METHODS OF LAND PREPARATION AND CROP ESTABLISHMENT
IN RAINFED LOWLAND RICE1
S.K. De Datta, R.A. Morris, and R. Barker2/
Under rainfed conditions, millions of experienced rice farmers
prepare land based on their judgement of soil texture, soil moisture
status and rainfall patterns, topography and degree of weed infes-
tation. The methods they chose must be within their labor and power
resource limitations. Field operations are carried out with the
objective of getting high returns. No matter how much stability of
yield one expects, yields vary from year to year and field to field,
regardless of the strategies used. Because of a dependence on
rainfall, rainfed rice yields are generally more variable than
yields of crops supplied with a more reliable water source. Thus
rainfed rice farmers are influenced by expected rainfall patterns
in these decisions on tillage and stand establishment methods.
For any crop, tillage practices usually have greatest effects
on plant growth during the early growing season during germination,
seedling emergence, and stand establishment stages (Larson, 1963).
What are the physical, chemical, and biological factors that affect
germination growth and stand establishment of seeds? They are
temperature, aeration, and moisture status of seedbed. Recent work
has been conducted to describe conditions under which seed germi-
nation and seedling growth are optimum.
Tillage practices that provide conditions for rapid water
intake and for temporary storage of water on the soil surface or
in the tilled layer are needed to prevent water runoff. Thus,
certain tillage practices reduce erosion and store moisture for
use during the dry part of the summer. Many practices developed
and used by farmers by centuries are based on sound reasons. But
one common reason for thorough land preparation is "pleasing
appearance to the eye". If a farmer is asked why he performs a
certain tillage operation, he can give reasons and may find the
tillage operation to be beneficial but for a wholly different set
of reason than cited. It is therefore imperative to seek answers
-Paper presented at the International Rice Research Conference in
the session on Crop management of rainfed lowland rice, International
Rice Research Institute, April 17-21, 1978, Los Baios, Philippines.
/Head, Department of Agronomy; Agronomist, Cropping Systems Program;
and Head, Department of Agricultural Economics, respectively, The
International Rice Research Institute (mail address: P.O. Box 933.
Manila, Philippines), Los Bafos, Laguna.
to the questions relating every tillage operation "Is it necessary?"
and 'Would the crop grow as well or perhaps better without it?" For
example, for upland soils, if the purpose of operations other than
plowing, roto-tilling or deep cultivation is to compact the soil in
only the top few centimeters thereby increasing bulk density then,
all tillage operations other than primary tillage must be scrutinized
and questioned. The question becomes "Is there value in plowing to
loosen the soil only to repack it to the same bulk density it had
before plowing? These and other questions should be carefully posed
if we have to select tillage practices with sound scientific
research (Blake, 1963).
For rice, the kind and degree of land preparation are closely
related to the method of planting rice and moisture availability.
The benefits cited for land preparation in rice are not greatly
different than for other crops. However, because rice is grown in
diverse land and water management systems, tillage practices for land
preparation vary with the systems of rice culture.
The discussion presented in this paper centers around land
preparation and crop establishment methods for lowland rainfed
rice. The scope of rainfed lowland conditions being considered
excludes rainfed deep-water and/or swamp rice culture conditions.
The biological, physical and economic environments found in deep-
water and swampy areas although overlapping in some respects with
more common rainfed rice -culture are sufficiently different to
justify exclusion from the general theme of this paper. Because the
primary focus of this paper is on rainfed lowland rice, an attempt
will be made to examine tillage operations and establishment methods
and discuss how they may vary depending on the moisture available
from rainfall, soil factors, and resources available to the farmer.
SYSTEMS OF LAND PREPARATION AND CROP ESTABLISHMENT
IN RAINFED LOWLAND RICE
Under rainfed rice culture, the onset of the monsoon deter-
mined the planting time and, thus, the day-length and solar
radiation available during the growing season. Usually, tillage
for rainfed lowland rice is either performed when soil water is
near or above the saturated state, hereafter referred to as wetland
tillage, or performed when soil water is drier than field capacity,
hereafter referred to as dryland tillage. A puddling seedbed is
the objective of tillage in the wet state whereas a loose, granular
seedbed is the objective of tillage in the dry state.
Wetland Tillage and Establishment Methods
Wetland tillage. It is important to understand the benefits of
land preparation by puddling before adoption of an alternative method
is advocated. Land preparation for wetland rice consists of plowing
and harrowing. Giessen (1943) observed that farmers in Indonesia
prepare land for wetland rice by repeated plowing and harrowing -
sometimes as many as four times. Land preparation can be divided
into two tillage operations between the harvest of the previous crop
and the planting of the next crop. In the first or primary tillage
operation, weeds and stubble are plowed under. Secondary tillage
operations consist of harrowings at shallow depths before rice
seedlings are transplanted. These operations further turns the
remaining weed seed and stubble into the reduced layer. The lack
of oxygen in the reduced layer inhibits weed germination.
The major advantages of rainfed wetland tillage are:
1. Improved weed control by primary and secondary tillage,
through puddling action.
2. Ease of transplanting.
3. Establishment of a reduced soil condition which improves
soil fertility and fertilizer management.
4. Reduced draft requirements for primary and.secondary tillage.
5. Reduced percolation losses, and therefore conservation of
rainfall as standing water over the duration of crop growth.
6. Reliability of monsoon rains by the time puddling operations
have been completed.
The major disadvantages of wetland tillage are:
1. Delayed planting and harvesting which may subject the crop
to late season drought stress or reduce the possibility of
successfully growing a second rice crop.
2. High transplanting labor requirement (where topographical
characteristics dictate transplanting must be practiced
to avoid seed and young seedling submergence).
3. Limitations on characteristics of power units suitable for
primary and secondary tillage in wet fields.
In tropical Asia, wetland tillage is the usual way to prepare
fields for rice planting. The traditional.method of preparing land
for transplanted paddy rice has been widely adopted partly because
it greatly reduces the weed population (De Datta, 1973). In addition,
puddling substantially increases the amount of water retained by the
soil (Fukuda and Tsutsui, 1968; Sanchez, 1973; De Datta and Kerim,
1974). Wickham and Singh (1977) have stated that "puddling reduces
percolation losses by decreasing aggregate cohesion, destroying
aggregates, essentially eliminating macroporosity, and increasing
Because puddling reduces the chance that yield reducing moisture
stress occurs between two successive rains, rainfed puddled systems
should result in better moisture conservation and possibly higher
yield than the rainfed nonpuddled system.
