, i.l., e ..... .. -19:- ... .1 I ..... .. 1 ..... ,
-:. :.: .., .. ... :: ..... .. ; 1. .? ,.- .. .Ii:iv .
...:.. ,.;: A:.. AH, K,-6-z ,,.Y, ;b .i ., ,.,
.. ..: .. .. -.,!.:. e;j, M ,., iKZ.j.
. .. .- .: ZA VKI..
N : .., iw, ... .N.,
..... i ,,iL I; .:.. ... 11.
:,;, I '... 1. .. ,, P, Y.w '..".1
.: .:: 1 V.- '.1 tk.,A ,
ie ... : .e;j .- 1- 1- l-, .1 I.. .. All w ........ W-- .. ... n. ;.. v i ,..,
: ,,::. :.,:, ... .. .. % ,., i% .. A % -,= : ; fl-. J.,.: .,.,, A PI, .
I V ... K H.".....'.".. ;
... v-:C 111-,* .. k.. .. g ,
,--, ... .-,,... ... ,- yii, -:.4.-::. ,-" .. 'm
,. e ,.. t ,., .. F jj
... ... I 1, -:v.17,1,.K. ... .;H.. .. 1, ... ., ;- I ., .,. I 11 -
1, I iiPl?:-.., ,A,. i, : A;A.J
.. ...,. ,A,_. ,.t. ,. ..: ... .... M 41, Q ..:
..... .. .. i% .::I._ .. ... ., HA
.1 ..m ,..:..,, iq .. .i..: ..:. I I ,. .,k .. ..
A" 'i% .1 %A...] ,... .1 I I 1. .. .., L,:,:
,:...:..... :.: ..... IZ A ....:.:.::..
.- i H, .,:. ,-..: .....
: :11,,%:_,,:k ,?ElxA.l "I 'A I . ... i,. ..:i. .... I.. 4.6 ...,...... ,,I,
1 ::- ..;:i... .,A:4 :: .. .,.
t. I r,, 1 ,. :. .... .....Mi... ... ` '. ,13,i, "".. -
I.- .. :x -.:... .i : :.-,.''i. ,
..... n..Z. .:: H.,.;M ..e..,, .,,-%,
.: 'A ... I ... .. 0 1 !,: V ,
;.. ,.. 3 ,... -:. .3. ..
... ,- ..:: 1:
.. ..., .......
.. .... ,,, .::: -e. %. I .. ...... R K::: 1 1.1-4- ,....l..,.. ..
,;z ,,.:, :i- ,,l .:.., .. M .
-1 ,:, ... .." ......... ::: -A::,v WA-1 ......-,.
.::T.I.- .... -::::i...
11 A ., ::, I ,,J. .., ...,. .. ,, ARE 1 ,:,:.., ,3,..,,., ,,:il ... ..
", I ; :. c-6:,, .. 14.1
.1 1 .. .. : ::
:g. .... .. V .." .. ,J. ... ,. .. .:w .. ."
I II ,:.,. .- I ..: .. .1 ... .;,.. 11,,:.::-,,i,,A:*
", ,., .., ... :. ol .1 ii, ... .1. k :,
.1 ., rm., ... .. I el.k.. ::q., ,: lk
,,, ,: 11 .. HE. .,,. l.k., ,,% :!: ... .1- I .:..,,:y N,,:, "A.,. :,!,,,!, I
.. 11 .. I ,:... "" ...... i.,.. ... ". V.. ,
... ... 1 .!.. i ,- .,
i: M R ...,.. '.. i,, ,..O .11
", .:: "" I ..... &.
", ... R A, ,- ., ., .,.A %;..... ....... U I .11 -,; ,
...... .. .I.. :: -,
...... ... .. .. .... -, I -:. .!. -, m -.,oti,I,,,:
-:,: ...... -- : ,,:. :-,.,,
...... ,- ...... ..., :..;";" ,.,,...,.:::i,,',.-A ... ..
,-1- ::....:. : :: : :, .. ,,. ,Z: ". -1,
,. : .. ., .,..,.. ..
... .. : . :
.. I : ..... k :,..,,,i,..:;.,..... ..
.. ... .. "El. -,.,.,,,a:! *- -I,.:,, ..-
:..,:,. .... 1:::ff .. ... 1".. .. ,
: ..:.., . . I -, ..::. :
.. I .. ....-. ... .. .. .. ....'. .. .. ., ,
... ... ....., I.... !.. ..... .,.:. .. w ;. .... Zp-ii- ,:
... ., -,< ,:.,::: .:A,:.
,.,,eix-l f '. .
,.: .. "vik .k
Ale,..?:-. -1:0L.:l;,.:::,i. .. ..:: i. 3 I ( I m -:-, I!- ,x,Xn, .,, .i.r
".. ... .. i N ,L.n, ....
.. W ..
... .,.. ,;,;:..-",.?[,,:..e :!! -. -1- & .. ... I ,- ...,V
:;A..,,:o:P. ..R- ,, I ,,.
...,.. .M. I-ZkAo, -
,2 tim A ... M 1 ,
,kflel, ,,,,Ie ,.- ', .. ...... ,- .A .,,,.S:r. ,:--,X.:?Z,,,AV iu,: .g ie ,!A i -:,:; .*
,- ,,.:, ,4 ,,h k x .li ..,. ., t, ,. ..
.4,.:, _,..,;...., xl. ,.:: I
I ::,, --e- ,- .. a I 11: ; m ..... ,..!i!
"A :,,.-, I" ,11 i.;, ,, ,,v % .4 A
-, ,Ik ... ,,, ..,: ... e a P m -j ,% t 4, ...
,x..& .k ,::.. ..,:..... I .. I .... .... -,: I ;
., v16 ,, pA,,,,::, ,,:: I .,,..% i,:,, .. I -,.. v T 61K.V.. ,,
... V I ... i ,..
.. ... %!,i: 'I'l-
...... li. v ...,N,!. ,., ... ..
I I .. .... '.. .
I.... I ,eL-..., n.. .. Z. :., N
.... ,:.... 4 .i ,6?., ;,,:- ,
... I ..... ., .. .... ,- i:Z,
;..:..: ., ., '. ,,. ., .
'I" IA- ,"I"....."k4l., : ,,iiA -:. ,0 :A!:,., P....;O,, ON &,Z-,,, .,
N %, ,:.;.tl -,,... ..., 7..
... .. .- -: ,,m .
... ..... .. W, ,: .... i:M
". lx ,,,,g ,,- .1 ........ -.... .,....,,A .l .... ... ,, ..... .,..:, I.- .". ,-- I.-
..: .... ,,.- .,., ..:: ,;.,. % I
', :. ? K :1, M.
.. I .. .. ... ... .. ... ; ':.:,., .
.. ,;: rA", -- ,,,, ...,.
,.,,.,..4....,. .;. ,,-..- ,;,z ,..: .?,. I li,, I 1 I % .., 'A ... I 111...,., "..
., ... ?, : .., ;A,;..,-, .. .. .. ...
...1- ., : `, i At: i:,ff. % .. Y. .: ... : ... lv I
,. .. 4 ,., .. ...... .. .. ..,,..,
.., .. .... ,..., .. ... 1 ,... Im ;
.; -i :.,. g ,,j,,.. :.: Z H::;
AW N. 2a ... M .,,o ..:-,i:T "L kM .
., .. ,,::, ,:A. ,; ... .,!.OZik ...
.. ; ...... '. .
I.? ,i ,.,. A,! ,:- e
... ?Q ,-.:','..:Ai..,:""j.:
-:...... ...... K, v,...,v;::,,,A., .... % ..
,. I ."A .; n.ll : N O -:.." ..?c ....
Lx I --f., :. !. -Q. ,- '... ..... .. .. n :-.,,,,,.:., i;!..,;-l.:,...:,_ A', ,: : ,W,,. ".:";!!.."..: !. .q..,Z,i.r..
", .. ... -.l,..-: I- -v. --l" .3 1 ..
:.., ... Y. F !.,
'' ..::,, jr Z;zl ... z,,ZiZ. : -. ,.Z..
-, lr.,,. I .- .!V.. -;Z- n.14ik..
,,,,...,:...,,:,.,,:,,, ..,. Z:,.9 ,,,&- ..-, :-......e i K.v ... -,
.., .1 .. .... ..... ... I .... :.: .. .,
I.:r i: .,: ...,: ,r. .:.3.....l..l,. ,,,.4,,,,., ,.-,? i L:: M rk ,Z, 'n. .
V. `1 ,-.- '... ..- .. ..... : im
,M, .4 .,R X -- k I.. A RA.. :...: ::
il,,:-X ...V, ",',', .11=1:11 .::;:- -. n.
.. ilr,. :E I 1.
.. I... % ...:-E.... ..". ., ,:. ... .- ,
.. .A- ..- ... .. -. 4"'...
-!,q .. .. ,.-i:kvZc,..Il ...
:.. : l,",.'.. ,- : .: ., :.- A.., i.., t-.X,,, ? -- ..: -.- ;. 4: .
I i i, 1 ... -
I ,,.,.!, .1; ... .. .... ...
.1 I, .,'l.'.,,,'.:,:..,.:: .111. .1 1. %': ; ,-,. ,-:::
v ,: NM ... ... .. ...... % ,
..... ."., I ... .. ... ,iF n ,... ....
.. ... g .,,3.. .,.. el, ........
.. ... .. .. .r.. > -K,".."
.,,i%p l..;. ,,,. .,; :. : ni. F '... : I
-.Iv,- e:..,,,,, .... A i- -,;,,,.J 11 .;...: X A,:Z'. ,:. ,.!>.
.:% 'I i: ,.,. -, ,::.. i;. !, .p,:,,l., 1. IF, ., .1
,. 11. A ,..,..... 1 :.., ...; .
iv I :,,,5...,,ir ...'..". J, ,.
: ... .n ."
..- .. a..,Yj t- :!, :. ,Y,,-?,, , ,,.. 4 I.A 71 '. ,
"'o .3 ...,
:... Z .... ..
,: .,.:; mi::Ti- .'. ... .. .....-
...... %W :.-,.:.-i .... .. Z.,
Z:.4, ,A., ... ; :.-..;. ,
.;. ,Z .. .., ,.: iA. :, ,e,. ,i. ,l%!,F, .E-..i .. -. Z,;. .., ,,.:,,
7,i .. -:. .::i.::.L,- ...... ..
..... ..."..." .S.-.fl,,! --, .-, ............. .... .. ... -,,n -....-,.,.. iz....... e. ..... .. ... .. .
A&.:,. 2, I :..,., %.4 4Q:e: &I ,., : ....k., .. :"
::- .,l -n:-.C _Zv x ., I @ ..
.,.;.'.. ,.". 1 e`eF:- e Rx 51, % ,iZ, ..... 1& 21N,,.
s.,Zv .::v ::,.;:..l .... ., 4... ok.. A A '... -.& ,j lz -Ai,..,.. ,.. I. .-I.v
.. -....... ,..i:q:,,,,:,.,.w ,O kel ... .,,3:,%..,: .... .-W.- .
I.., .. ..... ... ; A %N.. ,r,. k.. :!. ,. .
.. ...... .1. ?,, ... ,U ,i!:. .... ." ;.: ...:
.. ...... .,;.,:.,...., I -- .. ,.l. .,.. .
ii.......,_,-i -. -.,. %,,,.. ...., .,.;.... : ..
....,, ,:iv..:.U...:. V. ;Z ",A.,.., ... U '..q:. ,? '.. .:.,:..vl.-- ... -X ,-Z: ..... k,:;.: ,.: .:i:... .
4. ,,I.iv:.e ,.Z ..... .. -'..- Jv ... ,. .. ..,..-,,.. iv ,,,-,., A .... ". :.:::.,i,,,,.,,;:-
... :Iij: .. v .., ,i.::::.',:'.R::k, .. ,,,,.:.;,! ,Q. .:
..., -, ...,_ q
A x ,ii ,- .. ,::. VAI .. ... ..., ... Z,.,..,.,. ,
-.,il ., I ..... I --.- 1 I ... ..., ,%,:L;; %.. ..,:I:i-,-,.:
:i :--. .1 ,i,,..,,j
:1:1. %. '..- % 14."
.. e, -,:....- ,....,
,.; 4'. "'. rvii. ":;, -,.- :V- k. -. I .. .... It.."'.....,"..,.... ..
.i: , I I ,-
m ... Z....:.,.... ,-
I.. k ..,,
.... 1A .
:il......................,.,?4. .'-,: I-NiZ. .... <.F, .'. ZZi,- ::.... 1- .
... x. I w i;', ..,..:,.,:,.,..A- ,: '....
.. ,I Z ;;:Z:: ,. ni% :,A.. ., ,...,.,-,;%` -... I.I.... ., ", ...... ., .....
,:,... ,.,:;,..... ,.- ,.,, .. U.
,.l.. :, I ..
: A ,.:.,,,,...:. ,.. ,. ,Zg,,I : ,, xk &, `."Z
.,i:!:K.. .. X.. .. .... -- .....-"T ... ,- ... a -1-I.. IU :,, w W W, --.
:,"Ij Z&,,e ..,.., .. v ., ,. w..,. .i, .... ZZ:,- .;. ;
.. .... : X., .. ., ..g.:.,.
... ,.., k : q-12
z .1 j ,. %:& m 3A.., ,i, K" 41AV v-, ..%
..: g,-: 4, ,v.i "I., ..i.,... .. .... I..... 't ,% Z..b..,i,.. K.A., ...
.:: ,:... Z- :: 1,; I .-,... I ,,--<,,:Z-z .. .,%,i ......,
1. ,%,.- 1, ::i eur ` ... ..... ..,...,.-,. ,
....... :. 1: V W ... ,.. Z. M Z w.,.. ,wl $.,: 31. ,
.,.. ': ...... ..:...., ,. -:- .- .. ;
.e .:i.".. .. ,_4 z .6 I..,, .1 :m -%: Vi.!, !."L...,
..,k w,:.,X*;:, v:,Fji%.;F N-Hh- -..]:< I.. '...., ....... ;. ..... .. ....... ... y- .. .. ..%,;.-., .... ,. t. .:, H ,jgiZ i, :, ii"h.:!,;:,%I,,,,:...,..., ,:.
,g. A ,:,..:.. z.::. .e .. '. ..
V ..I, .. k I.Zll: I : ..., : "...-.- I Ziv..l. k A A U.. x .. ,.
.. "; "i-,", k.--,F,,,-,, :ie4k.5 I 1. M ,., .1 ,
41 '... ,-F, Pk,..: I.'
;:. ; N .1.11., -, ... J w a ,,N ,.:...2
.. ..... A :,. ""n- I., .., .. g. .. ,
.:... ... N,.: V i O', ".. .. -,, .,& :: : H ,.:J-,:,
.1 ...... ..,.....; .. %..:.,. ,,;, ,- ,uizrNz. g..;v ,,i. .. .1 jiw ..,V- i :: ?A,.-..,, ...
..i. ,.A,:,, .,3:i. .,.,;., ` . e ...,
:1 1 iigk.: e Rl.,,!, v 5 ,,M i!,!;
1: VAl.. ::::,:.Y;,, .....*, ,e Z r.,.,t i!: .. .::;:, :.., ... g .%, .... .. ,. .,
.:.::. .....'.11 .,: .1
,:,Zt% ,;Z, ,: I .. ..., %.- .. ..... I'll ... ... . I... ,"
qw, ,?%.: ,;: % ,
ZI., 1 9 ,;, '... .. m Z,,,:-,,:,,..
.: .. 4 ,N i,-,.. -.....,.- ,
... 1. ,,-,,. ,,i;;w ?!, Nl
., .. .;i i ,, .
-1. ,.:v..,;. ,i...:- .:, :,,- ... 1. ,,:l,:i ; Z:..,A -, k .1 "' ine.,i.,, : : m, A :,
C,:.,- .:-.- 3; .m ., ,i.% %. ..
.. -, P,4 N, ,,7 i% -.-: ,
Ki i ... % .,.. z -A .",?:?i- wll., :7,111, ", :i:.: I Zl *.,..Z,,,
,:.- -, ZZ, : W. ,:.e,..-
: .. ,.., ., ... .... .
.. .;.,.%.,;,::..:...:P.,,....". : .. ...: : ..
21, ... ZZ,:: ,;;
.. ,. .... ...... h; ,:ic.,:%c-
.. .1 ,,,? ...... .. .`.. Z %k.. ie :, :: *,l,lW,, V ,.,::
.. ...; ..-..?,,n.. %:E ... .1 M Ui I '
.1 :,, V, ,iE,,%k ,, 1. "I .
-,A,..-, "A I I ... ... -. 10. :1. I .- I -!,.., M ..!, ,.....,.-u "O.
.: :I ik:.%... Z&.: : : "
ll: "... -1
% ,? lji -Amii ,......x ..
.. Z.,. ,.. %. .. Tif, ;.;,F- ..
,-,, 4 't.
11,,.. ..,:, i, .; w
1 .. I N-`-, 'Zo : iZ .. ,., w e
... ... .. I .:, r. ..... z- -- 6; -:: .. -, < k. ... F ... .. L ,,
,. .... M .A -,:. ,:: .... :: ,:!, 4.::. ZI.: .
.,..... % ;: I .. ..:..: ll.:,:, I K: 1,
I I I ".- ,
....... .. I I MI. :;... .;.:, K.::: 1. ..., .,.,. ...%. ,,,
I 11 -. I "I'll" :" s : : I ..
... .. :K j v v ..: ...
... ,. .. I 11 I . H : e.
/..,. .. :, ,.: .. :6j4 :
". I 'I' J .... J ..: 4, ,
..... ... .. .1 : 1 ... ... :. ... ,..:."
.. ,. 1. .. .
... H ., .... .....
.. : :i;,: e,-:,;,
.:: : .1- 1 ., .... -,. i;--,-: Zi, >,: .
I. .1. ..xi.,.r ::N I.- I Z.: ..:::
.,,l;Z h:,.-: ivZ !- .1 1-0
-; I : K:: lv: '..".'e.... .. In H. ,,,-,.Xg.,,p.:Z-,, ...:
I I ,:,`,,,,. Za -..
.. ... Zi,, ..... .. -... .e l ....- .. .... Z ,..
41 .,L .,.h. .K:F%,,,, .. ,,91. .." .,i., ,.;:,,D,. k ,. ..
... .,:. .,.:.. ..,
C. w ms E.. i :,
ZZi.. .. .r .. : .,i .%-
I ,r? .. .. W .. ... I .;.... ,
,.,:...,. .,m ..e
.. .:. .. .... I .... K iil m A .-i.,oik
?ii ., -. ,%w-'. ,.... 1, ., q- w,4 ZX Q .....
... .-... % Zj. .. nil P-V-
.. :.- p ,.,:. ,Iii .1 iik>..K
:: .:.,;.,, .1 ZC .. V ...i.AQ
e Vi...., .:-T- ... lZ` v
.r 'Ei A -?...V ,,m:,.. .
... ... ,.,: .1 .. a ,
..Z, .1 ... 3.,, -.,.,.,! ,Pll-., .,.. ,,.A,.,. ,
:,.. : ......... V : ... I ,s,,:.:t .... .. : P.;:.:.:.,:.,. W. "I
,,-,,.,t, "'.- ,., .,; 4. :: Y .1 .% ,m :, ,g ,...., ..,
.. .. .... .. .. :.,... ..; .... '. ,..,
... .. .... ... : :., ,.. J ;NK e e,.v-.... .. R W......m ik.:.. .1 ,
,: .. ...,... ,. .ki ,.i...
V:. : : ,:" 1. I I......, ... .. .
.. .,: :. -:.. 1. s .: ,: ". :, .. .:... ., I I 0- ..: .. :-,...
.. I'll .. .. k
I ,. A ,.,..l...,x ..... ; I ,.
I-- ,. ., ., .. i ... .1 ..
5: ... 9 .1, .. ., .. I :!..',":.: ..,... I V ...,,..A,..,. .. ek;
., : ... ,,... i . I
,,. I ... .1 ...... %Z .,.4
-Jvl ,F ..4 .. ..: x ,Ft < ...-.,,,: .. -
I.- .. ,..,. .. 51 ;,,,... 2
.. ,,i,.v,,. 1ll..,..,-. 11 .. 1. ,, 4,- %i R ,- ... -.1, ,, : .... N.t,
1% I I V.,. i -,- ..,,.
.. .1. .i. ,.: ... 1 ?, I ... KIpx c ,V- -,Ii
,,,t,:i...,,: ", ,,? .. .....
:,:.... .,::: ,,;,, ;,.. I ,.,.,... A:i., .,k- ..... .. 1. .. I ..
n ,,z,, .1 ., A,,,W. Z.- ..", ... ,k. 1. T.
".. ". .""B'", .1. 1, ... ,3:,::, .. ... -, .. ,,, ,,,-,,: ,,,7 F.; .V v, .: k .. ,; : j "
,:?.!f,,::x.,......,,..",..'I .,Z %.,. .., x I I i- N: .:.
... 6 1 ,
,:.,,,-.i.i. ,i,. :, .,,.,.-.,.. W Z% ;K, % -, -.!- ,,,* ,s,;p .i:r:.i; .
,. .1 I ,:,... ..,.:,,: .. p .,. --. ,:. ,H. .... ll..%..,,A
,.,% P !iiim -:,:,Jv ,Iq. .,_. >v I -ZH w vi:%
i,; : : ., --,..., .,.: .
%, ... ,. X. i;, .5:
t ,...._, nk i K x,., '... ,,,.,. % i% .
I ,,,.' -, "j. ... _:- K.:...:-
... !.,.._ % V ... 9;..,,,% ........ ,, .q
Or- ?ii % ". 'a .... .. .... ..:. I....
:,:; :,:,.;. ,:, .K.. F _.:: .... .Z. -%, ,? .,
al lZ%, C, 4Zt: ,. ,: t" ;
.,- ,,.,ve. 1,. il '., e ... ,,,.% ..,. ,V. ,i
,,-, -.!-.:.v i -<.,& W 1:0
1. fMne6,.;:: ... ... ..... .... ....
., .. ...... v< : aw -,.S,: % X. ...:,,.;&:i:j,.-..,-.... ........ 1.
.. .,W .....- 1. 41, .,:.:,.:,v. ,.: R w
B "' ... ., ...,.. .......
..,, ... ..., .... .... ... ...
..,.., .... -61... ,A, .. .. ..; U.. :< ; Z "; I'l, ;,:
.. .. I e A XI ... k g.
6 .. .. O.W." % ...- .,i.. --:.
:,.. .. ,,i W .
jq .,.,: -w ..... I t .....
.... ., ?, .... I ".. n j i ,,. ,. %I '.-.,w ".. .: ',,.,. .... .11.1. -,..-..-.j ,11 P
.%.. 7., ;F, N, :,,,%.Nl,. Q ,&n.. Q W
......:,.,.::";"'6::I RU%dP: ., .... .;
.,,,..A..:. .:. G .,.V.i?,,,,,-. .l. ..
.:.. ;'..". ...... .. .... ... 11 .... ,i::Q,,, ,:.i, ;. .::,.V.,
,.r,,Kt:,,,,,1,10 EvI %M -,, ;: I c., 6 A I ,: .% Y,.,on.. .. 11 -1, ,,-Ipli. 211 .
... 'A ,:".y:,,.,,.,.:..,.,;.,..,,
..: ... ... : .. ... .. .. .. v: ,
.... .... 1. I ..... ... .... I 4 ;.: Z. Z.
..... ., 5xn .,rem l t fi ii .il .. ;LvI...,
-li..,.L '... ..
.. ..,. ,. I.,
.L. ,.tKal,,: .r ,
1,1 :Ii ,.m .. ... .. I i. .
L..:,- % M .... V. V ', ,,,,%..,,.A -:31 x ... .g.
,.a,3, Z-- : .... -:., .K..... ZK..-- : :.. O..... e,,,,,,:,-::v ..
.... 1111; :::i, ,.!,.. .
I ..r e I.. mffl,,,& KA ... ..... .. y,
,*g, A 'IE Q .. Rog : ,., y -;--,,.
H 1 I'M 14 ... ... 1: ,e,It -n :. :..:: ,,_,,.,F,...,,: lli:,:.,:,
M W -& 114.,.,., ... .. ,
ll,x-,Z ,g,,g,-*.:,Z%!.... -,. I...".... ".. ...
,::: : gi:,: Z;,: :": gq.m.,.j... yg n, p :2.,V-WM .,:-,,. ,.,., .1 .,.,-...,N.. .. ...
... .. .A --..,,- -.: im .. ...... 1. .. .... ..U.,< .. 'N 1- .. ... ...... : .......
ll p $g .:.m ,!IvZA-I, ,.:.., :1. ,g..., ".1-1. 1 1.- ... .
.. -... ,-, ......%-wZ, M. i q -;,i,::',
.... Zw .... .x .:,,:-- ,, -Z2, ,.. ::,:61 i: ..
... ....... .... ... .:2 ,:..... .. ... ;,..- ::.: ..... ..
: .. I.... ...,. ". "
........ Zn.., -.."..'..., ;;;.,
1-4--l,.- %i.M.U ...l. "...- %. I ..; .. .... ., I lil .. .......... 11.1 ... .. .... -..,
.. g z i.. -11 :, I I ..,
... .. .1 .. i .S I.e$h. ZwM ii.,lzl ,% 2::: -:,
.. .i i;:.,:..:..: ... .;,. % -%. :wP ...., V. :.:i.. ,
... :. i:,. !7:S'.. ...
"" ,e, !,,,!,.,,,.... .... ..:,,.?,I.ii:.,::.
,:-.. Imi; :1 'm .. .. ..
...:,, N Z, ... .: 'O .: 1, -, :,.:
.: .. I ZF I ,% "k
.; ..Ik ., ...,.-.- ...... I'- I.,V.- ,R ,i% 91 5!12:g g M i R a .;,... ,.,.:.. f .
.:..:,w Z,,..llK:,,:Z,.;eij... ....... -
7, M ." ... ..., CI&G ikek EY, e:,, .- ,-,"n, -:.4;,:%m- 1:.: ....... x4l. ,-
1. .. .Hi 7 ., &-,! 3,1 i :,::,:, .- ... .. : -.e.- ,-. ". ..
11 z ,,z,,,;: ? I I ,,.
;n I, vu : P .. ... 1. .
m6. ZOL Qk.;U I .. .. ..
,. ,,,- a .. -,q 5 %k -. ,:5-8: r, ,,;k,-,:...:. .::,-,;: ,,
-,,,: Q .. ... ; I .,--. 1 I., -:
,. ?'i.. .. I lll I .. 11
-..) ...., i. i n ,, ... ". .& IN I'li Fve .,.liz-.,.l, .11, ... .. i 1
511, llr,,E111-.&.-:,,,-,,, I ....... ., -,%,L,-... .
.. ... I .. ... .. ,.. ....
...;Ill:a ..o %"-" :.,.: .. 14 ,-
,:, .,Ii.. %..! 11,41 ..N....,,?, 1. 1,11-1.,,. 'M .. -, ...
.:,,,, -,: il:! %,,;::h,,,,,i:,i .k ..... ;. ..., P,
4 l4,I tz,; id
F-I t I . . t T
r "s- It Selta I
FARMING SYSTEMS RESEARCH:
A CRITICAL APPRAISAL****
Elon H. Gilbert*
David W. Norman**
Fred E. Winch***
*Visiting Professor, University of Florida.
**Professor, Kansas State University.
***Agricultural Economist, German Aid, Lesotho.
****This paper is published as part of the farming systems research
program being carried out by Michigan State University under the
"Alternative Rural Development Strategies" contract AID/ta-CA-3, U.S.
Agency for International Development, Development Support Bureau, Office
of Rural Development and Development Administration.
MSU is an Affirmative A ction/Equal Opportunity Institution
TABLE OF CONTENTS
PREFACE . . . ... .... .. .ix
ACKNOWLEDGEMENTS . . . . .. xiii
1. INTRODUCTION . . . .... ... 1
1.1 Definition of farming systems research . 2
1.2 Orientation and outline of paper . . 4
2. AN OVERVIEW OF FARMING SYSTEMS RESEARCH . . 6
2.1 Delineation of a farming system . . 6
2.2 "Upstream" and "downstream" FSR . . .. 10
2.3 Schematic framework for a "downstream" FSR program 11
2.4 Attributes of the "downstream" FSR approach .... 13
2.4.1 Consideration of family objectives ..... 13
2.4.2 Incorporating community and societal goals 13
2.4.3 Tapping the pool of knowledge of the society 14
2.4.4 Recognition of the locational specificity of
the technical, exogenous, and endogenous
factors . . . ... .16
2.4.5 The dynamic and iterative nature of FSR .. 17
2.4.6 The integrative and multidisciplinary nature
of FSR . . . . 18
2.4.7 Flexibility in accommodating both technical
and nontechnical improvements in farming
systems . . . . 19
2.4.8 Complementing existing traditional research
approaches . . . . 19
3. REVIEW OF CURRENT FARMING SYSTEMS RESEARCH PROGRAMS 21
3.1 Scope of activities of FSR-type programs . .. 21
3.2 "Upstream" FSR: programs aimed at producing
general prototype solutions . . .... .23
3.2.1 Research on component technology ...... .25
3.2.2 Operational linkages with national programs 26
3.3 "Downstream" FSR: programs focusing on immediate
solutions for specific local situations . .. 27
4. GENERAL METHODOLOGICAL AND IMPLEMENTATION ISSUES OF FSR 31
4.1 Mandates of institutions . . .... .32
4.2 Linkages . . . .. .33
4.3 Credibility . . . . ... 34
4.4 Efficiency of the research process . . .
4.5 Whose constraints should an FSR program concentrate
on? . . . . . .
4.6 Criteria for evaluating improved systems . .
5. METHODOLOGY OF FARMING SYSTEMS RESEARCH .. .
5.1 Descriptive and diagnostic stage .
5.1.1 Selection of the target area
5.1.2 Baseline data analysis ..
5.1.3 On-farm studies . .
5.2 Design stage . . .
5.3 Testing stage . . .
Trials at the farmers' level . .
Farmers' testing . . .
Intermediate types of trials . .
5.4 Extension stage . . . . .
5.4.1 Monitoring and evaluation . . .
5.4.2 Integrating FSR into projects . .
6. INSTITUTIONAL LINKAGES . . . . .
6.1 Problems in expanding FSR in national research
systems . . . . .. .
6.1.1 FSR and national agricultural research
institutions/universities . .
Should separate FSR units be established?
FSR and implementing agencies . .
FSR in relation to funding and planning
agencies . . . .
6.2 FSR at the regional and international centers .
6.3 Possible roles for developed country institutions
7. TRAINING IN FARMING SYSTEMS RESEARCH . . .
7.1 Requirements for FSR training . . .
7.2 FSR training as a part of degree programs . .
7.3 Non-degree training programs in FSR . .
8. CONCLUSIONS . . . . . .
8.1 Current status and potential of FSR . .
8.2 Concerns about FSR . . . .
8.3 Future directions .. : . . .
. . 44
A. Description of Selected Farming Systems Research Programs 86
Al. International Institutes . . ... .86
A1.1 CIMMYT, Mexico . . . . 86
A1.1.1 Basic orientation . . 86
A1.1.2 Program components . . .. 87
A1.1.3 Observations . . . 89
A1.2 ICRISAT, India . . . . 90
A1.2.1 Basic orientation .. . .. 90
A1.2.2 Program components . ... .91
A1.2.3 Observations . . .... .93
A1.3 IRRI, Philippines . . . 94
A1.3.1 Basic orientation . . .. 94
A1.3.2 Program components . ..... 95
A1.3.3 Observations . . .... .98
A1.4 IITA, Nigeria . . . . 98
A1.4.1 Basic orientation ... . 98
A1.4.2 Program components . . .. 99
A1.4.3 Observations ..... ...... 100
A1.5 CIAT, Colombia . . . . 101
A1.5.1 Current orientation . . .. 101
A1.5.2 Program components . ... 102
A1.5.3 Observations . . .... .103
A2. Regional Institutions . . . .... .105
A2.1 CATIE, Costa Rica . . . .. 105
A2.1.1 Basic orientation . . 105
A2.1.2 Program components . . .. 106
A2.1.3 Observations . . .... 107
A3. National Farming Systems Programs . ... 109
A3.1 ICTA, Guatemala . . . . 109
A3.1.1 Basic orientation .. . 109
A3.1.2 Program components . . .. 110
A3.1.3 Observations . . .... .112
A3.2 ISRA, Senegal . . . . 113
A3.2.1 Basic orientation . . .. 113
A3.2.2 Program components . ... 114
A3.2.3 Observations . . .... .115
A3.2.4 Reorganization of ISRA . ... 116
A3.3 ICA, Colombia . . . . 117
A3.3.1 Basic orientation . . .. 117
A3.3.2 Program components . ... 118
A3.3.3 Observations . . .... .119
B. Farm Management Research . . . ... 121
SELECTED BIBLIOGRAPHY ....... . . ... .123
LIST OF ACRONYMS . . . . ... ..... .135
LIST OF FIGURES
1. Schematic representation of some determinants of the
farming system 7
2. Schematic framework for farming systems research at the
farm level 12
3. Scope of activities of FSR programs in selected national,
regional and international institutes 24
This state-of-the-art paper is the second in a series of papers on
farming systems research (FSR) in the Third World. The objectives of
the paper are to: (a) review the literature on farming systems,
(b) evaluate farming systems research in international institutes
and in national agricultural research systems in the Third World, and
(c) recommend what can be done to improve and expand FSR in order to
develop technology that is appropriate for the majority of small
The authors discuss the confusion over the various definitions of
FSR. They recommend a working definition which includes a holistic
approach to diagnosing constraints faced by small farmers in site-
specific locations, and in carrying out farm trials of promising
technology. Stressing the strengths and weaknesses of current FSR
programs, they warn of the dangers of overselling FSR and of setting
up separate FSR departments; they advocate instead a close working
link between FSR and commodity research teams.
The lessons from agricultural sector studies in the 1970s should
be taken into account as FSR expands in the 1980s. Agricultural sec-
tor studies failed to gain credibility in the 1970s because the micro
research base was often inadequate to support macro models. Hence
FSR could easily lose its credibility if micro research is not
supplemented by macro research on the political, economic, and
institutional constraints on small farmers in the Third World.
A major section of the paper deals with rapidly evolving method-
ologies for carrying out FSR. FSR is not inexpensive. And critics are
raising valid questions about its cost effectiveness. But one can
only speculate on the costs and returns of FSR as compared with the
traditional "top-down" experiment station research approach that has
proven so capable of serving commercial farmers. For example, while
the top-down approach has been effective in serving Zambia's 300
commercial farmers, it has failed to reach Zambia's 500,000 small
farmers. The challenge is how to serve the majority of small farmers.
This question involves both sides of the equation--costs and returns.
Whereas the traditional research approach involves heavy capital
outlays for experiment station buildings and equipment, FSR requires
major recurrent costs to support site-specific research teams. MSU
will publish a paper in 1981 on the cost effectiveness of FSR com-
pared with the top-down research approach.
A number of innovations are currently reducing the cost of FSR.
In many countries FSR teams are using 2-4 week reconnaissance surveys
("sondeos") to identify the major problems facing small farmers.
But reconnaissance surveys must be supplemented by frequent interviews
of farmers ("cost route surveys") over a full year, followed by less
frequent interviews over 3-5 years as new technology is tested by
farmers. Since frequent interviewing techniques involve recurrent
costs that are increasing rapidly (e.g., petrol costs $3 to $4 a
U.S. gallon in many Third World countries), it is necessary to shift
to less intensive methods of data collection. For this reason, a
Michigan State University research team in Eastern Upper Volta has
recently shifted from weekly interviews of small farmers to interview-
ing rice farmers 14 times per year--once for each of the 14 activities
(e.g., planting, weeding) involved in the production of rice. The
results of the MSU survey will be available in mid-1981 and will
provide a comparison of the cost and accuracy of weekly interviews
with the "activity" approach.
Improvements are needed to speed up data processing and in publish-
ing results. The FAO1 is developing standardized terminology for farm
management concepts in French, English, and Spanish and has developed
pre-coded questionnaires and a standardized computer program. This
program can be used in different ecological zones and countries to
generate partial and whole farm budgets, and crop and livestock
enterprise tables on a farm by farm basis, as well as the usual
sample averages. A growing number of Third World countries are now
using micro-computers in farm surveys.' The strengths and weaknesses
K. H. Friedrich, Farm Management Data Collection and Analysis:
An Electronic Data Processing, Storage and Retrieval System. Rome:
FAO, 1977. For information about FAO's program of work, write to
Neal Carpenter, Chief Farm Management and Production Economics
Service, FAO, Via delle Terme di Caracalla, 00100 Rome, Italy.
of micro-computers will be assessed in a MSU Rural Development Paper
The third paper in this FSR series, "CIMMYT's Experience in
Facilitating Farming Systems Research in Eastern and Southern Africa,"
is being prepared by Michael Collinson and will be published in late
Carl K. Eicher, Director
Alternative Rural Development
The preparation of this paper has proved to be a valuable learn-
ing experience for us. In undertaking the assignment, which was
requested by Michigan State University, we found much of the material
required was not readily available. In some cases the material
available could not easily be traced to a written source, which is
not surprising because farming systems research is rapidly evolving.
So a great deal of emphasis has been placed on comments by individuals
currently engaged in, or interested in, farming systems research.
We would like to express our gratitude to the following who spent
so much time reviewing drafts of this monograph: D. Baker, J. Bingen,
D. Byerlee, C. Charreau, M. Collinson, E. Crawford, P. Crawford,
C. Davis, C. Eicher, J. Faye, S. Franzel, W. Freeman, H. Gilbert,
A. Hansen, H. Hays, P. Hildebrand, G. Johnson, J. Kampen, B. Michie,
L. Navarro, J. Ryan, J. Sanders, W. Shaner, J. Sjo, B. Stavis,
T. Van Schillhorn, W. Vincent and H. Zandstra. We may not have done
complete justice to all the comments of reviewers, so deficiencies
remaining in the manuscript should be attributed to us alone.
Contribution number 80-401-A, Department of Agricultural Economics,
Agricultural Experiment Station, Kansas University, Manhattan, KS 66506.
Increasing empirical evidence shows that the needs of small
farmers often have not been adequately addressed in development pro-
grams in the Third World over the past twenty years (Khan, 1978;
Poleman and Freebairn, 1973). Many development projects have been
introduced without sufficient understanding of the environment in
which small farmers operate. The chequered pattern of success is
traceable in part to the way research has been organized and under-
taken in low income countries (Longhurst, Palmer-Jones, and Norman,
1976). Public investment in agricultural research has not always
been spent with the needs of small farmers--who should be the major
customers of the results of such research--in mind.1 Instead allo-
cation of funds often has been based on:
(1) Expressed needs of more influential farmers2 who often hold
nonagricultural jobs in the society.
(2) Research that will appeal to professional "peer groups" of
(3) Types of technology that have been developed in high income
Therefore the link between the small farmers and the research
organizations has tended to be weak (Stavis, 1979). Traditionally
this interaction should have been facilitated via the extension
worker, but for a number of reasons this has not often worked. Two
possible reasons are:
(1) Institutional and administrative barriers which prevent
effective interaction between researchers, extension
workers and farmers.
1Anderson (1979) gave an excellent analysis of the factors
influencing misallocation of research resources in many LDC's.
Some have argued that this tendency has been present in
agricultural research at some of the Land Grant Universities in the
U.S. (Hightower, 1972; Heady, 1973).
(2) Researchers in the Third World often have higher academic
qualifications than extension staff, thereby reinforcing a
tendency toward top-down prescriptions--from research workers
to extension workers to farmers.
The quest for an efficient way of developing more relevant
research programs for small farms is analogous to the process used by
commercial firms producing a product for sale; that is, ascertaining
what the consumers or customers want. The farming systems research
approach starts with the farmer and provides a link between the farmer
and the research institution and funding agency, thus counterbalancing
the more conventional "top-down" experiment station research approach.,
The farming systems approach has the potential of providing the cus-
tomers, in this case small farmers, with an avenue for communicating
their needs, both to research workers and to funding agencies.
1.1 DEFINITION OF FARMING SYSTEMS RESEARCH
The primary aim of the FSR approach is to increase the produc-
tivity of the farming system in the context of the entire range of
private and societal goals, given the constraints and potentials of
the existing farming systems. Productivity can be improved through
the development of relevant technology and complementary policies which
increase the welfare of farming families in ways that are useful and
acceptable to them and society as a whole. Farming systems research
(FSR) has the following characteristics:2
(1) Farming systems research views the farm or production unit
and the rural household or consumption unit--which in the case
of small farmers are often synonymous--in a comprehensive
The FSR approach is, therefore, more realistic in orientation than
the more conventional reductionist approach exemplified by commodity
research programs, The reductionist approach involved studying one or
two factors at a time while attempting to control all others (Dillon, 1976).
2We are grateful for the help of Shaner Cpersonal communication) in
delineating these characteristics,
manner. FSR also recognizes the interdependencies and
interrelationships between the natural and human
environments.2 The research process devotes explicit
attention to the goals of the whole farm/rural house-
hold3 and the constraints on the achievement of these
(2) Priorities for research reflect the holistic perspective
of the whole farm/rural household and the natural and human
(3) Research on a sub-system4 can be considered part of the FSR
process if the connections with other sub-systems are re-
cognized and accounted for.
(4) Farming systems research is evaluated in terms of individual
sub-systems and the farming system as a whole.5
A variety of research and development activities falls under the defi-
nition of farming systems research. In addition some research programs
(e.g., commodity research programs) are not described as FSR programs,
but they exhibit most or all of the characteristics listed in our
definition. The focus of this paper is on research which includes the
four characteristics in our definition of farming systems research.
1As we discuss later (Section 2.2), we would prefer to confine the
use of the term FSR to research that has not only the characteristics
listed but also the active participation of the farmer in the research
2This ensures some consistency between the unit managed by the farm-
ing family and the unit studied in agricultural research programs (Hart,
31n the paper we use the term farming household or farming family to
stress the production and consumption interrelationships (see Section 2.1).
Sub-system implies a boundary separating the system from its
environment. Two systems may share a common component or environment
and one system may be a sub-system of another. So a farm system can
be broken down into a number of sub-systems--for example, crops, live-
stock, and off-farm--which may overlap and interact with each other
(Technical Advisory Committee, 1978).
The farming system reflects the resolution of the conflicts
between the goals of, and the constraints faced by, the farming household.
1.2 ORIENTATION AND OUTLINE OF THE PAPER
We have approached this review of FSR with definite notions about
the role FSR should play, the breadth of its activities, and its rela-
tionship to existing agricultural and rural development institutions.
We believe that:
C11 FSR is a unique and potentially significant approach that
can greatly increase the effectiveness of agricultural
research and development programs in the Third World. FSR
has antecedents in farm management activities in the U.S.
during the first half of the century Csee Appendix B) and
in the community development programs of the post-World War
II period CHoldcroft, 19781. FSR includes some charac-
teristics of both approaches.
C2) FSR concentrates on the individual farming family, which
necessitates a multidisciplinary team of researchers,
farmers, and extension workers interacting at the local level.
Thus the goals/objectives of the farming household tend
to take precedence in the process of designing improvement
measures. The importance of governmental policy--objectives
and societal concerns such as environmental quality--is
recognized, but to effectively incorporate the concerns
requires strong linkages with existing institutions that are
specifically responsible for such matters, including planning
ministries, ministries of agriculture and natural resources,
(3) Although FSR is holistic in its orientation, the degree of
comprehensiveness of FSR in practice is tempered by the state
of development of FSR methodology, resource availability, and
the limitations of agricultural-development planning
in the Third World,
C41 FSR has its institutional roots in the agricultural research-
institutes and thus has a bias toward bio-technical modifica-
tions in farming systems, although there is increasing recogni-
tion that changes in nontechnical factors such as markets,
pricing policy, institution, and infrastructure are often
C5) The operational perspective of our discussion ts that of the
researcher and rural development practitioner at the local
level rather than the theoretician. While we acknowledge
contributions made by researchers in understanding the nature
of agricultural systems through systems analysis, our focus
is upon FSR which forms a direct input Into the design and
implementation of development programs at the, local level.
C6) Effective FSR activities require close links with strong
commodity and disciplinary agricultural research programs,
We believe that the results of FSR will enrich commodity
and disciplinary research programs and provide inputs into
development programs designed by FSR teams at the local level,
FSR is not intended to replace either of these agricultural
(71 The focus should be on the possible and practical rather than
what would be ideal, For example, the conceptual framework
for FSR in Chapter 2 is couched more in terms of the desirable
and achievable rather than the perfect, Our review of exist-
ing FSR programs summarized in Chapter 3 reveals that some
existing programs already contain many features of our "ideal"
program. The discussion of methodology in Chapters 4 and
5 is not intended as a definitive and comprehensive treat-
ment of methodological problems and their solutions, but
rather a review--with commentary--on sometimes contrasting
approaches to specific aspects of FSR gleaned from the
published and unpublished observations of many practioners.
Similarly, the discussion of institutional issues and train-
ing activities in Chapters 6 and 7 focuses on the practical
problems of initiating and maintaining FSR programs within
the framework of existing institutions,
2. AN OVERVIEW OF FARMING SYSTEMS RESEARCH
In this chapter we define what a farming system is, offer a
schematic framework for conducting FSR, and discuss some of the dis
tinguishing features. The proposed approach discussed in this section
may differ from actual programs because of the compromises that must
necessarily be made in implementing a FSR program.
2.1 DELINEATION OF A FARMING SYSTEM
A system can be defined conceptually as any set of elements or
components that are interrelated and interact among themselves. Thus
a farming system is the result of interactions among several inter-
dependent components. At the center of the interactions are the
farmers themselves, whose households or families and means of live-
lihood are intimately linked and must not be separated. That is one
reason why we frequently refer to the farming family rather than just
the farmer.1 For achieving a specific farming system, farming families
allocate certain quantities and qualities of basic types of inputs--
land, labor, capital, and management--to which they have access, to
three processes--crop, livestock, and off-farm enterprises--in a manner
which, given the knowledge they possess, will maximize attainment of
the goals) they are striving for (Norman, Pryor, and Gibbs, 1979).
Figure 1 illustrates some of the possible underlying determinants
of the farming system, The "total" environment in which farming house-
holds operate can be divided into two parts: the technical
element and the human element (Norman, 1976).
The types of, and physical potentials of, livestock and crop
enterprises will be determined by the technical element, which reflects
what the potential farming system can be and therefore provides the
necessary condition for its presence. In the past the technical
element received most attention, particularly from technical scientists.
They have, within certain limits, been able to modify the technical
In addition there are often multiple decision makers within a
particular household (Newman, Ouedraogo and Norman, 1980.,
Broken lines represent results of farming system.
L___-- -------v -- -- -- -
Figure 1 Schematic Representation of Some Determinants of the Farming System
I I w
element and improve the potential farming system by developing technol-
ogies that partially alleviate the deficiencies in the technical
The technical element can be divided into two factors: physical
and biological.1 Physical factors are water, soil, solar radiation,
temperature, etc. Technical scientists, for example, can enhance water
availability through Irrigation (i.e., through the use of mechanical
techniques), or soil quality through fertilizer application (i.e.,
through the use of chemical techniques). Biological factors are crop
and animal physiology, disease, insect attack, etc. Examples of
limited intervention of technical scientists in this area would include
breeding early-maturing crop varieties and varieties that resist disease.
The farming system that actually evolves, however, is a subset of
what is potentially possible as defined by the technical element. The
determinant that provides the sufficient condition for the presence
of a particular system is the human element, characterized by two
types of factors: exogenous and endogenous.
The exogenous factors that largely influence the farming systems
in any given community are the social, economic, and political institu-
tions in the area--all largely outside the control of the individual
farming household. Yet all directly influence what the farming house-
hold or individual members can and cannot do, The exogenous factors
can be classified into three broad groups:
(1) Community structures, norms, and beliefs. Local institutions
and beliefs often directly affect the acceptability of speci-
fic development strategies. For example, processing of cer-
tain food crops may be the responsibility of the women, while
operating machinery is the responsibility of the men. In
The technical element can be considered as an exogenous factor
even though the "exogenous factors" in the text refer only to those
under the human element.
The technical element can affect the ways the human element
evolves. For example, in pastoral communities in Africa technical con-
siderations such as limited rainfall dictate the predominance of graz-
ing activities in certain areas, which in turn influence community
structures, norms, and beliefs and other exogenous factors, including
such a situation introduction of processing equipment is
faced with certain difficulties.
(2) External institutions. The two main types of institutions
influencing farming decisions are the input supply system and
markets where the farmers can sell or trade their commodi-
ties. On the input side, in the developing areas of the
world, programs such as extension, credit, and input distri-
bution systems are often financed and manned by government
and, therefore, reflect its policies. On the farm product
side, government may directly (e.g., marketing boards) or
indirectly (e.g., improving evacuation routes, transporta-
tion systems, etc.) influence the prices farmers receive.
C3) Other influences such as location and population density,
Endogenous factors, on the other hand, are those the individual
farming household to some degree controls, including the four basic
types of inputs mentioned earlier--land, labor, capital, and manage-
ment.1 It is important to recognize that these resources vary among
households, regions, and countries on the basis of both quantity and
quality, both of which influence the performance and potential of the
system. In addition these inputs or resources may or may not be owned
by the household, Access to one or more of these resources may be on
another basis of use, which may limit or restrict the ease or intensity
of use and thus, in turn, affect the goals and performance of the farm
Farmer goals and motivation are critical endogenous factors that
may profoundly affect the nature of the farming system, particularly in
situations where a range of options or enterprise combinations is con-
sistent with the existing technical element and exogenous factors.
Farmer goals and motivation are in another respect the motor that drives
the entire system--that gives it a dynamic dimension. Even where
changes in the technical element Ce.g., drought) and exogenous factors
(e.g., civil war) force alterations in the farming system, farmers still
1Management might be considered as a special type of input that
serves as a mechanism to implement decisions regarding farming activi-
ties made by the farm family as it selectively employs the other inputs--
land, labor, and capital.
have options, so the resulting choices are invariably strongly in-
fluenced by individual goals and motivation.
The farming system obviously is complex, which explains why some
technology thought to be relevant often has not been adopted, or when
it has, why the degree of adoption varied widely. Not considering
the human element in agricultural research has contributed to many
so-called "improved" technologies being irrelevant.
2.2 "UPSTREAM" AND "DOWNSTREAM" FSR
Two types of farming systems research programs have emerged in
recent years; namely "upstream" and "downstream". We believe there is
a fundamental difference between the objectives and nature of activi-
ties for the two types of programs. "Upstream" FSR seeks to generate
prototype solutions which will facilitate major shifts in the potential
productivity of farming systems. "Upstream" research often involves
several years of research, both on and off station, and is particularly
the concern of the International Agricultural Research Centers (IARCs)
and selected regional research programs. "Downstream" or site speci-
fic FSR programs are designed to rapidly identify and subsequently test
possible innovations which can be easily integrated into existing farm-
ing systems. "Downstream" FSR focuses on close interaction with farmers
via on-farm trials and draws selectively upon results from commodity,
discipline oriented research or "upstream" programs. Downstream FSR
programs are commonly carried out within the context of a national
agricultural development project or research institute.
In this paper we have chosen to discuss both "upstream" and "down-
stream" FSR programs while concentrating on "downstream" FSR. More
detailed discussion of the two types of FSR programs is contained in
Our own bias, which P. Crawford (personal communication) shares,
would be to confine the use of the term FSR to studies characteristic
of "downstream" FSR (i.e., those including the whole farm perspective
[Section 1.1] and the active participation of the farmer). The farmer
rarely participates actively in "upstream" FSR--particularly in the
early stages of the research process.' However, since the term FSR is
now commonly used to denote both the "upstream" and "downstream"
variants, we are reluctant to redefine it to suit our own bias.
2.3 SCHEMATIC FRAMEWORK FOR A "DOWNSTREAM" FSR PROGRAM
A schematic framework for a "downstream" FSR program is given in
Figure 2. Four stages of research can be delineated as follows:
(1) The descriptive or diagnostic stage in which the actual farm-
ing system is examined in the context of the "total" environ-
ment--to identify constraints2 farmers face and to ascertain
the potential flexibility in the farming system in terms of
timing, slack resources, etc. An effort is also made to
understand goals and motivation of farmers that may affect
their efforts to improve the farming system.
(2) The design stage in which a range of strategies are identified
that are thought to be relevant in dealing with the con-
straints delineated in the descriptive or diagnostic stage,
(3) The testing stage in which a few promising strategies aris-
ing from the design stage are examined and evaluated under
farm conditions, to ascertain their suitability for produc-
ing desirable and acceptable changes in the existing farming
system, This stage consists of two parts; initial trials at
the farm level with joint researcher and farmer participation,
then farmer's testing with total control by farmers them-
(4) The extension stage in which the strategies that were identi-
fied and screened during the design and testing stages are
In practice there are no clear boundaries between the various
stages. Design activities, for example, may begin before the descrip-
tive and diagnostic stages end and may continue into the testing stage,
as promising alternatives emerge during the trials at the farm level--
where farmers and researchers interact directly. Similarly, testing
by farmers may mark the beginning of extension activities.
1Although the primary focus of the schematic framework is the "down-
stream" FSR program, there are major similarities to "upstream" FSR.
The complexities surrounding the constraint issue are discussed
later (Section 4.5).
1. Description or
diagnosis of present
2. Design of improved
3. Testing of improved
4. Extension of improved
Experiment Station Trials < ---------> BODY OF
- - ---- -- -- -- -- -- >
Trials at Farm Level ---------------------
E ing ---------------------
MODIFIED FARMING SYSTEM ----------------------- j
. . . . .
Figure 2 SCHEMATIC FRAMEWORK FOR FARMING SYSTEMS RESEARCH AT THE FARM LEVEL
(Downstream Farming Systems Research)
2.4 ATTRIBUTES OF THE "DOWNSTREAM" FSR APPROACH
Some of the important attributes of "downstream" FSR are now discussed.
2.4.1 Consideration of family objectives
The objectives of the farmer (farming family) are directly incor-
porated into the designing and testing of strategies. An attempt is made
to understand the farmer's objective function in the initial descrip-
tive or diagnostic stage. The farmer directly participates in all
stages except possibly the design. This ensures evaluation criteria
relevant to the farmer, rather than simply the conventional returns-
per-unit-of-land so often used. Also the FSR approach recognizes that
farmer objectives may change over time. For example, as development
proceeds, the importance of community norms and beliefs in shaping
individual farmer goals may diminish. As the FSR approach is used in
designing successive generations of strategies, changes in farmer
objectives can be incorporated in the process.
2.4.2. Incorporating community and societal goals
The FSR approach views farmers both as individuals and as members
of the larger community and society, Thus the approach links the micro
perspective with broader societal considerations in the process of
designing development strategies, Such strategies may involve single
innovations proposed for adoption by farmers, such as improved seeds, or
policy changes that alter fertilizer subsidy levels.
Societal goals could include maintaining soil fertility to enable
the land resource to be used by future generations, avoiding an increase
in inequality of income distribution, and other goals. But it is
It is generally assumed that the objectives of the farmer--
usually the head of the family unit--reflect those of the farming
family as a whole, so the terms farmer and farming family are often
used interchangeably. However, that may not be true of other members
of the family with fields under their own control (Newman, Ouedraogo,
and Norman, 1979).
likely that such goals are not going to be achieved simply through the
development of improved technologies or practices that reflect the
heterogeneity that exists in the farming community. For example, it is
likely that, ceteris paribus, farming families with better quality
resources and easier access to external institutional support systems
will still progress more rapidly. Nevertheless the development of
improved practices relevant to farming families in less fortunate cir-
cumstances can at least slow down the increase in inequalities of in-
come distribution. In such cases a more positive effect of the applica-
tion of FSR may involve influencing changes to be made in agricultural
policy and in the operations of farmer-contact agencies, Although the
potential exists for FSR to be of value in such areas--in addition to
its current application in the development of improved practices--
there are as yet no examples where it has been systematically applied
in this fashion.
2.4.3 Tapping the pool of knowledge of the society
FSR recognizes that the potential benefactor (the farmer) must be
an integral part of the research process. The concept explicitly
recognizes the value of the farmers' experience (Swift, 19781 and their
traditionall experimentation (Johnson, 1972; Jodha, Asokan, and Ryan,
1977) as inputs into developing strategies for improving the produc-
tivity of existing farming systems.
Many changes envisioned in FSR involve small adjustments rather
than complete changes in the farming system. In addition, even greater
reality is encouraged in the research process through maximizing
research under actual farm conditions, When testing improved techno-
logy, the managerial input is initially provided by the research
worker--trials at the farmer's level (Figure 2); and then, often later,
by the farmer himself--farmer's testing.
The link with the extension worker in such work activity is vital
(Asian Cropping Systems Working Group, 1979; Navarro, 1979), Extension
Further discussion on this is presented in Sections 4,2 and
workers' knowledge about the local situation at the farm level and the
responsibilities they eventually will have for disseminating the
results of FSR, make it imperative that extension workers be involved,
or at least consulted, at each stage of the FS'R process, Interaction
with the research team has another benefit for extension workers. In
many countries they have been taught to tell farmers what they should
do rather than to listen and to help farmers through dialogue with them
(Belshaw and Hall, 1972), which is so important in the FSR approach,
Research workers often have cut themselves off from such valuable
knowledge and wisdom. As a result, researchers often spend considerable
time "rediscovering the wheel" rather than building on the knowledge
that farmers and extension workers already possess,
An example is the practice in many LDC"s of farmers growing crops
in mixtures; that is, more than one crop at the same time. For many
years that practice was considered by many agricultural scientists
and, for that matter, by officials in ministries of agriculture as
"primitive" and not compatible with "modern" agriculture, So it was
not considered worthy of serious research endeavor. However, efforts
in many parts of Africa to encourage farmers to plant single crops of
improved varieties alone often have failed. Why? The results of
surveys in northern Nigeria indicate that under indigenous technological
conditions it was rational for farmers to grow crops in mixtures when
either labor or land was limited. Mixed cropping proved to be more
profitable than single crops and to yield a more dependable return
(Norman, 1974). Belatedly, considerable interest in mixed cropping
with improved technology has developed amongst technical scientists
CMonyo, Ker, and Campbell, 1976). Many of the results confirm the
methods that farmers evolved over generations (Willey, 1978). Undoubt-
edly much more progress with mixed cropping could have been made if
the pool of knowledge possessed by farmers had been tapped earlier.1
A contribution from Collinson (personal communication) makes the
same point quite succinctly;
"I find both scientists and administrators don't really understand
what farmer participation can imply. I often give a hypothetical
dialogue between farmer and agronomist to show what it can be,
(continued on next page)
2.4.4 Recognition of the locational specificity of the technical,
exogenous, and endogenous factors
The FSR approach involves breaking heterogeneity into homogeneous
subgroups and developing strategies appropriate to each. The disaggre-
gation into homogeneous subgroups is first done according to ecologi-
cal systems or to differences in the technical element; then, if
further disaggregation is necessary, differences in the human element
may be basis for subgrouping CSection 5,1,31, The aim of such dissag-
gregation is that, in terms of interest to researchers, the variance
between subgroups be maximized and within them minimized, and that the
classification be useful as a guide to developing relevant strategies
(Technical Advisory Committee, 1978). The constraintCsl-most limiting
in the farming system of each subgroup as revealed by analyzing the
results then become the focus of research efforts.
Footnote 1 cont.
(a) Agronomist: We thought of having three cowpea to each
maize plant in this treatment.
(b) Farmer: What would I do with all those cowpeas,
there's no market and we only eat about 1/5 by weight of
(c) Agronomist: OK let's reduce to say equal maize and
cowpea to give a ratio close to the weights needed we
thought of putting the cowpeas in the row between the
maize plants which are one foot apart,
(d) Farmer: but that only leaves 6 inches between the maize
and the bushy cowpea plant how can we get our hoes
in for weeding?
(e) Agronomist: OK what about putting the cowpeas in be-
tween the plants within the maize rows?
(f) Farmer: Well we weed by putting the hoe between the
maize plants and pulling weeds into the interrow it
will slow us down a lot.
Cg) Agronomist: So what do you suggest?
(h) Farmer: Why not put the cowpea seed in the same hole
as the maize seed as we do now?"
2.4.5 The dynamic and iterative nature of FSR
The research process is recognized as being dynamic and iterative,
with backward linkages among farmers, research workers, and sponsoring
agencies rather than simply the presence of forward linkages character-
istic of the "top down" approach.
An example from northern Nigeria illustrates both the iterative
process and inefficiencies that can arise in allocating research
resources if a farming system perspective is not maintained. Tradition-
ally cotton, often grown in mixtures, is planted after the peak
labor demand period in June-July is partially past and priority has
been given to planting and weeding food crops. Growing cotton according
to the recommended practices--which were drawn up in the absence of a
farming systems approach--involved planting earlier sole stands,
and called for fertilizer and spraying six times with a knapsack sprayer
that used 225 litres of water per hectare each time. Ex post FSR
revealed why farmers in general were not adopting the recommendations
in their entirety (Beeden et al., 1976). On average the net return per
hectare was higher from growing cotton according to the recommended
practices, but the improved cotton technology needed to be planted
earlier in June and July when food crops were being planted and weeded.
Although the return per man-hour of labor on an annual basis was
higher, the return during the June-July labor bottleneck period was
lower for improved cotton than cotton grown according to traditional
practices. That, and the large amount of water required for spraying,
no doubt accounted for the farmers' lack of interest in the improved
cotton technology. The ex post FSR revealed that farmers were not
comparing the improved technology with traditional cotton technology,
but with the alternative of devoting labor to other enterprises--in
this case, food crops.
The above results of ex post FSR lent support to the development of
a modified technology under which cotton could be planted later when it
fitted in better with the traditional farming system. Also recommended
was replacing a water-based insecticide with an oil-based one that could
be applied with an ultra low-volume sprayer (Beeden, Hayward, and
Norman, 1976). Understanding of the farming system and the constraints
faced by farmers, as envisioned in a true FSR approach, would no doubt
have reduced the research resources devoted to developing a relevant
improved cotton technology.
2.4.6 The integrative and multidisciplinary nature of FSR
Most past agricultural research in developing countries has been
based on narrow disciplinary approaches, which left farmers the diffi-
cult task of integrating new information into their farming systems
(Technical Advisory Committee, 1978). Collinson (1979a) gave an example
of how impractical that sometimes can be. At one research institution
in East Africa, commodity-orientated research showed that the optimal
planting time for six crops grown by local farmers was the first week
after the rains began. Using hoes, farmers could prepare only one-
third hectare during that week. If farmers had stopped planting then,
their income would have dropped 80 percent.
FSR provides a means by which multidisciplinary teams of researchers
can examine problems of the farming system, including complementary and
supplementary relationships between resources and enterprises. With fos-
sil energy costs increasing, the possible ramifications of this are obvi-
ous. Such interactions have rarely been exploited in the reductionist
approaches to developing improved technology. It has been suggested that
if researchers overlook these interactions there may be adverse effects
on specific enterprises. For example, it was agreed at a recent workshop
(McDowell and Hildebrand, 1980) that small livestock have been adversely
affected in some research on cropping systems. The necessity of recog-
nizing and focusing on the interaction of the technical and human ele-
ments and fully appreciating the multiple use of resources requires a
multidisciplinary team working in an interdisciplinary manner. The
1An excellent example of the value of FSR in improving the effi-
ciency of research resources through systematic analysis of labor input
in relation to crops--in the case maize and sorghum--and rainfall
patterns is illustrated by work in central Tanzania (CIMMYT, 1977).
2Multidisciplinary suggests involving several disciplines while
interdisciplinary connotates the disciplines working together, rather
than independently, in solving a specific problem,
social scientist should play an ex ante role rather than simply the
traditional ex post role characteristic of the "top-down" approach,
For example, in India, the ex ante likelihood of labor bottlenecks with
improved watershed-based farming systems was demonstrated by Ryan
et al. (1979). The improved systems are now being tested on farmers
fields in cooperation with AICRPDA. The team, including both technical
and social scientists, needs to be involved at the first three stages of
the research process and possibly some in the fourth stage.
2.4.7 Flexibility in accommodating both technical and nontechnical
improvements in farming systems
Traditionally agricultural research has been rather narrowly
focused on yield-increasing technical innovations for specific commod-
ities. FSR is concerned with the productivity of the entire farming
system and, as a result, it will examine nontechnical changes that are
exogenous--factors like improving marketing arrangements for inputs
and outputs. The flexibility inherent in the FSR approach also assists
in linking macro and micro perspectives in designing strategies more
effective for specific rural areas or groups of farmers. National
policies like pricing and trade policies that affect agricultural
producers may be explicitly considered when diagnosing existing farming
systems and designing improvements, Changes in such policies may be the
most critical ingredient in efforts to improve the lives of small
2.4.8 Complementing existing traditional research approaches
The farming systems research approach is not intended to replace
basic and applied research or what can be described as the "body of
knowledge" (Figure 2). Also, the "body of knowledge" will be augmented
by FSR as follows. First, the results of the FSR approach in a specific
area may be applicable, with some modification, to other areas with
similar environments. Second, the variant of FSR that is "upstream"
can be used to develop prototype solutions, usually in the form of
packages of practices that address themselves to common constraints
facing a broad range of farming systems across one or more geographic
regions. For example, the Cropping Systems Program of IRRI seeks to
develop practices that will facilitate intensifying rice cropping
systems throughout South and Southeast Asia, as scarcity of land is an
overriding problem throughout most of those regions CIRRI, 1978).
Similarly, the Farming Systems Program at ICRISAT focuses on Improving
watershed-management practices because water is a critical common
constraint in the semi-arid tropics CKrantz, 1979). Such prototype
"solutions" become part of the "body of knowledge" applying the FSR
approach to a local situation,
The complementary nature of other research approaches in contribut-
ing to the "body of knowledge" is underlined in "downstream" FSR, which
draws upon this information in the process of designing practices or
recommendations suited to the specific local situation in the immediate
3. REVIEW OF FARMING SYSTEMS RESEARCH PROGRAMS
The upsurge in interest in FSR is largely a product of the 1960s
and 1970s. Thus, of the fewer than twenty FSR programs in the Third
World at present, most are fairly young and still in formative periods.
In many instances, programs are still in the process of defining a
research focus and developing methodologies. The differences among
existing FSR programs reflect in large part the diversity of the insti-
tutions involved: their histories, objectives, and scope of responsi-
bilities--national, regional, and commodity foci. While common fea-
tures emerge, certain important differences remain. This section ex-
amines the scope of FSR programs and distinguishes between the two
major types of programs, namely basic, general or "upstream" variety
and the site specific or "downstream" type. As noted earlier, this
paper focuses on "downstream" programs but the principal features of
both types are reviewed in this section. Summaries of FSR activities
at selected institutions including ICTA (Guatemala), ICA (Colombia),
ISRA (Senegal), CATIE (Central America), and selected IARCs (CrMMYT,
IRRI, IITA, CIAT, and ICRISAT) are in Appendix A,
3.1 SCOPE OF ACTIVITIES OF FSR PROGRAMS
Most FSR programs are still confined to developing technology
for the crop subsystem as a consequence of the crop mandate of some
research institutes (e.g., CIMMYT and IRRI), the current state of FSR
methodology, and the scarcity of researchers with FSR experience.
Some scholars (Boer and Welsch, 1977) have appealed for livestock to be
included in the process, but the approach has rarely been applied to
the livestock subsystem except where it impinges directly on the crop
subsystem. ILCA is now engaged in FSR on the livestock subsystem while
ISRA in Senegal and CATIE in Costa Rica are addressing crop-livestock
interactions. CIAT is pursuing limited FSR for both swine and cattle.
Excluding or assigning low priority to livestock research has been a
subject of considerable debate centering on the future of livestock in
the developing world under conditions of land scarcity, high population
growth rates, poverty, and staple food deficits. Since livestock is
an integral part of the farming systems of most of the world's popula-
tion, we believe it should be given due consideration in applying the
In addition to the subject focus of FSR programs is the issue of
sequencing research activities. Ideally, agricultural research might
be conceived of as a smooth continuum of interconnecting activities as
(1) Analyzing the existing situation.
(2) Initiating of basic lines of research.
(3) Developing broadly generalizable solutions.
(4) Adapting solutions to specific situations,
(5) Initiating of action programs,
In reality, the process is disjointed because of the fragmented insti-
tutional responsibilities among national, regional, and international
centers; variations in research methods required to generate solutions
for different problems; and differences in the state of knowledge with
regard to improved technologies for different commodities, enterprise
mixes and geographic areas, For example, the existing body of know-
ledge is more likely to offer readily available or easily adaptable
solutions for monocultural farming systems for major grain crops than
for complex intercrop situations involving a variety of less well
known commodities or both crops and animals.
The "ideal" program in any given situation will probably involve
some mixture of "upstream" and "downstream" features as determined in
part by the availability of innovations which can be easily and rapidly
integrated into existing farming systems. Where the pool of such inno-
vations is large, a "downstream" program can be an effective mechanism
to identify and adapt the most promising approaches. Conversely, where
significant research of a more basic or general nature is required, an
"upstream" program may provide an appropriate mode to organize the re-
search effort in a fashion which cuts across traditional disciplinary
and commodity lines. The appropriate mix may be achieved through
linkages between different programs rather than attempts to combine
For a discussion of integrating crop and animal production systems
in an FSR context see McDowell and Hildebrand (19801.
both dimensions in the same program. At a minimum, there should be a
two way flow of information from the farm level to research institu-
tions and from the research stations to the farmers-possibly via
"downstream" FSR programs--in the form of research results, In prac-
tice, links between both types of FSR programs on the one hand and
commodity/discipline oriented research on the other are likely to be
stronger than the links between "upstream" and "downstream" programs.
Figure 3 summarizes in a general fashion the scope of activities
of FSR-type programs at selected national, regional and international
agricultural research institutions,
3.2 "UPSTREAM" FSR: PROGRAMS AIMED AT PRODUCING GENERAL PROTOTYPE
The objective of "upstream" FSR programs is to find out how to
overcome major constraints common to a range of farming systems extend-
ing across one or more geographic zones. The partial or total removal
of a constraint such as water availability in arid areas and soil
fertility in the humid tropics can significantly expand the range of enter-
prises and techniques which can be potentially utilized by farmers.
Such programs mainly contribute to the "body of knowledge", rather
than develop practices specifically tailored to a local situation.
Prototype solutions produced by "upstream" FSR programs must be further
adapted by "downstream" FSR programs to specific local conditions.
Further, "upstream" programs may provide inputs into the establishment
of research priorities for commodity improvement programs, since the
"upstream" perspective is broader in terms of commodities and disci-
plines than commodity improvement programs. And their geographic
perspective tends to be broader than that of "downstream" programs.
Ultimately "upstream" programs should rely on feedback from "downstream"
programs to sharpen their own research priorities or objectives.
Extensive use of experiment station trials often characterizes "upstream"
Most of the "upstream" FSR activities are found in international
agricultural research centers (IARCs). Given the formidable array
of methodological problems involved in "upstream" FSR now and the strong
Type of Research Institution
Activity/Type of Knowledge,
c- L, Cc) -
t- 7n ii -
-< 73 3>-I 3=
-l '-i -l t -4
Component technology research
Research on prototype solu-
tions involving integration
of several components
Operational linkages with
Design/testing of immediate
solutions for local
Treatment of non technical
factors (eg., marketing,
a a a a a a
X X X X X X
X X X X X
X X X X X X
X X X X X
Sx x X X X
X X X
aVirtually all national institutions are engaged in some form of component technology research, although
not generally as part of FSR-type programs.
Figure 3 SCOPE OF ACTIVITIES OF FSR PROGRAMS IN SELECTED NATIONAL, REGIONAL AND INTERNATIONAL INSTITUTES
- -- -- '
comparative advantage of national programs in "downstream" FSR, such
a concentration may still be appropriate. Also, a large geographic zone
of potential applicability can better justify mounting relatively ex-
pensive "upstream" programs because such zones often extend across
national and regional boundaries.
Prominent examples of "upstream" FSR activities include the Farming
Systems Programs of IITA and ICRISAT and the Cropping Systems Program
of IRRI. For IRRI, the key constraint in the rice growing areas of
South and Southeast Asia is identified as land, and the solution is
crop intensification (Technical Advisory Committee, 1978), For ICRISAT,
the important constraint for the semi-arid tropics is identified as
water and the solution is better use of existing soil and water resources,
with the focus on watershed units CKampen, 1979b). For other centers,
it has been difficult to identify constraints around which research
programs could be built and which extend across a large area and
several farming systems. The two centers--IITA and CIAT--serving the
low land humid tropics of Africa and Latin America have had problems
in that regard that stem from wide diversity in farming systems in
their respective zones of responsibility. Difficulties at CIAT in
achieving some focus contributed to terminating the farming systems
program, although some of its activities have been integrated into the
commodity research programs (Technical Advisory Committee, 1978).
IITA is attempting to deal simultaneously with a broad range of con-
straints, including low solar radiation, erosion, drought stress,
intense weed competition, low and declining soil fertility, and seasonal
shortages of labor (IITA, 1979).
"Upstream" features are included in CATIE's mandate, and some in-
depth work on understanding existing farming systems in the Central
American Region has taken place (Hart, 1979a). However, CATIE remains
primarily an institution assisting "downstream" national programs in
its area of responsibility (Navarro, 1979).,
3.2.1. Research on component technology
Although "upstream" programs aim at producing prototype solutions
by integrating several components, much of the research to date has
been on individual components such as soil and water management,
mechanization, and agroclimatology, and tends to be organized along dis-
ciplinary lines--as at IITA and ICRISAT, while cropping systems work of
IRRI and CIAT is related to specific commodities,
Research at IITA and ICRISAT has taken place primarily within
the individual sub-program areas, which deal with specific components,
in part because of the need to assemble and analyze data on basic
factor relationships in the environment, Such work is regarded as a
necessary prerequisite to the design of prototype solutions, At
ICRISAT such prototype solutions are taking the form of improved
systems of soil and water management within Watershed units, ICRISAT
has initiated operational scale watershed-based, resource utilization
research that cuts across sub-program areas (Technical Advisory
3.2.2 Operational linkages with national programs
Since most "upstream" FSR programs are still relatively young,
they tend to be primarily in problem identification and solution design
stages, with limited testing of prototype solutions, mostly at the
research stations. At ICRISAT, specific soil and water management
practices have been tested in watershed units on site, and limited off-
site testing has been done via the All India Coordinated Research
Project for Dryland Agriculture (Technical Advisory Committee, 1978).
IITA has carried out tests of a variety of management practices on
hydromorphic soils on site (Menz, 1979).. Researchers in these programs
are beginning to work systematically with national programs in adapting
prototype solutions for possible eventual use in development programs
for specific areas.
Some of CIAT's farming systems type research is carried out in
cooperation with national programs, as is the case with the Beef Produc-
tion Systems Evaluation Project in the Cerrado of Brazil and
the Llanos of Colombia CCIAT, 19781, However, as noted previously,
the main emphasis of CIAT's farming systems activities is to influence
research priorities within the commodity improvement programs rather
than to design and test prototype solutions,
IRRI's Cropping Systems Program is closely linked with national pro-
grams via the Asian Cropping Systems Network (ACSN), which facilitates
extensive testing of prototype solutions in cooperation with national
programs. The ACSN also serves as a conduit for information on farm-
ing systems in various countries in the region for the Cropping Systems
Program, which assists in determining research priorities. A number of
practices developed at IRRI, centering on means of intensifying rice
cropping systems, have been adapted to local conditions in several
countries and are now being extended to farmers (Technical Advisory
3.3 "DOWNSTREAM" FSR: PROGRAMS FOCUSING ON IMMEDIATE SOLUTIONS FOR
SPECIFIC LOCAL SITUATIONS
As discussed earlier, "downstream" FSR programs begin with an
understanding of existing farming systems and the identification of key
constraints. However, in contrast to "upstream" programs, "downstream"
FSR does not always seek to significantly alleviate key constraints
in the short run, but instead identifies areas of flexibility in the
specific system through accommodating innovations to the reality of
existing constraints.1 In so doing "downstream" FSR, as emphasized
earlier, depends primarily on existing research results for testing and
incorporation directly--or with relatively minor modifications--into
farming systems. On-farm trials and direct or first hand interaction
with farmers predominate while experiment station research tends to be
minimal and restricted to adaptive rather than basic research.
"Downstream" FSR programs form part of the activities of the fol-
lowing institutions--ICTA (Guatemala), ISRA (Senegal), CIMMYT, IRRI,
CATIE, and ICRISAT. Since "downstream" FSR is the focus of other
sections of this review, it is not further elaborated here.
1The example of cotton in northern Nigeria is described in Section
That is, research in the same institution in the FSR program as
opposed to research in commodity programs, which might be primarily or
As already noted, existing FSR programs tend to focus on bio-
technical modifications of farming systems. "Downstream" programs
cover a wide range of approaches, from commodity focused programs to
programs that attempt to develop comprehensive solutions involving
a number of technical factors. The specific concern of IRRI is intensi-
fying rice cropping systems through such measures as shorter season
varieties, reducing turnaround time between crops, and fitting other
crops, such as legumes and vegetables, into annual rotations where
appropriate (IRRI, 1978). Partially because of similarities in the
rice cropping systems in the South and Southeast Asian region and a
narrow range of solutions, IRRI has been able to develop and use a
crop simulation model to best fit cropping patterns with soil and
climatic data. Cropping intensity is less a specific concern of
CIMMYT, which uses a broad range of possible improved practices in
designing solutions for specific situations for wheat, maize, barley
and triticale (CIMMYT, 1976).
ISRA's and ICRISAT's designs of practices extend into developing
a complete alternate farming system, involving several significant
modifications of existing practices, or introduction of new practices/
enterprises. For example, ISRA's improved systems have included
such new elements as animal traction and soil conservation practices
in addition to seed, fertilizer and pesticide (ISRA, 1977).
ICTA and CATIE represent intermediate approaches. Commodity
research priorities influence the geographic focus of FSR work. FSR
research concentrates on developing improved practices for the priority
commodities, but other recommendations will be developed as dictated
by the particular needs of the entire farming system, even though these
recommendations may not involve the selected commodities. ICTA does
not attempt to develop comprehensive solutions, but rather a few modi-
fications at a time, focusing on the key constraints (Hildebrand, 1979c).
One view is that farmers are not likely to adopt a whole range of
recommendations simultaneously, but are inclined to make progressive
modifications of existing practices
See also the discussion in Section 5.2 about single trait and
packages of practices.
There are a growing number of projects which incorporate "down-
stream" FSR type activities. They include the Caqueza project in
Colombia, the activities of Purdue University in the Sahelian countries,
the Michigan State University project in Upper Volta, the Central
Luzon State University/Kansas State University project in the Philippines,
and the Washington State project in Lesotho.
The Institut d'Economie Rurale in Mali recently initiated a
FSR program in the southern region of that country (Institut d'Economie
Rurale, 1977).2 A number of Asian countries, notably Malaysia, Indonesia,
and the Philippines, have started or expanded FSR type programs, in
some cases as a direct outgrowth of their participation in the Asian
Cropping Systems Network. Detailed information on most of these pro-
jects is very limited to date.
The focus of current FSR programs on designing and testing techni-
cal innovations stems from FSR programs emerging from and being currently
located in agricultural research institutions, with mandates restricted
to crops and livestock. Some researchers have included technical
aspects of processing, storage and marketing in their research mandate.
Although the importance of agricultural policy has long been recognized,
most agricultural research institutions have given scant attention
to policy research. But the need to address policy issues is appre-
ciated more now. The Economics Programs at both CIMMYT and ICRISAT
are addressing policy issues in their research, and other institutions
may incorporate policy issues in the future.
Few programs to date give explicit attention to the broader or
macro implications of specific development strategies that may emerge
from FSR programs. For example, a rapid adoption rate of an innovation
may lead to a significant increase in production. How will such an
increase affect prices in the short- and medium-terms? Are market
facilities adequate to absorb the increase? What special measures
The university projects are all supported by USAID.
2Another country which is currently reorganizing its research
structure to incorporate "downstream" FSR activities is Zambia. An
additional intriguing characteristic of the reorganization is their
plan for such activities to institutionalize more firmly the crucial
link between research and extension activities.
might be taken to guard against a short-term disruption of markets and
prices? Some of these questions have been addressed in the course of
certain national programs. The Central Luzon State University/Kansas
State University Technical Package Thrust project in the Philippines
is currently considering marketing specifically in the context of
an FSR-type approach. Several agricultural development projects have
included marketing and macro policy issues in the planning stages,
but examples of such research are scarce.
A qualified exception is the Caqueza project in Colombia (see
4. GENERAL METHODOLOGICAL AND IMPLEMENTATION
ISSUES OF FSR
As illustrated in the discussion of "upstream" and "downstream"
programs in the previous section, a broad range of activities is cur-
rently undertaken in the name of FSR. This variety of activities stems
in part from the holistic nature of FSR which involves a concept of
the "total" environment. Thus there is little activity concerned with
agricultural and rural development which cannot claim some relationship
with FSR, however tenuous.
Further, the breadth of activities included in FSR underlies both
the growing consensus about its desirability as well as the considerable
diversity of opinion about how it should be organized and undertaken.
The diverse opinions involve practical issues of methodology, implemen-
tation considerations, and resources available for research, which
individually or collectively may require some modification of the con-
cept of the "total" environment. Instead of assuming that all factors
determining the actual farming system can be potential variables,
operationalization of FSR may favor treating some or most factors as
parameters (Winkelmann and Moscardi, 1979),1 For example, the mandate
of a particular institution and the availability of research resources
may necessitate focus on a narrow range of variables such as agronomic
practices for one or two commodities.
In the following sections we examine a range of issues affecting
the focus and content of FSR programs, including mandates of institu-
tions, linkages among research and implementation agencies, professional
and practical credibility, efficiency and accountability of the research
process, selection of constraints and evaluation criteria. These
Zandstra (1979b) has expressed an analogous approach with respect
to cropping systems work in which plant growth and crop yield (Y) can
be considered to be the result of two multidimensional vectors: the
environment (E) and management (M).
Y = f (M, E)
In this relationship E are environmental factors (parameters)
that affect Y but are not subject to modification by M (variables). It
is in essence a default relation and reflects the researcher's decision
concerning the mix of M to E,
issues have their roots in two basic characteristics of "downstream"
FSR, First, FSR focuses on solving problems of small farmers--that is,
it is development oriented CNavarro, 1979-) and, second, it focuses on
adapting and using existing improved technology--putting something
together that can be used today--as opposed to science which involves
pushing back the frontiers of knowledge,2
4.1 MANDATES OF INSTITUTIONS
The mandates of the institution in which the FSR program is
located are obviously important in determining the scope of the FSR pro-
gram. For example, the mandate of IRRI requires a focus on rice crop-
ping systems, The methodological approaches chosen in such instances
will be sub-system specific to some extent. However, the research
mandate of a particular institution being somewhat narrowly focused
does not mean that a farming systems perspective cannot or should not
be used. An example of a broad mandate is ISRA's work on the integra-
tion of crops and livestock in Senegal. Also ICTA's research on the
crop sub-system in Guatemala has been broadened to include pigs. To
examine how a particular crop or specific improved practices for that
crop can fit into actual farming systems is an example of a more
restricted approach. An example of this would be the corn and wheat
work undertaken by the Economics Program at CIMMYT in Mexico (Byerlee
et al., 1979).
Since most FSR is underway in crop research institutes, it fol-
lows that FSR methodology is most advanced for the crop sub-system.
FSR programs involving the livestock sub-system have received relatively
little attention and off-farm production sub-systems have been ignored.
Most FSR programs are located in agricultural research organizations
that are committed to increasing production by developing improved
technologies. Unless linkages are well established with development
Science as defined here is more characteristic of the traditional
type of research and "upstream" FSR programs, both of which are involved
in creating the body of knowledge,
2McDermott (personal communication).
agencies Csee below), or the FSR project is located in a development
project or planning units or in "neutral" territory such as a uni-
versity,2 it is unlikely that nontechnical issues--such as policy and
institutional questions--will be satisfactorily addressed.
Because of the potential scope of FSR and the interdependencies
among the various stages of FSR (Figure 2), linkages become highly
significant in determining the success of the FSR approach. Although
this is a critical implementation issue, it also has important conno-
tations for methodology,3 The methodology used will be influenced by
the linkages the FSR program has with other research projects, both in
and outside the institution where it is located; commodity improvement
programs; policy making and rural development planning agencies; and
farmer contact agencies that include development projects. For example,
methodologies used should help articulate research priorities for other
research (Byerlee et al., 1979) and "upstream" FSR programs. Strong
links with other institutions can in essence widen the scope of the
FSR program and, as a result, make it possible to consider improvements
which may officially be outside the mandate of the institute or project
responsible for a FSR program. For example, in the Caqueza project in
Colombia the FSR group worked with credit institutions serving the pro-
ject in designing schemes to deal with farmer risk aversion (Zandstra,
Swanberg et al., 1979). However, linkages can increase the methodological
One of the few examples, as mentioned earlier, of such a program
is the ICA involvement in the Caqueza project in Colombia (see Appendix
2An example, also mentioned earlier, is the Central Luzon State
University/Kansas State University project in the Philippines, which
is looking at the whole food system: production, marketing and pro-
3Section 6.1. includes discussion of linkages from an implementa-
complexity of "downstream" FSR, since they tend to increase the ratio
of variables to parameters in the research program,
Linkages with extension services, delivery system agencies Ce.g.,
credit, fertilizer, etc.) and, where they exist, the management of
development projects, can be very important in determining both the
effectiveness of existing support systems -- external institutions in
Figure 2--and anticipated changes in the future. Incorporating in
the methodology a capacity to evaluate the support systems can be
important as an input in designing and testing potentially relevant
improved practices, Evaluation also is important, where linkages with
policy making agencies exist, in developing more appropriate develop-
mental strategies, In most countries micro-level information for
policy analysis is scarce, Therefore detailed information generated
through the FSR approach could be important for identifying changes
in policies that would complement the introduction of improved prac-
Since the FSR approach in the developing world has gathered momen-
tum only during the 1970s, credibility problems remain in both profes-
sional and practical senses,
Unlike the results of the Green Revolution, the results of FSR are
likely to be less spectacular because of the step-by-step modification
rather than a transformation of the farming systems. As a result the
credibility FSR achieves is likely to be heavily influenced by how
efficiently research funds are used Csee next section), Also, the prac-
tical nature of FSR may reduce peer respect and make it more difficult
An interesting example of this in ICRISAT has been provided by
Ryan Cpersonal communication). Researchers at ICRISAT have involved
bankers in the testing stage of their FSR program with a view to ob-
taining their assessments of the feasibility of the soil and water
management technology and, in particular, the prospects for loans to
finance items such as the tropiculteur. As a result, it is now an
approved item for credit in the Indian banking system,
2n aggregate the benefits of FSR may be significant due to large
numbers of farming families adopting the changes,
to recruit scientists to pursue FSR CNavarro, 1979), Finally the inters
disciplinary nature of FSR work causes problems related to the kind of
results considered publishablee". Often, "good" agronomic research is
that which produces a low coefficient of variation. An agronomist
setting up a program of field trials would, therefore, tend to favor
fewer trials and more replications per trial, An economist, on the
other hand, to achieve results representative over a wider area, would
tend to favor more trials and few replications--given limited research
4.4 EFFICIENCY OF THE RESEARCH PROCESS
Collinson 1979a) has contended that the major problem facing FSR
is funds and manpower too limited to deal with a large number of
farmers.2 Because of the specificity of FSR, with respect to both loca-
tion and stage of development of farmers (Harwood, 1979b), each FSR
effort deals with limited numbers of farmers and, therefore, appears to
be relatively expensive. Further, there is often a time lag between
the recognition of a problem, the finding of a relevant solution, and
its adoption by farmers, particularly where there is not an array of
readily available solutions which can be drawn from the "body of
knowledge" CFigure 2).
FSR is often perceived as being very expensive by researchers not
engaged in FSR and by funding agencies, in part because there are strong
vested interests in maintaining the status'quo of present research pro-
grams, Thus, unless payoffs from reorganization are perceived as being
high, it would be difficult to shift resources to FSR. Sunk costs and
low returns from past research endeavors are likely to be heavily dis-
counted or even ignored,
1E. Crawford (personal communication) citing Barker.
2For example, Ryan and Binswanger C19791 have calculated that in
the Semi-Arid Tropics research expenditures--presumably per year--
amount to only 0.008 cents per hectare of geographic area and only
0.14 cents per hectare of the five ICRISAT crops--sorghum, pearl mil-
let, pigeon pea, chickpea, and groundnut,
It would seem that the most logical way to compare the relative
merits of FSR programs and research programs of a more conventional
nature is to look at the costs in relation to returns, This is, of
course, an empirical question. Although we hypothesize that "downstream"
FSR will have a higher benefit/cost ratio in helping small farmers
than commodity and disciplinary research approaches, we are not sure
of the relevancy of the question. For reasons discussed earlier we
believe the two approaches are more complementary than competitive (see
In estimating the returns from FSR the obvious criterion is
measurement of the improvement in the welfare of farming families,
Measuring rural welfare, however, is very difficult. For example, "down-
stream" FSR may directly or indirectly increase the welfare of farming
families--indirectly by reorienting research priorities of other research
programs so they later contribute to increasing farmers" welfare.
Unfortunately, the potential of such feedback is often ignored in
evaluating "downstream" FSR contributions, possibly because it is
difficult, if not impossible, to quantify.
In spite of considerable potential benefits, efforts to reduce the
time and costs of producing FSR results are necessary if this approach
is ever to be applied to a significant portion of the farm population
in the LDC's.2 Three important principles are emerging in designing
cost and time efficient methodologies. They are:
(1) Reducing of time required to move through the four research
stages. The methodologies applied, in addition to ensuring
a fast turnaround, need to be practical, replicable and in-
expensive (Byerlee et al., 1979). Complex procedures that
require highly qualified individuals to collect and analyze
data and to design and test solutions, need to be avoided as
An alternative way of viewing the welfare discussion would be
immediate Cdirect) or future Cindirect) changes.
2This is particularly important since, as others have pointed out
CMenz and Knipscheer, 1979), "downstream" FSR raises the opportunity cost
of neglecting farmers not in the specific target groups. Lowering
explicit costs for specific target groups would enable work to be
undertaken with more target groups.
much as possible (Zandstra, 1979al,1 There are, however,
limits to reducing the length of time required to obtain
(2) Maximizing the return from such research by making results
more widely applicable. The extent to which improved sys-
tems can be transferred or extrapolated to other areas
directly affects efficiency.
(3) Using "second best" or "best of readily available solutions".
Traditionally research in agriculture has emphasized the
concept of developing optimal practices. When one considers
the heterogeneity existing in the "total" environment, how-
ever, costs in terms of finance and time to obtain optimal
recommendations for each type of variation would be astrono-
mical, Therefore, increasingly the emphasis of FSR is on
developing improved farming systems that are better but not
necessarily best, for each environment, In other words, the
process is "non-perfectabilitarian" and does not envision
developing optimal improved practices (Winkelmann and
4.5 WHOSE CONSTRAINTS SHOULD AN FSR PROGRAM CONCENTRATE ON?
The key to developing relevant strategies for improving the welfare
of farming families involves first obtaining information on the farming
systems practiced in terms of what is done and why it is done that way.
That information can help indicate the flexibility--for example, when
there are slack resources--and constraints that exist in the current
systems, Needs or constraints can be identified at three levels:
Cl) Those specifically mentioned by farming families,
(2) Those identified in a scientific manner by FSR workers,
(3) Those reflecting the interests of society as a whole,
This is true both for cost efficiency and replication. Skilled
personnel are characteristically in short supply in most LDC agricul-
tural research organizations,
Those specified by farming families themselves may be only what
they think can be solved with outside help CNair, 19611.1 Also, if
they are living near the survival level they may have a short-term
horizon and their expressed needs may conflict with the interests of
society as a whole. If conflicts exist between the two sets of needs,
in a society where voluntary change on the part of farming families is
permitted, societal needs are not likely to be met.
The constraints or needs identified by FSR workers are, by the
nature of FSR, likely to reflect needs of the farming families them-
selves. But because of their position, the researchers are more likely
to consider the potential societal impact of fulfilling farmers" needs,
Maximizing yields per hectare to satisfy short run private interests
at a long run cost to society by an irreversible drop in soil pro-
ductivity, for example, would hopefully be recognized by FSR researchers,
Their skill lies in devising strategies that meet the expressed needs
of farming families without exacerbating constraints of direct rele-
vance to such families but not explicitly mentioned by them, Also
they need to be sure their improved strategies do not violate the
interests of society as a whole. Unfortunately, this is easier said
than done. Because the model of FSR articulated to date has been based
on the individual farming family, the link to societal needs has not
been well established3--either conceptually or operationally. The
In a survey in Kenya, Shaner Cpersonal communications found
farmers tended to ask for those items such as schools, clinics and
roads which they thought government might be able to supply rather than
priority items based on their overall appraisal of needs.
2Nair's (1979) recent work is a good example of the need to adjust
government policies to bring about a convergence of private and societal
That, perhaps, is inevitable as most FSR work has been undertaken
in technical research institutes by technical and social scientists.
The latter, either because they feel they have neither the mandate to do
macro research nor the influence to change policy and the external
institutional environment, have focused almost exclusively on the micro
issue of understanding and changing farming systems within the present
environment CByerlee, personal communication). Interestingly, some tech-
nical scientists are now urging the social scientists to change the
policy environment so that they have less constraints in their own work.
mandates of the institutions with FSR programs and the linkages with
government and developmental agencies, will influence how much atten-
tion is likely to be placed on the link between private and societal
interests and whether potential conflicts may be resolved easily. We
believe that these micro-macro linkages will be of critical importance
in determining the long-term viability of FSR programs. Possible
broad societal concerns, which are handled by the commodity programs
and "upstream" FSR programs of research institutes in the case of
technical matters, and planning and policy making bodies in the case of
nontechnical matters, might be taken into consideration by having
these agencies pre-screen potential improvement strategies for compati-
bility with societal concerns.
The needs or constraints that arise from an investigation of
individual farming families may be technical, economic, or socio-
cultural. Several approaches are used in dealing with such con-
The first involves accepting the constraint and developing strate-
gies that exploit any flexibility in the current farming system with-
out further exacerbating the constraint. We think little can be or
should be done to change socio-cultural constraints unless they are
debilitating society--for example, deepening societal inequities.
Ways need to be devised to help improve the welfare of such groups of
people in a manner compatible with the constraints. For example, no
one should try to introduce hogs into Muslim areas.
The second approach to dealing with the constraints is to develop
improved strategies that will overcome the constraints, as is commonly
the focus of "upstream" FSR. For example, the FSR program at ICRISAT
is attempting to alleviate the water constraint through the development
of improved systems of soil and water management centering on watershed
units (Appendix A1.2). The removal or significant alleviation of con-
straints has to be viewed from more than simply the perspective of the
individual farming family. In India, for example, breaking of a labor
bottleneck period through mechanization and herbicides could have serious
consequences for society by decreasing employment opportunities of the
landless, laboring class.
The decision on which approach to use in dealing with constraints
will depend on their severity, the flexibility that exists in the
existing farming system, and the availability of potential improved
strategies that break the constraints or exploit the flexibility.
4.6 CRITERIA FOR EVALUATING IMPROVED SYSTEMS
It is important to evaluate the improved systems from both the
individual farming family's and society's point of view. The simplest
way to evaluate whether improved systems are suitable or relevant from
an individual or private perspective is to ascertain whether they are
adopted by farming households. Suitability can be assessed in an
ex post sense through various methods of acceptance such as adoption
indices. However, evaluating suitability that way creates two major
(1) To improve the efficiency of the FSR approach it is essen-
tial to use evaluation criteria that assess the potential
suitability of the innovation both for individual farmers
and the society as a whole.
(2) Additionally, the adoption indices give no indication as to
why some farmers did not adopt the improved system.
Both problems have important implications for developing suitable
In assessing whether the improved practices are potentially
suitable from the point of view of the individual farmer or farming
household, we suggest dividing the evaluation criteria into three groups
corresponding to the technical element, the exogenous factor, and endo-
genous factor. The first two constitute necessary conditions for the
adoption of improved practices--in other words, whether the farmer can
adopt it, if he is willing to.1 The endogenous factors, on the other
Although we broke the evaluation criteria into distinct groups,
we recognize that they are not mutually exclusive. For example,
willingness to adopt a particular technology will be partially deter-
mined by ability to do so. Also the profit--level and dependability--
of an improved system, which we consider a sufficient condition, will
be partially determined by the external institutions--such as prices
for the inputs and market for the product produced--which constitute
part of the necessary conditions.
hand, can be considered as providing the sufficient condition for adop-
tion; that is, they determine whether the farmer is willing to adopt it.
The necessary conditions for adoption of improved technology can
be specified by three evaluation criteria; technical feasibility,
societal acceptability, and compatability with external institutions
or support systems. The relative significance of the last two criteria
depend on the stage of development of agriculture in the area, and the
type of improved practices considered, Increasing contacts outside the
village and increased commercialization of agriculture--resulting in
increasing significance of economic forces--likely make social accepta-
bility in the community relatively less significant, while an appro-
priate support system becomes increasingly critical, The distribution
system must be able to provide the inputs required for adoption of the
improved technology, and a market for the product produced must be
Obviously, the improved practices must be compatible with the goal
or goals of the potential adopter, The objective function of farming
families likely will change as they move from self-sufficient subsis-
tence farming to commercial farming, In the case of the former, under-
standing the goals) may be a particularly complex task while in the
latter they are probably much easier to articulate--for example, as pro-
fit maximization, Most farming families, on the continuum between
the two extremes, are likely to have a hierarchical ranking of goals.
A commonly suggested ranking is food self-sufficiency first, then profit
maximization after food needs are met.2 The latter goal is easier to
examine by assessing profitability expressed in terms of the most
This is a critical issue in many situations and requires analysis
at the macro level. For example, Vincent (personal communication) em-
phasized that in one area of the Philippines an attempt was made to
help cabbage producers obtain higher prices for their product by con-
trolling the production of cabbage over time. While this was taking
place, farmers in another cabbage producing area took up the slack by
expanding production. This is a good example of the desirability of
an FSR program embracing not only production but also processing and
2One could argue that the participation of the farmer in the
research process will to some extent compensate for a complete and
detailed understanding of his/her family goals,
limiting factor of the improved practices compared with the those they
are designed to replace. Because of the relatively low living levels
and the desire for food self-sufficiency, avoiding risk by ensuring
dependability of return from an innovation should be an important
evaluation criterion CNorman and Palmer-Oones, 19771, For example, if
the improved practices can be proved to Be more profitable and as
dependable as those they replace, they are likely to be attractive to
Until now we have concentrated on evaluating the Improved systems
from the perspective of individual farming families, However, as
emphasized in the preceding section, attention also needs to be given
to its acceptability from a societal point of view. For example, if
food production were to decline, or if the technology adopted were to
result in degrading the natural resources base, or if increased in-
equality of income distribution were to arise, then short run private
returns would come at a long run cost to society. If at all possible,
divergence between private and societal interests needs to be avoided.
Unfortunately, looking at the improved.systems from a societal point of
view requires looking into the future--sometimes farther than the short
run, so uncertainty complicates the evaluation problems CFlinn, 1980).
The micro-macro linkages stressed earlier are very important but they
remain the weakest part of FSR programs. Because current FSR programs
concentrate on individual farming families, it is very difficult
operationally or even conceptually to link evaluation from the societal
point of view to evaluations for individual farming families. Such
linkage might take the form of the pre-screening of potential improve-
ment strategies by research institutes and planning agencies as
Profitability as a concept can be applied to production destined
not only for the market but also for home consumption. In the latter
case, the product price is what it would be necessary to pay to pur-
chase the product.
2We use societal as inferring some degree of aggregation of farm-
ing families. For example, in the current discussion it could mean the
community in which the farming families are located or the nation as a
3Michie Cpersonal communication)
suggested earlier (Section 4.5).
Currently, societal evaluations tend to be based on separate studies
that use aggregate measures and are often ex-post rather than ex-ante.
The micro-macro linkages need much more attention by researchers.
5. METHODOLOGY OF FARMING SYSTEMS RESEARCH
In this chapter we examine methodological issues involved at each stage
of the process: descriptive and diagnostic, design, testing, and extension,
with emphasis on "downstream" FSR. The way specific methodological issues
are resolved will depend on how the general issues outlined in the preceding
chapter are resolved. Since this chapter focuses on current FSR in the Third
World, most of the examples are drawn from research on the cropping sub-system.
5.1 DESCRIPTIVE AND DIAGNOSTIC STAGE
The objective of this stage is to pick target areas, describe the
present farming systems, ascertain major constraints on farming in the
area and discover the degree of flexibility in modifying the farming
5.1.1 Selection of the target area
The following three points need to be considered in selecting the
(1) A FSR program should not be implemented if it is incompati-
ble with government needs and priorities (Asian Cropping
Systems Working Group, 1979). Still, accepting government
priorities might lead to problems. For example, if develop-
ment priorities are designed to help the more commercialized
areas, these areas may not require FSR.
(2) The problem of obtaining credibility in reasonable time,
especially when research resources are scarce, means a bias
towards selecting an area not only consistent with national
development priorities but also one where tangible results
are potentially possible in a short time CNavarro, 1979).1
1For example, the Asian Cropping Systems Network picks areas desig-
nated as receiving priority in national development plans and that have
both potential for increased production and cropping intensity and an
adequate infrastructure support system (Zandstra, 1979b).
C3) The broader the target area the greater is the potential to
spread costs. Concentrating on small, unrepresentative areas
is likely to reduce the potential multiplier effect of FSR.
The criteria for delineating boundaries of the target area also
may be affected by political issues. The target area, for example, may
be demarcated by an administrative or political boundary and may
embrace a wide variety of farming systems. Boundaries delineated by
development projects may be useful in some circumstances as a compro-
mise because they have reasonably uniform farming systems.
In practice the procedure that often gives satisfactory results is
delineating an area where the majority of farmers follow similar
agricultural practices or a similar farming system (Hildebrand, 1979c).
Sometimes, however, in assessing the physical potential for parti-
cular enterprises--crops or livestock, for example--it is important to
delineate the target area on the basis of the characteristics of the
technical element or agro-climatologic features (Zandstra, 1979b).
5.1.2 Baseline data analysis
Baseline data analysis involves using available information.
In view of the time and cost of collecting primary data, available
secondary information should be exploited. Secondary data can be use-
ful in delineating the target area and in obtaining a preliminary under-
standing of existing farming systems.
The criteria for data to be used in baseline data analysis and in
the collection of data from on-farm studies should be the relevance
of the data in understanding existing farming systems, particularly
their constraints and flexibility and how to modify present systems
CTechnical Advisory Committee, 1978). Good data on the technical ele-
ment, particularly on such physical factors as land resource classifica-
tion and weather and climatic characteristics, can he particularly
1Mali, for example, is divided into development areas, each with
its own organization, delivery systems, etc., which emphasize different
crops--groundnuts in one area, cotton in another, etc.
valuable,1 It has been suggested that wherever possible, existing
methods of analysis should be used to classify soil, land, and climate
CTechnical Advisory Committee, 1978),2 Some information on various
exogenous factors can often be gleaned from reviewing secondary data
sources, but the common lack of detailed micro information usually means
that basic data concerning the endogenous factors are not available,
Therefore, as a rule, endogenous type data will be obtained through on-
The quantity and quality of secondary data available will determine
how well the objectives of the descriptive stage are achieved as a
result of baseline data analysis. The poorer the data and informational
base is, the more research at this stage becomes an art rather
than a science, and the more on-farm studies are needed to describe and
diagnose the area's characteristics and constraints.
5.1.3. On-farm studies
On-farm studies are important in disaggregating the target area's
environmental heterogeneity, Such studies should classify farming
families into homogeneous sub-groups or "recommendation domains" (CIMMYT
Economics Program, 1980). The sub-groups provide a focus for develop-
ing relevant strategies to improve their welfare, Effectively delinea-
ting such sub-groups depends on being able to isolate the factors in-
fluencing variation between groups of farmers and adopting a classi-
fication method that effectively weights the influence of the factors.
CIMMYT J1979) has suggested two types of divisions: "a locational
division by area and a hierarchical division between farmers in the
The international and regional agricultural research institutions,
with their resources and ecological focus, are in a good position to
set up data banks on such information, A number are now doing this
because such information has uses far beyond the specific needs of
2n recent years, however, IRRI has increasingly emphasized more
efficient ways of interpreting land and climate as they relate to
production alternatives CZandstra, Angus, and Tamisin, 1979; Angus and
same area." Three sets of factors are identified as contributing to
(1) Natural factors; climate, soil, topography.
(21 Historical factors: food preferences, social customs, pre-
sent technology, and tenurial arrangements.
C31 Institutional and economic factors: access to markets and
While sets (2) and C3) are relevant to both locational and hier-
archical divisions, set (1) is relevant only to locational divisions.
Dividing farmers into homogeneous sub-groups is a complex process and
includes consideration not only of differences in the technical element
but also of variations in the human element--which traditionally have
often been ignored.1 As a result of this classification, farming fam-
ilies in a particular sub-group will tend to have similar farming acti-
vities and include similar social customs, similar access to support
systems, comparable marketing opportunities, and similar technology
and resource endowment (Collinson, 1979a). Farming families
'within each specific sub-group should have the same problems
and development alternatives and should react in the same way to
Two major methods are generally used to obtain the necessary data
from on-farm studies to finish classifying farming families into homo-
geneous sub-groups: reconnaissance CHildebrand, 1979a) or exploratory
surveys (Collinson, 1979b), and formal surveys (Collinson, 1979a).
The reconnaissance or exploratory surveys are informal and consist
of field tours or sondeo (Hildebrand, 1979a). Multidisciplinary teams
1Economic and institutional factors heavily influence what will be
grown. Because they are perishable, vegetables will be grown near
urban markets even though technical factors in more remote areas may
be more suitable for growing them,
However, such divisions between the different sub-groups are
artificial, since interdependencies and interactions are likely to
exist between them, In evaluating the ex ante societal consequences
of improved strategies, it is important to understand such relation-
3These two types of surveys are complementary rather than competi-
tive, The former should nearly always precede the latter.
working in an interdisciplinary framework travel throughout the target
area talking with representatives of policy-making, farmer-contact
agencies and with community leaders and farm families. Such discus-
sions are to delineate sub-groups of farming families and to analyze
current farming systems and possibletypes of developmental strate-
gies potentially useful to farming families and consistent with their
goals. The exploratory surveys require interaction not only with
people in the target area but also among members of the FSR team
(Collinson, 1979a; Hildebrand, 1979a). Many FSR practitioners believe
the process leads to a partial but useful impression of the entire
farming system and helps classify farming families into sub-groups.
The extent to which these reconnaissance surveys can be carried out--
6 to 10 days in the case of the sondeo (Hildebrand, 1979a)--is
largely a function of the experience of the team in FSR and their
familiarity with the target area. Then more formal structured farm
surveys often are carried out among the target population to verify the
tentative insights from the exploratory survey. The surveys involve
trade-offs between cost and time efficiency on one hand and accuracy
on the other. For those concerned about efficiency, the formal survey
consists of a single interview with a representative sample of farmers.
Emphasis on accuracy, in contrast, calls for frequent interviews over
a long time--usually one year, particularly for data that are contin-
uous and non-registered such as labor flows, in contrast to those that
are single-point, registered in nature such as purchase of fertilizer
(Collinson, 1972; Lipton and Moore, 1972; Norman, 1977),
Single visit interviews of a large number of farmers are increas-
ingly being undertaken to minimize sampling errors. Such surveys can
be complemented by more frequent interviewing of a limited number of
farmers in order to minimize measurement errors. The frequent inter-
viewing approach (Hart, 1979b) is usually carried on concurrently with
later stages of the FSR program, Particular emphasis is usually
placed on including farming families who participate in the testing
stage of FSR, A combination of single interviews and frequent
CIMMYT calls the single-visit, formal survey a verification
interviews has these advantages:
1) It minimizes the delay in moving from the descriptive to the
design and testing stages, although some gamble is involved
in the sense that the needs or constraints emerging from the
in-depth study may not confirm the results from the earlier
single-interview survey that were fed into the design and
C21 It provides accurate quantitative information for comparing
results of the existing system with results of the improved
system, which is particularly useful during the testing
stage. Sometimes, in compromise, information is collected
on only the part of the farming system of direct interest
to the research mandate of the institution undertaking the
FSR program (Hildebrand, 1979b).1
Four basic methodological issues are involved in speed and effi-
ciency of carrying out the formal survey:
(11 Various sampling methods are available for selecting
farming families for study, The time involved and method
used will depend on whether or not a stratification procedure
is to be adopted and whether frames of farming families to
draw samples from are available (Bernsten, 1979; CIMMYT
Economics Program, 1980).
(C2 Criteria for collecting data often are poor.2 Too often the
criterion for the way in which data are collected is the ease
with which it can be collected accurately, rather than the
need for it to be collected accurately CCollinson, 1972),
The decision on how data should be collected should be based
Further efficiency in terms of cost is sometimes possible with
farming families keeping such records themselves (Hatch, 19801.
Literacy has permitted that in Guatemala CICTA), Philippines (IRRI).,
India CICRISAT), and elsewhere.
2Modeling through the use of simulation techniques has also some-
times been used to obtain an idea of the critical variables and hence
an indication of variables that need to be measured accurately. For
example, Brockington Cpersonal communtcationI has used such techniques
in looking at the dynamics of cattle-herd structures in Brazil.
on lowest cost commensurate with the understanding that is
necessary. Direct measurement techniques--like quantify-
ing actual seasonal labor flows--require expensive techni-
ques such as frequent interviewing over long periods.
CollinsonI suggests that the decision as to whether measure-
ment is necessary should be based on such considerations as
whether it improves understanding sufficiently and consistently,
whether it improves understanding enough to justify the extra cost
of measurement, and whether it improves understanding enough in
the light of opportunity costs forgone, such as working with
more sub-groups of farming families.2
(3) Related to the above and to efficiency is the idea that too
much emphasis may be placed on quantifying and too little on qual-
itative data. Qualitative information should include not only
attitudinal information but also types of data not essential
according to the criteria specified above. Limiting quanti-
fication to key characteristics reduces costs involved in
(4) Too often little consideration is given to increasing the
efficiency of the data collection-analysis link in surveys.
For frequent-interviewing surveys, processing should start
while data are still being collected. Also, all surveys,
no matter how they are undertaken, need to be designed to
facilitate quick processing--such as ease of transferring
data to computer-based systems.
Collinson (personal communication).
The same considerations apply to modeling techniques which are
based on accurate measurement. However, simulation techniques may be
useful under certain circumstances (E. Crawford, 1980).
The recent technological breakthroughs with pocket calculators
and mini-computers have increased the ease with which this can be done.
Purdue University is, for example, using mini-computers in the Sahelian
countries, and Michigan State University is using them in northern
5.2 DESIGN STAGE
Priorities for research should arise from the descriptive stage,
in terms of developing improved practices based on the needs of farm-
ing families and constraints they face. The design stage should pro-
duce a few sets of improved practices for testing at the farm level,
Collinson C1979a} suggests the following procedure for designing
Cl) The experimental variables should consist of practices in
which farmers' management is flexible and those where ex ante
evaluation suggests room for increased productivity.l
Flexibility in management is enhanced when there are under-
utilized resources, while increasing productivity of variables
is particularly important for those resources that are most
(2) The feasible range of treatments for such variables is set
by the flexibility that exists. Some flexibility could be
introduced, for example, by assuming the institutional sup-
port system could change--that is, be a variable rather than
a parameter, It could, for example, be assumed that an
institutional source of credit could be made available to
supplement the cash flow of the farm business. The above
remarks suggest that the development of improved practices
should usually consider the existing or definitely expected
infrastructural support system, However, that is now being
debated in some centers.
(3) The parameters in the experimental process should be those
not potentially subject to manipulation and as representative
1Such practices can be ascertained from investigating what is
available in the "body of knowledge" established as a result of commo-
dity research and "upstream" FSR programs,
Currently at CIMMYT considerable debate centers on how much they
should develop improved technologies given the existing infrastructure
support system and how-much they should be trying to change that
system as well (Byerlee, personal communicationJ.
as possible of practical farming conditions,1
The design stage is primarily implemented under station condi-
tions, Experimentation, according to the above specifications, is
essentially "downstream" FSR, Where the "body of knowledge" is not
sufficiently developed to provide adequate material for the design
stage of the "downstream" FSR program, relaxing the above experimental
constraints may be justified so an "upstream" FSR program may be ini-
tiated, To date much knowledge has been accumulated through the reduc-
tionist approach, usually without a systems focus, When interaction
is likely to be important--for example, in watershed management, soil
fertility, and mixed cropping, the above specifications may be relaxed
to build up the "body of knowledge" through an "upstream" FSR program,
Unlike the other stages of FSR, research methods for work on
experiment stations are somewhat better established (Technical Advisory
Committee, 19781. Usually conventional approaches can be used. How-
ever, complications are introduced when the research has more of a
systems focus and the ratio of variables to parameters is increased--
as is true in some "upstream" FSR programs,
In the design stage of "downstream" FSR programs the following two
issues have important methodological connotations:
01 The ever-present problem of minimizing costs of research has
two dimensions at the design stage.
Cal Computer modeling and simulation can have definite
advantages in meeting time limits, as with livestock
where relatively long cycles are the rule and where
topics such as rainfall/water balance/crop growth simula-
tion models help identify alternatives.3 However,
This may, as stressed earlier, also be determined in part by the
research mandate of the institution with the FSR program and the feasi-
bility of dealing with many variables,
2Hildebrand Cpersonal communications suggests that one should be
cautious about drawing a fine line between design and testing, In fact,
some design work can, and does, take place in trials at the farm level.
31LCA, for example, uses thts approach in their research on live-
stock, while ICRISAT applied it to their watershed management work,
complicated and nonstandardized modeling needs to be
undertaken cautiously with a full understanding of its
potential dangers (Technical Advisory Committee, 19781,
Too often such analytical tools have been used as a
substitute for, rather than a complement to, work of a
more pleblan nature, As a result they have often become
overly complex, expensive, and out of touch with
(b) Since ceteris paribus conditions are much greater on the
experimental station and the human element cannot
adequately be taken into account except as an input in
the initial experimental design, practitioners in
"downstream" FSR have reservations about spending much
research effort working on the experiment station.2
Generally, the greater the "body of knowledge", the
shorter is the time required on the experiment station to
complete the design stage in "downstream" FSR.
(2) Developing improved practices may involve incremental or
"single trait" changes instead of packages of practices.
Numerous studies have shown that where packages are intro-
duced, various components are adopted to various degrees
(Gerhart, 1975; Hildebrand, 1979c). The major advantages of
packages, of course, include the complementary or synergis-
tic effects or relationships among components. For example,
improved seeds respond better than indigenous varieties to
inorganic fertilizer. The major disadvantage of such pack-
ages involve complications due to the more complex methodo-
logies needed to put them together and problems or difficul-
ties in getting them adopted by farmers. It is likely to be
more difficult, for example, to convince farmers to adopt an
In addition to the institutions mentioned in the preceding foot-
note, some encouraging developments in simulation modeling are evolving
from work in the UK CSpedding and Brockington, 1976; Brockington, 1979-,
For example, ICTA's work on time devoted to research work off and
on the experiment station is a ratio of nine to one CBildebrand, 1979c),
improved package when few changes as a result of external
factors have been introduced into their farming systems.
Also, packages often imply more complex management
and more complex external institutional support systems,
Assuming ceteris paribus conditions, single trait changes
are obviously preferable (Bartlett and Ikeorgu, 1979). How-
ever, in theory at least, where single changes come at too
high a cost--private or social--as a result of Ignoring
synergistic effects, then packages of improved practices
should be developed. Accordingly, it could be argued that
improved packages of practices are likely to be the rule
rather than the exception CRyan and Subrahmanyam, 1975).
But in practice it can, and perhaps should, be argued that
because the farmer is unlikely to adopt the package in its
entirety, using an incremental approach is justified. In
other words, initial extension work might emphasize one or
two components with the rest to be added later. The crite-
rion for such an approach is that the changes being intro-
duced should be as many and as big as possible so long as
the farmer finds them acceptable.1 Instead the current
emphasis on packages has tended to result in offering, at
the design stage, only two possibilities--rejection or com-
plete acceptance of the whole package. Later the farmer
decides which parts of the package to adopt if initially he
can't accept the entire package. Byerlee believes that
research via the extension staff can provide such information
more efficiently than the usual approach described above, in
which a lot of valuable information is withheld from farmers
who discover it the hard way later.
1Collinson Cpersonal communications,
5.3 TESTING STAGE
The objective of this stage is to evaluate the improved practices
flowing from the design stage to the farm. The evaluation criteria
should be the same as those found to be important in the descriptive
stage, The testing stage consists of two parts;
(1) Trials at the farmer's level that use farmer's land and maybe
labor, but with the managerial input still provided by the
(2) Farmer testing with farm families providing their own land,
labor, capital,l and management. In essence the improved
technology is tested for compatibility with the technical,
exogenous, and endogenous factors.
Usually performance of the improved technology drops when it
moves from the somewhat artificial conditions of the experiment station
to trials at the farm level,2 and drops again at the farmer's testing
level where the improved technology is in effect being tested for com-
patibility with the current farming system and the managerial know-how
of the farmer.3
Two critical issues, both with important methodological connota-
tions, arise at the testing level:
(1) The issues of interaction between farmers and research wor-
kers,4 and the representativeness of farmers and farming
The amount of capital anticipated here is that already available
plus what could be derived currently or in the near future through
external institutions. Realization of the latter may require credit
or other inputs to be provided by the research organization in the
This corresponds to yield gap I as defined in the Rice Constraints
Studies undertaken at IRRI (JRRI, 1977),
This corresponds to yield gap II in the IRRI studies.
This interaction involving participation of the farmers has
important implications for the research process. The example given by
Collinson Cpersonal communication) cited in Section 2.4.3 indicates
just what it means and how significant it can and should be.
families. Conflict between the two desirable characteris-
tics is conceivable. Some research workers prefer to select
the better,l more responsive or more cooperative farmers to
participate in the testing stage, Using the cooperativeness
criterion has the advantage of maximizing interactions
between research workers and farmers. But there is the
potential problem that even when improved practices receive
a positive evaluation, they still may not be truly relevant
for the average farmer. The adoption process may be thus
biased towards farmers with the above characteristics and
cause inequalities in benefits in the long run. Other
research workers in FSR advocate selecting a cross section
of farmers representative of the subgroup or subgroups under
investigation. The possible disadvantage, that representative
farmers would not maximize interactions between farmers
and research workers, is offset by the big advantage of get-
ting a more satisfactory idea of whether the improved prac-
tices are likely to be suitable for the average farmer.
However, the bias that usually--and perhaps inevitably--occurs
is one of including only cooperative farmers at the testing
stage to ensure maximizing interactions between farmers
C21 The issue of transferability, Costs limit the number of
sites that can be included in the testing stage. So efforts
are needed to increase the multiplier effect by extrapolating
results to other areas, Chances to extrapolate or transfer
results to other areas'are, of course, increased if sites
for farm trials are picked to represent large areas, Possi-
bilities for extrapolation are increased by developing
1Shaner Cpersonal communication) has suggested that an advantage
of selecting such farmers is that they are an intermediate step between
the potential returns at the farm level and returns achieved by repre-
2Cooperation can be encouraged through a reward system, but
opinion is mixed regarding such.forms of encouragement.
technologies that are flexible in timing and other fac-
tors.1 At the farmers' testing level a detailed specifica-
tion of the proposed improved practices and conditions
under which they were tested is required to increase the
efficiency of extrapolation to other areas (Zandstra, 1979a).
Such specification should include CNormanand Palmer-Jones,
Ca) Delineation of what was actually done in describing
the proposed improved technology.
(b) Description of the technical environment where the
testing was undertaken, including location, avail-
ability and distribution of water, temperature, poten-
tial evapotranspiration, and soil type (i.e., physi-
cal and chemical properties that are likely to affect
tillage, nutrient and water characteristics, and ero-
(c) Economic specifications detailing output and input
totals and flows where relevant in both quantitative
and monetary terms. Also, ex ante evaluation criteria
should include more than just criteria relevant to the
specific test sites. Such specifications about improved
practices permit one to assess the suitability of the
improved technology to farming families adopting differ-
ent goals, families with wide variations in resources,
and those facing differences in the exogenous factors.
The ways the testing stage is conducted may vary widely. No
attempt is made here to discuss the various approaches that are used,
except in general terms,
Some FSR practitioners suggest that such strategies for Increas-
ing the multiplier effect from extrapolation to other areas should
not he pursued if it involves some sacrifice in refining the improved
technology to the specific area under investigation.
Obviously some of this information would be derived from the.
5.3.1 Trials at the farmerst ; evel,
Trials at the farmers' level, or research-managed trials as they
are sometimes called, can cover more treatments than those at the
farmers' testing stage. At the testing stage, treatments are
usually less complex than those undertaken on experiment stations dur-
ing the design stage because of costs, fields not being big enough
to carry out complex experiments--especially if replications are
involved, and the desirability of having some interaction between far-
mers and research workers. Interactions are less likely when experi-
ments become too complex.
The aim of such trials is to screen the improved technologies
arising from the design stage, to fine-tune them to the local situa-
tion, and to evaluate their potential both locally and for broader
regional coverage. Researcher managed,trials can consist of either
replications within fields or between fields--to check site variabili-
tyl The varied types of farm level trials2 can use experimental
designs similar to designs on experiment stations.
5.3.2 Farmers' testing3
Farmers' testing is the most rigorous test of the proposed im-
proved technologies. Three points need to be considered to derive
valid, useful data for evaluating the improved practices at this
1The pros and cons of each type of replication are discussed by
Bandong et al, (1977),
2For example, CIMMYT advocates three classes of on-farm trials:
yes-no trials, howmuch trials, and verification trials (Winkelmann
and Joscardi, 1979, and Appendix Al.1).
Farmers' testing in the context used here includes the pre-
production testing undertaken in the IRR1 Cropping Systems Program
(Asian Cropping Systems Working Group, 1979; Zandstra, 1979a.
(1) It is important that plots are large enough for the improved
technologies being tested. Labor is an important input, so
plots need to be large enough that labor inputs can be
accurately measured. Consequently, replications within the
field are not usually possible. However, the improved
technology may be replicated on fields of other farmers.
(2) Both the technical and human environments vary widely over
time. Testing for more than one year3 gives a better idea
of the level and stability of improved practices, partic-
ularly where there are substantial inter-annual variations
in the "total" environment,4 In effect, replications can be
increased by incorporating'the time dimension through using
the same improved practices in different years. But such a
replication should not preclude modifying the improved prac-
tices after obtaining results in earlier years.
(3) To provide valid evaluation of improved practices it is
important to obtain data that will assess compatibility of
the practices with other parts of the farming system. Two
IRRI, for example, suggests 100 sq. m. as a minimum (Zandstra,
1979) while Hildebrand (personal communication) advocates that at least
20% of the cultivated area of the farmer's farm should be devoted to
In any case, farmers providing the evaluation are not likely to
be interested in replications.
3Three years of on-farm testing are often advocated (Asian Cropping
Systems Working Group, 1979; Hart, 1979a). E. Crawford (personal
communication) has suggested that when a shorter evaluation period is
necessary, and when manpower and computational resources are available,
simulation offers a worthwhile method to assess the sensitivity of the
improved technologies under different assumptions. That, however,
raises the issue of accurate modeling mentioned earlier (see footnotes in
Section 5.1.31, It will probably be easier to investigate realistic
variations in the technical element than in the human element,
Another approach. that has sometimes been advocated but which has
some obvious problems is testing in slightly different "total" environ-
ments the same year to simulate differences between years.
alternative approaches may be used,1 One is to collect
data on all other parts of the farming system to assess
potential conflicts and compatibility. The alternative
often adopted to minimize costs is collecting data on only
the parts of the farming system that the improved practices
are likely to directly affect or replace, But caution is
needed if the flow and level of resources required to adopt
the improved practices differ substantially from those
required for the practices they are designed to replace,2
5.3,3 Intermediate types of trials
To encourage more farmer-research worker interaction and lower
costs by combining some of the characteristics of farm level trials and
farmer testing, two additional types of trials or experiments are
Cl) Trials superimposed at the farmer testing level; that is,
conducted on the same field where the farmers testing is
being undertaken--trials that reflect several factors
relevant to the farmer's situation (Zandstra, 1979al, For
example, IRRI's Cropping Systems Program specifies that
superimposed trials must include four levels: a simulation
of farmers' management with no purchased material inputs,
the level of component technology assigned to the cropping
pattern, and a level of component technology that will pro-
duce high yields in the cropping pattern, or will produce
similar yields with substantially lower input.
C2) To encourage more of a systems focus, unit farms3 have been used
As discussed earlier (Section 5,1,3), these can be combined
with the formal surveys,
2The example discussed earlier (Section 2,4.5) concerning improved
practices for cotton in northern Nigeria illustrates the potential
problems of such an approach,
3The use of unit-farms, of course, pre-dates the FSR approach
in the Farming Systems Program at TITA (Menz, 19791, They
have also been used to more realistically include the human
element with perhaps much more control than possible under
farming conditions at the village level. Menz (1979) sug-
gested that the degree of control imposed on the unit farmer
will depend on whether the technologies being developed are
still in the design stage, are trials at the farmer's
level, or have more farmer-testing orientation. The
inability to realistically incorporate the human element--
both exogenous and endogenous factors--supports the notion
that unit farms are more suitable in "upstream" FSR pro-
5.4 EXTENSION STAGE
The extension stage, because it provides vital information about
the effectiveness of the improved strategies from the earlier stages
of the FSR process, is an integral part of the FSR program. In addition,
assessment at this stage also provides information on changes taking
place and hence helps determine what new problems require FSR,
5.4.1 Monitoring and evaluation
Monitoring and evaluating activities serve as a management tool to
improve the effectiveness of on-going projects2 and provide important
input for the design of upcoming projects. Monitoring and evaluating
check the validity of the description, design and testing activities
of FSR so lessons from the project can be systematically incorporated
into the design of future projects in that area or similar areas.
1Because of the iterative and dynamic nature of FSR it is diffi-
cult to divorce its descriptive and extension stages,
2Although we are specifically referring to projects, which usually
contain monitoring and evaluation components, the same principles apply
to more general extension programs that are not project specific,
Monitoring projects while they are in progress helps provide
project managers with information they can use to improve the project.
Monitoring activities involve systematically overseeing the process of
change as a consequence of the project, Evaluation activities, on the
other hand, are more concerned with the overall impact or results
from a project,
Monitoring and evaluating the introduction of improved strategies
need to be looked at from the perspectives of research workers, farm-
ing families, and society as a whole, The research perspective is
reflected in the degree to which the needs of the individual farming
family and society are met,
(1) In monitoring it is important to determine the number of
individual farming families that have adopted the improved
technology, the degree to which they have adopted it, includ-
ing the different components of a package, and the reasons
for divergence from what was recommended. Some types of
information necessitate acceptability-testing procedures.
Acceptability or adoption indices like those suggested by
Hildebrand (1979a) can be a valuable aid.
Sometimes a distinction is made between on-going and ex post
evaluation activities CCernea and Tepping, 19771. On-going evaluation
provides direct input to project management and focuses on affecting
current project activities. Such evaluation cannot be readily distin-
guished from monitoring activities unless one conceives of monitoring as
simply collecting information and evaluation as both collecting data
and more than cursory analysis of certain problems. We use the term
monitoring to include on-going evaluation, and confine evaluation to
its ex post function.
2Vincent Cpersonal communication) pointed out a major problem of
most projects is that they are not designed to deal with learning from
either farmers or policy makers, The execution of projects and their
evaluation are usually tied to the initial objectives--often narrowly
defined--which, given the time from project paper stage to implementa-
tion stage and the interim changes in the political environment and
those from farmer participation in the project, do not allow for full
utilization of learning in the feedback loop. Frequently, Vincent
suggests, the feedback loop must be ignored because of time pressures
and implementation schedules, which seriously reduces the potential
effectiveness of the FSR program in many projects.
(2) Evaluating the impact of improved technology from the point
of view of society involves answering such questions as
the distribution of benefits from the adoption, stability of
the ecological base, and the general nutritional level,
Although methodological problems hinder investigation of the above
issues, Information collected during monitoring and evaluation activi-
ties can be important in giving credibility to FSR and in feeding back
into the FSR program problems that have arisen from incorrect or impre-
cise specification of the environment or evaluation criteria, or in
giving research priorities for future FSR work,
5.4.2 Integrating FSR into projects
Increasingly projects also include an adaptive research component
closely linked to monitoring and evaluation activities, Adaptive
research can upgrade recommended practices being extended by the pro-
ject and help anticipate and solve problems, particularly technical
ones, which inevitably arise during the course of a project, Monitor-
ing activities can serve as an early warning system, identifying pro-
blems when they first appear so they can be dealt with through adap-
tive research.1 Adaptive research personnel on a project, in turn,
need to have close links with research institutions where they can draw
materials and expertise on short notice as required.
Monitoring, evaluation, and adaptive research activities in a
project collectively provide an in-house capability of carrying out the
full range of FSR-type activities. Ideally some of the same research
personnel who would participate in the initial stages of the FSR
before a formal project is initiated2 should be available throughout
A common example is the sudden emergence of a disease like rice
blast in the variety being extended in a project area, Early detection
and solution of such problems is critical to the continuing progress
of the project,
In the next chapter we strongly recommend that FSR be-an integral
part of the project planning phase. Too often planners assume that
suitable improved technologies are already available for farmers in
the proposed project areas.
the life of the project,. These activities, collectively in effect,
become the FSR component of the project. Analysts, design, and test-
ing activities can improve the performance of extension activities and
lay the basis for future extension efforts in the same area,
Monitoring, evaluation, and adaptive research activities of on-
going projects face some of the same methodological problems discussed
in preceding sections, The problems are complicated by dealing with a
dynamic situation that should be progressing as projected. Research-
ers must follow and understand current situations with a view to improv-
ing project performance, often through a series of measures designed
and tested on very short notice. In addition, researchers, possibly
the same group of researchers, are often asked to anticipate effects
of changes and to identify future improvements that will maintain or
build upon the current project through the next generation of projects,
These activities obviously make significant demands for financial sup-
port and skilled personnel,
6. INSTITUTIONAL LINKAGES
Despite a growing consensus on both the desirability of FSR pro-
grams and the need for a division of labor in undertaking them, sev-
eral issues often adversely affect their implementation. This chapter
focuses on intra- and inter-institutional issues at the national
level, including universities, research institutions, and agricultural
development agencies. It also examines current and prospective roles
of the regional/international centers and DCIs in supporting FSR
activities at the national level.
FSR has been applied in relatively few areas of the Third World--
with limited results. Clearly major inputs of resources in FSR programs
will be required for any hope of significant impact on large numbers
of small farmers. Given the holistic nature of FSR and the fact that much
of the work is location specific, to provide even cursory coverage of major
regions of the developing world will require significant resources.
Such resources are unlikely to be available from existing national
agricultural research programs, which tend to be poorly staffed and
under-financed. The need for location specific research and the abil-
ity to command resources are key considerations in determining an
appropriate division of labor among various research institutions.1
Clearly national programs, including research institutes and
universities in developing countries, have a comparative advantage in
"downstream" FSR because they are closest to the local situations.
And, in theory, they have the most direct relationships with national
institutions charged with implementing agricultural development
projects. Regional and particularly international centers, on the other
hand, are best situated to mount "upstream" FSR programs, as their mandates
normally encompass large geographic areas, cutting across national
boundaries, with problems that provide a research focus for "upstream"
At issue is not only the relationship of national programs to
FSR, but also the appropriate division of responsibilities among
national, regional, and international centers across the entire range
of agricultural research activities. These broader issues are beyond
the scope of this study. For a useful discussion of agricultural
research in developing countries, see Moseman (19701,
FSR programs. Regional and international centers also tend to command
the required financial and personnel resources for "upstream" programs.
Finally, regional and international Institutes can play catalytic and
supporting roles in the development of "downstream" FSR programs at
the national level,
6.1 PROBLEMS IN EXPANDING FSR IN NATIONAL RESEARCH SYSTEMS
Despite the growing recognition of "downstream" FSR's value, few
FSR programs are yet in LDCs. Several FSR-type activities in national
systems are special projects funded by donor agencies, which, in many
cases, are not well integrated into the core activities of national
agricultural research institutions, Prospects for successfully intro-
ducing FSR programs at the national level are influenced by a complex
of intra-and inter-institutional relationships involving national
agricultural institutions and universities, implementing agencies,
including ministries of agriculture, natural resources, and rural
development, planning departments, and funding agencies,
6.1.1 FSR and national agricultural research institutions/universities
National FSR programs are commonly and logically associated with
existing agricultural research institutions. But some FSR activities
have not been readily accepted by such institutions for the following
C11 Resource limitations, National agricultural research organi-
zations in LDCs are generally thinly staffed, sometimes
include a high percentage of expatriates, are poorly supported,
and depend heavily on external donor agencies for assistance--
often even for some recurrent expenses. Such organizations
often hesitate to initiate FSR on their own account because
doing so diverts resources from resourcerstarved, on-going
national research activities.
C21 Reluctance to change. Most scientists at national insttu-
tions have been trained and have experience in disciplinary
and commodity research programs, so many have limited
understanding and mixed feelings about FSR,1 And research
institutions usually are set up along disciplinary or commo-
dity lines, so incorporating FSR can create jurisdictional
problems and formidable obstacles to redefining responsi-
C31 Self- sufficiency and professional image, People in many
developing countries resist looking to outside regional or
international institutions for research results that can
be adapted to local situations. They may think that "borrow-
ing technology" will relegate the in-research establish-
ments to permanent secondary or even tertiary status in the
hierarchy of agricultural research.
C41 Time required to establish an efficient and credible FSR
program. Even where existing agricultural research institu-
tions agree to initiate FSR-type activities, they may not
have the patience to allow them to become effective. Re-
searchers charged with implementing FSR programs character-
istically have little or no experience in multidisciplinary
team efforts.2 A FSR team gains experience and credibility
over time and through the continuity of staff. Further,
linkages with planning, funding, and implementing institu-
tions also take time to develop.3
The difficulties involved in mounting FSR-type activities in
connection with the Caqueza project in Colombia illustrate this pro-
blem CZandstra, Swanberg,et al., 1979).
2The initial years of an FSR-type program in Honduras illustrate
this problem, Despite a desire by researchers to work together, the
path of least resistance was to revert to traditional commodity and
discipline oriented experiments (J. Posner, personal communication, 1980).
For additional discussion of the efficiency and credibility of FSR,
see Chapter 4.
Among the critical ingredients for introduction of FSR into
national agricultural research institutions are both technical and
social scientists in the institution. Then FSR-type activities might
simply evolve naturally, via collaborative efforts among researchers,
without a special new program,l If the agricultural research organiza-
tion has no social science research responsibilities, special ad hoc
arrangements may be necessary to mount multidisciplinary research
efforts. Zandstra C1978) suggests that teams drawn from more than one
institution can complicate administrative lines and place extra demands
on team coordinators.
Universities in the developing world may increasingly be used'for
FSR programs, as all the necessary disciplines are located in the same
institution and more flexibility may exist in using existing research
resources.2 In some countries, universities and other training insti-
tutions have few or no formal research responsibilities, but staff
members frequently are interested and willing to participate in special
undertakings, if external funding is available. Such involvement might
be directly linked with FSR training activities as discussed in
On the negative side, FSR-type efforts in universities may create
problems related to tenure and promotion, which tend to favor publica-
tions from research oriented along disciplinary and academic lines.
And FSR-type programs in universities may be difficult to carry out on
a continuing basis and may not qualify for core budget support.
Collinson3 suggests that both positive and negative strategies
might be used in promoting FSR programs in LDCs. p the positive
side, technical scientists may appreciate receiving information which-
farmers can provide, through FSR approaches, on research priorities
for specific target groups, Extension recommendations can be reviewed
1The FSR-type activities at the Institute of Agricultural Research
at Ahmadu Bello University in Ntgerta evolved that way CNorman, 1973),
The Central Luzon State University project with Kansas State
University is an example of an FSR-type program located in a university,
with a view to eliminating components unacceptable to farmers or to
modifying them. On the negative side, the on-going research programs
can be critically reviewed from a FSR perspective specifically focus-
ing on issues relevant locally and acceptable to farmers,
6.1.2 Should separate FSR units be established?
As a general observation, we viewfSR as a process that can be
incorporated into existing research programs as a "philosophy" of
research or established as a separate administrative and substantive
unit within an agricultural research institute. It is not necessary,
nor perhaps even desirable, in many instances, to have an administra-
tively independent "farming systems research unit." Several agricul-
tural research institutes in LDCs already have quasi-FSR activities
that simply evolved from collaborative projects among researchers.
Such an evolutionary process-may be the-most effective way of promo-
ting FSR even if the activity does not bear the FSR label.1 Of course,
such an evolution may not emerge in some situations. When agricultural
research activities and development policies are not focused on the needs
of small farmers, FSR might take root only as part of a general reorien-
tation and reorganization of the total research system. In Guatemala,
FSR activities were initiated after a major reorganization of the
national agricultural research system (Fumagalli and Waugh, 1977).
6.1.3 FSR and implementing agencies
Closely related to the problems of introducing FSR nationally are
relationships between FSR programs and the various national and local
organizations charged with Implementing rural development projects and
programs. The following obstacles may prevent the development of FSR
programs to serve local projects,
We are grateful to Ryan Cpersonal communtcatton1 for this point,
which was made with specific reference to FSR-type activities in India.
(1) Conflict with national policies, National policies may
support commercial farming and the development of capital-
intensive technologies. FSR cannot thrive without a strong
commitment to rural development,
C21 Bureaucratic centralism, Even where there is a commitment
at the top, as in India, FSR programs may be frustrated by the
non-responsive government bureaucracies that look on the
central ministry headquarters as the source of all wisdom
and direction. The organization of the agricultural develop-
ment effort may already be so fragmented along regional,
commodity, discipline, and functional lines that opposition
to FSR programs--to say nothing of a reluctance to implement
results of FSR work in a particular area--may be great (Gupta,
(31 Conflict with authorities in development projects. Few parts
of the developing world are unscarred by development projects.
Old programs that failed often leave a residue of bitterness
and opposition among local residents to everything connected
with the government, FSR teams going into areas with unpopu-
lar on-going projects are faced with the worst of both
worlds, opposition of local people and suspicions of imple-
menting agencies that do not wish to be discredited. Yet on-
going projects often provide an opportunity for FSR to con-
tribute by identifying changes in recommended practices or
to provide evidence needed to terminate the project.
6,1,4 FSR in relation to funding and planning agencies
National agricultural development banks and donor agencies are a
potential ally of FSR at the national level. They have procedures for
identifying, designing, appraising, monitoring, and evaluating rural/
agricultural development projects, They also Kave policies that
The often tragic experiences in Bangladesh with various large
scale irrigation schemes during the 1960s are examples of situations
that might have been avoided by applying FSR (Thomas, 1972).
explicitly direct them to devote an increasing share of their resources
to assisting rural areas and the poorest of the poor. In many cases,
such agencies are actively seeking ways to improve their somewhat
mediocre performances since 1960. Ways might be sought to incorporate
FSR activities into the identifying, designing, monitoring and evalua-
ting activities of the agencies, However, some are staffed with vet-
erans of agricultural development who contend that FSR is too compli-
cated, costly, and time-consuming to be useful in preparing projects.
Preparing a project already is an involved process and FSR could become
another bottleneck, impeding efforts to "move" more resources to support
FSR programs may find an ally in national and regional planning
agencies, Planning agencies often are poorly staffed and not effec-
tively integrated into governmental decision-making processes. Yet
they often are given responsibility for vetting development projects
and generally assessing the merits of annual budgets. That makes them
receptive to mechanisms that can improve project designs and assist
them in monitoring/evaluating on-going projects. The FSR approach in
project design, monitoring and evaluating might be promoted by planning
agencies.2 To require all implementing agencies to use FSR in the
first instance might only create serious bottlenecks, because the capa-
city to provide such services is not likely to exist in most countries.
So a gradual and selective "imposition" of FSRis probably preferable.3
In summary, a range of inter- and intra-institutional issues at the
national level bear directly on the feasibility of FSR programs,
Resolving the institutional issues is the key to FSR's future success.
Examples of success either of functioning FSR teams composed exclusively
of nationals who are producing results or of successful development
efforts with the FSR approach as a major ingredient do not yet exist.
The success of Guatemala's functioning FSR-type program to date is
See Section 4.4 for further discussion of the efficiency issue.
2ICTA is thus involved with planning agencies in Guatemala,
See Sections 5,4.1 and 5,4,2 for further discussion of fSR in the
monitoring and evaluation of projects,
limited. Ironically, the conditions that have made increasing numbers
of institutions look to FSR as a way to improve agricultural develop-
ment in specific locations mitigate against achieving a spectacular
Green Revolution-type of breakthrough for large areas that would give
great impetus to the development and acceptance of FSR. The spectacu-
lar breakthroughs that took place in the relatively few well-endowed
areas of the developing world--such as Punjab--are not likely to be
repeated in less favored areas where smaller incremental changes are
more likely. In addition, FSR is by nature conservative because it is
linked to helping farmers in the context of existing farming systems.
6.2 FSR AT THE REGIONAL AND INTERNATIONAL CENTERS
The range of institutional issues confronting FSR programs at
regional and international levels is as complex as those at the national
level. But being relatively new and well funded gives international
and regional centers several advantages over national programs. Also,
FSR at the IARCs has recently been strongly endorsed (Technical Advisory
Committee, 1978). Still FSR programs at the centers vary considerably
in scope and quality. One participant-observer of the FSR scene com-
mented that; the 1978 TAC review reflected "the professional chaos
over the subject CFSR}; most centers doing different things and none
doing FSR as the Review Team defined it" (Collinson, 1979a).
FSR programs at regional and international centers are distin-
guished by being "upstream" or "downstream" in character, FSR activi-
ties in support of sharply defined commodity programs as in IRRI and
CIMMYT tend to have more sharply focused research activities and fewer
methodological problems. Although FSR initially examines the "total"
environment, the ratio of variables to parameters is quite low, with
variables limited to potential improvements in practices related to
target commodities, Programs with a regional focus, on the other hand,
have many variables and correspondingly more methodological problems,
Much farming systems work at IITA and TCRISAT, for example, is a
prerequisite component to developing prototype solutions that could
cut across disciplinary and sub-program lines, The tendency for much
of the work at the Centers to be organized along disciplinary lines
is further reinforced by scientists' training and experience in research
organized along disciplinary or commodity lines. Their backgrounds
and their need to remain viable in their respective disciplines by
producing publishable research results mitigate against change,
Critics of "upstream" FSR programs contend that the activities
are too academic, too removed from the real world, and the'results are
unlikely to be used readily in national programs, to say nothing of use by
farmers themselves. Such attitudes undoubtedly contributed to termi-
nating the "upstream" Farming Systems program at CIAT.1 "Downstream"
programs, on the other hand, are increasingly perceived as useful,
particularly because of the poor record of improved practices intro-
duced without being screened via "downstream" FSR programs. Although
FSR is unlikely to generate Green Revolution-type advances, it can
focus research on developing practices more acceptable to small farmers.
The orientation of the FSR programs at regional and international
centers has important implications for national programs, The "down-
stream" programs generally work directly with national programs while
the "upstream" programs develop sub-program areas and methodologies
that might be adapted by national programs to local conditions.
Often "downstream" activities in national programs are weak or
nonexistent. Regional and international centers have sought to assist
in developing national FSR capabilities through training and technical
assistance. The work of IRRI through the Asian Cropping Systems Network
is the most successful example.. While it is generally agreed that
stronger national FSR programs are desirable, opinions differ widely
over the roles that regional and international centers should play.
The key issue is the appropriate mixture between assistance to national
programs in the form of training and technical assistance, on the one
hand, and the production of research results on the other. An increas-
ingly prevailing view is that the two features are closely related in
both medium- and long-term perspectives. Strong national programs will
improve the quality of research at regional and international centers,
We further argue that strong national programs are essential both to
define problems for "upstream"' programs and to adapt prototype
See Appendix A1,5 for the history of FSR activities at CIAT.
solutions from "upstream" programs into local conditions.
Extensive involvement of the centers in promoting FSR programs at
the national level also has some problems. IARC cooperative programs
with national institutions, for example, tend to favor regions/sub-
regions where the conditions seem to fit the constraints and solutions
defined by a center's FSR program. So IRRI tends to work in areas
where crop intensification offers the most promise, and ICRISAT tends
to work in areas where soil and water management at the watershed level
appear feasible. In the short and medium run, FSR programs at the region-
al/international centers are likely to devote most of their efforts to
assisting in developing national FSR capacity and to developing FSR meth-
odology around certain assumptions about the constraints. They may also
provide services to the commodity improvement programs by testing the
technical and economic feasibility of certain innovations in a farming
systems context. That is particularly true of the economics sub-programs,
which in the case of IITA often do more work for commodity improvement
programs than strictly within the Farming Systems Program. As the FSR
programs mature, both at the IARCs and at national levels, a new set of
roles is likely to emerge. The focus of FSR work should increasingly re-
flect the results of FSR at the national level. Additionally, the FSR
program should increase inputs for determining research priorities of
IARC'S crop improvement programs. That is not currently the case. The
commodity improvement programs often have proven records. At best, many
scientists in the crop improvement programs look at the FSR program as a
service organization for them, certainly not as a source of ideas for
research priorities. Finally, the crop improvement programs often tend
to have stronger links with national crop programs than FSR programs do.
6.3 POSSIBLE ROLES FOR DEVELOPED COUNTRY INSTITUTIONS
Agricultural research and training institutions'in the developed
world have had a profound impact on the character of national, regional,
and international agricultural research centers that serve the develop-
ing countries. The basic agricultural research structure in developing
countries is largely the product of colonial inheritance, post-
independence technical assistance programs, and donor-dominated
consortium that govern the IARC's, Nationals of developed countries
continue to form a significant share of the agricultural research
staff in the developing world. Third World agricultural scientists
in most cases received their training at developed country institu-
tions CDCIsl, Advanced-degree training in universities in the develop-
ing countries is closely modeled after training programs in various
Despite considerable accomplishments, agricultural research in the
DCIs has been criticized because it tends to be oriented around Indivi-
dual disciplines that are often geared toward refining technologies
that are inappropriate for the ecological conditions and resources
of most of the developing world, further, DCIs have been a primary
source of a "top down" orientation in the design and extension of new
technologies. In addition, DCIs are even more physically removed or
isolated than the IARCs from the various local situations that are the
ultimate foci of FSR, With some notable exceptions--GERDAT in France,
for example, the primary clients of DCIs are the agricultural communi-
ties of the countries where the institutions are located. Yet the DCIs
possess resources and influence that can be of considerable assistance
in developing FSR, as illustrated by the following examples;
C11 DCIs are likely to continue to be major sources of technical
assistance and training in support of agricultural research
in developing countries, Incorporating FSR perspectives in
their efforts might enhance their effectiveness. A small
but growing number of technical scientists and social scien-
tists at DCIs have had FSR experience and are among its
most active proponents.
C21 FSR may provide an effective means of defining or focusing
research on energy conservation and environmental quality,
which became important issues during the 1970s in the
developed countries CCastle, 1977_, Specifically, FSR may
provide a framework. in which different disciplines can relate
to one another and interact with farmers, while designing and
testing improved practices with such issues in mind,
(3) DCIs may assist national, regional, and international agri-
cultural centers with the development of FSR methodology,
(4) For DCIs serving areas with large numbers of small farmers--
as is true for parts of the US.--FSR offers a way to assist
more effectively in rural development and "domestic" agricul-
7.. TRAINING IN FSR
A major problem involved in establishing farming systems research
programs is the lack of agricultural scientists and social scientists
with FSR training or experience. Few agricultural or social scientists
have any experience with interdisciplinary research or more than a
superficial understanding of the terminology and methodology of other
disciplines. If FSR programs are to grow and be effective, they must
be staffed by individuals with training and experience in FSR, which is
largely unobtainable outside of existing FSR programs.
We believe that the training needs for "downstream" FSR programs
can most effectively be met through intensive non-degree courses in
areas where participants are expected to work--or at least in an
area with similar farming systems, Although participants in FSR
activities should have at least a first degree in some agriculture-
related discipline, a separate degree program in FSR is not required.
At the same time first degree programs in specific disciplines might
be modified to include courses and research methodology for students inter-
ested in FSR as a career. Let us examine the requirements fo: FSR train-
ing and then training in FSR in relation to degree and non-degree programs.
7,1 REQUIREMENTS FOR FSR TRAINING
Few advocates of FSR argue that FSR is a separate and distinct
field, particularly at its present stage of development. At the level
of the national program, FSR is a methodology for more systematically
identifying constraints and designing and testing improved strategies
in various locations. The task of developing the component parts of a
new technological package can be addressed by training in the tradi-
tional disciplines. In addition, although there are certain core dis-
ciplines involved in FSR, one or more of a wide range of disciplines
may be required in particular situations. Participants in FSR pro-
grams should bring substantive competence in at least one discipline,
and the interdisciplinary dimension should involve a team with exper-
tise in the requisite disciplines, The FSR approach Is not simply a
collection of individuals working in their own fields of specialization
but a team that works together to produce a common product--improved
strategies suited to specific situations,
The objectives of training programs in FSR should be; first, to
help participants understand the basic features of the farming systems
in the areas were they will work; second, to instill a sense of multi-
disciplinary understanding and tolerance; and third, to encourage par-
ticipants to work creatively and efficiently with farmers and extension
workers. These objectives can bestibe met by having participants carry
out FSR in the field under the direction of trainers with FSR experience.
Field experience can be supplemented by lectures on such subjects as
experimental design for farm trials and budgeting. We believe an intensive
course of at least two weeks duration is necessary for the team to gain
experience in working together.
7.2 FSR TRAINING AS A PART OF DEGREE PROGRAMS
Few universities now provide training in FSR as part of first or
graduate degree programs. But several U.S. land grant institutions are
seriously considering modifying training programs to better suit stu-
dents interested in FSR. The nature of FSR, namely its locational
specificity and the need to modify methodologies to suit local condi-
tions, strongly favors training on location--in the developing world
for students seeking careers in agricultural development there, for
example. But currently few training institutions in the developing
world offer programs specifically related to FSR in the context of
regular degree programs,
The historical evolution of fonral degree training in agriculture
has involved increased specialization. Also, few university professors
have experience with FSR, So it is difficult to advise students seek-
ing careers in FSR and to find the necessary expertise to teach courses
and to direct research in farming systems,
The growing number of institutions of higher learning in agricul-
ture concerned with how to train students for careers in FSR seem to
agree on the following issues:
01 Competence in an existing discipline is required, Thus
training for careers in FSR should take place within degree
programs of the existing disciplines. FSR is not a separate
(2) Since interaction among disciplines is a key feature of FSR,
graduate students, in particular, should be conversant and
sensitive to the basic concepts, terminology, and methodo-
logy of the core disciplines involved in FSR. That might be
accomplished in one or two survey courses covering all the
C31 Identifying problems is a key ingredient of FSR, Students
should be able to diagnose a variety of situations and enter-
prise combinations--annual crops, perennials, multiple crop-
ping in its various forms, livestock, and non-farm activities--
in close cooperation with colleagues in other disciplines.
A special course may be required to handle the methodology
to deal with various enterprise combinations. Such a course
should include field research by students working together
in small interdisciplinary groups applying the FSR approach.
C4) Students with an interest in FSR should be made aware of the
heterogeneity of farming systems throughout the world.
Several institutions already offer courses that expose stu-
dents to the salient features of the principal types of
farming systems in the world. Such courses might be slightly
modified to form a sequence with a course on FSR methodology,
(5) The most important "modification" in existing degree programs
for students seeking careers in FSR is to gain field experience
in FSR, possibly through thesis or dissertation research.
However, that approach.faces two problems:
(a) The research should be carried out by a team, a common
product should be produced, and the product must be
somehow disaggregated to satisfy the thesis or disserta-
tion requirements of individual team members.
Cb) The direction of an FSR project for students, requires
a team of faculty supervisors from various involved
disciplines who are familiar with the approach and
willing and able to work closely together,
In some cases research at national, regional, and international
agricultural research institutions in the developing world can form
part of a degree program for individual students; most of the interna-
tional centers have such arrangements with nearby schools of agricul-
ture. Additionally, students pursuing degrees at institutions in
developed countries also have carried out dissertation research at
international centers in FSR related areas. Both Cornell and Kansas
State Universities have been involved in cooperative training pro-
grams with CIMMYT in Mexico, under which groups of masters and doctoral
students from various disciplines have spent time at CIMMYT working
together as a team, although FSR was not the specific focus, Such
arrangements can significantly enrich training experiences. However, an
FSR experience at an international center still requires a supervisor in the
degree granting institution who understands and appreciates FSR. And that
may be difficult to arrange. In addition, a fair amount of training or re-
search direction of students by scientists in FSR programs at international
centers is "upstream" and oriented along traditional disciplinary lines.
7.3 NON-DEGREE TRAINING PROGRAMS IN FSR
Training in FSR is still largely confined to non degree programs
at national, regional, and international agricultural research institu-
tions, where students can carry out research under the direction of
scientists in the FSR or Cropping Systems Programs.
ClMMYT has recently initiated a three-month, FSR-type training
program specifically for economists. Participants receive instruc-
tion and gain experience by using applied economic methodologies to
analyze specific farming systems. Particular attention is given to
sensitizing economists to the biological aspects of crop production.
Technological improvements in maize and wheat production are emphasized,
and policy issues also are considered, Activities include field work,
seminars, and independent work in collecting and analyzing information
on existing farming systems, developing research and testing plans,
on-farm experimentation, maize/wheat breeding and agronomy, and such
policy issues as organization of agricultural research and pricing/
marketing policies (CIMMYT, 1979),
Among the "upstream" FSR programs, the training activities at
IRRI are probably best developed (Technical Advisory Committee, 19781,
Virtually all IARC's have training programs of various types involving
personnel from the farming systems or cropping systems programs in
these institutions. However, muchrof the training is on an :individual
basis or oriented toward such specific sub-program areas as soils or
economics. It is expected that more FSR-type courses will be offered
as the FSR programs in these institutions develop.
A select number of national research institutions in the develop-
ing world also are mounting training programs in FSR. These are pri-
marily designed to serve the institutions involved; that is, training
as a prelude to actual FSR activities in the field. Their programs are
particularly valuable in this respect because FSR is fairly location-
specific. The general model at the national level must be adapted
to the needs and realities of specific'regional situations,
ICTA in Guatemala puts both production and new social science
staff members through a one year in-service training program that com-
bines field experience in FSR with classes in specific research tech-
niques, including statistical analysis. Teams of participants going
through the sequence of FSR activities analyze actual farming systems
and design and test improved crop and livestock practices. Individual
thesis work in connection with an advanced degree is part of the par-
1Hildebrand (personal communication),
8.1 CURRENT STATUS AND POTENTIAL OF FSR
In FSR the farm is viewed in a comprehensive manner and con-
straints in the farming systems, research priorities, and strategies
for improvement are evaluated in terms of the whole farming system.
The objective of "upstream" FSR isito develop prototype solutions,
primarily through experiment station work, in order to overcome gen-
eral constraints in the zone in which the upstream research is being
conducted. "Downstream" FSR is more applied, and includes the farmer
in the research process. "Downstream" FSR includes the selective use
of available information ("body of knowledge" in Figure 2) in the pro-
cess of designing practices or recommendations which are suited to a
specific local situation.
This review of FSR activities, by focusing primarily on "downstream"
programs, concentrates on how FSR can help generate technology appro-
priate to small farmers. The review does not give adequate attention
to marketing, rural small scale industry, or to national policies
and structural barriers to more effective participation of small
farmers in the developmental process. The shortcomings and omissions
stem partially from FSR's newness in many countries, especially in
some national research systems.
We think a compelling case can be made for incorporating FSR
in both design of rural development!efforts and in determining research
priorities in commodity and discipline programs. FSR explicitly
recognizes farmer goals and seeks to include community and societal
goals. The use of multidisciplinary teams of researchers facilitates
the interaction of technical and socio-economic perspectives, which com-
plements, rather than overrides, the wisdom and experience of farmers
and extension workers, Although current FSR activities focus primarily
on the range of technical solutions to improving agricultural produc-
tivity--particularly with reference to crops, increasing attention is
given to such nontechnical factors as input and output markets and
macro policies, Finally, FSR can complement and strengthen commodity
disciplinary research programs by increasing their relevance and effec-
Despite the theoretical attractiveness of FSR, it will take time,
resources, and improved understanding of the whole process before FSR
is operating on a broad scale. In short, FSR is relatively young and
is likely to undergo considerable refinement in the years ahead.
8.2 CONCERNS ABOUT FSR
We have three major concerns about FSR. The first is the possi-
ble incompatibility of private and societal interests. When FSR re-
sponds to the short-term needs of farming families, societal interests
need to be considered. But that is likely to be particularly difficult
because it calls for predicting what might happen in the future. Never-
theless, ignoring the broader macro and societal interests could have
an irreversible, deleterious impact in the long run, such as reducing
ecological stability, increasing income inequalities, etc. Second,
because the evolutionary character of "downstream" FSR is not likely to
generate the spectacular changes exemplified in the Green Revolution,
it may be difficult to secure the funding required to sustain FSR--es-
pecially at the national level--over time. Third, and perhaps most
important, FSR may not be given ample opportunity to prove itself.
FSR is rapidly gaining acceptance, particularly with the donor agen-
cies, which are encouraging its adoption by national research organ-
izations. Expectations are running high. FSR is regarded by some as
a panacea. But FSR clearly is not a panacea for solving all the pro-
blems facing small farmers. The hope is that sufficient progress
can be made to sustain FSR's credibility while it grows, in the face
of inevitable disappointments.
8.3 FUTURE DIRECTIONS
Most of the methodological and implementation issues discussed
in this review can be directly translated into an agenda of action for
proponents of FSR. On the methodological side, the cost effective-
ness of FSR will not be resolved until more information has been gener-
ated on its costs and benefits in different ecological zones. Meth-
odology needs to be developed for effectively incorporating livestock
systems and societal, environmental, and distributional impacts.
The interaction between "upstream" and "downstream" programs is
likely to become increasingly critical in the future, as further
improvement in agricultural productivity in certain areas will re-
quire major changes in farming systems.
Ultimately, FSR will be judged less by the "correctness" of
its methodology than by how much it contributes to rural and agri-
cultural development. Operational linkages are needed between FSR
activities and the entire range of agricultural research, develop-
ment planning, and program implementation.
We view FSR as a process, not a structure that should be esta-
blished as a separate unit in an agricultural research institution
or development project. However, major changes in the structure
and orientation of rural development efforts--research, planning,
and implementation--may be required in order to make effective use
of FSR in integrated rural development projects.
We started our review by discussing how FSR came into being
in response to shortcomings of commodity and disciplinary research
programs. Where FSR is going is more difficult to predict. Will it
prove to be a means by which small farmers can be helped in the fu-
ture? Or will it be a passing fad too difficult and too demanding in
personnel and in time and costs? We think FSR can make a modest but
significant contribution to improving the lives of small farmers in
the Third World.
A. Description of Selected Farming Systems Research Programs
B, Farm Management Research