Morphological and agronomical characterization and estimates of genetic parameters of Sesbania Scop. (Leguminosae) accessions

E.A. Veasey¹, E.A. Schammass¹, R. Vencovsky², P.S. Martins² and G. Bandel²

¹Centro de Forragicultura e Pastagens, Instituto de Zootecnia, Rua Heitor Penteado, 56, Caixa Postal 60, 13460-000 Nova Odessa, SP, Brasil.

²Departamento de Genética, ESALQ/USP, Caixa Postal 83, 13400-970 Piracicaba, SP, Brasil. Send correspondence to G.B.

ABSTRACT

Twenty-two accessions of seven Sesbania (Leguminosae) species: S. emerus, S. rostrata, S. tetraptera, S. exasperata (annuals), S. grandiflora, S. sesban and S. virgata (perennials), used for ruminant fodder, firewood, wood products, soil improvement, and human food, were investigated, with the aim of characterizing both inter- and intraspecific genetic variability, estimating genetic parameters for the characters evaluated and appraising the forage potential of the accessions. These were planted at the Instituto de Zootecnia, Nova Odessa, SP, Brazil, in a randomized complete block design with 22 treatments and four replications. Seventeen morphological and 17 agronomic characters were evaluated. Genetic parameters coefficient of intraspecific genetic diversity (b_i) and coefficient of intraspecific genetic variation (CV_gi) were obtained for the species represented by more than one accession. Highly significant differences were observed among as well as within species for most characters, showing considerable genetic variability. S. exasperata showed intraspecific genetic variability for the largest number of morphological characters. The same was observed for S. sesban for the agronomic characters. Most of the characters gave high b_i values, above 0.80, indicating the possibility of selecting superior genotypes. The CV_gi values, on the other hand, which indicate the magnitude of the existing genetic variability relative to the character mean, varied according to the species and character evaluated. Differences between annual and perennial species were observed, with higher biomass yields presented by the annuals at the first cut and by the perennials after the second cut, reaching the highest yield at the third cut. The annual species had higher seed production. Accession NO 934 of S. sesban gave the highest biomass yields and regrowth vigor, showing promise as a forage legume plant.

INTRODUCTION

The genus Sesbania Scopoli belongs to the subfamily Papilionoideae of the Leguminosae and the botanical tribe Robinieae. The approximately 60 tropical and subtropical species include annual and perennial herbs, shrubs or small trees and are distributed in four subgenera: Agati, Daubentonia, Pterosesbania and Sesbania (Monteiro, 1984). Their nitrogen fixing ability enables these plants to grow rapidly on nitrogen deficient soils and allows their utilization for green manuring in paddy fields, for intercropping and ground cover, and for agroforestry and wood production (Ndoye et al., 1990).

Subgenus Sesbania, originating from the Old World and later dispersed throughout the New World, has the greatest number of species, although among the perennial species of this subgenus, only S. sesban has been widely utilized in agriculture for livestock forage and green manure (Brewbaker et al., 1990). This is a highly variable diploid species with two subspecies: sesban, with four varieties, and punctata (Monteiro, 1984; Bray, 1994). The varieties sesban and bicolor are very similar except for flower color. These two varieties as well as var. nubica give vigorous growth and high yields. The zambesiaca variety and ssp. punctata are less well known (Brewbaker et al., 1990). Among the annual species, subgenus Sesbania is represented throughout the Americas by S. bispinosa, S. emerus, S. oligosperma and S. exasperata (Monteiro, 1984).

The tetraploid S. grandiflora of subgenus Agati, originating in the Old World (Australia and Asia), together with S. sesban have been utilized extensively in traditional agroforestry systems (Bray, 1994). Their leaves and young twigs are used as high protein fodder for ruminants, while the thick branches and stem provide fuelwood and construction material. These species are also used to improve soil fertility and to reduce soil erosion (Heering et al., 1996a). The species with winged-pods were divided into two subgenera: subgenus Daubentonia, a native of the New World with five species, including the species S. virgata and S. punicea, and subgenus Pterosesbania, represented by only one species, S. tetraptera (Monteiro, 1984).

Agronomical and phenological evaluations of Sesbania sp. germplasm collections have been conducted by the University of Hawaii and International Livestock Centre for Africa (ILCA) in Ethiopia and Tanzania (Mengistu, 1990; Bray, 1994), the Commonwealth Scientific and Industrial Research Organisation (CSIRO) in Australia (Wood and Larkens, 1987), and the International Centre for Research in Agroforestry (ICRAF) in Kenya (Ndungu et al., 1994). In Brazil, Rocha et al. (1979) collected accessions corresponding to four species of Sesbania located mainly in damp areas near river banks in the States of São Paulo, Minas Gerais and Rio de Janeiro, from 1976 to 1978.

Evaluation of the genetic variability found in these collections represents an important step for breeding programs. Wood and Larkens (1987) presented plant growth data and general characteristics of 192 accessions of the genus Sesbania. The collection of 161 accessions of Sesbania spp. was reported by Mengistu (1990), who observed considerable polymorphism in S. sesban, occurring in a wide range of environments. The classification of a collection of 108 accessions of S. sesban based on morphological and agronomical attributes was done by Heering et al. (1996a,b), who observed distinct groups within the collection.

The objective of this study was the characterization of genetic variability based on phenological, morphological and agronomical data of 22 accessions belonging to seven Sesbania species, representing the four subgenera Sesbania, Pterosesbania, Agati and Daubentonia, as well as the estimation of genetic parameters for the characters evaluated, and an appraisal of their forage potential.

