J Agro Crop Sci (2016) ISSN 0931-2250

HEAT STRESS

Response of Mung Bean (Vigna radiata (L.) R. Wilczek) to an

Increasing Natural Temperature Gradient under Different
Crop Management Systems
M. A. P. W. K. Malaviarachchi1,2, W. A. J. M. De Costa2,3, J. B. D. A. P. Kumara2,4,
L. D. B. Suriyagoda2,3 & R. M. Fonseka2,3
1 Field Crops Research and Development Institute, Mahailluppallama, Sri Lanka
2 Postgraduate Institute of Agriculture, University of Peradeniya, Peradeniya, Sri Lanka
3 Department of Crop Science, Faculty of Agriculture, University of Peradeniya, Peradeniya, Sri Lanka
4 Faculty of Agriculture, Sabaragamuwa University of Sri Lanka, Belihul Oya, Sri Lanka

Keywords Abstract
adaptation systems; climate change;
integrated pest management; mulching; Increasing temperatures pose a significant threat to crop production in the tro-
temperature sensitivity pics. A field experiment was conducted with mung bean at three locations in Sri
Lanka representing an increasing temperature gradient (24.4–30.1 °C) during
Correspondence two consecutive seasons to (i) determine the response of mung bean to increasing
W. A. J. M. De Costa
temperature and (ii) test a selected set of crop management practices aimed at
Department of Crop Science
decreasing essential inputs such as water, synthetic pesticides and inorganic nitro-
Faculty of Agriculture
University of Peradeniya gen fertilizer. The control treatment (T1) consisted of standard crop management
Peradeniya 20400 including irrigation, chemical crop protection and inorganic fertilizer applica-
Sri Lanka tion. Adaptation system 1 (T2) included mulching with rice straw at 8 t ha 1
Tel.: +94(0)714430572 with 30 % less irrigation and crop protection and nutrient management as in T1.
Fax: +94(0)812395110 Adaptation system 2 (T3) included crop protection using a pretested integrated
Email: [email protected]
pest management package with water and nutrient management as in T2. In
adaptation system 3 (T4), 25 % of the crop’s nitrogen requirement was given as
Accepted March 17, 2015
organic manure (compost) at 0.8 t ha 1 while 75 % was given as inorganic fertil-
doi:10.1111/jac.12131 izer with water management and crop protection as in T3. Durations of both pre-
and post-flowering phases were reduced with increasing temperature. In the
warmer (25.4–30.1 °C) yala season, seed yield (Y) of T1 decreased with increasing
temperature at 366 kg ha 1 °C 1. However, in maha season, Y did not show a
significant relationship across the narrower temperature gradient from 24.4 to
25.8 °C. Pooling the data from both seasons showed a second-order polynomial
response with an optimum temperature of 26.5 °C. In addition to shortened
durations, reduced crop growth rates and reduced pod numbers per plant were
responsible for yield reductions at higher temperatures. In yala, yields of all adap-
tation systems at all locations were on par with yields of the respective controls.
Furthermore, yala yields of T2 and T3 were less sensitive than T1 to increasing
temperatures (265 and 288 kg ha 1 °C 1). In maha, T3 and T4 had greater yields
than the control at the relatively cooler site while having lower yields than the
control at the warmer site. Maha yields of T2 were on par with the control at both
temperature regimes. While demonstrating the significant temperature sensitivity
of mung bean yields, results of the present work showed that components of the
tested adaptation systems could be promoted among smallholder farmers in Asia,
especially in view of their long-term environmental benefits and contributions to
sustainable agriculture in a warmer and drier future climate.

© 2015 Blackwell Verlag GmbH, 202 (2016) 51–68 51

Malaviarachchi et al.

agronomic management practices aimed at decreasing essen-

Introduction
tial inputs in crop management was evaluated at different
Pulse crops provide ample opportunities to diversify crop- growing season temperatures. These included decreasing the
ping systems and to assist in managing the risks involved in irrigation water input by conserving soil moisture via mulch-
subsistence agriculture in the tropics amidst unpredictable ing, decreasing the use of synthetic pesticides by an integrated
weather and market patterns (Siddique et al. 2012). Mung pest management (IPM) package and decreasing the use of
bean, one of the important pulse crops in Asia, occupies a inorganic nitrogen fertilizer by replacing part of the crops’
prominent place in a large number of tropical cropping nitrogen requirement by organic manure. Evaluation of these
systems (Weinberger 2003), because of its importance as a modified agronomic practices, referred to as ‘adaptation sys-
source of high-quality protein in the cereal-based diets of tems’ in the present study, becomes relevant because adapta-
many people (Bogahawatte and Kailasapathy 1986, Thi- tion to climate change in relation to agriculture, especially in
rumaran and Seralathan 1988, Tharanathan and Maha- the low-income farming systems in the tropics, necessitates
devamma 2003). It is predominantly grown in the the adjustment of agronomic practices, agricultural processes
subhumid and semi-arid tropics in low productive rainfed and capital investments in response to climate change threats
farming systems (Lawn and Ahn 1985) where yield varia- (Turner 2004, Easterling et al. 2007, Turner 2008, Siddique
tion is high due to many biotic and abiotic stresses such as et al. 2012). If effective, these practices could form the core
drought (De Costa et al. 1999), a consequence of erratic of an agronomic package of increased resilience to future cli-
rainfall, high temperature (Cooper et al. 2009) and pest, mate change to ensure sustainable crop production in
disease and weed incidences (Allen and Lenn
e 1998, Siddi- legume-based cropping systems. Introduction of such culti-
que et al. 2012). Thus, even though mung bean has a high vation technology is an urgent need in view of the fact that
potential yield, farmers have failed to realize this potential yields of legumes in the tropics have not shown sustained
yield at the field level (Hewavitharana et al. 2010, Siddique increasing trends during the last few decades despite the
et al. 2012). These stresses are highly likely to be aggravated availability of improved varieties and more resource-efficient
in the future with long-term climate change (Tubiello et al. agronomic practices (Siddique et al. 2012).
2007, Lobell and Gourdji 2012). Temperature increases will
affect a wide range of physiological (Prasad et al. 2003,
Materials and Methods
Porter and Semenov 2005, Hatfield et al. 2011, Lobell and
Gourdji 2012) and phenological (Summerfield and Lawn
Experimental sites
1987, Rawson 1992, Summerfield et al. 1997, Lobell and
Gourdji 2012) processes of the crop and increase the fre- A multilocation field experiment was conducted in the sub-
quency of heat stress (Gourdji et al. 2013, Teixeira et al. humid zone of Sri Lanka during two consecutive seasons
2013). In addition, temperature variations will cause from November 2012 to February 2013 (known locally as
changes in biotic factors such as pest and disease dynamics the maha season) and from May 2013 to August 2013 (yala
(Daugherty et al. 2009, Karuppaiah and Sujayanad 2012). season). The experimental locations were the Experimental
Furthermore, rates of microbial decomposition of organic Farm of the University of Peradeniya, Kundasale (KD)
matter would increase with increasing soil temperatures (07.28°N, 80.69°E, altitude 367 m above mean sea level),
(Davidson and Janssens 2006) thus influencing soil fertility. the Field Crops Research and Development Institute,
Therefore, development of agronomic management pack- Mahailluppallama (MI) (08.60°N, 80.27°E, 137 m amsl)
ages to increase productivity and resilience of tropical crop- and the Regional Agricultural Research and Development
ping systems, which include mung bean, to the adverse Centre, Kilinochchi (KN) (09.35°N, 80.41°E, 15 m amsl).
impacts of climate change is of paramount importance The three locations represented three different agro-ecolog-
(Siddique et al. 2008). ical regions of Sri Lanka (Punyawardane 2008), with KD in
Even though the environmental control of the flowering IM3c (mid-elevation subhumid zone), MI in DL1b and KN
process of mung bean has been studied extensively in con- in DL3 (both in the low-elevation subhumid zone). The
trolled environmental conditions (Summerfield and Lawn annual average rainfalls of the above agro-ecological
1987, Imrie and Lawn 1990) and in the field (Ellis et al. regions are 1100–1200, 900–1100 and 800–1100 mm,
1994), limited experimental work has been carried out on the respectively (Punyawardane 2008). The two cropping sea-
effects of increasing temperature on growth, biomass parti- sons differed in terms of their rainfall. Maha is the main
tioning and yield of field-grown mung bean in the tropics. cropping season receiving rains from the north-east mon-
Therefore, the primary objective of this study was to deter- soon with the seasonal rainfall ranging from 300 to
mine the response of growth, biomass partitioning and yield 800 mm depending on the location within the subhumid
of mung bean to increasing growing season temperature. As zone of Sri Lanka (Panabokke and Punyawardane 2009). In
a secondary objective, effectiveness of a selected set of contrast, yala is the minor cropping season with a lower

