Department of Agricultural 
Economics and Rural 
Sociology
Agricultural Economics 
Series No. 38 
October 1985
.Alabama Agricultural 
Experiment Station 
Auburn University
David H. Teem, 
Acting Director 
Auburn University, 
Alabama
Data 
Collection 
in
Subsistence 
Farming 
Systems
A Handbook
James L. S 
tallings
7 
r,
II
*4 4
-A
- 4~~
-4 
'-4,
~
~
~ & ~
7"-~
-, - 4t '5
~ ~,*
~
t 'Nots ~ -
il
"I
t7
"' ~~ ~~~ ~ ~ ~ "#". ., _ -e r?"-u' ". _ r
.",. x ir x
. "- a swr ~ 
'AM Ake
a 'M
- 7 "'A\SC;
U
S'
r
-00

FOREWORD AND ACKNOWLEGMENTS
This handbook summarizes a 1980 
United States Agency for
International Development 
(USAID) Personal Service 
Contract
No. AID/afr-C-1607 (Revised); 
Project No. 698-0135, "Collab-
orative Survey and Analysis of 
Agriculture--Rwanda" (9). As
part of this contract, the 
author was directed to "prepare a
state-of-the-arts paper on 
the techniques for measuring 
area
and production in intercropped or 
associated cultivation and
the enumeration of livestock 
numbers and livestock products,
as well as summarize the information 
available in the best
written material on the subject" 
(9). Dr. John M. O'Sullivan,
presently at North Carolina 
A & T, also contributed to this
handbook through a paper written 
under the author's direction
while he was a M.S. student 
at Auburn University (8).
Many sources, published and informal, 
were used in writ-
ing this report. A complete list of these 
is contained in the
two papers mentioned above. 
Only portions of these will be
cited here since this handbook 
is in outline form. However, 
a
few were used more extensively 
than others and deserve special
mention. Collinson's book (2) 
was drawn upon heavily, as was
Hunt's (5). Dalyrymple's book 
(3) was very useful, as were
selected papers presented at 
the Workshop on Field Data Col-
lection in Rural Areas of 
Africa and the Middle East 
(1, 4,
10). Other important sources 
were from Kearl (6), and 
Norman
(7). There has been no attempt 
to research further refer-
ences after the two main 
source publications were 
written in
1980 in compiling 
this handbook. This 
handbook was written
mainly to service requests 
for the more bulky 
source publica-
tions mentioned above 
(8, 9).
While many of the recommendations 
in this paper are the
result of review of literature 
for the above-mentioned 
USAID
contract, they are 
also from the author's 
own experience in
using students for 
conducting surveys 
in Tanzania in the 
late
1960's while on the 
faculty of Morogoro 
Agriculture College,
as well as his experience 
in the U.S. Census 
of Agriculture
in the early 
1960's.
CONTENTS
Page
FOREWORD AND ACKNOWLEDGMENTS 1
I. INTRODUCTION 5
II. THE COMPLETE PLANNING OF A SURVEY 
6
III. RECOMMENDATIONS FOR SELECTEI PROBLEMS 
9
A. DETERMINING CROPS GROWN AND CROP AREA 
9
1. Determining General 
Land Use 
9
2. Measuring Area on a Slope 
12
3. Boundary Problems 
13
4. Non-Productive Areas 
Within Plots (Roads,
Ditches, Erosion Barriers, 
and Weed
Patches) 
13
5. Planting Concentration 14
6. Idle and Fallow Areas 
15
7. Prepared vs. Planted Areas 
15
8. Planted vs. Harvested 
Areas 16
9. Different Varieties of 
the Same Crop 16
10. Pasture and Land from 
Which Forage is
Harvested 
17
11. Forest and Land from 
Which Forage is
Harvested 18
12. The Farmstead, Kitchen 
Garden, and Other
Non-Crop Areas 
19
13. Minimum Plot Size 
for Recording Area 20
14. Multiple Cropping, 
Double Cropping and
Triple Cropping 
21
15. Row or Sub-Plot Intercropping 
22
16. Mixed Intercropping 
23
17. Relay or Sequence Intercropping 
28
18. Changing Area and Location 
of Plots in
the Holding Within the Year 
29
19. Wild Economic Production 
Areas 30
20. Community, Multiple 
Person Partnerships,
and Other Tenure and Ownership 
32
21. Ponds, Lakes, and 
Other Water Areas in
the Holding 
33
22. Special Crops and Livestock 
Grown or
Raised by Only a Few Farmers 
34
B. DETERMINING CROP PRODUCTION, 
CONSUMPTION, AND
SALES 35
1. A Common Unit for Measuring 
Production 35
2. Techniques for Physically 
Measuring or
Obtaining Production Data 36
3. Non-Homogeneous Production 
37
4. Stage of Maturity and Intermittent
Harvesting, Consumption, and 
Sales 37
5. Production from Intercropped 
Areas 39
6. Unharvested and Incompletely Harvested
Production 40
7. Production in Progress but not Mature in
Survey Year 40
8. Wild Production 41
9. Production from Community, Multiple-
Person, or Partnership Holdings 41
10. Production from the Farmstead, Kitchen
Garden, and Other Non-Crop Areas of the
Holding 41
11. Accounting for Home Consumption vs. Sales
vs. Manufacturing vs. Kept for Seed 42
12. Production from Multiple Product Crops 43
13. Production of Forage Crops 43
14. Production of Forest Products 43
15. Production from Ponds, Lakes, and Other
Water Areas on the Holding 44
16. Crops Grazed by Humans 44
C. ENUMERATING LIVESTOCK NUMBERS AND PRODUCTION 44
1. Determining Kinds of Livestock, Ownership,
Numbers, and Disposition 45
2. Migratory Herds 46
3. Owners of Livestock Herded by Someone
Else 47
4. Frequency and Time of Enumeration 48
5. Enumeration of Numbers of Poultry and
Small Animals 48
6. Animal Products 49
7. General Conclusions 49
SELECTED REFERENCES 51
Information contained herein is available to all without
regard to race, color, sex, or national origin.
DATA COLLECTION IN SUBSISTENCE FARMING SYSTEMS --
A HANDBOOK
James L. Stallings*
I. INTRODUCTION
Surveying farming systems in subsistence agriculture is
more difficult, in many ways, than conducting a similar survey
in a highly developed economy. Unlike the agriculture of
developed economies, subsistence agriculture usually involves
many special problems. These include complex and non-standard
cropping systems, intercropping, and double cropping. There
is usually a high degree of illiteracy which precludes use of
mail questionnaires. Records are poor if any are available at
all. There is usually poor recollection of harvested or col-
lected amounts of crops and livestock products, as much of the
production is harvested intermittently for home consumption.
Enumerators encounter poor roads, poor accommodations, and
poor communications. There is no cadre of educated personnel
which could be hired to assist in enumeration. Boundaries are
uncertain, as well as ownership patterns. Also, farm and off-
farm work patterns tend to confuse the type of farm-to-farm
enumeration common in western agriculture. Finally, migratory
livestock systems pose special problems not usually encoun-
tered in settled agriculture.
* Associate Professor, Department of Agricultural Econom-
ics.
The purpose of this 
handbook is to aid the 
various re-
searchers, both expatriates 
and locals, who must attempt 
to
conduct surveys in such 
situations. While most of the 
sugges-
tions here can be found in some 
form in the various literature
reviewed for the two source 
publications, the information 
is
scattered and not easily found 
under one cover, and much 
of it
is unpublished material. 
This handbook attempts 
to condense
some of the more important 
of these sources, plus reporting
the author's own experiences.
II. THE COMPLETE PLANNING 
OF A SURVEY
Before discussing some of the specific 
problems encoun-
tered in surveying the complete 
farming system, it might be
well to review what a "complete" 
survey could involve. Ob-
viously a survey could vary from 
a small sample survey, with
limited objectives, for one enterprise, 
in a limited area, to
a massive national census involving 
an attempt at complete
enumeration of all rural residence 
and producing units. The
planning would necessarily be 
more elaborate for the latter.
