Directory UMM :Data Elmu:jurnal:A:Animal Feed Science and Technology:Vol87.Issue1-2.Sept2000:

Animal Feed Science and Technology
87 (2000) 1±27

Review article

A review of the potential of Lathyrus sativus
L. and L. cicera L. grain for use as animal feed
C.D. Hanburya,*, C.L. Whiteb, B.P. Mullanc, K.H.M. Siddiquea,c
a

Centre for Legumes in Mediterranean Agriculture (CLIMA), University of Western Australia,
Nedlands 6907, Australia
b
CSIRO, Division of Animal Production, Private Bag, PO Wembley 6014, Australia
c
Agriculture Western Australia, Locked Bag No. 4, Bentley Delivery Centre, Bentley 6983, Australia
Received 28 December 1999; received in revised form 3 July 2000; accepted 19 July 2000

Abstract
The use of two closely related species, Lathyrus cicera and L. sativus, as grain legumes for
human and animal consumption, dates to the Neolithic period. Due to its tolerance to harsh

environmental conditions L. sativus is still used widely for human food in Ethiopia and the Indian
sub-continent, although cultivation has diminished in many other regions.
The grain of both L. cicera and L. sativus contains a neurotoxin, 3-(-N-oxalyl)-L-2,3-diamino
propionic acid (ODAP), which can cause a paralysis of the lower limbs (lathyrism). Due to the
occurrence of lathyrism in humans recent plant breeding has produced cultivars with low ODAP
concentrations. The susceptibility of animal species to lathyrism is poorly understood, although
horses and young animals are more susceptible. Older published animal feeding studies are of
limited use, since the presence and role of ODAP was unknown until the 1960s. More recent
feeding studies indicate that low ODAP lines of L. cicera or L. sativus can be safely incorporated at
inclusion rates up to 40, 30 and 70% of the diet of poultry, pigs and sheep, respectively, without
growth reductions.
The compositions of both L. cicera and L. sativus are similar to other commonly used feed grain
legumes, respective protein contents are 25 and 27%. Antinutritional factors (ANFs), other than
ODAP, are present in both L. cicera and L. sativus at concentrations similar to those found in other
grain legumes; including trypsin inhibitors, chymotrypsin inhibitors, amylase inhibitors, lectins,
tannins, phytate and oligosaccharides. The effect of ANFs in L. cicera and L. sativus on animal
performance is not well understood and sometimes confounded with ODAP effects. Heating of
grain will reduce levels of the proteinaceous ANFs and in some cases ODAP as well.
Variation recorded in the germplasm of L. cicera and L. sativus has not been greatly utilised in
plant breeding to lower levels of ANFs, with the exception of ODAP, leaving considerable potential

for rapid improvement of cultivars. L. cicera and L. sativus are low production cost legumes
*

Corresponding author. Tel.: ‡61-8-9368-3744; fax: ‡61-8-9368-2165
E-mail address: chanbury@agric.wa.gov.au (C.D. Hanbury).
0377-8401/00/$ ± see front matter # 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 3 7 7 - 8 4 0 1 ( 0 0 ) 0 0 1 8 6 - 3

2

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

adapted to low rainfall environments and have considerable potential as good quality, cheap protein
sources. As world demand for legume feed protein is likely to increase, due to increasing demand
for animal food products, both L. cicera and L. sativus are crops that should be considered in
regions with suitable environments. # 2000 Elsevier Science B.V. All rights reserved.
Keywords: Lathyrus; Protein; ODAP; Antinutritional factors; Lathyrism

1. Introduction
The use of legumes as sources of protein for the animal feed industry is expected to

increase further in the near future. Rising incomes in the Asian region are increasing the
demand for meat products, and hence the requirement for animal feeds. There have been
changes in public perception and some unfortunate developments, such as the
consequences of `mad cow' disease (i.e. bovine spongiform encephalopathy or BSE) in
UK. This has resulted in many feed compounders either choosing to, or being banned
from, using animal by-products as a source of protein (Farrell, 1997). The amino acid
requirement of animals often differ with species and bodyweight, hence no single source
of plant protein will provide the exact amino acids required for all animals. It is,
therefore, preferable to include a range of protein sources in diet formulations, each
complementing the other. For these reasons, the demand for grain legumes, such as
Lathyrus spp., by the feed industry is expected to increase. Any feedstuff is likely to be
used in diets for animals if it supplies the required nutrients, if it is cost competitive with
other available ingredients, and if the user is con®dent it will produce the desired result.
Soybean (Glycine max) meal is used widely as a source of protein for animal feeds, and
the price of most other protein meals and grain legumes are set relative to this
commodity. In Europe there has been increasing emphasis on local production of legumes
for animal feed in order to supply some of this protein demand (Gatel, 1994), rather than
relying on imported soybean meal. Subsidies have resulted in increasing production of,
particularly, ®eld peas (Pisum sativum) and faba beans (Vicia faba). This expansion has
been partly at the expense of previously grown legume crops, such as Lathyrus spp.,

which do not have subsidies (Franco Jubete, 1991).
It has been demonstrated in recent studies that two Lathyrus spp. (L. sativus L. and L.
cicera L.) have considerable potential as grain legume crops on ®ne textured, neutral to
alkaline, soil types in southern Australian Mediterranean-type environments (Hanbury
et al., 1995; Siddique et al., 1996, 1999; Siddique and Hanbury, 1998). In Australia the
adaptation of the two species is slightly different, L. cicera seeming better adapted to
lower rainfall (250±350 mm annually) regions and L. sativus to medium rainfall (350±
600 mm annually). They are being evaluated as low input, multi-purpose crops for green
manuring, animal feed and fodder. In Australia, animal feed is seen as the primary use of
the grain of these species, both on and off farm. Animal feeding with both species has a
long history and is still practised in some parts of the world. However, published studies
are scattered and frequently dif®cult to access. This review aims to summarise the animal
feeding literature and demonstrate the potential for the two species as animal feed.

