• TABLE OF CONTENTS
HIDE
 Front Cover
 Title Page
 Introduction
 1988 on-farm test of effect of...
 1988 on-farm test of millet variety...
 1988 on-farm test of sorghum seed...
 1988 on-farm test of sorghum...














Group Title: 1988 annual report Testing & Liaison Unit (TLU) : Agronomic Research Center, Maroua
Title: 1988 annual report Testing & Liaison Unit (TLU)
CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00073335/00001
 Material Information
Title: 1988 annual report Testing & Liaison Unit (TLU) Agronomic Research Center, Maroua
Alternate Title: 1988 on-farm tests Testing & Liaison Unit (TLU)
Agronomic Research Center, Maroua
Physical Description: 24 leaves : ill. ; 28 cm.
Language: English
Creator: Russell, John
Fobasso, Martin
Cameroon -- Ministry of Higher Education, Computer Services, and Scientific Research
Publisher: MESIRES
Place of Publication: Cameroon
Publication Date: 1988
 Subjects
Subject: Agricultural development projects -- Cameroon -- Maroua   ( lcsh )
Agricultural extension work -- Cameroon -- Maroua   ( lcsh )
Genre: non-fiction   ( marcgt )
Spatial Coverage: Cameroon
 Notes
Statement of Responsibility: John Russell and Martin Fobasso.
General Note: "National Cereals Research and Extension Project, IRA/IITA/USAID."
General Note: At head of title: Republic of Cameroon, MESIRES, IRA.
 Record Information
Bibliographic ID: UF00073335
Volume ID: VID00001
Source Institution: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 76932803

Table of Contents
    Front Cover
        Front Cover
    Title Page
        Page 1
    Introduction
        Page 2
    1988 on-farm test of effect of single super phosphate on peanut
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
    1988 on-farm test of millet variety and date of seeding
        Page 8
        Page 9
        Page 10
        Page 11
    1988 on-farm test of sorghum seed treatments
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
    1988 on-farm test of sorghum varieties
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
Full Text

.I.,
FEPLM.-JC-^`-. Ak^w


RA


- 5K..M- d


JES


ING


a LAW U-TNT


Natn ealesea),cha- atA~nD s
Project WCit), IR-9-WTA-AID


~L~WWYI4-~;~-~-L--'~ -- -~.-~ ~-~---


~~ CeU7fR, I~-:


ic_ -
r'i





REPUBLIC OF CAMEROON
MESIRES
IRA






1988 ANNUAL REPORT
TESTING & LIAISON UNIT (TLU)
AGRONOMIC RESEARCH CENTER, MAROUA







John Russell and Martin Fobasso
National Cereals Research and Extension
Project (NCRE), IRA/IITA/USAID





- 2 -


INTRODUCTION

The 1988 season was one of change for TLU Maroua. The
expatriate extension agronomist who had been in Maroua since
December 1983 left in mid-August, his replacement having arrived
the first week of June. Change will continue in the 1989 season
as TLU Maroua, thanks to the addition of an agricultural
economist, attempts to continue and improve its on-farm agronomic
research efforts while developing a truly interdisciplinary
approach to farming systems research in northern Cameroon.

The most salient characteristics of the 1988 campaign were,
as usual, a function of rainfall. Good rains began in mid-May,
but were followed by a prolonged, serious dry spell in June.
Once the rains returned for good, rainfall was very high, even
excessive, through the end of September. Yields of food crops
were lower overall than might be expected from the total
rainfall; contributing factors were loss of entire stands during
the dry period in June (due to drought and/or armyworms),
prolonged periods of day-long overcast, flooding, erosion, and
lodging.

Processing of samples and data management were hindered by
the unusually large number of samples and concomitant databases,
which often exceeded our computing capacities, resulting in some
data remaining still to be fully exploited. That the tests were
successfully carried out (although with a somewhat high
proportion of failed sites), and the samples and data accurately
handled, is a tribute to the training and experience of the TLU
staff and its collaborators in SODECOTON.

