1

Potato Crop Response to Genotype

and Environment in a Subtropical Highland
Agro-ecology

L. Molahlehi & J. M. Steyn* & A. J. Haverkort

Abstract Potato response to environment, planting date and genotype was studied
for different agro-ecological zones in Lesotho. Field experiments were conducted at
four different sites with altitudes ranging from 1,655 to 2,250 m above sea level
during the 2010/2011 and 2011/2012 summer growing seasons. Treatments consisted
of three cultivars that varied in maturity type, two planting dates and four sites
differing in altitude and weather patterns. Various plant parts were measured peri-
odically. To understand and quantify the influence of abiotic factors that determine
and limit yields, the LINTUL crop growth model was employed which simulated
potential yields for the different agro-ecological zones using weather data collected
per site during the study period. Observed actual crop yields were compared with
model simulations to determine the yield gap. Model simulations helped to improve
our understanding of yield limitations to further expand potato production in sub-
tropical highlands, with emphasis on increasing production through increased yields
rather than increased area. Substantial variation in yield between planting date,
cultivar and site were observed. Average tuber dry matter (DM) yields for the highest
yielding season were above 7.5 t DM ha−1 or over 37.5 t ha−1 fresh tuber yield. The
lowest yield obtained was 2.39 t DM ha−1 or 12 t ha−1 fresh tuber yield for cultivar
Vanderplank in the 2011/2012 growing season at the site with the lowest altitude.
Modelled potential tuber yields were 9–14 t DM ha−1 or 45–70 t ha−1 fresh yield.
Drought stress frequently resulted in lower radiation use efficiencies and to a lesser
degree harvest indices, which reduced tuber yield. The site with the lowest altitude
and highest temperatures had the lowest yields, while the site with the highest altitude
had the highest yields. Later maturing cultivars yielded more than earlier maturing
ones at all sites. It is concluded that the risk of low yields in rain-fed subtropical

L. Molahlehi : J. M. Steyn * : A. J. Haverkort

Department of Plant Production and Soil Science, University of Pretoria, Private Bag X20, Hatfield,
Pretoria 0028, South Africa
e-mail: [email protected]

A. J. Haverkort
Plant Research International, Wageningen University and Research Centre, P.O. Box 616, 6700
AP Wageningen, The Netherlands
 2

highlands can be minimised by planting late cultivars at the highest areas possible as
early as the risks of late frosts permit.

Keywords Harvest index . Potential yield . Radiation use efficiency . Subtropical

highlands . Tuber dry matter . Yield gap analysis

Introduction

The Kingdom of Lesotho is a small (30,000 km2), landlocked country which is an

enclave surrounded by South Africa (Fig. 1). Potatoes are widely grown in Lesotho
and it is an important food crop. However, the area cropped with potato is still far
below that of other staple crops such as maize, wheat, sorghum and pulses (Bureau of
Statistics 2008). Potato production takes place mainly under subsistence farming

Fig. 1 Map of Lesotho showing the four agro-ecological zones and the four experimental sites (Depart-
ment of Geography, University of Pretoria, Pretoria, South Africa)
 3

practices with low levels of input use and under rain-fed conditions. Farming
practices such as planting, ridging and harvesting are done manually and with the
use of draught animals in some incidences. The use of mechanisation is very minimal
as there is no large-scale commercial production. The crop is grown in the lowlands
and highlands of Lesotho (above 2,000 m a.s.l.) in summer, as winters are too cold
with frequent snow and night frost due to the high altitude. Weather patterns are
extremely erratic with periods of excessive rainfall and prolonged droughts, which
are both deleterious to crop growth. The former lead to periods with low solar
radiation and leaching of nutrients, especially nitrogen, and the latter lead to lack of
water available to the plants (Vos and MacKerron 2000). As a result, Lesotho is not
self-sufficient in potato and a substantial proportion of the potatoes consumed in the
country is imported.
Despite these limitations, the climatic conditions that prevail in the country are
considered to be suitable and have the potential for high yields when the crop is
managed adequately (Bureau of Statistics 2008). Lisinska and Leszczynski (1989)
stated that the effect of locality on potato tuber yield and quality is generally
connected to climatic conditions prevailing in a given area. Lesotho can be divided
into four agro-ecological zones: the lowlands, foothills, Maluti mountains and Senqu
River Valley (Fig. 1).
The climatic conditions of Lesotho, of warm summers and cold winters, allow for
only one production season throughout the whole country. Potato is grown as a
summer crop and planting can start anytime from mid-September to December, al-
though there is a risk of late frost incidence up to the beginning of October. However,
planting normally only starts in October since summer rains usually only occur from
then. The country has a typical monsoon or Sahel-type climate, with dry winters and
rainy summers. Although potatoes are produced in all parts of the country, the
most suitable growing conditions—with relatively low temperatures—are those at
the higher altitude mountain zones rather than the foothills and lowlands. Seed
potato production, which is considered of economic importance in the country, is
confined to isolated areas in the cooler high mountain valleys. Advantages of the
subtropical highland summer crop compared to subtropical lowland winter crop is
that the low night temperatures assure adaptation of the crop while the long days assure
high levels of solar radiation and as such potentially high dry matter
production (Haverkort 1990). The relatively cool conditions also mean that there is
relatively low incidence of diseases and pests.
Seed production is mainly carried out by farmer groups who normally buy
their seed stock together. The seed potatoes are obtained from South Africa and
farmers usually buy generation two or three, and in some cases generation four
seed, and then multiply the seed for one season. The seed produced is then sold
to farmers who produce ware potato. The cultivars commonly used are BP1 and
Up-to-date, with BP1 having higher preference. Ware potato is commonly pro-
duced in the foothill and lowland areas, with generally low typical yields (Zones of
Production 2008). According to FAOSTAT DATA (2013), the area planted to potato
increased from less than 500 ha in 1961 to 6,850 ha in 2011, while for the same period
the average yield has not increased significantly and is currently at about 14 t ha−1 (fresh
tuber yield). With a total annual production of around 85,000 t and a population of 1.7
million, the average consumption is 50 kg per person per annum of potato grown in
 4

Lesotho itself. In addition, a considerable amount is imported from South Africa, but this
amount is not known.
Given the availability of suitable climatic conditions that exist in the country,
consumer preferences and the desire to be more self-reliant regarding food, there is a
need to further expand and improve potato production in Lesotho with more empha-
sis on increasing production through increased yields rather than through increased
area. To understand and quantify the influence of abiotic factors that determine and
limit yields (Van Oort et al. 2012), there was a need to elaborate the potato crop
ecology and physiology through field trials and a modelling approach for the various
agro-ecological conditions prevailing in the country. The methodology followed
would enable us to establish actual and potential yield (Caldiz et al. 2001) and
quality levels and allow us to analyse the yield and quality gaps. The knowledge of
these yield levels and how they are established as interaction between genotype and
environment will help to establish the role of cultivars, environments and
management practices that can be employed to enhance production.
Based on this background, the current study was conducted for different agro-
ecological zones, planting dates and different cultivars. The aim was to establish
the actual yields that can be achieved using best practices under the different
environments and weather conditions. Secondly, the study could help identify
which production and management practices need to be optimised to make the
crop more robust in erratic environmental conditions. To this end, we planted
three cultivars varying in maturity type in two different years, which experienced
different weather conditions, at four sites differing in altitude and weather
patterns, and on two planting dates to further increase variation in the conditions
crops were subjected to. The seed for the two planting dates was acquired at the
same time and this further enhanced variation in genotype response, since the
second planting was a month later and therefore the seed was physiologically
older. Observed crop parameters were compared with those of the LINTUL-
Potato model (Kooman and Haverkort 1994) which calculates the potential yield.
This study is significant as it was the first crop ecological analysis for this agro-
ecology. It will help to design more robust potato cropping systems for Lesotho
in terms of planting time, cultivar selection and management practices for the
different agro-ecological zones, in order to enhance potato yields. The results of
the study will also be significant for other potato cropping systems with sub-
tropical highlands similar to that of Lesotho, such as the Southern Sahel in North
Africa.

