Chapter 5 5

Seed Testing
Robert P. Karrfalt

Mr. Karrfalt is the director of the USDA Forest Service’s National Seed Laboratory, Dry Branch, Georgia.

Contents
Introduction 98 Excised Embryo Testing 110
Sampling 98 X-Radiography 111
Sample Identification 100 Other Quick Tests 112
Moisture Tests 100 Sowing Rates 112
Purity, Noxious Weed Content, and Seed Weight Tests 102 Buying and Selling Seeds 113
Purity Analysis 102 Test Limitations and Variation 114
Noxious Weed Examination 103 Scheduling Seed Tests 114
Seed Weight Determination 103 Commercial Trade of Tree, Shrub, and Native
Germination Testing 104 Plant Seeds 114
Vigor Testing 107 References 115
Chemical Staining for Viability 109

Chapter 5: Seed Testing • 97

 Introduction Then the gates should be opened gently and the probe shak-
5 en gently back and forth, so that seeds caught in the gates
Seed testing is the cornerstone of all other seed tech-
will slip down into the probe and the gates can be safely
nologies. It is the means by which we measure the viability
closed. Finally, the probe should be emptied by pouring the
and all the physical factors that regulate the use and mainte-
seeds out the top of the probe and into a second container
nance of seeds. Everything that is done with seeds should
(figure 3). This sample is the first primary sample.
have some test information to guide the work and ensure
If there is only 1 container, primary samples should be
high quality. Seed tests tell if a crop of seeds is worth col-
taken until there are 5 of them. When more than 1 container
lecting, if handling procedures are correct, and how many
holds the seedlot, at least some of the other containers must
potential seedlings are available for regeneration.
be sampled. When there are between 1 and 5 containers, all
The earliest form of seed analysis, the cut test, is still
containers should be sampled, at least 1 probe from each
often used today. Before seeds are collected in the field,
container. When there are more than 5 containers, 5 of the
some seeds are cut open with a knife or razor blade to see if
containers plus 10% of the remaining ones should be sam-
their internal tissues are fully developed and undamaged.
pled. It is never necessary to sample more than 30 contain-
This analysis is made more accurate in some cases by the
ers. (It would be rare that a forest seedlot would need 30
use of a hand lens. It is also used for simple analysis during
containers or more, or possibly even 20.) All of the primary
extraction and cleaning, or after germination to determine if
samples are then placed together to make up the composite
the ungerminated seeds have deteriorated or remained dor-
sample.
mant. Although the cut test is often very good at producing
Sampling by hand is sometimes necessary when the
some information quickly, it is limited in the amount of
seeds will not flow into the probe because of their size,
information it can supply and it lacks accuracy compared to
shape, or surface texture. Sampling by hand can be done by
more sophisticated procedures. Therefore, it should never
inserting the open hand (figure 4) into the seeds, closing it
be taken as a substitute for a formal laboratory analysis.
once the point of sampling is reached, and then withdrawing
Sampling it closed. The seeds are then placed in a second container to
form the composite sample, just as in sampling with the
Formal seed analysis begins with the sampling of the
probe. At least 5 handfuls must be taken, and all levels must
seedlot. The Rules for Testing Seeds (AOSA 1996) and the
be sampled. When the hand cannot be inserted into the
International Seed Testing Rules (ISTA 1996) both give
seedlot, the seeds can be poured from one container into a
instructions on how to draw samples from a seedlot so that
second. The tester then should stop at a minimum of 5 even-
the sample is representative of the entire seedlot.
ly spaced intervals and remove a handful of the seeds for
Representative means that any tests conducted on this sam-
the composite sample.
ple will accurately estimate the mean value of the lot
quality. The composite sample, whether taken with a probe or
Sampling can be done with the hand or with a seed by hand, is usually too large to submit to a seed laboratory
probe, also known as a trier (figure 1). If a probe is used, it for analysis. The composite sample is, therefore, mixed and
must be long enough to reach to the farthest edge of the divided to obtain a submitted sample. This procedure is very
container. A probe has gates that prevent seeds from enter- important and must be done correctly for the results to be
ing until the probe is inserted the full dimension of the con- accurate.
tainer. The probe should be inserted into the seed container The composite sample can be mixed either mechanical-
with these gates closed. Otherwise, seeds from the upper ly or by hand with rulers. Hand-mixing the composite sam-
layers will fill the probe as it is inserted and the bottom lay- ple is done by pouring the seeds into a cone on a flat, clean
ers will not be sampled. Once the tip reaches the bottom or surface. An open file folder makes a good work surface that
far side of the container, the gates should be opened and the can be picked up to return the seeds to a container. With one
probe gently turned back and forth to help the seeds fall in. ruler held stationary against the seeds, the second ruler is
Then the gates should be closed gently, not forced, so that used to pull the outer edge of the pile up to the top of the
any seeds that are caught in the opening and are preventing pile, allowing the seeds to roll down the sides and over the
the gates from closing fully (figure 2) are not crushed. top of the stationary ruler (figure 5). The full pile is thor-
(Mechanically damaged seeds would bias the sample.) After oughly turned over and all layers mixed together. This pro-
the probe has been withdrawn from the seed container, it cedure should be repeated for 1 full minute. Then the pile
should be held horizontally, with the gates facing upward. should be divided by cutting the cone in half and then into