52 © 2015 Blackwell Verlag GmbH, 202 (2016) 51–68

 Temperature Response of Mung Bean

seasonal rainfall of 100–400 mm from the south-west mon- around them to drain excess water from high-intensity
soon, which is weakened when it reaches the subhumid rains especially during maha. The planting dates in maha
zone. The soils at MI, KN and KD are Rhodustalfs, Eutrus- were 4th January 2013, 30th December 2012 at KD and MI
tox and Rhodudults, respectively, with a moderately well- while they were 9th May 2013, 10th May 2013 and 10th
drained sandy clay loam texture (Panabokke 1996). The June 2013 in yala at KD, MI and KN, respectively. Even
initial soil N, P, K and organic matter content at a depth of though crops were established at KN in maha also, they
0–30 cm were 0.089 %, 24.8, 120 mg kg 1 and 1.14 % at were destroyed by one event of excessive rainfall at the early
KD, 0.099 %, 38, 176 mg kg 1 and 0.97 % at MI, and vegetative stage. An inter-row spacing of 30 cm and intra-
0.068 %, 29, 292 mg kg 1 and 0.9 % at KN. row spacing of 10 cm were used. Two seeds per hill were
sown at the beginning and thinned out to have a plant den-
sity of 33.33 m 2. An application of fertilizer (35 kg ha 1
Experimental treatments and design
of urea, 100 kg ha 1 of triple super phosphate and
The experimental treatments were four different crop man- 75 kg ha 1 of muriate of potash) at sowing and a top
agement systems in a Randomized Complete Block Design dressing (30 kg ha 1 of urea) was applied at 50 % flower-
with three replicates at each site. T1, which was the control, ing in all experiments for treatments 1 (T1), 2 (T2) and 3
included the current recommended management practices (T3). In contrast, 25 % of urea was reduced at each dress-
in terms of irrigation, fertilization and crop protection by ing for treatment 4 (T4) while applying 0.8 t ha 1 of com-
the Department of Agriculture, Sri Lanka. T2, T3 and T4 post as a replacement for the reduced N from urea. Weed
were different adaptation systems consisting of modified control was performed manually. Depending on the experi-
agronomic packages. Adaptation system 1 (T2) included mental treatment, pests and diseases were controlled by
mulching with rice straw at 8 t ha 1 for the conservation periodic spraying of recommended chemicals or by an IPM
of soil moisture with current recommended fertilization package (explained in the following section). Because of
and crop protection practices which are predominantly asynchronous pod maturity of mung bean, the crops were
based on inorganic fertilizer and synthetic pesticides. Adap- harvested in several picks with the final harvests taken on
tation system 2 (T3) included mulching with modified crop 15th March 2013 and 10th March 2013 in maha at KD and
protection to include an IPM package with recommended MI and on 13th July, 9th July 2013 and 4th August 2013 in
nutrient management practices. Adaptation system 3 (T4) yala at KD, MI and KN, respectively.
included mulching, modified crop protection (i.e. IPM)
and modified nutrient management. Modified nutrient
Measurements and data analysis
management included the addition of 25 % of the crop’s
nitrogen requirement through organic manure in the form Daily weather data were recorded using an automated
of compost (0.8 t ha 1) while providing 75 % of the nitro- weather station (Watch dog, 2900 ET) located at the exper-
gen requirement through inorganic fertilizer. Because of imental sites. A destructive sampling of four randomly
the conserved soil moisture due to mulching, all adaptation selected plants per plot was taken to measure the leaf area
systems received 30 % less irrigation than the control. The index at 50 % flowering (LAI50fl), total biomass at 50 %
IPM package, designed after preliminary laboratory testing, flowering (TBM50fl) and total biomass at harvest (TBMh)
included a soil application of a 2 % bleach solution after while an area of 1 m2 from the middle of the plot was used
crop establishment for eliminating adverse effects of soil- for measuring seed yield and yield components. LAI50fl was
borne pathogens, weekly foliar applications of a baking measured using an automatic leaf area meter (LI 202, CID
soda (NaHCO3) solution, monthly applications of a talc- Bio-Science Inc., Camas, WA, USA). Dry weights of roots,
based biopesticide (Bacillus megaterium) to the foliage, a stems, leaves and pods were measured by oven drying at
one-time application of neem seed kernel extract solution 60 °C to a constant weight. Crop growth up to 50 % flow-
for the control of insect pests and establishment of two ering (CGR50fl) was calculated as the ratio between TBM50fl
border rows of maize around each plot. The plots contain- and the duration from germination up to 50 % flowering.
ing the IPM treatments (i.e. T3 and T4) were separated Post-flowering crop growth rate (CGRpfl) was calculated as
from plots containing conventional crop protection (T1 the ratio between total biomass increment from 50 %
and T2) by a 20 m wide barrier of maize rows. flowering up to final harvest (i.e. TBMh–TBM50fl) and the
duration from 50 % flowering to final harvest.
The count data were subjected to nonparametric analysis
Crop establishment and maintenance
(PROC CATMOD) while parametric data were analysed
Mung bean (variety MI 6) was established by direct seeding using the general linear model (PROC GLM) using the
after disc-ploughing and harrowing the experimental field. software, STATISTICAL ANALYSIS SYSTEM (SAS), version 9.0 (Cary,
Plots (5 m 9 5 m) were prepared with shallow drains NC, USA). Since the season was nested within locations

© 2015 Blackwell Verlag GmbH, 202 (2016) 51–68 53

Malaviarachchi et al.

and treatments (i.e. crop management systems) were nested (a)

within locations and seasons, it was first tested whether
effects of season and effects of crop management systems
were significant within the nested system using ‘Season
(Location)’ and ‘Treatment (Season Location)’ as error
terms in the analysis of variance. Thereafter, individual sea-
sonal effects, location effects and treatment effects were
examined. Duncan’s new multiple range test (DNMRT)
was used to examine the significance of differences among
(b)
treatments at P = 0.05.
The responses of growth and yield parameters to the
temperature gradient represented by the different sites were
quantified by fitting linear or second-order polynomial
functions against the mean location temperature during
each season using regression analysis. To examine the
impacts of temperature extremes, similar regression analy-
ses were performed with mean seasonal minimum and
maximum temperatures as the respective independent vari-
ables. Linear correlation analysis was carried out to exam- Fig. 1 Seasonal variation of daily rainfall at Kundasale (a) and Mahail-
ine the contribution of different yield components to the luppallama (b) during maha 2012/2013 and yala 2013. Kilinochchi did
not receive any rainfall during yala 2013.
observed variation in seed yield across the temperature gra-
dient as represented by different locations and seasons.
Thermal durations for 50 % flowering and maturity were of phenology, growth and yield formation of mung bean
calculated using daily mean temperatures and a base tem- crops among the different sites primarily represented the
perature of 8 °C (Ellis et al. 1994). The dates of final har- response of these processes to variation in the site tempera-
vest were taken for the calculation of thermal durations. ture regimes. KD experienced a lower daily mean tempera-
The relationships between leaf area index at 50 % flowering ture regime than MI and KN in both seasons (Fig. 2a). KN
(LAI50fl) and total biomass at harvest (TBMh) and between had a higher temperature regime than MI during the sec-
LAI50fl and seed yield across the different locations and sea- ond season. The respective seasonal mean temperatures at
sons were quantified by fitting second-order polynomial KD and MI were 24.4 (range 20.3–27.0 °C) and 25.8 °C
functions with the intercept fixed at zero. The LAI50fl at (23.1–28.3 °C) during the maha season while it was 25.4
which maximum TBMh and seed yield were achieved (i.e. (22.5–27.9 °C), 28.5 (25.2–30.4 °C) and 30.1 °C (29.8–
the optimum LAI50fl) was estimated by taking the first 31.3 °C), respectively, at KD, MI and KN in the yala sea-
derivative of the polynomial function and equating it to son. The respective mean minimum seasonal temperatures
zero. (Tmin) also were higher in yala than in maha and showed
an increasing trend KD < MI < KN in both seasons
(Fig. 2b). Mean maximum seasonal temperatures (Tmax) in
Results
yala also showed a similar trend among locations (Fig. 2c).
However, in maha, KD had a higher mean seasonal Tmax
Variation of meteorological conditions during the two
(31.3 °C) than MI (30.0 °C).
seasons at different locations
Except in the yala season at KD, the seasonal totals of
In maha 2012/2013, a well-distributed rainfall pattern was precipitation and seasonal means of daily mean tempera-
experienced at both sites (Fig. 1a,b). In contrast, yala 2013 tures experienced in the two seasons of the present experi-
was characterized by substantially lower and sporadic rain- ment were within the range of variation shown during the
fall. The total rainfall during the first cropping season at last decade at the respective locations (Table 1). In com-
KD and MI was 530 mm and 325 mm with 22 and 27 rain parison with the respective decadal means and ranges, the
days, respectively. The corresponding values for the second yala at KD experienced less rainfall, but was cooler.
season were 2 and 26 mm with 3 and 13 rain days at KD Daily irradiance of total solar radiation showed substan-
and MI, respectively. No rainfall was received at KN during tial fluctuations during the course of both seasons at all
yala 2013. Even though there was substantial variation in sites (Fig. 3), with the yala season showing a higher level
seasonal rainfall among sites in both seasons, supplemen- (10.82–27.47 MJ m 2 day 1) than the maha season (8.98–
tary irrigation ensured that the crops were grown without 23.40 MJ m 2 day 1). Despite the absence (KN) or very
any limitation of water. Therefore, the observed variation low (KD and MI) rainfall in yala, there were an appreciable

54 © 2015 Blackwell Verlag GmbH, 202 (2016) 51–68

 Temperature Response of Mung Bean

(a) in yala than in maha at KD and MI. In contrast, the post-

flowering total irradiance was 16 % and 20 % lower in yala
as compared to maha at KD and MI. In comparison with
KD and MI, KN received lower seasonal total radiation in
yala. Both pre- and post-flowering irradiance totals were
also lower at KN than at KD and MI, mainly because of the
shorter durations of both phenological stages (Table 3).