MATERIAL AND METHODS

Seven Sesbania species (22 accessions) from the germplasm bank of the Instituto de Zootecnia, in Nova Odessa, São Paulo, were evaluated (Table I), representing four different subgenera: subgenus Sesbania (S. emerus, S. exasperata, S. rostrata, S. sesban), originating from the Old World and later dispersed throughout the New World; subgenus Daubentonia (S. virgata), from the Americas; subgenus Agati (S. grandiflora), from Australia and Asia, and subgenus Pterosesbania (S. tetraptera), originating from Africa (Monteiro, 1984).

The experiment started in October 1994 at the Instituto de Zootecnia at an elevation of 550 m, S. lat. 22º42' and W. long. 47º18'. The soil was a reddish yellow podzolic, with a pH of 5.3 (CaCl₂), 3.33% organic matter, 0.19 meq/100 cm³ K, 3.80 meq/100 cm³ Ca, 1.15 meq/100 cm³ Mg, 3.68 meq/100 cm³ Al, and 25.77 mg/dm³ P. The mean annual maximum and minimum temperatures during the experimental period were approximately 29 and 17ºC, respectively, with a mean annual rainfall of 1560 mm (mean of three years).

The experiment was set up as a randomized complete block design, with four replicates and 22 treatments (accessions). Scarified seeds were sown in 67 x 34 cm-plastic trays, containing a mixture of sand and vermiculite as substrate, maintained in a greenhouse. Two months after sowing, well-developed seedlings were transplanted to the field. Plots consisted of single 16-m long rows of eight plants, spaced 2 m apart with 3 m between rows. Lime was applied at 1.92 t/ha. At planting, fertilizer was bandplaced at rates equivalent to 500 kg/ha single superphosphate, 150 kg/ha potassium chloride, 500 g/ha sodium molybdate, 8 kg/ha zinc sulfate, 4 kg/ha copper sulfate and 4 kg/ha borax. Two side dressings were applied during the experimental period, at the same rates described above, in September/95 and February/96.

Seventeen morphological, including two phenological characters were evaluated (Table II) in the first four plants of each plot. A pachymeter was used to measure the characters stem diameter (SD), pod width (PW), seed length (SL) and seed width (SW). Seventeen agronomical characters (Table III) were evaluated. With the exception of characters plant height at flowering (PHF), plant height (PH), seed production (SP) and plant survival of uncut plants (PS1), evaluated in the first four plants of each plot, all the others were evaluated in the four remaining plants. These plants were submitted to four cuts (April, August and December 1995, and April 1996), at a height of 1 m. At each cut, the plant material was separated into two fractions: leaves (including small stems < 6 mm) and stems ( > 6 mm in diameter), representing edible and inedible components, respectively.

DM, Dry matter.

Univariate analyses were undertaken, with the partitioning of the accessions within species and sources of variation, according to the mathematical model:

Y_ijlk = m + p_i + a_l(i) + b_j + e_(ijl) + d_k(ijl)

where m = overall character mean; p_i = fixed effect of the ith species (i = 1,2,...,7); a_l(i) = fixed effect of the lth accession, within the ith species (l = 1,2,...,7); b_j = random effect of the jth block (j = 1,2..,4); e_(ijl) = random effect of plot error with the lth accession of the ith species in the jth block, and d_k(ijl) = random effect of kth individual, in the plot with the lth accession of the ith species in the jth block (k = 1,2,....,4).

The analyses were conducted with plot means (Vencovsky and Barriga, 1992), obtaining the sums of squares and mean squares for blocks, species, accessions within species and error. The program SAS Version 6 (SAS Institute, 1993) was used for these analyses. Mean squares within plots were independently obtained through the means, weighted by the degrees of freedom, of the individual variance estimates within plots (Geraldi, 1977), such as:

QM_d = å (GL)_ij(QM_d)_ij / å (GL)_ij = å (SQ_d)_ij /å (GL)_ij

where QM_d = mean squares within each plot; SQ_d = sum of squares within each plot, and GL = degrees of freedom within each plot. The mean squares within plots (QM_di) for each of the four species with more than one accession was also obtained.

The analysis of variance and expected mean squares are summarized in Table IV, where = phenotypic variance among plants within plots; = variance of the error among plots; V_ai = measure of genetic diversity among accessions of the ith species (i = 1, .., 4); V_p = measure of genetic diversity among species; J = replicate number; = mean number of plants per plot; = mean number of accessions per species (arithmetic mean), and = coefficient obtained to correct for the different number of accessions per species.

There was a different number of plants per plot as a consequence of plant death; therefore, the mean number of plants per plot () was obtained through the harmonic mean, as follows:

= N / ån(1/k), where N = total number of plots; n = number of plots with k plants, and k = number of plants per plot (k = 1, ..., 4). Similarly, the mean number of plants per plot for each species was obtained separately (_i). Due to the different number of accessions per species, a coefficient ( ) was calculated in order to correct for this problem, according to the following equation described by Weir (1996):

where p = number of species and ci = number of accessions of the ith species.

The phenotypic variance within species ( ), error variance among plots ( ), genetic variation among species and among accessions within the species S. emerus ( ), S. exasperata ( ), S. sesban ( ) and S. virgata ( ) were estimated from the mean squares, according to the equations:

These quantities permitted the estimation of intraspecific phenotypic variance (), intraspecific genotypic determination (b_i), and intraspecific genetic variation coefficient (CV_gi), according to the following equations:

RESULTS AND DISCUSSION

Considerable interspecific variability was observed for all the characters evaluated (P < 0.001), which is expected since the comparison is between annual and perennial species, belonging to different subgenera. In the measures of intraspecific variation, 76.5, 100, 82 and 71% of the morphological characters showed significant differences (P < 0.05) between accessions of S. emerus, S. exasperata, S. sesban and S. virgata, respectively. S. exasperata had the highest intraspecific variation, followed in decreasing order by S. sesban, S. emerus and S. virgata (Tables V and ^VI). Intraspecific variation for S. grandiflora, S. rostrata and S. tetraptera could not be estimated as these species were represented by only one accession in this study.

