Professional Documents
Culture Documents
ABSTRACT
Not much information has been publicly available on the diff usion of modern varieties (MVs) from the 1990s and on current
diversity of rice (Oryza sativa L.) varieties in farmer fields. Using national household surveys of Philippine growers conducted in
1993, 1997, and 2002, this study describes the adoption of specific MVs and variety groups in farmer fields; measures the adop-
Rice
tion rates of newly released varieties; and analyzes the spatial diversity of varieties planted in farmer fields. Data showed that
IR64, IR74, IR42, and Burdagol are enduring varieties from 1970s and 1980s, and PSB Rc18, PSB Rc10, PSB Rc28, Masipag,
PSB Rc14, and PSB Rc82 are the most commonly planted new varieties. The MV3 varieties (released mid-1980s to mid-1990s)
were popularly planted with an observed decrease in the use of MV2s (released mid-1970s to mid-1980s) and increase in the use
of MV4s (released after 1995). Around 30 to 40% of the total rice area is planted to new rice varieties and the aggregate replace-
ment period of rice varieties is around 8 to 11 yr, with faster adoption rates during dry season (DS) in irrigated areas. On average,
only 10 varieties occupy around 70% of the rice area in major rice-producing provinces in one season. Indices of spatial diversity
show wide variability across provinces in terms of the richness of diversity, dominance of specific varieties, and the equality of
abundance, but no clear trend across periods. Policymakers are thus encouraged to continually support and strengthen the cur-
rent efforts of public rice breeding research and extension, even as research beyond analysis of on-farm morphological diversity
is recommended.
1380 A g r o n o my J o u r n a l • Vo l u m e 10 0 , I s s u e 5 • 2008
persal in the major rice-producing regions of the country would variety groups; (ii) measure adoption rates of newly released vari-
indicate to some extent the genetic diversity of the rice crop eties; (iii) analyze the spatial diversification of varieties planted
being planted in the country. Finally, information on rice vari- in farmers’ fields using both actual spatial data tabulation and
ety diff usion by province, season, and ecosystem can provide spatial diversity indices; and (iv) draw policy recommendations
basis for decision making of researchers, extension workers and from the results of the study.
local policymakers in their research and development activities,
particularly in location-specific variety development, seed pro- MATERIALS AND METHODS
duction, and targeted variety promotion programs. Data Set
The major players in rice breeding research in the Philippines The study used data from the rice-based farm house-
are the International Rice Research Institute (IRRI), hold survey (RBFHS) being conducted every 5 yr by the
Philippine Rice Research Institute (PhilRice), Bureau of Plant Socioeconomics Division (SED) of PhilRice. To our knowl-
Industry (BPI), University of the Philippines at Los Baños edge, this is the most useful source of information on specific
(UPLB), and recently, some private seed companies. These variety-use represented by the most number of rice-producing
institutions develop and submit promising lines to the Rice provinces in the Philippines. The Bureau of Agricultural
Varietal Improvement Group (RVIG), a multi-disciplinary, Statistics (BAS) conducts a regular quarterly survey on rice
multi-agency body that tests, evaluates, and recommends to the production for all provinces but they are only able to classify
National Seed Industry Council (NSIC) varieties for release area harvested to MVs and traditional varieties (TVs), and
either as a commercial variety for all regions of the country or very recently, hybrid rice varieties. Thus, despite the drawbacks
for specific regions or sites. Upon approval of release, breed- on the RBFHS data set as will be mentioned later, this study
ing or host institutions submit breeder seeds to PhilRice and used the RBFHS cross-section data sets covering 1992–1993,
its network for the production of the foundation seeds (FS) 1996–1997 and 2001–2002 crop years, with a quick reference
of these released varieties. The FS are distributed to the seed to some reports of the earliest 1988 survey in several regions.
production network (SeedNet), which produces the registered The RBFHS in 1992–1993 covered 15 major rice-producing
seeds and sell to other seed growers who then produce the certi- provinces with 1100 respondents sampled purposively. The
fied seeds released to farmers. Since 1968, 138 varieties have been total of the rice areas in these major rice-producing provinces
released, of those, 92 for irrigated, 19 for rainfed, 13 for upland, comprise around 50% of the country’s total rice area harvested.