The following outline, therefore, 
is an attempt to cover most
aspects of planning a survey 
from start to finish, keeping 
in
mind that seldom, if ever, 
will all steps be used.
A. PLANNING THE SURVEY
1. Determine the objectives 
of the survey
a. Get advice from potential 
users
b. Consult with experts
c. Review literature and 
previous surveys
2. Decide on whether the survey should be a sample 
or a
complete enumeration
3. Determine kinds and forms of data needed
a. Consider the objectives
b. Consider all user needs
c. Set up dummy tables with tabs, 
intervals, etc.
4. Determine statistical 
procedures
a. Determine the sample frame--this 
may require a
preliminary survey(s)
b. Determine the sample size
c. Determine error control
d. Determine measurement equipment needed, 
if appli-
cable
5. Draw the sample, if applicable
6. Design questionnaires and other 
enumeration instru-
ments
7. Write training manuals for 
such participants as su-
pervisors, crew leaders, and enumerators, 
as applica-
ble
8. Recruit supervisors, crew leaders, 
and enumerators,
as appropriate
9. Train supervisors, crew leaders, 
and enumerators, as
appropriate
B. CONDUCT PUBLIC RELATIONS 
CONCERNING THE SURVEY
C. CONDUCT A PRETEST OF 
INSTRUMENTS AND PROCEDURES
1. Conduct actual field enumeration pretest with limited
staff and enumerators
2. Return and hold critique of all procedures, instru-
ments, and manuals
3. Revise procedures based on the critique
4. Conduct further, more elaborate, "pilot survey," if
appropriate, before beginning a large, extensive
survey
5. Revise again as appropriate after pilot survey
D. CARRY OUT SURVEY--CONTENT WILL VARY BUT MAY INCLUDE:
1. Determination of soil types and general land use--
agriculture and non-agriculture--by types of farming
areas
2. Nature of agricultural holdings, including ownership
patterns
3. Sociological data concerning the family, household,
and associated institutions
4. Nature of the cropping system(s), including crop
calendars
5. Nature of the livestock system(s), including calen-
dars
6. Income sources, including crops and livestock and
their products, gifts and other exchanges, crafts,
forest products, off-farm work, and other miscella-
neous
7. Farm inputs, implements, facilities, and practices,
including labor use, improved 
varieties, and erosion
control
8. Prices received, market channels 
and methods, by
enterprises
E. ANALYSIS OF DATA
1. Receiving, editing, precoding, 
and coding
2. Data processing, tabulating, 
and generating tables
3. Computation of enterprise cost 
and returns, as appli-
cable
4. Computation of measures of 
farm income, as applicable
5. Calories per capita, per hectare, 
and other appro-
priate computations, as applicable
F. PUBLISH RESULTS
G. USE THE RESULTS FOR PLANNING AND POLICY 
FORMULATION
H. CRITIQUE RESULTS FOR FUTURE 
SURVEYS
III. RECOMMENDATIONS FOR SELECTED 
PROBLEMS
A. DETERMINING CROPS GROWN AND 
CROP AREA
1. Determining General Land Use
In farming systems research in subsistence 
agriculture,
it is usually desirable to know both 
general land use for the
1
area or country as well as area of 
agricultural holdings,
1
Throughout this article the term "holding" will be used
to designate the home or farmstead 
and any associated agri-
cultural land, contiguous or not, 
operated as one unit by a
farmer and the extended family. 
The term "subsistence agri-
culture" implies the holding mainly 
is engaged in production
for home consumption, but does not 
exclude off-farm sales.
including areas of 
specific plots and 
crops. However, gener-
al land use is quite 
a different problem 
from enumeration of
specific plots or holdings 
and will not be treated 
here.
For areas of specific 
plots, holdings, or crops, 
however,
there are several approaches. 
The most desirable, 
if possi-
ble, would be to simply 
ask the farmer, using 
a mailed ques-
tionnaire or personal enumeration. 
This is usually not possi-
ble in subsistence 
agriculture for two 
reasons: a large
portion of the farmers 
may be illiterate and 
could not read
and fill out a mailed questionnaire; 
and, they do not have in
mind, in many cases, a concept 
of area in the holding 
and of
specific crops. Even if 
a farmer did know the area 
of the
holding or crops, the unit 
may or may not be the 
international
or national unit being used 
and it will be necessary 
to use a
conversion rate.
Other problems of 
getting area, to be 
discussed in more
detail later, include: 
distortions due to slope, 
undefined
boundaries, non-productive 
areas within a plot, 
planting
concentration and what 
constitutes a stand, 
idle and fallow
lands, prepared vs. 
planted acres, planted 
vs. harvested
acres, variety differences, 
intercropping, sequence 
cropping,
wild economic production, 
and community production.
Other than directly 
asking the farmer, 
however, there
are several objective 
measurement approaches 
to determining
area as follows.
10
a. If fields have parallel 
sides, one of each of 
the
two sets of parallel sides 
can be measured and area
calculated as the product 
of the two.
b. If there are more than three sides, the 
area can be
converted into a polygon 
and broken into triangles.
The area of each triangle can 
then be determined as
1/2 the base x height 
to apex.
c. The area of a polygon can 
also be determined by
traversing the circumference, 
measuring the differ-
ent segments, and taking bearings 
to determine ap-
propriate polygons or triangles.
d. The area of a polygon can also 
be determined by
plotting it on a map using bearings 
from a single
point and distance measurement 
of two points on the
polygon without traversing 
the circumference.
e. Aerial photography 
will be mentioned as 
a possibil-
ity here but will not be pursued. 
While possibly
appropriate in helping determine general 
land use of
a political unit, it has usually been 
of little use
on small subsistence areas having many 
of the prob-
lems already mentioned.
Because of the difficulties of any 
other approach in
subsistence agriculture, it 
appears that these direct mea-
surement technique would be 
more appropriate than simply
asking the farmer. Unless 
the fields or plots are perfect
rectangles, all fields or plots 
should be forced into the
11
shape of a polygon, and sides and angles should be measured
by traversing around them rather than choosing a method
involving cutting across plots. Cutting across plots may
involve the chance of destroying crops, in some cases, and
might be difficult because of obstacles in other cases.
2. Measuring Area on a Slope
This is a problem concerning the actual area of sloping
land projected on the face of the earth as if the area were
flat. This is a problem sometimes mentioned in literature,
but rarely are any recommendation made. To quote one author,
"It seems rarely to be a serious issue in practical working
conditions, though adjustments are discussed occasionally.
Even in a 25 percent slope, the difference between projected
area and actual area is only about 5 percent, and this is
probably among the least of the uncertainties attaching to a
good deal of this work in developing countries," (5, p. 46).
If the problem were in accounting for all land in general land
use in a country, slope might be an important consideration.
Most studies, however, deal only with samples of holdings.
These holdings usually account for only a certain percentage
of all land use. General land use is usually determined by
more macro methods (aerial photos, etc.). In subsistence
agriculture, however, actual area in holdings and in specific
crops is usually more important than area on the face of the
earth in enumerating.
An appropriate recommendation, therefore, would be to do
12
as good a job as possible 
in measuring actual area 
planted,
harvested, and prepared as 
if the area were flat, and not 
be
concerned with projected 
area. Even if the differences 
were
great, the concept sought 
is actual area of crops.
3. Boundary Problems
In many cases in subsistence 
agriculture, the specific
boundary of a plot will 
not be clear. This is especially
true of areas where clean cultivation 
is not practiced and
crops are planted in trash 
and among weeds and other non-
economic plants and 
trees.
There appears to be no 
good solution to this problem
except judgment. If, in the 
enumerator's judgment, the land
planted has essentially been 
cleared, and in any way 
culti-
vated as a single intact plot, 
the boundaries can be esti-
mated and forced into a polygon 
shape for measurement. Other
approaches will be treated later.