The second major hinderance to maintaining a large, widely
distributed group of on-farm tests was the lack of adequate
facilities for maintenance of vehicles in the Extreme North. As
two of the TLU vehicles enter their fourth and fifth years of on-
farm work, this constraint is expected to worsen.

As in past years, the cornerstone of TLU activities was on-
farm testing of varieties and agronomic practices. In 1988,
there were four on-farm tests:

1) Single Super Phosphate Fertilization of IB-66 Peanut

2) Pearl Millet Variety by Date of Seeding

3) Sorghum Seed Treatments


4) Sorghum Variety.





- 3 -


1988 ON-FARM TEST OF EFFECT OF SINGLE SUPER PHOSPHATE ON PEANUT

Introduction

In 1984 and 1986, on-farm tests were conducted in the
Centre-Nord which included phosphate fertilization as a factor.
Both triple super phosphate and single super phosphate were shown
to increase yields significantly, but to be uneconomical. The
1988 peanut agronomy test, with two replications per site, was
designed to allow estimation of whether single super phosphate
fertilization might be economical in at least some areas of the
Extreme North and the Department of Mayo Louti.

Materials and Methods

The TLU conducted the test in collaboration with the local
cotton and rural development organization, SODECOTON. In early
April, 1988 meetings were held to explain the test to SODECOTON
Chefs de secteur, Chefs de zone and monitors.

The TLU team visited each site several times throughout the
season, to verify choice of field and take soil samples; to check
layout of plot, seeding and spreading of fertilizer; to follow
development of crop and collect data recorded by monitors; to
distribute materials for harvest; and to weigh the harvest and
collect samples.

The test was planned for 28 on-farm sites in the peanut-
growing zones of the Centre-Nord. The design was a randomized
complete block design (RCBD) with 2 replications per site.
Treatments were four levels of single super phosphate
fertilization (0, 50, 100, 150 kg SSP/ha). Each field measured
2500 m2 (50 m x 50 m); each plot consisted of 15 rows, 0.4 m
apart, and measured 6.4 m by 50 m; 5 central rows were harvested.
Plowing was required. The variety grown was IB-66, an early (105
day), upright Virginia type with medium size seed and dormancy up
to 2 weeks past maturity. The density was 100,000 plants/ha (one
plant per pocket, 0.25m between pockets). Seed was treated with
Thioral.

Seeding was between the 1st and 15th of June, depending on
first good rains. Single super phosphate was applied in bands at
seeding. Herbicide was applied by monitors according to
SODECOTON recommendations for intensive peanut. Missing pockets
were reseeded 7 days after seeding, and two weedings were done,
15 and 30 days after seeding. Field observations included
number of pockets reseeded, number of plants 15 and 30 days after
seeding, and harvest weight of pods. Samples from each plot were
taken to Maroua and shelled; shelling percentage, seed yield, and
millipede loss were calculated.





-4 -


Results and Discussion


Two of the original 28 tests were lost during the season due
to poor choice of fields resulting in excessive and uneven
erosion or in inundation; three more were lost at harvest by
mistakes in harvesting or in labelling sacks by farmers or
extension monitors.




Table 1. Analysis of Variance Over 23 Sites


Seed Yield


Pockets Reseeded


Site
Rep/Site
Rate
R x S

C.V. =


**
**


**


Table 2. ANOVA 12 Eastern Sites


Table 3. ANOVA 11 Western Sites


Pod Yield


SOV


Pod Yield


Site
Rep/Site
Rate
R x S


C.V. = 12 %





Table 4.