Field studies by Choon (1974) have shown how drastically
penetration resistance decreases after a soil has been soaked only
a few days. This resistance decrease has two important impli-
cations. First, draft requirements for primary tillage come into
a range suitable for draft animals and power tillers. Second, the
load bearing capacity is decreased and rolling resistance is inc-
reased, both of which hamper tillage by large tractors.
When clay soils are plowed and harrowed at about soil satu-
ration, several changes take place in the soil structure of which
marked reduction in air-filled pore volume, lowering of permeability,
increase in moisture suction, lowering of resistance to raindrops
and increase in deformability are some examples (Koenigs, 1963).
An important difference between a dry soil and a puddled soil is the
presence of oxidized and reduced soil layer. Ponnamperuma has dis-
cussed electrochemical changes of submergence and the beneficial
effects of reduction on the availability of nutrients to rice (1977).
Although puddling, as practiced in much of tropical Asia,
involves a great amount of labor, the method has been widely
adopted primarily because its compatibility with other components
of production technology and economic conditions.
Crop establishment methods following wetland tillage. Once a
soil has been puddled, possible methods of crop establishment are
restricted, whether under irrigated or rainfed conditions. Under
rainfed conditions, manual transplanting of 20-40 day old seedlings,
is by far the most common establishment method. Direct-seeding of
pregerminated seeds ranks a distant second. Other methods such as
mechanical transplanting, dapog transplanting and drilling of seeds
in rows are not suitable methods or have not been perfected for use
in rainfed fields.
Water control is more critical for broadcast-seeded than trans-
planted rice. Broadcast seeding of pregerminated seed is limited
in adaptation under rainfed conditions because it requires good
drainage control, which means the rice fields must be located on
land having sufficient slope so that undesired early heavy rains
can be drained away. Furthermore, with the direct-seeding method,
achievements of uniform stands require that the puddled surface be
level and smooth so water does not accumulate in depressions, thereby
drowning seeds or seedlings shortly after germination.
Where direct-seeding on puddled soils is practiced, it is pri-
marily done by broadcast method. Such is the case in Sri Lanka,
India, Bangladesh, and the Philippines. Fields are prepared under
wet conditions and the degree of puddling depends on the amount of
moisture accumulated from the rain. Oftentimes, the stand establish-
ment is poor because of poor land preparation and due to insufficient
In the direct-seeded method of stand establishment weeds grow
more vigorously than transplanted rice (De Datta and Bernasor, 1973).
Therefore, weed control is an important consideration in determining
whether to grow transplanted or direct-seeded rice as a method for
Under satisfactory conditions, as are found at the Philippine
Bureau of Plant Industry-IRRI cropping systems site in Iloilo,
farmer-cooperators have achieved high yields for the first crop in
a rainfed two crop patten over two years using adapted broadcast-
seeding methods (Table 1). The farmer have found the broadcast
method reduces labor input and eliminates the difficulty in timing
seedling age to coincide with completion of puddling for
Based on this discussion, it seems that the importance of
puddling depends on a number of factors, but for tropical Asia,
the system evolved because of necessity. On dry soils, it is
difficult to transplant rice, to perform tillage operations and to
control weeds. If there is a switch to direct-seeded rice on dry
soil than the puddling process must be evaluated against salient
land and rainfall factors. Several relevant land and rainfall
factors are covered by Morris and Zandstra (1978). The puddling
process should also be evaluated in the context of the entire rice-
based cropping system rather than on an individual crop (De Datta,
Dryland Rainfed Cultivation
Dryland tillage. For several important reasons, dryland tillage
has been adopted over a limited area for lowland rice production in
tropical Asia. However, recent technological advances may cause the
area over which dryland cultivation is suitable to increase.
The major advantages of dryland tillage for rice are:
1. Initial crop growth is obtained from early monsoon
rainfall, whereas the rainfall is used in for land-
soaking and puddling under wetland preparation.
2. Labor constraints associated with seedbed preparation-
land preparation-transplanting operations are reduced.
3. Large power units can be employed for primary and
secondary tillage operations.
4. Where a non-rice crop follows an unpuddled rice crop,
soil structure is in a more favorable state for stand
establishment and root development of the following
5. Insect and disease pressures which build-up on alternate
hosts during the period required for wetland soaking and
puddling, are avoided.
The major disadvantages of dryland preparation are:
1. Draft power is high and often beyond the reach of poorly
endowed rice farmers in the Asian tropics.
2. Early weed control requirements are comparatively exacting
3. Percolation losses are comparatively high, leaving the
crop more susceptible to periodic drought stress.
4. The crop may be exposed to several soil inhabiting insects
and to blast prior to accumulation of standing water.
5. Fertilizer requirements will often be higher.
For many years, farmers in Bangladesh, northeast India, and
Indonesia have grown rainfed wetland rice seeded directly on non-
puddled soils in bunded fields at the beginning of the rainy season.
This method of rice culture is known as aus cropping in Bangladesh
and northeast India, and as gogo-rancah in Indonesia. The method
takes advantage of early rainfall and mainly allows a second crop
of rice to be grown. Under the system, the land is prepared by
animal-drawn implements or manually. Often, the land is not prepared
into fine surface, and as a result, germination of rice is not
uniform and weeds grow vigorously. Drought damage is also critical.
Undoubtedly, the aus and gogo-rancah tillage and planting systems
have evolved in response to local biological, physical and economic
factors. However, several new technological developments may make
the application of dryland tillage techniques economically feasible
over a wider area. These developments are: (a) early-maturing
modern rice cultivars that provide, under some rainfall patterns, an
opportunity to grow two rice crops if the first crop can be harvested
sufficiently early to complete a second crop, (b) power tillers and
small tractors can be manufactured in several tropical Asian coun-
tries at comparatively low costs and used for tillage and inter-row
cultivation and (c) cheap herbicides which can be used as aids in
combating early weed growth. Another factor, expanding population
pressure, which simultaneously increases food demand and increases
the agricultural labor force, will put more pressure on the land for
increased food production. Where conditions are favorable, some of
the production increase should come from rice sown after dryland
preparation. However, the most difficult problem of using the aus
and gogo-rancah systems over wider areas is quick tillage of a dry
In recent years, a method known as sabog-tanim has been tried
in limited areas in the Philippines. In applied research trials
in Bulacan, Philippines, it was observed that where land was prepared
at the end of the previous wet season, the newly emerged rice crop
was able to survive a period of drought. By contrast, a crop seeded
at the same time but following land preparation at the beginning of
a current wet season suffered from a considerable moisture stress.
