Breadmaking Use of Andean Crops Quinoa, Kañiwa, Kiwicha, and Tarwi
Cristina M. Rosell,1,2 Gladys Cortez,3 and Ritva Repo-Carrasco3
ABSTRACT
The effect of addition of flours from the highly nutritious Andean crops
quinoa (Chenopodium quinoa), kañiwa (Chenopodium pallidicaule),
kiwicha (Amaranthus caudatus), and tarwi (Lupinus mutabilis) has been
investigated in wheat doughs and fresh bread quality. The thermome-
chanical profile of wheat doughs and bread quality has been explored by
increasing substitution of wheat flour at 0–100% by Andean crop flours.
Dough blends were evaluated using the Chopin Mixolab device, whereas
bread quality assessment comprised sensory (overall acceptability) and
physicochemical (moisture, specific volume, texture, color) determina-
tions in composite breads. In general, no breads with aerated crumb struc-
Western diets based on wheat-derived breads are much less sa-
tiating than those based on more traditional grains in less devel-
oped countries. Particularly, some alternative crops (buckwheat,
oat, quinoa, amaranth grain, etc.) are of nutritional interest for
developing healthier and typical regional foods (Berghofer and
Schonlechner 2000; Berti et al 2005).
Quinoa (Chenopodium quinoa), kañiwa (Chenopodium pallidi-
caule) and kiwicha (Amaranthus caudatus) are indigenous pseudo
cereals from the Andean region. Tarwi (Lupinus mutabilis) is a
legume cultivated in Peru, Bolivia, and Ecuador since pre-Inca
times. All these crops are highly nutritious and environmentally
adaptable. The protein content of quinoa and kañiwa is high and
they have balanced amino acid composition (Repo-Carrasco et al
2003). Kiwicha proteins have also an excellent nutritional value,
especially when combined with other cereals (Pedersen et al
1987). Tarwi is extremely rich in protein (45% protein content)
(Schoeneberger et al 1987) and oil (16% oil content) (Gross et al
1988), but due to the presence of bitter and toxic alkaloids must
be debittered before consumption. Quinoa contains saponins that
also have to be eliminated before use as human food. Addition-
ally, high dietary fiber content and natural antioxidants such as
phenolic compounds gives these Andean grains even greater po-
tential as functional and bioactive ingredients in food products
(Gorinstein et al 2007). Partial substitution of wheat flour by
flours from Andean crops could improve the nutritional quality of
wheat bread because they are rich in lysine and other essential
amino acids present in scarce amounts in wheat flour (Berghofer
and Schonlechner 2000). These Andean crops contain no gluten,
which seriously constrains the technological performance of the
baking process. However, recently in Europe, quinoa has attracted
attention for people with celiac disease as an alternative to the
common cereals like wheat, rye, and barley, all of which contain
gluten (Schoenlechner et al 2008). The use of composite flours
also offers economic advantages for countries like Peru where the
cultivation of wheat is very scarce for geographical and climate
reasons. Andean grains have potential agronomic importance
world-wide because they easily adapt to different environmental
conditions. They can be cultivated on poor soils and at high alti-
tudes. In particular, specific ecotypes of quinoa are adapted to
1 Cereal
Group, Institute of Agrochemistry and Food Technology (IATA-CSIC),
P.O. Box 73, 46100 Burjassot, Valencia, Spain.
2 Corresponding author. Phone: +34 963900022. Fax: +34 963636301. E-mail:
crosell@iata.csic.es
3 Universidad Nacional Agraria La Molina, Facultad de Industrias Alimentarias,
Av. La Molina s/n, Lima, Peru.
doi:10.1094 / CCHEM-86-4-0386
© 2009 AACC International, Inc.
386 CEREAL CHEMISTRY
Cereal Chem. 86(4):386–392
ture could be obtained from 100% Andean crop flours, with the exception
of quinoa breads that had overall sensory values about half a completely
perfect score, and which were not significantly different from the breads
made from a 50:50 blend of wheat and quinoa. Replacement of wheat
flour by ≤12.5% (tarwi), 25% (kañiwa), and 50% (kiwicha), respectively,
still produced breads with good sensory acceptability but variable color
and doughs with acceptable thermomechanical patterns. Partial substitu-
tion of wheat flour by Andean crop flours constitutes a viable option to
improve the nutritional value of the breads, with acceptable technological
performance of dough blends and composite breads.
diverse conditions of humidity and temperature (Jacobsen 2003).
The crop was recently introduced to various European countries
and also to North America, Asia, Africa, and Australia (Jacobsen
2003). Quinoa was nominated by FAO as a crop most likely to
significantly contribute to global food security in the 21st century
(FAO 1998).