98 • Woody Plant Seed Manual

Figure 1—Chapter 5, Seed Testing: seed probes are used Figure 3—Chapter 5, Seed Testing: the seed probe is
to sample free-flowing seeds. emptied by pouring the seeds out the top. 5

Figure 2—Chapter 5, Seed Testing: seeds caught in the Figure 4—Chapter 5, Seed Testing: an open hand is
gates of the seed probe must not be cut when the gates are inserted into a seedlot to take a sample for testing.
closed.

quarters. The quarter is then weighed to see if it is enough must not be turned on until the seeds have been poured
for the sample. If not, then another quarter, an eighth, or a completely into the hopper. Once the seeds are cleared out
smaller fraction is taken until the minimum weight is of the machine, the motor must be turned off before the
obtained (figure 6). seeds are poured back into the hopper for the next pass. The
Hand-mixing can be replaced by mixing with either a seedlot is then divided in half, then quarters, eighths, and so
soil divider or a gamet divider (figure 7). These devices can forth, to obtain the correct weight for the submitted sample,
save a substantial amount of time and also, by reducing the just as in the hand mixing and dividing.
tedious nature of the work, increase the likelihood of doing The size of the submitted sample for some species is
a quality job. The seedlot needs to be poured through the stated in the Rules for Testing Seeds (AOSA 1996) and is
divider 3 times. When the gamet divider is used, the motor twice as large as the minimum amount for the purity test.

Chapter 5: Seed Testing • 99

 This amount is different for each species and the rules need to the rules. If a species is not listed in the rules, an amount
5
to be consulted to be sure the correct amount is submitted that contains 2,500 seeds should be taken. This amount can
for purity tests that are to be done according to the rules. A be estimated by counting out 100 seeds and multiplying
smaller sample of seeds can be submitted, but the test will their weight by 25. Under the AOSA rules, samples can be
not be according to the rules and the accuracy cannot be as small as 600 seeds when only germination is tested. It is
assured to the same degree as a test that is done according important to work quickly when drawing the sample, if the
submitted sample is to be tested for moisture content. This
will prevent the gain or loss of moisture from the air. Once
Figure 5—Chapter 5, Seed Testing: seed can be hand- obtained, the submitted sample should be put in a moisture-
mixed before withdrawing a submitted sample from a proof container to maintain its true moisture content until it
composite sample.
is sampled and tested at the laboratory. Plastic bottles with
tight-fitting lids or tightly closed plastic bags of at least 0.1
mm (4 mil) thickness are adequate. Metal containers can be
used but are harder to find. Glass containers should not be
used; they easily break in transport, allowing the samples to
be exposed to the air or, worse, mixed together.

Sample Identification
Assignment of a test number is the first step in handling
every seedlot that is received in the laboratory. This number
allows for the orderly tracking of the test sample among the
other samples in the laboratory. A typical test number indi-
cates the test year and an accession number. For example,
the 300th test conducted in 2005 would have a number such
as 05-300.
Figure 6—Chapter 5, Seed Testing: the composite
sample is divided systematically into quarters, eighths, Moisture Tests
sixteenths, and smaller fractions to obtain the submitted Moisture tests must be the first tests conducted on sam-
sample at the seed storage plant or the working sample in ples when they arrive at the seed laboratory. Once a sample
the laboratory. container is opened and work begun, the seeds will likely

Figure 7—Chapter 5, Seed Testing: a soil divider (left)

and a gamet divider (right), devices that systematically mix
and divide seed samples.