4. Non-Productive Areas 
within Plots (Roads,
Ditches, Erosion Barriers, 
and Weed Patches)
The question may arise as 
to whether the area of a 
plot
should include such things 
as paths, ditches, erosion 
bar-
riers, and weed patches in 
what would otherwise be known 
as a
separate plot; or, should 
only the portion of land on 
which
crops are actually grown 
be counted? The usual 
recommenda-
tion seems to be as follows: 
If the non-productive area 
is
man-made and definitely 
associated with production 
of crops
(i.e, irrigation ditches, 
access strips, paths or 
roads, and
13
erosion barriers) it should be counted as land 
in the plot.
For other natural, non-productive areas (i.e., weed 
patches,
and boulders) judgment will have to be used by the 
enumera-
tor. A rule could be made, if desired, to exclude 
any natu-
ral non-productive area from the plot 
area if it constitutes
more than 5 percent of the total plot (or other 
appropriate
percentage as agreed upon).
5. Planting Concentration
It is common in some areas of subsistence agriculture to
scatter plantings of individual plants among trash and in
wooded areas so that it is impossible to define a plot or
field. The problem then becomes one of whether to try to
estimate an area for such production at all and, if so, how?
In estimating areas, the enumerator needs to address the
concept of what concentration of plantings of a particular
crop is necessary before it can be considered as a solid
stand.
If, in the enumerator's judgment, the area in question
is not a solid stand or an intact plot for measurement 
pur-
poses, it is suggested that production from such areas be
recorded throughout the enumeration year (keeping it separate
from "wild" production) and coverted to crop area later using
standards for solid stands of different crops 
developed for
that purpose. If this production has been 
planted outside
the boundaries of the contiguous area easily 
recognizable as
"the holding," the equivalent area may be added 
to the area
14
in the holding.
6. Idle and Fallow Areas
In subsistence agriculture, as in 
any other agriculture,
there are frequently areas in the 
holding which are being
prepared during the enumeration 
year, or in some recent
period, but are not used for 
agricultural purposes in the
year of enumeration.
If these areas are suitable for cropping 
(as opposed to
rough, non-usable land), they may be 
appropriately categorized
and the area recorded, as 
they represent potential future
production. Categories may include: 
cropland suitable only
for pasture, further broken down 
into "improved" and "unim-
proved" cropland suitable for cultivated crops; 
and, any other
category desired. If the land has 
not been prepared, culti-
vated, or otherwise used for agricultural 
purposes in the
enumeration year, a rule may be made that if it has 
been used
or cultivated within some reasonable 
recent time, say the last
5 years, it should be counted. All other 
land can be classed
as "idle land," "homestead," "roads, paths, 
etc.," or other
category, as appropriate.
7. Prepared vs. Planted Areas
Should the whole portion of land 
prepared for a crop be
included in the area of a crop 
or only that area actually
planted? Because preparation 
of land involves labor and,
perhaps, other inputs, it is 
suggested that this area must 
be
accounted for in some way, even 
if left fallow after prepared
15
and not used in the enumeration 
year. One suggestion is that
after determining an appropriate boundary for planted area,
the rest of the area may be recorded as "fallow" cropland.
Also, fallow land for part of a year may become planted land
later in the year. If so, it may eventually 
be included in
planted and/or harvested cropland.
Any land formally prepared for planting may be measured
and allocated either to fallow cropland or to planted and/or
harvest area. "Formally prepared" may mean 
clearing by any
method and/or plowed or otherwise cultivated in any 
way.
8. Planted vs. Harvested Areas
Should planted or harvested acres be measured and re-
corded, or both? As with prepared vs. planted areas, plant-
ing involves labor and other inputs and most sources indicate
that it is valuable to record planted areas as well as har-
vested areas. This is particularly true if crops are planted
in the enumeration year, but will not be harvested until the
next year. Both planted and harvested acres are also useful
in estimating the extent of drought or other natural disaster
in the enumeration year.
9. Different Varieties of the Same Crop
The question becomes one of when are two or more varie-
ties, of what would be otherwise known as the 
same crop,
enough different to be enumerated as separate 
crops? For
example, should cooking and sweet varieties 
of bananas be
enumerated simply as bananas, or should they 
be separated as
16
different crops? Should 
two or more types of beans 
be enu-
merated simply as beans or 
as separate crops?
These are judgments for which 
general recommendations
cannot be made without specific 
information as to the objec-
tives of the survey. However, the 
criteria suggested is as
follows: Count as separate 
crops any variation in 
varieties
of what would otherwise be known 
as the same crop, if there
are significant differences 
in either cultural practices,
yields, use, or marketing methods. 
Rules will have to be made
for specific cases and included 
in appropriate instructions 
or
manuals.
10. Pasture and Land from Which 
Forage is Har-
vested
This category of land can vary widely from 
migratory
ranges with sparse vegetation 
and not owned or controlled by
any one holding, person, or persons, 
to improved or unim-
proved communal grazing in a 
definitely defined area, to the
cutting of forages or grazing off 
the contiguous holding, to
different categories on the holding 
itself. If the areas
harvested are on the contiguous 
holding, there is little
problem other than devising 
appropriate categories for enu-
meration. If off the holding, 
however, enumeration becomes
more difficult.
A suggestion is to not 
include in the holding any 
pasture
land on which migratory 
herds graze and which is 
not owned or
controlled by any one holding. 
Similarly, do not include 
in
17
the holding definitely defined areas 
of communal grazing.
Within the holding, however, enumerate different categories of
pasture and forage land as desired, such as land from which
hay or other forage is harvested (along with estimates of
production) versus land used only for pasture, broken down, if
desired, into "improved," and "unimproved." Off the holding,
enumerate equivalent areas of pasture or land from which hay
or other forage is harvested for a specific holding. This
should be estimated by standards developed and included in
enumerator's instructions if the areas are not formally de-
fined and not capable of being measured.
11. Forest and Land from Which Forest Products
are Harvested
This category of products has caused trouble in many
surveys and censuses throughout the world. The problem is
that much of the forest products of the world are harvested
from areas not traditionally thought of as "farms" or "hold-
ings." Many large forest operations are owned and managed by
either the government of a country or a large private com-
pany. This would cause no problem if all forest products
came from operations of this type. However, there are sev-
eral instances where there are forested areas or trees on
holdings which are considered more traditionally "agricul-
ture." For instance, a farmer who grows traditional agricul-
tural crops and/or raises livestock also may harvest timber
or sell firewood or charcoal. These products may come from
18
the formal holding area or may be harvested 
off land not
owned or controlled by the farmer or holding. 
In some in-
stances this is handled by not considering forest 
products as
"agricultural" products and they are simply 
ignored. How-
ever, in farming systems research, they should 
be accounted
for as part of the effort expended 
and income obtained by the
farm family in the total 
farming system in the same 
manner as
off-farm work. Labor used for tending and 
harvesting forest
products also should be accounted 
for in recording general
labor use. Forested area on holdings should 
be enumerated
and recorded, but no equivalent 
area off the holding need 
be
estimated and added to the area of the 
holding as with tradi-
tional agricultural crops. There 
is an exception to this
which will require judgment. 
If the forest products are of a
cultivated nature (i.e., capok trees 
from which capok is
harvested or other forest enterprises 
grown in rows and
cultivated), they should be treated 
as any other crop in the
holding. Production from wild 
or non-cultivated areas, ei-
ther on or off the holding, need 
only be recorded.
12. The Farmstead, Kitchen Garden, and 
Other
Non-Crop Areas
In subsistence agriculture throughout 
the world, there
are usually areas on which the 
home stands, yard areas, lanes,
paths, corrals or livestock 
holding areas, area occupied 
by
out-buildings, etc. These are 
usually lumped into an area
called the "farmstead" or "homestead" 
in most of the world
19
after accounting for all 
other economic and idle 
or fallow
farmlands.
