Lathyrus Lathyrism Newsletter 4 (2005)

Characterization of grass pea (Lathyrus sativus L.) entries

by means of agronomically useful traits
G. B. Polignano1 1*, P. Uggenti1, G. Olita1, V. Bisignano1, V. Alba2 and P. Perrino1
1. Istituto di Genetica Vegetale, C.N.R., Via Amendola 165/A, 70126 Bari, Italia
2. Dipartimento di Biologia Difesa e Biotecnologie Agro Forestali, Università della Basilicata, Potenza

*E-mail: [email protected]

Introduction distribution of the residual genetic variation is

Grass pea or chickling pea (Lathyrus sativus L.) is a recommended for the benefit of both direct users and
diploid (2n=14), self-pollinated annual with branched, crop improvement programmes (12).
straggling, or climbing habit, blue (sometimes violet
or white) flowers and characteristic smooth seed with In this regard it is important to underline that the
pressed sides (16). The center of origin and consumption of L. sativus seeds by humans and
diversification of the Lathyrus gene pool is in the animals has been limited by presence of a neurotoxin
Mediterranean region (15). The earliest archaeological known as β-N-oxalyl-L-α,β-diaminoproprionic acid
remains of Lathyrus appear in the Neolithic in the (β-ODAP) in the seeds, which when taken in large
Balkans and Near East of Bulgaria, Cyprus, Iraq, Iran quantity can lead to “lathyrism” a disease causing
and Turkey (4). According to Kupicha (8), the genus paralysis of the limbs (3). With that premise, breeding
Lathyrus contains about 150 species but only L. programmes evolving genotypes combining high yield
sativus is widely cultivated for human consumption, with high protein content and low neurotoxin (ODAP)
particularly in Bangladesh, China, Ethiopia, India, are in progress all over the world (7). At the same time
Nepal and Pakistan (10). In Italy, this crop has mostly it was felt that it was necessary to evaluate and
disappeared and today it is no longer seen as the ‘food describe the genetic diversity available in the grass
of the poor’ as it was in the past. Fortunately, grass pea collections (2,6,10,11,12). In other words there is a
pea is still used by local populations in marginal areas, need to survey, collect, conserve and characterise the
and sold in some marketplaces. valuable resources of the Lathyrus species germplasm
for the benefit of both users and crop improvement
Emerging global and national strategies on sustainable programmes.
farming systems, sustainable development and the
preservation of biological diversity reflect concern at The main objective of the present research was to
adequate quantification of local biodiversity. study the variation in a collection of grass pea entries
Consequently, researchers, farmers and policy makers with respect to yield capacity and other important
have focused their attention on the neglected and/or agronomic traits (such as biomass) with the aim of a
underutilised crops to improve the food security, direct utilisation of the most promising material and
nutrition and economic welfare of humans all around their use in cross combinations for breeding purposes.
the world (5). In Italy, among these species the grain
legume grass pea has received renewed attention as a Material and Methods
local and typical product, it is becoming an exclusive Seventy-six grass pea entries of different geographical
and fashionable food for which discerning consumers origin were used. These were subset of the whole
are prepared to pay a higher price than for other pulse collection including entries characterized by desirable
products. In addition, the most interesting agronomical traits: erect plants, high podded node, early flowering,
feature of the species are drought tolerance, resistance high seed yield, big and light seeds, high biomass, low
to pests and diseases, adaptability to different types of ODAP content and high protein content. All entries
soil as well as to adverse climatic conditions (9). were grown in 2002-2003 winter season on clay-sand
Despite these and other advantages, L. sativus is soil at the experimental farm “Pantanello”, belonging
inadequately exploited and studied. In fact, it is well to the Basilicata region, at Metaponto (Matera) in
known that this crop is grown mainly as landraces; southern Italy. Generally the climate in the Metaponto
their genetic diversity is used and maintained largely area (0-300m a.s.l.) is a strong Mediterranean type
by a small number of farmers in very limited areas of with an annual rainfall less than 600 mm and an
central southern Italy. In other words, valuable genetic annual temperature trend consisting of mild or absent
resources of L. sativus are exposed to the threat of winters and hot summers. Sowing was done in mid
genetic erosion and disappearance. Therefore November after a deep summer plowing and two
collection and storage of germplasm and deeper secondary tillages. During summer tillage 120 kg/ha
knowledge of the nature, entity, and geographical