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

3

The genus Lathyrus is a large one, comprising 189 species and sub-species according
to Allkin et al. (1985), and approximately 150 species according to Kupicha (1983). Of

these only a small number are cultivated. The closely related L. sativus (grass pea) and L.
cicera (dwarf chickling) both belong to the same section: Lathyrus. Jackson and Yunus
(1984) suggest that the similarities between semi-domesticated L. cicera and
domesticated L. sativus may be a result of hybridisation or common ancestry. Some
interspeci®c crosses between the two have been successful (Yunus and Jackson, 1991).
Archaeobotanical evidence shows that both L. sativus and L. cicera were cultivated on the
Iberian peninsula in the Neolithic period (PenÄa-Chocarro and Zapata PenÄa, 1999).
Evidence also suggests that L. sativus is possibly the most ancient domesticated crop in
Europe, the Neolithic expansion of its cultivation into what is now Spain led to the
cultivation of a local native species, L. cicera (Kislev, 1989). Erskine et al. (1994) suggest
that L. sativus was originally domesticated as a secondary crop as a result of being a weed
of lentil (Lens culinaris) crops.

2. Lathyrism
Lathyrus species, particularly L. sativus, have been known since classical times to be
implicated in a paralysis of humans and animals (Hugon et al., 2000) known as
``lathyrism'' or more speci®cally ``neurolathyrism''. Both ruminants and monogastric
species can be affected, some literature indicates that monogastrics can be more affected.
It was only in the later half of the 20th century that the compound responsible was
identi®ed (Murti et al., 1964; Rao et al., 1964).

There are two forms of lathyrism, neurolathyrism and osteolathyrism. Osteolathyrism
is characterised by skeletal deformities and can be caused by consumption of the species
L. odoratus (sweet pea), L. hirsutus, L. pusillus and L. roseus (Roy, 1981). Osteolathyrism
has been recorded experimentally in a wide range of animals (Barrow et al., 1974). The
principal compound responsible was found to be b-aminopropionitrile (BAPN; Fig. 1),
although the related nitriles aminoacetonitrile (AAN) and methylene aminoacetonitrile
(MAAN) also have some osteolathyritic activity (Barrow et al., 1974). Although BAPN is
not found in either L. sativus or L. cicera (Bell, 1962, 1964), there is evidence that a
BAPN precursor (2-cyanoethyl-isoxazolin-5-one) is present in L. sativus seedlings but not
in seed (Lambein et al., 1993). Consumption of L. sativus seedlings and shoots as
vegetables has been blamed as the cause of osteolathyritic symptoms found in a small
proportion of people with chronic neurolathyrism (Haque et al., 1997). Incidents of
osteolathyrism from feeding of L. sativus or L. cicera have not been reported in animal
studies, either under natural grazing or experimental conditions, and consequently the
following discussion will focus on neurolathyrism.
Neurolathyrism is the term used to describe the symptoms shown after heavy
consumption of several different Lathyrus species and some Vicia species. The symptoms
are weakness of the hind limbs and paralysis or rigidity of the muscles. Within the
Lathyrus genus, the category of neurolathyrism has been further divided into two subcategories. One is caused by the compound L-2,4-diaminobutyric acid (DABA; Fig. 1),
primarily in the perennial species L. sylvestris (Foster, 1990) and L. latifolius (Barrow

4

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

Fig. 1. Chemical diagrams of b-aminopropionitrile (BAPN), L-2,4-diaminobutyric acid (DABA), 3-(-N-oxalyl)propionic acid (ODAP or BOAA) and glutamate.

L-2,3-diamino

et al., 1974). However, DABA is not found in L. cicera or L. sativus (Bell, 1962, 1964;
Padmanaban, 1980). The form of neurolathyrism most pertinent to this discussion is that
caused by the non-protein amino acid 3-(-N-oxalyl)-L-2,3-diamino propionic acid (ODAP,
also referred to as b-N-oxalylamino-L-alanine or BOAA; Fig. 1): which has been recorded
in humans and animals following consumption of L. sativus, L. cicera, L. ochrus and L.
clymenum (Barrow et al., 1974; Padmanaban, 1980; Franco Jubete, 1991). The seed of a
number of other uncultivated Lathyrus species have been found to contain ODAP (Bell,
1962, 1964). Historically the consumption of L. sativus has been most often linked with
lathyrism in humans and animals, primarily because of all Lathyrus species it is the most
widely utilised as grain and fodder. Lathyrism is the term mostly used to refer to
neurolathyrism caused by ODAP, therefore this term will be used in this review from this

point onward.
Lathyrism in humans has received more attention than that in animals, due to the social
cost. Symptoms in humans are most often initial painful spasms in the muscles of the
lower limbs with accompanying weakness, followed by chronic spastic paraplegia of
various degrees (Spencer et al., 1986), and can lead to total loss of use of the legs (Attal
et al., 1978). The paralysis is rarely reversible (and then only in early stages of the
symptoms; Hugon et al., 2000) and the consequences for poor communities who depend
upon L. sativus as a primary food source at times of food scarcity can be devastating.
Lathyrism still occurs, with a 1997 outbreak during food shortages in Ethiopia crippling
2000 people (Getahun et al., 1999). Lathyrism is endemic to the areas of the world which
have signi®cant areas of L. sativus cultivation; India, Bangladesh, Ethiopia and Nepal.
However, in the 20th century outbreaks were also reported in Afghanistan, Algeria,
China, France, Germany, Italy, Pakistan, Romania, Russia, Spain and Syria (Trabaud and
MouhaÈrram, 1932; Barrow et al., 1974; Roy and Spencer, 1989; Hugon et al., 2000).

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

5

Boiling has been found to reduce ODAP levels in several cases, however, there are mixed

reports on other forms of cooking (Tekle Haimanot et al., 1993; Akalu et al., 1998).
Padmajaprasad et al. (1997) reported that boiling grain and discarding the water reduced
ODAP levels by up to 90%.