Variation of crop phenology across different locations and

(b) cropping seasons
Crop management systems and their interaction with sea-
son and location did not show significant variation with
respect to days to 50 % flowering and days to final harvest
whereas there was a significant location effect on these phe-
nological parameters (Table 3). Mung bean at KD reached
50 % flowering in 38 and 36 days in maha and yala,
respectively. In comparison, the same developmental stage
(c) was reached within a shorter period (31 and 33 days) at MI
and 30 at KN. However, the total crop duration was the
same (70 days) at both sites in the first season. In the sec-
ond season, KD reported the longest crop duration
(65 days) while MI and KN recorded lower durations (60
and 55 days, respectively). The post-flowering duration
(i.e. that from 50 % flowering to final harvest) in maha
was greater at MI (39 days) than at KD (32 days) thus
showing an increase with increasing mean location temper-
ature. However, the opposite trend was shown in yala with
Fig. 2 Seasonal variation of daily mean (a), minimum (b) and maximum the post-flowering duration decreasing from 29 days at KD
(c) temperatures during maha 2012/2013 and yala 2013 at the different to 25 days at KN (Table 3). The respective thermal dura-
experimental sites.
tions for 50 % flowering and crop maturity showed only
narrow variations across locations and seasons. The ther-
number of days of relatively lower irradiance mal duration for 50 % flowering varied between 493 and
(<15 MJ m 2 day 1) due to cloudiness. Because of the 550 °Cd with an average of 521 °Cd in the first season
higher irradiance levels in yala, the seasonal total solar radi- while it was 575–599 °Cd with an average of 585 °Cd in
ation was greater in yala as compared to maha at KD and the second season (Table 3). On the other hand, the corre-
MI (Table 2). Notably, total irradiance from sowing to sponding thermal durations for maturity ranged from
50 % flowering was, respectively, 16 % and 52 % greater 1101 °Cd at KD to 1207 °Cd at MI with an average of

Table 1 Mean and range of seasonal total rainfall and seasonal mean daily mean temperature at the different experimental sites during the period
from 2004 to 2013

Mean seasonal total rainfall (mm) Mean seasonal daily mean temperature (°C)

Site Maha Yala Maha Yala

KD 288.9 (530.0)1 218.5 (2) 24.4 (24.4) 26.1 (25.4)

(74.1–659.2)2 (91.2–430.6) (23.7–24.9) (25.7–26.9)
MI 302.1 (325) 79.1 (26) 26.5 (25.8) 28.8 (28.5)
(85.3–618.4) (11.7–153.1) (25.9–27.2) (28.4–29.2)
KN * 29.3 (0) (0–95.7) * 28.9 (30.1) (29.1–31.1)

KD, Kundasale; MI, Mahailluppallama; KN, Kilinochchi.

1
Value observed in the present experiment is given in italics.
2
Range during the last decade.
*Experiment was abandoned due to heavy rains.

© 2015 Blackwell Verlag GmbH, 202 (2016) 51–68 55

Malaviarachchi et al.

1154 °Cd in the first season while it was 1052–1145 °Cd that of the control in terms of LAI50fl and CGRpfl. During
with an average of 1108 °Cd in the second season. the second season (i.e. yala), at MI growth parameters of
crops in all adaptation systems were not significantly differ-
ent from those of the control. At KN also, during yala crop
Variation of vegetative growth among different crop
growth in all adaptation systems showed LAI50fl and
management systems in different locations and cropping
CGR50fl which were on par with the control. During the sec-
seasons
ond season, at KD also, several adaptation systems showed
The seasonal effect was significant (P < 0.05) for the growth performances which were not significantly different
majority of growth and yield parameters (Table 4), with from the control. In particular, T4 showed similar growth as
crop growth rate and total biomass at 50 % flowering the control in terms of all three growth parameters. More-
(CGR50fl and TBM50fl) and harvest index (HI) being the over, T3 at KD in yala had similar growth as compared to
exceptions. The effect of different crop management sys- the control up to 50 % flowering while T2 had similar
tems was significant for all growth and yield parameters growth during the post-flowering period. In general, during
except HI. As the treatment effect was nested within differ- yala, a majority of crops growing under the three adapta-
ent locations and seasons, the observed variation of growth tion systems showed growth, which was on par with the
and yield parameters is presented for each location and sea- control at their respective locations. Location mean growth
son separately. performance as measured by all three parameters was supe-
Crop growth in different adaptation systems (i.e. T2, T3 rior in yala as compared to maha at both KD and MI. Such
and T4) in comparison with the control (i.e. T1) differed a seasonal comparison of crop growth was not possible for
with season and location (Table 5). For example, at KD KN as the maha crops were destroyed by excessive rain.
during the first season (i.e. maha), T3 treatment (i.e. mulch- Notably, all three growth parameters showed decreasing
ing + IPM + standard nutrient management) showed trends at the warmer locations (i.e. MI and KN relative to
LAI50fl and CGR50fl values which were not significantly dif- KD, Table 4) indicating negative impacts on mung bean
ferent from those of T1, which had the highest among treat- growth with increasing temperature.
ments. Both T3 and T4 (i.e. mulching + IPM + 25 % N
from organic source) had substantially greater CGRpfl than
Variation of biomass partitioning and yield among
the control in maha at KD. On the other hand, at MI during
different crop management systems in different locations
the first season, crop growth in T2 (mulching + standard
and cropping seasons
crop protection and nutrient management) was on par with
In the maha season at KD, two adaptation systems (i.e. T3
and T4) produced significantly higher seed yields (Y) than
the control (Table 6). This was primarily due to their
greater total biomass at harvest (TBMh), which was mainly
because of their greater post-flowering crop growth rates
(CGRpfl) (Table 5). Conversely, at MI during maha, the
same two adaptation systems (i.e. T3 and T4) had lower
yields than the control. Here also, lower TBMh brought
about by lower CGRpfl were responsible for the lower Y of
T3 and T4. In contrast to the maha season, during yala,
Fig. 3 Seasonal variation of daily incident solar radiation during maha yields of all adaptation systems at all locations were not sig-
2012/2013 and yala 2013 at the different experimental sites. nificantly (P = 0.05) different from the respective control

2
Table 2 Cumulative totals of incident solar radiation (MJ m day 1) during pre- and post-flowering stages of mung bean grown at different
locations and cropping seasons

From sowing to 50 % flowering From 50 % flowering to final harvest From sowing to final harvest

Location Season 1 Season 2 Season 1 Season 2 Season 1 Season 2

Kundasale 646.38 747.66 575.41 486.04 1221.79 1233.69

Mahailluppallama 481.72 731.51 701.16 561.43 1182.88 1292.94
Kilinochchi * 561.43 * 459.64 * 1065.24

*Experiment was abandoned due to heavy rains.

Season 1 – Maha 2012/2013; Season 2 – Yala 2013.

56 © 2015 Blackwell Verlag GmbH, 202 (2016) 51–68

 Temperature Response of Mung Bean

Table 3 Days to 50 % flowering and final harvest of mung bean grown at different locations and cropping seasons

Thermal duration1 for Days to final Thermal duration for

Days to 50 % flowering 50 % flowering (°C days) harvest2 final harvest (°C days)

Location Season 1 Season 2 Season 1 Season 2 Season 1 Season 2 Season 1 Season 2

Kundasale 38 36 549.7 580.8 70 65 1101.4 1051.6

Mahailluppallama 31 33 493.1 598.5 70 60 1207.3 1145.4
Kilinochchi * 30 * 574.7 * 55 * 1125.7

*Experiment was abandoned due to heavy rains.

1
Thermal duration was calculated using a base temperature of 8 °C (Ellis et al. 1994).
2
Final harvest of several picks.

Table 4 Probability levels of significance of the season and treatment effects in the nested treatment structure for the measured growth and yield
parameters of mung bean crops grown under different crop management systems, locations and seasons

Probability

Source of variation LAI50fl TBM50fl TBMh CGR50fl CGRpfl Yield HI

Season (location) 0.0003 0.222 0.034 0.100 0.024 0.048 0.582

Treatment (location season) 0.0012 0.027 0.001 0.004 0.009 0.021 0.601

TBM50fl and TBMh, total crop biomass per unit land area at 50 % flowering and final harvest, respectively; CGR50fl and CGRpfl, respective crop
growth rates from germination up to 50 % flowering from 50 % flowering up to final harvest; HI, harvest index.
Each value is the probability of a given source of variation being due to random variation.

Table 5 Variation of LAI at 50 % flowering (LAI50fl) and crop growth rates up to 50 % flowering (CGR50fl) and during the post-flowering period
(CGRpfl) in mung bean crops growing under different crop management systems at the three experimental locations
2
LAI50 fl CGR50fl (g m day 1) CGRpfl (g m 2
day 1)

Site System Season 1 Season 2 Season 1 Season 2 Season 1 Season 2

Kundasale T1 2.48 a 5.57 a 6.78 a 9.77 ab 7.55 b 23.49 a

T2 1.93 b 2.72 b 5.46 ab 4.53 b 8.13 b 19.91 ab
T3 2.19 ab 5.15 a 7.25 a 11.49 a 13.08 a 12.50 b
T4 1.93 b 3.59 b 4.97 b 8.18 ab 14.76 a 18.42 ab
Mean 2.13 B 4.14 A 6.11 B 8.37 A 10.88 B 18.11 A
CV (%) 9.8 22.7 27.4 33.7 25.4 23.2
Mahailluppallama T1 1.83 a 2.43 a 6.10 a 5.83 a 11.61 ab 14.12 a
T2 1.88 a 2.55 a 4.58 b 5.03 a 13.29 a 11.93 a
T3 1.07 b 2.10 a 3.33 c 6.58 a 7.89 b 13.13 a
T4 1.25 b 2.31 a 4.05 bc 4.94 a 8.54 b 12.66 a
Mean 1.51 B 2.35 A 4.51 B 5.60 A 10.33 B 12.96 A
CV (%) 12.8 13.7 9.6 22.6 21.6 12.5
Kilinochchi T1 * 1.61 a * 4.55 a * 15.02 a
T2 * 1.87 a * 5.44 a * 13.13 b
T3 * 1.72 a * 5.05 a * 6.44 c
T4 * 1.60 a * 4.19 a * 7.89 c
Mean 1.70 4.81 10.79
CV (%) 16.7 16.8 6.3
Rate of change with increasing temperature (°C 1)1 0.45 0.51 1.14 0.80 ns 1.71
R2 0.53 0.58 0.40 0.50 – 0.53

*Experiment was abandoned due to heavy rains.

Within each cropping season and location, treatment (i.e. crop management systems) means with the same lower-case letter are not significantly
different at P = 0.05. Within each location, seasonal means with the same upper-case letter are not significantly different at P = 0.05.
1
Slopes of linear regressions of treatment means against mean seasonal location temperature. A negative slope indicates significant reduction per
unit increase in temperature. ns – No significant relationship. CV - Coefficient of Variation.