After all economic crops, 
forest area, pastures, 
water
areas, and other areas directly 
associated with agricultural
production have been accounted 
for, all other area definitely
enclosed in the contiguous 
holding can be classed as 
"farm-
stead," "homestead," or by some 
other appropriate term. If
the home is definitely separate 
from all or most agricultural
plots, it should be included 
in the holding but will be 
only
an estimate of the living 
area directly contiguous to 
and
associated with family living, 
processing, and otherwise
working around the home. 
This will be a judgment in many
cases.
13. Minimum Plot Size for Recording 
Area
This is a definitional problem. 
It should be determined
by the purpose of the survey 
and/or nature of the crop(s)
being enumerated. For instance, 
if the enumeration was pri-
marily in range land, 1 hectare 
may be the minimum necessary
to be enumerated, or even a larger 
area. However, if the crop
was intensive and of high value, 
such as a vegetable crop,
even 1/10 hectare may be large 
enough to be significant in a
survey. While some sources 
might advocate different minimums
for different crops, that, 
too, causes a problem and 
may not
be worth the extra effort. 
Before a minimum is set, 
some
preliminary study may have 
to be made as to how much, 
what
crops, and what value would 
be missed at decreasing 
minimum
20
sizes. Then, a decision should be made commensurate with the
budget and trouble involved with enumerating increasingly
smaller sizes. It may not be worth it, or appropriate, for
instance, to enumerate area of a crop grown as an herb 
in the
yard or farmstead area, although production may be recorded.
However, if a crop with small area per holding is 
widespread
throughout the area, a significant contribution to the 
agri-
culture of the area may be missed by setting 
the minimum too
high per holding.
Considering that most holdings in subsistence agricul-
ture are small, and areas of individual crops are even
smaller, with high value per unit of area, the recommendation
most often suggested is to enumerate and record area of any
plot of any one crop or mixture if the area is as much as 5
square meters. Production should be recorded, but 
area
should be designated by some symbol, with 
a footnote "less
than 5 square meters or 0.0005 hectare if the area is less
than 5 square meters.
14. Multiple Cropping, Double Cropping and 
Triple
Cropping
In cases where the mixture of crops in any one enumera-
tion year on the same plot of land follows in a definite
sequence, with one crop harvested before the other is planted
(or, with only a minimum of overlap as in the case where the
new crop is planted in the rows or other area of the previous
crop just before harvest), there is little problem of enumera-
21
tion. The only requirement is that the areas and 
production
of each crop in the sequence be recorded. There is another
figure which is useful, however, and that is the area within
an enumeration year which is double or triple cropped. These
areas, along with all single cropped and other areas, are
necessary to account for total area of each crop and area in
the holding. Otherwise, if all areas of crops in double or
triple cropped sequences were simply added up, double and
triple counting would occur and estimates of the 
total area in
the holding would be distorted.
15. Row or Sub-Plot Intercropping
In cases where the mixture of crops in multiple cropping
is planted in definite rows or sub-plots, there is usually
little problem involved in estimating area other than deciding
on the boundary between them if plants are of similar size.
The sub-plots can be measured as separate plots and the
boundaries can reasonably become half the distance between the
definite plots. In the case of equally spaced rows, the whole
area can be measured and allocated to each crop by counting
rows.
If rows are not equally spaced, or plants are not of
similar size, areas of row-planted crops may become more of 
a
problem, especially between crops of widely different growth
patterns and heights (For example, wheat in narrow rows 
or
planted solid abutting maize planted in rows of 2 or 3 
feet,
or fruit trees 15 to 20 feet apart next to the wheat.) 
In
22
cases of tree crops, an estimated 
point below the last outer
branches can reasonably 
be taken as a point of measurement 
to
the next crop such as wheat. 
If the area of wheat or cotton
extends somewhat under the 
branches of a tree crop, the 
judg-
ment decision will have to be made as 
to whether it will be
called intercropping. 
If not considered intercropping, 
the
boundary may be estimated 
at some average of the 
overlapping
area where the two crops 
meet. In summary, if areas 
are
capable of being definitely 
separated and are not judged 
to
fall under the classification 
of intercropping, determine
boundaries between areas 
planted as half the distance 
between
the last plants of each 
sub-plot, or at the average 
of the
overlapped area, if minor. This will 
require judgment in many
cases.
16. Mixed Intercropping
In farming systems in tropical 
areas, multiple cropping
frequently consists of 
interplanting the various 
crops by a
scattering of the plants 
of one crop, which is clearly 
domi-
nant, with a few scattered 
plants of one or more others. 
In
other cases there are mixtures 
of two or more crops which 
are
part of a regular and traditional 
pattern of cropping in the
neighborhood. In other 
cases, there appears to 
be no definite
pattern and the plant 
mixture is non-homogeneous 
between
farms, with no one species 
dominant in all cases.
Assuming that areas 
of individual crops in 
a mixture
could somehow be estimated, 
there is the initial 
question as
23
to whether or not it would make 
sense to aggregate data from
pure stands of a crop with that 
from a mixture. However, if
it is assumed, for the sake of 
analysis, that the total area
of the plot containing a mixture of 
crops must be allocated
among the different crops, there 
are several approaches which
have been suggested in different 
references as follows:
a. The "Pure Stand Approach": 
This approach requires
development of a set of standards 
for the different
plants contained in mixtures for 
their density in
some "average pure stand." This 
may be difficult if
one or more of the crops 
in a mixture are seldom or
never grown in pure stands 
in the area. However,
some estimate of the number of plants 
in a "pure
stand" must be estimated in some way. The number 
of
the plants of each variety 
in the plot must then be
actually counted or estimated 
in some way and the
areas determined for each 
as if they were in a pure
stand. If the total areas of 
the different plants
add to more or less than the 
area of the plot, the
area of the plot can then be allocated 
in proportion
to the areas of individual crops 
in the added total.
This can be done in the office 
after enumeration is
completed.
b. The "Proportion of the 
Area Occupied Approach":
This is similar to the pure 
stand approach, but is
only applicable if it is possible 
to estimate the
24
proportion of the ground 
covered by each crop. 
It
would not work well in cases 
where one or more crops
are grown in the shade 
of other crops. This 
ap-
proach requires "judgment" 
on the part of the ob-
server as to the proportion 
of the area occupied by
the respective crops. With 
practice, this approach
may be accurate enough for 
some mixtures.
c. The "Each Crop Occupying 
the Whole Area Approach":
This is applicable if there 
is no single dominant
crop and the concept of total 
acres of crops grown
adding to more than total 
land area can be accepted
as with double or triple 
cropping. This may be 
more
applicable in cases where only 
one crop is being
studied in a survey. 
It poses problems, however, 
for
a general survey of all 
agricultural operations 
where
there is a need to have a concept 
of specific acres
of specific crops at any 
one time adding up to total
cropland in the holding.
d. The "Mixed Crop Approach": 
This approach is par-
ticularly applicable when there 
are certain tradi-
tional well-defined mixtures 
grown by many farmers
in an area and the proportion 
of each crop in the
typical mixture is about 
the same among many farms
(for example, bananas 
shading coffee, with sweet
potatoes planted under 
these in some areas of 
East
Africa). In this approach, 
the familiar mixture 
may
25
be given a name and counted as a single 
crop such as
"Banana-Coffee-Sweet potato Mixture." 
Unfortu-
nately, there usually comes a time 
when a national
government, the U.N., etc. wants area of individual
crops for a particular use. This approach 
would
then break down and area would have to be estimated
in some way. This could be done after the fact in
the same manner as in any of the three approaches 
in
a, b, or c above.
e. The "Area from Harvested Production" Approach: this
approach is much like the "Pure Stand Approach" 
in
that it requires development of standards for aver-
age production per unit of area from a pure stand
for the area and year of enumeration. In this ap-
proach, area is determined for the whole field 
plot
at any time, but it is not allocated 
to specific
crops in the mixture until harvest 
of all crops is
complete. Then the standards are applied 
to esti-
mate area which would have been required to produce
this amount from a "pure stand" with "average
yields" given the weather conditions, cultural prac-
tices, etc. of the particular year of enumeration.