Materials and Methods

Environments of the Field Experiments

Field experiments were conducted during the 2010/2011 and 2011/2012 summer
growing seasons at four different sites, representing the four agro-ecological zones
in Lesotho. All experiments were planted under dry-land conditions, as almost no
irrigation is practiced in the country. The experiments were planted on research
stations and farms of the National Agricultural Research Institute (Department of
 5

Agricultural Research) and the National University of Lesotho (NUL). Given that
the terrain of the country does not allow off-road exploration, the sites—besides being
representative of the various zones—were selected based on accessibility to
allow frequent monitoring and data collection. The four sites (Fig. 1) are
geographically separated from each other and have different micro-climates in terms
of temperature, rainfall and solar radiation intensity. Rainfall and temperatures tend
to decrease with increasing altitude. Table 1 shows the names of the sites, their
altitude, coordinates, soil type and pH, and which zone they represent.
The long-term monthly weather data (daily maximum and minimum temperatures
and total monthly precipitation) of each site are shown in Fig. 2. The long-term data
lack evapotranspiration data, which was not available. The lowland and foothill area,
represented by Roma and Nyakosoba, respectively, clearly have the highest temper-
atures, while temperatures in the highland areas of Thaba Tseka and Mokhotlong are
slightly lower (Table 1; Fig. 2). Temperatures can drop below zero during winter
months, especially in the highlands (Fig. 2). The total monthly precipitation is highest
in the lowland and foothill areas of Roma and Nyakosoba, compared with the
highlands (Table 1).
Daily weather data between planting and final harvest of the trials for each of the
two seasons were collected with an automatic weather station (AWS) located within
about 100 m from each of the trial sites, except for Roma where data were supplied by
the Geography Department at NUL. The location details of each weather station are
given in Table 1. The AWS instruments consisted of an LI200X silicon pyranometer
(LI-COR, Lincoln, Nebraska, USA) to measure solar irradiation, an electronic cup
anemometer (R.M. Young, Inc., USA) to measure wind speed, an electronic tipping
bucket rain gauge (TE525 Texas Instruments, USA), a Vaisala HMP50 electronic
temperature and humidity sensor and a CR10X datalogger (Campbell Scientific, Inc.,
Logan, Utah, USA). Weather data were offloaded once every 3 weeks using a laptop
computer.

Table 1 Trial sites with their respective agro-ecological zoning, altitudes, soil characteristics and long-
term average annual temperature and rainfall

Site name Agro-ecological Coordinates Soil Soil Long-term Long-term

zone and altitude type pH average annual average annual
(m a.s.l.) Latitude Longitude temperature (°C) rainfall (mm)

Roma Lowland; 1,655 29.443° S 27.723° E Sandy 5.0 Max. 22.40 779
Loam Min. 8.80
Nyakosoba Foothill; 2,034 29.522° S 27.778° E Clay 4.7 Max. 22.08 778
Loam Min. 7.56
Thaba Highland; 2,250 29.526° S 28.612° E Sandy 6.6 Max 19.01 631
Tseka Clay Min 5.48
Loam
Mokhotlong Highland; 2,170 29.289° S 29.079° E Clay 6.9 Max 20.02 596
Loam Min 4.78
 6

Fig. 2 Long-term monthly rainfall, maximum temperature (Tmax) and minimum temperature (Tmin)
patterns for the four trials sites (Roma, 26-year temperature and rainfall averages, 1985–2011; Nyakosoba,
44-year temperature averages, 1967–2011; 61-year rainfall averages, 1950–2011; Thaba Tseka, 35-year
temperature and rainfall averages, 1976–2011; Mokhotlong, 44-year temperature averages, 1967–2011; 81-
year rainfall averages, 1930–2011) (Lesotho Meteorological Services 2012)

Experimental Details

Field experiments were laid down in a randomised complete block design with
four replicates during the 2010/2011 and the 2011/2012 growing seasons. The
treatments consisted of three cultivars (BP1, Mondial and Vanderplank) and two
planting dates: the first in mid-November and the second planting 1 month later
in mid-December. Cultivars were selected based on their maturity types as follows:
Vanderplank (short growth cycle), Mondial (medium to long cycle) and BP1 (long
cycle cultivar) (Visser 2012). BP1, a South African cultivar, is the most
commonly planted cultivar in Lesotho because of its good adaptability and
relatively high yields. The soils were fertilised with mineral fertilisers according
to soil analysis results and fertilisation guidelines (Steyn and Du Plessis
2012). Plants were sprayed three-weekly with fungicides to control late blight
caused by Phytophthora infestans.
Plots were 30 m2 in size (6×5 m) and plant spacing was 100 cm between the rows
and 25 cm within the row. Each plot consisted of five rows and 24 plants per row (120
 7

plants per plot). The seed tubers, which originated from seed producers in South
Africa, ranged in mass between 60 and 90 g each and were well sprouted.
Periodic harvests for growth analyses were carried out three-weekly throughout
the growing season and the final harvest was conducted at the end of March for both
planting dates and seasons. At each periodic harvest, ten plants per plot (two plants
per row) were sampled, leaving a single guard plant between samples. No guard rows
were left between adjacent plots due to the limited number of plants available for
sampling. Although this could have affected the results, the effect was assumed to be
consistent across all cultivars. Fewer plants were sampled when plants were missing
due to non-emergence. The following crop characteristics were observed at each
periodic harvest: number of main stems; stem length and number; fresh and dry mass
of leaves, stems and tubers; and tuber number. Subsamples of about 300 g each were
taken from the fresh foliage, stems and tubers. These were then weighed, dried at
70 °C for 48 h and re-weighed to determine their dry matter contents. After decline of
the foliage had commenced, no further above-ground samples were taken. At crop
maturity, the tubers of 20 plants per plot (four per row) were harvested to determine
final tuber yield.
The percentage ground cover with green leaves was determined every 3 weeks at
three spots in each plot. A transparent plastic grid of 19×27.5 cm (GC calculator),
which is divided into 100 rectangles with black lines, was used to estimate the
percentage cover of green leaves, as described by Kooman et al. (1996). The grid
was handheld and pointed downward at such distance from the eye that a represen-
tative area of about 1 m2 was observed and the number of rectangles with over 50%
of leaves were counted.

Crop Model

The LINTUL crop growth model used in the present study is similar to that used by
Franke et al. (2011). The model calculates potential potato dry matter production
from the amount of solar radiation intercepted by the green foliage and a conversion
factor (radiation use efficiency, RUE) (Spitters 1990), and follows the approach of
Kooman and Haverkort (1994) to calculate temperature-dependent phenological
development of the potato crop. Higher temperatures lead to earlier crop emergence
and a more rapid initial leaf growth, resulting in increased interception of solar
radiation at early stages of crop growth, a rapid maturation of the crop and a reduced
length of the growing cycle from planting to harvest. Unfavourably high temperatures
reduce photosynthesis and thereby biomass accumulation (Levy and Veilleux
2007).
We simulated shoot growth, foliar expansion, biomass accumulation and tuber
growth on a day-to-day basis. Climate input data required by the model include daily
minimum and maximum temperatures, incoming solar radiation and rainfall and
reference evapotranspiration. Management input data include the depth and date of
planting. Accumulated degree days from planting (with a base temperature of 2 °C)
determines the time to crop emergence, leaf area development and the time of crop
termination. The leaf area index (LAI) increases exponentially from crop emergence
until a leaf area index of 0.75 is achieved; whereafter, development rate depends on
temperature and water availability until a full crop cover is reached (LAI >3). Daily
 8

biomass growth is calculated from the LAI, solar radiation interception (using an
extinction coefficient of 1 (Spitters and Schapendonk 1990)) and RUE (1.25 g dry
matter MJ−1 of intercepted solar radiation). In the model, photosynthesis capacity is
reduced when the average daily temperature falls below 16 °C or when the maximum
temperature exceeds 30 °C and is completely halted at temperatures below 2 °C and
above 35 °C (Kooman and Haverkort 1994). The harvest index for all cropping
situations was set at 0.75 (Kooman and Haverkort 1994) and simulated yields are
presented as tuber dry and fresh matter, assuming a dry matter concentration of 20%.