6 for cool-elevated, and 8 for saline-prone rice areas (Padolina, On the other hand, the 1996–1997 and 2001–2002 surveys
1995; PhilRice, 1998, 2007). About 70% were developed for the covered 30 and 33 provinces with 2239 and 2474 respondents,
irrigated lowland rice ecosystem (including hybrid rice varieties), respectively. The sum of the rice areas represented by the prov-
generally for transplanted rice. The NSIC released an average of inces covered in the 1996–1997 and 2001–2002 surveys are
around four varieties every year since 1968 (Fig. 1). around 70% of the country’s total rice area (Fig. 2). In the 1997
Most rice varieties were originally released for irrigated and 2002 surveys, the province served as domain of the study
lowland, rainfed-lowland, and upland ecosystems, until after and all barangays (villages) in the province served as sampling
1990 when IRRI and PhilRice began breeding rice varieties population. This study used a two-stage sampling selection.
specifically for cool elevated, saline-prone, and transplanted The first stage is the barangay (equivalent of a village), selected
and direct-seeded rainfed-lowland. The early varieties released by using systematic random sampling, and the second stage
were bred with the objective of creating a plant type that would unit is the rice farm household, selected using the right cover-
be resistant to lodging and would make efficient use of solar age method (PhilRice, 1997). This method is used in ensur-
energy and fertilizer to achieve high yields (Barker et al., 1985). ing random sampling even without the list of farmers in the
Later, breeders incorporated in their variety development char- barangay. A data collector starts from a designated landmark
acteristics such as resistance to pests and diseases and improved (for e.g., barangay hall, church, or school) and interviews a
grain quality. Much later, breeders bred specifically for some qualified respondent every three houses from the landmark.
adverse environments such as the cool-elevated area and the The data collector always moves in the right direction until he
saline-prone environments. The first varieties for
cool-elevation and saline-prone areas were released
by 1995. Although breeding for hybrid rice started
much earlier and the first hybrid rice was released
in 1994, and then followed by the PSB Rc72H
(Mestizo) in 1997, most of the hybrid rice varieties
were commercially released after 2002, which was
after the national program for hybrid rice commer-
cialization was launched in 2001.
This study was conducted to analyze the adoption
of these officially released rice varieties in farmers’
fields using national rice farm household surveys
conducted in 1993, 1997, and 2002. The specific
objectives of the study were to: (i) describe the
recent diffusion adoption of specific rice varieties
in farmer fields, and compare the adoption rate by Fig. 1. Yearly release of rice varieties in the Philippines (1968–2006).
Data Analyses
The study approached adoption and diversification analy-
sis descriptively since it is simple to execute while giving a
clear picture particularly on specific varieties. Percentage
distribution of respondents by specific variety planted, and
share to total area planted by variety were used to describe
the recent diffusion adoption of specific rice varieties. In
Fig. 2. Survey municipalities in the Rice-based Farm addition, because there are more than 100 varieties released
Household Survey (2001–2002).
since 1968, we also tried to group divide these varieties into
four distinct groups based on the dates of release and their
reaches the sample quota. Appendix Table 1 (available by distinct characteristics following Estudillo and Otsuka
request from the author or the journal) shows the distribu- (2006), and distinguishing as fourth-generation varieties
tion of the number of barangays per province covered in the (MV4) the varieties released after 1995 most of which were
three RBFHS. Despite the asymmetric sourcing of data, developed for adverse production environments (Table 1).
particularly between the 1993 survey and the 1996 and Following Estudillo and Otsuka (2006), the first-generation
2002 surveys, the data sets were used as is treating them MVs (MV1) consisting of IR series from IR5 to IR34 devel-
as cross-section data in the discussion of variety adoption. oped by IRRI and C4 series developed by the University
The total area planted of respondents was always used in of the Philippines were released from the mid-1960s to the
the derivation of indices, and in the case of 1996–1997 and mid-1970s. The MV1 varieties were potentially higher yield-
2001–2002 surveys, the difference in area is not significant ing than traditional varieties (TVs) under ideal conditions
so any anticipated variation in results may not be very sig- but not necessarily so on farmers’ fields due to its suscepti-
nificant to affect the trends. bility to pests and diseases. The MV1 varieties are more fer-
tilizer responsive than TVs because they are short
Table 1. Commercially released rice varieties† in the Philippines, 1968–2005. in stature and have stiff straw that enable them to
Total no. of bear more grains with greater fertilizer applica-
Variety Irrigated- Rainfed- Cool- Saline- varieties tion. Second-generation MVs (MV2) consisting
group‡ lowland lowland Upland Hybrid elevated prone released
MV1§ 16 1 1 0 0 0 18
of IR36 to IR62 were released from the mid-
MV2¶ 18 3 6 0 0 0 27 1970s to the mid-1980s. The MV2 varieties were
MV3# 20 8 1 1 2 2 34 designed to ensure yield stability by incorporating
MV4†† 30 7 5 7 4 6 59 multiple pest and disease resistance. The third-
Total 84 19 13 8 6 8 138 generation MVs (MV3) consisted of IR64 to IR72
† Varieties approved for release by the National Seed Industry Council (NSIC), previously the
Philippine Seed Board.
and PSBRc2 to PSBRc74 were released from the
‡ As defined by Estudillo and Otsuka (2006) with an additional category included (MV4 defined as
mid-1980s to the late 1990s. The MV3 varieties
those released after 1995). incorporated better grain quality and stronger host
§ MV1, IR series from IR5 to IR34; C4 series from the mid-1960s to the mid-1970s. plant resistance. For policy purposes, we classified
¶ MV2, IR36 to IR62 released from the mid-1970s to the mid-1980s. hybrid varieties as a separate category. Traditional
# MV3, IR64 to IR72; PSB Rc varieties released from mid-1980s to mid-1990s. varieties and varieties bred by the nongovernment
†† MV4, varieties released after 1995. organization Magsasaka at Siyentipiko para sa pag-
Shannon Both richness D = –∑pilnpi The pi is the proportion, or relative The pi is the area share occupied by the ith variety
and relative (D =>0) abundance of a species
abundance
† Source: Smale et al. (2003); Mathematical construction and explanation defined by Magurran (1988).
unlad ng Agrikultura (MASIPAG) were also under separate where R it is the age of the variety in terms of the number of
categories. A percent distribution of respondents by variety years (at time t) since the release of variety i. Th is measure
groups was used to compare adoption by variety groups. avoids the use of an arbitrary defi nition of “new” or “recent”
Conceptually, a promising line undergoes a series of field varieties (Brennan and Byerlee, 1991). The oldest and new-
performance tests and evaluation trials before it can be released est varieties planted were also identified in the three survey
as a commercial variety. In the Philippines, a promising line periods to provide an additional indicator of how fast newly
coming from preliminary yield trials is subjected to national released varieties are observed in farmers’ fields.
cooperative testing and multi-adaptation trials, and the check To analyze the spatial diversity of varieties in farmers’
variety usually used is the best available existing variety. fields, both actual spatial data tabulations and spatial diver-
Therefore, newly released varieties are expected to perform sity indices were derived. The estimated proportion of area
similar to if not better than the existing varieties in the given planted to only one or few varieties in a province and the
environment for which it is recommended. When a farmer number of released varieties planted by farmers in a province
chooses to adopt a new variety in place of an older variety, it at a given year and season were also described. Spatial diver-
reflects the farmer’s judgment that the new variety offers some sity was also analyzed using diversity indices used by ecolo-
net benefit or advantage (Evenson and Gollin, 2003). With this gists following Benin et al. (2004) and Smale et al. (2003).
in mind, we also calculated simple indices to understand more Table 2 shows the concepts and mathematical formulae used
the dynamics of rice variety adoption, especially of the newly in deriving the indices adapted from Smale et al. (2003) but
released varieties. Indices such as the proportion of recent vari- applying it in the case of rice.
eties and weighted average age of varieties were calculated to
determine the adoption rate of the newly released varieties fol- RESULTS AND DISCUSSION
lowing Brennan (1984). Modern Rice Varieties Planted in Farmers’ Fields
The proportion of recent varieties (Brennan, 1984) is the It did not take long for farmers in the Philippines to adopt
proportion of the total area that is planted to recently released MVs. After their introduction in the mid-1960s, more than
varieties. This index, qit, is computed as follows: 50% of the total rice area was planted to MVs by 1970, espe-
cially in the irrigated ecosystem (Fig. 3). This adoption rate
qit = pit if year of release ≥t – m, steadily grew, and by 1980, around 90% of the irrigated areas
qit = 0 if year of release <t – m, were planted to MVs and 80% in the rainfed areas. The 2002
data showed that only <5% of the total area harvested was not
where pit is the proportion of the area sown to variety, i, in year, planted with MVs and in 2004, almost 100% of the total pro-
t; m is the number of years used to define “recent.” Then duction in the irrigated areas came from MV production.
Using the three cross-section survey data, Table 3 shows the
It = ∑q it most common varieties planted in the years after 1990 based on
i
number of farmers reporting. Enduring varieties released in the
where I is the proportion of the total area that is sown to varieties 1970s and 1980s are IR64, IR74, and IR42 and Burdagol later
released in the previous m years. In this particular study, we assumed released as PSB Rc34. IR64 still ranked first or second until
a lag of 2 yr between the release of a variety and its availability to the 2002 survey. In the 1997 and 2002 surveys, PSB Rc10, an
farmers, and defined recent variety as a variety that is available to extremely early-maturing variety released in 1992, and PSB
farmers for five more years thus in this study we used m = 7. Rc18, released in 1994 and known for its resistance to pests and
The weighted average age of varieties (following Brennan and diseases particularly during the wet season (WS), were widely
Byerlee, 1991), on the other hand, consists of the weighted aver- used along with IR64. PSB Rc14, released in 1992, and PSB
age age of varieties grown by farmers in a given year, measured Rc28, released in 1995 were also widely used in both seasons
in years from varietal release and weighted by the proportion of and ecosystems largely for its high yield potential and good
area sown to each variety at that time. This index, WAt, is com- grain quality. Masipag varieties bred by a nongovernment orga-
puted for a given year, t, as follows: nization largely through farmers’ selection strategy also ranked
high. PSB Rc80 and PSB Rc82 were widely used among the
WAt = ∑p R it it
varieties released in 2000. PSB Rc82 has the highest maximum
i
Table 5. Indices of adoption rate of new varieties, by season and farm type, Philippines.
Proportion Most recent variety used Oldest variety used
No. of sown to Weighted
Year released recent average age
and varieties varieties‡ of varieties§ Year
Season Ecosystem planted† (It/%) (WAt/yr) Variety Year released Variety released
1992 WS irrigated 39 32.9 9 PSBRc6,-10,-12 May 1992 IR8, C4–63G, BPI76 1968
rainfed 37 20.1 10 PSBRc2,-4 Nov. 1991 IR8,-5, C4–63G, BPI76 1968
1993 DS irrigated 34 37.0 8 PSBRc8,-10,-12,-14 May 1992 IR8, C4–63G, BPI76 1968
rainfed 8 10.3 14 IR72 1988 IR36 1976
1996 WS irrigated 53 51.4 9 PSBRc5,-54,-56,-60 Nov. 1997 IR20 1969
rainfed 45 43.5 10 PSBRc3,-5,-60,-72H Nov. 1997 IR20 1969
1997 DS irrigated 49 54.7 8 PSBRc3,-5,-54,-60 Nov. 1997 IR20 1969
rainfed 38 57.4 9 PSBRc5,-60,-68 Nov. 1997 IR20 1969
2001 WS irrigated 47 23.5 11 NSIC Rc122, PSB Rc 90,-94 Dec. 2003, Jan. 2001 BPI Ri3 1973
rainfed 49 28.9 11 PSB Rc90/-94 Jan. 2001 IR29, IR32 1975
2002 DS irrigated 54 36.0 9 PSBRc90,-94,-98,-100 Jan. 2001 IR29, IR32 1975
rainfed 49 32.0 10 NSIC Rc122, PSB Rc 94,-100 Dec. 2003, Jan. 2001 IR42 1977
† This is the total number of released varieties mentioned by farmers. This does not include varieties which are not released by the Philippine seed board or now, the
National Seed Industry Council, such as the Masipag varieties or other selections or lines; traditional varieties; or farmer-selections or farmer-named varieties.
‡ Proportion of recent varieties is an index showing the proportion of the area planted to varieties released in the previous 7 yr.
§ Weighted average age of varieties is an index of the average age of released varieties grown by farmers, measured in years from varietal release and weighted by the
proportion of area sown to each variety.
The wide difference in the diversity index across provinces therefore that future analyses consider the parental lines
is also reflected in the calculated Shannon index–for example, or the genetic makeup of the current rice varieties planted
minimum of close to zero (0.13) and maximum of 2.85 in in the farmer fields to determine the exact state of genetic
2001 WS. A close to zero Shannon index means that the diversity in rice.
area share distribution of the varieties is not even. The mean
Shannon index did not significantly change from the 1996– CONCLUSIONS
1997 to 2001–2002 period, although there is an observed The main contribution of this paper is to document and
slight decrease implying a lesser degree of equality in the describe, using three sets of nationwide farmer surveys, the
area share distribution of varieties in the farmers’ fields. recent adoption of rice varieties in farmer fields, especially the
The calculated spatial diversity indices show no clear later generation varieties. Data showed that although enduring
trend across periods except a relative increase in the coef- varieties released in the 1970s and 1980s such as IR64, IR74,
ficient of variation of the indices by season from 1997 to and IR42 and Burdagol are still planted in farmer fields, there
2002, and some indications of relatively increasing richness is also evidence of the fast adoption of selected newly released
and lower dominance especially during the DS where farm- varieties such as PSB Rc18, PSB Rc10, PSB Rc28, Masipag, PSB
ers are more adventurous in trying more varieties. Spatial Rc14, and PSB Rc82. A progression in the use of MVs from the
diversity also varies across provinces in terms of richness, early to the later generation is also evident from the decreasing
relative abundance, and evenness. These results can be used trend in the use of MV2s and the increasing trend in the use of
in further understanding the concern on genetic diversity as MV4s. This information can be used to encourage policymak-
a factor to vulnerability to pests and diseases. We assume, ers to continually support and strengthen the current efforts
however, that farmers choose varieties based on their of public rice breeding research and extension. The higher
observable traits or genetic expression rather than on genetic adoption rate in irrigated areas and during the DS also points
composition at a molecular level, which they cannot see to the importance of sufficient irrigation and less risk of pests
(Smale et al., 2003; Rubenstein et al., 2005). It is important and diseases in the early adoption of new varieties. In addition,