As in the "Pure Stand Approach" if 
total estimated
area of individual crops add to more 
or less than
the area of the plot, the area of the 
plot can then
be allocated in proportion to the 
areas of individ-
26
ual crops in the 
added total. This 
can be done in
the office after enumeration 
is completed.
f. The "Seeding 
Rate" Approach: If 
it is possible to
know how many seeds 
or plants were planted 
in a
mixed stand, standards 
for planting rates 
for "typ-
ical" plantings 
could be used to allocate 
area to
each crop in the 
mixture. This is 
frequently not
possible, however, 
and seems applicable 
only to
special cases.
It would be much better 
to be able to pretest 
each of
these methods for various 
kinds of mixtures which 
include
tree crops, row 
crops, and solid plantings 
before making a
recommendation. In 
fact, it may be found 
from pretest that
different methods 
should be applied 
to different types 
of
mixtures if they can 
be well defined. 
It is also anticipated
that, for some uses, 
there will eventually 
be a need for
allocating area in 
mixed intercropping 
to individual crops,
even though the area 
in mixed crops themselves 
may be the most
useful figure from 
an area standpoint 
and will be obtained 
in
the original measurement 
in any case.
Considering ease of 
enumeration and the 
eventual need
for area of individual 
crops, no matter 
how artificial the
concept, the following 
approach is recommended 
for multiple
cropping or mixed 
intercropping: Get 
area of the overall
plot first or early 
in the year of enumeration. 
Get produc-
tion from all the 
different crops 
throughout the enumeration
27
or harvest period, keeping the record 
separate for that plot
or plots. Then allocate area of the plot(s) 
to individual
crops using the "Area from Harvested Production" 
approach (e)
using standards developed for that purpose. 
Standards can be
developed as detailed as needed 
by areas, elevation, soil
types, etc. from measured pure stands 
in the enumeration year
or estimated from mixed stands for crops seldom 
occurring in
pure stands. Actual computations of 
areas of individual
crops can be done after the enumeration 
in the office.
17. Relay or Sequence Intercropping
Subsistence agriculture is often found in 
tropical areas
without great variations in rainfall and 
with growing seasons
where crops may need to be planted and 
harvested in overlap-
ping sequences rather than planted 
and harvested all at one
time. For example, bananas may be planted a 
few at a time
almost constantly throughout the year 
with maturity and har-
vest occurring accordingly. Beans, maize, 
potatoes, cassava,
and other crops may be planted in the same 
way. This makes it
extremely difficult for field staff to keep track 
of sizes of
plots and areas of specific crops throughout the 
enumeration
year.
Complicating the problem is the fact that 
certain annual
crops may be planted in sequence 
among perennial tree and
other crops which continue on year 
after year. Recording
areas of individual crops, then, 
will result in more area of
28
crops per year than 
there is of actual 
land area. This
should cause no problem 
as long as data are 
collected and
published for both 
land area in general 
and area of multiple
cropping. While the 
concept is not difficult, 
the carrying
out of the enumeration 
is.
Whenever this type 
of production occurs, 
and it appears
commonly in subsistence 
agriculture, there 
appears to be no
possibility for complete 
enumeration at only 
one time of the
year, or even at a few 
times a year. The only 
solution ap-
pears to be to intensively 
enumerate a manageable 
sample of
farms, commensurate 
with the budget, 
so that the enumerators
can keep detailed track 
of this almost constant 
planting and
harvesting. Ideally, 
where this practice is 
very prevalent,
the enumerator should visit 
each sample holding once a 
week to
note new plantings and 
to record harvesting. 
If this is not
possible within the budget, 
2-week or 1-month intervals 
may be
sufficient, but the 
time interval should 
probably be no more
than 1 month between 
visits at the most.
18. Changing Area and 
Location of Plots in the
Holding Within the Year
It is common in subsistence 
agricultural areas to change
areas where plots are 
located within the enumeration 
year.
For instance, all harvesting 
may be concluded on 
a certain
plot at some time 
during a year and others 
started in another
area with the original 
plot being abandaoned. 
This may occur
some distance away 
from the holding and 
the farmer may not
29
think to mention the fact to 
the enumerator. This is par-
ticularly bothersome when 
the enumerator must go 
out early in
the enumeration year and 
initially get area of all 
plots in
the holding. Considerable 
area and production on a particu-
lar holding may be missed unless 
the enumerator is constantly
alert to increased cultivation being 
started during the year
and other plots being finished 
and abandoned.
As with the relay or sequence 
problem, the only solution
appears to be to visit each sampled 
holding as often as
possible, preferably once a week 
or, at least, no less than
once a month, making sure each time to 
ask the farmer for any
action concerning new planting, and proceeding 
to measure and
add this area to the existing plot areas. 
Any areas either
prepared, planted, or harvested during 
the enumeration year
should be included in the total crop area in 
the holding at
the finish of the enumeration year 
even if it was not cleared
and planted until near the end of 
the year or was harvested
and abandoned early in the year. Judgment 
will have to be
used when preparation or harvesting 
is occurring at the
beginning or end of an enumeration 
year as to whether or not
to include the area in the holding.
19. Wild Economic Production Areas
In subsistence agriculture, 
it is common for local people
to harvest the production from 
trees and plants in certain
areas not contained in formal 
holdings, without having any
formal rights to them. This 
may represent a significant
30
amount of food consumption 
at certain times of the 
year in
some cases, and products 
sold for cash in others. This 
is
similar to certain areas in the 
Western United States, par-
ticularly the Sand Hills of Nebraska, 
where hay is harvested
by farmers from roadsides 
and railroad right-of-ways. 
If not
accounted for in "area in this 
farm" it can seriously distort
average yield figures if allocated 
to only area owned or
formally controlled by the farmer. 
This also is true of pecan
production in some Southeastern 
U.S. States where certain
pecan producers have pecan routes 
and pick up and add produc-
tion bought from other people to their production for 
selling
in one lot or for processing. This 
production may be from
trees on city lots and from 
other people not ordinarily
classed as farmers. Production of 
this type is certainly
agricultural production and should 
be accounted for if impor-
tant to an individual holding or if 
significant to an area.
The question then becomes one 
of whether or not to account for
it in formal holdings or in general 
in some way.
One recommendation is to ask if any 
agricultural prod-
ucts are harvested and consumed 
from areas not in his "hold-
ing" during the enumeration year. 
This should be done
weekly, or once a month at the 
most, so that seasonal con-
sumption of certain plants, 
fruits, and other products 
can be
accounted for. The "Area 
from Harvested Production" 
approach
should then be used to estimate 
equivalent land area neces-
sary to grow that production 
and this area may be added 
to
31
area in the formal "holding" as if it 
were a part of it.
This is similar to the hay cut from roads 
or railroad rights-
of-way mentioned above. It may be very difficult in the case
of seasonal browsing of the family from mango 
trees, etc. to
determine the amount consumed or picked. 
Also, a typical
mango tree and production would be difficult 
to determine, as
would be the area it occupies if in a pure 
stand, which is
usually not the case. Much judgment will be involved in 
these
cases.
20. Community, Multiple Person Partnerships, and
Other Tenure and Ownership Arrangements (In-
cluding Rights to Tree Production not in Hold-
ing)
The type of tenure and ownership arrangements listed in
the title are not uncommon in subsistence agriculture 
and may
result in under or spurious enumeration if not 
handled prop-
erly. For instance, a farmer may, quite honestly, 
forget to
tell an enumerator about an interest he has in a crop with
another farmer in another area or ownership or harvest 
rights
he has retained in perennial tree production from an area
where he has lived before.