Daily Penman–Monteith grass reference evapotranspiration (ETo) values were

calculated using measured daily weather data as input parameters (Smith et al.
1996). Daily potential evapotranspiration (PET) for potatoes was calculated by
multiplying ETo with a crop-specific coefficient (Kc) according to the procedure
recommended by Allen et al. (1996). Actual (water-limited) crop transpiration
rate (T) was calculated as the PET multiplied by a drought stress factor, which
is a function of the plant available water in the soil, as described by Franke
et al. (2011).
Evaporation from the soil (E) was quantified following the procedure of Ritchie
(1972), who calculated that a soil with an average water holding capacity that is
wetted every 4 days by irrigation or rain has an evaporation rate that is one third of
ETo until emergence of the crop. Thereafter, evaporation from the soil decreases
linearly with ground cover (calculated from LAI) to 10% of ET at full ground cover
(LAI >3).

Data Analysis

Data analyses were conducted using the SAS software computer programme (SAS
9.3). ANOVA was carried out to determine significant differences, and where there
were differences, mean separations were done using LSD. Correlations were also
done to show relationship between the different variables (Gomez and Gomez 1984;
Petersen 1994).

Results and Discussion

The weather data recorded from planting until harvest for the two cropping seasons
and four sites are presented in Fig. 3 and Table 2. For both planting seasons, the
lowest altitude site—Roma—usually had the highest day and night temperatures
compared to sites at the foothills (Nyakosoba) and higher elevations (Thaba Tseka
and Mokhotlong).
The solar radiation levels were higher in the highland areas than at the foothills and
lowland sites and seasonal averages varied between 19 and 25 MJ m−2 day−1. For
both seasons, there was more precipitation in the lowland and foothills than in the
highlands. Precipitation levels for both seasons ranged between 300 and 900 mm
during the growing seasons and there was more precipitation in the first season than
in the second for all the sites. ETo was higher in the highland areas, especially for
Mokhotlong, compared to the lowlands and foothills.
 9

Rainfall Radiation T-avg Rainfall Radiation T-avg

Temperature (oC) & Radiation (MJ m-2 d-1)

Temperature (oC) & Radiation (MJ m-2 d-1)

30 360 30 360

25 300 25 300

20 240 20 240

Rainfall (mm)
Rainfall (mm)
15 180 15 180

10 120 10 120

5 60 5 60

0 0 0 0
Nov Dec Jan Feb March Nov Dec Jan Feb March
Roma (1,655 masl) season 2010/11 Roma (1,655 masl) season 2011/12

Rainfall Radiation T-avg Rainfall Radiation T-avg

Temperature (oC) & Radiation (MJ m-2 d-1)

Temperature (oC) & Radiation (MJ m-2 d-1)

30 360 30 360

25 300 25 300

20 240

Rainfall (mm)
20 240 Rainfall (mm)

15 180 15 180

10 120 10 120

5 60 5 60

0 0 0 0
Nov Dec Jan Feb March Nov Dec Jan Feb March
Nyakosoba (2,034 masl) season 2010/11 Nyakosoba (2,034 masl) season 2011/12

Rainfall Radiation T-avg Rainfall Radiation T-avg

Temperature (oC) & Radiation (MJ m-2 d-1)
Temperature (oC) & Radiation (MJ m-2 d-1)

30 360 30 360

25 300 25 300

20 240 20 240

Rainfall (mm)
Rainfall (mm)

15 180 15 180

10 120 10 120

5 60 5 60

0 0 0 0
Nov Dec Jan Feb March Nov Dec Jan Feb March
Thaba Tseka (2,250 masl) season 2010/11 Thaba Tseka (2,250 masl) season 2011/12

Rainfall Radiation T-avg Rainfall Radation T-avg

Temperature (oC) & Radiation (MJ m-2 d-1)
Temperature (oC) & Radiation (MJ m-2 d-1)

30 360 30 360

25 300 25 300

20 240 20 240
Rainfall (mm)
Rainfall (mm)

15 180 15 180

10 120 10 120

5 60 5 60

0 0 0 0
Nov Dec Jan Feb March Nov Dec Jan Feb March
Mokhotlong (2,170 masl) season 2010/11 Mokhotlong (2,170 masl) season 2011/12

Fig. 3 Actual weather data (total rainfall, average radiation and temperature (T-avg)) recorded from
planting until harvest date during the 2010/2011 and 2011/2012 planting seasons for the four localities.
Altitudes of localities in metres above sea level are given in parentheses

Observed and modelled yields and other parameters recorded in the field trials are
presented in Table 3. In the first season (2010/2011; S1) total dry matter and tuber
yields were significantly higher for the first planting than the second planting at all
sites, except for Nyakosoba (Table 3). Average tuber dry matter (DM) yields across
 10

Table 2 Actual weather data recorded from planting until harvest date during the 2010/2011 and 2011/2012 planting seasons for each of the four localities

Parameter Site Site

Roma Nyakosoba Thaba Tseka Mokhotlong Roma Nyakosoba Thaba Tseka Mokhotlong
Planting 1 Planting 2

Season 1, 2010/2011
Tmax °C 25.4 22.9 23.1 24.5 25.4 23.2 23.2 24.6
Tmin °C 13.5 11.1 9.7 10.0 13.9 11.4 10.0 10.2
Rad MJ 19.1 19.1 22.0 22.8 19.0 19.0 20.2 21.7
Rain mm 857 891 598 502 637 625 482 358
ETo mm 615 589 628 793 465 447 468 549
Planting 12 November 2010 08 November 2010 10 November 2010 11 November 2010 16 December 2010 13 December 2010 14 December 2010 15 December 2010
date
Final 15 April 2011 12 April 2011 20 April 2011 18 April 2011 15 April 2011 12 April 2011 20 April 2011 18 April 2011
harvest
date
Season 2, 2011/2012
Tmax °C 27.3 24.4 24.6 25.3 27.1 24.5 24.7 25.6
Tmin °C 13.9 11.2 9.2 9.2 14.3 11.4 9.4 9.5
Rad MJ 19.4 19.4 23.7 24.8 19.4 19.4 23.1 24.3
Rain mm 425 508 408 333 386 403 335 310
ETo mm 625 589 635 726 487 466 496 585
Planting 07 November 2011 08 November 2011 09 November 2011 10 November 2011 05 December 2011 06 December 2011 07 December 2011 08 December 2011
date
Final 26 March 2012 27 March 2012 29 March 2012 28 March 2012 26 March 2012 27 March 2012 29 March 2012 28 March 2012
harvest
date
 11