© 2015 Blackwell Verlag GmbH, 202 (2016) 51–68 57

Malaviarachchi et al.

Table 6 Variation of seed yield (Y), total biomass at final harvest (TBMh) and harvest index (HI) in mung bean crops growing under different crop
management systems at the three experimental locations

Seed yield (kg ha 1) TBMh (kg ha 1) HI

Site System Season 1 Season 2 Season 1 Season 2 Season 1 Season 2

Kundasale T1 2041 b 3627 a 4992 b 8628 a 0.409 a 0.409 a

T2 1908 b 2969 a 4677 b 7405 a 0.408 a 0.399 a
T3 2998 a 2791 a 6939 a 7762 a 0.433 a 0.360 a
T4 2786 a 3189 a 6611 a 7230 a 0.422 a 0.443 a
Mean 2433 B 3144 A 5805 B 7756 A 0.418 A 0.403 A
CV (%) 14.1 26.4 8.4 16.2 13.3 13.7
Mahailluppallama T1 2113 a 2167 a 6416 a 6723 a 0.326 a 0.325 a
T2 1996 a 2057 a 6604 a 5714 a 0.304 a 0.369 a
T3 1115 b 1752 a 4108 b 6637 a 0.284 a 0.268 a
T4 1232 b 1724 a 4586 b 5933 a 0.271 a 0.290 a
Mean 1614 B 1925 A 5429 B 6252 A 0.296 A 0.313 A
CV (%) 16.1 13.0 16.9 14.0 17.7 21.6
Kilinochchi T1 * 1983 a * 5420 a * 0.368 a
T2 * 1743 a * 4748 ab * 0.370 a
T3 * 1469 a * 4098 ab * 0.365 a
T4 * 1423 a * 3389 b * 0.423 a
Mean 1655 4414 0.382
CV (%) 17.2 21.5 18.8
Temperature sensitivity (°C 1)1 585 327 ns 681 0.087 2

R2 0.45 0.85 0.79 0.93 0.60

*Experiment was abandoned due to heavy rains.

1
See the footnote of Table 4 for an explanation of mean separation symbols and temperature sensitivity.
2
Significant second-order negative polynomial relationship with a minimum estimated HI of 0.333 at 27.8 °C.

yields (Table 6). With the exception of TBMh of T4 at KN, ature gradient (P = 0.0044 and P = 0.002, respectively).
both TBMh and HI of all adaptation systems at all locations TBMh and seed yield decreased by 452 and
in yala were on par with their respective controls. Mean 366 kg ha 1 °C 1 across the increasing temperature gradi-
yields across different management systems were signifi- ent (Fig. 4a,b). However, when the linear regressions were
cantly greater in yala as compared to maha at both KD and carried out across the mean seasonal temperatures of the
MI. These superior yields in yala were due to the greater two locations in maha, no significant (P = 0.05) relation-
TBMh, which was in turn brought about by greater CGR50fl ships were found (Fig. 4a,b) probably because of the very
and CGRpfl (Table 5). In both seasons, seasonal means of narrow range of mean seasonal temperatures in maha. On
seed yield, TBMh and HI were lower in the warmer loca- the other hand, the HI of T1 showed significant (P < 0.05)
tions (i.e. MI and KN as compared to KD). Relationships decreasing trends with increasing temperature in both sea-
of these parameters with temperature will be examined in sons (Fig. 4c) at 0.115 and 0.0236 °C 1 in maha and yala,
the next section. respectively. Pooling the data from both seasons showed
second-order polynomial response patterns for both Y and
TBMh against mean seasonal temperature (Fig. 4a,b). The
Response of growth, biomass partitioning and yield to the
estimated optimum temperatures for Y and TBMh were
temperature gradient across locations during different
26.5 and 27.3 °C, respectively. The estimated maxima for Y
cropping seasons
and TBMh at the respective optimum temperatures were
Crops growing under currently recommended standard man- 2671 and 7095 kg ha 1. HI also showed a negative second-
agement practices (T1) order polynomial response with a minimum HI of 0.35
Variation of crop growth and yield among locations in the being reached at 28.0 °C (Fig. 4c).
control treatment (i.e. T1) was considered to explore the
response of mung bean to a mean seasonal temperature Response of growth, biomass partitioning and yield of crops
gradient as soil moisture deficits were avoided through irri- growing under different adaptation systems to the tempera-
gation. In the yala season, TBMh and seed yield of T1 had ture gradient during different cropping seasons
significant linear relationships with mean seasonal temper- Within a given season, the rates of biomass and yield
ature across the experimental sites representing the temper- reductions of the different adaptation systems (i.e. T2, T3

58 © 2015 Blackwell Verlag GmbH, 202 (2016) 51–68

 Temperature Response of Mung Bean

(a) (b)

(c)

Fig. 4 Relationships between seasonal mean

location temperature and total biomass at har-
vest (TBMh) (a), seed yield (Y) (b) and harvest
index (HI) (c) in mung bean crops grown with
standard recommended management prac-
tices (i.e. the control treatment, T1). ○ – Maha
2012/2013 season’s data; ▲ – Yala 2013 sea-
son’s data. Each data point is a value for a rep-
licate plot at a given location. Solid lines
represent linear temperature response func-
tions of different seasons. Curves represent
polynomial temperature response functions
for pooled data from both seasons.

and T4) in response to increasing temperature were TBMh and seed yield in maha (P = 0.016 and P = 0.0003,
observed to be within a narrow range (Fig. 5a–d). The respectively) and in yala (P = 0.0005 and P = 0.0005,
respective rates ranged between 563 and 763 kg ha 1 °C 1 respectively) seasons (Fig. 6 with polynomial relationships
and 265 and 389 kg ha 1 °C 1 for TBMh and yield, respec- not shown). Furthermore, when the data from both seasons
tively, during yala 2013. However, in the maha season, were pooled, LAI50fl showed highly significant (P < 0.0001)
TBMh and yield of T3 and T4 decreased at comparatively second-order polynomial relationships with TBMh and
higher rates (2022, 1446 kg ha 1 °C 1 and 1345, seed yield (Fig. 6a,b). Both TBMh and seed yield reached
1110 kg ha 1 °C 1). In contrast, in T2, TBMh increased by their respective maxima of 8137.7 and 3122.0 kg ha 1
1377 kg ha 1 °C 1 whereas yield did not vary significantly when the LAI50fl reached 4.30 and 4.70, respectively.
with temperature. On the other hand, in all adaptation
systems, HI showed significant (P < 0.05) negative second-
Yield components and their relationships to yield
order polynomial relationships with mean location temper-
ature in yala (Fig. 5f). The respective temperatures at All three yield components showed significant variation
which HI was minimum were 28.0, 27.9 and 27.6 °C for between the two seasons (Table 7). On the other hand, the
T2, T3 and T4, respectively, with the estimated minimum different crop management systems (i.e. the treatment
HIs being 0.38, 0.31 and 0.27. In contrast to yala, in maha effect) had significant effects on the pod number (PN) per
only T2 showed a significant negative linear relationship plant only. As the treatment effect was nested within loca-
( 0.036 °C 1) with increasing temperature (Fig. 5e). tions and seasons, variation of yield components in differ-
ent crop management systems is shown separately for
different locations and seasons (Table 8). At KD, PN
Relationships between leaf area index at 50 % flowering,
showed significant variation among treatments in both sea-
total biomass at harvest and seed yield across different
sons. While T3 had significantly greater PN than the con-
experimental treatments, locations and seasons
trol in maha, PN of the other adaptation systems (i.e. T2
When pooled across experimental treatments and loca- and T4) was not significantly different from that of the con-
tions, LAI50fl, one of the principal crop growth parameters trol. In yala at KD, all adaptation systems had PN which
had significant second-order polynomial relationships with were equal to the control. In contrast to KD, at MI, PN did

© 2015 Blackwell Verlag GmbH, 202 (2016) 51–68 59

Malaviarachchi et al.

(a) (b) (c)

(d) (e) (f)

Fig. 5 Relationships between seasonal mean location temperature and total biomass at harvest (TBMh) (a and b), seed yield (Y) (c and d) and harvest
index (HI) (e and f) in mung bean crops grown with different adaptation systems (□ – T2, ○ – T3, ▲ – T4) in maha 2012/2013 (a, c and e) and yala
2013 (b, d and f) seasons. Each data point is a value for a replicate plot at a given location.