10
 Lathyrus Lathyrism Newsletter 4 (2005)

of P2O5 was applied. Harvest occurred at the end of of correlation among traits for these entries. The first
June at full maturity stage. principal component accounted for 31% of variation
reflected mostly influence on pod and seed traits. The
Five randomly chosen plants from each entry from a second component accounts for 17% of the variance
single row plot were scored for the 18 quantitative and and thus is comparable in importance to the first. The
qualitative descriptors reported in Table 1. Frequency traits with the largest coefficients and which
distributions for the qualitative descriptors flower and contribute to it are the length of longest stem, biomass,
seed colour were also determined. Multivariate data seed yield and pedicel length. Time to flowering, leaf
analysis followed three steps: a) estimation of length, seed thickness, length of internode, leaf width,
standardized entry means of 16 quantitative traits; b) time to emergence and height of first podded node
derivation of orthogonal, uncorrelated traits for each have some importance in the other components.
entry using Principal Component Analysis (PCA); c) Although there is no clear demarcation between
clustering entries into similarity groups using important and unimportant principal components, it is
uncorrelated traits (PCA coefficients). The SAS interesting to note that yield and some yield
procedure PRINCOMP computed the correlation components appear strongly in the first two
matrix and determined the principal components (SAS components.
Institute, 1987). The sum of the first eight PC axes,
representing 88% of total variation were used in Clustering entries based on similarity of the first eight
subsequent analyses. Entries were clustered using principal components identified five large groups
Ward’s minimum-variance method (SAS Institute, accounting for a 31% share of variance. Cluster
1987). The cluster routine was stopped to form five memberships are reported in Table 3. Cluster I
discrete clusters looking for a consensus among the included the highest number of Italian entries; while a
four statistics R2 (RSQ), cubic clustering criterion large number of Cyprus entries were in cluster II.
(CCC), pseudo-F (PSF) and pseudo-t2 (PST2). Results Entries from the other less represented origins spread
of this clustering were combined with results of the over all five groups.
PCA analysis as a visual aid in discerning clusters.

Results and Discussion Table 1. Means, minimum and maximum values,

Means, minimum and maximum values, coefficients and coefficient of variation (C.V.) for 16
of variation for 16 quantitative traits and the frequency quantitative descriptors and frequency
distributions of the flower colour and seed colour are distributions for 2 qualitative descriptors observed
reported in Table 1. Entries showed a wide range of in 76 grass pea landraces.
variation as evidenced by coefficients of variation.
The most variable traits were seed yield, biomass, leaf Descriptora Mean Min Max C.V.
width and seeds/pod; and the lowest values of Time to emergence (d) 23.9 21.0 33.0 5.8
Time to flowering (d) 79.5 74.0 90.0 3.1
variation were estimated for time to flowering and Length of longest stem (cm) 74.0 25.0 98.0 18.9
time of emergence; all the other traits showed Height first podded node (cm) 22.2 8.0 48.0 24.1
intermediate values. The variability of means for yield Length of internode (cm) 3.9 2,0 9.0 24.6
components was lower than variability for seed yield. Leaf lenght (cm) 7.9 0.7 9.9 23.9
Leaf width (cm) 0.7 0.2 1.5 34.3
Extreme values for seed yield and biomass were 7.0- Pod lengthb (cm) 3.8 2.7 5.2 12.4
214.0 and 13.0-481.0g respectively. The distribution Pod widthb (cm) 1.3 0.8 2.0 17.7
in frequency classes for flower colour showed that Pedicel lengthb (cm) 5.0 1.7 8.5 23.0
nearly 51% of entries were characterized by violet Seeds/podb (no.) 2.6 1.0 5.0 33.8
Seed lengthc (cm) 0.8 0.4 1.5 25.0
flowers; while the prevalent seed colour was beige Seed widthc (cm) 0.8 0.4 1.3 23.0
(68.4%). Seed thicknessc (cm) 0.5 0.2 1.0 16.0
Seed yield (g) 82.4 7.0 214.0 48.7
The eigenvalues representing the variance of the Biomass (g) 201.7 13.0 481.0 44.0
Colour
principal components, and the cumulative percent of Flower colour White Violet Pink -
the eigenvalues indicating percentage contribution to Frequency (%) 37.6 51.1 11.3 -
the total variance attributable to each principal
component are given in Table 2. Eigenvectors Seed colour White Beige Brown Green
-grey
indicating the degree of association among original
Frequency (%) 3.9 68.4 5.5 7.9
data and each principal components are also reported. a
Data collected on single plant; b Average of 5 dry pods/plant;
The first two PC axes accounted for >48% of the c
Average of 5 seeds/ plant.
multivariate variation among entries and the first eight
axes >88% of variation, indicating a moderate degree