100 • Woody Plant Seed Manual

gain or lose moisture in exchange with the ambient air. The Figure 9—Chapter 5, Seed Testing: a convection oven
and desiccator are used to conduct a seed moisture test. 5
standardized laboratory test for moisture content is the oven
method (ISTA 1996). This procedure was determined, after
many years of research, to be a best estimate of moisture for
general testing work (Bonner 1972, 1981, 1984, 1992;
Buszewicz 1962; Hart and Golumbic 1966). This test is
made on 2 subsamples containing 3 to 5 g of whole seeds.
These 2 samples are placed in containers with lids and
weighed to determine the wet weight (figure 8). Then they
are placed in a forced-draft drying oven (figure 9) for 16 to
18 hours at 105 ± 2 °C. The lids are removed during drying
but are also placed in the oven. The samples are then placed
in a desiccator to cool for about 20 minutes before being
weighed a second time to determine their dry weight. The
lids are placed on the cans while cooling and weighing. The
loss of weight represents the weight of water in the undried
sample. This water weight is divided by the wet weight to
obtain the percentage moisture content on a wet-weight Some larger seeds and seeds with impermeable seed-
basis. The percentage moisture is expressed on a wet-weight coats need to be cut to make an accurate test (Bonner 1974,
basis because this value most accurately represents how 1981, 1992). If the seed is not cut open, the moisture is not
much of the seedlot is water. Therefore, the buyer knows the freely released, and the moisture content is underestimated
weight of seeds and the weight of water that are purchased. (figure 10).
For example, when the price per weight is the same, a The oven method is not a direct measure of the content
pound of seeds at 7% moisture content is a better value than of water. It measures weight loss that is assumed to be due
a pound of seeds at 9% moisture content. For example, a to the loss of water. A basic analytical procedure is required
100-kg seedlot (or a 100-lb seedlot) at 7% moisture contains to verify the temperatures and length of drying. The current-
93 kg of seeds, whereas the lot at 9% moisture contains 91 ly accepted procedure is the Karl Fisher procedure (figure
kg of totally dry seeds, 2 kg less. 11) (Hart and Golumbic 1962). The moisture committee of
the ISTA uses this procedure in its work to standardize and
validate the oven procedures.
Another widely used method to measure seed moisture
Figure 8—Chapter 5, Seed Testing: 2 seed samples are
is the electronic moisture meter. Although there are numer-
tested to measure the moisture content of a seedlot.
ous brands of electronic moisture meters on the market, not
all of them will work for tree and shrub seeds (figure 12),
and those that do will not have calibrations for tree seeds.
Therefore, conversion charts must be developed for them by
testing samples with high to low moisture contents with
both the meter and the oven. A linear regression between
the oven and meter readings is calculated, and the conver-
sion chart predicted from this regression (Bonner 1981; Hart
and Golumbic 1966; Jones 1960; Karrfalt 1987; Lanquist
1965). These meters provide quick results, are nondestruc-
tive to the seed, and are usually accurate to within ±1% of
the moisture estimated by the oven method.

Chapter 5: Seed Testing • 101

 Figure 10—Chapter 5, Seed Testing: cutting large seeds Figure 12—Chapter 5, Seed Testing: electronic moisture
5 open before drying them in the moisture test is necessary testers can give a quick and reasonably accurate estimate of
to release all the moisture. seed moisture.