SV

NF

(No.)

PDL

(cm)

FL

(cm)

PL

(cm)

PW

(cm)

SL

(cm)

SW

(cm)

NSP

(No.)

FDS

(%)

Species (V_p)

152.4164^a

0.2040^a

0.9629^a

55.8596^a

0.1095^a

0.0317^a

0.0160^a

189.4384^a

135.6366^a

Accessions within:

S. emerus (V_a1)

2.8782^b

0.0091^b

0.0150^b

0.1655

0.0022^a

0.0079^a

0.0006^a

79.3061^a

37.1697^c

S. exasperata (V_a2)

4.9572^a

0.3041^a

0.0452^a

7.4416^a

0.0002^b

0.0085^a

0.0001^b

37.0672^a

127.8382^a

S. sesban (V_a3)

7.3191^a

0.0305^a

0.0164^a

3.8018^a

0.0003^b

0.0015^a

@ 0.0000

33.7496^a

@ 0.0000

S. virgata (V_a4)

4.0263^a

0.0003

@ 0.0000

0.0075

0.0011^a

0.0001^b

0.0001^a

@ 0.0000

1.2994

Mean

14.9 ±

0.1

0.9 ±

0.0

2.0 ±

0.0

14.0 ±

0.1

0.6 ±

0.0

0.5 ±

0.0

0.3 ±

0.0

19.8 ±

0.2

83.8 ±

0.6

CV (%)

7.47

5.19

3.71

6.42

2.67

2.32

3.10

7.81

6.03

b₁

0.9193

0.9407

0.9131

0.4384

0.9714

0.9948

0.9630

0.9916

0.8828

b₂

0.9511

0.9973

0.9707

0.9719

0.8078

0.9959

0.8313

0.9828

0.9374

b₃

0.9638

0.9807

0.9196

0.9452

0.8319

0.9767

0.1608

0.9800

-

b₄

0.9121

0.3892

-²

0.0399

0.9380

0.7619

0.8497

-

0.1869

CV_g1 (%)

42.68

14.92

7.08

2.17

14.83

23.95

11.40

27.00

6.63

CV_g2 (%)

41.51

34.82

5.99

12.79

3.31

20.40

3.68

21.04

16.14

CV_g3 (%)

30.29

20.38

6.21

12.34

4.82

10.18

0.97

21.75

-

CV_g4 (%)

6.72

2.84

-

1.48

3.57

1.58

2.47

-

1.24

In general, the genotypic determination coefficients (b_i) were high, above 0.80, indicating the possibility of selection for these characters. CV_gi values, however, varied considerably according to the character and the species analyzed. For example, the character SD presented a high CV_gi value of 89.78% for S. emerus and also a high b_i value of 0.9977, indicating the possibility of selection for this character. The averages presented by the two S. emerus accessions for SD were highly contrasting (Table VII). For S. sesban, the b_i value for this same character was high, 0.9181, but the CV_gi of 16.96% was low, with little genetic variability existing for selection to act upon this character. Another character that had a high b_i as well as a high CV_gi value was floral initiation (FI) for S. emerus (Table V), with means varying from 42.1 to 175 days (Table VII), indicating precocity for accession No. 1. The characters pedicel length (PDL), pod length (PL) and percentage of fully developed seeds per pod (FDS) presented low values for both b_i and CV_gi for S. virgata, indicating low genetic variability for these traits. Variance was nil for the characters flower length (FL) and number of seeds per pod (NSP) for S. virgata, leaflet width (LFW) for S. emerus, and SW and FDS for S. sesban (Tables V, ^VI, VII and ^VIII).

No.

Species

NF

(No.)

PDL

(mm)

FL

(mm)

PL

(cm)

PW

(mm)

NSP

(No.)

FDS

(%)

SL

(mm)

SW

(mm)

1

S. emerus

2.7 ^j

7.1 ^fg

16 ^gh

18.3 ^cd

3.5 ^e

26.6 ^d

87.2 ^abc

4.3 ^d

2.2 ^f

2

"

5.2 ^hij

5.7 ^h

18 ^fg

19.2 ^cd

2.8 ^f

39.3 ^a

96.6 ^ab

3.1 ^g

1.9 ^g

3

S. exasperata

4.3 ^ij

19.7 ^a

36 ^ab

24.0 ^a

4.6 ^d

31.5 ^c

83.6 ^bcd

5.0 ^c

2.7 ^cde

4

"

4.5 ^ij

17.0 ^b

35^b

20.1 ^bc

4.5 ^d

24.7 ^de

90.7 ^abc

5.0 ^c

2.6 ^de

5

"

3.9 ^ij

18.9 ^a

38 ^a

18.0 ^cd

4.8 ^d

23.1 ^de

83.0 ^cde

4.9 ^c

2.8 ^cd

6

"

8.8 ^fg

7.7 ^ef

33 ^c

23.2 ^a

4.4 ^d

36.3 ^ab

63.7 ^f

3.1 ^g

2.5 ^e

7

S. grandiflora

-¹

-

-

-

-

-

-

-

8

S. rostrata

10.6 ^f

14.1 ^c

22 ^d

21.8 ^ab

4.6 ^d

34.8 ^bc

64.7 ^f

3.6 ^f

2.9 ^c

9

S. sesban

14.3 ^e

8.2 ^ef

22 ^d

13.2 ^f

3.3 ^e

21.6 ^e

69.5 ^f

3.8 ^f

1.9 ^g

10

"

