44 140 185
85 142 2.1 2.8
2.2 2.6 HD912
26 42
23 38
95 153 89 117
2.0 3.9 2.3 2.5
Riscio 34
34 38
42 12
96 190 125 218 38
2.3 6.2 1.3 4.3
1.7 B9
49 56
42 54
8 104 150
89 150 22 0.3 1.4
0.3 1.9 0.4 H523
47 58
137 255 3.2 3.6
Venda 79
89 159 256
1.6 2.7 Polluce
66 76
177 278 3.9 4.0
114 32
36 83 174
2.3 3.5 929
68 65
163 223 4.4 6.3
H101 33
38 119 187
1.5 2.1 125
27 36
119 164 2.6 5.4
519 33
42 118 146
1.8 5.2 Castore
111 128
229 270 2.0 2.8
H531 46
74 137 178
1.7 2.3 H529
40 60
101 149 2.0 2.3
H913 67
107 171 207
1.3 1.7 LSD
c 12
10 9
21 23
22 1.0
0.7 0.5
Mean 48
56 47
58 40
62 124 212 141 192 113 158
2.6 4.0 1.9 3.6
2.1 2.2 Source of variation
d Location L
NS Genotype G
L ×G
NS NS
NS CV
16 14
12 9
10 11
22 19
15 a UF: University farm; L: Loudias.
b Between genotypes common in the two or three years at p0.05. c Between genotypes and locations for the same year at p0.05.
d NS, p0.05; p0.05; p0.01.
raceme was higher compared to the variation of The highest number of secondary racemes was
observed in 929 4.4 on the University farm and plant height.
All genotypes produced additional secondary 6.3 at Loudias and the lowest in B9 0.3 on the
University farm and 1.4 at Loudias. racemes, and the over genotypes mean number
was higher at Loudias compared to the University farm in 1995 and 1996 Table 2. In 1997, Pronto
3.3. Yield and yield components and Negus produced significantly higher number
of secondary racemes on the University farm and The genotypes differed in total yield Table 3.
The highest total yield was obtained with H531 in B9 and H531 at Loudias, whereas the other geno-
types produced the same in both locations. This 1997 4.35 Mg ha−1, and the lowest total yield
was with Negus in 1996 2.46 Mg ha−1 on the resulted in the same over genotypes mean number
of secondary racemes in both locations in 1997. University farm. At Loudias, the highest total
232 S.D. Koutroubas et al. European Journal of Agronomy 11 1999 227–237
Table 3 Yield of the first and secondary racemes and total yield of 19 castor plant genotypes
a
Genotype Yield of the first raceme Mg ha−1
Yield of the secondary racemes Mg ha−1 Total yield Mg ha−1
1995 1996
1997 LSD
b 1995 1996
1997 LSD
b 1995 1996
1997 LSD
b UF
L UF
L UF
L UF
L UF
L UF
L UF
L UF
L UF
L Negus
1.56 1.26 1.64 1.86
1.63 1.45 0.19 1.69 2.34
0.82 2.32 2.02 2.13 0.33
3.25 3.60 2.46 4.18 3.65 3.58 0.31
Pronto 1.97 2.08
2.22 1.70 1.60 1.91
1.26 2.54 1.13 2.88
2.32 1.45 3.23 4.62
3.35 4.58 3.92 3.36 H530
1.77 2.33 2.27 1.92 0.35
1.20 2.81 2.01 2.08 0.36
2.97 5.14 4.28 4.00 0.37 H526
2.29 2.15 2.26 2.10
0.87 2.72 1.63 1.99
3.16 4.87 3.89 4.09 HD912
2.06 1.93 1.98 1.89
0.93 3.02 1.62 2.14
2.99 4.95 3.60 4.03 Riscio
1.74 1.26 1.37 1.26
0.28 1.45 2.84
1.11 2.49 0.78
3.19 4.10 2.48 3.75
0.49 B9
2.87 2.99 3.43 2.86 0.56
1.12 1.16 0.32 2.22 0.57
3.99 4.15 3.75 5.08 0.57
H523 1.72 1.74
1.85 2.21 3.57 3.95
Venda 1.81 1.69
1.29 2.61 3.10 4.30
Polluce 1.36 1.29
1.75 1.96 3.11 3.25
114 1.87 1.36
1.18 2.33 3.05 3.69
929 1.49 1.06
1.56 1.93 3.05 2.99
H101 1.85 1.49
0.96 2.09 2.81 3.58
125 1.42 1.36
1.40 3.04 2.82 4.40
519 1.28 1.45
1.35 3.18 2.63 4.63
Castore 1.75 2.21
1.24 2.50 2.99 4.71
H531 2.65 2.51
1.70 1.42 4.35 3.93
H529 2.52 2.52
1.27 1.31 3.79 3.83
H913 2.82 2.48
1.24 1.55 4.06 4.03
LSD c
0.24 0.30
0.34 0.39
0.34 0.33
0.48 0.57
0.40 Mean
1.82 1.64 1.76 1.77
2.35 2.18 1.46 2.21
1.10 2.70 1.57 1.81
3.28 3.85 2.87 4.50 3.92 3.99
Source of variation d
Location L NS
NS NS
Genotype G L
×G CV
10 12
11 15
13 14
10 11
7
a UF: University farm; L: Loudias. b Between genotypes common in the two or three years at p0.05.
