Agriculture Reference
In-Depth Information
College of Agriculture and are used here with permission from Dr. Kanyama-Phiri. D2
data were collected by us and local assistants employed by Bunda College of Agriculture
for the complete duration of the project.
9.5.4 Maize yields
Mature maize was harvested from subplots at the center of each plot. For D1, grain yields
were taken when plants had dried in the field, and 12% moisture content was assumed for
the grain (Kamanga et al., 1999). For D2, we harvested maize slightly earlier as per farmers'
requests and therefore calculated the actual dry weight based on a subsample of the fresh
material
.
Although the study began in 1995, we only present phase 1 data from the 1999
and 2000 maize harvests and phase 2 data from the 2003 and 2004 maize harvests since
these were years with the most complete data sets and greatest number of replicates avail-
able. Also, 2001-2002 was a transition period between the two designs. Maize yield data, as
well as other forms of data, were not available for all farms in any given year for a variety
of reasons. For instance, many farmers harvested maize early due to concerns of crop theft
from their fields, and sometimes fields had been set ablaze by youth trying to capture mice
as a protein source. Occasionally, some farmers applied small quantities of fertilizer in
unknown quantities, precluding their inclusion in the analysis. These kinds of problems
are to be expected in on-farm participatory work with resource-limited farmers, and we
acknowledge that this poses challenges for data analysis. The resulting unbalanced data
structure was considered throughout our statistical analyses in both our choice of meth-
ods and in evaluating the assumptions for each test. Data are presented for 29 farmers in
1999, 28 in 2000, 21 in 2003, and 14 in 2004. For each design, maize yields were evaluated
using two-way analyses of variance (ANOVAs) (*
P
< 0.05), with cropping system and year
as the main factors and maize yield as the response variable. For both D1 and D2, Tukey's
post hoc tests (*
P
< 0.05) were employed to compare treatments.
9.5.5 Soil analysis
Soil was sampled from each treatment with the exception of D1 legume treatments receiv-
ing full fertilizer (the latter were not collected due to a variety of logistical constraints). A
composite soil sample of six subsamples was collected from 0 to 10 cm depth in 1994 (prior
to the establishment of treatments), 2000, and 2004. Soils were air dried, passed through a
2-mm sieve, and analyzed for pH, extractable Bray-P (Diamond, 1995), and percentage total
C and percentage total N using a Carlo Erba 1108 elemental analyzer (Smith and Tabatabai,
2003). Soil data were analyzed from a subset of 12 farmers, balanced across the three land-
scapes. The same set of farmers was used for 1994, 2000, and 2004.
9.5.6 Maize foliar analysis
Foliar tissue testing was performed for D2 years 2003 and 2004. Samples were collected using
a composite of 12 ear-leaf blades located immediately above the primary ear node (Jones and
Eck, 1973) when maize was at peak tassel. They were then analyzed for percentage C, N, and
S (Smith and Tabatabai, 2003) using high-temperature dry combustion. Foliar percentage P
was determined using a nitric acid/hydrogen peroxide microwave digestion and analysis
by inductively coupled plasma atomic emission spectrometry (Meyer and Keliher, 1992).
Maize foliar data were collected from 21 farmers in 2003 and 12 farmers in 2004.
Search WWH ::
Custom Search