11
 Lathyrus Lathyrism Newsletter 4 (2005)

Table 2. Principal component analysis (PCA) of descriptors associated with 76 grass pea landraces showing
eigenvalues and proportion of variation associated with the first eight axes and eigenvectors of descriptors.

PC axis
1 2 3 4 5 6 7 8
Eigenvalues 5.01 2.61 1.67 1.38 1.06 0.98 0.71 0.62
Variation (%) com. 31 48 58 67 73 79 84 88
Descriptora Eigenvectors
Time to emergence (d) -0.12 0.15 0.40 0.27 0.17 0.44 -0.45 0.35
Time to flowering (d) -0.11 0.04 0.61 0.04 -0.07 0.09 0.31 0.05
Length of longest stem (cm) 0.17 0.45 -0.18 0.00 0.11 0.00 -0.03 -0.40
Height of 1° podded node (cm) 0.25 0.13 0.18 0.36 0.28 0.20 -0.16 -0.49
Length of internode (cm) 0.11 0.21 -0.23 0.23 0.64 -0.14 0.26 0.20
Leaf lenght (cm) 0.08 0.11 0.21 -0.52 0.20 0.42 0.52 -0.13
Leaf width (cm) 0.06 -0.05 -0.41 -0.34 -0.01 0.60 -0.31 0.03
Pod lengthb (cm) 0.38 -0.04 0.13 -0.15 -0.06 -0.10 -0.10 0.14
Pod widthb (cm) 0.42 -0.09 0.07 -0.06 0.00 -0.07 -0.05 0.01
Pedicel lengthb (cm) -0.03 0.42 -0.01 -0.31 0.24 -0.19 -0.11 0.50
Seeds/podb (n.) -0.33 0.31 -0.07 0.02 -0.08 0.08 0.04 -0.08
Seed lengthc (cm) 0.42 -0.11 0.06 -0.01 0.01 -0.01 -0.04 0.11
Seed widthc (cm) 0.41 -0.14 0.07 -0.00 0.04 0.02 0.03 0.13
Seed thicknessc (cm) 0.12 -0.05 -0.31 0.46 -0.21 0.38 0.44 0.32
Seed yield (g) 0.18 0.43 -0.02 0.12 -0.40 -0.01 0.11 0.07
Biomass (g) 0.20 0.44 0.06 -0.05 -0.40 -0.02 -0.08 0.01
a
Data collected on single plant; b Average of 5 dry pods/plant; c Average of 5 seeds/ plant.

Table 3. Cluster memberships: entry number and Table 4. Cluster means of 16 quantitative
geographical origin of 76 grass pea landraces. descriptors observed in 76 grass pea landraces.