Somewhat different from the partnership and ownership
rights mentioned above is community or tribal 
areas farmed in
common and the production allocated in some 
way not necessar-
ily according to labor or other inputs supplied, 
and in which
the farmer or his family has no direct rights--only 
the
32
obligation to supply 
labor or other inputs. 
In this case,
they may receive output 
from this area, or they 
may not, and
the amount received may 
have no relationship to 
labor or
other inputs supplied.
If the area is farmed in 
partnership and/or if the
farmer has some definite 
proportional right to the 
production
from an area or trees, 
the area can be estimated 
using the
"Area from Harvested Production" 
approach and added to the
area of the formal holding. 
If the farm family is only
supplying labor or other 
inputs to a communal operation, 
this
area should not be included in the 
holding and any benefits
received should be accounted 
for as "other income," as with
off-farm work and income, when accounting 
for income of the
farm family as a whole or, in 
consumption figures if these
details are obtained 
in the survey.
21. Ponds, Lakes, and Other Water 
Areas in the
Holding
Normally, such areas would 
not be included in areas 
in a
holding. They would, however, 
be accounted for in general
land use for the whole nation 
or other political unit. In-
creasingly, however, there 
is an interest in water area
available on a holding for various 
purposes, including for
stock watering, for drinking 
and other home uses, and, 
in-
creasingly, for confined 
fish production. Increasingly, 
man-
made or other water areas 
are being stocked with 
fish and
fish harvested and consumed 
or sold.
33
Areas of water on a formal holding, or controlled by the
farmer, should be measured or estimated as the area which
would normally never go dry during the year. This would
include all non-flowing lakes and ponds controlled or on the
holding, but only flowing streams if within the holding.
22. Special Crops and Livestock Grown or Raised by
Only a Few Farmers
These are categories for which sampling or a formal
survey is not appropriate. While there may be only one or a
few holdings in a country growing or raising some special
crop or livestock, the production may be quite important for
a country or political unit. In other cases, while area or
value may be small, the consequence of missing the production
of some special crop or livestock in a sample survey may have
a highly unfavorable effect on the user's confidence in the
data in general. If it is well known, for instance, that
only three farmers in a political unit grow orange trees, but
a spurious amount or none is recorded when the data are pub-
lished, the usefulness of the data may be undermined, even
though expanded data from samples for more widely distributed
crops and livestock are quite accurate.
For such cases, where sampling is inappropriate, a sys-
tem of "Special Farms Cards" is recommended, compiled from
whatever manner available or appropriate. These would then
be given to the enumerators for the areas involved and they
wouldbe required to include them in their enumeration. The
34
enumerators, in each case, should ask 
the respondent of each
special farm for names of similar 
cases or of others who grow
or raise the same special crop 
or livestock. The list of
special farms should be completely 
enumerated and continu-
ously updated until no more are found.
B. DETERMINING CROP PRODUCTION, 
CONSUMPTION, AND SALES
1. A Common Unit for Measuring Production
Most sources caution that units known locally 
must be
used in obtaining production rather 
than international or
national units desired by the final users 
of the data. The
information collected from farmers can then be converted to
the desired unit using conversion rates 
supplied, either to
the enumerators in the field, or the work can 
be done after
the data reach the processing point.
Another problem, in addition to the final 
unit itself,
is that the output may not be harvested 
at the time or in the
form chosen as the standard unit. 
Again, the recommendation
is usually to get production at the time and 
in the form
harvested and convert to the standard unit 
using prepared
conversion factors.
A usual recommendation is to determine 
the unit used
locally for each crop being enumerated 
and the form in which
it is usually harvested. Then 
develop conversion tables for
the enumerator to use for each 
crop and farm situation likely
to be encountered. Enumerate 
in the common unit and form and
have the enumerator convert 
to the standard unit and form 
in
35
the field before sending the data to the processing point.
2. Techniques for Physically Measuring or Obtain-
ing Production Data
The actual physical measurement of production is a prob-
lem in itself and will probably vary with the crop being
recorded and the nature of its production, consumption, and
sales. Alternatives for measuring or otherwise determining
physical output include:
a. Ask the farmer
b. Weigh or count units of crops harvested at some point
c. Eye judgment by experienced reporters
d. Crop sampling and cutting or harvesting by the enu-
merator
e. Utilization tables
In the above, asking the farmer, eye judgment by experi-
enced reporters, and utilization tables are commonly used in
advanced economies; but most sources seem to feel they have
little place in subsistence agriculture. This leaves only the
crop sampling and actual weighing or counting technique; even
this approach has its drawbacks because of the practice in
subsistence agriculture of starting harvest before maturity
for many crops. In spite of this drawback, it seems that
actual weighting or counting of output is most desirable, if
possible. This usually must be done at the time of actual
harvesting by the farmer because of the practice of harvesting
piecemeal at different stages of maturity. It also must be
36
done at sufficiently frequent intervals 
so that accurate re-
call is no problem. Standard sized measuring 
containers must
be provided for some crops. It also would 
be desirable to
have some technique for the farmer or family 
to record small
lot harvesting between visits of the enumerator, such as a
form with pictures on it for making marks or tallys.
3. Non-Homogeneous Production
This problem results from the practice 
of harvesting at
various stages of production, as mentioned above, as well as
from different sizes and weights of units, such as tubers 
of
cassava and sweet potatoes, making counting irrelevant. It
also results from different varieties of what would otherwise
be thought of as the same crop having different 
characteris-
tics, such as cooking and sweet bananas.
The common form in which the crop is usually harvested
locally will have to be determined and conversion 
factors
devised, when appropriate. When not appropriate, as in 
the
case of cassava, sweet potatoes, bananas, which are 
non-
homogeneous, and other crops such as beans where counting 
is
not applicable, some form of weighing procedure or common
sized plastic or metal containers may be devised and used 
for
each crop. Conversion factors will then have to be devised
and applied, preferably by the enumerator 
in the field before
sending the data for processing.
4. Stage of Maturity and Intermittent Harvesting,
Consumption, and Sales
37
Quantity, and even the nature and quality, of a crop will
differ greatly depending upon whether it is harvested early or
at the time accepted as "maturity." The point of harvest or
enumeration will affect such aspects of the product as:
a. The carbohydrate, protein, and vitamin content
b. The water content and, therefore, density or weight
of the output per cubic volume
c. Pest damage before harvest from birds, insects, and
animals
d. Field losses due to shattering at harvest, transport
to storage, and
e. Losses in storage from pests, rot, or other
deterioration
For example, maizeeaten as green corn before harvest is
quite a different product from mature maize, so much so that
underdeveloped countries have been known to discourage the
practice of pulling and eating green maize before maturity
because total food value was greater at maturity and the prac-
tice reduced the country's food supply. Similarly, the people
may dig and start eating the immature cassava roots as early
as 9 months after planting and continue digging up to as much
as 2 years or more. Obviously the young tubers have a dif-
ferent nutrient, fiber, and water content Fromn the older
tubers. Also, total food yield from an area will be less if
harvest occurs too early before total maturity.
The question becomes, then, one of definition. Is the
38
desired figure for production 
the actual weight of 
the product
harvested regardless of the 
stage of maturity, the potential
amount at some chosen point of 
harvest or maturity, or some
amount actually reaching consumption 
after various losses and
deteriotation. If there were no complications 
due to harvest
and consumption before "maturity," 
most sources in developed
countries opt for the concept 
of production reaching the
initial storage point immediately 
after harvest and before any
appreciable loss due to storage 
occurs. But, because harvest
and consumption may often occur 
before maturity in subsistence
agriculture, the recommendation is 
usually to record produc-
tion as at the first point of storage just 
after harvest, plus
estimated amounts harvested before maturity, 
if the product is
defined as the same. In cases where the 
product is quite
different (i.e., green maize vs. 
mature dry maize), judgment
will have to be used as to whether they should 
be defined as
two different products (i.e., a vegetable 
vs. a field crop).
If defined differently, the areas of 
each will have to be
allocated accordingly for reporting 
purposes.
5. Production from Intercropped Areas
Production from intercropped areas 
should cause no par-
ticular problem except that associated with 
units, form,
time, and maturity of production 
already mentioned. Record
as recommended above, but keep 
production amounts by specific
plot of intercropped area, especially 
mixed intercropping and
sequence intercropping, in order 
to be able to estimate area
39
of each crop later.
6. Unharvested and Incompletely Harvested Produc-
tion
For various reasons, certain crops may reach 
maturity and
remain unharvested or incompletely harvested. 
These crops may
have taken the same amount of labor and other inputs to grow
as in other areas which are completely 
harvested. A good
example of this is cassava, which is frequently 
planted as a
"reserve" crop in case of a drought and not 
harvested, or
incompletely harvested, if not needed. 
The problem is a
conceptual one. Are we interested in actual 
production or
only that harvested?
While estimation of potential production is done 
in some
countries for some purposes, most 
sources would estimate only
production actually harvested in the enumeration 
year. Area
prepared, planted, and harvested in 
the enumeration year also
will have been recorded, however, as will 
areas simply car-
ried over from previous years as with perennials.
7. Production in Progress but not Mature in 
Survey
Year
This problem is common for such crops as 
tree crops,
pineapples, bananas, and certain other 
crops. It is also
common because of the nature of year-round production in
certain tropical areas. The recommendation is usually to
record production when it occurs and not "potential" 
produc-
tion. Area should be accounted for in land use data, how-
40
ever.
8. Wild Production
This is very important in some areas and in some seasons
of the year in subsistence agriculture. The recommendation
is for the enumerator to visit the holding frequently, taking
care to ask each time about this type of production. He
should anticipate it also from his knowledge of the area.
This is an argument for picking local enumerators as far as
possible.
9. Production from Community, Multiple-Person, or
Partnership Holdings
This may account for a large amount of the nutrients
consumed or sold by holdings and would be an important omis-
sion if missed in many cases. If equivalent area can be
estimated, the recommendation is usually to count this pro-
duction as from the holding. If the farmer and his family do
not have definite exclusive rights to the output, but re-
ceived output in some manner not related to labor or other
inputs, consider anything received as "other income." Ac-
count for it in family consumption, however, if applicable.
10. Production from the Farmstead, Kitchen Garden,
and Other Non-Crop Areas of the Holding
It may be that some production does not come from areas
identifiable as area of cultivated crops because the area
falls below an arbitrar minimum area to record. Also, impor-
tant production may come from small "kitchen gardens," from
41
random trees in the farmstead and along boundaries, and from
vines on roofs and in trees. All of this type of production
may be very important or it may be insignificant. The ques-
tion becomes one of choosing a criteria for deciding when it
is important enough to record and how to record it.
In many counties, especially highly developed 
countries,
if this production is all for home consumption, it is usually
not recorded. However, in subsistence agriculture, where
production for home consumption is usually the most 
important
category for disposal of production, the recommendation 
is to
record productions of specific crops if it 
meets one of two
criteria: (1) It would convert 
to the minimum area to be
recorded from an area standpoint, or (2) 
it is a special crop
grown by only a few holdings, or 
having a high value, and to
miss it would result in lack 
of usefulness in the survey by
users. If it would convert to less than 
some minimum area,
record production anyway with appropriate notation.
11. Accounting for Home Consumptions 
vs. Sales
vs. Manufacturing vs. Kept 
for Seed
This is a matter of design of the 
questionnaire or other
recording instrument. The recommendation 
is to design col-
umns opposite each crop in each section 
of a questionnaire to
reflect the above breakdowns of use of output as appropriate.
Tailor each section for different types of crops to 
their
particular characteristics.
42
12. Production from Multiple Product Crops (Greens
and Roots from Cassava, etc.)
It is common in parts of Africa to harvest the leaves of
the cassava for greens several times during the life of the
plant and then harvest the tubers. There may be other crops
of this nature. If preliminary surveys indicate that two
products from the same plant are important, the questionnaire
will have to be designed to reflect this. However, if it is
necessary to estimate area from production, estimates should
be made from the most important product of the two, perhaps
from a value standpoint if sold.
13. Production of Forage Crops
This is usually not an important activity in much of
subsistence agriculture. However, if it occurs, area and
production should be handled as for any other crop. The
recommendation is to design a question on the questionnaire
as for any other crop if pretest indicates that it is impor-
tant enough to warrant it. Otherwise, record in an area on
the questionnaire designed for "Other Field Crops."
14. Production of Forest Products
Many kinds of forest products, such as firewood, char-
coal, and poles, will not be formally grown or produced on a
holding but will usually be gathered off the farm in "wild"
areas. The recommendation is to record only production of
forest products that are sold from wild production. Area may
be accounted for as part of the holding, however, for certain
43
cultivated forest products, such as capok, which can be
treated as any other tree fruit or product from the holding if
the tree is not destroyed in harvesting.
15. Production from Ponds, Lakes and Other Water
Areas on the Holding
This would be fish or any other water-living aquatic
product. The recommendation is to have a screening question
if there is water area on the holding to ask if there are
fish or other aquatic products. If so, record in an area of
the questionnaire designed for these products.
16. Crops Grazed by Humans (Eaten Off the Trees or
in the Field)
This is a difficult problem in subsistence agriculture.
While it may constitute an important source of nutrients
during certain times of the year, there is usually no way to
accurately estimate amounts. For instance, in many tropical
areas mangoes are grazed off the tree in large amounts during
the season and constitute an important part of the diet.
Such production must be estimated in some way other than from
a survey. Perhaps this information would be obtained in more
"in-depth" anthropologic studies which record activities,
food eaten, etc. of families.
C. ENUMERATING LIVESTOCK NUMBERS AND PRODUCTION
There are many livestock systems in subsistence agricul-
ture, some of which are easy to enumerate and some which are
not. If livestock are confined on the holding, or if the
44
holding is the base to which the livestock are brought at
night, there is usually little problem of enumeration. The
problem becomes one of questionnaire design and technique of
asking the questions. If the livestock system is nomadic,
however, and the economic value of livestock is secondary to
its social or prestige value, enumeration becomes more of a
problem. Not the least of the problems of nomadic herders is
that there is no home base or holding which can be located on
a map and enumerated. Also, unlike with crops, there are
many taboos and fears of revealing livestock numbers, as well
as problems of absentee owners who do not want their inter-
ests known. In many cases, there is a shared arrangement
between an owner and a herder, with the herder getting the
first born calf, milk, etc.
This section is not as long as the sections on crops and
their problems of enumeration, not because the problems of
enumeration of livestock and their products are not diffi-
cult, but because the difficult problems can be summarized in
fewer categories. Also, no attempt will be made to break out
the problems of specific breeds of livestock where the prob-
lem is the same for all.
1. Determining Kinds of Livestock, Ownership, Num-
Regardless of the breed of animal and nature of produc-
tion, the following information is needed on the question-
naire, or other instrument, in order to cross-check accuracy
45
and to obtain information usually desired by data users for
each type of livestock:
- Inventory at beginning of year
- Number born during the year
- Number killed for home consumption
- Number bought (or obtained in some way)
- Number sold
- Number died, lost, or stolen during the year
- Inventory at end of year
In addition, age and sex groupings, as well as a distinction
as to work stock, breeding stock. etc. may be obtained.
2. Migratory Herds
Obtaining any kind of information on migratory herds is
one of the more difficult problems associated with enumeration
of livestock. In addition to being difficult to find, there
are superstitions concerning giving livestock numbers as well
as fear of the tax collector in many parts of the world.
Traditional enumeration of holdings will usually not be appro-
priate for these herds and some other method may be necessary.
Aerial photos may help, but they have the disadvantage of not
being able to determine ownership and other characteristics of
a herd. Government veterinary programs to control different
pests may also give fair estimates of numbers in some cases,
as most migratory people love their animals and frequently
participate in such programs. Also, special enumerators may
be needed for this type of livestock production and, possibly,
46
all livestock production. 
Most references and 
experience
indicates that it 
takes a special type 
of person, interested
in livestock, to 
be a good livestock 
enumerator. Also, most
references indicate 
that inquiries about 
migratory herds
should be an entirely 
separate activity from 
enumeration of
other types of subsistence 
agricultural holdings.
Migratory herds are 
not now as important 
as they once
were in developing 
countries. However, 
to the extent that
they exist, they should 
be considered much as 
the "special
farms" mentioned earlier 
and not sampled and enumerated 
in
the usual way. Rather, 
a livestock "expert" 
should be as-
signed to the enumeration 
for this class of livestock, 
deter-
mining the sample frame 
from whatever sources 
are possible,
and proceeding with a 
complete enumeration of 
these herds at
intervals of, as near 
as possible: (1) 
beginning of the
year, (2) April 1, (3) July 
1, (4) October 1, and (5) 
end of
year. Otherwise, the 
information collected 
should be the
same as for livestock 
on regular holdings.
3. Owners of Livestock 
Herded by Someone Else
It is common for people 
in cities in some developing
countries to have interest 
in or own livestock 
which are kept
on holdings by someone 
else or are kept in 
migratory herds.
The recommendation 
is to enumerate livestock 
as "on this
holding" or "in this 
herd" regardless of 
ownership. The key
should be, who has 
active control over 
the animals as to
husbandry, etc. Ownership 
may be an item obtained 
on a
47
questionnaire, but all livestock 
should be considered 
a part
of a holding or managed 
herd regardless of ownership.
4. Frequency and Time 
of Enumeration
This will vary by type 
of livestock and whether 
or not
there are livestock products 
to enumerate such as milk and
eggs. It will also vary by holdings 
and by whether or not
herds are migratory. 
Livestock and products raised 
on, or
associated with, a holding 
which also produces 
crops should
be enumerated in much the same 
manner and at the same time
the crops are enumerated--about 
every week or two or, at 
the
most, no less than once 
a month. For migratory 
herds, they
should be enumerated by 
an "expert" separately. 
Any products
from these operations should 
be estimated at time of 
inter-
view, which may be approximately 
quarterly (January 1, April
1, July 1, October 1, and 
December 31).
5. Enumeration of Numbers of Poultry 
and Small
Animals
Poultry and small animal 
production have different 
char-
acteristics from large animal 
production. While inventory 
at
even one time may give a 
good indication of the relative
importance or numbers 
of large animals throughout 
the year,
it may not adequately characterize 
small animal numbers
raised, because more 
than one batch may be 
born, raised, and
consumed or sold between 
the beginning and end 
of a year.
The recommendation is 
to enumerate poultry 
and small animal
numbers at 1-or 2-week 
intervals at the same 
time that
48
crops are enumerated, obtaining:
- Inventory January 1 by categories desired
- Numbers every 2 or 3 weeks including:
-- Numbers on hand by categories
-- Numbers born or hatched
-- Numbers consumed by categories
-- Numbers sold by categories
-- Numbers died, lost, or stolen, by categories
- Inventory December 31 by categories
6. Animal Products (Milk, Eggs, Meat, Skins, Dried
Fish, Honey, etc.)
These products may cause some problem on subsistence
farms because they are usually for home consumption and the
farmer and his family may have no idea as to quantities. The
important thing is to enumerate frequently (every 1 or 2
weeks) to facilitate recall. Production should be in the
units known, or standard measures should be provided for milk,
honey, etc., with provisions for the farmer to record amounts
in some way between visits. Eggs and skins can be counted.
Meat and fish can be weighed, but this could be very difficult
and may only be determined easily if sold. If used for home
consumption, amounts would probably need to be estimated for
fish. The whole animals used for home consumption would,
otherwise, be recorded only by number.
7. General Conclusion on Livestock and Livestock
Products Enumeration
49
Authors which treat the subject are frequently frustrated
at the inadequacy of livestock and livestock product enumera-
tion in subsistence agriculture. To quote one, "These are not
very satisfying proposals for handling livestock statistics,
but to judge from reports of statistical work in this field in
various developing countries, satisfaction is rare" (5, p.
81). On another page Hunt (5) states, "Collection of live-
stock statistics in developing countries is usually either a
reasonably straightforward process or an intensely difficult
one." In another case he states, "The history of work in this
field suggests that in any but straightforward cases, where
the herds are part of the regular system of a settled agricul-
ture, the prospects of initial success are small" (5, p. 78).
While Hunt may seem unduly pessimistic, he represents most
authors.
This author's own ideas on the subject of livestock and
livestock products enumeration in subsistence agriculture
were confirmed by review of literature and by personal expe-
rience. It seems that livestock and products take a special
enumerator with special knowledge of the livestock systems in
the country. While this is also true of crops, it is true to
a much greater extent with livestock. It appears that enu-
merators chosen for livestock enumeration must be a "special
breed" with a "special knack"; and, if the enumerator does
not have this "feel" for livestock, he is doomed to failure
from the start.
50
SELECTED REFERENCES
(1) Ashcroft, Joseph, 1974. 
On the Art and Craft of 
Col-
lecting Data in Developing Countries. 
Paper presented to
the Workshop on Field Data 
Collection in Rural Areas of
Africa and the Middle East, Beirut, 
Lebanon, December 8-
14.
(2) Collinson, M. P., 1972. Farm 
Management in Peasant Agri-
culture, A Handbook for Rural 
Development. New York,
Praeger.
(3) Dalrymple, Dana G., 1971. 
Survey of Multiple Croppng 
in
Less Developed Nations. ERS, 
USDA, Foreign Ag Econ
Report No. 91, Oct.
(4) El Hadari, A. M., 1974. Problems 
of Field Data Collec-
tion in Rural Areas. Paper presented 
to the Workshop on
Field Data Collection in Rural Areas 
of Africa and the
Middle East, Beirut, Lebanon, 
Dec. 8-14.
(5) Hunt, K. E., 1969. Agricultural 
Statistics for Develop-
ing Countries. Oxford.
(6) Kearl, Bryant, editor, 
1976. Field Data Collection 
in
the Social Sciences: Experiences 
in Africa and the
Middle East. Agricultural 
Development Council, N.Y.
(7) Norman, David W., 
1977. Problems Associated 
with the
Gathering of Technical, Economic, 
and Social Data on
the Farm Level in West 
Africa. Paper presented 
at the
International Conference 
on the Economic Development 
of
Sahelian Countries, University 
de Montreal, Oct. 13-14.
51
(8) O'Sullivan, John M., 1980. Agricultural Survey 
Methods
in Less Developed 
Countries. Dept. 
of Ag. Econ., Auburn
University, and 
Center for Rural Development, 
Tuskegee
Institute, unpublished 
paper.
(9) Stallings, James L., 
1980. A State-of-the-Arts 
Paper
(SOAP) on Techniques 
of Enumeration of Intercropping 
and
Associated Cultivation 
and Livestock-Numbers 
and Products
in Subsistence Agriculture 
in LDC's. Personal Service
Contract No. AID/afr-C-1607 
(Revised), Project No. 
698-
0135 "Collaborative 
Survey and Analysis 
of Agriculture--
Rwanda."
(10) Ucherdu, Victor 
C., 1974. Rapid 
Survey Techniques 
for
Interdisciplinary 
Research. Paper presented 
to the Work-
shop on Field Data 
Collection in Rural 
Areas of Africa
and the Middle East, 
Beirut, Lebanon, 
Dec. 8-14.
52