12 Eastern Sites


**


Site 10
Rep/Site 11
Rate 3
Rx S 30


C.V. = 13 %






Mean Pod Yields by Groups of Sites


11 Western Sites


Pod Yield (kq/ha)


1553 a
1487 ab
1439 bc
1351 c


Rate


150
50
100
0


Pod Yield (ka/ha)


1393
1386
1375
1313


LSD (0.05)= 107 ns


OVr


**


12.5 %


13.6 %


SOV


15.0 %


Rate


150
100
50
0


S% (I A f V A VJ--d Seed Yield


I-


I-


Table 4


df Pod Yield


LSD (.05) = 104




- 5 -


Over all 23 sites, there was a significant response of pod
yield and seed yield to only the first 50 kg of single
superphosphate. The levels of response to 100 kg/ha are very
similar to the results of previous TLU phosphate fertilization
tests (Figure 1). At the current cost of 7000 FCFA per 50 kg
sack of SSP, and a general average price of 75 FCFA per kilo of
unshelled peanuts at time of harvest, the approximately 80 kg/ha
increase in pod yield would not justify the fertilizer expense.

The relatively small number of rates tried, and the
concentration of most response in the initial 50 kg/ha also
prevented estimation of either a linear or a quadratic response
curve.

The lack of a significant rate-by-village interaction in the
combined ANOVA for all 23 sites means that the response of pod
and seed yields to rate of SSP did not vary significantly from
site to site (Table 1). Since it was not clear whether this
meant that there would also be no interaction of fertilizer rate
by groups of sites, a combined analysis of variance was done with
12 eastern sites as one group and the remaining 11 western sites
as the other group. The rationale for this attempt was based on
indications that soils in the western part of the Extreme North
are generally lower in phosphorous. The computer used had
insufficient memory to run the General Linear Model combined
analysis of this unbalanced design; as a compromise, two separate
ANOVAs were done, one for each group of sites (Tables 2 and 3).

Although the response of pod yield (and also seed yield, not
reported here) seems to vary from one region to the next, being
significant in the east but not in the west--therefore supporting
the supposition of differences in initial P levels--the response
in the east still appears to be uneconomical (Table 4).

Conclusions

SODECOTON is planning on recommending 50 kg/ha SSP as an
option on its fiche technique for intensive peanut. Based on
analysis by groups of sites, this recommendation should be
limited to the regions of Kaele and Mayo Danay. Further analysis
of 1988 as well as past years' data will be necessary to try to
account for differences in response based on initial soil
fertility and soil type. Future research on fertilization of
peanut should try to account for effect on production of
vegetative matter as well as on pod and seed yields.





- 6 -


RESPONSE OF IB-66 PEANUT

TO SINGLE SUPER PHOSPHATE

-- POD YIELD -+ SEED YIELD


-!:I I!~~i~.


SINGLE SUPER PHOSPHATE (KG/HA)

FIGURE 1, FROM 23 ON-FARM SITES


b C C;


1400

1300

1200

1100

1000

900-


800


a- a a
b







- b


____ _il_---


-----




- 7 -


1988 PEANUT AGRONOMY TEST

AINIALL. (rmm)


00 .


80-


60 -


IlM
MalMaMa3 Jul ju2 Ju3


Jll Ji2


JI3 Aul Au2 Au3 8e1 u2i Se3 Ocl Oc200c3


DECADE


FI 2.: AVERAGE OF 23 ON-FARM SITES


--


--- ----








- 8 -


1988 ON-FARM TEST OF MILLET VARIETY AND DATE OF SEEDING

Introduction

On-farm tests in 1986 on the pearl millet variety IKMV,
compared to local varieties, showed yields of the new variety to
be considerably lower. One explanation offered was that
coincidence of flowering and periods of heavy rains resulted in
pollen washing of IKMV. The objective of the 1988 test was to
compare IKMV at three dates of seeding to the local variety at
its optimum date of seeding, in order to identify the optimum
date of seeding of IKMV.

Materials and Methods

The test was conducted in collaboration with SODECOTON, as
described in the Materials and Methods section of the 1988 Peanut
Fertilization Test above. It was planned for 15 on-farm sites in
the northernmost, i.e. pearl millet growing areas of the Centre
Nord. The design was a randomized complete block design (RCBD)
with each site representing a single block. There were four
treatments per site: 1) Farmer-selected local variety seeded on 1
July; 2) IKMV 8201 seeded on 1 July; 3) IKMV 8201 seeded on 15
July; 4) IKMV 8201 seeded on 31 July.