Cluster I (n=19) Descriptor Cluster

Entry1 Origin2 Entry Origin Entry Origin I II III IV V
Time to emergence (days) 24 24 23 24 24
Time to flowering (days) 80 99 79 79 79
100263 ITA 112411 CYP 115243 ITA
Length of longest stem (cm) 67 74 75 79 76
100288 ESP 106531 AUS 113090 ITA Height of 1st podded node (cm) 20.7 19.2 21.8 25.0 25.0
115099 ITA 112252 ITA 100290 ITA Length of internode (cm) 3.7 3.9 3.7 3.9 4.5
100291 MAR 103641 ITA 112390 CYP Leaf length (cm) 7.3 8.3 11.0 8.0 7.5
115242 ITA 103203 ITA 113874 ITA Leaf width (cm) 0.7 0.7 0.8 0.8 0.6
115795 BGR 100041 Unk.3 100042 Unk.3 Pod length (cm) 4.0 3.5 3.9 4.0 4.0
100287 AUS Pod width (cm) 1.3 1.0 0.8 1.2 1.2
Cluster II (n=16) Pedicel length (cm) 4.8 5.4 5.1 4.9 5.0
Seeds/pod (n.) 2 3 2 3 2
106529 AUS 111982 ITA 112401 CYP Seed length (cm) 0.9 0.6 0.9 0.9 0.9
112414 CYP 112418 CYP 116171 ALB Seed width (cm) 0.9 0.6 0.8 0.8 0.9
112403 CYP 112399 CYP 112407 CYP Seed thickness (cm) 0.5 0.5 0.4 0.5 0.4
112415 CYP 112417 CYP 112419 CYP Seed yield (g) 83.5 81.9 73.1 94.9 70.0
112408 CYP 112412 CYP 116170 ALB Biomass (g) 195 197 206 228 163
112410 CYP
Cluster III (n=15) The mean values of the original traits for each cluster
110434 ITA 110435 ITA 110437 ITA are listed in Table 4. For some traits it was impossible
110492 ITA 115833 HUN 110955 ITA to clearly differentiate the phenotypic diversity among
109680 ESP 110262 ITA 111986 ITA
clusters, such as the time to emergence and seed
100289 RUS 111985 ITA 115097 ITA
115834 HUN 100043 Unk.3 103585 ETH thickness; while, by the other traits the clusters were
Cluster IV (n=17) better differentiated. In particular, means indicate
112400 CYP 112416 CYP 113873 ITA shorter plants with larger pods and seeds in cluster I.
115653 ITA 103244 ITA 103376 ITA Entries in cluster II with smaller seeds and shorter
103468 ETH 103579 ETH 113949 HUN pods flower nearly three weeks later. Highest yield
115093 ITA 115094 ITA 100293 Unk.3 and biomass means characterize entries in cluster IV,
103212 ITA 110957 ITA 112251 ITA which also showed taller plants and higher first
116250 ALB 113089 ITA podded node. On the contrary, cluster V grouped
Cluster V (n=9) entries with longer internode and lower mean values
100044 Unk. 112413 CYP 115096 ITA
for yield and biomass. Entries in cluster III had larger
100292 FRA 103237 ITA 106385 AUS
106434 ITA 106530 AUS 115241 ITA leaves; indicated by leaf length and width.
1
Mediterranean Germplasm number (MG); 2ISO Country
code ; 3Unknown

12
 Lathyrus Lathyrism Newsletter 4 (2005)

Plot of PCAs and clusters using the first two axes, PRIN1 and PRIN2

-1

-2

-3
-3 -2 -1 0 1 2 3 4

Cluster 1 Cluster 2 Cluster 3 Cluster 4 Cluster 5

Fig 1. Plot of principal component analysis and clusters using the first two axes PRIN1 and PRIN2. Each
cluster is represented by a different symbol.

A combined spatial distribution of entries and clusters production (for example MG 113089, MG 112416) or
can be represented in two-dimensional scatter to use them in future selection programs. As reported
diagrams as shown in Figure 1. in previous experiences (1,14) grouping germplasm
entries into morphologically similar and presumably
Conclusions genetically similar groups is useful when little is
The phenotypic diversity among grass pea entries was known about the crop history and the population
well defined by both Principal Component and Cluster structure; as is the case for the grass pea collection. It
analyses. Considering the different morpho-bio- is evident that entries clustered together are more alike
agronomic descriptors, it has been possible to observe than entries from other clusters. However, the data
a remarkable inter and intra-group diversity. The must be considered with caution because they are an
covariation structure in the material studied revealed a expression of linked genetic and environmental
different association between traits. The traits with effects. So, it is important to emphasise that the groups
dominant roles in the first two components are closely were defined on the basis of results from in a single
related to yield and yield components; while, location of southern Italy. The clusters could be
vegetative traits like flowering time, height of first different if the examination took place elsewhere.
podded node and leaf traits were separately linked to
other components. This suggests the possibility of There seems to be no significant differences in
obtaining, though selection, suitable genotypes relation to the origin of entries, most of which were
combining high yield with desirable traits for direct from Italy and distributed in all groups. However, the
release as cultivars in marginal areas of southern Italy. entries from Cyprus with some exception were found
to form a distinct group (II), which was characterized
Cluster analysis also helped us to differentiate entries by entries showing smaller pods and seeds, longer
on the ground of their different levels of similarity. pedicels and later flowering time. A more detailed
Five groups were identified with clear-cut differences geographical differentiation was impossible with
according to the first principal component, which many origins underrepresented, with the exception of
mostly accounted for yield and yield components. Italy and Cyprus. In fact, at the level of five clusters,
Smaller differences among groups were seen the proportion of variance accounted for by the
according to the second principal component, which clusters is 31%. This is a low percentage for the
accounted for vegetative traits. Compared with other variance explained by the identified groups. Thus the
groups, group IV show highest mean values for yield, differentiation according to these clusters can only be
biomass, seed size and height of first podded node, considered as a preliminary approach until more
which are useful agronomic traits to breed new grass detailed analysis and information is available. Finally,
pea varieties. With the exception of group V, which the observed wide diversity in the Italian grass pea
included the less productive entries, groups I, II and entries distributed in all groups suggests the use of this
III were moderately similar. Among the entries tested adapted material to breed new improved grass pea
the analysis provides useful information in order to varieties.
utilise directly the most promising materials for

