?7- /3 i --l-! :i '
Potential Errors in Determining Grain
Drying Rate of Corn With Electronic Moisture Meters'
F. M. Rhoads2
Determination of grain drying rate a few weeks before harvest allows
prediction and scheduling of a harvest date which will minimize weather and
insect damage to grain while in the field. A linear increase of dry-matter
content in corn ears during most of the grain filling period has been reported
(Friedrich et al., 1979). Two moisture measurements taken 5 or 10 days apart
before physiological maturity provide sufficient information for determining
grain drying rate in the linear portion of the grain drying curve of corn.
Electronic grain moisture testers require about two ears of corn for each
sample and are convenient for determining the moisture content of grain.
However, the moisture range in which these instruments are calibrated is below
about 40%. The normal range for harvesting high moisture grain is 25 to 30%
moisture. Therefore, lead time for harvest scheduling is short with a system
that cannot measure grain moisture until it reaches 40%. Moisture can be
determined at any stage of grain filling with the proper drying and weighing
facilities but this is a time consuming process. Since the Burrows model 700
digital moisture computer has a digital readout range of 0 to 99.9%, it might
be assumed that this instrument is suitable for measuring grain moisture at
any level. Unless the operator's manual is referred to, the above assumption
is a potential source of error in the determination of grain moisture with
electronic procedures. Another potential source of error is using a sample of
grain smaller than the amount required for accurate measurement of grain
INFREC, Quincy Research Report NF 87-13.
2Professor of Soil Science, Univ. of Florida, NFREC, Quincy, FL 32351.
moisture. Some electronic grain moisture meters have a built in weighing
system which may be ignored if the grain sample is too small.
The objectives of this research were to (1) compare electronic and gravi-
metric methods for determining grain drying rate of corn, (2) determine magni-
tude of errors with improper use of electronic grain moisture meters, and (3)
compare the drying rate of a full season corn hybrid with that of an early
season corn hybrid.
MATERIALS AND METHODS
Two corn hybrids (Sunbelt brand '5613' and '1882') were planted on 7
April, 1987 in four separate plots (10' x 20') for each hybrid. Sunbelt brand
'5613' is an early season hybrid and Sunbelt brand '1882' is a full season
hybrid. Recommended cultural practices were applied uniformly to all plots.
Silking date was determined as the date when 50% of plants of each hybrid
silked. Silking dates were 5 and 9 June, respectively, for the early and full
Two ears of grain from each plot were randomly harvested at weekly
intervals starting 22 June and continuing to the last sampling date on 3
August. Samples were immediately transported to the laboratory, hand shelled,
and the grain placed in zip lock plastic bags. Moisture readings were
determined separately for 250 g and 100 g subsamples of grain on each field
sample with a Burrows model 700 digital moisture computer and a 100 g
subsample of grain was immediately weighed for drying at 700C in a heated
cabinet. Samples were dried to constant weight and the dry weight recorded.
Moisture % was calculated as % moisture = Field weight-dry weight/Field weight
Analysis of variance procedures were performed on the data for each
sampling date and a paired comparison test was used to determine if difference
between electronic and gravimetric procedures were significant. Regression
analyses for electronic versus gravimetric procedures were performed
separately on data for grain moisture above 40% and for grain moisture below
40%. Regression analyses for gravimetric % moisture versus time were perform-
ed separately on data for individual hybrids. All statistical procedures are
described in Steel and Torrie (1960).
RESULTS AND DISCUSSION
Values from the electronic determination of grain moisture were signifi-
cantly (P<0.05 and 0.01) different from those found with gravimetric proced-
ures in the moisture range above 40% (Table 1.). However, there was no dif-
ference (P>0.1) between the two procedures in the moisture range of 24 to 40%.
This is in agreement with specifications outlined in the operator's manual for
the digital electronic moisture computer (Burrows Equip. Co., 1977). Highest
values (>97%) were obtained electronically in the gravimetric moisture range
of 45 to 55%.
Table 1. Grain moisture in two corn hybrids 13 to 59 days after silking as
measured with a Burrows model 700 digital electronic moisture
computer and by gravimetric (weighing and drying) procedures.
Fall Season Hybrid Early Season Hybrid
(Sunbelt '1882') (Sunbelt '5613') Probability
Days after Elec- Gravi- Days after Elec- Gravi- of larger
silking tronic metric silking tronic metric t
-% moisture- -% moisture-
13 87.0 86.5 17 86.6 78.8 <0.05
20 88.3 72.4 24 87.6 68.8 <0.01
27 89.2 59.9 31 97.2 54.8 <0.01
34 97.2 45.9 38 64.4 39.2 <0.01
41 54.8 40.5 45 37.3 38.3 >0.20
48 34.7 34.7 52 29.7 30.1 >0.30
55 30.2 31.7 59 24.6 24.8 >0.10
Student's t was used to test for differences between electronic and
gravimetric methods of measuring grain moisture (Steel and Torrie, 1960).
Data from individual hybrids were pooled together for statistical
Regression analysis showed that the regression coefficient (b), or the
slope of the regression equation for electronically measured % moisture versus
gravimetric % moisture, was significantly (P<0.01) less than one in the
electronically measured % moisture range of 64 to 97% (Table 2.). Further-
more, there was not a significant (F<1.0, r=0.176) correlation between the two
methods of measuring grain moisture during the first 38 days following
silking. However, b was not significantly (P>0.1) different from one and
electronically measured % moisture was correlated (F=263.80**, r=0.967) with
gravimetric % moisture between 45 and 59 days after silking (24 to 38 elec-
tronically measured % moisture).