sites were in the range of 6.96–8.65 t ha−1 (average 7.60 t ha−1) for the first planting
(P1), while the range was 3.60–6.07 t ha−1 (average 4.57 t ha−1) in the second
planting (P2). No significant differences in harvest index (HI) values occurred
between the first and second plantings. HI values ranged between 72.4 and 76.7%
(average of 73.8%) for P1 and 67.8–79.3% (average of 71.5%) for P2 (across all
sites). There was also little variation in the calculated average RUE across sites in P1
(range of 0.84–0.91 g MJ−1 and average of 0.88 g MJ−1) and between the two planting
dates. However, a wide range of RUE values was recorded between localities in the
second planting (range 0.53–1.22 g MJ−1 and average of 0.83 g MJ−1). Tuber dry matter
contents (DM%) were quite stable across planting dates, with no significant differences
recorded, except for Nyaksoba where the values were lower in the second planting.
Average tuber dry matter contents across sites were in the range of 19.4–20.9% (average
20.3%) for P1 and 19.2–21% (average 20.1%) for P2.
In the second season (2011/2012; S2), there were no differences in average tuber
yields between the first and second plantings. Tuber yields across localities ranged from
3.60 to 5.52 t DM ha−1 (average 4.64 t DM ha−1) for P1, and for P2, the range was similar
(3.23–5.86 t DM ha−1 and average 4.64 t DM ha−1). The same trend was observed for
harvest index (range 60.4–72.6% and average 68.0% for P1; range 57.1–75.7% and
average 68.8% for P2) and radiation use efficiency (range 0.45–0.62 g MJ−1 and
average 0.52 g MJ−1 for P1; range 0.41–0.57 g MJ−1 and average 0.51 g MJ−1 for
P2). There were no differences in tuber DM% between plantings, with a range of
19.5–22.9% (average of 21.1%) for P1, and 19.5–23.4% (average 21.3%) for P2.
Overall, the tuber yields were highest in the first season (S1 average 6.1 t
DM ha−1), and yields were especially high for P1 (average 7.6 t DM ha−1),
compared to the second season (S2, average 4.6 t DM ha−1). The same pattern was
observed for radiation use efficiencies, which were relatively high in S1 (average
0.85 g MJ−1) compared to S2 (0.51 g MJ−1). There was little variation in harvest
indices between the two seasons (average 72.6% in S1 and 68.4% in S2). The lowest
HI of 57% was recorded for Roma in S2. Similarly, average tuber DM% showed little
variation, with an average of 20.2% in S1 and 21.2% in S2. The radiation use
efficiencies achieved in this study were generally low, probably because production
was under dry-land conditions and the crops were frequently exposed to stress
periods, similar to the findings of Kooman and Rabbinge (1996).
Arab et al. (2011) found that planting date influences leaf area index and, there-
fore, also the amount of intercepted solar radiation, which is an important factor that
determines potato yield. Planting date directly affects the weather conditions that a
crop will be exposed to, for example the rainfall, temperature and amount of solar
radiation (Kawakami et al. 2005). Hassanpanah et al. (2009) found that marketable
tuber yields were affected by planting date and that early planting leads to more time
for plant growth and higher yields. The long-term weather data for Lesotho (Fig. 2)
show that for all four of our sites substantial summer rains that would allow for early
planting only start from October onwards. November to March are the months with
the highest rainfall that will best support dry-land crop production. The best period
for planting therefore seems to be between mid-October (after the first good rains)
and mid-November. Later planting may firstly pose a risk of too wet soils that will
result in rotting of tubers in poorly drained or heavier soils. For example, the low
yields recorded for Roma (P2 in S1) (Fig. 3) resulted from erratic intense rainfall in
 12

Table 3 Observed total dry matter (DM) yield, tuber dry matter yield, LINTUL model simulated potential
tuber dry matter yield, tuber dry matter content, harvest index (HI) and radiation use efficiency (RUE) per
site, planting time and season (average values across cultivars)

Roma Nyakosoba Thaba Tseka Mokhotlong

Parameter Unit Planting Planting Planting Planting

1 2 1 2 1 2 1 2

Season 1, 2010/2011
Total DM yield t ha−1 9.64 a 4.70 b 9.55 a 8.55 a 10.08 a 7.23 b 11.70 a 5.29 b
Tuber DM yield t ha−1 7.43 a 3.70 b 6.96 a 6.07 a 7.36 a 4.91 b 8.65 a 3.60 b
Tuber dry matter % 20.1 a 20.5 a 20.9 a 19.6 b 20.8 a 21.0 a 19.4 a 19.2 a
HI % 76.7 a 79.3 a 72.4 a 70.8 a 72.8 a 68.2 a 73.1 a 67.8 a
Calculated RUE g MJ−1 0.89 0.53 0.91 1.22 0.84 0.74 0.87 0.83
Potential yield t ha−1 13.74 11.67 13.51 10.32 14.34 9.47 14.33 9.66
Actual/potential 0.54 0.32 0.52 0.59 0.51 0.52 0.60 0.37
Season 2, 2011/2012
Total DM yield t ha−1 5.80 a 5.49 a 6.37 a 6.41 a 7.57 a 7.76 a 7.22 a 6.98 a
Tuber DM yield t ha−1 3.60 a 3.23 a 4.52 a 4.75 a 5.52 a 5.86 a 4.91 a 4.72 a
Tuber dry matter % 19.5 a 19.8 a 20.4 a 19.5 a 21.5 a 22.3 a 22.9 a 23.4 a
HI % 60.4 a 57.1 a 70.8 a 74.4 a 72.6 a 75.7 a 68.2 a 68.0 a
Calculated RUE g MJ−1 0.62 0.57 0.53 0.52 0.47 0.52 0.45 0.41
Potential yield t ha−1 9.14 6.75 12.81 10.21 13.15 10.49 12.99 11.27
Actual/potential yield 0.39 0.48 0.35 0.47 0.42 0.56 0.38 0.42

For each parameter, means followed by the same letter within the same row (across planting dates) for each
site and season are not significantly different at the p≤0.05 level of probability, according to the LSD test

December 2010, which resulted in hampered emergence, rotting of seed tubers and
leaching of nitrogen from the soil. This was followed by a period of prolonged
drought in both seasons at the time of the sensitive tuber bulking stage (Adams and
Stevenson 1990; Onder et al. 2005).
Secondly, late planting results in a shorter growth period, which results in less
radiation intercepted (Hassanpanah et al. 2009) and increased frost and drought risk
from April onwards, which will decrease tuber yield. Therefore, farmers planting later
than these prescribed dates (mid-October–mid-November) are facing an increased
risk of yield loss. In the present study, late planting indeed resulted in a higher
drought risk towards the end of the growing season. This was especially evident in
S1, when yields were substantially lower for the second planting, partly due to the
drier second half of the growing season, resulting in water stress during the sensitive
tuber bulking stage (Van Loon 1981; Steyn et al. 1998).
The S2 of the present study was generally substantially drier than S1 and the long-
term average (Figs. 2 and 3), especially early in the growing season (November
2011), which resulted in delayed crop emergence, development and radiation inter-
ception (data not presented). Van Loon (1981) also reported that the planting of tubers
 13