Each field measured 2500 m2 (50 m x 50 m); each plot
consisted of 15 rows, 0.8 m apart, and measured 12 m by 50 m.
The six central rows of each plot were harvested. The plots were
separated by a line of peanuts. Farmers were asked not to plow.
The density was 62,500 plants/ha (three plants per pocket, 0.60m
hb twilon pnrcket n) Stid wi tri n r aI l d within TIlli t l .

No fertilizer or pesticides were applied. Missing pockets
were reseeded 7 days after seeding, and two weedings were done,
at 15 and 30 days after seeding. Field observations included
number of pockets reseeded, number of plants 15 and 30 days after
seeding, number of panicles harvested, and harvest weight of
panicles. Samples from each plot were taken to Maroua and
threshed; threshing percentage, grain yield and loss to birds
were estimated.

Results and Discussion

Thirteen tests were actually seeded correctly; successful
results were obtained from only ten; reasons for lost sites
include flooding of fields and/or excessive lodging, almost total
failure of IKMV to set seed, or almost total loss to birds,
causing farmer to abandon test.




- 9 -


Table 5. Analysis of Variance for grain yield

SOV df

Site 9 **
Treatment 3 **

C.V. = 55.7 %


The same factors that led to abandonment of some tests
resulted in high coefficients of variation. In addition, both
farmers and extension agents found the staggered operations of 3
different dates of seeding difficult to manage. Only the 15 July
date of seeding for IKMV was not significantly different in terms
of grain yield from the local variety, although mean yields in
this test were more than 200 kg/ha less (Figure 3). Both the
early and late dates of seeding of IKMV gave significantly lower
yields than the loral, but yields were not significantly
different from Lhuse of Lhe 15 July seedigy.

In general, local varieties appear more robust and tolerant
of the environmental extremes seen this year. Although the 15
July date was the best of the three dates of seeding this year,
it is uncertain if the advantages of short cycle and drought
tolerance of IKMV would allow IKMV planted at this date to
outyield local varieties. Differences between performance of
IKMV under the higher fertility of the research station and
performance in two years of on-farm tests, plus evidence in the
North Province that this variety responds to fertilization with
brewery residues, suggest that higher fertility may also be
required for IKMV to realize its maximum yield potential.

Conclusions

Based on the results of 1986 and 1988 on-farm tests, IKMV
was not recommended to the development agency, SODECOTON. It was
suigq.sted by TIT,I during the annual TRA/SODECOTON planning meeting
in Maroua in January 1989 that future efforts in development of
improved pearl millet varieties should include multi-locational
trials of varieties in the pearl millet areas of the extreme
North. The sorghum and pearl millet breeder at IRA Maroua has
incorporated this step in his 1989 work plan. More detailed
descriptive study of variability among local varieties and
adaptation to different zones is needed; again, the sorghum/pearl
millet breeder is planning on incorporating local varieties in
1989 multi-locational trials. Finally, on-farm testing should
begin with researcher-managed trials, followed by farmer-managed
ones.










- 10 -




- 11 -


1988 PEARL MILLET TEST

RAINFALL (MM)


MA1MA2MA3 JU1 JU2 JU3 JL1 JL2 JL3 AU1 AU2 AU3 SE1 SE2 SE3 001002003


DECADE


FIG 4, AVERAGE OF 10 ON-FARM SITES


_ ___ __ _~





- 12 -


1988 ON-FARM TEST OF SORGHUM SEED TREATMENTS

Introduction

One of the major oonstrainto to improved production of
sorghum in northern Cameroon is poor establishment of plant
stands, usually a function of the erratic and unpredictable
nature of early-season rains. Research on-station has indicated
that the seed treatment Marshal can increase sorghum yields by
improving initial plant stands. Marshal is a systemic pesticide,
which lasts in the plant as long as one month after seeding, and
is thought to confer protection against a range of soil-borne and
seedling insect pests. The objective of this test was to compare
Marshal as a seed treatment to treatment with the currently
recommended product, Thioral, and to no treatment, both with
local varieties and with the currently recommended improved
short-cycle variety S35.