13
 Lathyrus Lathyrism Newsletter 4 (2005)

References
1. Alba E, Polignano GB, Notarnicola L. 1997. 9. Noto F, Poma I, Gristina L, Venezia G, Ferrotti
Diversity analysis in some Amaranthus entries. F. 2001. Bioagronomic and qualitative
Agr. Med. 127, 198-204. characteristics in Lathyrus sativus lines. In:
2. Bisignano V, Della Gatta C, Polignano GB. Proceedings 4th European Conference on Grain
2002. Variation for protein content and seed Legumes (eds. AEP), 8-12 July 2001, Cracow,
weight in grass pea (Lathyrus spp.) germplasm. Poland. P 183.
Plant Genetic Resources Newsletter 132, 30-34. 10. Polignano GB, Uggenti P, Perrino P. 2001.
3. Eyzaguirre PB, Paludosi S, Hodgkin T. 1999. Phenotypic diversity in Bari grass pea (Lathyrus
IPGRI’s strategy for neglected and underutilized spp.) collection. In: Proceedings 4th European
species and the human dimension of Conference on Grain Legumes (eds. AEP), 8-12
agrobiodiversity. In: Priority-setting for July 2001, Cracow, Poland. Pp 184-185.
underutilized and neglected plant species of the 11. Polignano GB, Uggenti P, Bisignano V, Alba E.
Mediterranean region (Paludosi S, ed.). Report of 2003. Patterns of variation in Lathyrus sativus and
the IPGRI Conference, 9-11 February 1998, some related species. Agr. Med. 133, 1-9
ICARDA, Aleppo, Syria. IPGRI, Rome, Italy. Pp 12. Sarker A, Robertson D, Campbell CG. 2000.
1-20. Lathyrus spp.: Conserved Resources, Priorities
4. Enneking D, M. Wink. 2000. Towards the for collection and future prospects. In: Linking
elimination of anti-nutritional factors in grain Research and Marketing Opportunities for Pulses
legumes. In: Linking Research and Marketing in the 21st Century (R. Knight, ed.), Kluwer. Pp
Opportunities for Pulses in the 21st Century 645-654.
(Knight R, ed.), Kluwer. Pp 671-681. 13. SAS/STAT Guide for Personal Computers.
5. Erskine W, Smartt J, Muehlbauer FJ. 1994. Version 6. SAS Institute Inc., Cary, NC USA,
Mimicry of lentil and the domestication of 1987.
common vetch and grass pea. Economic Botany 14. Souza E, Sorrells ME. 1991. Relationships
48, 326-332. among 70 North American Oat Germplasm: I.
6. Granati E, Bisignano V, Chiaretti D, Polignano Cluster analysis using quantitative characters.
GB, Crinò P. 2002. Characterization of Italian and Crop Sci 31, 599-605.
exotic Lathyrus germplasm for quality traits. 15. Zeven AC, Zhukovsky PM. 1975. Dictionary of
Genetic Resources and Crop Evolution 50, 273- cultivated plants and their centres of diversity.
280. Centre for Agricultural Publication And
7. Kupicha FK, 1983. The infrageneric structure of Documentation, Wageningen, The Netherlands.
Lathyrus. Notes from the Royal Botanic Garden 16. Zohary D, Hopf M. 1988. Domestication of
Edinburgh 41, 287-326. Plants in the Old World. Clarendon, Oxford.
8. Hanbury CD, White CL, Mullan BP, Siddique
KHM. 2000. A review of the potential of
Lathyrus sativus L. and L. cicera L. grain for use
as animal feed. Animal Feed Science and
Technology 87, 1-27.

14