in a dry soil could delay emergence and root growth and, accordingly, shorten the
vegetative and tuber formation periods, which will reduce production. As a result of
the much lower rainfall early in S2 (Table 2; Fig. 3), yields recorded for P1 in S2 were
substantially lower than those recorded for P1 in S1 (Table 3).
RUE is a measure of how efficiently the solar radiation intercepted by the canopy is
converted into dry matter (Hammer and Wright 1994; Kooman and Rabbinge 1996;
Lindquist et al. 2005). In this study, the average RUE values recorded were substantially
lower (0.85 g MJ−1 in S1 and 0.51 g MJ−1 in S2), compared to the value of 1.25 g MJ−1
solar radiation (or 2.5 g MJ−1 PAR), which was used in our model, and which is typically
cited for potato (Kooman and Haverkort 1994). Since all production aspects, except for
water, were optimised, the lower RUE values recorded here could probably be attributed
to water stress that interfered with normal production (Kooman and Rabbinge, 1996).
Similarly, the HI gives an indication of the proportion of dry matter that was allocated
to reproductive organs (tubers) (Vos 1997). A low harvest index is usually indicative
of a remarkably shortened growth cycle as the crop was not allowed to grow to
maturity (Kooman et al. 1996). In this study, HI values were mostly close to typical
values (70–80%) reported in literature (Kooman 1995; Victorio et al. 1986), except for
S2, when HI values were substantially lower for both planting dates (P1 and P2)
at Roma and Nyakosoba (Table 3). It is therefore clear that the very dry conditions
towards the end of the growing season (January–March 2012) at these sites (Fig. 3)
resulted in premature crop senescence and lower HI values.
Regarding cultivars, in S1, there were significant differences in their total yields
per site, except for Mokhotlong, where no differences occurred between cultivars
(Table 4). In this season, tuber yield differences between cultivars were signifi-
cant at all sites, except Thaba Tseka. BP1 and Mondial produced higher yields
than Vanderplank at all the other sites. Average tuber DM yields across sites were in the
range of 6.44–7.35 t ha−1 for BP1 (average 6.91 t ha−1), 5.52–7.20 t ha−1 (average
6.49 t ha−1) for Mondial and 4.80–5.50 t ha−1 (average 5.20 t ha−1) for Vanderplank.
Harvests indices, radiation use efficiencies and tuber dry matter contents per site
hardly differed between cultivars, except for Nyakosoba, where DM% values were
the lowest for Mondial. Harvest indices ranged between 65 and 74.5% (average
70.3%) for BP1, 68.0 and 82.4% (average 74.0%) for Mondial, and 71.9 and
79.0% (average 74.2%) for Vanderplank. Radiation use efficiencies were in the range
of 0.72–1.02 g MJ−1 (average of 0.81 g MJ−1) for BP1, 0.60–1.00 g MJ−1 (average
0.83 g MJ−1) for Mondial and 0.78–1.16 g MJ−1 (average 0.92 g MJ−1) for
Vanderplank. Tuber dry matter contents tended to be highest for BP1 (range 20.0–
21.2%, average 20.6%), followed by Vanderplank (19.3–21.5%, average 20.5%) and
Mondial (range 18.7–20.3%, average 19.3%).
In S2, tuber yields were substantially lower than in S1 and significant cultivar
differences occurred. In almost all instances, except for Nyakosoba, Vanderplank again
had lower tuber yields (range 2.4–5.4 and average 4.1 t DM ha−1), compared to BP1
(range 3.7–5.7 and average 4.8 t DM ha−1) and Mondial (range 4.1–6.0 and average 5.0 t
DM ha−1) (Table 4). Harvest index values of cultivars did not differ for all the sites (except
for Thaba Tseka). The HI ranged from 55.6 to 76.9%. Mondial had the highest average HI
(70.7%), followed by Vanderplank (68.5%) and BP1 (66.1%). Radiation use efficiencies
showed little variation per site and ranged between 0.41 and 0.58 g MJ−1 (average
0.51 g MJ−1) for BP1, 0.39 and 0.56 g MJ−1 (average 0.47 g MJ−1) for Mondial and
 14

Table 4 Observed total dry matter (DM) yield, tuber dry matter yield, LINTUL model simulated potential
tuber dry matter yield, tuber dry matter content, harvest index (HI) and radiation use efficiency (RUE) per
cultivar, site and season (average values across planting times)

Season 1, 2010/2011 Season 2, 2011/2012

Parameter Unit BP1 Mondial Vanderplank BP1 Mondial Vanderplank

Roma
Total DM yield t ha−1 9.21 a 6.81 ab 5.94 b 6.35 a 6.61 a 3.98 b
Tuber DM yield t ha−1 6.85 a 5.52 a 4.80 a 3.73 a 4.12 a 2.39 b
Tuber DM % 20.5 a 19.4 a 20.6a 20.6 a 18.2 b 20.2 a
HI % 74.5 a 82.4 a 79.0 a 58.3 a 62.3 a 55.6 a
RUE g MJ−1 0.75 0.60 0.78 0.55 0.56 0.67
Potential yield t ha−1 13.48 13.48 11.16 7.94 7.94 7.94
Actual/potential yield 0.51 0.41 0.43 0.47 0.52 0.30
Nyakosoba
Total DM yield t ha−1 10.43 a 9.80 a 7.22 b 7.24 a 6.61a 5.31 b
Tuber DM yield t ha−1 7.35 a 7.20 a 5.22 b 5.06 a 4.77 ab 4.07 b
Tuber DM % 21.2 a 18.8 b 20.8 a 20.7 a 18.9 b 20.1 ab
HI % 69.8 a 72.6 a 72.4 a 69.5 a 71.9 a 76.3 a
RUE g MJ−1 1.02 1.00 1.16 0.58 0.45 0.54
Potential yield t ha−1 13.10 11.79 10.85 11.51 11.51 11.51
Actual/ potential yield 0.56 0.61 0.48 0.44 0.41 0.35
Thaba Tseka
Total DM yield t ha−1 9.55 a 9.27 a 7.15 b 7.92 a 7.87 a 7.20 a
Tuber DM yield t ha−1 6.44 a 6.68 a 5.28 a 5.67 a 6.04 a 5.36 a
Tuber DM % 20.8 a 20.3 a 21.5 a 23.8 a 20.9 b 21.0 b
HI % 65.0 a 73.1 a 73.5 a 71.4 b 76.9 a 74.3 ab
RUE g MJ−1 0.73 0.83 0.81 0.48 0.49 0.50
Potential yield t ha−1 12.09 12.09 11.52 11.82 11.82 11.82
Actual /potential yield 0.53 0.55 0.46 0.48 0.51 0.45
Mokhotlong
Total DM yield t ha−1 9.70 a 8.97 a 7.73 a 7.58 a 7.09 ab 6.62 b
Tuber DM yield t ha−1 7.01 a 6.56 a 5.50 a 4.90 ab 5.07 a 4.47 b
Tuber DM % 20.0 a 18.7 a 19.3 a 24.0 a 23.4 ab 22.1 b
HI % 71.7 a 68.0 a 71.9 a 65.0 a 71.5 a 67.7 a
RUE g MJ−1 0.72 0.90 0.92 0.41 0.39 0.49
Potential yield t ha−1 12.45 12.45 11.08 12.13 12.13 12.13
Actual /potential yield 0.56 0.53 0.50 0.40 0.42 0.37

For each parameter, means followed by the same letter within the same row (across cultivars) for each site
and season are not significantly different at p≤0.05 probability level, according to the LSD test

0.49 and 0.67 g MJ−1 (average 0.55 g MJ−1) for Vanderplank. Significant differences in
tuber DM% occurred at Roma and Thaba Tseka, where Mondial had the lowest values.
Like in S1, BP1 had the highest average DM% values (range 20.6–24.0% and average
22.3%), followed by Vanderplank (range 20.1–22.1% and average 20.9%), while Mondial
had the lowest DM% values (range 18.2–23.4% and average 20.4%).
 15

Tuber DM% gives an indication of quality and is generally dependent on the

cultivar and environment—especially high temperatures can affect DM% negatively
(Haverkort and Verhagen 2008). In this study, locality (as an indicator of environ-
ment) did not have a clear effect on DM% in S1, but in S2, Roma tended to have
lower DM% values. This could be explained by the fact that the average daily
maximum temperatures at Roma were the highest of all sites (over 25 °C), while
night temperatures were about 4 °C higher than at the other sites. Cultivar clearly
played the dominant role in determining the DM%, with BP1 having the highest
average DM% values and Mondial the lowest.
As was the case with planting date, tuber yields per cultivar were slightly lower in
S2, than in S1 (Table 4). In the first season average yields across sites were 6.91 t
DM ha−1 for BP1, 6.49 t DM ha−1 for Mondial and 5.20 t DM ha−1 for Vanderplank
(overall average of 6.20 t DM ha−1 for all three cultivars), while in S2, the average
yield for BP1 was 4.84 t DM ha−1, 5.00 t DM ha−1 for Mondial and 4.07 t DM ha−1
for Vanderplank (overall average of 4.64 t DM ha−1 for all three cultivars).
In the second season, tuber yields did not vary much among cultivars for the
different sites. BP1 and Mondial were not that different to Vanderplank in yield
levels. As already mentioned, early planting was affected by water stress at the
beginning of the season and also during the critical stages of crop development.
The delayed establishment affected the long season cultivars as they started late and
their yields were reduced as their growth period was shortened.
Selection of appropriate cultivars for planting in a given region is of vital impor-
tance for production of high yields and quality (Dehdar et al. 2012). In general, the
longer season cultivars (BP1 and Mondial) performed better in terms of total and
tuber dry matter yields. Similar to the findings of Hassanpanah et al. (2009), these
longer season cultivars were able to establish well at the beginning of the season and
had a longer time to develop and mature (especially for P1). White and Sanderson
(1983) found that early establishment of medium to long season cultivars such as
Mondial and BP1 gives sufficient time for full bulking after tuberisation. They
indicated that shorter season cultivars (such as Vanderplank) were still able to reach
its potential even when established later. However, in the present study, Vanderplank
lost its leaves, matured before the end of the growing season and always had the
lowest yields, independent of planting date.
Even when planted later (P2) and under dry conditions (e.g. in S2), Vanderplank
was not able to escape drought (a well-known drought avoidance mechanism of early
cultivars) to produce better yields. Kooman and Rabbinge (1996) mentioned a self-
destruction process where at the end of the season when the root zone is depleted
from nutrients, tubers start to withdraw nitrogen from the foliage. They found that
since early cultivars have a shorter leaf longevity, this reallocation of nitrogen from
foliage is faster in early maturing cultivars with smaller foliage than in late cultivars
with a large amount of foliage and longer leaf longevity.
Simulated potential tuber yields for each season and planting date are presented in
Tables 3 and 4. For calculation of the theoretical potential yields, a default RUE value
of 1.25 g MJ−1 intercepted solar radiation and a HI of 0.75 were used. Subsequently,
the ratio of actual tuber yield observed was divided by the simulated yield for both
planting date and cultivars per site, which gave the fraction of potential yield achieved
for each treatment combination (cultivar × plant date × season). When actual observed
yields are
 16

close to the values calculated with the crop growth model, the observed harvest index
should be close to 0.75 and the observed RUE should be close to 1.25 g MJ−1. Table 3
shows the average values per planting across cultivars. The first planting (P1) of S1
showed a HI of about 0.74 and RUE of 0.88 g MJ−1. The second planting (P2), which
was planted a month later showed similar, but slightly lower HI (0.72) and RUE values
(0.83 g MJ−1). During the second planting, plants apparently suffered more from adverse
drought conditions and as a result the RUE and HI values were lower, resulting in lower
actual/potential yield ratios for P2 (0.45), compared with P1 (0.54).
In S2 (2011/2012), planting date had no effect on HI and RUE, and almost
identical values were recorded for P1 and P2. However, both the average HI and
RUE were substantially lower in S2 (HI 0.68 and RUE 0.51 g MJ−1), compared to S1
(HI 0.73 and RUE 0.85 g MJ−1). This resulted in lower average actual/potential yield
ratio (0.43) for S2, compared to S1 (0.50).
Table 4 shows actual and potential yields, HI, RUE values and actual/potential yield
ratios per cultivar for different seasons and sites. The lowest yield of Vanderplank was to
be expected, since it was the earliest maturing cultivar of the three. Vanderplank also
showed the lowest relative performance with actual/potential ratios of 0.45 in the
2010/2011 and 0.35 in the 2011/2012 season (average of 0.42). This is also expected
as earlier maturing cultivars suffer more from transient drought stress periods than later
maturing cultivars. BP1 and Mondial showed similar average actual/potential yield
ratios of 0.50. The actual/potential yield ratio depends on various environmental and
management factors, but typically it is in the range of 0.6 (i.e. actual yields are about
60% of the potential). The lower values reported here could probably be attributed to the
less favourable growing conditions and water stress under dry-land production.
The extensive data set of 3 cultivars × 2 planting times × 4 replicates × 4 sites × 2
seasons gave a total of 196 observations per crop and plant characteristic (variable).
When averaging the four replicates, 48 data points remain, allowing a useful correlation
exercise among some of these variables. Figure 4 shows the relationship between RUE
and tuber yield of the three cultivars planted at four sites, two years and two planting
times. It was expected that there would be a good correlation between these variables, as
yield is one of the components needed to calculate RUE (auto-correlation). However, the
correlation per variety was low for BP1 (R2=0.37) and even lower for the two other
varieties. This shows that factors other than the RUE affected final yield, implying that
the calculated RUE is influenced by factors not included in the model. One possibility is
that nitrogen was leached due to intense rainfall events, leading to variation in the ability
of leaves to photosynthesise adequately. Similarly, prolonged periods of drought
negatively affected the RUE. Moreover, there seems to be two clusters of points in
Fig. 4; across clusters, there is a positive correlation, but within the clusters it is negative.
The lower left cluster is associated with season 2 and the upper right cluster is associated
with season 1. The negative correlation between RUE and yield within a cluster can
probably be attributed to the fact that higher yields are associated with a longer growing
period (later varieties), which also means longer exposure to possible negative
influences such as nitrogen depletion and drought on RUE.
An example of a very good correlation is that between tuber dry matter yield and
total dry matter yield (Fig. 5a, b). The slope (0.7765) represents the harvest index.
Figure 5a clearly shows that the first planting (filled symbols) generally gave higher
yields than the second planting (open symbols), due to better rainfall distribution. Site
differences were
 17

less distinct (Fig. 5a), except for Roma with the majority of observations at the low
end. Figure 5b shows that cultivar Vanderplank had the lowest yields and BP1 the
highest, whereas Mondial yields were least stable and cover the whole range between
250 and 1,050 g m−2. The first season (filled symbols) clearly showed higher yields
than the second season (open symbols), due to higher rainfall (Fig. 5b).
The 48 data points (when the average value of the four replicates were taken) of
each observation allowed for a meaningful regression analysis. Table 5 presents
correlations between the various parameters that were measured or calculated. Cor-
relations are shown for all cultivars and all plantings at all sites (bold text), and for
one cultivar only (BP1 as an example). There were relatively good correlations
(R2>0.5) among several of the different parameters for both BP1 and for all the
cultivars together in both seasons. Significant positive correlations were shown for
– Stem length versus foliar dry mass, tuber yield, total dry matter and tuber number
per plant
– Total dry matter versus foliar dry mass, tuber yield and tuber number per plant
– Tuber yield versus tuber number per plant and harvest index
– Radiation use efficiency and stem length, tuber yield, tuber dry matter concen-
tration, total dry matter and tuber number per plant.
Apparently conditions that lead to higher radiation use efficiency (more rain, lower
light intensity), to longer stems and to higher leaf and stem mass lead to higher tuber
yields. With unchanged or increased HI, this is well expected and explained. The
relationship with a higher tuber number is also expected as in these rain-fed condi-
tions higher yields are associated with more rain and favourable soil water conditions
during stolon and tuber formation, leading to higher stolon and tuber numbers, as was
demonstrated by Haverkort et al. (1990) in controlled and field conditions. The
similarity of the data for BP1 and for all cultivars together (bold) in Table 4 shows
that these correlation values also hold for each individual cultivar.

Fig. 4 Relationship between actual tuber dry matter yields and RUE for the different cultivars across sites,
planting dates and seasons
 18

Roma-P1 Roma-P2 Nyakosoba-P1 Nyakosoba-P2

Thaba Tseka-P1 ThabaTseka-P2 Mokhotlong-P1 Mokhotlong-P2

1200
y= 0.7765x- 47.044
1000 R² = 0.962
Tuber yield (g m-2)

800

600

400

200

0
0 200 400 600 800 1000 1200 1400 1600
Total dry matter (g m-2)

BP1-S1 BP1-S2 Mondial-S1

Mondial-S2 Vanderplank-S1 Vanderplank-S2
1200
y = 0.7765x- 47.044
1000 R² = 0.962
Tuber yield (g m-2)

800

600

400

200

0
0 200 400 600 800 1000 1200 1400 1600
Total dry matter (g m-2)
Fig. 5 Relationship between total dry matter and tuber dry matter yields for (a) the different planting date ×
site and (b) cultivar × season treatment combinations

Concluding Remarks and Relevance of the Findings for Potato Production

Potato production in sub-Saharan Africa has more than doubled since 1994 (FAOSTAT
DATA 2010) but more so due to increased area than through increased yield and more
farmers growing the crop. In Lesotho, the area increased from 500 ha in 1961 to 6,500 ha
in 2009. In most African potato-producing countries, the output of potato increases
faster than the population growth. Rwanda is another example: in 1980, the country
produced 270,000 tons of potatoes on 45,000 ha (6 t ha−1) and currently 2 million tons
are produced on 150,000 ha (13 t ha−1), a 7.4-fold increase, whereas the population
increased about 2.5-fold. Yields on average may not have increased, which is mainly
due to the fact that increasing the area necessarily means producing the crop on more
 19

Table 5 Correlation diagram of ten crop observations

Stem Tuber % dry Foliar Tuber Total dry Tuber number per Average weight per Stem Harvest Radiation use
length matter mass yield matter plant tuber number index efficiency

Stem length 1.0

Tuber % dry matter −0.261 ns 1.0
−0.074 ns
Foliar mass 0.663** −0.091 ns 1.0
0.622** −0.097 ns
Tuber yield 0.587** 0.005 ns 0.460** 1.0
0.547** 0.091 ns 0.401**
Total dry matter 0.670** −0.032 ns 0.669** 0.967** 1.0
0.649** 0.036 ns 0.659** 0.953**
Tuber number per 0.568** −0.318** 0.415** 0.565** 0.590** 1.0
plant 0.625** −0.197** 0.414** 0.560** 0.596**
Average weight per 0.115 ns 0.352** 0.123 ns 0.527** 0.476** −0.338** 1.0
tuber −0.110 ns 0.322** 0.003 ns 0.468** 0.385** −0.380**
Stem number −0.078 ns −0.117 ns −0.065 ns −0.075 ns −0.075 ns 0.019 ns −0.011 ns 1.0
0.107 ns −0.024 ns 0.002 ns 0.083 ns 0.071 ns 0.068 ns 0.001 ns
Harvest index 0.033 ns 0.168 ns −0.374** 0.608** 0.402** 0.191 ns 0.455** −0.021 ns 1.0
−0.074 ns 0.210** −0.463** 0.527** 0.280** 0.120 ns 0.479** −0.002 ns
Radiation use 0.587** −0.404** 0.386** 0.509** 0.537** 0.691** −0.146 ns 0.165 ns 0.143 ns 1.0
efficiency 0.169* −0.281** 0.204** 0.298** 0.306** 0.380** −0.035 ns 0.116 ns 0.053 ns

Regular figures are for BP1 only and bold for all cultivars in both seasons, both planting dates and all four sites
ns non-significant at p≥0.05
*p≤0.05 (significant); **p≤0.01 (highly significant)
 20

marginal land for potato (lower altitudes). This brings down the average yields but there
are also many examples of growers that strongly increase their yields by proper
fertilisation and crop protection techniques. This, combined with the use of certified
seed, will still increase yields further.
Potato is a knowledge-intensive crop and this study contributed to this knowledge.
The situation described here for the conditions of Lesotho (summer rains, cold dry
winters with too low temperatures for potato growth): high elevation with high solar
radiation levels and large differences between daily maximum and minimum tempera-
tures are typical for many other high altitude areas at latitudes between approximately
15° and 30° North and South of the equator. Such conditions are also encountered in,
e.g., northern Ethiopia, some highlands in West Africa such as in Burkina Faso and
Mali. In South America, such conditions prevail in Mexico (Toluca) and in southern
Peru and Bolivia. Globally, subtropical summer highland crops approach an area of half a
million hectare, thereby contributing substantially to food security. In the future, the crop
will strongly benefit from the expected increase in CO2 concentration of the air (Haverkort
et al. 2013) in most of these areas and, with winters warming up a few degrees in the
decennia to come, the longer growing seasons in these areas and growers adapting their
cropping season (Franke et al. 2013) will further contribute to the success of the crop.
Our study revealed a number of findings of relevance for crop ecological science and
for strategies of potato production in subtropical highlands in summer. The current
actual yields in Lesotho are about 13 t ha−1, while the average tuber yields in our field
trials (first planting in the first season) were well above 7.5 t ha−1 dry matter or over
37.5 t ha−1 fresh yield. The strong variation among planting dates and seasons resulted in
average fresh yields of about 23 t ha−1, while the lowest yield obtained was 2.39 t ha−1
dry or 12 t ha−1 fresh tubers for cultivar Vanderplank in the 2011/2012 growing season at
the Roma site. The calculated potential yields were between 9 and 14 t ha−1 tuber dry
matter or between 45 and 70 t ha−1 fresh yields. The main abiotic factor limiting
production proved to be a shortage of water supply, resulting primarily in lower
radiation use efficiency and to a lesser degree a lower harvest index.
The study showed that there is a strong degree of variation in yields across sites,
planting time and cultivar, which can be attributed to:

– Site: Roma lying lowest with the highest maximum temperatures and especially
minimum temperatures and highest amount of long-term rainfall, but with great
degree of erraticness of rainfall pattern had the lowest yields and tuber DM%,
while Thaba Tseka, the highest and coolest site with adequate rainfall, had the
highest yields.
– Planting time: The long-term weather averages show sufficient rain for crop
growth ‘on average’ between October and March (Fig. 2). However, the rainfall
distribution between seasons and within a season is very erratic (2010/2011 and
2011/2012), with excess rain on 1 day (with only part of the water infiltrating into
the soil and the rest running off), followed by subsequent dry spells that can
sometimes last for weeks (Fig. 3). This illustrates that it is opportune to plant
early in summer to assure that enough soil water is collected and the growing
season is long enough for optimal crop growth during the growing season.
– Cultivar: In some conditions, having an early cultivar can be a strategy to escape
droughts that are more frequent towards the end of the growing season. This
 21

usually applies to winter and spring crops in Mediterranean climates. Our study
showed that such a strategy is not opportune for rain-fed summer crops in
subtropical highlands. The early cultivar Vanderplank performed poorly in con-
ditions with transient dry periods. A late cultivar such as BP1 may shed some
leaves when it is dry for a prolonged period but recuperate when it rains again.
It can hence be concluded that the lowest risk of low yields in rain-fed subtropical
highlands is to plant late cultivars as soon as the weather permits it in the highest
areas possible. If options for irrigation can be realised, actual yields of say 15 t ha−1
can be raised to yields of about 25 t ha−1 in dry seasons and well over 37 t ha−1 in
seasons with adequate water supply from rainfall and irrigation, as was illustrated for
experimental conditions in our trials.

References

Adams S, Stevenson W (1990) Water management, disease development, and potato production. Am J Pot
Res 67:3–11
Allen RG, Smith M, Pruitt WO, Pereira LS (1996) Modifications to the FAO crop coefficient approach.
Proc. Int. Conf. Evapotranspiration Irrigation Scheduling, San Antonio, Texas, USA, pp 124–132
Arab HR, Afsharil H, Daliri MS, Laei G, Toudar SR (2011) The effect of planting date, depth and density
on yield and yield components of potato in Shahrood (Iran). J Res Agric Sci 7:141–149
Bureau of Statistics (2008) Lesotho agricultural situation report. Ministry of Finance and Development
Planning, Maseru
Caldiz DO, Gaspari FJ, Haverkort AJ, Struik PC (2001) Agro-ecological zoning and potential yield of
single or double cropping of potato in Argentina. Agric For Met 109:311–320
Dehdar B, Asadi A, Jahani Y, Ghasemi K (2012) The effect of planting and harvesting dates on yield and
vegetative growth of two potato cultivars in Ardabil Region. Int J Agron Plant Prod 3:675–678
FAOSTAT DATA (2013) Agricultural production domain. http://faostat3.fao.org/home/index.html.
Accessed 5 July 2013
Franke AC, Steyn JM, Ranger KS, Haverkort AJ (2011) Developing environmental principles, criteria,
indicators and norms for potato production through field surveys and modelling. Agric Syst 104:297–
306
Franke AC, Haverkort AJ, Steyn JM (2013) Climate change and potato production in contrasting South
African agro-ecosystems. 2. Assessing risks and opportunities of adaptation strategies. Potato Res
56:51–66
Gomez KA, Gomez AA (1984) Statistical procedures for agricultural research, 2nd edn. Wiley, New York
Hammer G, Wright G (1994) A theoretical analysis of nitrogen and radiation effects on radiation use
efficiency in peanut. Crop Pasture Sci 45:575–589
Hassanpanah D, Hosienzadeh AA, Allahyari N (2009) Evaluation of planting date effects on yield and yield
components of Savalan and Agria cultivars in Ardabil region. J Food Agric Environ 7:525–528
Haverkort AJ (1990) Ecology of potato cropping systems in relation to latitude and altitude. Agric Sys
32:251–272
Haverkort AJ, Verhagen A (2008) Climate change and its repercussions for the potato supply chain. Potato
Res 51:223–237
Haverkort AJ, Van De Waart M, Bodlaender KBA (1990) The effect of early drought stress on numbers of
tubers and stolons of potato in controlled and field conditions. Potato Res 33:89–96
Haverkort AJ, Franke AC, Engelbrecht FA, Steyn JM (2013) Climate change and potato production in
contrasting South African agro-ecosystems. 1. Effects on land and water use efficiencies. Potato Res
56:31–50
Kawakami J, Iwama K, Jitsuyama Y (2005) Effects of planting date on the growth and yield of two potato
cultivars grown from microtubers and conventional seed tubers. Plant Prod Sci 8(1):74–78
Kooman PL (1995) Yielding ability of potato crops as influenced by temperature and daylength. Ph.D.
Thesis, Landbouw Universiteit Wageningen, Netherlands
 22

Kooman PL, Haverkort AJ (1994) Modelling development and growth of the potato crop influenced by
temperature and daylength: LINTUL-POTATO. In: Haverkort AJ, MacKerron DKL (eds) Ecology and
modeling of potato crops under conditions limiting growth. Kluwer Academic Publishers, Dordrecht,
pp 41–60
Kooman PL, Rabbinge R (1996) An analysis of the relation between dry matter allocation to the tuber and
earliness of a potato crop. Ann Bot 77:235–242
Kooman PL, Faham M, Tegera P, Haverkort AJ (1996) Effects of climate on different potato genotypes.
Radiation interception, total and tuber dry matter production. Eur J Agron 5:193–205
Lesotho Meteorological Services (2012) Agrometeorology section. Ministry of Energy, Meteorological and
Water Affairs, Maseru
Levy D, Veilleux RE (2007) Adaptation of potato to high temperatures and salinity—a review. Am J Potato
Res 84(6):487–506
Lindquist JL, Arkebauer TJ, Walters DT, Cassman KG, Dobermann A (2005) Maize radiation use
efficiency under optimal growth conditions. Agron J 97:72–78
Lisinska G, Leszczynski W (1989) Potato science and technology. Elsevier, New York
Onder S, Caliskan ME, Onder D, Caliskan S (2005) Different irrigation methods and water stress effects on
potato yield and yield components. Agric Water Manag 73:73–86
Petersen RG (1994) Agricultural field experiments—design and analysis. Marcel Dekker, New York
Ritchie JT (1972) Model for predicting evaporation from a row crop with incomplete cover. Water Resour
Res 8:1204–1213
Smith M, Allen RG, Pereira LS (1996) Revised FAO methodology for crop water requirements. Proc. Int.
Conf. Evapotranspiration and Irrigation Scheduling, San Antonio, Texas, USA, pp 133–140
Spitters CJT (1990) Crop growth models: their usefulness and limitations. Acta Hortic 267:349–368
Spitters CJT, Schapendonk AHCM (1990) Evaluation of breeding strategies for drought tolerance in potato
by means of crop growth simulation. Plant Soil 123:193–203
Steyn JM, Du Plessis HF (2012). Chapter 8. Fertilisation and nutrient deficiencies of potatoes. In: Denner
FDN, Venter SL, Niederwieser JG (eds.) Guide to potato production in South Africa. ISBN 978-1-868-
49-422-4. pp 107–121
Steyn JM, Du Plessis HF, Fourie P, Hammes PS (1998) Yield response of potato genotypes to different soil
water regimes in contrasting seasons of a subtropical climate. Potato Res 41:239–254
Van Loon CD (1981) The effect of water stress on potato growth, development and yield. Am J Potato Res
58:51–69
Van Oort P, Timmermans B, Meinke H, Van Ittersum M (2012) Key weather extremes affecting potato
production in the Netherlands. Eur J Agron 37:11–22
Victorio RG, Moreno U, Black CC (1986) Growth, partitioning, and harvest index of tuber-bearing
solanum genotypes grown in two contrasting Peruvian environments. Plant Physiol 82:103–108
Visser AF (2012). Chapter 6. Selecting a suitable potato cultivar. In: Denner FDN, Venter SL, Niederwieser
JG (eds.) Guide to potato production in South Africa. ISBN 978-1-868-49-422-4. pp 61–98
Vos J (1997) The nitrogen response of potato (Solanum tuberosum L.) in the field: nitrogen uptake and
yield, harvest index and nitrogen concentration. Potato Res 40:237–248
Vos J, MacKerron DKL (2000) Basic concepts of the management of supply of nitrogen and water in potato
production. In: Haverkort AJ, MacKerron DKL (eds) Management of nitrogen and water in potato
production. Wageningen Academic Publishers, Wageningen, pp 15–33
White R, Sanderson J (1983) Effect of planting date, nitrogen rate, and plant spacing on potatoes grown for
processing in Prince Edward Island. Am Potato J 60:115–126
Zones of production. http://www.lanra.uga.edu/potato/Africa/Lesotho.htm. Accessed 28 Nov 2008