Materials and Methods

The test was conducted in collaboration with SODECOTON, as
described in the Materials and Methods section of the 1988 Peanut
Fertilization Test above. It was planned for 23 on-farm sites in
22 SODECOTON sectors of the Centre Nord. The design was a split
plot, with each site representing a single block. Main plot
treatments were varieties: 1) Farmer-selected local variety; 2)
S35. Sub-plot treatments were seed treatment: 1) Thioral; 2)
Marshal; 3) none.

Each field measured 2500 m2 (50 m x 50 m); each plot
consisted of 10 rows, 0.8 m apart, and measured 8 m by 50 m. The
four central rows of each plot were harvested. Farmers were
asked to choose a field that had been in cotton the previous
season. The density was 62,500 plants/ha (two plants per pocket,
0.40m between pockets). Seed was treated with Thioral and
Marshal by the monitors just before seeding.

Urea was applied, banded, at the rate of 50 kg urea/ha, 30
days after seeding. Missing pockets were reseeded 7 days after
seeding, and two weedings were done, at 15 and 30 days after
seeding. Field observations included number of pockets
reseeded, number of plants 15 and 45 days after seeding, number
of plants and panicles harvested, number of striga plants between
the 2nd and 3rd central rows, and harvest weight of panicles.
Samples from each plot were taken to Maroua and threshed;
threshing percentage, grain yield and loss to birds were
estimated.

Results and Discussion

Successful results were obtained from 21 of the 23 sites
originally planned. Although there existed considerable
variability within several fields, farmers expressed an ability
to distinguish an effect of seed treatment on initial plant




- 13 -


stands. A prolonged dry spell in June led to widespread attacks
of the caterpillar Spodoptera exempta; again, field observations
indicated that attacks were less severe in treated plots.



Table 6. GRAIN YIELD

.nv,, Mean Grain Yield (kc/ha) = 1498


SOV
Site
Variety
Site*Var
Trt
Var*Trt


df
20 **
1 ns
20 **
2 ns
2 ns


C.V. trts = 16.4 %


Table 7. DENSITY 15 DAYS AFTER SEEDING (plants/ha)

Anova Means of Varieties


SOV
Site
Variety
Vill*Var
Trt
Var*Trt


S35
Local


df
20 **
1 *
20 ns
2 **
2 ns


C.V. trts = 17.5 %


44377 a
41860 b


LSD (0.05) = 2669

Means of Treatments


Marshal
Thioral
None


46263 a
43489 a
39602 b


LSD (0.05) = 3269



Table 8. NUMBER OF POCKETS RESEEDED (no./ha)


Means of Treatments


None
Thioral
Marshall


SOy df
Site 20 **
Variety 1 ns
Site x Var 20 ns
Treatment 2 **
Var*Trt 2 ns

C.V. trts = 42 %


10697 c
8539 b
6493 a


LSD (0.05)-= 1564


-*_~-r-" --,-*--* ir-r.~----' -- -" --~---*---..,


J4W a





- 14 -


Table 9. GRAIN MOLD (Rated 0-3)