(a) (b)

Fig. 6 Relationships between leaf area index at 50 % flowering (LAI50fl) and total biomass at harvest (TBMh) (a) and seed yield (Y) (b) of mung bean
crops grown with standard crop management (i.e. T1) and in different adaptation systems (i.e. T2, T3 and T4) across the temperature gradient repre-
sented by different locations and seasons. Each data point is a value for a replicate plot at a given location. ○ – Maha 2012/2013 season’s data;
▲ – Yala 2013 season’s data.

not differ significantly among crop management systems in than in maha. In contrast to PN, the seed number (SN) per
both seasons. At KN in yala, T3 and T4 had greater PN than pod did not show significant variation between different
the control. Notably, when averaged across treatments, treatments in all sites and seasons. When averaged across
mean PN of both KD and MI was greater (P < 0.05) in yala treatments, mean SN of maha was greater than that of yala

60 © 2015 Blackwell Verlag GmbH, 202 (2016) 51–68

 Temperature Response of Mung Bean

at MI. However, at KD, mean SN did not differ between (P = 0.05) correlations with Y in both seasons. In addition,
the two seasons. Individual seed weight (ISW) did not dif- several significant negative correlations were observed
fer significantly between treatments at all sites and seasons between different yield components in different seasons.
with the exception of yala at MI, where T4 showed lower These included the negative correlations between PN and
ISW than the control. At both KD and MI, seasonal mean SN and between SN and ISW in the maha season and that
ISW was greater in maha than in yala. Of the three yield between PN and ISW in yala.
components, only PN showed significant declining trends Different yield components showed significant positive
with increasing seasonal mean temperatures across loca- correlations with different vegetative growth parameters.
tions in both seasons (Table 8). These correlations were different in the two seasons. For
The PN showed highly significant (P < 0.0001) positive example, PN was positively correlated to LAI50fl, TBM50fl
correlations with seed yield (Y) in both seasons (Table 9). and TBMh in both seasons (Table 9). On the other hand,
The SN was positively correlated (P < 0.0001) to Y in the SN and ISW were not correlated to any of the above growth
yala season only. The ISW did not have significant parameters in maha. However, in yala, SN showed signifi-
cant positive correlations with LAI50fl and TBM50fl while
ISW was negatively correlated to TBMh.
Table 7 Probability levels of significance of the season and treatment
Similarly, correlations of yield components to the HI
effects in the nested treatment structure for the yield components of
were also different in the two seasons. While none of the
mung bean crops grown under different crop management systems,
locations and seasons yield components were significantly correlated to the HI in
maha, PN and SN had significant positive correlations with
Probability HI in yala (Table 9). In both seasons, HI did not show sig-
Pods per Seeds per Individual seed nificant correlations with any of the growth parameters.
Sources of variation plant pod weight Interestingly, a significant positive correlation was observed
between Y and HI in the yala season only. In contrast, Y
Season (location) 0.006 0.001 <0.0001
showed highly significant positive correlations with TBMh
Treatment (location season) 0.012 0.717 0.531
and all other growth parameters (i.e. LAI50fl and TBM50fl)
Each value is the probability of a given source of variation being due to in both seasons. When the data for both seasons were
random variation. pooled, Y showed a much stronger positive correlation with

Table 8 Variation of pod number per plant (PN), seeds per pod (SN) and individual seed weight (ISW) in mung bean crops growing under different
crop management systems at the three experimental locations
1 1
PN pl Seeds pod ISW (mg)

Site System Season 1 Season 2 Season 1 Season 2 Season 1 Season 2

Kundasale T1 13.1 bc 20.8 ab 8.9 a 9.8 a 59.8 a 51.3 a

T2 12.1 c 21.3 a 8.1 a 8.3 a 61.8 a 51.0 a
T3 18.3 a 16.0 b 9.4 a 10.3 a 58.2 a 50.7 a
T4 16.8 ab 19.3 ab 9.2 a 9.8 a 61.8 a 50.7 a
Mean 15.1 B 19.4 A 8.9 A 9.6 A 60.4 A 50.9 B
CV (%) 13.5 12.2 13.5 14.8 4.1 4.2
Mahailluppallama T1 10.4 a 14.3 a 10.7 a 8.8 a 57.2 a 55.8 a
T2 10.5 a 14.3 a 10.1 a 8.7 a 58.0 a 52.4 ab
T3 7.9 a 11.0 a 10.4 a 9.1 a 58.2 a 52.8 ab
T4 9.3 a 12.3 a 10.3 a 8.5 a 56.8 a 50.4 b
Mean 9.5 B 13.0 A 10.4 A 8.8 B 57.5 A 52.9 B
CV (%) 21.9 13.5 8.7 4.9 1.9 4.2
Kilinochchi T1 * 12.7 a * 9.1 a * 54.5 a
T2 * 12.3 a * 9.3 a * 55.1 a
T3 * 9.3 b * 8.4 a * 56.4 a
T4 * 10.3 b * 9.0 a * 54.7 a
Mean 11.2 9.0 55.2
CV (%) 6.8 10.0 4.2
Temperature sensitivity (°C 1)1 4.0 1.8 1.1 ns 2.0 0.9
R2 0.67 0.85 0.79 – 0.61 0.64

*Experiment was abandoned due to heavy rains.

1
See the footnote of Table 4 for an explanation of mean separation symbols and temperature sensitivity.

© 2015 Blackwell Verlag GmbH, 202 (2016) 51–68 61

Malaviarachchi et al.

Table 9 Linear correlation coefficients between seed yield, yield components, total biomass and harvest index of mung bean crops growing under
standard crop management and different crop adaptation systems across different locations in maha 2012/2013 (above the diagonal) and yala
(below the diagonal) seasons
1 1
Yield HI TBMh LAI50fl TBM50fl Pods pl Seeds pod ISW

Yield – 0.053ns 0.774*** 0.678*** 0.651*** 0.896*** 0.232ns 0.113ns

HI 0.580*** – 0.207ns 0.043ns 0.075ns 0.107ns 0.219ns 0.083ns
TBMh 0.783*** 0.042ns – 0.488* 0.365ns 0.563** 0.076ns 0.104ns
LAI50fl 0.554*** 0.174ns 0.572*** – 0.825*** 0.592** 0.390ns 0.096ns
TBM50fl 0.532*** 0.163ns 0.592*** 0.950*** – 0.651*** 0.374ns 0.117ns
Pods pl 1 0.877*** 0.522** 0.703*** 0.370* 0.351* – 0.479* 0.175ns
Seeds pod 1
0.620*** 0.392* 0.289ns 0.588*** 0.550*** 0.295ns – 0.584**
ISW 0.315ns 0.157ns 0.499** 0.261ns 0.250ns 0.489** 0.051ns –

HI, harvest index; TBMh and TBM50fl, total biomass at harvest and 50 % flowering; LAI50fl, leaf area index at 50 % flowering; ISW, mean individual
seed weight.
Each correlation had 24 and 36 data points in maha and yala, respectively. ns – Non-significant at P = 0.05; * – Significant at P < 0.05;
** – Significant at P < 0.01; *** – Significant at P < 0.001.

TBMh (r = 0.778 with P < 0.0001) than with HI data for the two seasons were pooled, HI did not show a
(r = 0.284 with P = 0.028). significant variation across the range of mean seasonal
Tmax.
Effects of seasonal temperature extremes
Discussion
Effects of minimum and maximum temperatures were
examined on mung bean crops growing under standard The principal objective of the present work was to assess
crop management (i.e. T1) (Table 10). When the data for the sensitivity of mung bean yields to increasing tempera-
both seasons were pooled, both TBMh and Y showed signif- ture across different locations representing a natural tem-
icant second-order polynomial relationships with mean perature gradient over two contrasting cropping seasons.
seasonal minimum temperature (Tmin) across different In terms of the seasonal mean temperature, the first season
locations. The respective optimum Tmin for maximum (i.e. maha 2012/2013) had a narrower temperature range of
TBMh and Y were 23.0 and 23.4 °C. Across the two 1.4 °C (from 24.4 to 25.8 °C) across the different locations.
seasons, the HI also showed a negative second-order poly- In contrast, crops in the second season (yala 2013) experi-
nomial response to mean seasonal Tmin with the minimum enced a broader temperature range of 4.7 °C (25.4–
HI at 24.6 °C. When the responses to mean seasonal Tmin 30.1 °C). Comparison with the meteorological data at the
were examined separately for the two seasons, negative lin- locations during the last decade (Table 1) showed that the
ear relationships were shown for Y and HI. Seed yield (Y) two cropping seasons were largely representative of the cli-
showed a greater sensitivity (as indicated by the slope of matic conditions experienced in these locations. A primary
the fitted linear function) to increasing Tmin during the assumption in this multilocational experimental set-up was
warmer yala season with a yield reduction of that temperature would be the principal environmental fac-
422 kg ha 1 °C 1 as compared to the cooler maha season tor controlling growth, biomass partitioning and yield of
( 26 kg ha 1 °C 1). However, the HI showed a greater mung bean crops. Impacts of precipitation variation
sensitivity to increasing Tmin during the maha season. In among locations were largely eliminated by irrigating the
contrast to Y and HI, TBMh showed a second-order poly- crops. Excess water from high-intensity rainfall was
nomial response to increasing Tmin during yala while show- removed by drains between plots. Any effects on crop
ing a positive response in maha. growth and yield formation due to soil variation among
In contrast to the response to Tmin, TBMh and Y showed sites were eliminated by providing recommended doses of
negative linear responses to increasing mean seasonal maxi- nutrients as inorganic (T1, T2 and T3) or a combination of
mum temperatures (Tmax), for both seasons separately and inorganic and organic (T4) fertilizer. Any variation among
when the data for the two seasons were pooled (Table 10). sites in pest and disease incidence was controlled via con-
The sensitivity to increasing mean seasonal Tmax was ventional chemical-based (T1 and T2) or IPM (T3 and T4)
greater in yala for Y. However, the opposite was observed methods. Therefore, it was assumed that the effects of any
for TBMh. The HI showed a positive response to increasing uncontrolled variations among different sites would not
Tmax in maha, but a negative response in yala. When the override and obscure the principal response to the

62 © 2015 Blackwell Verlag GmbH, 202 (2016) 51–68

 Temperature Response of Mung Bean

Table 10 Fitted relationships of total biomass at harvest (TBMh), seed yield (Y) and harvest index (HI) with mean minimum and maximum seasonal
temperatures (Tmin and Tmax) for mung bean crops growing under standard crop management (T1) across three experimental locations and two crop-
ping seasons

