L16 |_0-1096
ILLINOIS UBRAff
- URBANA-CHAMPA'C
JUL 2 q 1997
^(CULTURE LIBRARY
CIRCULATING
MJjKIUUL I URt.
i m - ' *. *"" J* * ' "-"^ Vfl\l
Changes in Corn
During Export from New Orleans to Japan
Lowell D. Hill, Marvin R. Paulsen, Barry J. Jacobsen, Ri
Bulletin 788 A
COMPANION TO THIS VOLUME
This research report is available in a twenty-
page Summary Version illustrated with color
photographs. The summary version focuses on
courses of action to improve the quality of corn
that importers receive. Order Bulletin No. 788B,
Changes in Corn Quality During Export from New
Orleans to Japan: Summary Version.
Changes in Corn Quality
^^ r^nTirtrr T7vT"*r\t*f riv^-m M/aTAr f^Yrlaano 4-y\ T^r-*^*^
During Export from New Orleans to Japan
Lowell D. Hill, Marvin R. Paulsen,
Barry J. Jacobsen, Richard A. Weinzierl
Bulletin 788A *- College of Agriculture
University of Illinois at Urbana-Champaign
Agricultural Experiment Station >- December 1990
The University of Illinois provides equal opportunities in programs
and employment.
Acknowledgments
The research reported in this publication
represents the joint efforts and cooperation of
many individuals and agencies. The Japan
Feed Trade Association (JAFTA) and several of
its members helped us obtain access to the
operating facilities in the United States and
Japan. It was agreed that the names of the
cooperating firms would be kept confidential,
but they provided excellent cooperation,
assistance, and hospitality on both sides of the
ocean. JAFTA and the Japan Grain Inspection
Agency provided direct help with their person-
nel and services. The cooperation and assis-
tance of the USDA's Federal Grain Inspection
Service, Lutcher Office, was essential in obtain-
ing samples and making on-site analysis, and
we offer a special thanks to Conrad Herndon
and Robert Alice. The stevedores, who were in
charge of loading and unloading the vessel
during both shipments, made every possible
effort to accommodate our sampling plans.
Their cooperation made the task possible; their
accommodating help made our task easier.
The authors express a special thanks to the
research teams, who worked around the clock
in New Orleans with little food and no rest for
nearly 30 hours, sampling the barges and
vessels during loading. This team consisted of
Lowell Hill, Marvin Paulsen, Barry Jacobsen,
Gene Shove, Richard Weinzierl, Donald
Uchtmann, Larry Pruitt, Terrence Kuhn, Martin
Patterson, Alan Tumblin, Wojciech Florkowski,
Brian Cratz, Todd Ringhouse, Robert Spangler,
Grear Kimmel, William Creswell III, Terry
Taylor, and Dean Gordin.
The appearance and accuracy of charts and
tables were immensely improved through the
efforts of Daniel C. Marriott. Terrence Kuhn
and Robert Spangler contributed many hours of
work analyzing samples and organizing data.
Joe Vercimak assisted in statistical analysis.
Thanks to Dee Lambert for her patience in
typing the innumerable revisions and drafts
and her expertise in format and organization.
Funds for this research were provided by the
Illinois Corn Marketing Board, the National
Corn Growers Foundation, and the Illinois
Agricultural Experiment Station. This research
is a contribution to regional project NC-151,
"Marketing and Delivery of Quality Cereals
and Oil Seeds in Domestic and Foreign Mar-
kets," and Illinois Agricultural Experiment
Station Hatch Project 05-371.
Contents
Introduction 1
Organization of the Study 2
Sampling Procedures 3
Sampling at origin
Sampling at destination
Sample Analyses 8
Moisture content
Particle size and percent of whole kernels
Broken corn and foreign material (BCFM)
Breakage test and stress cracks
Density
Determination of insect infestation levels (1986 only)
Mold identification (1986 only)
Results 10
Changes in grade factors
Changes in nongrade factors
Mycological analysis of samples
Segregation and variability
Temperature changes during transit
Summary of Results 33
Conclusion 35
Tables 1-20 37
Appendix A:
Lower Limits of Moisture (%) Needed for Growth of Major Fungi in Cereals 57
Appendix B:
Methodology for Individual Kernel Moisture Determination for Corn 58
Notes... ,...60
Figures
Tables
Figure 1. Two-probe sampling patterns for each layer, New
Orleans to Japan, 1985 and 1986.
Figure 2. Single-probe sampling patterns for each layer,
New Orleans to Japan, 1985 and 1986.
Figure 3. Placement of temperature sensors in hold 5 of the
Century Progress in New Orleans.
Figure 4. Sampling patterns for barges and coasters in
Japan. (Fewer samples were taken in barges than in
coasters.)
Figure 5. Comparison of percentage BCFM in the peri-
meter and inside sectors (average of layers 1-4, hold 1,
New Orleans, 1986).
Figure 6. Frequency distribution of individual kernel
moisture in origin samples, New Orleans, 1985.
Figure 7. Frequency distribution of individual kernel
moisture in vessel samples, Japan, 1985.
Figure 8. Frequency distribution of individual kernel
moisture in incoming barges, New Orleans, 1986.
Figure 9. Frequency distribution of individual kernel
moisture in sublets, New Orleans, 1986.
Figure 10. Frequency distribution of individual kernel
moisture in Japan coaster, 1986.
Figure 11. Temperature changes within the vessel during
transport from New Orleans to Japan, 1985.
Figure 12. Temperatures (' C) averaged for 24-hour
periods from hold 5 located 20 feet forward of aft bulkhead
wall, 1985.
Figure 13. Average daily temperatures (* C) measured
near the surface during shipment from New Orleans to
Japan, 1986.
Figure 14. Average daily temperatures (" C) near the floor
during shipment from New Orleans to Japan, 1986.
Figure 15. Temperature profiles in the vessel at selected
depths, Japan, 1986.
Figure 16. Location of hot spots in hold 3, Japan, 1986.
Figure 17. Frequency distribution of individual kernel
moisture in vessel hot spot, Japan, 1986.
Table 1. Number of samples by hold and layer in the New
Orleans to Japan shipment, 1985 and 1986.
Table 2. Changes in grade factors in the market channel
from New Orleans to Japan, 1985 and 1986.
Table 3. Average oven moisture at various locations in the
market channel from New Orleans to Japan, 1985 and 1986
(%).
Table 4. Particle-size distribution and percentage of whole
kernels at selected points in the market channel from New
Orleans to Japan, 1985 and 1986.
Table 5. Breakage susceptibility and stress cracks at selected
points in the market channel from New Orleans to Japan,
1985 and 1986.
Table 6. Changes in the percent of kernels infected with
selected fungi during ocean transport, 1986.
Table 7. Variability in percentage of BCFM among samples
at selected points in the market channel from New Orleans
to Japan, 1985.
Table 8. Variability in percentage of BCFM among samples
at selected points in the market channel from New Orleans
to Japan, 1986.
Table 9. Variability in percentage of moisture content
among samples at selected points in the market channel
from New Orleans to Japan, 1985 and 1986.
Table 10. Changes in moisture and BCFM among layers
during transit from New Orleans to Japan, 1986.
Table 11. Moisture variability by 1-foot increments in the
hold, Japan, 1986.
Table 12. Variability in percentage of moisture content for
single kernels at origin and destination, 1985.
Table 13. Variability in percentage of whole kernels among
samples at selected points in the market channel from New
Orleans to Japan, 1985 and 1986.
Table 14. Variability in percentage of kernels with stress
cracks at selected points in the market channel from New
Orleans to Japan, 1985.
Table 15. Variability in percentage of kernels with stress
cracks at selected points in the market channel from New
Orleans to Japan, 1985 and 1986.
Table 16. Variability of breakage susceptibility at origin and
destination, New Orleans and Japan, 1985.
Table 17. Variability of particle sizes at origin and destina-
tion, New Orleans and Japan, 1985.
Table 18. Variability of particle size at origin and destina-
tion, New Orleans and Japan, 1986.
Table 19. Corn temperatures (' C) at selected depths in
random locations in the ocean vessel at Japan, 1986.
Table 20. Temperature and quality characteristics identified
in holds 3 and 1 at various depths in the ocean vessel at
Japan, 1986.
Introduction
Changes in the quality of corn between U.S.
origins and foreign country destinations have
been a frequent explanation of complaints by
foreign buyers about low quality in their
processing plants. Since quality at the time of
loading is certified by federal inspectors, and on
the average is above the minimum quality
requested by the foreign buyer [9], the problem
has been presumed to be caused by handling
and shipping practices. Past studies have
shown that the number of broken kernels
(recorded as BCFM in U.S. corn grades) in-
creases during loading and unloading of the
vessel after the grade is certified [2, 3, 4, 5] and
that heating of the corn can occur during transit
on the ocean, especially through warm waters
[8]. This study was designed to (a) establish the
extent of quality-loss under different environ-
mental conditions, (b) identify the causes of
quality-loss, and (c) suggest possible solutions
to reduce the loss of quality during shipment.
Two shipments of corn, both bound from U.S.
ports in the New Orleans area to Japanese
ports, were selected for this study. The first
was loaded in May 1985, the second in March
1986. By coincidence, the same vessel was
nominated for both shipments. Different
exporters, buyers, and final users were involved
in each shipment; this diversity offered a useful
range of technology and loading conditions but
added to the complexity of organizing identical
tests and sampling procedures.
The identity of the corn in the 1985 shipment
was maintained from the outbound belt at
export to the barges in Japan, and samples were
taken at each point where the grain was
handled physically or transferred. In the 1986
shipment the identity was maintained from
each barge loaded at midwestern river elevators
to the barges and coasters in Japan. Origin
quality in the barges loaded at river elevators in
the 1986 shipment was based on records
provided by the river elevators. All other data
were based on samples obtained under supervi-
sion of the researchers, supplemented by
quality data reported by the Federal Grain
Inspection Service of the USDA, the Japan
Grain Inspection Association, and the export
elevator.
Organization of the
Study
A comparison of corn quality at origin and
destination, based on extensive sampling,
requires the cooperation of many firms and
individuals. The first step was to obtain a
commitment from the Japanese importers, who
selected the export elevator and nominated the
vessel. A commitment of time, resources, and
cooperation was essential. Many other people
and agencies were involved in implementing
the research plans:
> in Japan, customs officials, stevedores,
inspection agencies, the quarantine author-
ity, and a team of Japanese samplers;
>- in the United States, the Federal Grain
Inspection Service (FGIS), export brokers,
plant managers, stevedores, and the Animal
Plant Health Inspection Service (APHIS),
which provided permission to return
samples to the United States; and
> on the vessel, the owners of the vessel, the
captain, and the chief engineer.
Questions of insurance, liability, customs
clearance for equipment and personnel travel-
ing into and out of both countries, safety with
respect to the installation of equipment and the
logistics of personnel and equipment (e.g.,
before the loading, 4,000 feet of temperature
cables were cut to lengths to match the location
of each sensor) all added to the complexity of
planning and implementation. Any delays in
loading that resulted from sampling or from the
installation of equipment represented large
financial losses to the exporter. Logistics,
sequencing, and speed with safety were all
given high priority.
Before loading of the vessel, the research team,
stevedores, the plant manager, the ship's
master, and FGIS personnel met to facilitate the
coordination and sequencing of loading and
sampling. The required grain depth for sam-
pling needed to be coordinated with the
stowage plan in each hold, the balance of
weight and pressure among holds in the vessel,
the loading schedule of the elevator, and the
schedule of stevedore shifts. The success of the
sampling program required a willingness to
adapt and adjust. The selection of holds for
sampling had to be adjusted to match all the
factors above. The timing of sampling and
analysis in the elevator, on the outbound belts,
and from the inbound barges was scheduled to
keep a full sampling team available in the hold;
twenty minutes was often the maximum time
available for sampling, bagging, and tagging
sixteen two-probe samples. Accurate records
and identification of every sample had to
accompany all of this activity.
Similar coordination and planning was re-
quired when the vessel reached its destination.
Before the vessel arrived in its berth, all sam-
pling and analysis was coordinated with
stevedores, inspection agencies, and elevator
managers. When the vessel was several hours
from the harbor, the vessel masters were
contacted by ship-to-shore communications to
plan the unloading sequence of each layer and
each hold. Receivers, as well as barge owners
and operators, were contacted for permission to
take samples of their grain. More than one
destination was receiving grain from the vessel,
and the logistics of which holds were to be
sampled in which sequence without creating
imbalance on the vessel added to the complex-
ity. Pneumatic unloaders had to be removed
from the hold at each 10-foot level before the
sampling team could enter the hold. And in
1986, after all plans were finalized, a typhoon
required the vessel master to close the hatches
and move back out to sea. This disrupted the
entire sequence and generated a second round
of adapting research plans to the complex
world of moving corn from vessel to barge.
Despite this difficult and shifting planning
process and the associated negotiations, the
researchers did not compromise the important
basic principles: (a) samples had to be repre-
sentative of the sublot from which they were
taken; (b) sample identification had to be kept
inviolate and the samples kept intact; (c)
samples had to be selected randomly and
without bias with respect to quality; (d) analysis
had to follow established laboratory procedures
and statistical principles; and (e) the objective of
origin-to-destination comparisons had to be
fulfilled by matching holds and samples. The
cooperation and adaptability on the part of
everyone involved in the research made the
project exciting and rewarding, and helped
provide answers to the questions posed at the
beginning of the project.
Sampling Procedures
Researchers obtained samples at each location
in the market channel: from the inbound
barges during unloading at the export elevator
(only in 1986); from the mechanical cross-
cutting diverter on automatic timers in the
export house; from the vessel during loading,
using 10-foot compartmentalized grain probes;
and from the vessel, barges, and coasters in
Japan, using the same grain probes. The source
of each sample was recorded, and the corn was
sealed in a plastic bag and sent to the Univer-
sity of Illinois Agricultural Engineering Grain
Quality Laboratory (AEGQL) for cold storage
until tests could be completed. Some analyses
were completed on site during the sampling to
preclude the possibility of quality changes in
the corn while it was in transit from Japan to
the University of Illinois.
Sampling rates and patterns were selected to
provide the most accuracy and to minimize the
possibility of bias, within the restrictions
imposed by time, costs, and the configuration of
vessels and barges. Barges at the export
elevator were sampled during unloading using
diverter samplers, which take approximately
125 g from a cross-section of the full grain-
stream at intervals determined by a timer.
Several hundred "cuts" of the diverter were
composited into one sample per barge. Eleva-
tor personnel controlled the operation of the
sampler according to USDA procedures.
Samples from the outbound belt were obtained
from a similar automatic sampler under the
supervision of FGIS. A composite sample from
every 15,000 bushels was sent to the University
of Illinois.
The only acceptable method for sampling in the
hold of the vessel was with grain probes.
Official sample analysis requires a minimum of
1,600 g; a two-probe sample provided approxi-
mately 1,800 g of corn. Therefore, except for an
intensive single-probe sampling in one hold to
determine variability extremes, each sample
consisted of the contents of two probes. Each
hold was divided into 16 sectors, and the two-
probe samples were taken at random locations
in each sector. Because the probes were 10 feet
in length, each hold was sampled in layers after
loading approximately 10 feet of grain. Barges
and coasters in Japan were sampled with 10-
foot probes following the official pattern
developed by the USDA for obtaining represen-
tative samples from barges.
Duplicate analyses by other agencies and
laboratories were performed on some samples
to verify accuracy, repeatability, and represen-
tativeness of the samples and to provide
supplemental information. All samples were
tested at the University of Illinois AEGQL.
Selected tests on grade factors were performed
by the Federal Grain Inspection Service (FGIS)
of the USDA at its Lutcher office in Louisiana;
by the Japan Grain Inspection Association
(JGIA) in Chiba, Japan; and by the Champaign-
Danville Grain Inspection Department (CGID)
in Champaign, Illinois. In the 1986 test, analy-
ses for mold and insect infestation were per-
formed on site and at the University of Illinois
Entomology and Plant Pathology laboratories,
using a randomly selected portion of samples
collected at the origin and destination. In the
1986 test, insect-infestation levels were assessed
at loading and again at destination to determine
any increase in detectable insect numbers
following a 4-week incubation period during
transport.
Changes in the corn's quality between the
various points in the market channel were
estimated by comparing the results of these
analyses, on the assumption that the various
sampling procedures yielded equivalent results.
Previous research has indicated that analytical
results may be influenced by the sampling
method. For example, compartmentalized
grain probes generally result in higher levels of
broken corn and foreign material (BCFM) than
do di verier samplers [7]. Because both methods
are accepted by the USDA for issuing grade
certificates, however, no adjustments were
made in the data from samples taken by
diverter in the export house or from probe
samples taken in the vessel and barges. The
number of samples and the sampling proce-
dures provided composite samples considered
to be representative of the bulk grain lot from
which the samples were taken. The percentage
of broken corn and foreign material (BCFM)
was determined by the Carter-Day dockage
tester in all tests performed by the Federal
Grain Inspection Service (FGIS). The Garnet
shaker, with 12 /64-inch sieves and 30-stroke
cycles, was used to determine BCFM at the
University of Illinois AEGQL. Results from the
two testing methods differ, but no adjustments
were made in the BCFM data. Statistical tests
were used to identify significant differences in
quality. The 99 percent level of probability was
used for all comparisons, except as noted in the
text and tables.
Sampling at Origin
In 1985, the vessel Century Progress was loaded
with 56,895 metric tons (2.24 million bushels) of
U.S. No. 3 yellow corn; in 1986 it was loaded
with 55,006 mt (2.17 million bushels). Maxi-
mum moisture was 15.0 percent on both
contracts. No additional quality factors were
specified in the contracts.
In the first test the vessel was loaded May 22
and 23, 1985, in the New Orleans port area and
arrived June 28 in Chiba, Japan, via the Panama
Canal and Long Beach, California. The second
test followed a similar path but took slightly
less time in transit. In the second test the vessel
was loaded March 20 and 21, 1986, in the New
Orleans port area and arrived April 22 in
Kashima, Japan. The 1985 experiment started at
the export elevator with official samples from
the diverter sampler taken during loading of
the vessel. FGIS provided duplicate samples
from each 15,000-bushel (381 mt) component
used to construct the official 60,000-bushel
(1,524 mt) sublot sample. The FGIS loading log
provided inspection information on each
sublot. Samples were taken with grain probes
from holds 5 and 6 from each of the 16 sectors
in 10-foot layers according to the pattern shown
in Figure 1. The width and length of each hold
were divided into four equal parts and identi-
fied by letters and numbers. A sample con-
sisted of the corn from two probes taken in each
sector in each layer. Each sample was bagged
and identified by sector, layer, and time of
sampling. All five layers were sampled in hold
5; only the bottom three layers were sampled in
hold 6 (see Table 1).
Data collection from the 1986 shipment started
farther back in the marketing channel. Origin
data were provided on 41 barges that were
inspected when they were loaded at midwest-
ern river elevators. Export-elevator personnel
used diverter samplers to sample 40 of these
barges and record the grade information; the
remaining portion of each sample was split
using a Boerner divider, and approximately
2,000 g of each were sent to the University of
Illinois. AEGQL received intact samples from
39 of these barges. The grain from each barge
was placed in a separate bin as it was unloaded,
to preserve its identity so that the loads could
be blended to the correct contract specification.
FGIS obtained official samples from each
30,000-bushel (762 mt) component, using a
diverter sampler located on the belt between
the storage bins and shipping bins. Samples
from the two components were composited to
provide an official sample for the inspection log
for 60,000-bushel (1,524 mt) sublets. After the
corn was loaded into the Century Progress,
samples were taken in 10-foot layers in holds 1,
3, 5, and 7 (see Table 1). After a 10-foot layer
was loaded, the loading spouts were moved to
another hold while that layer was sampled.
After the sampling of each layer was com-
pleted, the spouts were returned to the test
holds. In holds 3, 5, and 7, two-probe samples
were taken from each of the 16 sectors in each
layer (see Figure 1). Hold 1 was sampled using
single-probe samples from each of 36 sectors
per layer to provide more detailed information
on segregation of quality within the hold (see
Figure 1. Two-probe sampling patterns for each layer, New Orleans to Japan, 1985 and 1986.
A B C D
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
y
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Figure 2. Single-probe sampling patterns for each layer, New Orleans to Japan, 1985 and 1986.
A B C D E F
2
3
4
5
6
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Figure 2). Once the hold was filled, there was
not enough space between the corn surface and
the deck to permit sampling the 20 sectors on
the perimeter of hold 1 using the 36-sector
sampling pattern. Only the center 16 sectors
were sampled in the top 10-foot layer in hold 1.
The moisture variability within the samples
was evaluated by measuring the moisture of 48
individual kernels selected at random from
each sample during the loading process. One
cut of the diverter sampler (approximately
125 g) was taken as a sample from which to
randomly select 48 kernels. Each kernel was
hermetically sealed in a glass vial. Later, the
air-oven method was used to determine the