Posted November 9, 2004, Chapingo, Mexico.
Dr. Roberto Garcia opens the trunk of his car and reveals
piles of cloth bags of bean seed, labeled with such names
as Atimegaya, Hogar, and Huesonapan. “This is how I
get the beans from my program to the farmers. I am the extension
system as well as the researcher,” says Garcia. This
is surprising, considering the impressive data he has shown
me on the yields of the disease resistant bean cultivars (field
beans, Phaseolus vulgaris) from his 10-year old breeding
program. Most successful crop breeding programs like his have
an extension system, or at least a full time employee, for
getting seed out to farmers.
Garcia, professor at the Colegio de Postgraduados in Chapingo,
outside of Mexico City, has no extension system support and
no funding for seed dissemination (or for other activities
fundamental to a breeding program) because he sits, despite
successful results, on the wrong side of an unnecessary chasm
that divides two different approaches to plant breeding.
In a society that calls itself democratic, open, and merit-based,
it is always surprising, even shocking, to hear about a situation
in which a proven technology or a solution to one of society’s
problems has been ignored or even suppressed because it didn’t
fit into the existing philosophy. Of course, those of us in
the organic farming and food movement are all too familiar
with such scenarios, but when I hear about it happening, as
a scientist, it still pushes me hard.
Garcia’s development of high-yielding bean varieties
that need no inputs of pesticides has been accomplished
using an approach to plant breeding that literally turns the
green revolution approach on its head. His approach is called
horizontal or quantitative resistance breeding. (Terminology
note: The mainstream plant breeding community uses the term
quantitative when referring to this aspect of plant genetics.
However, as a gesture of respect to those scientists who have
used the term horizontal resistance breeding and who have
persevered in establishing it as an indispensable approach
to crop improvement, I am keeping the term horizontal.)
Horizontal resistance breeding can be of enormous value to
the North American organic and sustainable agriculture community,
and could prove to be the basis of the development of many,
if not most organic crop lines in the future.
learn more about horizontal resistance
Robinson's classic text on plant breeding to reduce
pesticide use, Return to Resistance,
is out of print. In keeping with the principles
of participatory plant breeding and democratic
information sharing, however, Robinson has recently
made several of his works available through the
Canadian free e-books site, www.sharebooks.ca.
The books carry copyright protection but can be
downloaded for personal, non-commercial use.
Included on the site are Return to Resistance
in both English and Spanish versions; Self-Organising
Agro-Ecosystems, described as the technical
companion volume to Return to Resistance;
and the Amateur Potato Breeder's Manual.
Also available is the Amateur Plant Breeder's
Handbook 2004, an expanded, cross-linked
version of the glossary of technical terms included
in the original print edition of Return to
If you just can't give up the pleasures of books
in book form, copies of Robinson's books can also
be found through on-line second-hand book sites
On the other side of the chasm from the horizontal approach
to plant breeding sits (with all the money) what is known
as vertical or qualitative resistance, dominant for some 100
years over the crop breeding milieu. Vertical disease resistance
generally involves one or just a few genes, and the vertical
approach to breeding involves the transfer of one or a few
genes from plant to plant to bring about desired traits like
This approach to crop development is the basis for the genetic engineering
industry, and is heavily invested in a proprietary orientation
to crop development (ownership of genes and cultivars).
Garcia has seen his funding pulled by both international
and Mexican funding sources, despite developing bean cultivars,
via his horizontal resistance breeding program, that have
increased yields by 2-5 times that of the existing cultivars
in the Mixteca bean growing region of Mexico. The funding
cuts came despite the fact that Garcia not only eliminated
the need for pesticides, but showing that his new cultivars
consistently yield on par or better than the newest “improved”
green revolution varieties, all of which need pesticides.
The Mexican agricultural ministry and the international funding
agencies are all staffed by mainstream plant breeders, invested
in the vertical resistance paradigm, according to Garcia.
“They simply refuse to believe that horizontal resistance
breeding has any value,” says Garcia.
Garcia’s disease resistant bean data, showing yields
of 1,500 to 2,400 kg of beans per hectare, with no pesticides,
vs. former yields of 350 to 600 kg per hectare, are not just
one year’s data. These are consistent yields from replicated
trials, year after year for the past half dozen years. And
the yields of the horizontal resistant cultivars are still
climbing with each breeding cycle of hand cross-pollinations
A separate but related horizontal resistance breeding program
at the University of Chapingo, under the direction of Dr.
Luis Serrano, has achieved the similar levels of bean yields,
also showing superior yields to the green revolution varieties.
History of the chasm between vertical
and horizontal resistance breeding
The history of the chasm between horizontal resistance breeders
and the dominant alternative, vertical resistance breeding,
goes back to the first decade of the 20th century, to the discovery
of the neglected work of Gregor Mendel. Mendel had discovered
that single genes could be inherited as recessive or dominant
alleles as part of homologous chromosomes, forming the basis
for single gene selection and breeding of organisms for traits
based on selected genes. The transfer into crop cultivars of
genes with characteristics such as disease resistance or short
stature became the focus of crop breeding programs.
The problem is that a majority of the time that a gene for
disease resistance was bred into a crop, the disease pathogen
would develop new races that were virulent against the newly
bred “resistant” cultivar. The breakdown of resistance
based on single genes has been a continual problem in crops
since vertical resistance breeding began. In wheat, resistance
to leaf rust, stem rust, stripe rust (Puccinia spp.)
and scab (Fusarium) have broken down one after another.
The most recent breakdown of stem rust in wheat was announced
in 1999 when a virulent biotype developed on wheat in Uganda,
causing concern for wheat crops around the world.
The traits that were of interest to the Mendelian disease
resistance breeders were generally for genes with total resistance
to disease. However, it was known, even before the discovery
of Mendel’s work, that there was a quantitative aspect
to disease resistance in plants, and that within populations
there could be found plants with anything from zero to near
total resistance to a disease and everything in between. This
quantitative type of disease resistance was not as spectacular
as single gene resistance and receded into the shadows of
In 1963 a South African plant pathologist, J.E. Vanderplank,
elucidated these two types of disease resistance, vertical
and horizontal, in his classic book Plant Diseases: Epidemics
and Control. The terms vertical and horizontal were based
on graphical depiction of the two phenomena (Fig. 1).
The vertical resistance system is actually a system of matching
co-evolved genes between the plant and the pathogen. Horizontal
resistance is polygenic; many genes are involved in the resistance
to the disease, thus the terms vertical/qualitative and horizontal/quantitative.
Both vertical and horizontal disease (and insect) resistance
are found as integrated mechanisms for defense within a plant
population in the wild – they are strategies evolved by
the plant species to stabilize its relationship with pathogens.
Dr. Raoul Robinson, British/Canadian plant pathologist and
author of Return to Resistance, probably the staunchest
defender and strongest promoter of horizontal/quantitative
resistance breeding of crops, uses various metaphors to describe
vertical and horizontal resistance. In my own adaptation of
his description, vertical resistance can be likened to a door
lock on a house, where the pathogen has to have the key to
enter. But the house has other doors through which other races
of the pathogen may enter, if they find the plant, in which
case vertical resistance breaks down.
Horizontal/quantitative resistance is akin to having unlocked
doors on the house, with soldiers in the house to fight off
the pathogens when they enter. It is characterized by many
genes which can fight off any local races of pathogen. Horizontal
resistance, according to Robinson, does not break down, because
there is no lock and key gene-for-gene mechanism as with vertical
resistance. It can vary from very low levels of resistance
to nearly complete, and thus is known as quantitative resistance.
When vertical resistance is isolated by taking a narrow genetic
slice of a spectrum of genes from a wild population, and then
bred into a crop, as is done in mainstream resistance breeding,
the resistance, in isolation, cannot be sustained, and in
most cases eventually breaks down. Potato illustrates another
case. Early in the 20th century potatoes were bred almost
entirely for vertical resistance to late blight, the devastating
fungus that hit Ireland in the 1840s. It was found however,
that it took the late blight fungus about four years to overcome
the vertical resistance while it took eight years to breed
the vertical resistance. After that, resistance breeding was
largely given up and control of late blight brought about
by copper compounds.
Dr. John Niederhauser, working in Toluca, Mexico, discovered
that polygenic, quantitative (horizontal) resistance was much
more successful and sustainable against late blight of potato.
In 1991 he was awarded the World Food Prize for his pioneering
The chasm between horizontal and vertical breeding is not
all black and white, and many crop breeding programs incorporate
quantitative resistance into their regimens. However, most
are based substantially on gene transfer and vertical resistance,
while horizontal/quantitative resistance remains a minor player.
Much of the alienation of the Raoul Robinson-trained horizontal
resistance breeders from the mainstream may be due to Robinson´s
approach of eliminating vertical resistance from
crop lines in resistance breeding programs (described below
and outlined in his books Return to Resistance and
in Self-Organizing Agroecosystems).
Robinsons’s practice of completely eliminating vertical
resistance from parental lines appears to have merit, however.
Vanderplank discussed in his book--and Niederhauser further
showed--how in potato, vertical resistance to late blight
could, over many breeding generations, cause a near complete
loss of horizontal resistance to late blight, just as happens
when fungicides are consistently applied over many sexual
generations. The plant, minus any pressure from late blight,
in a sense says “Why do the work to produce these (horizontal
resistance) defense mechanisms when there is no late blight
attacking me?” When the vertical resistance broke down
in this particular variety of potato, called Vertifolia, late
blight completely and utterly devastated it, as it now had
neither vertical nor horizontal resistance. This Vanderplank
called the Vertifolia Effect. It is for this reason that the
Robinson school of plant breeding first eliminates vertical
resistance from parental lines of crops, as they don’t
want any masking of horizontal resistance.
How scientists breed for horizontal/quantitative
The methods used for breeding for horizontal resistance are
such a radical departure from conventional plant breeding
that the reader is likely to do a double-take when introduced
to them. They literally turn the classical plant breeding
paradigm on its head. (Note: Not all crops are conducive
to horizontal resistance breeding, a subject that Robinson
discusses in his book Return to Resistance.)
After collecting as wide a spectrum of crop germplasm as
possible, the collected lines are planted, and inoculated
with or exposed to two or three of the worst diseases or insect
pests for that particular crop. The next step is the controversial
one. The surviving plants (the disease and pest inoculations
should be designed to be sub-lethal) are then selected for
on the basis of susceptibility. This may be completely
counterintuitive, but it is one of the most important steps
in horizontal resistance breeding. The objective is to completely
eliminate any vertical resistance, which is easy to identify
because the plants will be completely free of the disease
In Garcia’s breeding program in Mexico, he selected
71 of the best bean cultivars from the Mixteca region, exposed
them to three of the worst bean diseases - bean common mosaic
virus (BCMV), the common blight bacterium (Xanthomonas),
and the fungus Macrophomina, a soil-borne root pathogen.
He then selected seven of the most diseased lines
out of the 71. The seven lines were so diseased that some
of them had just one or two pods.
When Drs. Garcia and Robinson were working together at the
susceptibility selection stage in the first year of project,
they visited the selection site with the field operations
manager. Diseased plants were everywhere.
“What a mess!” said the field ops guy. “On
the contrary, it’s a beautiful sight!” countered
After the initial stage of selection for susceptibility,
every subsequent generation is subject to selection for resistance,
measured by yield, since disease pressure is intense due to
the pathogen inoculations. The common bean is 95% self-pollinated.
Therefore, starting with the seven selected susceptible lines,
hand cross-pollination was carried out to cross each of the
selected lines with the other, an extraordinarily laborious
task. The progeny lines (there are some 49 families, including
self-pollinations) are grown out and subjected to the same
disease screening, this time for resistance. Only 1% to 10%
of plants are selected (each family has hundreds of plants,
and there is high variability between the plants within those
families.) This is known as recurrent mass selection, and
the resistance selected for is entirely horizontal/quantitative.
With each subsequent generation, all selected lines are crossbred
with each other to generate new lines, exposed to the pathogen
screening, and 1-10% of plants selected.
On the average there is a 20% increase in yield from the
lines selected in each subsequent breeding cycle (Fig. 2).
The breeding cycle often takes more than one year, since it
involves crosses which yield just a few seeds. Those seeds
need to be increased for one or two generations to obtain
enough seed for another round of selection.
After the 3rd breeding cycle, the selected cultivars are
taken from the university experiment station to the bean growing
region, where dozens of bean diseases and insect pests are
endemic, and are planted in the most disease and insect pest
prone areas that can be found. Therefore, according to Garcia,
selection from the 4th cycle onwards was not only for resistance
to the inoculated diseases, but also for other pests and diseases
as well. This is the great strength of horizontal resistance
breeding. Resistance to all locally active pests
and diseases is selected for after the 3rd breeding cycle.
Since at least half of the generations occur in the target
location of the crop, horizontally resistant crop lines are
generally regional and local in their resistance
characteristics. Transfer to another region which has a different
complex of pests and diseases will give variable results,
depending on how much the pests and diseases vary from the
original site of selection. Thus, horizontal breeding programs
have a local and regional focus. This is a distinct difference
from the green revolution approach, which attempts to develop
crops that have resistance to a few major diseases and pests
for widespread inter-regional use.
Unplanned but later built into Garcia’s breeding program
was horizontal resistance to soil borne diseases – the
classic heavy hitters of so many crops – Fusarium,
Pythium, and Rhizoctonia. These pathogens
hit the research plots during the 4th breeding cycle, nearly
wiping out the entire experiment. Garcia then went on to select
for resistance to the soil borne diseases, in addition to
BCMV and blight. In subsequent years, these pathogens were
inoculated to screen for resistance. Garcia now has bean lines
that are completely resistant to these soil-borne diseases.
Garcia loses funding because he doesn’t
fit into the dominant paradigm
When Garcia’s project, which was funded by the International
Development Research Center of Canada, came up for review,
the referees who were sent to review his program were all
from the CIAT (Centro Internacional de Agricultural Tropical)
bean program, a green revolution center. According to Garcia,
the scientist referees, who had spent their entire careers
developing vertical resistance and gene-transfer type breeding
programs, were threatened by his success in breeding high
yielding varieties that competed with the green revolution
varieties. Either that or, for inability to perceive
the benefits of horizontal resistance breeding, they had the
program funding ended.
According to Garcia, the basic premise of the vertical resistance
breeders’ objections to the horizontal resistance approach
is that it is characterized by “Garbage in, garbage
out”. They maintain that you simply cannot base a breeding
program on crop lines whose original parents were selected
for on the basis of susceptibility to disease. It
is simply antithetical to them.
This is a perfect example of how philosophical blinders work
in science, as the data shows unequivocally it is
possible to develop resistance from such parental lines.
Garcia counters this criticism: “We didn’t start
with garbage. We started with 71 of the best varieties, out
of hundreds, in the Mixteca region. The seven cultivars we
selected in the susceptibility stage were susceptible to diseases,
but they had good genes.”
When Garcia put together an international conference on bean
pathology, his approach to bean improvement, despite data
that showed success, was largely ignored. “The mainstream
breeders had such an arrogant attitude,” says Garcia,
“they declined to even to take a short field trip to
look at my trials.”