Figure 11—Chapter 5, Seed Testing: the Karl Fisher contains the minimum weight for conducting a purity analy-
apparatus is used as the analytical standard for determining sis. Each species has its own specified minimum weight,
seed moisture content. which has been determined to contain 2,500 seeds. The
mixing and dividing should be done in the same way as
described in the sampling section for drawing the submitted
sample from the composite sample. However, at this point it
is necessary to be very close to the minimum weight for
2 reasons. First, the analyst does not want to examine more
seeds than necessary, and second, the accuracy of the test is
evaluated using tolerance tables that were developed using
these minimum weights. Using substantially more seeds than
the minimum will invalidate the use of these tables.
Purity is determined differently by each of the 2 major
testing organizations. The ISTA rules specify a 3-part purity
and the AOSA rules specify a 4-part purity. The ISTA purity
values report percentages of pure seeds, other seeds, and
inert materials. The AOSA purity values report percentages
of pure seeds, weed seeds, other crop seeds, and inert mate-
Purity, Noxious Weed Content, and Seed rials. The pure-seed fraction consists of all those seeds that
Weight Tests are of the kind specified on the seedlot’s label. Specific
descriptions in the rules define “pure seeds,” but basically
Purity, noxious weed content, and seed weight tests are
the pure-seed fraction comprises whole seeds and seeds that
sometimes called physical tests because they do not relate to
are not more than half broken away. “Other seeds” in the
viability. These tests are described individually as follows.
ISTA rule are all kinds of seeds other than those listed on the
Purity Analysis label. The AOSA rule makes a distinction between “crop
After samples for the moisture-content test are with- seeds” and “weed seeds” and uses a detailed list (AOSA
drawn, the remainder of the submitted sample should be 1995) to specify when a species is a weed and when it is a
mixed and divided to obtain the working sample, which crop. Weed seeds are mainly a problem in lots collected
from nets or directly from the ground. Contaminated clean-
ing equipment can also result in weed seeds entering a seed-

102 • Woody Plant Seed Manual

lot. “Inert matter” is all other material that is not classified as Figure 13—Chapter 5, Seed Testing: a purity sample is
divided into its component parts. 5
crop seeds or other seeds. It could include soil particles,
stones, wire, small pieces of broken seeds, or other plant
parts. Purity is calculated by dividing the weight of the of
pure seeds by the total weight of all the fractions in the sam-
ple (figure 13) and is expressed as a percentage.
Purity work can often be tedious and very technical.
Devices such as the mechanical purity board (figure 14) can
speed up the procedure. The analyst must understand impor-
tant taxonomy principles and accurately use the seed herbari-
um (figure 15) to identify all the kinds of seeds in the
sample.

Noxious Weed Examination

The noxious weed exam is a specialized purity examina-
tion. It is not a test traditionally associated with forest seeds Figure 14—Chapter 5, Seed Testing: a mechanical purity
but may become more common as the commercial exchange board can reduce the time required to conduct a purity
analysis.
of native plants increases. A noxious weed is a highly
aggressive competitor or a plant with other highly objection-
able characteristics, such as being poisonous. It is so offen-
sive it has been put on a noxious weed list compiled by an
individual state or the federal government. A noxious weed
exam is made solely to identify the number of noxious weed
seeds found in the sample. Nothing else is noted in this
exam. The presence of any noxious weed seeds makes it ille-
gal to sell the seeds until the noxious weeds have been
removed. The sample size for a noxious weed examination is
25,000 seeds.

Seed Weight Determination

The number of seeds per unit weight (kilogram and gram
or pound and ounce) is determined on the pure-seed fraction Figure 15—Chapter 5, Seed Testing: seed herbaria are
from the purity test. This test is called the seed weight deter- used to make positive identification of the species of seed
mination in the ISTA rules. It is made by counting out 8 tested.
replicates of 100 seeds and weighing them to the same preci-
sion as the weights for the purity test. The coefficient of
variation for these 8 values is computed. This coefficient
cannot be greater than 6 for chaffy seeds or greater than 4
for all other seeds. Otherwise, an additional 8 replications
need to be counted and weighed and combined with the first
8 weights. All 16 weights are then used to compute the
mean. Any weight diverging from the mean by more than 2
standard deviations is discarded; only the remaining weights
are used to compute the number of seeds per unit weight.
Seeds can be counted by hand, with a counting tray, a
shutter box, or a vacuum counter (figure 16). When seeds are
counted by hand, it is usually best to count out the appropri-
ate number of piles of 10, 20, or 50 seeds, in order not to
lose one’s place. A counting tray is simply a block of wood