Season TBMh (kg ha 1) Y (kg ha 1) HI

Tmin
Maha TBMh = (234*Tmin) + 556 (R2 = 0.348) Y=( 26*Tmin) + 1547 (R2 = 0.008) Y = ( 0.059*Tmin) + 1.603 (R2 = 0.497)
Yala TBMh = ( 512*Tmin2
) + (24034*Tmin) + 273395 Y=( 422*Tmin) + 12993 Y = ( 0.029*Tmin) + 1.151
(R = 0.644) (Optimum Tmin = 23.47 °C)
2
(R2 = 0.533) (R2 = 0.446)
Both1 TBMh = ( 121*Tmin2
) + (5580*Tmin) + 57277 Y=( 2
74*Tmin ) + (3336*Tmin) + 34492 Y = (0.0025*Tmin
2
) (0.1228*Tmin) + 1.91
(R2 = 0.402) (Optimum Tmin = 23.01 °C) (R2 = 0.227) (Optimum Tmin = 23.42 °C) (R2 = 0.08)
(Tmin for minimum HI = 24.56 °C)
Tmax
Maha Y = ( 488*Tmax) + 20273 (R2 = 0.348) Y = ( 54*Tmax) + 3743 (R2 = 0.008) Y = (0.122*Tmax) 3.326
(R2 = 0.497)
Yala Y = ( 416*Tmax) + 19874 (R2 = 0.572) Y = ( 341*Tmax) + 13469 Y = ( 0.022*Tmax) + 1.156
(R2 = 0.508) (R2 = 0.390)
Both1 Y = ( 207*Tmax) + 12543 (R2 = 0.112) Y = ( 234*Tmax) + 9738 (R2 = 0.24) Y = ( 0.0082*Tmax) + 0.6935
(R2 = 0.025)
1
Data from both seasons pooled.

temperature gradient. Accordingly, results of the present scales. As most of the published studies have focused on
study are especially relevant in view of the conclusion of major food crops of global importance, to our knowledge,
Lobell and Burke (2008) that uncertainty in the sensitivity the present work constitutes the first published report on
of crop yields to temperature variation rather than to pre- the sensitivity of mung bean yields to seasonal warming
cipitation variation constitutes a greater proportion to the across a multilocational temperature gradient. The esti-
uncertainty of estimated impacts of climate change on crop mated temperature optimum of 26.5 °C for mung bean in
yields. the present work is higher than the ranges of 22–23 °C esti-
Our results, especially those of the second season where mated for soya bean (Hatfield et al. 2011, Lobell and Gour-
the crops experienced a broader temperature range, show dji 2012) and 23–24 °C for peanut (Prasad et al. 2003) and
that despite being a warm-season grain legume that is red kidney bean, that is Phaseolus vulgaris (Laing et al.
adapted to higher temperature regimes (Rachie and Rob- 1984, Prasad et al. 2002).
erts 1974), mung bean yields are reduced significantly when Analysis of yield responses to seasonal means of mini-
growing temperatures increased from ca. 25–30 °C mum (Tmin) and maximum (Tmax) temperatures
(Fig. 3). It is notable that this range of mean seasonal tem- (Table 10) showed that yields of mung bean (Y) are more
peratures is well within the temperature range across which sensitive to Tmax than Tmin. This is in agreement with the
mung bean is currently grown in different parts of the results of Lobell and Field (2007) who found greater
world (Lawn and Ahn 1985). Furthermore, the current impacts of increasing temperatures on yields of wheat and
growing season temperatures in many parts of the tropics maize when Tmax and Tmin were used instead of seasonal
where mung bean is grown exceed the estimated optimum means of daily mean temperatures (Tavg). This is demon-
temperature of 26.5 °C for maximum yield (Christensen strated by the lower optimum for Tmin (i.e. 23.4 °C) as
et al. 2007, Lobell et al. 2008). Therefore, it is highly likely compared to the optimum for Tavg (26.5 °C) and by the
that future increases in air temperature due to the negative linear response of Y to Tmax. Our analysis also
enhanced greenhouse effect (IPCC 2013) would have nega- showed that mung bean yields are more sensitive to
tive impacts on future mung been yields in Sri Lanka and increasing Tmin and Tmax in the warmer yala season than in
South-East Asia. This is in agreement with the findings of the relatively cooler maha.
many studies in which negative impacts of future warming Results of the present experiment show that high tem-
on the yields of several major food crops (e.g. wheat, maize, perature-induced yield reductions in mung bean occur due
rice and soya bean) have been shown both at the global to reductions in both total biomass at harvest (TBMh) and
(Lobell and Field 2007, Lobell and Gourdji 2012, Teixeira HI. This was shown especially in the second season when a
et al. 2013) and regional (Lobell and Asner 2003, Lobell wider temperature range was experienced. This indicates
et al. 2008, Cooper et al. 2009, Schlenker and Roberts that higher temperatures influence both vegetative and
2009, Schlenker and Lobell 2010, Hatfield et al. 2011) reproductive growth and also both developmental and

© 2015 Blackwell Verlag GmbH, 202 (2016) 51–68 63

Malaviarachchi et al.

growth processes. These observations agree with the period from planting up to 50 % flowering, which allowed
expected responses based on the effects of increasing tem- greater radiation interception and biomass production.
peratures on basic physiological processes (Hatfield et al. The observed relationships of LAI50fl with TBMh (Fig. 6a)
2011, Lobell and Gourdji 2012). The influence of tempera- and seed yield (Fig. 6b) demonstrate the importance of
ture variation on developmental processes could be early vegetative growth in yield determination of mung
observed in the variation of durations to 50 % flowering bean. Therefore, greater irradiance levels during the vegeta-
and maturity, with increasing temperatures clearly hasten- tive phase enabled the yala crops to achieve greater yields
ing the flower initiation and maturity. These shortened than in maha despite the warmer temperatures. This find-
durations clearly contributed to the yield reductions ing indicates that, if irrigation is available, growing crops
observed at higher temperatures. This is in agreement with during periods of greater irradiance, especially during the
the conclusions of reviews of previous experimental work vegetative stage, enable achievement of higher yields at
by Hatfield et al. (2011) and Lobell and Gourdji (2012) higher temperatures.
and the specific predictions of Cooper et al. (2009) for The stronger correlations between yield and TBMh as
yields of peanut and pigeon pea at higher temperatures. compared to those between yield and HI (Table 9) indi-
The estimated thermal durations for flowering and matu- cate that increasing temperatures have a greater influence
rity in the present work were comparable to those of other on processes responsible for biomass production than on
work on mung bean. For example, Chauhan et al. (2010) processes determining biomass partitioning to yield. How-
estimated the thermal time for flowering and maturity for a ever, this does not mean that the reproductive processes
typical mung bean variety (Emarald) grown in Australia to of mung bean are not sensitive to increasing temperatures.
be 595 and 1200 °Cd, respectively, at a base temperature of Observed variations in the duration from germination to
7.5 °C, which are within the range of our values, which 50 % flowering (Table 3) showed that flower initiation
were estimated at a base temperature of 8 °C (Table 3). was sensitive to even the narrow temperature range
The small seasonal differences in the thermal durations between locations in the first season. The clear reductions
observed in the present study might have been caused by in PN per plant with increasing temperatures across loca-
photothermal variations as even slight variations in photo- tions (Table 8) indicate the possible adverse impacts of
period can alter the initiation of flowers in mung bean heat stress on the number of flowers formed, their fertil-
(Summerfield et al. 1997). Similar photothermal sensitivity ization and retention of the fertilized flowers and young
of flowering has been shown in other grain legumes such as pods. These results of the present study affirm the high
soya bean (Hatfield et al. 2011) and for annual crops in sensitivity of crop reproductive processes of legumes to
general (Craufurd and Wheeler 2009). higher temperatures (Gross and Kigel 1994, Prasad et al.
The increasing temperatures across the locations clearly 2000, 2001, 2002, 2003, Lobell and Field 2007, Hatfield
reduced vegetative growth as shown by the decreasing sea- et al. 2011). The highly significant positive correlations
sonal means of LAI50fl, CGR50fl, CGRpfl (Table 5) and that were observed between seed yield and PN in both sea-
TBMh (Table 6). These reductions reflect the negative sons showed that PN is a key determinant of yield in
impacts of increasing temperatures on the processes of leaf mung bean. Thus, increasing temperatures exert a key
expansion, radiation interception, photosynthesis and bio- negative impact on mung been yields by decreasing PN
mass accumulation (Squire 1990, Porter and Semenov (Table 8). Furthermore, the significant positive correla-
2005, Hatfield et al. 2011, Lobell and Gourdji 2012) indi- tions of PN with all growth parameters (i.e. LAI50fl,
cating that the temperature range experienced across loca- TBM50fl and TBMh) (Table 9) indicate that the magnitude
tions in the present study exceeded the respective optimum of vegetative growth partially influences the number of
temperatures for the above processes. While the high tem- pods retained probably by controlling the assimilate
perature-induced reductions in LAI50fl and TBMh were supply to developing pods. In addition to PN, SN per pod
partly caused by the shortened durations of the pre- and (SN) also exerted a significant influence on yield determi-
post-flowering stages (Table 3), the reductions in CGR50fl nation when the crops were subjected to a higher temper-
and CGRpfl showed that the rates of key physiological pro- ature range in the yala season (Table 9). This was
cesses responsible for growth were also reduced at the probably because of adverse impacts of higher tempera-
supra-optimal temperatures experienced by crops of the tures on fertilization and seed formation (Prasad et al.
present work. However, it is notable that within a given 2001, 2003, Hatfield et al. 2011). The significant negative
location, all the above growth parameters were higher in correlations that were observed between different yield
the warmer yala season as compared to the cooler maha components in the two seasons (Table 9) indicated that
season (Tables 4 and 5). This difference between the two the yield formation process of mung bean possesses a cer-
seasons was most probably caused by the greater irradiance tain degree of flexibility of adapting to higher tempera-
levels in the yala season (Table 2), especially during the tures by increasing SN and ISW when PN is decreased.