7.8 ^fgh

10.4 ^d

20 ^de

15.2 ^ef

3.2 ^ef

23.8 ^de

71.9 ^def

3.9 ^ef

2.0 ^fg

11

"

7.9 ^fgh

8.5 ^e

19 ^ef

-

-

-

-

-

-

12

"

8.0 ^fgh

10.8 ^d

19 ^ef

17.7 ^d

3.5 ^e

26.4 ^d

70.0 ^ef

4.2 ^de

2.0 ^fg

13

"

9.2 ^fg

7.2 ^efg

20 ^de

-

-

-

-

-

-

14

"

6.5 ^ghi

6.3 ^gh

22 ^d

17.1 ^de

3.6 ^e

35.0 ^bc

68.8 ^f

3.2 ^g

2.0 ^fg

15

S. tetraptera

10.3 ^f

14.8 ^c

16 ^h

17.0 ^de

12.0 ^a

21.8 ^e

97.6 ^a

4.3 ^d

2.7 ^cde

16

S. virgata

30.5 ^abc

5.7 ^h

11 ⁱ

6.3 ^g

9.4 ^b

4.6 ^f

94.0 ^abc

7.3 ^ab

4.4 ^b

17

"

31.8 ^ab

6.2 ^gh

11ⁱ

5.1 ^g

8.5 ^c

4.1 ^f

94.8 ^abc

7.1 ^b

4.6 ^ab

18

"

29.0 ^bcd

5.6 ^h

11ⁱ

5.8 ^g

9.1 ^b

4.6 ^f

88.2 ^abc

7.2 ^ab

4.4 ^b

19

"

33.1 ^a

5.9 ^gh

11 ⁱ

6.1 ^g

9.2 ^b

4.9 ^f

89.1 ^abc

7.1 ^b

4.5 ^ab

20

"

26.9 ^d

5.8 ^h

11ⁱ

6.0 ^g

9.2 ^b

4.5 ^f

90.9 ^abc

7.2 ^ab

4.5 ^ab

21

"

28.5 ^cd

5.8 ^h

11 ⁱ

6.3 ^g

9.5 ^b

4.7 ^f

93.5 ^abc

7.4 a

4.6 ^ab

22

"

29.2 ^bcd 6.4 ^h 11 ⁱ 5.4 ^g 9.1 ^b 4.2 ^f 95.0 ^abc 7.1 ^ab 4.7 ^a

Significant differences (P < 0.05) between accessions of S. emerus, S. exasperata, S. sesban and S. virgata were presented by 80.0, 77.8, 86.7 and 33.3% of the agronomical characters, respectively (Tables IX and ^X Table X - Estimates of variance among (Vp) and within (Va) Sesbania spp., coefficients of variation (CV), coefficients of intraspecific genetic variation (CVgi), and coefficients of intraspecific genotypic determination (bi), for the agronomic characters: stem DM yield of the 4th cut (SDMY4), crude protein percentage of the 1st, 2nd, 3rd and 4th cuts (CP1, CP2, CP3, CP4), plant height at flowering (PHF), plant height (PH), and seed production (SP). ). S. sesban had the greatest intraspecific variation in the agronomic characters. After the second cut, high values were observed both for b_i and CV_gi in S. sesban for the characters leaf dry matter yield (LDMY) and stem dry matter yield (SDMY).

SV

SDMY4

(g)

CP1

(%)

CP2

(%)

CP3

(%)

CP4

(%)

PHF

(m)

PH

(m)

SP

(g)

Species (_Vp)

20016.46^a

8.8976^a

29.9055^a

1.9446^a

0.6618^c

0.5899^a

0.4011^a

11028.75^a

Accessions within:

S. emerus (V_a1)

-²

26.1378^a

-

-

-

1.8750^a

-

5727.41^c

S. exasperata (V_a2)

-

16.6858^a

29.4715^a

10.6282^a

-

0.3082^a

-

25779.69^a

S. sesban

(V_a3)

12797.96

2.5888^b

8.3486^a

4.3558^a

2.8439^b

0.1254^a

1.2046^a

@ 0.00

S. virgata (V_a4)

2365.76

@ 0.0000

3.1200^b

2.0285^b

@ 0.0000

0.0071

0.0312

1174.34

Mean

252.7 ± 11.4

25.8 ± 0.19

17.4 ± 0.20

19.1 ± 0.20

16.0 ± 0.21

2.2 ± 0.02

3.8 ± 0.05

122.3 ± 7.29

CV (%)

34.19

6.74

9.79

8.54

10.30

9.80

9.37

51.98

b₁

-

0.9706

-

-

-

0.9935

-

0.8573

b₂

-

0.9562

0.9765

0.9206

-

0.9542

-

0.9594

b₃

0.9005

0.7685

0.9177

0.8733

0.8170

0.9194

0.9713

-

b₄

0.5414

-

0.8236

0.7590

-

0.4184

-

0.5558

CV_g1 (%)

-

19.61

-

-

-

51.96

-

81.09

CV_g2 (%)

-

15.93

43.33

19.83

-

16.49

-

73.73

CV_g3 (%)

86.74

6.24

21.40

11.28

10.86

17.79

26.90

-

CV_g4 (%)

13.14

-

7.76

7.10

-

5.06

-

42.37

S. sesban, represented by six accessions, showed considerable genetic variation for several characters, such as first mature pod (FMP), number of leaflet pairs (NLP), leaflet length (LFL), inflorescence length (IL), number of flowers (NF), PDL, PL, SD, LFW, FL, NSP, SL, LDMY (2nd to 4th cut), CP (1st to 4th cut), PHF and PH (Tables VII, ^VIII, ^XI and ^XII). Significant differences (P < 0.05) were also observed between accessions of S. sesban 56 days after sowing (DAS) for the characters plant height and top growth dry weight (Veasey et al., 1997). At this stage, the plants had already attained heights varying from 59 to 81 cm, reflecting the rapid growth of these species compared with the performance of Leucaena leucocephala, which varied from 18 to 22 cm.