3. ODAP toxicity
Following its isolation and identi®cation (Murti et al., 1964; Rao et al., 1964) the
neurolathyritic action of ODAP was soon demonstrated in adult monkeys (Macaca
radiata; Rao et al., 1967). Cheema et al. (1969) administered ODAP intraperitoneally to
rats. Young rats showed lathyrism symptoms and had 0.11 mmol gÿ1 ODAP in the brain,
adult rats showing trace or nil ODAP and no symptoms. Olney et al. (1976) found some
indication of exclusion of ODAP by the blood-brain barrier in mice. Padmanaban (1980)
suggested that the hypothesis of less ODAP exclusion by the blood-brain barrier in young
animals should be re-examined, as greater excretion of ODAP by older animals may be
an important factor. Spencer et al. (1986) showed unequivocally that ODAP, either
naturally present in L. sativus or when added to other food sources, was the cause of
corticospinal dysfunction in monkeys (Macaca fascicularis), with symptoms of hind limb
motor dif®culties.
ODAP acts as a glutamate (Fig. 1) analogue in the nervous system and probably acts by
binding strongly to a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)type glutamate receptors. Permanent damage probably occurs with excitotoxic
degeneration of neurons, although there are other possible neurotoxic effects. The

ultimate fate of ODAP and the distribution in the brain and spinal cord is not known
(Hugon et al., 2000). ODAP was not detected in pig loin tissue following feeding for an
extended period (Castell et al., 1994; see Section 4.3).
In human populations young men are widely reported as the most susceptible to
lathyrism (McCarrison, 1926; Shourie, 1945; Attal et al., 1978; Hamid et al., 1986;
Getahun and Tekle Haimanot, 1997), although the reasons for this are not understood
(Hugon et al., 2000). The production and susceptibility to ODAP may be linked to Zn
de®ciency in plants and humans, respectively (Lambein et al., 1994; Lambein and Kuo,
1997). ODAP has also been found to inhibit growth of some insects and yeasts (Rao et al.,
1964; Mehta et al., 1972), and so may have a plant protective role.

4. The nutritive value of Lathyrus for animals
4.1. Chemical composition of the seed
Note: All concentrations in grain are expressed as received unless speci®ed otherwise.
4.1.1. Proximate composition
The proximate compositions of L. cicera and L. sativus are generally very similar to
®eld pea and faba bean (Table 1). Both Lathyrus spp. have low fat and high starch

6

L. cicera
i

iii

iv

1

4

1

Component (% DM)
Protein
21.7
Ash
2.9
Fat
1.4
Crude ®bre
7.3
ADF
10.7
NDF
22.1
Lignin
0.6
Starch
44.2

27.2
3.1
0.7
6.7
10.6
24.3
0.2
±

33.0
3.8
±
±
10.7
17.8
±
±

Dry matter (% ar) ±

89.7

90.3

No. lines

a

ii

L. sativus

1

Mean

iii

iv

±

8

1

26.4
3.1
±
±
11.0
18.2
±
±

29.6
3.5
1.1
7.0
10.7
19.4
0.4
44.2

34.3
3.9
±
±
9.0
15.5
±
±

89.5

90.1

90.9

v

vi

vii

viii

ix

3

1

3

1

1

30.1
3.1
±
±
12.2
16.0
±
±

30.9
3.3
0.9
±
±
±
±
±

26.4
2.8
1.7
6.0
±
±
±
±

32.6
2.6
5.3
8.3
±
±
±
±

26.3
3.2
0.7
5.5
±
±
±
±

31.3
3.1
1.0
10.0
±
±
±
±

89.0

90.0

±

±

90.0

87.6

x

Faba bean

Lupin

i

i

i

xi

xii

1

1

±

1

27.3
2.0
1.4
8.3
±
±
±
41.2

35.9
2.7
1.2
5.3
±
±
1.5
±

26.9
2.9
0.8
5.9
8.3
±
0.8
±

29.4
2.6
1.6
8.0
9.3
15.6
1.2
41.2

21.0
3.3
1.7
7.2
8.1
14.6
1.1
45.3

25.7
2.8
1.2
6.6
10.3
14.7
0.6
±

91.9

±

91.1

91.3

±

90.2

25

Mean

Field pea
xiii
24±3788

xiii
1

xiii

5±355 1

111±3782

23.7
3.8
1.4
10.0
13.1
20.2
2.4
40.0

26.9
3.0
1.4
9.4
11.0
14.3
±
±

29.1
2.6
7.2
16.1
22.9
26.9
2.8
0.8

35.1
3.0
6.5
16.8
21.6
25.8
0.9
±

±

89.7

±

91.1

Source: (i) Abreu and Bruno-Soares (1998), (ii) Hanbury (unpublished), (iii) Aletor et al. (1994), (iv) Farhangi (1996), (v) Adsule et al. (1989), (vi) Infascelli et al.
(1995), (vii) Shobhana et al. (1976), (viii) Dhiman et al. (1983), (ix) Latif et al. (1975), (x) Urga et al. (1995), (xi) Kuo et al. (1995), (xii) Low et al. (1990), (xiii) Mean
values from Petterson et al. (1997).

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

Table 1
Composition of L. cicera and L. sativus in comparison to ®eld pea (P. sativum), faba bean (V. faba) and lupin (L. angustifolius)a

7

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

Table 2
Fatty acid composition of L. cicera and L. sativus (compared to ®eld pea (P. sativum), faba bean (V. faba) and
lupin (L. angustifolius))a
L. cicera
i
No. lines

L. sativus
ii

2

ii

2

3

Fatty acid (% of total fats)
Mystiric
±
Palmitic
15.4
Palmitoleic
±
Stearic
7.9
Oleic
12.1
Linoleic
46.2
Non-adecanoic
0.8
Linolenic
8.6
Arachidic
2.4
Eicosadienoic
2.7
Behenic
±