Table 2. Regression analysis for electronically measured % moisture (EMPCM)
versus gravimetric % moisture (GPCM) showing students t for b (slope
of regression line) =1, F test for significance of regression, and
correlation coefficients (r).
Days after Regression t F
silking equations value value r
13 to 381 EMPCM=78.2+0.140(GPCM) 6.056** 0.97 0.176
45 to 592 EMPCM=1.85+0.919(GPCM) 1.434 263.80** 0.967
1Range for meter per cent moisture = 64.4 to 97.2.
Range for meter per cent moisture = 24.6 to 37.3.
**b significantly different from 1 and regression is significant (P<0.01).
Drying rates were calculated for each hybrid from regression analyses of
gravimetric % moisture versus days after silking (d) (Table 3.). The average
drying rate for both hybrids between 13 and 59 days after silking was about
1.3% per day. However, the drying rate between 13 and 59 days after silking
for each hybrid was best described with a quadratic equation having a signifi-
cant (P<0.01) quadratic component (Fig. 1). Grain moisture was predicted from
the quadratic equations to be 31.9% 55 d after silking for the full season
hybrid and 25.5% 59 d after silking for the early season hybrid. The quad-
ratic equations for the two hybrids were nearly parallel until about 40 days
after silking when they began to diverge. Drying rate between 40 and 60 days
after silking appeared greater for the early season hybrid than for the full
season hybrid. Regression analyses between 13 and 40 days after silking
revealed a significant (P<0.01) linear regression without a significant
quadratic component. The relationship between 40 and 60 days was also linear
with no quadratic component. Therefore, the quadratic equation can be
effectively divided into two linear segments, one from silking to 40 days and
the other from 40 to 60 days after silking. Students t revealed that both
hybrids had the same drying rate (1.9% per day) between 13 and 40 days after
silking but the drying rate for the early season hybrid between 40 and 60 days
after silking was about 1.5 times (P<0.01) that of the full season hybrid.
Table 3. Regression analyses for gravimetric per cent moisture (GPCM) in corn
grain versus days after silking (d) for a full season hybrid (FSH)
and an early season hybrid (ESH), showing student's t for comparing
linear regression coefficients (b) between hybrids, F tests for
significance of regression, and coefficients of determination (r2).
Days after Regression t F
silking Hybrid equations value value r
Total Seed Growth Period
13 to 55 FSH GPCM=98.1-1.32d n.d. 365.29** 0.934
17 to 59 ESH GPCM=97.4-1.30d n.d. 390.81** 0.938
13 to 55 FSH GPCM=123-3.08d+0.0259d2 n.d. 94.89**. 0.986
17 to 59 ESH GPCM=122-2.81d+0.0199d2 n.d. 26.33** 0.970
Early Seed Growth---
13 to 34 FSH GPCM=111-1.92d 1.23 723.34** 0.981
17 to 38 ESH GPCM=113-1.90d 1.23 322.98** 0.959
Late Seed Growth
41 to 55 FSH GPCM=65.7-0.63d 16.44** 32.22** 0.763
45 to 59 ESH GPCM=81.2-0.96d 16.44** 111.81** 0.918
n.d. = not determined. ** (F) regression significant with P<0.01.
** (t) b (FSH) < b (ESH) with P<0.01.
1F test for quadratic component only.
5 A % = 123-3.08d+0.0259d2 B
B % = 122-2.81d+0.0199d2
10 20 .30 40 50 60
Days after Silking
Figure 1. Moisture content of corn grain versus time
(days after silking), showing mean values from
four samples on each date for two hybrids and re-
gression curves for individual hybrids. A = full
season hybrid and B = early season hybrid.
Use of a 100 g sample of grain did not significantly (P>0.10) influence
moisture readings above 40%, but it reduced readings by about half for
moisture levels below 40% (Fig. 2). Student's t was used to compare effect of
sample size. None of the meter readings agreed with the gravimetric procedure
at grain moisture levels above 40%. Therefore, it is important to use the
correct sample size and to only consider moisture readings below 40% when
using an electronic grain moisture computer for determining grain moisture.
A a B B
s 90- A A
6 60- A
S30. A 250 gms grain
Bg B 100 gms grain
B B B B
I I I I I I
24 36 48 60 72 84
Gravimetric % moisture
Figure 2. Effect on readout of reducing the amount of
grain placed in an electronic moisture meter that
is calibrated for 250 gms of grain per sample.
Solid line represents identical values for meter
and gravimetric procedures.
Results of this research indicate that drying rate of grain cannot be
determined electronically with a Burrows moisture computer when moisture
content of grain is above 40%. However, satisfactory results with the
electronic meter were obtained in the 20 to 40% moisture range. The
operator's manual specifies the range of accurate moisture determination to be
5 to 40%. Therefore, it should not be assumed that levels of grain moisture
higher than 40% can be determined electronically just because the readout
range is 0 to 99.9%.
Burrows Equip Co. 1977. Burrows model 700 digital moisture computer. AP2M.
Friedrick, J.W., L.E. Schrader, and E.V. Nordheim. 1979. N deprivation in
maize during grain-filling. I. Accumulation of dry matter, nitrate-N,
and sulfate-S. Agron. J. 71:461-465.
Steel, R.G.D., and J.H. Torrie. 1960. Priciples and procedures of
statistics. McGraw-Hill. New York.