Despite the potential improvement that these Andean grains
could represent when looking for healthier diets, scarce studies
have been focused on the development of bakery goods. Up to
20% replacement of wheat flour by quinoa or kiwicha has been
reported (Bruemmer and Morgenstern 1992; Chauhan et al 1992;
Morita et al 2001) for making wheat-based breads. Even the use
of germinated quinoa as breadmaking ingredient in wheat bread
formulation has been proposed (Park and Morita 2005). Addition-
ally, a breadmaking trial showed that wheat flour replacement by
10% tarwi gave acceptable bread quality (Gross et al 1983; Lo-
renz and Coultier 1991). No information has been found regard-
ing the use of kañiwa in breadmaking processes.
A comparative study on the potential breadmaking properties of
these four Andean crops is described for extending their applica-
tion and improving the nutritional value of the existing bakery
products. The effect of increasing substitution of wheat flour by
Andean crop flours from 0 to 100% on rheological characteristics
of bread dough and fresh bread quality is presented. Rheological
behavior of the bread doughs was assessed by defining the ther-
momechanical profile using the Mixolab device and fresh bread
quality was assessed by measuring sensory (overall acceptability)
and physicochemical (moisture, specific volume, texture, crumb
and crust color) parameters.
MATERIALS AND METHODS
A commercial blend of breadmaking wheat flour and the fol-
lowing varieties of the Andean crops were used: Rosada quinoa
(Chenopodium quinoa) from Huancayo location, Cupi kañiwa
(Chenopodium pallidicaule) from Puno, Centenario kiwicha
(Amaranthus caudatus) a new cultivar developed by the National
Agrarian University from Ancash, and tarwi (Lupinus mutabilis)
from the leguminous program of the National Agrarian University
La Molina (Lima, Peru). All the Andean crops were grown in
2006 season and were used in 2007. The grains were stored in
sealed plastic containers in a cool, dark, dry cabinet until use.
Standard bread formulation using baker’s compressed yeast and
salt were followed.
Preparation of Flours of Quinoa, Kañiwa, Kiwicha, and Tarwi
Debittering of quinoa to remove saponins was achieved by
washing in water then drying at 50°C for 4 hr. Quinoa grains were
milled in a laboratory mill (Cyclotec 1093, Foss, Denmark). Kañi-
wa and kiwicha grains were cleaned and milled with the Cyclotec
mill. Tarwi grain was dehulled then debittered to remove alkaloids
following the modified Cusco method described by Gross et al
(1983). The debittered tarwi was dried at 50°C for 12 hr in a hot-
air cabinet dryer and then milled as previously described.
Flour Characteristics
The fflours used were wheat (Triticum aestivum), quinoa, ki-
wicha, kañiwa, and tarwi. Moisture, protein, ash, and fat were
determined following standard methods (ICC, Vienna). The colors
of milled grain samples were measured directly in the powder by
using a colorimeter (Chroma Meter CR-400/410, Konica Minolta,
Japan) after standardization with a white calibration plate (L =
97.64, a = –0.02, b = 1.77). The color was recorded using Hunter
L,a,b uniform color space (HunterLab), where L indicates white-
ness, a indicates hue on a green (–) to red (+) axis, and b indicates
hue on a blue (–) to yellow (+) axis.
Thermomechanical Behavior of Flours
Mixing and pasting behavior of doughs from wheat and Andean
crops blends were studied using the Mixolab device (Chopin,
Tripette et Renaud, Paris, France). The real time torque (ex-
pressed in Nm) measured for the dough indicates physical pa-
rameters including the quality of the protein network and starch
behavior during mixing, heating, and cooling (Bonet et al 2006;
Collar et al 2007, Rosell and Collar 2008). Flour blends (50 g) of
increasing proportions of the Andean crop flours (0, 12.5, 25, 50,
and 100%) were evaluated by the Mixolab with a constant addi-
tion of water. The mixing settings used were 6 min at 30°C, heat-
ing rate of 4°C/min until 90°C, 7 min holding at 90°C, cooling
rate of 4°C/min until 55°C, then 5 min holding at 55°C. The mix-
ing speed during the entire assay was 75 rpm. The recorded curve
shows consistency during mixing, overmixing, pasting, and gel-
ling.
Breadmaking Process
Wheat-Andean crop flour blends (0, 12.5, 25, 50, and 100%)
were used for making small pan breads. Blended flours (300 g)
were mixed with water (59% v/w, blend basis), salt (1.8% w/w,
blend basis) and yeast (2% w/w, blend basis) in a 300-g Bra-
bender farinograph bowl for 4 min (wheat flour and wheat-quinoa
blend), 6 min (wheat-tarwi blends), or 8 min (wheat-Kiwicha
blends and wheat-kañiwa blends). After kneading, the doughs
were divided into 50-g pieces and individually placed into alumi-
num pans. Fermentation was done in a proofing cabinet for 45
min at 30°C and 85% RH. The dough pieces were baked for 35
min at 165°C in an electric oven and the loaves were removed
from the pans and cooled at room temperature for 30 min.