Anova


SOV
Site
Variety
Site*Var
Trt
Var*Trt


Means of Varieties


S35
Local


df
20 *2
1 **
20 **
2 *
2 ns


1.68 a
0.41 b


LSD (0.05) = 0.18

Means of Treatments


C.V. trts = 47 %


None
Thioral
Marshal


1.19 b
1.02 ba
0.92 a


LSD (0.05) = 0.21



Table 10. LOSS TO BIRDS (ka/ha) From 18 sites

Anova Means of Varieties


Site
Variety
Site*Var
Trt
Var*Trt


S35
Local


17 **
1 **
17 **
2 ns
2 ns


66.1 a
20.8 b


LSD (0.05) = 15.1


C.V. trts = 88 %


In terms of final grain yield, there were no significant
differences either between varieties or among treatments; nor was
there any interaction of these factors (Table 6). This lack of
effect on yield can probably be explained by the fact that
farmers and extension monitors were asked in the test protocol to
reseed missing pockets seven days after seeding; some farmers
also transplanted to achieve good stands. These operations no
doubt mitigated the effect of differential initial plant stands
on final yield.

That seed treatment affected plant stands is clear. Marshal
and Thioral both gave significantly better stands 15 days after
seeding than no treatment, though there was no difference between
the two. S35 had significantly higher stands 15 d.a.s. than the
local varieties (Table 7). Marshal gave significantly higher
plant stands at 45 d.a.s than Thioral or no treatment (Figure 5).
Marshal was also significantly better in terms of numbers of
pockets that had to be reseeded than Thioral, which was superior
to no treatment (Table 8 and Figure 6).
As would be expected, and as seen in past years, S35
suffered significantly higher losses to birds than the local


---''~--- -- ~~;---~---


`~'"''"






- 15 -


varieties, indicating yet again that when this variety cannot be
seeded late it must be protected against birds
(Table 10). The data seems also to indicate that seed treatment
had an effect on grain mold (Table 9, Figure 7), but considering
the high C.V.s and the lack of any clear biological rationale for
this response, the indication should merely be noted with
reservations.


Conclusions

Marshal seems superior to Thioral in terms of assuring good
initial plant stands; both products are more effective than no
treatment. In many parts of northern Cameroon farmers do not
reseed unless plant stands are practically decimated; to estimate
the effectiveness of these treatments in increasing yields from
improved plant stands, they should be tested either with no
reseeding (except where initial seeding fails completely) or with
farmers themselves deciding whether and to what extent to reseed.
Since Marshal is a more expensive product, and farmers in the
Extreme North tend to resist increased input costs in sorghum,
reduced rates of Marshal should be tested.





- 16 -


SORGHUM SEED TREATMENT

PLANT STAND AT 45 DAYS


DENSITY (PLANTS/HA) (Thousands)


60


50 ...


40 -..


30 -..


20 --


10 .............


0


b
... ......-.. .- ...........


NONE


THIORAL


MARSHAL


SEED TREATMENT

FIG. 5 FROM 21 ON-FARM TRIALS


........,.,...,
.............. ...... ... .........> :X '. .. .
..........-......
............
.. ...........
.'.........
....X.......
i li~iiii~i...........
..........


....... ..........................




- 17 -


SORGHUM SEED TREATMENT
No. OF POCKETS RESEEDED
POCKETS RESEEDED (Thousands)


NONE THIORAL MARSHAL
SEED TREATMENT
FIG. 6 FROM 21 ON-FARM SITES





- 18 -


SORGHUM SEED TREATMENT
INCIDENCE OF GRAIN MOLD
GRAIN MOLD SCORE (0-0)
1.4
b
1.2 .. ...


0 8 ::
0 ....... ...... ..........
o ,4 .. ......... ............................................ -........................... ............. .
04 .....2 ......... ....... ......... .......... ...... ..

0.0
NONE THIORAL MARSHAL
SEED TREATMENT
FIG. 7 FROM 21 ON-FARM TESTS 1988


1988 SORGHUM SEED TREATMENT TEST


MAIMA2MA JUIJU2 JUL lJL ..2 dJL AIJ8E18E28E30O10C20C3
DECAL-
FIG 8. AVERAGE ACROSS 21 ON-FARM SITES




- 19 -


1988 ON-FARM TEST OF SORGHUM VARIETIES

Introduction

Sorghum varieties have been tested on-farm in northern
Cameroon for a number of years. The tests have usually been
designed with a farmer-selected local variety as check and with
one replication per site. The use of single.replications limited
the extent to which variety-by-site interactions could be
estimated; although the TLU extension agronomist suspected that
different sorghum genotypes would differ in their adaptability to
the various agro-ecological zones of the far North, this
supposition could not be tested.