c Between genotypes and locations for the same year at p0.05. d NS, p0.05; p0.05; p0.01.
yield was
obtained with
H530 in
1996 on the University farm, whereas the others had a
higher yield at Loudias. In 1997, the over geno- 5.14 Mg ha−1 and the lowest with 929 in 1995
2.99 Mg ha−1. In 1995 and 1996, most of the types mean yield of secondary racemes was the
same in both locations. The locations ×genotypes
genotypes tended to have a higher yield at Loudias compared to the University farm, although the
interaction in total yield in 1997 was due to the higher yield of B9 and HD912 at Loudias and of
interaction locations ×genotypes was significant.
The increased yield was mainly the result of the Pronto on the University farm.
The percentage of the total yield that was higher yield of secondary racemes Table 3. As
mentioned above, the over genotypes mean produced in the first raceme was affected by both
locations and genotypes, and the locations ×
number of secondary racemes was also higher at Loudias. The differences between locations in the
genotypes interaction
was also
significant Table 4. The over genotypes mean yield percen-
yield of the first raceme were smaller, and some genotypes had a higher yield on the first raceme
tage in the first raceme was higher on the
233 S.D. Koutroubas et al. European Journal of Agronomy 11 1999 227–237
Table 4 Yield of the first raceme to total yield and thousand seed weight of 19 castor plant genotypes
a Genotype
Yield of the first racemetotal yield Thousand seed weight g
1995 1996
1997 LSD
b 1995
1996 1997
LSD b
UF L
UF L
UF L
UF L
UF L
UF L
Negus 0.48
0.35 0.67
0.44 0.45
0.41 0.05
359 342
365 312
358 351
34 Pronto
0.61 0.45
0.66 0.37
0.41 0.57
356 369
377 354
360 364
H530 0.60
0.45 0.53
0.48 0.07
390 375
370 377
27 H526
0.72 0.44
0.58 0.51
395 410
397 388
HD912 0.69
0.39 0.55
0.47 366
339 354
354 Riscio
0.55 0.31
0.55 0.34
0.11 325
307 318
299 25
B9 0.72
0.72 0.91
0.56 0.11
335 334
325 317
22 H523
0.48 0.44
364 371
Venda 0.58
0.39 297
314 Polluce
0.44 0.40
368 368
114 0.61
0.37 321
305 929
0.49 0.35
324 311
H101 0.66
0.42 364
308 125
0.50 0.31
341 328
519 0.49
0.31 363
326 Castore
0.59 0.47
349 368
H531 0.61
0.64 341
334 H529
0.66 0.66
350 382
H913 0.69
0.62 416
438 LSD
c 0.06
0.06 0.06
20 30
21 Mean
0.55 0.42
0.61 0.39
0.60 0.55
339 336
363 342
363 367
Source of variation d
Location L NS
NS Genotype G
L ×G
NS NS
CV 9
8 8
4 6
4 a UF: University farm; L: Loudias.
b Between genotypes common in the two or three years at p0.05. c Between genotypes and locations for the same year at p0.05.
d NS, p0.05; p0.05; p0.01.
University farm compared to Loudias. B9, which racemes, but in some genotypes, the year also
affected the yields of the first raceme Table 3. had the lowest number of secondary racemes com-
pared to all other genotypes, produced the 56– The weight of 1000 seeds was affected mainly
by the genotypes Table 4. Differences between 91
of the seed yield in the first raceme. In the other genotypes, this percentage ranged from 31
locations were observed only in 1996. Also, the genotypes behaved in the same way in the two
to 72 .