64 © 2015 Blackwell Verlag GmbH, 202 (2016) 51–68

 Temperature Response of Mung Bean

Performance of adaptation systems The estimated yield reductions per unit of increased tem-
The ‘adaptation systems’ that were tested in the present perature (Figs 4b and 5c,d) indicated the sensitivity of dif-
study were aimed at reducing the inputs of water, synthetic ferent crop management systems to increasing
pesticides and inorganic nitrogen fertilizer. As such the spe- temperature. In the warmer yala season, which experienced
cific adaptations tested here are components of ecosystem- a wider temperature range, yields of T2 and T3 showed
based approaches and conservation agriculture (Siddique lower temperature sensitivity (Fig. 5d) than the control
et al. 2012), which are part of innovative cultivation tech- (Fig. 4b) while the sensitivity of T4 was on par with that of
nologies especially aimed at resource-limited smallholder the control. In the cooler maha season, which experienced
farmers in the tropics. A summary of seed yields of the a narrower temperature range, T2 showed a substantial
adaptation systems in comparison with the respective con- positive yield response to increasing temperatures (Fig. 5c).
trols (i.e. the standard crop management practices) The beneficial effects of mulching in making a cropping
(Table 6) show that with the exception of the maha season system more resilient to increasing temperature are in
at MI, performances of the three adaptation systems were agreement with the simulated beneficial impacts of mulch-
either superior (e.g. T3 and T4 at KD in maha) to or on par ing on crops in the semi-arid tropics (Cooper et al. 2009).
with the control (e.g. T2, T3 and T4 at all locations in yala). In contrast, yields of T3 and T4 showed substantially greater
When the reduced cost of inputs and the long-term bene- temperature sensitivity than the control. This was because
fits to the soil (through mulching in T2 and organic matter of the significantly lower yields of T3 and T4 at the warmer
addition in T4) and the environment (through reduced site (i.e. MI) in maha (Table 6). Despite this, on the whole,
application of synthetic pesticides and inorganic nitrogen results of the present work indicate the likelihood that these
fertilizers) are taken in to consideration, the modified agro- adaptation systems have a greater resilience to the expected
nomic and crop protection practices which formed the temperature increases in a future climate.
adaptation systems can be recommended to upland crop-
ping systems in the tropics. Characters of mung bean varieties suitable for a warmer and
The success of IPM in the present work in either obtain- drier future climate
ing higher yields over chemical-based pest and disease con- Temperature responses of the measured growth parameters
trol or maintaining the same yield levels is a notable and yield components and their interrelationships provide
achievement, especially in view of the pesticide-related indications to plant breeders about characters to be incor-
environmental and health hazards that have been reported porated in future mung bean varieties to achieve higher
in Sri Lanka (Van der Hoek et al. 1998, Aponso et al. 2003, yields in a warmer and drier climate. Varieties with higher
Marasinghe et al. 2011). Similar success stories of IPM in crop growth rates, both during pre- and post-flowering
legumes have been reported for chickpea in India (Singh stages need to be developed to compensate for reduced
et al. 2003) and northern Bangladesh (Harris et al. 2008) crop durations. For this, varieties which are able to achieve
and soya bean in Brazil (Bueno et al. 2011). Adaptation the optimum LAI (leaf area index) of ca. 4.5 (Fig. 6) within
systems T3 and T4 at MI in maha represented the only situ- the shortest possible duration and maintain it for a maxi-
ation where the performance of an adaptation system was mum duration need to be bred. The strong positive corre-
inferior to that of the control (Table 6). This was because lation between seed yield and PN means that future
of the higher incidence of the mung bean pod borer (Maru- varieties should be capable of maximum pod initiation and
ca vitrata), which could not be controlled by the IPM pack- retention in warmer temperatures.
age practised in these two adaptation systems. This is in
agreement with the general consensus about IPM that its
Acknowledgements
success is not universal and could be location and season
specific (Landa et al. 2004, Van den Berg 2004). This work was funded by the Higher Education for Twenty
The effectiveness of mulching to conserve soil moisture First Century (HETC) Quality and Innovation Grant Win-
and reduce the irrigation requirement was demonstrated dow 3 (QIG-3) project. The Integrated Pest Management
by the observation in the present study that seed yield of (IPM) package was formulated by Dr. Devika M. De Costa
the T2 treatment (i.e. the adaptation system in which with assistance from Nadeeka Dharmadasa.
mulching and reduced water input was the only agronomic
modification from the control treatment) was always on
References
par with that of the respective control (T1) (Table 6). This
finding is especially relevant in the subhumid zone of Sri Abeyrathna, E. F. L., 1956: Dry land farming in Ceylon. Trop.
Lanka where soil water deficits constitute a major con- Agric., CXII, 191–229.
straint to increasing the productivity of upland cropping Allen, D. J., and J. M. Lenn
e, 1998: The Pathology of Food and
systems (Abeyrathna 1956). Pasture Legumes. CABI, Wallingford.

© 2015 Blackwell Verlag GmbH, 202 (2016) 51–68 65

Malaviarachchi et al.

Aponso, G. L. M., C. Magamage, W. M. Ekanayaka, and G. M. 2007: Impacts, Adaptation and Vulnerability. Contribution of
Manuweera, 2003: Analysis of water for pesticides in two Working Group II to the Fourth Assessment Report of the
major agricultural areas of the dry zone. Ann. Sri Lanka Dept. Intergovernmental Panel on Climate Change, pp. 273–313.
Agric. 5, 7–22. Cambridge University Press, Cambridge, UK.
Bogahawatte, C., and K. Kailasapathy, 1986: The consumption Ellis, R. H., R. J. Lawn, R. J. Summerfield, A. Qi, E. H. Roberts,
patterns of food in Sri Lanka with special reference to protein- P. M. Chay, J. B. Brouwer, J. L. Rose, S. J. Yeates, and S. San-
calorie nutrition. Agric. Syst. 22, 289–303. dover, 1994: Towards the reliable prediction of time to flower-
Bueno, A. de F., M. J. Batistela, R. O. de F. Bueno, J. de B. ing in six annual crops. IV. Cultivated and wild mung bean.
Francßa-Neto, M. A. N. Nishikawa, and A. L. Filho, 2011: Expl. Agric. 30, 31–43.
Effects of integrated pest management, biological control and Gourdji, S. M., A. M. Sibley, and D. B. Lobell, 2013: Global crop
prophylactic use of insecticides on the management and sus- exposure to critical high temperatures in the reproductive per-
tainability of soybean. Crop Prot. 30, 937–945. iod: historical trends and future projections. Environ. Res.
Chauhan, Y. S., C. Douglas, R. C. N. Rachaputi, P. Agius, W. Lett. 8, 1–10. 024041. doi:10.1088/1748-9326/8/2/024041.
Martin, K. King, and A. Skerman, 2010. Physiology of mung Gross, Y., and J. Kigel, 1994: Differential sensitivity to high tem-
bean and development of the mung bean crop model. In: Pro- perature of stages of the reproductive development of com-
ceedings of the 1st Australian Summer Grains Conference, mon bean (Phaseolus vulgaris L.). Field Crops Res. 36, 201–
Gold Coast, Australia, 21st–24th June 2010. http:// 212.
www.grdc.com.au/uploads/documents/2010AS- Harris, D., C. Johansen, and A. M. Musa, 2008: Farmer-friendly
GCEditedPapersPDF/Chauhan_MungbeanCropMod- technologies to improve chickpea production in the High Ba-
el_edited_paper.pdf [last accessed 17 Aug 2014]. rind tract. In: C. Riches, D. Harris, D. E. Johnson, and B.
Christensen, J. H., B. Hewitson, A. Busuioc, A. Chen, X. Gao, I. Hardy, eds. Improving Agricultural Productivity in Rice Based
Held, R. Jones, R. K. Kolli, W.-T. Kwon, R. Laprise, V. Systems of the High Barind Tract of Bangladesh, pp. 119–133.
Maga~ na Rueda, L. Mearns, C. G. Men
endez, J. R€ais€anen, A. International Rice Research Institute (IRRI), Los Ba~ nos, Phil-
Rinke, A. Sarr, and P. Whetton, 2007: Regional climate pro- ippines.
jections. In: S. Solomon, D. Qin, M. Manning, Z. Chen, M. Hatfield, J. L., K. J. Boote, B. A. Kimball, L. H. Ziska, R. C. Iza-
Marquis, K. B. Averyt, M. Tignor, and H. L. Miller, eds. Cli- urralde, D. Ort, A. M. Thomson, and D. Wolfe, 2011: Climate
mate Change 2007: The Physical Science Basis. Contribution impacts on agriculture: implications for crop production.
of Working Group I to the Fourth Assessment Report of the Agron. J. 103, 351–370.
Intergovernmental Panel on Climate Change, Cambridge Uni- Hewavitharana, H. V. C., H. U. Warnakulasuriya, and G. B. S.
versity Press, New York, NY. Wajira Kumara, 2010: Constraints to expansion of cowpea
Cooper, P., K. P. C. Rao, P. Singh, P. S. Traore, K. Rao, P. Dixit, and mung bean under rainfed farming in Anuradhapura dis-
and S. J. Twomlow, 2009: Farming with current and future trict. Ann. Sri Lanka Dept. Agric. 12, 91–104.
climate risk: advancing a ‘Hypothesis of Hope’ for rainfed Imrie, B. C., and R. J. Lawn, 1990: Time to flowering of mung
agriculture in the semi-arid tropics. J. SAT Agric. Res. 7, 19. bean (Vigna radiata) genotypes and their hybrids in response
http://ejournal.icrisat.org/aerpaperv7.htm. to photoperiod and temperature. Exp. Agric. 26, 307–318.
Craufurd, P. Q., and T. R. Wheeler, 2009: Climate change and IPCC, 2013: Summary for policymakers. In: T. F. Stocker, D.
the flowering time of annual crops. J. Exp. Bot. 60, 2529– Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J. Boschung, A.
2539. Nauels, Y. Xia, V. Bex, and P. M. Midgley, eds. Climate
Daugherty, M. P., D. Bosco, and R. P. P. Almeida, 2009: Tem- Change 2013: The Physical Science Basis. Contribution of
perature mediates vector transmission efficiency: inoculum Working Group I to the Fifth Assessment Report of the
supply and plant infection dynamics. Ann. Appl. Biol. 155, Intergovernmental Panel on Climate Change, Cambridge
361–369. University Press, Cambridge, UK and New York, NY, USA.
Davidson, E. A., and I. A. Janssens, 2006: Temperature sensitiv- Karuppaiah, V., and G. K. Sujayanad, 2012: Impact of climate
ity of soil carbon decomposition and feedbacks to climate change on population dynamics of insect pests. World J. Ag-
change. Nature 440, 165–173. ric. Sci. 8, 240–246.
De Costa, W. A. J. M., K. N. Shanmugathasan, and K. D. S. M. Laing, D. R., P. G. Jones, and J. H. Davis, 1984: Common bean
Joseph, 1999: Physiology of yield determination of mung bean (Phaseolus vulgaris L.). In: P. R. Goldsworthy, and N. M.
(Vigna radiata (L.) Wilczek) under different irrigation Fisher, eds. The Physiology of Tropical Field Crops, pp. 305–
regimes in the dry and intermediate zones of Sri Lanka. Field 351. John Wiley & Sons, New York, NY.
Crops Res. 61, 1–12. Landa, B. B., J. A. Navas-Cort
es, and R. M. Jim
enez-D
ıaz, 2004:
Easterling, W. E., P. K. Aggarwal, P. Batima, K. M. Brander, L. Integrated management of Fusarium wilt of chickpea with
Erda, S. M. Howden, A. Kirilenko, J. Morton, J.-F. Soussana, sowing date, host resistance, and biological control. Phytopa-
J. Schmidhuber, and F. N. Tubiello, 2007: Food, fibre and for- thology 94, 946–960.
est products. In: M. L. Parry, O. F. Canziani, J. P. Palutikof, Lawn, R. J., and C. S. Ahn, 1985: Mung bean (Vigna radiata (L.)
dervan Linden P. J., and C. E. Hanson, eds. Climate Change Wilczek/Vigna mungo (L.) Hepper). In: R. J. Summerfield,