Chapter 5: Seed Testing • 103

 or plastic with impressions drilled into it to hold the seeds as Seed weights are sometimes determined with an elec-
5
they are poured across the plate. The counting tray may or tronic counter (figure 18). The ISTA rule calls for counting
may not be faster than counting by hand, depending on the all pure seeds in the working sample when this is done. No
seeds’ size and shape, which determine how many double- error-check is then made. A recent internal report made by
or triple-seeded holes must be thinned by hand. Using the the Seed Count Committee of the Association of Official
shutter tray is similar to using the counting tray, but the shut- Seed Analysts, augmented by the author’s personal observa-
ter tray is emptied by sliding the bottom tray rather than tions, suggests caution in the use of electronic counters for
turning it over. For uniformly sized seeds, using either the seed weight determinations. A high potential for error in
counting tray or the shutter tray can speed up counting con- counts exists. If carefully calibrated, these machines can
siderably. count quite accurately, but the machines need to be adjusted
The vacuum counter is probably the fastest and one of and used correctly. A thorough evaluation of the degree of
the most common ways to count seeds in the laboratory. The desired accuracy and the amount of time required to achieve
vacuum counter is made of an acrylic (or sometimes metal) it needs to be made before deciding to use the electronic
plate that is drilled with 25, 50, or 100 holes and attached to counter.
a hollowed-out second plate. A vacuum line is attached to
the head and a shut-off valve controls the application of the Germination Testing
vacuum. To use the vacuum counter, the seeds are spread out Germination testing is designed to estimate the maxi-
loosely in a 1-seed-deep layer, the counter is placed on top mum number of seeds that will produce a normal seedling
of the seeds, and the vacuum is then turned on. Moving the and to give results that are as repeatable as possible. Without
counting head back and forth for about the diameter of a uniform procedures, there would be no standard on which to
seed will bring the seeds into contact with a vacant hole. base the value of seedlots for commercial transactions and
With the vacuum still on, the seeds can be transferred to a the seed trade would be chaotic and filled with dispute.
dish for weighing or to a germination container (figure 17). Germination also tells a grower about a seedlot’s potential.
Some users of vacuum counters report a tendency for lighter A seedlot with 80% germination cannot produce more than
seeds, such as empty or partially filled seeds, to be picked 80 seedlings per 100 seeds. Therefore, if 100 seedlings are
up in preference to heavier seeds. To eliminate this problem, needed, a minimum of 125 seeds must be planted (100/0.80
this device must be used according to the procedure = 125). How to use test data to compute sowing rates is pre-
described above. sented in detail in chapter 7 (Nursery Practices) and later in
this chapter in the section on the use of test data.

Figure 16—Chapter 5, Seed Testing: seeds can be count- Figure 17—Chapter 5, Seed Testing: a vacuum counter is
ed sometimes more quickly using a counting tray, a shutter often used to count out seeds for weight determinations
box, or a vacuum counter. and for planting germination tests

104 • Woody Plant Seed Manual

 The germination test is conducted on the pure-seed frac- Counting devices are described in the seed weight discus-
tion from the purity test. Both the AOSA and ISTA pre- sion above. Seeds should be hand-planted only when count- 5
scribe the use of 4 replications of 100 seeds. These replica- ing devices cannot be used in order to save time.
tions can either be planted 1 to a container (figure 19), 2 to Seeds can be germinated on various media. Sand, sand
a container, or all on 1 tray. Alternatively, the 4 replications and perlite mixtures, potting mixtures, soil, and various
can be further divided into smaller replications, but the total papers—blue blotters, white blotters, or crepe-cellulose
number of seeds tested must remain 400 to remain in com- papers (such as Kimpak®)—can be used (figure 20). Testing
pliance with the rules. If fewer than 400 seeds are available, rules, however, specify what is an acceptable medium for
then the number of seeds per replication should be reduced the kinds of seeds tested. Specifying the medium helps
so that an equal number of seeds is present in each of the 4 assure uniformity in test results. The blotters resist penetra-
replications. Using fewer than 100 seeds in a replication is tion by the roots of the plants, whereas the crepe-cellulose
not according to the rules, and the test would thus be unoffi- paper allows for root penetration. Blotters offer the advan-
cial. However, it is better statistically to have 4 replications tage of keeping the roots where the analyst can actually see
of 50 seeds each rather than 2 replications of 100 seeds them for evaluation, but if a seedling is very large it will fall
each. The 4 replications are then placed under optimal ger- over and tangle with other seedlings, making counts diffi-
mination conditions for the period specified in the rules. cult. The media also differ in their water-holding ability.
Germination is the number of normal seedlings produced Blotters usually need to be watered several times during the
from 100 pure seeds expressed as a percentage. A normal test, whereas crepe-cellulose paper, sand, sand mixtures,
seedling has all the essential plant structures necessary for potting soils, and soil are absorbent enough to hold all the
the plant to continue to grow normally under favorable con- water the seeds need for up to 3 months, if kept in a mois-
ditions (AOSA 1996; ISTA 1996). ture-proof container. Watering the medium can be done by
Seeds can be planted in a number of ways. They can be hand or by machine. Watering by hand is usually done using
scattered or placed one at a time with forceps, although a squeeze bottle or a small hose from the tap and requires
more generally a vacuum counter or other type of planting subjectivity on the analyst’s part to estimate that the correct
plate is used for speed and to ensure even spacing of the amount of water has been applied. Too much or too little is
seeds. The vacuum counter is the most expedient technique, harmful, but in most cases there is wide latitude in the
because it can handle a variety of seed sizes (figure 17). amount that will give optimal results (Belcher 1975).

Figure 18—Chapter 5, Seed Testing: an electronic seed Figure 19—Chapter 5, Seed Testing: 2 germination tests,
counter is sometimes used to estimate the number of each composed of 4 dishes containing 100 seeds each, the
seeds per weight (in either kilograms or pounds). dishes are stacked for transport and prechilling.

Chapter 5: Seed Testing • 105

 Machines for watering include automatic pipetting machines Figure 20—Chapter 5, Seed Testing: seeds are germinat-
5 (figure 21) or small traveling spray booms. Both save a great ed on various media, from right to left: crepe-cellulose
amount of time if many tests are conducted and, once adjust- paper (such as Kimpak®), blue blotters, sand–perlite mix-
tures, and potting soil.
ed, take all the guesswork out of applying the correct
amount. These machines should be checked periodically,
however, to verify that they are in fact applying the desired
amount of water.
Germination tests should be run in cabinets or rooms
that meet exacting requirements for temperature and light
control in order to make accurate and repeatable estimates.
Temperatures should be carefully checked throughout the
chamber at the level of the substrate to be sure there are no
places that deviate from the desired temperature by more
than 1 °C. Poor air circulation and hot spots from lights or
light ballasts are the most common causes of temperatures
that are too high or too low. The temperature at which the
germination chamber is set depends on the species being
tested. Many species do well at an alternating 20 and 30 °C.
For this regime, the chamber is held for 16 hours at 20 °C
and for the remaining 8 hours of the day at 30 °C. Other Figure 21—Chapter 5, Seed Testing: an automatic pipet-
possibilities are constant temperatures of 15, 20, or 22 °C, ting machine can help to uniformly and rapidly water
with light usually supplied for either 8 or 16 hours. When germination dishes.
temperatures alternate, the light is provided during the high-
er temperature to follow a natural cycle of light and tempera-
ture. Sources of light need to contain abundant amounts of
blue and/or red light but not far-red light because far-red
light is known to inhibit germination. Cool white flourescent
lamps are most commonly used. The temperature/light
regime used for a germination test is determined by experi-
ments that germinate the same seedlot at different tempera-
ture/light combinations. The combination that supports the
highest percentage of germination in the most reasonable
time is the one that is then adopted in the rules for testing.
Dormancy is the condition of a seed that prevents it
from germinating when it is placed in conditions that are
favorable for germination. (For a discussion of dormancy,
seeds are soaked for 16 to 48 hours in water to become fully
see chapter 1.) Dormancy must be overcome in order to
imbibed, placed in a moisture-proof container, held in the
conduct the germination test, just as when trying to grow
cold for the specified period, and then planted on the germi-
seedlings. Pre-germination chilling (commonly called
nation medium. This last procedure is sometimes called
“prechilling” and traditionally called “stratification”) is the
naked stratification, because no moisture-holding medium is
procedure most used for breaking dormancy in forest seeds.
used (figure 22). How long seeds are held in prechill varies
The seeds are held in moist conditions at temperatures
widely by species and genetic source of the seedlot. The
between 0 and 3 °C. Pre-germination chilling can be accom-
period can range from 10 days to many months. For some
plished in 1 of 3 basic ways. In the first, the seeds can be
species, a warm period preceding the cold period is required.
planted on moistened germination medium in sealed contain-
This is called “warm stratification” or “warm incubation.”
ers and then put in the cold. In the second, the seeds can be
Western white pine (Pinus monticola Dougl. ex D. Don)
placed in a moist medium, placed in the cold, and then at the
(Anderson and Wilson 1966) and European ash (Fraxinus
end of the prechilling period, planted on the germination
excelsior L.) (Piotto 1994) have been reported as requiring
medium. In the third method (similar to the second), the
this warm-cold stratification .