From the over years analysis of genotypes used locations, except for that of Negus, H101 and 519
in 1996, which produced heavier seeds on the for more than 1 year, differences were found in
seed yield between years, which depended on the University farm compared to Loudias. The over
locations mean heaviest seeds were produced by locations and genotypes. These differences mainly
followed the trend of the yield of the secondary Polluce 368 g in 1995, H526 403 g in 1996 and
234 S.D. Koutroubas et al. European Journal of Agronomy 11 1999 227–237
H913 427 g in 1997 and the lightest by Venda compared to 1997. The over genotypes mean oil
content was higher on the University farm com- 306 g, Riscio 309 g and B9 321 g in 1995,
1996 and 1997, respectively. pared to Loudias. The oil content varied on the
University farm from 46.3 Pronto in 1997 to
54.2 Castore in 1996 and 44.5
Pronto in 3.4. Oil content and oil yield
1997 to 53 Castore in 1996 at Loudias, but
the differences among genotypes were larger at Oil content was affected by both locations and
genotypes in all years and the locations ×
Loudias compared to those at the University farm. The highest oil content was obtained in both
genotypes interaction was significant in 1996 and 1997 but not in 1995 Table 5. Differences in oil
locations with Polluce in 1995, Castore in 1996 and H913 in 1997. The oil content was significantly
content were also observed between years, and in most genotypes, the oil content was higher in 1996
correlated with the seed yield only at Loudias in
Table 5 Oil content and oil yield of 19 castor plant genotypes
a Genotype
Oil content Oil yield Mg ha−1
1995 1996
1997 LSD
b 1995
1996 1997
LSD b
UF L
UF L
UF L
UF L
UF L
UF L
Negus 52.5
49.0 50.2
48.8 50.5
47.6 1.9
1.71 1.77
1.23 2.04
1.84 1.70
0.19 Pronto
49.0 47.3
49.9 47.5
46.3 44.5
1.58 2.19
1.67 2.18
1.81 1.50
H530 52.2
50.9 49.1
48.8 1.9
1.55 2.62
2.10 1.95
0.21 H526
53.5 52.3
50.6 49.1
1.69 2.55
1.97 2.01
HD912 53.2
50.0 51.4
50.0 1.59
2.47 1.85
2.02 Riscio
51.9 48.7
49.9 47.5
2.1 1.66
2.00 1.24
1.78 0.28
B9 52.1
49.6 48.8
48.6 1.3
2.08 2.06
1.83 2.47
0.34 H523
50.4 48.0
1.80 1.90
Venda 51.7
49.2 1.60
2.12 Polluce
53.0 51.2
1.65 1.66
114 52.6
48.5 1.60
1.79 929
53.1 50.8
1.62 1.52
H101 51.2
44.8 1.44
1.60 125
50.9 50.0
1.44 2.20
519 51.4
48.6 1.35
2.25 Castore
54.2 53.0
1.62 2.50
H531 49.3
47.7 2.15
1.88 H529
51.5 51.0
1.95 1.95
H913 52.5
51.9 2.13
2.09 LSD
c 1.5
1.9 0.5
0.26 0.30
0.24 Mean
51.8 49.1
51.7 49.2
50.0 48.8
1.70 1.89
1.48 2.21
1.96 1.95
Source of variation d
Location L NS
Genotype G L
×G NS
CV 2
3 1
10 11
9 a UF: University farm; L: Loudias.
b Between genotypes common in the two or three years at p0.05. c Between genotypes and locations for the same year at p0.05.
d NS, p0.05; p0.05; p0.01.
235 S.D. Koutroubas et al. European Journal of Agronomy 11 1999 227–237
Table 6 Contribution of the seed yield and the seed oil content to the sum of squares of the oil yield in castor plants
Resultant trait Component trait
∑x i
y ∑y
2 Location
University farm Loudias
1995 1996
1997 1995
1996 1997
Y log oil yield g m−2 X
1 log seed yield g m−2
0.985 0.838
0.818 1.145
0.724 0.783
X 2
log seed oil content g g−1 0.015
0.162 0.182
−0.145 0.276
0.217
1995 and 1996 but not in 1997. The correlation tion of castor seeds, a soil temperature of 17
°C is required Weiss, 1983. This means that the recom-
was negative in 1995 r =−0.773, p0.05 and
positive in 1996 r =0.769, p0.01.
mended time of sowing for the area is after the middle of April, when these soil temperatures are
The oil yield followed the variations in seed yield in all genotypes due to the remarkably low
usually reached. Delayed sowing until the end of the April is not expected to have any serious
variation of seed oil content CV less that 3 ,
Table 5. The over genotypes mean oil yield was consequences because the growing period in both
locations was found to be long enough for the higher at Loudias compared to University farm in
1995 and 1996 and the same in both locations in ripening of the first raceme and a number of
secondary racemes depending on the genotype. 1997. On the University farm, the highest oil yield
was obtained with H531 2.15 Mg ha−1 in 1997 The proposed period for harvesting is between the
end of September and the middle of October, and the lowest with Negus 1.23 Mg ha−1 in 1996
and at Loudias with H530 2.62 Mg ha−1 in depending on the environmental conditions of the
autumn. During this period, the first raceme and 1996
and Pronto
1.50 Mg ha−1 in 1997, respectively.
most of the secondary racemes reached maturity. The increased number of mature secondary
In Table 6 the analysis of the log of seed oil yield Y as a function of the sum of logs of seed
racemes is desirable since, in some genotypes, their contribution to total yield was more than 50
. yield X
1 and seed oil content X
2 is given. The
contribution of the seed yield to the variation of The genotypes differed in plant height and
height of insertion of the first raceme, and in most the oil yield among genotypes was the most impor-
tant and was higher than 70 in all years. It was
cases, both characteristics had higher values at Loudias compared to University farm. At Loudias,
more significant in 1995 compared to 1996 and 1997. In contrast, variation in seed oil content was
the soil organic matter and the water table were higher than on the University farm. Also, the area
small and accounted for less than 30 of the
variation of oil yield. at Loudias, where the experiments were estab-
lished, has a microclimate with a high relative humidity RH , and many fields in the immediate
neighbourhood are cultivated with rice. These con-
4. Discussion