It was suggested that crop survival was due to availability of soil
moisture conserved by the presence of a soil mulch in the dry season.
Bolton and De Datta (1977) conducted a field experiment at
IRRI in both dry and wet seasons to test the conservation of soil
moisture by soil mulching, and to determine whether the conserved
moisture was due to control of weeds or to the dry soil mulch. They
also compared the time saved by the early dry-seeding of a rice crop
after the dry soil mulch versus transplanting a crop after traditional
wetland tillage. In plots where land preparation was completed at
the end of the previous wet season, a dry soil mulch was maintained
from February to the first week of May. The soil moisture tension
under the mulch did not exceed 33 centibars, despite a water table
within 1 m of the soil surface during the whole period.
In contrast the soil moisture tension in the weedy fallowed
plots rose to 5 bars at a depth of 15 cm at the end of the dry
season. Weed-free plots conserved some moisture, but not as much
as did the dry soil mulch (Fig. 1). It is obvious that the dry
soil mulch has potential as a practice to conserve soil moisture
for early direct-seeded rice.
During the 1977 crop seasons, experiments were repeated under
lowland rice culture (Singh and De Datta, 1978). The six dry season
treatments are summarized in Table 2. Experiments had the following
tillage treatments: shallow tillage (10 cm deep), deep tillage
(20 cm deep), 5 t/ha rice straw incorporated at 10 cm deep soil,
surface spreading of 5 t/ha of rice straw without incorporation,
weed-free fallow using herbicides, and weedy fallow. All these
treatments were followed by a dry-seeded crop during the wet season
except one transplanted treatment used as a control, i.e., a weedy
fallow followed by puddling followed by transplanting, a common
system practiced by most rainfed rice farmers in tropical Asia.
The transplanted rice control treatment was not part of the main
experiment as it was not randomized within the main experimental
field. However, the transplanted treatment was imposed adjacent
to the main experimental plots and, therefore, some comparisons can
be useful to evaluate advantage of dryland seeding over transplanting
in puddled soil. Four rices were selected of which three were
drought tolerant rices (IR1529-430-3, IR2035-117-3, and IR9575) and
one was drought-suceptible (IR20). These four cultivars were either
dry-seeded or transplanted, depending on tillage treatment, as sub-
plots on each of the main tillage-muleh plots. The rice matured at
different dates, mostly due to differential maturity (Table 2).
After harvesting the first crop, plots were plowed and harrowed
under wetland conditions and transplanted with the same varieties
used in the first crop (Table 2).
Soil moisture tension (SNT) was measured daily, or as weather
conditions permitted, by means of tensiometers and gypsum blocks.
In plots where land preparation was completed at the end of previous
wet season with shallow tillage, deep tillage and straw incorporation,
the SMT did not exceed 40 centibars (Fig. 2). By comparison, the
SMT in the weedy-fallow plots at the end of dry season rose to
5 bars at a depth of 15 cm (Fig. 3). The weed-free fallow and straw
mulch treatments having no tillage operations during the dry season
gave somewhat lower SMT values (Fig. 4) than weedy fallow (Fig. 3).
By the end of the dry season, the SNT at 15 cm rose between 2 and 3
bars indicating some moisture conservation by eliminating or mini-
mizing weed infestation. At 30 cm, there was some conservation of
moisture by the control of weeds. These findings suggest that
keeping the fields weed-free or covering the soil by straw mulch
during the dry season conserved same soil moisture but not as much
that was conserved by shallow tillage, deep tillage and straw
incorporation treatments. In shallow tillage, deep tillage, and
straw incorporation treatments, there was a saving in seeding time
of 2 days as compared with weed-free, 10 days as compared with straw
mulch and as much as 24 days as compared with weedy fallow treatments.
The land preparation for the transplanted crop was started only in
the beginning of July after sufficient rains accumulated for thorough
puddling and it was late by about one month for transplanting compared
with the dry-seeded crop. Crops on the dry soil mulch and straw
incorporated plots matured earliest on Sept. 10 (IR20, IR9575, and
IR1529-430-3) and latest on Sept. 28 (IR2035-117-3). On the other
hand, weedy fallow during the dry season followed by direct-seeding
caused delay in harvest of the first rice crop. Weedy fallow followed
by the puddled transplanted system delayed harvest even further,
primarily because of delay in land preparation (Table 2). Thus, only
one crop of transplanted rice was raised on puddled soil while in the
dry-seeded plots a second crop of transplanted rice was established.
For the first crop, straw incorporation gave an average yield
of 3.7 t/ha, while deep tillage, shallow tillage and straw mulch
treatments produced yields of 3.6, 3.6 and 3.4 t/ha, respectively
(Table 4). The lowest yield of 2.5 t/ha was recorded in weedy fallow
that was dry-seeded in the beginning of wet season. Varietal diffe-
rences were also apparent. IR9575 yielded highest (5.5 t/ha) with
an average yield of 4.3 t/ha, IR1529-430-3 and IR20 yielded about
2.4 t/ha. Among the rices tested, IR1529-430-3 was most severely
affected by brown planthoppers (BPH) which reduced yields.
These studies demonstrated moisture conservation in a mont-
morillonite clay soil with a dry-soil mulch during the dry season.
Most of the moisture was conserved by the soil mulching, not by
control of weeds. This stored moisture may be of value to rice
seedlings encountering drought at the onset of the wet season. It
was noted that an additional benefit of land preparation at the end
of the previous wet season was time saved for the following wet-
season crop; earlier and more timely establishment of the first
crop facilitates more intensive cropping later in the wet season.
The second.crop grown on puddled plots produced considerably lower
grain yield (Table 5) than the first rainfed crop (Table 4).
However, IR9575 produced the highest yield compared with the other
varieties, similar to the crop yields obtained during the first crop.
Two crops of rainfed dry-seeded rice gave 2 t/ha higher yield
than one transplanted crop grown on puddled soil (Table 6). Weedy
fallow during the dry season, followed by first a dry-seeded and
then a transplanted crop produced poor yieldA. IR9575 produced
about a ton/ha more rice than the other varieties (Table 6).