No.

Species

LDMY1

(g)

LDMY2

(g)

LDMY3

(g)

LDMY4

(g)

SDMY1

(g)

SDMY2

(g)

SDMY3

(g)

SDMY4

(g)

1

S. emerus

160 ^cde

-1

-

-

21 ^cd

-

-

-

2

"

193 ^cde

51 ^g

-

-

208 ^ab

71 ^e

-

-

3

S. exasperata

631 ^a

322 ^fg

-

-

264 ^a

30 ^e

-

-

4

"

538 ^ab

389 ^efg

-

-

207 ^ab

99 ^e

-

-

5

"

242 ^cde

269 ^fg

-

-

88 ^cd

38 ^e

-

-

6

"

160 ^cde

186 ^g

384 ^d

44 ^h

122 ^bc

252 ^de

345 ^bdce

135 ^cde

7

S. grandiflora

17 ^e

104 ^g

491 ^cd

173 ^efgh

8 ^d

51 ^e

243 ^cde

195 ^bcde

8

S. rostrata

674 ^a

-

-

-

285 ^a

-

-

-

9

S. sesban

277 ^cd

2123 ^a

3006 ^a

1146 ^a

42 ^cd

1336 ^a

1076 ^a

321 ^abc

10

"

240 ^cde

1268 ^bc

1455 ^bc

210 ^efgh

44 ^cd

772 ^b

586^abcde

56 ^de

11

"

98 ^de

732 ^def

671 ^cd

107 ^fgh

23 ^cd

266 ^de

122 ^de

-

12

"

250 ^cde

1115 ^bcd

1158 ^bcd

285 ^efgh

56 ^cd

643 ^bc

375 ^bcde

87 ^de

13

"

90 ^de

709 ^def

660 ^cd

69 ^gh

24 ^cd

214 ^de

72 ^e

12 ^e

14

"

246 ^cde

1353 ^b

1678 ^b

706 ^bcd

62 ^cd

1177 ^a

748 ^abc

247 ^bcd

15

S. tetraptera

364 ^bc

-

-

-

46 ^cd

-

-

-

16

S. virgata

147 ^cde

481 ^efg

2022 ^b

488 ^bcde

23 ^cd

77 ^e

931 ^ab

381 ^ab

17

"

164 ^cde

820 ^cde

1180 ^bcd

832 ^ab

33 ^cd

340 ^cde

549^abcde

501 ^a

18

"

126 ^cde

453 ^efg

1669 ^b

411 ^defg

15 ^d

69 ^e

688^abcde

336 ^abc

19

"

145 ^cde

334 ^fg

1792 ^b

512 ^bcde

18 ^d

44 ^e

698 ^abcd

389 ^ab

20

"

95 ^de

458 ^efg

1658 ^b

459 ^cdef

13 ^d

88 ^e

808 ^abc

316 ^abc

21

"

160 ^cde

302 ^fg

1430 ^bc

338 ^efgh

22 ^cd

71 ^e

687^abcde

309 ^abc

22

"

215 ^cde

1046 ^bcd

1339 ^bcd

785 ^abc

37 ^cd

454 ^bcd

748 ^abc

358 ^ab

No.

Species

CP1

(%)

CP2

(%)

CP3

(%)

CP4

(%)

PHF

(m)

PH

(m)

SP

(g)

PS1

(%)

PS2

(%)

1

S. emerus

22.4 ^gh

-1

-

-

1.7 ^ghi

-

35.3 ^cd

0.0

0.0

2

"

29.7 ^abc

-

-

-

3.6 ^a

-

151.4^bcd

0.0

0.0

3

S. exasperata

23.8 ^defgh

9.4 ^e

-

-

3.9 ^a

-

391.3 ^a

0.0

0.0

4

"

22.9 ^efgh

10.9 ^e

14.9 ^e

-

3.7 ^a

-

288.6 ^ab

0.0

0.0

5

"

24.1 ^defg

9.2 ^e

-

-

3.3 ^ab

-

183.1 ^bc

0.0

0.0

6

"

31.9 ^ab

20.8 ^bc

20.7 ^abc

14.8^ab

2.6 ^cd

2.8 ^fg

7.9 ^d

0.0

0.0

7

S. grandiflora

32.5 ^a

22.1 ^ab

20.2 ^abc

18.1 ^a

1.0 ^j

3.2 ^efg

-

43.7

0.0

8

S. rostrata

19.4 ^hi

-

-

-

3.0 ^bc

-

305.6 ^ab

0.0

0.0

9

S. sesban

28.2 ^abcd

13.0 ^de

17.5^bcde

13.3 ^b

1.9 ^efghi

5.1 ^a

33.8 ^cd

81.2

0.0

10

"

25.3 ^cdefg

13.4 ^de

19.7 ^abc

15.9^ab

2.2 ^defg

4.4 ^abc

3.3 ^d

68.7

6.2

11

"

25.8 ^cdefg

17.1 ^cd

20.9 ^abc

17.1^ab

1.6 ^ghi

2.6 ^g

-

43.7

0.0

12

"

26.0 ^cdefg

9.5 ^e

16.8 ^cde

14.7^ab

2.3 ^def

4.8 ^ab

27.7 ^cd

93.7

33.3

13

"

26.6 ^cdefg

16.8 ^cd

20.5 ^abc

18.3 ^a

1.5 ^hij

2.8 ^fg

-

6.2

0.0

14

"

22.6 ^fgh

11.1 ^e

15.4 ^de

13.9 ^b

2.5 ^cde

4.8 ^ab

11.9 ^d

87.5

37.5

15

S. tetraptera

17.4 ⁱ

-

-

-

2.1^defgh

-

316.7 ^ab

0.0

0.0

16

S. virgata

27.1 ^cdef

23.7 ^ab

19.3^abcd

15.8^ab

1.6 ^hij

3.8 ^cde

82.9 ^cd

100.0

100.0

17

"

26.8 ^cdefg

20.1 ^bc

22.2 ^a

16.6^ab

1.8 ^fghi

3.6 ^cdef

53.7 ^cd

87.5

87.5

18

"

27.8 ^bcd

24.1 ^ab

18.1^abcde

16.6^ab

1.6 ^hij

3.7 ^cde

52.0 ^cd

100.0

100.0

19

"

27.0 ^cdefg

25.2 ^a

21.8 ^a

16.6^ab

1.7 ^fghi

3.4 ^defg

69.9 ^cd

100.0

100.0

20

"

27.5 ^bcde

22.4 ^ab

18.7^abcde

15.9^ab

1.5 ^ij

3.8 ^cde

52.9 ^cd

100.0

100.0

21

"

26.9 ^cdefg

23.6 ^ab

19.0^abcde

15.4^ab

1.6 ^ghi

4.0 ^bcde

71.1 ^cd

100.0

81.2

22

"

27.9 ^abcd

20.2 ^bc

21.2 ^ab

17.4^ab

1.9 ^efghi

4.2 ^abcd

183.5 ^bc

100.0

81.2

Mengistu (1990), evaluating 161 accessions of Sesbania spp. collected in various ecological zones of Tanzania, observed considerable polymorphism for S. sesban, with variations in size, seed production and leafness, occurring over a wide range of altitudes (200-2200 m), rainfall (600-1400 mm, mean annual), vegetation (savannah, steppe, rain forest, semiarid) and a variety of soils. Variation in leaf dry matter (DM) yield and seed production was also observed by Heering (1995) for S. sesban. Karachi et al. (1994) observed significant differences (P < 0.05) between accessions of S. sesban for height, basal collar diameter, primary branching, number of primary, secondary and tertiary branches, number of days to flowering and fodder and wood DM production. Heering et al. (1996a,b) classified a collection of 108 accessions of S. sesban into distinct groups based on morphological and agronomical data, separating accessions of the varieties sesban and bicolor from accessions of var. nubica. These two groups were separated principally because of the morphological characters leaflet size and number, seed size and color, and growth habit (Heering et al., 1996a). Considerable variation was also observed by Karachi and Matata (1997) in forage yield between accessions of S. sesban and S. macrantha.

S. emerus was represented by two accessions only, that contrast greatly in various characters. Indeed, because of the highly significant differences (P < 0.001) observed between them for characters FI, FMP, SD, leaf length (LL), NLP, PDL, PW, NSP, SL, SW, SDMY1, CP1 and PHF (^{Tables VI}, ^VIII, ^XI and ^XII), a hypothesis that these accessions belong to different species was considered. However, the results obtained in the isozyme analyses, conducted in parallel to this trial (Veasey, 1998), confirm the identification of these two accessions as belonging to the same species, as the band patterns for both accessions were very similar. Consequently, this confirms the existence of great genetic variability in S. emerus, which corroborates the extensive geographic distribution of this species. This last factor, associated with the morphological plasticity expressed in different areas of its occurrence, also generated a complex taxonomic history for this species (Monteiro, 1984).

The genetic variability observed for S. exasperata (Tables V, ^VI, IX and ^X Table X - Estimates of variance among (Vp) and within (Va) Sesbania spp., coefficients of variation (CV), coefficients of intraspecific genetic variation (CVgi), and coefficients of intraspecific genotypic determination (bi), for the agronomic characters: stem DM yield of the 4th cut (SDMY4), crude protein percentage of the 1st, 2nd, 3rd and 4th cuts (CP1, CP2, CP3, CP4), plant height at flowering (PHF), plant height (PH), and seed production (SP). ) is mostly due to accession No. 6, which differs significantly (P < 0.05) from other accessions of this species for the characters FI, FMP, LL, NLP, IL, NF, PDL, FL, NSP, FDS, SL, CP1, CP2, CP3, PHF and SP (Tables VII, ^VIII, ^XI and ^XII). This accession was classified in isolated groups in cluster analyses with morphological, agronomical and isozyme data (Veasey, 1998).

S. virgata also presented significant differences (P < 0.05) between accessions for characters FI, FMP, LL, LFL, LFW, IL, NF, PW, LDMY2, LDMY4 and SDMY2 (Tables VII, ^VIII and ^XI). This species belongs to subgenus Daubentonia and is native to the New World (Monteiro, 1984), commonly occurring along river banks and waterlogged areas in various regions of Brazil. The seven accessions evaluated in this study were collected in the States of São Paulo, Minas Gerais, Rio de Janeiro and Mato Grosso do Sul (Table I), including the Mato Grosso swampland, which indicates the wide geographic distribution of this species in Brazil. Its use as a forage plant is doubtful as it is suspected that it has anti-nutritive and toxic factors. Tsuhako et al. (1989) reported toxic factors in seed of S. virgata. The use of this species to recuperate poor soils and for erosion control was evaluated by Franco et al. (1990) in soils whose horizons A and B had been removed, obtaining better results with the species Mimosa caesalpiniaefolia and Acacia auriculiformis than with Sesbania virgata and Gliricidia sepium.

As for the crude protein percentage, with the exception of the annuals S. rostrata and S. tetraptera, high values were observed at the first cut, above 22%, for all accessions (^{Table XII}). However, there was a sharp drop after the second cut, except for S. virgata, which maintained high values up to the third cut. Accession No. 9 of S. sesban, which stood out in terms of LDMY and SDMY, presented a higher value at the first cut but showed a strong reduction following the second cut, with values of 28, 13, 17 and 13% from the first to the fourth cut, respectively, and a mean value of 18% of CP.

A small number of accessions per species was evaluated in this study, especially S. emerus with only two accessions, which is far from an ideal number to represent a species. However, even with such small accession numbers, high levels of genetic variability were observed.

Tables XIII and ^XIV show the amplitude of variation for each character evaluated for morphological data compared with data observed by Monteiro (1984) for S. emerus, S. exasperata, S. sesban, S. virgata and S. grandiflora and by Gutteridge et al. (1995) for S. sesban. Some observations come close to data obtained by Monteiro (1984), such as characters FL and PL for S. emerus, LL, LFW, PDL, PW and SW for S. sesban and PW and SL for S. virgata, but the majority of the observations differ from data obtained by Monteiro (1984). Character LL in S. exasperata, for example, varied from 10.5 to 26.6 cm in this study and from 20 to 40 cm in Monteiro (1984). Therefore, S. exasperata has high variability for LL. Another example would be the character NSP in S. virgata, varying from one to eight seeds in this study and from five to nine in Monteiro (1984). Monteiro (1984) based his taxonomic review for Sesbania species from the New World on herborized material and some morphometric restrictions always exist in dehydrated specimens. However, a wide geographical range was considered in this review.

¹Characters described in Table II. ²Days after sowing. ³Data not obtained.

Character1

S. grandiflora

S. rostrata

S. tetraptera

Veasey

et al.

Monteiro

(1984)

Veasey

et al.

Veasey

et al.

FI

(DAS)²

-³

-

128-139

119-139

FMP

(DAS)

-

-

174-184

163-188

SD

(cm)

8.9-18.0

-

17.8-27.9

7.4-23.2

LL

(cm)

21.7-33.5

25-30

10.8-22.8

12.7-20.2

LFL

(mm)

28 – 43

25-35

0.9-3.2

1.8-2.8

LFW

(mm)

10 – 15

7-10

0.4-0.7

0.5-0.7

NLP

(nº)

14 – 25

10-25

18-27

15-28

IL

(cm)

-

10-20

2.2-6.6

5.3-17.5

FL

(mm)

-

50-100

2.0-2.7

1.4-2.0

NF

(nº)

2-4

5-15

5-15

-

PDL

(mm)

-

20

1.1-1.8

1.0-2.0

PL

(cm)

-

-

15.6-26.2

13.4-19.4

PW

(mm)

-

-

0.39-0.55

1.01-1.43

SL

(mm)

-

6

0.28-0.37

0.40-0.55

SW

(mm)

-

4

0.17-0.24

0.23-0.30

NSP

(nº)

-

25-40

13-46

16-26

FDS

(%)

-

-

43.7-88.0

90.1-100.0

Similarly, Gutteridge et al. (1995) observed a variation of 10 to 50 NSP for S. sesban, while in this study it varied from 11 to 46 NSP. Gutteridge et al. (1995) also observed a range in NLP from six to 27, while this study revealed a range of nine to 30 NLP.

As for phenological data, S. sesban presented floral initiation in the months of March and April, varying from 128 to 195 days after sowing (DAS) (Table XIII). The first mature pod for S. sesban accessions appeared after 162 to 378 days. Sedi and Humphreys (1992) established that flowering of S. sesban var. nubica is sensitive to temperature but not to daylength. Wood and Larkens (1987) observed greater variability for FI between accessions of S. sesban, from 75 to 218 days. Karachi et al. (1994) evaluated more precocious accessions flowering from 59 to 97 DAS, while Heering (1994) evaluated accessions flowering from 102 to 153 DAS when 50% of the plants had flowered.

The most precocious of all species was S. emerus, whose accession No. 1 flowered from 38 to 45 DAS. Flowering for this species occurred in the months of December to April and presented the FMP from February to June. S. virgata presented the latest flowering accessions, with FI varying from 121 to 384 DAS, from February to November, with the FMP occurring from April to December. S. exasperata flowered in the months of February to June, from 119 to 248 DAS. S. tetraptera flowered from February to March, from 119 to 149 DAS. S. rostrata initiated flowering in March, from 128 to 139 DAS, and seeding in April, from 174 to 184 DAS (^{Table XIV}). Earlier flowering was observed by Joshua and Ramani (1993) for S. rostrata, 35 to 40 DAS for both control and M₂ plants in a mutation induction trial with gamma rays.

Wood and Larkens (1987) evaluated only one accession of S. virgata, that initiated flowering at 115 DAS, of S. exasperata, which flowered at 106 DAS and started seeding at 141 DAS, and of S. tetraptera, which flowered at 111 DAS and initiated seeding at 158 DAS. The authors also observed FI at 112 DAS and FMP at 173 DAS for S. rostrata.