Bread Quality Determination
Bread quality was evaluated by assessing volume (rapeseed
displacement), weight, specific volume, crust and crumb color,
TABLE I
Proximate Composition (g/100 g, as-is) and Color Tristimulus Parameters of Sample Floursa
Wheat Quinoa
Moisture content 14.21 ± 0.09d 8.31 ± 0.06b
Proteinb 9.81 ± 0.10a 13.83 ± 0.23c
Ash 0.53 ± 0.08a 2.09 ± 0.09c
Fat 0.92 ± 0.09a 5.04 ± 0.14b
Color
L 89.96 ± 2.09d 81.74 ± 2.51c
a –0.73 ± 0.09a 0.35 ± 0.10b
b 10.48 ± 0.23a 14.48 ± 0.81b
a Mean values in a row followed by different letters are significantly different (P < 0.05). Data expressed on moisture basis (n = 3).
b Protein conversion for wheat was N × 5.7 and N × 6.25 for all others.
moisture content, and crumb hardness. Crust and crumb color
were determined using a colorimeter (ChromaMeter CR-400/410,
Konica Minolta, Japan). Moisture content was determined follow-
ing the ICC Method. Texture profile analysis (TPA) of the bread
crumbs was performed by a texture analyzer (TA-XT2i, Stable
Micro Systems, Surrey, UK). A bread slice 1-cm thick was com-
pressed up to 50% of its original height at a crosshead speed of 1
mm/sec with a cylindrical stainless steel probe (10 mm diameter).
All measurements were performed immediately after the 30-min
cooling period. Sensory perception was performed by a trained
panel of judges who scored the overall acceptability of the breads
using a semistructured scale (0 extremely dislike, 10 extremely
like).
Statistical Analyses
Values were the mean of three replicates, unless otherwise de-
scribed. A multiple sample comparison procedure was used to
determine statistically significant differences among the means
(Plus v.7.1, Statistical Graphics Corporation, UK). Fisher’s least
significant difference (LSD) test with 95% confidence was used
to discriminate among means.
RESULTS AND DISCUSSION
The proximate composition and color tristimulus parameters of
the flours used in this study are presented in Table I. Compared
with wheat, all the Andean cereals had significantly (P < 0.05)
high protein content, particularly tarwi, which was extremely high
(57.4%). Fat and ash content of the four Andean crops were sig-
nificantly (P < 0.05) higher than in the wheat flour, again particu-
larly the fat content of tarwi (24.5%). These levels for quinoa,
kañiwa, and kiwicha were similar to those previously reported
(Repo-Carrasco 1992; Repo-Carrasco et al 2003; Park et al 2005).
Protein and fat content of tarwi flour agree with previous data
reported for tarwi or lupine flour obtained after debittering by
using the Cusco method (Gross et al 1983). The L (whiteness)
value of all the Andean flours was lower than that observed for
wheat flour, particularly for kañiwa, which showed a strong
brown color. The a value for the wheat flour was negative (green
hue), but positive (red hue) for all Andean flours, and most sig-
nificantly for the brown kañiwa. All b values were positive (yel-
low hue); tarwi had extremely high value and a characteristic
yellowish color. As a consequence, Andean grain-wheat flour
blends will produce a range of colored bakery products.
Thermomechanical Behavior of Andean Grains-Wheat
Flour Blends
Plots of the thermomechanical behavior of doughs obtained
from wheat-Andean crop blends recorded in the Mixolab device
are shown in Fig. 1. Additionally, plots obtained for wheat flour
and Andean crops flour were included for gaining information
about the Andean crops viscometric profile associated to the heat-
ing-cooling processes. The first part of the curve corresponds to
Kañiwa Kiwicha Tarwi
11.46 ± 0.10c 11.57 ± 0.09c 6.14 ± 0.07a
14.75 ± 0.19d 12.34 ± 0.09b 57.36 ± 0.59e
3.27 ± 0.06e 1.73 ± 0.08b 2.61 ± 0.09d
6.40 ± 0.15c 6.36 ± 0.12c 25.40 ± 0.33d
44.70 ± 1.96a 76.12 ± 2.41b 73.92 ± 2.09b
5.01 ± 0.08e 1.90 ± 0.09c 2.68 ± 0.15d
13.50 ± 0.94b 17.09 ± 1.03c 70.99 ± 2.15d
Vol. 86, No. 4, 2009 387
mixing and initial heating (up to 52–58°C) and is associated with
proteins undergoing hydration, changes induced by mechanical
input, thermally induced aggregation and unfolding (Rosell et al
2007). Further heating and cooling results in starch gelatinization
and gelling, respectively (pasting), with a consequent increase in
dough consistency initially due to starch swelling and then due to
starch recrystallization (Rosell et al 2007). The Andean crops all
had very different thermomechanical behaviors to wheat flour,
and to one another. Significant differences were detected in the
part of the curve associated to starch changes, except for tarwi
where the differences were more pronounced in the part of the
curve associated with protein changes.
The plots for quinoa-wheat dough (Fig. 1A) showed that blends
had intermediate behavior between pure wheat flour and quinoa
flour, proportional to the level of added quinoa. Pure quinoa flour
gave a very low consistency dough, requiring a higher tempera-
ture than wheat for its starch pasting (minimum torque observed
at a dough temperature of 57°C, compared with wheat at 55°C).
However, significant genetic variation for both thermal and past-
ing properties of quinoa starch has been reported (Lindeboom et
al 2005). Maximum peak torque during heating was reached at
72°C for the wheat dough, followed by a pronounced torque de-
crease. This fragility of the swollen starch granules is reflected in
the swelling after they break down under continuous stirring (Tip-
ples et al 1980). The quinoa flour gave a very low consistency
during cooling, which indicates very low recrystallization or set-
back (Rosell et al 2007; Collar and Rosell 2009). Dough stability
during pasting decreased significantly upon addition of 12.5%
quinoa but only slightly more up to 50%. There was a dramatic
drop in stability between the 50% blend and pure quinoa. The
same trend was observed for the maximum torque produced dur-
ing the heating stage, with a drop on addition of any quinoa, only
a small difference up to 50%, but a dramatic decrease to pure
quinoa. Similar trends were observed for the torque measured
after cooling at 50°C.
Fig. 1. Thermomechanical behavior plots of dough obtained from wheat-Andean crop blends (0, 12.5, 25, 50, 100%) recorded in the Mixolab device. A,
Quinoa; B, Kañiwa; C, Kiwicha; D, Tarwi.
388 CEREAL CHEMISTRY
Very similar patterns were observed for the blends with kañiwa
and wheat flour (Fig. 1B), despite the pure kañiwa dough giving
only a very low torque during the entire test period. The greatest
difference was observed during mixing, overmixing, and initial
heating, where dough containing kañiwa-wheat blends showed
lower torque than wheat flour dough, indicating that kañiwa pro-
teins decrease dough stability. Kañiwa showed very low mechani-
cal stability and consistency, which could only be recorded during
a short initial period. During the overmixing and heating phases
no torque could be detected, indicating no starch gelatinization.
The doughs with pure kañiwa were very fragile and easily disin-
tegrated, yet the blends up to 50% gave similar doughs to those of
quinoa.
The kiwicha-wheat blends generated dough with mechanical
behavior similar to wheat flour (Fig. 1C) and overlapping
Mixolab plots were obtained for all blends during the mixing,
overmixing and initial heating phases. Divergence started when
starch gelatinization began and in contrast with the quinoa and
kañiwa-wheat blends, the effect was more pronounced depending
on the level of kiwicha added. Low pasting viscosity and setback
has been associated with the low amylose contents of kiwicha
starch (3.0–12.5%) (Uriyapongson and Rayas-Duarte 1994; Kong
et al 2009). In contrast to the other Andean grains, kiwicha flour
was able to develop dough with measurable torque, but like
kañiwa, as heating started its consistency became undetectable.
The tarwi-wheat blends showed opposite behavior to the other
Andean crops (Fig. 1D). Increasing the amount of tarwi flour in
the blends augmented dough consistency, particularly through the
initial mixing phase. This was expected considering the protein
content of this flour (57.4% tarwi vs. 9.8% wheat). During the
first part of the curve (mixing, overmixing, and initial heating),
plots are governed by changes to proteins and the extent of the
modification is greatly dependent on the protein content and na-
ture (Bonet et al 2006). Generally, the tarwi doughs were difficult
to manipulate due to high consistency. Again, in contrast to the
other Andean crops, the presence of tarwi up to 25% hardly af-
fected wheat starch gelatinization. From 50% addition, the curve
during pasting did drop significantly from the wheat curve. The
effect of wheat starch dilution was not evident in the presence of
tarwi, likely masked by the proteins. Protein type can modify
pasting and gelling properties of cereal starches (Marco and
Rosell 2008), and in turn be influenced by thermal-induced ge-
latinization processes leading to modified dough rheological be-
havior (Ngarize et al 2004). The blends with up to 25% tarwi,
slightly increased the torque during cooling compared with wheat,
probably due to the interactions between the wheat amylose and
tarwi lipids (the second major component of the tarwi flour), as it
has been previously observed for surfactant-supplemented wheat
doughs (Collar 2003). The 50% blend showed a completely dif-
ferent thermomechanical plot, with extremely high dough consis-
tency during mixing but a more pronounced drop during the
pasting phase. Pure tarwi showed very erratic behavior: low
torque during mixing, then almost being undetectable during heat-
ing and cooling.
Bread Quality Parameters
Breads obtained by substituting wheat flour with increasing
Andean crops flour produced a range of loaves with different
sensory and technological characteristics (Fig. 2).
The color tristimulus parameters of the breads obtained from
wheat-Andean crops flour blends are detailed in Table II. Crust
whiteness (L) decreased when the proportion of the Andean flours
increased in the blends. Breads with higher proportions of Andean
flours had darker crust (Fig. 2) due to the darker color of the An-
dean crops flour and also the increased content of reducing sugars
and proteins with lysine residues present in these flours (Repo-
Fig. 2. Breads obtained from wheat-Andean crops flour blends (0, 12.5, 25, 50, 100%).
Carrasco et al 2003) that produce nonenzymatic Maillard brown-
ing during the baking process. The a values for the bread crusts
were always positive (red hue) but with the exception of breads
containing kañiwa, no clear trends were observed with increasing
substitution of wheat flour. The b values for bread crusts were
also positive (yellow hue), and a generally significant decrease
was observed with increasing the level of Andean crop flours in
the blends.
For the crumb, the L values showed generally significant reduc-
tion whit increasing proportion of Andean crop flours, yielding
darker crumbs. However, the level of additional Andean crop
flours needed to significantly darken the crumb from pure wheat
varied between them, from >50% with quinoa and tarwi to 12.5%
with kañiwa (Fig. 2). The a values ranged from negative values
(green hue) to positive values (red hue), increasing significantly
when the proportion of Andean flour was >25%, except for
kañiwa where the 12.5% blend was significantly redder than pure
wheat. The b values were all positive, indicating yellowish color.
The pure Andean crop breads were all significantly higher than
pure wheat bread (with tarwi the highest) except for kañiwa
which was significantly lower (Table II, Fig. 2).
The moisture contents of the Andean crop breads are given in
Fig. 3. Kañiwa and kiwicha breads showed increasing moisture
content with increasing wheat flour substitution but the pure
breads decreased to a level similar to pure wheat for kañiwa and
significantly lower for kiwicha. Breads containing up to 25%
tarwi flour showed increasing moisture content, but it dropped at
higher levels of wheat replacement until pure tarwi was signifi-
cantly below pure wheat. For quinoa, yet another pattern was
evident, with no increase in moisture content up to 25% substitu-
tion, then a significant jump but no difference between the substi-
Vol. 86, No. 4, 2009 389
 TABLE II
Color Tristimulus Parameters of Breads from Wheat Flour Blends with Andean Crops (0, 12.5, 25, 50, 100%)a
Crust Color Crumb Color
Sample and Blend (%)b L a b L a b
Quinoa
0 55.9d 10.5b 23.9d 60.3c –1.0a 11.4a
12.5 39.7c 12.7c 16.5c 60.4c –0.8a 13.6b
25 37.1c 12.7c 14.9c 59.5c –0.2b 15.9c
50 34.0b 12.2c 13.2b 55.7b 1.2c 17.8d
100 21.1a 8.5a 9.4a 46.7a 1.8d 16.1c
Kañiwa
0 55.9d 10.5c 23.9d 60.3e –1.0a 11.4b
12.5 43.1c 10.8c 17.7c 45.7d 1.9b 10.5b
25 34.5b 8.4b 12.5b 38.2c 3.8c 10.8b
50 37.8b 6.6a 12.3b 30.1b 4.9c 8.2a
100 23.0a 6.4a 8.3a 23.2a 5.0c 7.9a
Kiwicha
0 55.9c 10.5a 23.9c 60.3c –1.0a 11.4a
12.5 42.5b 12.7b 18.4b 58.4c –0.5a 12.4a
25 37.7b 11.7b 15.3a 54.8b –0.4a 13.0a
50 34.3a 11.9b 13.4a 54.4b 1.0b 15.9b
100 32.8a 9.9a 14.1a 45.2a 3.2c 16.0b
Tarwi
0 55.9b 10.5b 23.9b 60.3c –1.0a 11.4a
12.5 49.3a 12.5c 22.4b 63.0c –0.8a 17.3b
25 47.6a 12.9c 22.3b 63.4c –0.7a 21.4c
50 51.2a 10.6b 24.3b 56.6b 3.5b 23.8c
100 45.3a 6.6a 20.0a 45.3a 6.6c 20.0c
a Mean values in columns within each Andean blend with different letters are significantly different (P < 0.05) (n = 4).
b Andean crops flour in the blend.

Fig. 4. Specific volume of breads obtained from wheat-Andean crops

flour blends (0, 12.5, 25, 50, 100%). Error bars indicate standard devi-
Fig. 3. Moisture content of breads obtained from wheat-Andean crops ation.
flour blends (0, 12.5, 25, 50, 100%). Error bars indicate standard deviation.

although kañiwa and kiwicha flours with similar proximate com-

tution levels from >25% to pure quinoa. Likely the distinct flour
composition, rich in proteins and lipids, was responsible of this
result. Wheat flour substitution by Andean crops flour also af-
fected bread specific volume (Fig. 4). The 12.5% Andean crop
flours did not show any effect, except for kañiwa where this level
produced an increased specific volume. Kañiwa flour contains
more fermentable sugars than quinoa and kiwicha, favoring yeast
fermentation and consequently increased gas production (Repo-
Carrasco et al 2003). At 25% substitution, kiwicha was still simi-
lar to pure wheat and the kañiwa blend had dropped back to the
wheat level. For quinoa and tarwi, substitution levels >12.5%
resulted in significant reductions in the specific volume of their
loaves, dependant upon substitution level. It was previously re-
ported that substitution of 7.5–10% wheat flour with quinoa in-
creased loaf volume, but further substitution at >15% resulted in
decreased volume (Morita et al 2001). The reduction of the gluten
content in the dough might be responsible for the volume effect,
position to quinoa do not conform with this explanation. Previous
microscopic analysis of doughs containing 30% quinoa flour re-
vealed they were unable to form a well-developed gluten matrix
involving aggregated starch granules (Park et al 2005).
Results of texture profile and sensory perception (overall ac-
ceptance) of bread crumbs with increasing proportion of Andean
crop flours are given in Table III. The presence of 25% quinoa in
the blends caused significantly harder crumbs compared with the
control (wheat bread) and that hardness progressively increased at
higher substitution levels. Similar findings were reported by Park
et al (2005), where up to 30% quinoa substitution of wheat flour
produced a poor extensible gluten network that lead to hardening
of the bread crumb. For kañiwa and kiwicha, a significant harden-
ing did not occur until 25% substitution and for tarwi crumb
hardness increased at all substitution levels. Tarwi flour has a
higher percentage of proteins and lipids and therefore relatively
limited starch compared with the other Andean crop flours, ex-

390 CEREAL CHEMISTRY

 TABLE III
Texture Profile Analysis and Sensory Perception of Crumbs from Breads Obtained Wheat Flour Blends with Andean Crops (0, 12.5, 25, 50, 100%)a
Sample and Blend(%)b Hardness (g force) Springiness Cohesiveness Chewiness (g force) Resilience Overall Acceptancec
Quinoa
0 508a 0.999b 0.811b 414a 0.446b 7.3b
12.5 461a 0.999b 0.838b 462a 0.470b 8.0b
25 1,013b 0.999b 0.808b 817b 0.434ab 7.5b
50 2,815c 0.938a 0.702a 1,859c 0.362a 5.5a
100 4,249d 0.967a 0.769a 3,160d 0.390a 4.5a
Kañiwa
0 508a 0.999a 0.811c 414a 0.446b 7.3c
12.5 497a 0.999a 0.821c 405a 0.367b 7.5c
25 645a 0.996a 0.773b 478a 0.388b 7.0c
50 3,886b 0.954a 0.585a 2,163b 0.221a 3.8b
100 nd nd nd nd nd 0.7a
Kiwicha
0 508a 0.999c 0.811b 414a 0.446b 7.3c
12.5 455a 0.999c 0.803b 366a 0.410b 8.3d
25 454a 0.998c 0.805b 258a 0.394b 7.3c
50 1,520b 0.901b 0.646a 880b 0.183a 5.8b
100 2,599c 0.602a 0.540a 829b 0.153a 2.0a
Tarwi
0 508a 0.999c 0.811b 414a 0.446b 7.3d
12.5 1,014b 0.983b 0.667a 665a 0.315b 6.3c
25 2,427c 0.944b 0.647a 1,483b 0.280ab 3.3b
50 6,105d 0.874a 0.637a 3,422c 0.241a 1.0a
100 nd nd nd nd nd 0.3a
a Mean values in columns within each Andean blend with different letters are significantly different (P < 0.05) (n = 4); nd, not determined.
b Andean crops flour in the blend.
c Scale of 1–10.

plaining its difference in this hardness trend. In all cases, there
was a general trend for decreased springiness, cohesiveness, and
resilience when the proportion of Andean crop flours increased,
although often not significantly between adjacent blend levels.
For chewiness, this trend was reversed, generally increasing with
rising percentage of Andean crop flours in their breads, although
again, often not significantly between adjacent blend levels. Qui-
noa was the only Andean grain in this set that produced a bread in
the complete absence of wheat flour, although with very compact
macrostructure. The other Andean grains, gave very hard dough
unable to retain the gas released during fermentation, and thus
very compact and hard bread structures (Fig. 2).
Only pure quinoa bread gave a sensory acceptance score that
was a least close to its pure wheat comparative (Table III). How-
ever, at least some of the blends of all the Andean crops gave ac-
ceptable bread, except for tarwi. The highest overall acceptance
was given by the 12.5% kiwicha blend and blends of up to 25%
were not significantly different to wheat bread for them all, again
except for tarwi. For tarwi, each level of substitution caused a
significant decrease in overall acceptability. This result contrasts
with those of Gross et al (1983), who found that 10% supplemen-
tation of wheat flour by either whole lupine flour or dehulled
lupine flour produced an acceptable product with both Lupinus
albus and L. mutabilis (tarwi), although smell and taste were af-
fected. They further determined that at this level of lupine substi-
tution, the protein efficiency ratio increased from 28% (wheat
flour) to 76% (with 10% lupine flour) compared with 100% ca-
sein.
CONCLUSIONS
Wheat flour replacement at different levels (0–100%) by flours
from different Andean crops produced doughs with different
thermomechanical profiles and breads with variable sensory ac-
ceptability and physicochemical features, depending on both the
degree of wheat flour substitution and on the Andean crop used.
No breads with aerated crumb structure could be obtained from
100% Andean crop flours, with the partial exception of quinoa
bread that produced an overall sensory score of 4.5/10 for pure
quinoa (compared to 7.3/10 for pure wheat bread), and no signifi-
cant differences compared with wheat in blends up to 25%. De-
spite the very poor performance of pure kañiwa and kiwicha
breads, blends of up to 25% gave acceptable sensory scores, but
variable thermomechanical and color patterns. Only tarwi gave
universally poor results for most of the blend levels tested. If
these variable quality attributes can be tolerated addition of An-
dean crop flours to wheat flours, with the possible exception of
tarwi, to wheat flours constitutes a viable option to improve the
nutritional value of breads for quantitative and qualitative protein
composition and bioactive components.
ACKNOWLEDGMENTS
This work was financially supported by Spanish Ministry of In-
novación y Ciencia (AGL2005-05192-C04-01 and AGL 2008-00092ALI),
Consejo Superior de Investigaciones Científicas (CSIC) and a joint
CYTED project (PANXTODOS P106AC0301). We would like to thank
C. Collar for technical assistance.
LITERATURE CITED
Berghofer, E., and Schonlechner, R. 2000. Utilization of amaranth and
quinoa in Europe. Ernahrung 24:303-308.
Berti, C., Riso, P., Brusamolino, A., and Porrini, M. 2005. Effect on appe-
tite control of minor cereal and pseudocereal product. Br. J. Nutr.
94:850-858.
Bonet, A., Blaszczak, W., and Rosell, C. M. 2006. Formation of homo-
polymers and heteropolymers between wheat flour and several protein
sources by transglutaminase catalyzed cross-linking. Cereal Chem.
83:655-662.
Bruemmer, J. M., and Morgenstern, G. 1992. Baking properties of the
pseudocereals amaranth and quinoa. Getreide Mehl Brot. 43:78-84.
Chauhan, G. S., Zillman, R. R., and Eskin, N. 1992. Dough mixing and
breadmaking properties of quinoa-wheat flour blends. Int. J. Food Sci.
Technol. 27:701-705.
Collar, C. 2003. Significance of viscosity profile of pasted and gelled
formulated wheat doughs on bread staling. Eur. Food Res. Technol.
216:505-513.
Collar, C., and Rosell, C. M. 2009. Relationship between Mixolab and
other devices. In: Mixolab Handbook. A. Dubat, C. M. Rosell, M. Tul-
bek, eds. AACC International: St. Paul, MN.

Vol. 86, No. 4, 2009 391

Collar, C., Bollaín, C., and Rosell, C. M. 2007. Rheological behaviour of
formulated bread doughs during mixing and heating. Food Sci. Technol.
Int. 13:99-108.
FAO. 1998. Under-utilized Andean Food Crops. FAO: Rome.
Gorinstein, S., Medina Vargas, O., Jaramillo, N., Arnao Salas, I., Martinez
Ayala, A., Arancibia-Avila, P., Toledo, F., Katrich, E., and Trakhten-
berg, S. 2007. The total polyphenols and the antioxidant potentials of
some selected cereals and pseudocereals. Eur. Food Res. Technol.
225:321-328.
Gross, U., Galindo, R. G., and Schoeneberger, H. 1983. The development
and acceptability of lupine (Lupinus mutabilis) products. Qual. Plant-
Plant Foods Hum. Nutr. 32:155-164.
Gross, R., von Baer, E., Koch, F., Marquard, R., Trugo, L., and Wink, M.
1988. Chemical composition of a new variety of the Andean lupin (Lu-
pinus mutabilis cv. inti) with low-alkaloid content. J. Food Comp.
Anal. 1:353-361.
Jacobsen, S.-E. 2003. The worldwide potential for quinoa (Chenopodium
quinoa Willd.). Food Reviews Int. 19:167-177.
Kong, X., Bao, Y., and Corke, H. 2009. Physical properties of Amaran-
thus starch. Food Chem. 113:371-376.
Lindeboom, N., Chang, P. R., Falk, K. C., and Tyler, R. T. 2005. Charac-
teristics of starch from eight quinoa lines. Cereal Chem. 82:216-222.
Lorenz, K., and Coulter, L. 1991. Quinoa flour in baked products. Plant
Food Human Nutr. 41:213-223.
Marco, C., and Rosell, C. M. 2008. Effect of different protein isolates and
transglutaminase on rice flour properties. J. Food Eng. 84:132-139.
Morita, N., Hirata, C., Park, S. H., and Mitsunaga, T. 2001. Quinoa flour
as a new foodstuff for improving dough and bread. J. Appl. Glycosci.
48:263-270.
Ngarize, S., Adams, A., and Howell, N. K. 2004. Studies on egg albumen
and whey protein interactions by FT-Raman spectroscopy and rheol-
ogy. Food Hydrocolloids 18:49-59.
[Received December 22, 2008. Accepted March 9, 2009.]
392 CEREAL CHEMISTRY
Park, S. H., and Morita, N. 2005. Dough and breadmaking properties of
wheat flour substituted by 10% with germinated quinoa flour. Food
Sci. Technol. Int. 11:471-476.
Park, H. S., Maeda, T., and Morita, N. 2005. Effect of whole quinoa
flours and lipase on the chemical, rheological and breadmaking charac-
teristics of wheat flour. J. Appl. Glycosci. 52:337-343.
Pedersen, B., Hallgren, L., Hansen, I., and Eggum, B. 1987. The nutritive
value of amaranth grain (Amaranthus caudatus) as a supplement to
cereals. Plant Foods Human Nutr. 36:325-334.
Repo-Carrasco, R. 1992. Andean crops and infant nourishment. Institute
of Development Studies Report B 25. University of Helsinki: Finland.
Repo-Carrasco, R., Espinoza, C., and Jacobsen, S. E. 2003. Nutritional
value and use of the Andean crops quinoa (Chenopodium quinoa) and
kaniwa (Chenopodium pallidicaule). Food Reviews Int. 19:179-189.
Rosell, C. M., and Collar, C. 2009. Effect of temperature and consistency
on wheat dough performance. Int. J. Food Sci. Technol. 44:493-502.
Rosell, C. M., Collar, C., and Haros, M. 2007. Assessment of hydrocol-
loid effects on the thermo-mechanical properties of wheat using the
Mixolab. Food Hydrocolloids 21:452-462.
Schoeneberger, H., Moron, S., and Gross, R. 1987. Safety evaluation of
water debittered Andean lupins (Lupinus mutabilis): 12-week rat feed-
ing study. Plant Foods For Human Nutr. 37:169-182.
Schoenlechner, R., Siebenhandl, S., and Berghofer, E. 2008. Pseudocere-
als. In: Gluten free cereal, products and beverages. E. K. Arendt and F.
Dal Bello, eds. Elsevier: London.
Tipples, K. H., D’Appolonia, B. L., Dirks, B. M., Hert, R. F., Kite, F. E.,
Matsuo, R. R., Patton, J., Ranum, P., Shuey, W. C., and Webb, B. D.
1980. Uses and applications. Page 12 in: The Amylograph Handbook.
W. C. Shuey and K. Tipples, eds. AACC International: St. Paul, MN.
Uriyapongson, J., and Rayas-Duarte, P. 1994. Comparison of yield and
properties of amaranth starches using wet and dry-wet milling proc-
esses. Cereal Chem. 71:571-577.