Using a somewhat different approach to examine the effect of
early tillage on rice yields, Hasselbach and van Amson (1965)
working in a rainfed peasant farm sector of Surinam, found diffe-
rences due to early and late fallow period disking on soils with a
long history of rice production even though puddling was practiced
in all treatments for rice (Table 7). In a follow-up test, both
late disking and wet-fallow produced statistically significant
responses in an additive manner on yellow mottled and sandy clay
soils (Table 8). Lack of statistical significant effects on brown
mottled clay soils, sandy loams, and loamy sands are thought to be
due to the few number of observations in each case. The experiments
suggest that other factors in addition to soil water conservation
may lead to increased rice yields following dryland preparation.
Crop establishment methods following dryland tillage. As with
puddling, possible methods of crop establishment are restricted
once a field has been prepared in the dry state. Common establi-
shment alternatives are (a) broadcasting on a levelled field,
(b) broadcasting over shallow furrows and passing a spike-toothed
harrow at 450 to concentrate seed in rows, (c) hand "drilling"
(placing seeds in shallow furrows opened by an animal-drawn
apparatus), and (d) dibbling seeds so that seedlings emerge in
hills at uniform spacings. To the vast majority of tropical Asian
farmers, mechanical drilling is not a foreseeable alternative.
Partly because dry land preparation is not common, relatively
few experiments have been conducted to compare dry establishment
methods under rainfed conditions. However, one such comparison
showed very little difference between grain yields obtained by
broadcasting on levelled field or on furrows opened with a lithao
(Table 9). Although the lithao is satisfactoy for loose upland
soils, it has not worked well on heavy, cloddy lowland soils.
Furthermore, the lithao method also requires extra field operations.
To gain the benefits of row-placement of seeds following dryland
tillage, hand-drilling by dropping seeds in a furrow opened with an
animal-drawn plow produced average yields exceeding 4 t/ha on 48
farmer-cooperators fields during the 1977 wet season in Pangasinan
(Table 10). Although the data presented in Table 10 are combined
across rainfed and partially irrigated fields, the effect of irri-
gation was nil during the wet season and in fact rainfed yields
were higher. Of the 48 observations, the first 24 farmers to plant
obtained yields which averaged 0.4 t/ha more than the remaining
24 farmers. Creation of rows through the row-placement method
aids in weed control and results in a less suitable microenvironment
for insects and diseases. Moreover, placement of the seed below the
soil surface may reduce lodging susceptibility.
Two experiments conducted under rainfed conditions to determine
how far in advance of rainfall onset rice could be seeded in dry
soil without adverse emergence, stand establishment and yield effects,
were conducted in Pangasinan and Iloilo during the 1977 wet season.
Results indicate that seed could remain in the soil for several
weeks without major adverse effects, implying farmers could plant
well ahead of anticipated rainfall onset if land can be prepared
(Table 11). Regardless of planting date, seeds planted.prior to
rainfall onset emerged in response to initial soaking rains.
Dibbling in dry soil is practiced in parts of Java. Experi-
ments reported by CRIA scientists indicate that yield exceeding
4 t/ha, can be achieved if adopted varieties are grown with
appropriate fertilization and weed control (Fagi, 1974; Saefuddin
et al., 1978). Dibbling is labor-intensive, but as water accumulates
in the field, the hill arrangement allows for two-direction rotary
or landak weeding which improves weed control.
Increasing yields of the second rainfed rice crop. With the
advent of modern short duration rice varieties, the possibility
of increasing rice production by growing two rice crops under
rainfed conditions becomes a viable possibility in regions receiving
6-7 months or more of rainfall exceeding 200 mm/month, and perhaps
even possible under some conditions where there are only 5-6 months
of rainfall. However, the performance of the second rice crop is
heavily dependent on avoidance of late drought stress. Three major
considerations arise in attempts to reduce stress on the second crop:
(1) earliness in planting of the first crop, (2) second crop esta-
blishment method and (3) reduced (and faster) tillage for the second
Early first crop plantings can be achieved through dry-seeding
as mentioned in previously. Figure 5 shows the differences in time
of harvest in relation to dry-seeding of IR1529-680-3 and trans-
planting of IR20 in a Central Luzon study (IRRI, 1974). Under
prevailing conditions, transplanting could not be done until
sufficient water had accumulated in late July to allow puddling.
Water management studies show that 200 to 700 mm of water are used
during the land preparation operations prior to transplanting
depending on field conditions, duration of preparation, assumptions
in computation and other factors (Wickham, 1977). Analysis of
rainfall data indicated that the dry-seeded rice crops in the
Pangasinan and Iloilo experiments had completed approximately 50
days and 30 days of growth respectively, before 400 mm of rainfall
had accumulated in those two localities. Dry-seeding results in
net time savings for the second rice crop of over four weeks in
Pangasinan but only one week in Iloilo assuming 3-week old seedlings
had been transplanted as the first crop after the fields were puddled.
Similar advantages attributable to dry-seeding establishment of
the first crop were obtained by Singh and De Datta (1978), showing
that by delaying to transplant the first crop on a puddled soil,
insufficient time remain to plant a second crop within the wet
Transplanting may reduce the field duration of second crop by
two to three weeks, whichwill lower the risk of late season drought
stress during the critical grain-filling period. Evidence of this
effect appears in Table 1, in the form of higher yields and lower
yield variability associated with transplanted second crops in
comparison with.direct-seeded second crops. Furthermore, the
number of complete crop failures under direct-seeding was appro-
ximately twice the frequency under transplanting.
Another factor responsible for delayed second rice crop
plantings is the long time taken to prepare land for the second
crop. Previous studies demonstrate that the degree and duration
of tillage operations may be substantially reduced without adverse
effects on crop growth and yield of rice. This has been amply
demonstrated in several rice growing regions of Asia, including
Sri Lanka (Mittra and Pieris, 1968), Malaysia (Seth et al., 1971),
Indonesia (Varley, 1970), and the Philippines (Moomaw et al., 1968;
De Datta et al., 1977). However, these trials were conducted on
experimental farms and with mechanized tillage operations. Except
for a few trials, most of these were conducted under good water
condition. It is important that such possibilities of reduced
tillage operations and reduced turnaround time should be evaluated
under rainfed rice systems with farmer's resources and tillage
ECONOMIC FACTORS AFFECTING CHOICE OF LAND PREPARATION
AND CROP ESTABLISHMENT
It has already been noted that due to the wide diversity of the
rainfed rice growing environment, a wide range of land preparation,
and crop establishment practices are being utilized. Ultimately,
the choice of appropriate land preparation and crop establishment
in a given area is determined by the benefits and costs associated
with alternative methods. The benefits can be measured in terms
of higher yields, greater cropping intensity, or the reduction in
risk of yield loss. The costs include power for tillage, planting/
transplanting labor and weed control. In this section, we call
attention to some of the important factors that influence the
benefits and costs and hence the decision to use a particular land
preparation or crop establishment procedure in a given area.
In much of the rainfed rice growing environment, delayed
planting results in a loss in yields due two factors: (i) greater
probability of loss in yield due to drought in the ripening stage
and (ii) a difference in environmental factors such as solar energy
and temperature. The pattern of yield reduction will vary from
year to year, location to location and even by varietal type. An
example of the types of yield reduction encountered however, is
illustrated in Figure 6 based on 3 surveys run in a rainfed area
of Central Luzon over a 9-year period. Modern varieties were
grown on 2/3 of these farms even as early as 1971. The average
yield reduction in this location appears to be about 30 kg/day as
the harvest extends into November and December. However, the
introduction of IR36, an early maturing variety (100-110 days for
IR36 vs. 120-140 days for IR8 and and IR20) has had a dramatic
impact on production and yield by allowing farmers to advance
their harvesting date.
It takes about a month after the onset of the rains before
there is sufficient moisture for puddling. Thus, rainfed farmers
typically transplant and harvest a month later than many of the
better irrigated farmers. Depending on the rate of onset of the
monsoons and the degree of flooding, further delays may also be
encountered if conventional puddling and transplanting method are
used. It is practical in many rainfed areas to consider alternative
methods of land preparation and planting.
Let us begin by a .consideration of the power requirement. A
recent study was conducted to determine the power requirements and
yield differences of paddy when fields were prepared using
different combinations of animal and tractor power on 3 irrigated
and one rainfed site (Tables 12 and 13). It cannot be assumed,
from these results that rainfed areas typically require more power
input per hectare. However, cone penetrometer readings of 35 psi
were obtained at about 6 cm soil depth on the rainfed soil and at
almost twice the depth in the irrigated soil indicating that the
rainfed fields were not as well soaked. If a water buffalo works
6 hours per day, a hectare of rainfed Kapalangan soils can be
prepared in about 3 weeks. But the working animals on rainfed
farms in this area range from 0.5 to 1 per hectare. Thus with
available horsepower, the time required to prepare land on a farm
may extend to a month or more.
An economic survey conducted in the same rainfed village
indicated that by employing a four-wheel tractor for primary
tillage, land preparation time was reduced from37 to 27 days
(Bautista and Wickham, 1974, Table 7). Assuming that this
reduction in time could be translated into higher yields (Fig. 6)
the value of the 300 kg additional yield (p300) is only slightly
higher than the custom rate for rotovating one hectare (P220).
A few rainfed farmers in this area do hire tractors, not on a regular
basis, but principally when planting is delayed due to late rains.
The average rainfed farmers who grow only one crop of rice and obtain
yields of less than 2 t/ha cannot afford the power investment needed
to reduce the time required for puddling and transplanting the
monsoon crops. The introduction of a short season variety such as
IR36 clearly has a much greater impact on production and profits
than the increase in farm power. The situation may differ in areas
where double cropping is feasible.
Iloilo Cropping Systems site (municipality of Oton and Tigbauan
in Iloilo Province) provides one example of a rainfed area part of
which can be double cropped (Tinsley et al., 1977). Introduction of
early maturing varieties was the main factor responsible for the
change. Lands prepared before the first of July are sown to early
maturing varieties using the wet-seeded method (broadcasting on
puddled soil). The wet-seeded method saves both time and labor and
thus facilitates the growing of two crops (Table 14). There seems
to be little difference in yield for wet-seeded vs. transplanted
rice, although this is not well documented. On lands prepared after
the first of July, late maturing varieties are transplanted and
only a single crop is harvested. The Iloilo case shows a particularly
skillful adjustment on the part of rainfed farmers to make the most
efficient utilization of limited power and water supplies. The
potential for increasing productivity by double cropping may encourage
mechanization of land preparation in the future, although animals are
the principal source of power at present.
In much of the Central Plain of Thailand where water control is
poor, lands are planted dry and seeds are sown on dry soil before
the rains come. Under this method, weed control is a serious problem,
and as a result yields are substantially lower. Where water control
is adequate, farmers normally shift to transplanting. However, a
recent study by Tipaporn (1976) shows that some farmers are switching
from transplanted to wet-seeded rice. The situation appears to be
similar to Iloilo in that two crops of rice are grown. The major
reasons given by farmers for broadcasting and transplanting are
shown in Table 15. It is possible that some situations are better
suited for one method than for the other. Alternatively, some
farmers may be slow to make the transition.
Perhaps the most dramatic shift in method of stand establishment
in recent times has occurred in Burma. The motivating economic force
has been up until the early 1970's a rise in the wage rates for labor
relative to the farm rice price. Richter (1976) reports that as a
result of these price changes by 1971-72 an estimated 1.4 million
hectares of formerly transplanted land was being broadcast with an
estimated reduction in yield of 0.5 t or more per hectare.
The decision on method of land preparation in rainfed rice is
also related to weed control. Land preparation is in fact a form
of weed control. More thorough land preparation reduces the weed
control problem later on. But we have clearly observed that
rainfed farms are likely to be underpowered. Furthermore, the
intermittent rains that lead to the drying out of the paddy not
only encourage weed growth but also make it impossible to recommend
the inexpensive preemergence herbicide treatments which are being
used in irrigated areas. Inspite of these difficulties, weed control
is less of a problem in transplanted than in broadcast rice. If
weeds are not adequately controlled in the early stages of stand
establishment, the yield losses in broadcast rice may be very severe.
This is one of the risks associated in particular with dryland
In summary, we have observed that changes in technology and
factor price relationships have resulted in significant changes
in land preparation and crop establishment methods. Because of the
heterogeneity in the rainfed environment there can be no typical
set of recommended practices.
Bautista, F. and T. Wickham. 1974. The tractor and the carabao:
A socio-economic study of choice of power source for land
preparation in Nueva Ecija. Saturday seminar, Int. Rice Res.
Inst., July 27, 1974.
Blake, G.R. 1963. Objectives of soil tillage related to field
operations and soil management. Neth. J. Agric. Sci. 11(2):
130-139. (Special issue).
Bolton, F.R. and S.K. De Datta. 1977. Dry soil mulching to conserve
soil moisture in tropical rainfed rice culture. Paper presented
at the 8th Annual Sci. Mtg. Crop Sci. Soc. Phil. 1977 (3). p. 7.