0.4
5.3
0.3
18.7
57.7
12.1
±
0.7
1.2
±
0.9

0.6
8.1
0.4
13.8
58.3
14.1
±
1.1
0.5
±
0.4

Total

97.3

97.3

96.1

iii
1

iv

v

Field
pea
vi

Faba
bean
vi

Lupin
vi

1

1

5±16

3±8

3±174

±
25
±
2
1
67
±
3
±
±
Trace

0.8
14.8
0.3
7.5
16.7
56.0
±
2.2
±
±
±

0.5
16.8
±
4.6
18.6
38.9
±
8.0
±
±
±

0.3
12.5
±
1.2
25.1
42.3
±
9.7
0.7
±
0.3

0.5
14.0
±
2.3
21.0
45.0
±
4.7
1.8
±
0.9

0.1
11.0
0.1
3.7
33.5
37.1
±
5.3
0.9
0.4
1.9

98

98.3

87.4

91.8

89.3

92.1

a

Source: (i) Hanbury (unpublished), (ii) Senatore and Basso (1994), (iii) Choudhury and Rahman (1973),
(iv) Kuo et al. (1995), (v) Grela and GuÈnter (1995), (vi) Mean values from Petterson et al. (1997).

contents, similar to ®eld pea and faba bean, conversely lupin has high fat and low starch
content. The composition of the lipid fraction (Table 2) shows in most cases that the fatty
acid pro®le is similar to other grain legumes, concentrations of stearic and linoleic acid
are a little higher and oleic acid slightly lower. The data of Senatore and Basso (1994)
differ from other data in both L. cicera and L. sativus, they found levels of oleic acid
considerably higher and levels of linoleic acid considerably lower than all other reports.
Whether this difference is due to the methods of Senatore and Basso (1994) or a
difference in the Italian Lathyrus ecotypes examined is unclear, their results also differ
from those for other grain legumes species (Table 2).
4.1.2. Mineral content
The data on L. cicera mineral content are more complete than those for L. sativus
(Table 3). Mineral contents of both species are similar and compare to other agriculturally
important grain legumes. On the basis of the available data it does not appear that either
species will be markedly different to the other grain legumes widely used.
4.1.3. Protein content and quality
The mean protein content in L. cicera and L. sativus is 25 and 27%, respectively, from
samples across a wide range of locations (Table 4). These are higher than protein contents
in ®eld pea (23%) or faba bean (24%), but lower than in lupin (32%) (Petterson et al.,
1997) or soybean (42%; Ravindran and Blair, 1992). Chandna and Matta (1994) found
the composition of seed protein in L. sativus to be: albumins (14%), globulins (66%),

8

L.cicera
i
No. of lines

ii
2

Mineral (mg/kg)
Se
0.12
Cu
7.6
Fe
70
Mn
11
Zn
22
B
9

Field pea Faba bean Lupin

L. sativus

1
±
5.6
78
12
15
14

Mean
±
0.12
6.9
73
12
20
11

ii

iii

iv

v

vi

vii

1

1

25

1

1

3

±
8.2
38
15
27
11

±
±
±
±
±
±

±
±
95
±
±
±

±
±
±
±
±
±

±
7.7
63
±
±
±

±
±
74
±
±
±

Mean

viii

viii

viii

±

5±84

4±23

355±677

±
8.0
89
15
27
11

0.07
4.8
53
14
30
±

0.05
10.3
77
30
28
±

0.08
4.9
75
17
35
±

Mineral (%)
P
K
Na
Ca
Mg
S

0.30
0.88
0.07
0.16
0.12
0.16

0.27
0.86
0.06
0.27
0.13
0.17

0.29
0.87
0.07
0.20
0.13
0.17

0.34
±
0.02
0.12
0.12
0.18

0.31
±
±
0.14
±
±

0.44
±
±
0.16
±
±

0.26
±
±
0.28
0.11
±

0.32
0.64
0.04
0.09
0.09
0.14

0.41
±
±
0.18
±
±

0.42
0.64
0.03
0.16
0.11
0.16

0.34
0.91
0.01
0.07
0.12
0.18

0.41
0.96
0.01
0.12
0.10
0.13

0.30
0.81
0.05
0.22
0.16
0.23

Ca:P

0.53

1.0

0.69

0.53

0.45

0.36

1.1

0.28

0.43

0.38

0.21

0.29

0.73

a
Source: (i) Hanbury (unpublished), (ii) Farhangi (1996), (iii) Low et al. (1990), (iv) Urga et al. (1995), (v) Latif et al. (1975), (vi) Adsule et al. (1989), (vii)
Shobhana et al. (1976), (viii) Mean values from Petterson et al. (1997).

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

Table 3
Mineral content (as received) of L. cicera and L. sativus (includes comparison with ®eld pea (P. sativum), faba bean (V. faba) and lupin (L. angustifolius))a

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

9

Table 4
Protein concentrations (as received) reported in L. cicera and L. sativus
Species

Mean protein (%)

No. lines

Location

Source

L. cicera

25
26
23
24
27

128
51
20
17
16

Spain
±
Australia
Australia
Syria

Franco Jubete (1991)
Petterson et al. (1997)
Laurence (1979)
Hanbury (unpublished)
Aletor et al. (1994)

L. sativus

24
28
29
26
31
25
30
25
25
27
29

114
76
41
40
36
25
15
12
10
3
3

Bangladesh
Chile
±
Australia
Syria
Ethiopia
India
Australia
Spain
Canada
India

Kaul et al. (1982)
Tay et al. (2000)
Petterson et al. (1997)
Laurence (1979)
Aletor et al. (1994)
Urga et al. (1995)
Ramanujam et al. (1980)
Hanbury (unpublished)
Franco Jubete (1991)
Rotter et al. (1991)
Shobhana et al. (1976)

glutelins (15%) and prolamins (5%); similarly; Duke (1981) also quotes values of 26, 53,
15 and 6%, respectively.
The amino acid pro®les of L. cicera and L. sativus (Table 5) are similar to those
reported for many grain legumes (Petterson et al., 1997; Ravindran and Blair, 1992). For
monogastric species most grain legumes are de®cient in the sulphur containing amino
acids (methionine and cystine) but are rich in lysine (Gatel, 1994; Ravindran and Blair,
1992), this is also the case in L. cicera and L. sativus. In mixed diets, grain legumes,
therefore, complement cereals, which have higher levels of methionine and cystine but
lower levels of lysine. The mean lysine concentration in L. cicera is slightly lower than in
L. sativus, 6.16 (n ˆ 3) cf. 6.8 g/16 g N (n ˆ 12). Lysine contents per 16 g N are 5%
lower than in ®eld pea, and 30% higher than in lupins (Table 5). On the basis of the
protein composition both Lathyrus spp. have similar application to other legumes used as
animal feed.
Little information is published on L. sativus and L. cicera amino acid availabilities in
monogastric and protein degradability in ruminant species. The protein degradability
estimates from in sacco studies in three ruminant species (Table 6) showed both L. sativus
and L. cicera to be similar to ®eld pea and faba bean. Protein degradabilities in both
Lathyrus spp. were usually slightly greater than lupin; and usually slightly less than in
faba bean or ®eld pea.
4.1.4. Energy
For both L. cicera and L. sativus, measures of energy are similar to those for many
other common feed grain legumes. However, L. cicera and L. sativus have consistently
lower gross energy (GE) than lupin (which has a much higher fat content, Table 1) but are
similar to ®eld pea and faba bean (Table 7).

10
Table 5
Amino acid concentrations (g/16 g N) and protein (as received) in L. cicera and L. sativus (includes comparison with ®eld pea (P. sativum), faba bean (V. faba) and lupin
(L. angustifolius))a

i
No. of lines

L. sativus
ii

2

1

Mean
±

ii
1

iii

iv

1

3

Amino acids (g/16 g N)
Cystine
1.26
Aspartic acid
9.54
Methionine
0.75
Threonine
3.31
Serine
4.58
Glutamic acid
16.26
Proline
4.10
Glycine
3.74
Alanine
3.67
Valine
4.30
Isoleucine
3.68
Leucine
6.50
Tyrosine
2.93
Phenylalanine
4.11
Lysine
5.98
Histidine
2.18
Arginine
7.90

±
11.81
±
3.99
4.95
17.46
±
4.00
4.31
4.66
4.08
6.53
±
±
6.52
±
7.96

1.26
10.30
0.75
3.54
4.70
16.66
4.10
3.83
3.88
4.42
3.82
6.51
2.93
4.11
6.16
2.18
7.92

±
10.45
±
3.55
4.75
16.37
±
3.43
3.62
4.00
3.69
5.76
±
±
5.37
±
7.28

1.39
11.8
0.82
4.08
4.73
17.43
4.00
4.20
4.53
4.90
4.41
6.90
2.45
4.49
6.73
2.61
8.04

Protein (% ar)

23.6

25.7

26.8

24.5

a

26.8

1.53
±
1.00
4.04
±
±
±
±
±
±
±
±
±
±
7.10
±
±
26.9

v
1

vi

vii

viii

4

1

1

Mean
±

Field pea

Faba bean Lupin

ix

ix

ix

37

6

30

±
8.53
0.24
2.59
±
13.40
3.07
3.45
3.20
3.91
3.41
5.93
2.39
3.26
4.08
2.82
6.13

1.2
±
0.6
2.6
±
±
±
±
±
4.4
5.0
6.6
±
4.2
7.0
2.5
8.0

±
14.6
0.61
5.15
5.08
17.47
4.42
3.91
2.19
5.08
4.82
8.60
2.92
3.89
6.27
3.47
6.11

±
9.97
0.59
3.82
4.40
13.99
3.50
3.91
3.92
5.88
3.89
6.42
1.44
2.95
9.65
2.70
3.29

1.4
11.07
0.7
3.5
4.74
15.73
3.75
3.78
3.49
4.6
4.5
6.7
2.30
3.9
6.8
2.7
7.0

1.49
10.16
0.85
3.35
4.13
15.88
4.24
4.13
4.00
4.29
3.89
6.54
2.87
4.17
6.81
2.37
10.04

1.37
10.53
0.78
3.54
5.04
16.03
3.82
4.20
4.17
4.30
3.80
7.27
3.39
4.12
6.29
2.54
9.46

1.48
9.29
0.72
3.36
4.85
20.77
4.28
4.12
3.19
3.91
3.97
6.61
3.46
3.65
4.66
2.41
12.03

27.4

±

32.3b

25.6

27.2

23.0

24.1

32.2

Source: (i) Hanbury (unpublished), (ii) Farhangi (1996), (iii) Low et al. (1990), (iv) Rotter et al. (1991), (v) Latif et al. (1975), (vi) Adsule et al. (1989), (vii) Kuo
et al. (1995), (viii) Ronda Lain et al. (1963), (ix) Mean values from Petterson et al. (1997).
b
Estimated as 0.90% DM.

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

L. cicera

11

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

Table 6
In sacco degradability parameters for protein and dry matter in L. cicera and L. sativus grain fed to ruminant
species (compared to ®eld pea (P. sativum), faba bean (V. faba) and lupin (L. angustifolius))a
Animal

Protein

Cattleb

Cattlec

Buffalod

Sheepd

Dry matter

Cattleb

Cattlec

Buffalod

Sheepd

Feed grain

Parameter

Degradability
(at r ˆ 0:05)

Soluble
fraction (a)

Potentially
degradable
fraction (b)

Rate of
degradation
(c)

L. cicera
Field pea
Faba bean
Lupin
L. cicera
Field pea
Faba bean
Lupin
L. sativus
Field pea
Faba bean
Lupin
L. sativus
Field pea
Faba bean
Lupin

0.53
0.56
0.59
0.51
0.39
0.52
±
0.38
0.52
±
0.79
0.34
0.62
±
0.70
0.52

0.46
0.44
0.41
0.50
0.61
0.48
±
0.62
0.48
±
0.18
0.66
0.37
±
0.28
0.46

0.36
0.35
0.39
0.22
0.20
0.18
±
0.17
0.17
±
0.10
0.17
0.14
±
0.11
0.12

0.93
0.95
0.95
0.92
0.78
0.87
±
0.84
0.88
±
0.91
0.84
0.89
±
0.89
0.84

L. cicera
Field pea
Faba bean
Lupin
L. cicera
Field pea
Faba bean
Lupin
L. sativus
Field pea
Faba bean
Lupin
L. sativus
Field pea
Faba bean
Lupin