The 1988 Sorghum Variety test was designed to evaluate the
performance of a wide range of sorghum genotypes across a large
numlbr' of Nitos to identify thy vwriuty -by-location interactions
that existed, with the view in mind to establish "recommendation
domains" in the far North.

Material and Methods

The test was conducted in collaboration with SODECOTON, as
described in the Materials and Methods section of the 1988 Peanut
Fertilization Test above. It was planned for 67 on-farm sites in
22 SODECOTON sectors of the Centre-Nord. The design was a
randomized complete block (RCBD), with three blocks per site.
Four varieties were initially seeded at each site: 1) CS54
(improved, open-pollinated); 2) Walaganari (local); 3) Djigari
(local); 4) Gueling (local).

On the 28th of June, seed for a fifth genotype, the hybrid
CSH5, arrived in Maroua from India. Since this genotype would
provide a wider overall range of genotypes than a group including
three local varieties, the predecessor of the current extension
agronomist took the decision to seed CSH5 in place of Walaganari.
Although some sites had not been seeded or had lost almost the
whole stand to the dry spell in June, this decision meant that
some farmers were asked--not required--to remove Walaganari that
had already emerged and replace it with CSH5. The decision
displeased some farmers and development agency personnel, so
farmers who had removed well-established Walaganari were given an
indemnization to offset possible loss.

Each field measured 2500 m2 (50 m x 50 m); each plot
consisted of 5 rows, 0.8 m apart, and measured 4 m by 50 m. The
three central rows of each plot were harvested. Farmers were
asked to choose a field that had been in cotton the previous
season. The density was 62,500 plants/ha (two plants per pocket,
0.40m between pockets). Seed was treated with Thioral.

Urea was applied, banded, at the rate of 50 kg urea/ha, 30
days after seeding. Missing pockets were reseeded 7 days after
seeding, and two weeding were done, at 15 and 30 days after





- 20


seeding. Field observations included number of pockets reseeded,
number of plants 15 and 45 days after seeding, number of plants
and panicles harvested, number of striga plants between the 1st
and 2nd central rows, and harvest weight'of panicles. Samples
from each plot were taken to Maroua and threshed; threshing
percentage, grain yield and loss to birds were estimated.

Results and Discussion

Successful results were obtained from 60 tests. One of the
most salient observations made during the growing season was that
CSH5 did poorly in almost all locations, chiefly due to failure
to set seed. Whether this was due to insect attack or other
causes is uncertain. Early development of CSH5 before flowering
was also poor, perhaps due to effects of low on-farm fertility.



Table 11. Combined ANOVA for 3 Varieties Common to 60 Sites


Gimay i Vi d1


M %&. & a M A rA & a iab.4.


Site 5!
Rep/Site 12
Variety
Site x Var 11

C.V. Varieties


9
0
2
8


**


19 %


82 %


Table 12. Means for Grain Mold (0-3) for 3 Varieties. 60 Sites

CS54 1.18 a
Gueling 0.25 b
Djigari 0.20 b

LSD (0.05) = 0.09





Table 13. Means for Grain Yield (kg/ha), 3 Varieties, by Region


KAnlA Mnay flanav


1179 a
1042 b
830 c


1220 a
1087 b
1048 b


Manv TLoti


1513 a
1340 b
1220 b


Morn Mokolo


2010 a
1924 a
1686 b


LSD (0.05)= 120


niamaren


Gueling
Djigari
CS54


1626 a
1431 b
950 c


--- e ----- -- -K,- ---v -T ---- -------


cr\V


Gr in M l <4


77 90


172 153




- 21 -


Analysis of Variance over all 60 sites, for the 3 varieties
common to those sites, showed--not surprisingly--significant
effects of variety and variety-by-site interaction for both grain
mold and grain yield (Table 11). Across-all sites, grain mold
for CS54 was greater than for the local varieties.