Dr. Luis Serrano of the University of Chapingo, whose bean
horizontal resistance breeding program has achieved very similar
results as those of Garcia at Colegio de Postgraduados, has
the same things to say about the chasm between his program
and the mainstream breeders. His funding was ended by the
Mexican government, whose agronomists and geneticists are
all mainstream, green revolution types. Serrano now manages
his program with his own salary and with the scores of students
that come through his classes. The plant breeding and agronomy
courses at the university involve dozens of hours of field
and fieldhouse work by students to manage the breeding program.
Serrano showed me boxes of envelopes of seeds, 400 bean lines
in each of six types of beans – black, pink,
yellow, tan, white and mottled. He is working on horizontal
resistance to the bean’s worst pest in Latin America,
the Mexican bean beetle. He also showed me data in which his
horizontally resistant bean cultivars are compared in replicated
trials to the best green revolution and commercial cultivars.
His horizontally resistant cultivars came out well above the
others, with yields of 2,000 to 2,400 kg per hectare compared
to 1,500 for the commercial and green revolution varieties.
The CIAT bean program web page shows that yields of their
improved varieties in commercial plantings in Mexico and Central
America range from 800 to 1,500 kg per ha, with low to moderate
levels of pesticide inputs.
The entire agricultural extension system in Mexico is, for
the most part, set up to promote green revolution and capital
intensive type approaches to agriculture. The bean lines that
Garcia and Serrano have developed do not interest them.
The inability to patent and make proprietary (to own) the
genes and cultivars in a horizontal breeding system, which
is polygenic, is likely another reason for the rejection of
horizontal resistance breeding by mainstream agriculture.
The entire basis of the genetic engineering industry is the
ownership and ability to transfer of one or a very few genes
from crop to crop and cultivar to cultivar.
Additionally, programs which develop crop lines that don’t
need agrichemicals are not going to get much support in an
agricultural system that is heavily invested in the use of
Other crops are currently the target of horizontal disease
and pest resistance breeding at the University of Chapingo
– mainly onions and, just at its inception, tomatoes.
Niederhauser’s former program for potato horizontal
resistance apparently continues in Toluca, southwest of Mexico
City, and has spawned at least one potato breeding program
in the US at Michigan State.
Bean breeders from international centers and universities
who I contacted had not heard of the Mexican bean horizontal
resistance breeding program. At least one had heard of the
Raoul Robinson horizontal resistance approach, but had only
cursory knowledge of it.
A student plant breeding club to the rescue
Traditionally, beans are a central part of the diet in Mexico
and Central America. Along with corn they form the main food
source of the poor. In the category of basic grains, beans
are second only to corn in area planted, and are the number
one source of farm income in Mexico. The market for beans
has grown substantially in recent decades as urban populations
swell with migrants from the countryside.
Nutritionists consider beans to be an almost perfect food –
high in complex carbohydrates and protein. A single serving
provides at least half the USDA´s recommended daily allowance
of folic acid (very important for pregnant women), 25-30% of
recommended iron, as well as magnesium, copper, potassium and
zinc. Bean nutritional qualities make them particularly beneficial
in the diets of women and children.
With the help of Raoul Robinson, Serrano has developed a
rather innovative way of getting his bean lines out into the
field all over the country – the student breeding club.
Students from farms all over the country come to Serrano’s
program, participate in the development of bean cultivars,
and when they go home, take many of the cultivars with them,
as well as bring local cultivars back to Chapingo for putting
into the screening and breeding program.
“Our bean lines have been planted in a dozen states
all over Mexico via what is now officially called the Raoul
Robinson Breeding Club, and we have collected a very large
assortment of bean lines via the students,” says Serrano.
“It’s not a perfect extension system, because
many times we don’t hear back from the students once
they have gone with our beans, but it works at our level of
“Plus, we are putting out hundreds of students who
know the value of horizontal resistance breeding, not just
from our telling them, but from firsthand experience. This
is good for the future of horizontal resistance breeding.”