0.45
0.50
0.47
0.37
0.34
0.41
±
0.33
0.43
±
0.6
0.17
0.59
±
0.64
0.14

0.48
0.49
0.46
0.61
0.62
0.59
±
0.67
0.54
±
0.34
0.80
0.37
±
0.30
0.83

0.25
0.22
0.29
0.13
0.16
0.13
±
0.14
0.12
±
0.09
0.18
0.07
±
0.16
0.14

0.85
0.90
0.86
0.81
0.72
0.80
±
0.80
0.81
±
0.82
0.80
0.81
±
0.81
0.75

a
The parameters describe the non-linear equation: fractional loss ˆ a ‡ b…1 ÿ eÿct †; t is the time (h).
Degradability is determined at 0.05 fractional rumen out¯ow rate (r) according to:
degradability ˆ a ‡ bc=…c ‡ r†, where a is the water soluble fraction, b the potentially degradable fraction, c
the rate of loss of b and t is the time (h).
b
White (unpublished).
c
Guedes and Dias da Silva (1996) incorporates a lag phase and is recalculated from r ˆ 0:044 to r ˆ 0:05.
d
Infascelli et al. (1995).

12

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

Table 7
Gross energies (GE; MJ/kg as received) of L. cicera and L. sativus compared to ®eld pea (P. sativum), faba bean
(V. faba) and lupin (L. angustifolius)
L. cicera
a

Field pea

L. sativus

Faba bean

Lupin

Source

a

16.6
16.6a
16.7
16.2
±
±
±
±

±
±
16.9
16.7
15.9
16.1
16.1
±

±
16.6a
17.1
16.6
16.2
16.6
±
16.8

17.0
16.3a
16.7
16.7
15.9
16.5
±
16.8

±
18.1a
18.4
18.3
±
±
±
18.1

Flores and Castanon (1991)
Abreu and Bruno-Soares (1998)
Farhangi (1996)
Hughes (unpublished)
Duke (1981)
Guada Vallepuga (1972)
Low et al. (1990)
Petterson et al. (1997)

16.5

16.3

16.6

16.6

18.2

Mean

a

Estimated as 0.90 MJ/kg DM.

The metabolisable energy (ME) of L. cicera may be slightly higher than L. sativus
(Table 8). The ME for both Lathyrus spp. in sheep are approximately 13 MJ/kg DM,
similar to ®eld pea and faba bean. The sheep ME data of Farhangi (1996) are consistently
lower than all other sources, the in vitro dry matter digestibility (DMD) method used to
calculate ME is not recommended for grain (SCA, 1990). The single record for ME in
cattle shows L. sativus to be similar to ®eld pea and faba bean (Table 8). In chickens,
Table 8
Metabolisable energy (ME), digestible energy (DE), available metabolisable energy (AME) and true
metabolisable energy (TME) of L. cicera and L. sativus, all in MJ/kg DM, for sheep, cattle and poultry
(compared to ®eld pea (P. sativum), faba bean (V. faba) and lupin (L. angustifolius))a
Energy measure
(MJ/kg DM)

Source

L. cicera

L. sativus

Field pea

Faba bean

Lupin

ME sheep

i
ii
iii
iv
v
vi
vii
v
vi
viii
iii
ix
x
iv
xi

11.4
14.2
±
±
±
±
16.2
±
±
±
±
±
13.4
±
11.5

11.0
±
14.0
±
14.4
12.9
±
17.5
15.8
±
12.9
11.3
11.3
±
±

11.0
±
13.9
12.0
13.1
11.8
16.5
15.8
14.2
13.5
12.7
±
12.0
11.7
±

11.4
±
13.3
11.5
13.7
12.3
16.3
16.7
15.0
13.1
12.7
±
±
11.2
12.5

10.7
14.4
±
12.2
±
±
16.5
±
±
±
±
±
7.9
10.4
±

DE sheep

ME cattle and sheep
ME cattle
ME poultry
AME poultry
TME poultry

a
Source: (i) Farhangi (1996), (ii) White (unpublished), (iii) Kearl (1982), (iv) Petterson et al. (1997), (v)
Guada Vallepuga (1972), (vi) Zorita et al. (1972), (vii) Abreu and Bruno-Soares (1998), (viii) ARFC (1993), (ix)
Latif et al. (1975), (x) R.J. Hughes, personal communication, (xi) Flores and Castanon (1991).

13

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

Table 9
In vitro dry matter digestibilities (DMD), in vitro organic matter digestibilities (OMD), in vivo DMD and in vivo
OMD in sheep of L. cicera and L. sativus (compared to ®eld pea (P. sativum), faba bean (V. faba) and lupin
(L. angustifolius))a
Digestibility
measure (%)

Sourceb

L. cicera

L. sativus

Field pea

Faba bean

Lupin

In vitro DMD

i
ii
i
ii
i
iii
iv
iii
iv
v

82.1
88.6
85.5
85.5
93.6
93.0
±
94.5
±
90.0

79.9
92.9
89.3
90.3
±
±
95.9
±
95.6
±

82.7
±
85.4
±
±
±
85.7
±
86.9
91.1

80.3
±
83.2
±
±
±
90.9
±
91.8
91.7

78.7
±
85.4
±
96.7
95.5
±
96.5
±
86.0

In vitro OMD
In vivo DMD

In vivo OMD

ab
a
a
a
a
a
a
a

ab
b
b
b

a
a

a
a

b
b
a

ab
c

b
ab
a

b
a
a
a
a
b

a

Within each row the values with different letters are signi®cantly different (P < 0:05).
Source: (i) Farhangi (1996), (ii) Aletor et al. (1994), (iii) White (unpublished), (iv) Zorita et al. (1972), (v)
Abreu and Bruno-Soares (1998).
b

available ME is slightly higher in the Lathyrus spp. than in lupin, but otherwise similar to
®eld pea and faba bean (Table 8).
4.1.5. Digestibility
Only a few studies have determined in vitro DMD and organic matter digestibilities
(OMD). Similarly, little has been done on in vivo digestibilities. Generally, but not
consistently, digestibilities of L. cicera are slightly lower than for L. sativus. Both are,
however, broadly similar to other commonly used grain legumes (Table 9). In vivo
measures were usually higher than in vitro measures.
Several studies that have examined in sacco dry matter degradabilities have shown that
L. cicera and L. sativus have similar values to other commonly used grain legumes
(Table 6). However, the parameters differ, particularly with lupin, which had a small
water soluble fraction.
4.2. Antinutritional factors
In common with all grain legumes there are a range of antinutritional factors (ANFs)
found in L. cicera and L. sativus grain. The ANFs commonly found in grain legumes
include: tannins, phytic acid, oligosaccharides, protease inhibitors (trypsin and
chymotrypsin inhibitors), amylase inhibitors and lectins (Liener, 1989). ODAP is also
an ANF and is almost unique to the Lathyrus genus. There are only a small number of
published studies of levels and activities of ANFs, other than ODAP, in L. sativus (Latif
et al., 1975; Deshpande and Campbell, 1992; Aletor et al., 1994; Urga et al., 1995;
Srivastava and Khokhar, 1996; Wang et al., 1998), and less on L. cicera (Aletor et al.,
1994).

14

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

4.2.1. Proteinaceous ANFs
The proteinaceous ANFs are (i) trypsin and chymotrypsin inhibitors (both protease
inhibitors), respectively measured as trypsin inhibitor activity (TIA) and chymotrypsin
inhibitor activity (CTIA), (ii) amylase inhibitors and (iii) lectins. ODAP will also be
included under this heading, although ODAP is a non-protein amino acid and not strictly
a proteinaceous ANF.
4.2.1.1. ODAP. Until recent times animal feeding studies with L. sativus or L. cicera have
been performed with no knowledge of the role of ODAP in lathyrism. Therefore, in all
older studies of animal feeding (i.e. pre 1960s) the concentration of ODAP is unknown.
Recent studies have shown that ODAP concentrations can vary widely both within and
between the two species, however, environmental conditions are not as important as
genotype (Hanbury et al., 1999). Nonetheless, stresses such as salinity and drought
(Hussain et al., 1997) have been found to increase ODAP concentrations but are little
understood. Generally L. cicera has lower seed ODAP concentrations than L. sativus
(Table 10). Concentrations of ODAP in the seed can be particularly high in L. sativus land
races, up to 1.50% (Table 10).
1. TIA and CTIA
These inhibitors are destroyed in the rumen and so are not a problem for ruminant
animals. In monogastric species they can result in hypertrophy of the pancreas if
present in suf®cient quantities. There is often an increased production of S-containing
enzymes which are lost due to forming indigestible complexes. Animal growth rate is
commonly depressed by TIA and CTIA (Deshpande and Damodaran, 1990).
Due to different assay conditions, making comparisons between reported levels of
TIA and CTIA (Table 11) are dif®cult. The reported ranges are lower in L. cicera than
L. sativus (Aletor et al., 1994). Urga et al. (1995) claimed that measured TIA in

Table 10
ODAP content (% as received) mean and range of a number of lines of L. cicera and L. sativus grown at various
locations
L. cicera

L. sativus

Mean (range)
0.15
0.16
0.13
0.18
±
±
±
±
±

a

(NA)
(0.10±0.22)
(0.09±0.16)
(0.08±0.34)

0.16
a

No.

Mean (Range)

No.

128
24
16
96
±
±
±
±
±

0.20 (0.16±0.25)
0.49 (0.07±0.75)
0.49 (0.33±0.59)
0.39 (0.04±0.76)
0.88 (0.45±1.40)
0.72 (0.37±1.04)
0.44 (0.28±1.50)
0.32 (0.18±0.52)
NA (0.08±0.99)

10
70
36
407
172
10
1187
76
73

0.46
Not available.

Location

Source

Spain
Syria
Syria
Australia
Bangladesh
Ethiopia
India
Chile
China

Franco Jubete (1991)
Abd El-Moneim (1994)
Aletor et al. (1994)
Hanbury et al. (1999)
Kaul et al. (1982)
Tekle Haimanot et al. (1993)
Pandey et al. (1995)
Tay et al. (2000)
Chen cited by Campbell (1997)
Grand mean

15

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27
Table 11
Measured TIA and CTIA (units mgÿ1 DM) of L. cicera and L. sativus

TIA

CTIA

L. cicera

No. lines

L. sativus

12.6±20.4
9.15±15.1
±
±
21±31
±

16
2

20.1±44.1
±
16.7±26.2
133±174
±
0±23

2

No. lines
36

Source
Aletor et al. (1994)
Hanbury (unpublished)
Urga et al. (1995)
Deshpande and Campbell (1992)
Hanbury (unpublished)
Deshpande and Campbell (1992)

25
100
100

L. sativus (Table 11) was considerably lower than for soybeans, common beans and
cowpeas, but higher than found in chickpeas, although the values in these species were
not reported.
The TIA data can be compared to that of other feed legumes (measured by similar
methods) to estimate a ranking for both L. cicera and L. sativus (Table 12). In
summary, the species rank in likely order of increasing TIA: ®eld pea, faba bean,
L. cicera, L. sativus, P. vulgaris, soybean. Soybean is widely used for animal feeding
but must be heated to destroy the protease inhibitors prior to feeding to monogastric
animals.
2. Amylase inhibitor activity
Amylase is the enzyme primarily involved in starch digestion in mammals.
Amylase inhibitors are thought to reduce amylase activity, but the extent to which they
are important is debated (Deshpande and Damodaran, 1990). Deshpande and
Campbell (1992) found in 100 lines of L. sativus that the range of amylase inhibitor
activity (AIA) was 3.6±91.4 units gÿ1 DM, substantially lower than the 330±
675 units gÿ1 DM found in P. vulgaris cultivars (Deshpande et al., 1982).

Table 12
Relative comparisons (%) of measured TIA of L. sativus and L. cicera in relation to ®eld pea (P. sativum), faba
bean (V. faba), Phaseolus vulgaris and soybean (G. max) (each relative comparison is made between similar
measurement techniques)a
Relative comparison

1
2
3
4
5
6
7

Species TIA (%)
Field pea

Faba bean

L. cicera

L. sativus

P. vulgaris

Soybean

±
±
7
18
7
13
±

±
±
18 (iv)
11 (v)
±
5 (vii)
±

±
32 (ii)
±
±
±
22 (viii)
±

30 (i)
61 (ii)
±
±
±
±
56 (ix)

36
±
28
44
±
100
100

100
100
100
100
100
±
±

(iv)
(v)
(vi)
(vii)

(i)
(iv)
(v)
(vii)
(x)

(i)
(iii)
(iv)
(v)
(vi)

a
Source: (i) Latif et al. (1975), (ii) Aletor et al. (1994), (iii) Smith et al. (1980), (iv) Deshpande and
Damodaran (1990), (v) Elkowicz and Sosulki (1982), (vi) Saini (1989), (vii) Petterson et al. (1997), (viii)
Hanbury (unpublished), (ix) Deshpande and Campbell (1992), (x) Deshpande et al. (1982).

16

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

3. Lectins
Lectins are present in most legumes (Liener, 1989), they interfere with nutrient
digestion and absorption and increase wasteful protein synthesis, resulting in reduced
ef®ciency of nutrient utilisation. Levels in L. cicera and L. sativus are unknown,
however, Srivastava and Khokhar (1996) detected lectins in all of four lines of
L. sativus.
4.2.1.2. Proteinaceous ANF conclusions. Rotter et al. (1990) found that autoclaving feed
containing 82% L. sativus increased feed consumption and increased the efficiency of feed
utilisation by chickens. In a separate sample autoclaving for 2 h decreased the concentration of ODAP from 0.24 to 0.11%. Such effects of heating are commonly known to occur
for lectins, TIA and CTIA (Saini, 1989; Wiryawan and Dingle, 1999). Latif et al. (1975)
found that heating could totally inhibit TIA in L. sativus. The extrusion of L. sativus (which
involves heating) before feeding to pigs removed any inhibitory effect on proteolytic
activity, including trypsin activity, in pancreatic homogenates (Kapica et al., 1998).
The effect of various kinds of heating on the levels of ODAP are not clear. Boiling
treatments seem to consistently reduce ODAP concentration by 30±90% (Tekle Haimanot
et al., 1993; Srivastava and Khokhar, 1996; Padmajaprasad et al., 1997; Akalu et al.,
1998). However, the effect of boiling is largely (though not wholly) due to the water
solubility of ODAP. Roasting has been reported to both increase ODAP concentration
(Tekle Haimanot et al., 1993) and to reduce it by 87% (Akalu et al., 1998). Akalu et al.
(1998) found that heating induced isomerisation of b-ODAP (the neurotoxically active
form) to a-ODAP (the benign form), however, there appeared to be an equilibrium of
60% of total ODAP in the b-ODAP form, irrespective of heating time. From the human
nutrition perspective further research on reducing ODAP content during Lathyrus food
preparation is warranted. Given the variable results and the techniques required (possibly
boiling) the pre-treatments are probably not practical for animal feeding purposes. The
availability of low ODAP cultivars also reduces the need for such pre-treatment of animal
feeds.
The results of heat treatment of L. sativus indicate that the protease inhibitors are
inactivated as observed in other grain species. It would be desirable not to require heat
treatment, however possible reductions in ODAP content could also occur. The presence
of lines of negligible TIA indicates the potential for further reductions through breeding.
There is insuf®cient data on the range of CTIA in either L. sativus or L. cicera, although
with wider testing variation is highly likely to be found, given the variation present in
other grain legume species. Such variation can be exploited in breeding programs.
4.2.2. Tannins
Tannins are polyphenolic compounds of two classes: low molecular weight
hydrolysable and higher molecular weight non-hydrolysable (or condensed). It is
postulated that condensed tannins bind to proteins in the digestive tract and form
complexes which are frequently indigestible (Marquardt, 1989). The hydrolysable tannins
are often found to have little effect on digestibility.
Tannins in faba beans and ®eld peas are frequently localised in the seed coat
(Marquardt, 1989; Gatel and Grosjean, 1990), with high tannin levels in darker seed coats

C.D. Hanbury et al. / Animal Feed Science and Technology 87 (2000) 1±27

17

Table 13
Condensed tannins contents (catechin equivalents, % as received) in L. cicera and L. sativus
Species

Condensed tannins

No. lines

Mean

Range

L. cicera

0.36
0.68

0.27±0.55
0.59±0.77

16
2

L. sativus

0.12
0.21
0.64
0.31a

0.00±0.44
0.00±0.50
0.46±0.77
0.08±0.47a

100
36
25
9

a

Source

Aletor et al. (1994)
Hanbury (unpublished)
Deshpande and Campbell (1992)
Aletor et al. (1994)
Urga et al. (1995)
Wang et al. (1998)

Estimated as 0.90 of % DM.

than lighter ones. Similarly, Deshpande and Campbell (1992) found that white or cream
coloured seeds of L. sativus were associated with low tannin levels (both condensed and
total), whereas seed with darker seed coats generally had high tannin levels. Similar
observations regarding L. sativus were made by Urga et al. (1995) and Wang et al. (1998).
In L. sativus lighter seeds are associated with white ¯ower colour (Jackson and Yunus,
1984), consequently the selection of white ¯ower colour could be used to reduce tannin
contents.
The range of condensed tannins in the literature is from undetectable to 0.77%
(Table 13), with a smaller range in L. cicera than in L. sativus. This may be a result of the
smaller number of measurements and/or little selection of L. cicera in comparison to
L. sativus lines. Unlike L. sativus, ¯ower colour in L. cicera does not vary greatly (Franco
Jubete, 1991; Hanbury et al., 1995) and may be related to the small variation in tannin
content. There is considerable scope for improvement of varieties with negligible condensed
tannin levels. In the case of L. sativus, much suitable germplasm is already identi®ed.
4.2.3. Phytate
Phytate is a cyclic compound that chelates with mineral ions (e.g. Ca, Mg, Zn, Fe) and
forms compounds not readily absorbed in the intestine (Liener, 1989), thereby reducing
animal performance. It is destroyed in the rumen and so is not a nutritional problem for
ruminants. The range 0.49±1.09% in the two Lathyrus spp. (Table 14) is high in
comparison to ®eld pea (0.15±0.70%) and faba bean (