Choon, Y.K. 1974. Soil bearing pressure studies in the Muda Irri-
gation Scheme. MARDI Res. Bull. 2:58-70.
De Datta, S.K. 1973. Principles and practices of rice cultivation
under tropical conditions. ASPAC, Food & Fert. Techn. Center
Extn. Bull. 33. 28 p.
De Datta, S.K. and P.C. Bernasor. 1973. Chemical weed control in
broadcast-seeded flooded tropical rice. Weed Res. 13:351-354.
De Datta, S.K., F.R. Bolton and W.L. Lin. 1977. Prospects for using
minimum and zero tillage in tropical lowland rice. Paper
presented at the 8th Annual Conference of the Pest Control
Council of the Philippines, Bacolod City, Philippines, May 18-20,
1977. (Unpubl. mimeo).
De Datta, S.K. and M.S.A.A.A. Kerim. 1974. Water and nitrogen
economy of rainfed rice as affected by puddling. Soil Sci.
Soc. Am. Proc. 38(3):515-518.
Fagi, A.M. 1974. Some aspects of research on gogo-rancah rice.
Paper presented at Loka-Korya NRRP, Bogor, Indonesia,
April 15-16, 1974.
Fukuda, H. and H. Tsutsui. 1968. Rice irrigation in Japan. Food
and Agricultural Organization (FAO), Rome. 60 p.
Giessen, C. van der. 1943. Rice cultivation in Java and Madura.
Cont. Chuo Noozi Sikenzyoo, No. 11, Bogor, Indonesia. 80 p.
Hasselbach, O.E. and F.W. van Amson. 1965. A study of preseasonal
cultivation methods in rice. Landbouwproefstation Suriname.
Bull. No. 85. Poramaribo, Surinam, S.A.
International Rice Research Institute. 1974. Annual report for
1973. Los Baios, Laguna, Philippines.
Koenigs, F.F.R. 1963. The puddling of clay soils. Neth. J. Agric.
Larson, W.E. 1963. Important soil parameters for evaluating tillage
practices in the United States. Neth. J. Agric. Sci. 11(2):
100-109. (Special issue).
Lokaphadhana, Tipaporn. 1976. Economic comparison of broadcasting and
transplanting dry season rice in Thailand. Unpublished M.S.
Thesis, Faculty of Economics, Thammasat University. Bangkok,
Mittra, M.K. and J.W.L. Pieris. 1968. Paraquat as an aid to paddy
cultivation. pp. 668-672. In Proc. 9th British weed control
Moomaw, J.C., S.K. De Datta, D.E. Seaman and P. Yogaratnam. 1968.
New directions in weed control research in tropical rice.
pp. 675-681. In Proc. 9th British weed control conference.
Morris, R.A. and H.G. Zandstra. 1978. Soil and climatic determi-
nants in relation to cropping patterns. Paper presented at
the International Rice Research Conference, International
Rice Research Institute, April 17-21, 1978. Los Bafos, Laguna,
Orcino, N. and B. Duff. 1974. Experimental results from alternative
systems of land preparation. Saturday seminar, Int. Rice Res.
Inst., July 27, 1974.
Roxas, N.M., M.P. Genesila and E.C. Price. 1977. Farmer's responses
to direct seeding technology in a cropping system rainfed lowland
rice area in Iloilo. Int. Rice Res. Inst., Los Baios, Laguna,
Saefuddin, S.W.A., S. Tajudin, A. Hidayat and S. Effendi. 1978.
Methods of N application and weed control in gogo-rancah rice
culture. IRR Newsletter 3:1.
Sanchez, P. 1973. Puddling tropical soils. 2. Effects of water
losses. Soil Sci. 115(4):303-308.
Seth, A.K., C.H. Khaw and J.M. Fua. 1971. Minimal and zero tillage
techniques and postplanting weed control in rice. Proc. 3rd
Asian-Pacific Weed Sci. Conf., Kuala Lumpur, Malaysia 1:188-200.
Singh, M.P. and S.K. De Datta. 1978. Effect of dry soil mulch on
moisture conservation in rainfed lowland and upland rice.
Saturday seminar, Int. Rice Res. Inst.., April 1, 1978.
Tinsley, R.L., E.C. Price and H.G. Zandstra. 1977. Dynamics of
land-use in rainfed production. Int. Rice Res. Inst. Unpubl.
Varley, J.E. 1970. Gramaxone minimum tillage. Central Research
Institute of Agriculture, Bogor, Indonesia, Staff meeting
Wickham, T. 1977. Water management for lowland rice: Water require-
ments and yield response. Paper presented at the symposium on
"Soils and Rice", Int. Rice Res. Inst., Los Bafos, Laguna,
Philippines, September 20-23, 1977.
Wickham, T. and V.P. Singh. 1977. Water movement through wet soils.
Paper presented at the symposium on "Soils and Rice", Int. Rice
Res. Inst., Los Banos, Laguna, Philippines, September 20-23, 1977.
Table 1. Farmer-cooperator yields of first rainfed direct-seeded (broadcast) and second rainfed
direct-seeded and transplanted rice crops, 1976-77 and 1977-78 wet seasons. Iloilo
First crop Second crop
Wet season Direct-seeded Transplanted
No. of Yield SD1 No. of Yield SD No. of Yield SD
observations (t/ha) (t/ha) observations (t/ha) (t/ha) observations (t/ha) (t/ha)
1976-77 41 5.3 1.5 37 2.2 1.6 4 3.5 1.4
1977-78 22 5.9 1.1 13 1.4 1.8 9 1.7 1.2
Table 2. Summary of treatments used in 1977 soil water conservation experiments.
FIRST CROP (wet season)
Dry season operations Wet season operations
Treatment Tillage operations Planting operations Planting date Harvest date
Dry soil mulch,
10 cm tillage
Dry soil mulch,
20 cm tillage
(5 t/ha rice straw,
10 cm tillage)
Straw mulch (5 t/ha
Rotovated 3x in
rotovated again on
Rotovated 3x between
April 30 and planting
Rotovated 2x between
April 30 and planting
Plowed Ix on May 9,
rotovated 3x between
May 10 and May 24
Plowed and harrowed
2x between July 1
and July 20
Dry seeded furrows
spaced at 25 cm,
Transplanted in 20 cm x
20 cm hills, 28-day old
seedlings, 2-3 seed-
Sept. 17-Oct. 4
Sept. 29-Oct. 15
Table 3. Summary of treatments used in 1977-78 soil water conservation
SECOND CROP (wet season)
Dry season operations Wet season operations
Treatment Tillage operations date Harvest date
Dry soil mulch, Plowed and harrowed 2x Oct. 3 Jan. 11-Jan. 23
10 cm tillage between Sept. 29 and
Dry soil mulch,
20 cm tillage -do- Oct. 3 -do-
(5 t/ha rice straw,
10 cm tillage) -do- Oct. 3 -do-
Straw mulch (5 t/ha
rice straw) Plowed and harrowed 2x Oct. 10 Jan. 17-Jan. 29
between Oct. 5 and
Weed-free fallow Plowed and harrowed 2x Oct. 3 Jan. 11-Jan. 23
(herbicide treated) between Sept. 30 and
Weedy fallow, Plowed and harrowed 2x Oct. 20 Jan. 27-Feb. 9
dry-seeded between Oct. 16 and
Table 4. Grain yield of rainfed-bunded rice as influenced by different dry soil mulch and straw mulch
treatments. IRRI, 1977 wet season.
1 Grain yield (t/ha)
Variety Shallow Deep Straw Straw Weed- Weedy Variety Weedy
tillage tillage incorporation mulch free fallow (mean) fallow
(10 cm deep) (20 cm deep) (5 t/ha) (5 t/ha) fallow dry-seeded transplanted3
IR1529-430-3 2.5 b 2.5 c 2.6 c 2.5 c 2.3 b 2.0 c 2.4 2.2 d
IR2035-117-3 4.4 a 4.2 b 4.3 b 3.9 b 4.1 a 3.3 a 4.0 4.0 a
IR9575 4.7 a 5.2 a 5.5 a 4.5 a 4.0 a 2.7 b 4.4 3.4 b
IR20 2.6 b 2.5 c 2.4 c 2.7 c 2.3 b 1.9 c 2.4 2.9 c
Mulching (mean) 3.6 3.6 3.7 3.4 3.2 2.5 3.1
CV for mulching
CV for varieties
1Average of four replications.. 21n a column, any means followed by the same letter are not. significantly
different at the 5% level. 3Not included in the analysis but tested in adjacent areas.
Table 5. Grain yield of rainfed-bunded transplanted rice as influenced by different dry soil mulch
and straw mulch treatments. IRRI, October 1977-January/February, 1978. (Second rainfed
1 Grain yield (t/ha)
Variety Shallow Deep Straw Straw Weed- Weedy Variety
tillage tillage mulch incorporation free fallow (mean)
(10 cm deep) (20 cm deep) (5 t/ha) (5 t/ha) fallow
IR1529-430-3 1.1 c 1.1 c 1.0 c 1.3 c 1.2 c 0.7 c 1.1
IR2035-117-3 1.7 b 1.6 b 1.4 b 1.8 b 1.7 b 1.2 b 1.6
IR9575 2.0 a 2.3 a 1.8 a 2.5 a 2.4 a 1.5 a 2.1
IR20 1.5 b 1.6 b 1.2 bc 1.6 b 1.3 c 0.9 c 1.3
Mulching (mean) 1.6 1.6 1.4 1.8 1.7 1.1
CV for mulching
Average of four replications.
significantly different at the
CV for varieties
2In a column, any means followed by the same letter are not
Table 6. Grain yield total of rainfed-bunded rice-rice rotation as
influenced by different dry soil mulch and straw mulch
treatments. IRRI, 1977-78.
Grain yield (t/ha)
Treatment dry-seeded transplanted Total
rice crop rice crop
Apr.-Nov. 1977 Oct. 1977-Feb. 1978
Shallow tillage (10 cm deep) 3.6 1.6 5.2
Deep tillage (20 cm deep) 3.6 1.6 5.2
Straw mulch (5 t/ha) 3.4 1.4 4.8
Straw incorporation (5 t/ha) 3.7 1.8' 5.5
Weed-free fallow 3.2 1.7 4.9
Weedy fallow, dry-seeded 2.5 1.1 3.6
Weedy fallow, transplanted2 3.1 3.1
IR1529-430-3 2.4 1.1 3.5
IR2035-117-3 4.0 1.6 5.6
IR9575 4.4 2.1 6.5
IR20 2.4 1.3 3.7
Mulching treatment average is based on sixteen plots and variety average
on twenty-four plots. Not included in the statistical analysis but
tested in adjacent areas.
Table 7. Grain yields (t/ha) from 4 observations on the effect of disking
during the fallow period prior to puddling and transplanting.
(Hasselbach and von Amson, 1965).
Soil texture 1 early disking fb
Puddled Early disking Late disking 2x late disking fb
only fb puddling fb puddling puddling
Heavy clay 1.9 2.4 2.6 3.7
Heavy clay 2.8 2.9 3.5 3.9
Heavy clay (new
ricefield) 4.3 4.0 4.0 4.3
Sandy loam 1.2 1.8 1.9 2.6
fb = followed by.
Table 8. Effects of dry fallow disking and preseasonal water management
on yield responses. (Hasselbach and von Amson, 1965).
significance of Best approx. of
Soil group No. of effects and treatment
replications interactions effect (kg/ha)
G W WG G W
Brown mottled clay 3 ns ns ns +219 -153
Yellow mottled clay 23 ** ** ns +302 +239
Sandy clay 9 ns 4-260 +303
Sandy loam plus
loamy sand 4 ns ns ns +347 +278
G = Disc-harrowed once during late fallow period.
W = 2-3 months "flooded" in the minor rainy season.
* = Significant at level P = 0.05.
** = Significant at level P = 0.01
ns = Not significant.
Table 9. Grain yields (average of IR1561-228-3 and IR30)
as affected by seven patterns of rainfall1 in
the Philippines and by two methods of stand
establishment by dry seeding.
Rainfall pattern Grain yield (t/ha)-
A 3.2 3.3
B 1.9 2.3
C 4.2 3.8
D 2.1 2.1
E 2.9 2.9
F 5.3 5.5
G 4.5 4.8
Average 3.4 3.5
In the Philippihes, rainfall pattern Type A and Type B
each have a long wet season (WS) during the high sun
period (HSP) (April through September); pattern A has a
5-month dry season (DS) and pattern B has a 4-month DS.
Type C and Type D each has a short DS (1 to 3 months) .
The DS of Type C occurs during the low sun period (LSP)
(October through March), while the DS of Type D is during
the HSP. Type E and Type F each have a marked WS of heavy
rainfall; the WS of Type E occurs during the LSP and the
WS of Type F is during the HSP; Type G has an even distri-
bution of rainfall with no marked seasonality.
DSL = dry-seeded in rows with lithao (a carabao-drawn
farm implement with five teeth spaced at 20 cm, used
primarily to make furrows in planting upland rice, and
is indigenous to the provinces of Cavite and Batangas,
3DSB = dry-seeded broadcast.
Table 10. Yields of dry-seeded rice, 48 farmers' fields
combined rainfed and partially irrigated obser-
vations. Crop year 1977-78, Manaoag, Pangasinan.
Yield range No. of Average yield
(t/ha) observations (t/ha)
< 3.5 7 3.1
> 3.5 < 4.5 16 4.2
> 4.5 < 5.5 17 5.0
> 5.5 < 6.5 5 6.1
> 6.5 3 6.9
Effect of seeding date on yield of dry-seeded rice in
Pangasinan and Iloilo, 1977 wet season.
Seeding date Yield (t/ha) Seeding date Yield (t/ha)
April 22 4.5 May 6 3.6
April 27 5.0 May 11 4.2
May 2 4.9 May 16 4.0
May 7 4.6 May 21 3.7
May 12 5.1 May 26 3.8
May 17 4.6 May 31 3.7
May 22 5.0 June 5 4.5
May 27 3.9 June 10 4.5
Treatment MS 652899ns 537695ns
s- 300 371
CV (%) 13 18
Alternative land preparation treatments, 3 barrios,
Nueva Ecija, 1973 wet season.
Land preparation method
Treatment Primary Secondaryl
Power source Implement Power source Implement
1 65 hp tractor rotary carabao comb
2 14 hp tiller rotary carabao comb
3 7 hp tiller moldboard 7 hp tiller comb
4 water buffalo moldboard 7 hp tiller comb
5 water buffalo moldboard carabao comb
iSecondary tillage consists of two passes over the field repeated
three times at one week intervals.
Source: Orcino and Duff, 1974.
Table 13. Site characteristics, soil conditions and level of inputs
used in land preparation trials. 3 barrios, Nueva Ecija,
1973 wet season.
Site/treatment Labor input (hr/ha) readingI Mean2
Plow Harrow Total (depth in cm for yield
Plow35 psi) t/ha)
35 psi) (t/ha)
Baluarte (shallow hardpan)
Pulo I (medium hardpan)
Pulo II (deep hardpan)
1At beginning of land preparation.
2Averaged over the
Source: Orcino and Duff, 1974.
Table 14. Most important advantages and disadvantages of
planting wet-seeded rice reported by 39 farmer-
respondents in Iloilo, 1976.
Labor is less
Cost is lower
Double cropping possible
Water availability early in season
Risk is great
High seed requirement
High fertilizer requirement
Source: Roxas et al., 1977.
Major reason given for either broadcasting or
transplanting rice by a sample of 94 farmers.
Amphoe Bang Nam Prieo, Changwat Chachoengsuo,
Broadcast only (78%)
Cost is lower
Cost is lower
Labor is less
and yield is same or higher
Transplant only (10%)
Sure that yield will be higher
Weeds easier to control
Less fertilizer is required
Source: Lokaphadhana, 1976.
SOIL MOISTURE TENSION (bars)
DRY SOIL MULCH (no weeds)
FEB MAR APR MAY JUNE JULY
--Tillage---Dry soil mulch I I Dry-seeded rice crop I
i-----Weed free- I -I Dry- seeded rice crop I
We y fallow Tillage
FEB MAR APR MAY JUNE JULY
I- Weedy fallow----- lloge-t---Dry-seeded rice crop----
Fig. 1. Soil moisture tension under a dry soil mulch
did not exceed 33 centibars at any time during the
1976 dry season. Some moisture was conserved by
keeping the plots weed free, but moisture tension
went higher than in the soil mulched plots. Plots
that were fallow and on which weeds were allowed
to grow had 5 bars of soil moisture tension at a
15-cm depth by the end of the dry season. IRRI,
1976. (Bolton and De Datta, 1977).
Soil moisture tension (bars)
Feb Mar Apr May Jun Jul
Fig. 2. Effect on soil moisture tension of a dry soil mulch generated by
shallow tillage (10 cm deep) and deep tillage (20 cm deep) and straw
incorporation at 5 t/ha (10 cm deep) during the dry season followed by (fb)
a dry-seeded rice crop. Tillage consisted of 3 rotovations in January/
February fb 1 rotovation in April. Rices were dry-seeded on 30 April.
IRRI, 1977 dry and wet seasons. (Singh and De Datta, 1978).
Soil moisture tension (bars)
Mar Apr May Jun Jul
Fig. 3. Effect on soil moisture tension of a dry-season weedy fallow
followed by a transplanted rice crop. Tillage consisted of 2 plowings
and 2 harrowings. The seedbed was sown 22 June and seedlings were
transplanted 20 July. IRRI, 1977 dry and wet seasons. (Singh and
De Datta, 1978).
Soil moisture tension (bars)
Weed free fallow
Fig. 4. Effect on soil moisture tension of straw mulch at 5 t/ha (without incorporation) and weed-free fallow treatments
during the dry season followed by a dry-seeded rice crop. Tillage consisted 3 rotovations in straw mulch and 2 rotovations
in weed-free fallow plots. Rices were dry-seeded on 10 May and 2 May in straw mulch and weed-free fallow plots, respectively.
IRRI, 1977 dry and wet seasons. (Singh and De Datta, 1978).
Se PI F H
i i I I Direct seeded
Late planting of IR20 I
So T H
I I- Early planting of IR20
A M J J A S 0 N D J F M
Fig.,5. Time of seeding and harvesting of direct-seeded
IR1529-680-3 and transplanting of IR20 in relation to the
amount-of rainfall. San Rafael, Bulacan, 1973 wet season.
' I !
Groin yield (t/ha)
Ist 2nd 3rd 4th Ist 2nd 3rd 4th Ist 2nd 3rd 4th Ist 2nd 3rd 4th
I- Sept -- I Oct Nov I Dec-
Date of harvest
Average rice yield of the most widely planted variety in each of
by date of harvest in rainfed areas of Bulacan and Nueva Ecija,
1977. (Figure above each bar refers to percent of area harvested.)