A combined Analysis of variance, grouping sites by region
and treating region as a source of variation, was not possible
since number of sites per region was quite unequal, and the TLU
computer had insufficient memory to run a General Linear Model
ANOVA on such a large unbalanced data set. Separate ANOVAs by
region indicated that variety-by-site interaction was still
significant within each of the regions. Nevertheless, in an
attempt to see if an important variety-by-region interaction
existed, comparison of mean separations of grain yield by region
(Table 13 and Figure 9) suggested that even if such an
interaction is statistically significant, the comparative
rankings of varieties indicate that this interaction might not be
useful in determining either recommendation domains or research
domains for sorghum.

This conclusion is supported by Modified Stability Analysis
of the three varieties across sites, which shows no cross-over of
the responses lines of yield of each variety regressed on the
environmental index for each site (Figure 10).

Despite the low yields of CS54, and its susceptibility to
grain mold and loss to birds, it was highly appreciated by many
farmers (Figure 11), due to its earliness, quality of grain and
flour, and a perception by some farmers that it was tolerant of
striga. A remarkable result of this test was the consistently
high yields of the local variety Gueling across environments
(Tables 11 and 13 and Figure 10) and the fact it was also highly
appreciated by farmers (Figure 11). This variety, recommended to
the TLU team by farmers in Kaele and grown in the eastern parts
of the Extreme North, was thought by some farmers not to be
"afraid of striga." Not much is known about Gueling, since it
was not described in early reports of collections of local
sorghum varieties, but based on this year's results, it deserves
further study.

Conclusions

Based on 1988 results, "recommendation domains" for sorghum,
i.e. contiguous geographical zones of differential adaptation of
groups of genotypes, cannot yet be identified on the basis of
this year's results. Reexamination of past data by a variety of
analytical techniques may identify such domains across years;
efforts should focus on the possibility of such agronomic
recommendation domains based on agro-ecological zones,.but final
recommendation domains for varieties will probably need to take
into account socio-economic considerations as well as agronomic
ones.





- 22 -


I

SORGHUM YIELD
1988 SORGHUM VARIETY TEST

GRAIN YIELD (KG/HA)
2600






1000 .-....... ..........

50 0 -. ....iiiiii ......


DIAMARE KAELE MAYO DANAY MAYO LOUTI M. MOKOLO
SODECOTON REGIONS


M 0864


DJIGARI Eli GUELING


FIG. 9 FROM 60 ON-FARM SITES





- 23 -


1988 SORGHUM VARIETY TEST

MODIFIED STABILITY ANALYSIS

OiAiN iEELD (TONS/HA)

YLD GUELING= 106.5 + 1.04 El
YLD DJIGARI -15.3 + 1.02 El
YLD CS54 -91.3 + 0.92 El









0 500 1000 1500 2000 2500 3000 3500
ENVIRONMENTAL INDEX
O CS54 --r- DJIGARI -- GUELING


FIG. 10 FROM 60 ON-FARM SITES


I--- -
I---------------- ----- I





- 24 -


SORGHUM PREFERENCE
"BEST" VARIETY


CS-54 17


GIlE L NG 14 \\\\\\\\ \\
\\\\\\\\y 7

DJICARI
1988 ON-FARM VARIETY


l"SH-5 1
WAL AGANAHI 3


10
TEST


I i( 11 FIOM 4h INA1MF-I I rIpnN2:!-


SORGHUM PREFERENCE
"WORST" VARIETY

FUELING
4
DJIGARI CS54
7 8


VMLA(3ANARI


CSH-5
17
1988 ON-FARM VARIETY TEST

FIG. 12 FROM 41 FARMER RESPONSES


II------------~ -~-~- -~---*--




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs