The document summarizes the Smallholder Farmer Innovation Programme in South Africa, which aims to optimize conservation agriculture systems for non-commercial and semi-commercial smallholders from 2013-2015. It provides an overview of the methodology, equipment, soil recommendations, monitoring and evaluation, partnerships, and the scaling model used. Key points include: the programme worked with over 200 smallholder farmers across several villages in 2013-2015; it used a "farmer level experimentation" approach to test different planting and crop options; soil health was monitored using tests like Solvita and Haney; and visual indicators were used for local monitoring of conservation agriculture practices.

1. GRAIN-SA SMALLHOLDER FARMER
INNOVATION PROGRAMME
Optimising the Conservation Agriculture system for non-
commercial and semi-commercial smallholders
2013-2015

2. WHAT IS THE SMALLHOLDER SYSTEM?
• Programme overview; methodology process, experimentation
• Equipment
• Soils: fertilizer recommendations, soil health, crop diversification,
spacing
• Monitoring and Evaluation

3. • Partnership: Mahlathini Organics, SaveAct Trust, Lima RDF, Siyazisiza... Grain- SA
– Saving and credit groups and organised farmers organisations and cooperatives to
work within the whole value chain- Inputs- production- storage- marketing.
• Horizontal scaling model starting with a nodal village in each area and expanding within
and between villages:
– 2013: 3 Villages Matatiele and 3 in Bergville (total trial participants: 50
– 2014: 7 villages Matatiele, 9 villages Bergville (Total trial participants: 100
– 2105: 4 villages Matatiele, 2 Mt Fere, 2 Mt Ailiff, 10 Bergville, 2 Nkandla (total trial
participants: 210)
• Farmer volunteers (mentor 5 farmers each), local facilitators (coordinate in each village),
farmer centres (for each node surrounded by 3-5 villages close by for input provision
and production advice)
Description of Model and Process

4. Scale of farming operation
CATEGORY Non commercial
smallholders
Semi commercial smallholders Commercial smallholders in
loose value chains
% of people 72 23 5
Farmer priorities For household
Minimal inputs bought
For household and some selling which
supplements household income.
Minimal inputs bought
Consumption and sale in various
percentage mixes but moving to more
sales. More inputs bought
Traction Hand cultivation, limited labour Hand cultivation, animal traction, some
family labour
Animal traction, tractors, family and
hired labour
Land size > 0.1ha 0.1-1ha. 1-2.5ha
Farm productivity,
labour access
Inter cropping, micro-dosing with
fertilizers, herbicides, pesticides,
cover crop rotations
Inter cropping, micro-dosing with
fertilizers, herbicides, pesticides, cover
crop rotations
Inter cropping, micro-dosing with
fertilizers, herbicides, pesticides, cover
crop rotations
Access to improved
agricultural tech and
information
Hand planters,
Learning groups, working groups
Hand planters animal drawn planters
Learning groups, working groups
Animal drawn planters, two row
planters
Learning groups, working groups
Access to financial
services and local
organisation
SCGs, CIGs - SCGs, CIGs
Bulk buying
SCGs, CIGs, formal loans and bank
accounts (group and individual
Bulk buying

5. Farmer level experimentation
• Year 1: Predefined with the research team:
– Choice of planting method; hand hoes, hand
planters, animal drawn planters
• Year 2: Choices and options within the same
overall design:
– Different varieties maize (white yellow, OPV,
hybrid)
– Different varieties and types of legumes
– Early planting
– Manure and fertilizer combinations
– Targeted fertility regimes and pest control
measures
• Year 3: Own design of experiments by
participants :
– Intercropping vs crop rotation options
– Summer and winter cover crops
– As well as options for year 2.
Incremental change in yr 1,2,3

6. Trial layout and design
• Herbicide application (Round Up or
Gramoxone, Dual Gold)
• Close spacing of tramlines (basins
and furrows)
– 50x50 maize
– 25x10 beans
• Lime application in planting lines
• Micro dosing of fertilizer; generic
recommendation
• One weeding is enough before
canopy cover is reached. Must be
early weeding: 2-3 weeks.
• Winter cover crop relay planting
(Feb-March.
Intercropping and close spacing

7. • Maize and bean yields somewhat lower in 2nd season- due to
adverse weather conditions
• CA trial plot yields consistently higher than control plot yields
1,24
0.26
2.25
2,06
3,74
3,63
1,14
0 0.5 1 1.5 2 2.5 3 3.5 4
Year 1
Year 2
Average yields for maize and beans in Bergville; 2013-2015
Average of (Cowpea (1-1.2 t /ha), - Trial plot
Average of Maize (4-10t/ha) - Trial plot
Average of Maize (4-10t/ha) - Control
Average of Yields: (Beans 1-2.5 t /ha), - Trial
plot
Yields comparison over 2 seasons in Bergville

8. Trial summaries -2 seasons; Matatiele and Bergville
Bergville Matatiele
2014 2015 2014 2015
No of villages 3 9 4 10
No of trial participants 28 83 23 (61) 16 (63)
Area planted (trials) 13,1ha 42,6 ha 0,36ha 0,3ha
Average yield maize 3,74 t/ha 3,63 t/ha 0,95t/ha 0,7 t/ha
Min and max yield maize 2-4,3 t/ha 1-6,7 t/ha 0,25-1,65 t/ha 0,34-1,84 t/ha
Actual amount of maize pp 233kg 576kg 15kg 64kg
Rand replacement value R 1 600,00 R 4 492,80 R102,75 R 499,20
Average yield beans 1,24t/ha 0,26t/ha 1,26t/ha 0,34t/ha
Trial summaries over 2 seasons; Matatiele and Bergville

9. SOILS
• Fertility and fertilizer recommendations
• Soil health
• Crop diversification, moisture and run off control, spacing
• Bergville: Avalon Clay-loams- (30-50% clay)
• Matatiele: Sandy Avalons/Longlands (5-12% clay)

10. •Based on average of 8-12 soil
samples from an area.
•To make it as simple as possible
for participants to remember
•Use as few different options as
possible
•Stick to single nutrient options
for a closer fit to the average
•Has worked well except for a few
outliers- related in Bergville to
high acidity issues and in
Matatiele to extreme infertility
and badly damaged soils.
Generic fertilizer
recommendations
Fertilizer and Manure Quantities
NAME Element Amount/ha Amount/m2
Measuring
LAN
(Basal)
28% N 100kg/ha =
355kg LAN
35.5g/m2
18 g /meter of row
9 g/planting basin
2 match boxes
1 matchbox
½ match box
LAN (top
dress)
28%N 40kg/ha
~140kg LAN
10g/m2
5g/meter row
2.5g/planting basin
2 cool drink bottle tops
1 coo drink bottle top/ 1
teaspoon
½ cool drink bottle top/ ½
teaspoon
Supers 10.5% P 40kg/ha =
380kg supers
39g/m2
20g/meter row
10 g/planting basin
2 match boxes
1 match box
1/2 matchbox
MAP (33)
(supplies
N and P)
11% N,
22% P
55 kg/ha P,
30kg/ha N =
250kg MAP
18g/ m2
8,3g/meter row
4,2g/planting basin
1 match box / 1 tablespoon
½ match box
1 cool drink bottle top / 1
teaspoon
KCL 50% K 20kg/ha =40
kg KCL
4g/m2
2 g/meter of row
1g/planting basin
1 cool drink bottle top
½ bottle top (1/2 teaspoon)
¼ bottle top (1/2 teaspoon)
Lime 1ton/ha 1kg/m2
500g/meter row
250g/planting basin
1 large jam tin
1/2 large jam tin ( or 1 normal
food tin)
1/4 large jam tin (1/2 food tin)
MANURE:
Cattle/
goat
5kg/t N
2kg/t P
3kg/t K
10 tons/ha 10kg/m2
5kg/ meter row
2.5kg/planting basin
4 heaped spades full
2 spades full
1 spade full/ 4 large jam tins

11. • Soil health:
– Availability of organic and inorganic fractions of nutrients
– microbial activity in the soil
• Increase ratio of organic: inorganic, build reserve –
through diversity
• Reduce inorganic fertilization
• Solvita Test measures soil CO2 respiration (in ppm)
– The estimated plant useable N in the soil is 50% of the CO2-
C expressed in lb/a (x 1,12=kg/ha)
– The quantity of soil microbial biomass is roughly 20 x the
CO2-C rate
Soil Health, Example
Range of values
>100: High N - sufficient for
crops. Biomass .2500ppm. Well
supplied organic matter
61-100: Mod-high N - limited N
required. Adequate organic
matter
31-60: Mod. Supplement with
N. Requires appl of stable org
matter
6-30:Low-Mod. Apply N.
Biomass<500ppm. Supply org
matter
0-5: Significant fertilization
needed. Very inactive soil.
Biomass<100ppm
AREA NAME DATE SAMPLE Soil pH
CO2-C
(ppm C)
MATATIELE (EC)
Sehutlong
Lelatsa
Thuso Jul-15
1st year intercrop
trial 5,6 65,7
Lelatsa
Thuso Jul-15
1st year intercrop
trial with cover crop 5,6 134,1
Lelatsa
Thuso Jul-15 1st year control 5,7 74,6
Sehutlong Jul-15
Veld baseline
samples 5,9 141,8
Soil Health
Score
11,01
14,98
9,03
14,55

12. • Haney test: Soil health score
– CO2-C/C:N + Organic C/100
+Organic N/10 = Soil health Score
• Low C:N ratio -> N released – good for
cropping
• Build N reserve
Soil Health
CO2-C
(ppmC)
Organic C:N
ratio
Total Org C
(ppm)
Org N ppm
Soil Health
Score
1st year control 69.6 10.6 140.7 13.2 9.4
1st year intercrop trial 155.6 9.2 213.0 23.2 21.4
1st year intercrop trial with cover crop 107.2 9.7 124.3 12.7 13.1
Veld baseline samples 141.8 13.2 217 16.4 14.55
0.0
50.0
100.0
150.0
200.0
250.0
Haney Test: EC (Matatiele) N=6
•Soil health highest for
intercropping
•Soil health lowest for
‘normal’ or control
cropping
•Need more organic
matter in the soil to
show full effect of cover
crops

13. • Haney test: Soil health score
– CO2-C/C:N + Organic C/100 +Organic N/10 = Soil
health Score
• Low C:N ratio -> N released – good for
cropping
• Build N reserve
Soil Health
CO2-C
(ppmC)
Organic
C:N ratio
Total
Org C
(ppm)
Org N
ppm
Soil
Health
Score
2nd year intercrop trial 120.7 12.1 141.1 12.2 12.9
Veld baseline sample 77.7 19.5 318.0 16.8 8.6
0.0
50.0
100.0
150.0
200.0
250.0
300.0
350.0
Haney Test: Bergville (Emmaus) N=5•Soil health higher
for intercropping;
(yr2)
•Need more
organic matter in
the soil to show
full effect

14. Soil Health
•Intercropping
with cover crops
builds N reserve
over time
•Balance with N
release for best
soil health score
•Favour legumes in
cover crop mixes
0
2
4
6
8
10
12
14
16
18
30/70
40/60
50/50
60/40
70/30
30/70
40/60
50/50
60/40
70/30
30/70
40/60
50/50
60/40
70/30
N released N reserve Soil Health Score
Bergville: N released and reserve for trials and veld baseline N=8
2nd year intercrop trial
2nd yr intercrop and cover
crops
Veld baseline sample Bgvl
• N reserve: Humus and reserve for release as organic N.
– Build this up while balancing with N release

15. Soil Health
•Here little N reserve
– even for veld
baseline
• Intercropping starts
to build the reserve
while increasing N
release.
•Favour legumes
strongly in cover
crop mixes
• N reserve: Humus and reserve for release as organic N.
– Build this up while balancing with N release
0
5
10
15
20
25
30/70
40/60
50/50
60/40
70/30
80/20
30/70
40/60
50/50
60/40
70/30
80/20
30/70
40/60
50/50
60/40
70/30
80/20
N released N reserve Soil Health Score
Matatiele: n relseased and reserve for controls, intercrop trials,
with cover crops and veld baseline N=6
1st year control
1st year intercrop trial
1st year intercrop trial with
cover crop
Veld baseline samples

16. • To ensure soil organic matter build
up
• Improve soil fertility
• Improved soil structure and soil
health
• Reduce prevalence and types of
weeds
• Management of
pests and disease
incidence and severity
Crop diversification

17. • Temperature
and moisture
control
• Reduces runoff
And close spacing
Mrs Simephi Nkosi- Emmaus Jan 2015 Intercrop and GM mono crop
Mr Mtoleni Dlamini- Stulwane Jan 2014 Traditional planting and intercrop

18. MONITORING AND EVALUATION
• CA indicators and scoring
• VSA- Visual Soil Assessment
• PES- Payment for Ecosystem services

19. Local monitoring of visual indicators
• Indicators have been chosen
that can be monitored visually
and throughout the growth
season of the crop. The
intention is twofold:
• To assess different indicators
for their reliability ,robustness,
sensitivity to change in
management practices and
the envirnoment
• To design a system that can be
locally implemented by
facilitators and farmer.
• % soil cover at planting (From 0% - no
cover to 100% full cover); Cover of the soil
looking from above- can be crop residue,
weeds, mulch, grass etc)
• % crop canopy cover at 6-8 weeks (From
0% - no cover to 100% full cover); Cover of
the soil looking from above- crop cover/
canopy) – a new indicator
• % Weed infestation (0%- very high weed
incidence, complete yield loss; to 100%- no
weeds zero yield loss)
• % Pest occurrence (0%- very high infestation
, complete yield loss, to 100%- no insect
pests and zero yield loss)
• % growth ; (germination, colour, height,
health)
• % growth of cover crop (from 0% – not
planted or not germinated to 100% full
germination and excellent growth) – A new
indicator that has been included this year,
but not yet added to the overall score for
each participant