S. grandiflora did not flower throughout the experimental period and presented low leaf and stem DM yields when compared with the other perennial species, indicating that it is not suitable for this region. However, before a more definitive conclusion can be drawn, more accessions of this species should be tested.

The difference between annual and perennial species can be observed in Figure 1, where at the first cut the annual species S. exasperata (accessions Nos. 3 and 4), S. rostrata (No. 8) and S. tetraptera (No. 15) were the most productive. The annual S. emerus, however, presented lower forage yields. After the second cut, the superiority of the perennial S. sesban, followed by S. virgata, was observed, as well as the reduction of LDMY of the annual species. The peak of production for the perennial species was reached at the third cut, made in December 1995, one year after the seedlings were transplanted to the field. The superiority of accessions No. 9 and 14 of S. sesban, both from subspecies sesban var. sesban, was noted, principally accession No. 9 (NO 934). At the fourth cut, in April 1996, a sharp fall was observed when compared with the yields at the third cut, clearly showing the weak perenniality of the Sesbania species.

Heering (1995) established that the majority of the S. sesban accessions evaluated attained their maximum production before or at the second cut, with DM yields decreasing rapidly soon afterwards. However, some accessions of S. sesban evaluated by Heering et al. (1996b), classified in a separate group in the cluster analysis, maintained high productivity even after the second cut. Moisture is one of the factors limiting production in S. sesban and S. macrantha, according to Karachi and Matata (1997), who concluded that the usefulness of sesbania as a potential source of fodder is limited to wet season growth. Bray et al. (1997) also observed less perenniality for S. sesban and S. grandiflora, as well as for Cajanus cajan, Codariocalyx gyroides and Desmodium discolor. These authors commented on the important role that these species can play when cultivated with more perennial species, as early sources of animal feed in the form of protein banks, during the period in which the perennial species are developing before they attain their peak of production.

Another comparison between annual and perennial species can be made for seed production which was much greater for the annual species S. emerus (No. 2), S. exasperata, S. rostrata and S. tetraptera than for the perennials (^{Table XII}). Reis (1984) also observed greater production of seeds, bracts and flowers by the annual Stylosanthes humilis when compared with the perennial S. guianensis, S. hamata and S. scabra, in agreement with the r and k selection theory (MacArthur and Wilson, 1967), which attributes greater reproductive effort for shorter life-cycle populations of precocious sexual maturity than perennial, late flowering populations subject to more stable environments.

A plant survival evaluation was carried out at the end of the experiment, two years and four months after sowing (^{Table XII}). The annual species did not survive up to that time. In general, plants that did not undergo cutting presented greater persistence, with high variation being observed for S. sesban, from 6.2 to 93.7%.

ACKNOWLEDGMENTS

The authors are grateful to Professor Reinaldo Monteiro from the University of the State of São Paulo, Rio Claro, for identification of Sesbania accessions and reviewing the manuscript; to Dr. Joaquim Carlos Werner and Dr. Darcy A. Beisman, from the Instituto de Zootecnia, Nova Odessa, for technical assistance; to Centro Nacional de Recursos Genéticos e Biotecnologia (CENARGEN/Embrapa), Brasília, The Henry Doubleday Research Association, Coventry/UK, Companhia do Vale do Rio Doce, Linhares, and Instituto de Zootecnia, Nova Odessa, for providing the seed; and to Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) for financial support. Publication supported by FAPESP.

RESUMO

Sete espécies de Sesbania (Leguminosae): S. emerus, S. exasperata, S. rostrata, S. tetraptera (anuais), S. grandiflora, S. sesban e S. virgata (perenes), utilizadas como forrageira para ruminantes, madeira para lenha e construção, melhoramento do solo e alimentação humana, foram avaliadas neste trabalho, num total de 22 acessos, com o objetivo de caracterizar a variabilidade genética tanto inter como intraespecífica, estimar parâmetros genéticos para os caracteres avaliados e avaliar o potencial forrageiro dos acessos. O experimento foi conduzido no Instituto de Zootecnia, em Nova Odessa, SP, em blocos ao acaso com 22 tratamentos e quatro repetições. Foram avaliados 17 caracteres morfológicos, incluindo dois de fenologia, e 17 caracteres agronômicos. Foram estimados os parâmetros coeficiente de diversidade genotípica intraespecífica (b_i) e coeficiente de variação genética intraespecífica (CV_gi) para as espécies representadas por mais de um acesso. Diferenças altamente significativas foram observadas tanto entre como dentro de espécies para a maioria dos caracteres avaliados, mostrando a grande variabilidade genética do material em estudo. S. exasperata apresentou variação intraespecífica para o maior número de caracteres morfológicos, o mesmo sendo observado para S. sesban para caracteres agronômicos. Valores elevados de b_i, acima de 0,80, foram obtidos para a maioria dos caracteres, indicando possibilidade de seleção de genótipos superiores. Já os valores de CV_gi, que indicam a magnitude da variabilidade genética existente com relação à média do caráter, variaram de acordo com a espécie e o caráter avaliado. Observaram-se diferenças entre espécies anuais e perenes, com as anuais apresentando maior produção de biomassa no primeiro corte e as perenes a partir do segundo corte, atingindo o pico de produção no terceiro corte. As espécies anuais apresentaram maior produção de sementes. O acesso NO 934 de S. sesban se destacou quanto à produção de biomassa e vigor de rebrota, sendo indicado como promissor para uso como leguminosa forrageira.

(Received January 22, 1998)

Publication Dates

History