66 © 2015 Blackwell Verlag GmbH, 202 (2016) 51–68

 Temperature Response of Mung Bean

and E. H. Roberts, eds. Grain Legume Crops, pp. 584–623. ambient and elevated carbon dioxide. Glob. Change Biol. 9,
William Collins, London. 1775–1787.
Lobell, D. B., and G. P. Asner, 2003: Climate and management Punyawardane, B. V. R., 2008: Evolution of climatic zones in Sri
contributions to recent trends in U.S. agricultural yields. Sci- Lanka. In: Agro-Climatological Zones and Rainfall Pattern in
ence 299, 1032. Sri Lanka (In Singhalese), pp. 44–113. Department of Agricul-
Lobell, D. B., and M. B. Burke, 2008: Why are agricultural ture, Peradeniya, Sri Lanka.
impacts of climate change so uncertain? The importance of Rachie, K. O., and L. M. Roberts, 1974: Grain legumes of the
temperature relative to precipitation. Environ. Res. Lett. 3, lowland tropics. Adv. Agron. 26, 1–132.
1–8. 034007. doi:10.1088/1748-9326/3/3/034007. Rawson, H. M., 1992: Plant responses to temperature under
Lobell, D. B., and C. B. Field, 2007: Global scale climate–crop conditions of elevated CO2. Aust. J. Bot. 40, 473–490.
yield relationships and the impacts of recent warming. Envi- Schlenker, W., and D. B. Lobell, 2010: Robust negative impacts
ron. Res. Lett. 2, 1–7. 014002. doi:10.1088/1748-9326/2/1/ of climate change on African agriculture. Environ. Res. Lett. 5,
014002. 1–8. 014010. doi:10.1088/1748-9326/5/1/014010.
Lobell, D. B., and S. M. Gourdji, 2012: The influence of climate Schlenker, W., and M. J. Roberts, 2009: Nonlinear temperature
change on global crop productivity. Plant Physiol. 160, 1686– effects indicate severe damages to US crop yields under cli-
1697. mate change. Proc. Natl Acad. Sci. USA 106, 15594–15598.
Lobell, D. B., M. B. Burke, C. Tebaldi, M. D. Mastrandrea, W. P. Siddique, K. H. M., C. Johansen, J. V. D. K. Kumar Rao, and M.
Falcon, and R. L. Naylor, 2008: Prioritizing climate change Ali, 2008: Legumes in sustainable cropping systems. In: M. C.
adaptation needs for food security in 2030. Science 319, 607– Kharkwal, ed. Food Legumes for Nutritional Security and Sus-
610. tainable Agriculture. Proceedings of the Fourth International
Marasinghe, J. P., C. Magamage, M. G. D. Shiromi, and A. G. P. Food Legumes Research Conference (IFLRC-IV), Vol. 1, pp.
Aravinna, 2011: Organophosphate pesticides residues in food 787–819, October 18–22, 2005, New Delhi, India. Indian Soci-
commodities in Sri Lanka. A review. Ann. Sri Lanka Dept. Ag- ety of Genetics and Plant Breeding, New Delhi.
ric. 13, 81–94. Siddique, K. H. M., C. Johansen, N. C. Turner, M.-H. Jeuffroy,
Panabokke, C. R., 1996: Soils and Agro-Ecological Environments A. Hashem, D. Sakar, Y. Gan, and S. S. Alghamdi, 2012: Inno-
of Sri Lanka. Natural Resources, Energy and Science Authority vations in agronomy for food legumes. A review. Agron. Sus-
of Sri Lanka, Colombo, Sri Lanka. 220 pp. tain. Dev. 32, 45–64.
Panabokke, C. R., and B.V.R. Punyawardane, 2009: Climate Singh, A., S. Singh, and S. N. Rao, 2003: Integrated pest manage-
Change and Rain-fed Agriculture in the Dry Zone of Sri ment in India. In: K. M. Maredia, D. Dakouo, and D. Mota-
Lanka. In: A. Evans and K. Jinapala, eds. Proceedings of the Sanchez, eds. Integrated Pest Management in the Global
National Conference on Water, Food Security and Climate Arena, pp. 209–221. CABI, London.
Change in Sri Lanka, Vol. 2, Water quality, environment and Squire, G. R., 1990: The Physiology of Tropical Field Crop Pro-
climate change, pp. 141–46, 9–11 June 2009, BMICH, duction. C.A.B. International, Wallingford, UK. 236 pp.
Colombo, Sri Lanka. International Water Management Insti- Summerfield, R. J., and R. J. Lawn, 1987: Environmental modu-
tute, Colombo, Sri Lanka. lation of flowering in Mung Bean (Vigna radiata): a reap-
Porter, J. R., and M. A. Semenov, 2005: Crop responses to cli- praisal. Exp. Agric. 23, 461–470.
matic variation. Philos. Trans. R. Soc. Lond. B Biol. Sci. 360, Summerfield, R. J., R. H. Ellis, P. Q. Craufurd, Q. Aiming, E. H.
2021–2035. Roberts, and T. R. Wheeler, 1997: Environmental and genetic
Prasad, P. V. V., P. Q. Craufurd, R. J. Summerfield, and T. R. regulation of flowering of tropical annual crops. Euphytica 96,
Wheeler, 2000: Effects of short episodes of heat stress on 83–91.
flower production and fruit-set of groundnut (Arachis hypo- Teixeira, E. I., G. Fischer, H. van Velthuizen, C. Walter, and F.
gaea L.). J. Exp. Bot. 51, 777–784. Ewert, 2013: Global hot-spots of heat stress on agricultural
Prasad, P. V. V., P. Q. Craufurd, V. G. Kakani, T. R. Wheeler, crops due to climate change. Agric. For. Meteorol. 170, 206–
and K. J. Boote, 2001: Influence of high temperature during 215.
pre- and post-anthesis stages of floral development on fruit- Tharanathan, R. N., and S. Mahadevamma, 2003: Grain legumes
set and pollen germination in peanut. Aust. J. Plant Physiol. – a boon to human nutrition. Trends Food Sci. Technol. 14,
28, 233–240. 507–518.
Prasad, P. V. V., K. J. Boote, L. H. Allen, and J. M. G. Thomas, Thirumaran, A. S., and M. A. Seralathan, 1988: Utilization of
2002: Effects of elevated temperature and carbon dioxide on mung bean. In: S. Shanmugasundaram, ed. Proceedings of the
seed-set and yield of kidney bean (Phaseolus vulgaris L.). Glob. Second International Conference on Mung Bean, pp. 470–
Change Biol. 8, 710–721. 485, AVRDC, Shanhua, Taiwan.
Prasad, P. V. V., K. J. Boote, L. H. Allen, and J. M. G. Thomas, Tubiello, F. N., J.-F. Soussana, and S. M. Howden, 2007: Crop
2003: Supra-optimal temperatures are detrimental to peanut and pasture response to climate change. Proc. Natl Acad. Sci.
(Arachis hypogaea L.) reproductive processes and yield at USA 104, 19686–19690.

© 2015 Blackwell Verlag GmbH, 202 (2016) 51–68 67

Malaviarachchi et al.

Turner, N. C., 2004: Agronomic options for improving rainfall- Global IPM Facility. Wageningen University, Wageningen, the
use efficiency of crops in dryland farming systems. J. Exp. Bot. Netherlands.
55, 2413–2425. Van der Hoek, W., F. Konradsen, K. Athukorala, and T. Wani-
Turner, N. C., 2008: Adaptation to climate change: research gadewa, 1998: Pesticide poisoning: a major health problem in
needs for smallholder farmers growing sorghum and pigeon Sri Lanka. Soc. Sci. Med. 46, 495–504.
pea in Eastern and Southern Africa. Report commissioned by Weinberger, K., 2003: Impact analysis on mung bean research in
ICRISAT, East Africa, 29 pp. South- and South-east Asia. Final Report GTZ Ei-
Van den Berg, H., 2004: IPM Farmer Field Schools: A genmassnahme No. 99.9117.5.AVRDC, Shanhua, Taiwan.
Synthesis of 25 Impact Evaluations. Prepared for the

68 © 2015 Blackwell Verlag GmbH, 202 (2016) 51–68