20. Values
Row Labels Runoff
Average of
Cover
crops
Average of
%Cover
(at
planting)
Average of %
Weeds (0%)
– high weed
infestation –
to (100%) no
weeds,
Average
of % pests
(0%) bad-
(100%) –
Average of %
growth
(germination,
colour,
height,
health)
Average of
Overall
score (10)
Lelatso Thuso 1 70% 10% 75% 85% 30% 5
Bulelwa Dzingwa 1 10% 15% 75% 85% 65% 6
Mahutlong Dodo 0 1% 10% 65% 90% 70% 5,9
Majaokbo Sabasaba 0 0% 8% 15% 90% 20% 3,3
Mamolelekeng Lebuoea 1 100% 10% 90% 80% 80% 6,5
Manapo Moshoeshoe 1 1% 5% 55% 80% 60% 5
Manyalleng Sikhosana 1 70% 10% 70% 85% 70% 5,9
Matshepo Futhu 1 60% 10% 55% 85% 65% 5,4
Mohajane Kanetsi 0 1% 10% 15% 80% 75% 4,5
Mongezi Bhekaphezulu 2 0% 8% 40% 80% 55% 4,6
Nkosiyamankwali
Maqungo 0 0% 5% 25% 80% 40% 4
Nokuphiwa Phekula 1 0% 5% 10% 90% 35% 3,5
Nthabiseng Moshoeshoe 1 0% 10% 35% 80% 65% 4,8
Siyabonga Maqungo 0 0% 5% 10% 80% 10% 2,6
Thabiso Dihollo 0 0% 13% 45% 78% 75% 5
Tsolonae Mapheele 2 55% 15% 63% 85% 58% 5,5
Grand Total 10 21% 10% 45% 83% 54% 4,8
CA Scores for Matatiele; 2015 (N=17)

21. • Control plot scores are lower
on average than the trial plots
• Trial plot scores for the 2nd
year are higher than for the
1st year. This is expected due
to better management of
trials by participant farmers as
they get used to the process
and are able to improve
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
CA scores for control and trial plots; Stulwane
2 seasons (N=11)
Control - Year 1
Control - Year 2
Trial plot - Year 1
Trial plot - Year 2
A comparison of scores over 2 seasons for Stulwane; Bergville

22. Visual indicator of Soil Quality Visual Score (VS)
0 = Poor conditions
1 = Moderate conditions
2 = Good conditions
Weighing VS Ranking
Soil Structure  3
Soil porosity  3
Soil colour  2
Number and colour of soil mottles  1
Earthworm counts  2
Soil cover at planting  2
Crop cover at 6-8 weeks  2
Soil depth  2
Run-off  2
Ranking Score (sum of VS rankings) Max =38
Crop emergence (% germination)  3
Crop growth and height (%; overall
growth and colour- relative
height at time of assessment)
 3
Weed infestation  2
Crop yield  3
Size and development of root
system
 2
Surface ponding / water infiltration  2
Production costs  2
Ranking Score (sum of VS rankings) Max =22
VSA; Visual soil assessment

23. CA monitoring
scores
VSA Soil scores VSA plant scores Yields
Above average ≥7 >28 >15 3-8.9 tons/ha
Average 5-6.9 11-28 7-15 1-2.9tons/ha
Below average 3-4.9 <11 <7 ≤1ton/ha
17
17
22
20
20
12
19
16
19
19
16
22
29
31
31
29
19
29
29
24
26
29
6.5
7.625
8.625
7.96
8.125
4.75
8.25
7.875
5.45
7.95
8.625
0 5 10 15 20 25 30 35
2,7
3,6
4,8
6,7
5,7
2,2
3,9
1
5,1
3
1,8
Year2Year2Year2Year2Year2Year2Year2Year2Year2Year2Year2
Bange
ni
Dlami
ni
Cazile
Zimba
Cupile
Buthel
ezi
Dlezak
he
Hlong
wane
Khule
kani
Dladla
Landil
e
Nsele
Maket
hi
Dladla
Mthol
eni
Dlami
ni
Phasa
zile
Sitheb
e
Thulisl
ie
Hlong
wane
Zaman
i
Dladla
Trial plot - Sum of Overall score
(10)
Trial plot - Sum of VSA Scores soil
Trial plot - Sum of VSA scores
plant
Combination of yields, CA scores and VSA

24. Thank You
Erna Kruger
November 2015