106 • Woody Plant Seed Manual

 A species that does not require prechilling is called non- it is impossible to predict the weather with the necessary
dormant. If 10 to 14 days of prechilling are needed, the dor- 5
precision. Vigor testing is one possible solution. The vigor
mancy would be considered light. If 30 to 60 days of test does not predict performance for a particular set of fluc-
prechilling are required to break the dormancy, it would be tuations; rather, it predicts the general ability of a seedlot to
considered moderate. More than 60 days of prechilling clas- germinate normally over a range of adverse conditions. Its
sifies the seedlot as highly or strongly dormant. The degree purpose is to differentiate seedlots, with essentially equal
of dormancy varies within the seedlot of even lightly dor- germination, according to their ability to germinate well in
mant species; some seeds germinate without prechilling, spite of adversity. Figure 23 illustrates the relationship
whereas other seeds in the same lot will not germinate until between vigor and germination. As seeds age and begin to
they are prechilled. However, the term variable dormancy is weaken and die, vigor declines before germination test
usually reserved for seedlots in which some seeds germinate results decline (Belcher 1978; Justice and Bass 1978).
during prechilling, whereas other seeds in the same lot will Like germination tests, vigor tests are conducted under
not germinate even after being placed in favorable germina- standardized conditions in order for the results to be repeat-
tion conditions. Species that fit the deep and variable dor- able and useful in the field. A vigor test cannot make up for
mancy category are Rocky Mountain juniper (Juniperius poor practices that unnecessarily increase environmental
scopulorum Sarg.) and basswood (Tilia americana L.). variation in the field; such poor practices can be major
Because of the above-mentioned variation in dormancy, sources of disparity between laboratory and field germina-
seedlots will often be tested with and without prechilling or tion. Uniform sowing depth and watering, as well as sowing
with varying lengths of prechilling. Such tests are referred to only on soil at the minimum acceptable soil temperature—
as paired or double tests; usually only 2 tests are done. More all help make field germination more predictable.
tests, of course, can be and are done with some seedlots. The most common vigor tests in agriculture are the cold
This type of testing can determine the presence of dorman- test, the accelerated aging test, the conductivity test, and the
cy, the strength of dormancy, or a weakness in the seeds tetrazolium test. These 4 tests have not been used very much
(Belcher 1995). When the seedlot has the same germination for forestry. In addition to these tests, speed of germination
with and without prechilling, it is said to be nondormant. as expressed in a number of formulas has been put forward
When the germination is increased with prechilling, the for use in forestry as a vigor test. Despite the potential bene-
seedlot is classified as dormant; the longer the prechilling fit for tree seed nurseries, the science and technology are not
period needed, the stronger the dormancy is said to be. A advanced enough to permit the practical application of vigor
decrease in germination with prechilling is an indication of testing with forest species.
weakness in the seeds. This last condition is similar to the
situation of the type of vigor test known as the cold test,
which is described in the following section. Figure 22—Chapter 5, Seed Testing: seeds can be
Prechilling is not the only treatment to break dormancy. prechilled on a germination medium, in separate medium, or
Light is useful to break dormancy and can reduce the need in a plastic bag.
for prechilling. Birches (Betula L.) and loblolly pine (Pinus
taeda L.) are prime examples where light helps break dor-
mancy. Seedcoat dormancy is treated by scarifying the seed-
coat with either acid, bleach or mechanical means. Chemical
stimulates such as gibberellins or potassium nitrate have
been little used with forest tree seeds.

Vigor Testing
Sometimes standardized laboratory germination proce-
dures are criticized as not predicting field performance very
well (Moreno 1985; Stein 1967). These critics suggest using
a variety of test conditions to find an optimum for each
seedlot. The problem in predicting field germination is that

Chapter 5: Seed Testing • 107

 The cold test is done by planting seeds in damp soil and This greater leakage causes the water to have a higher con-
5 then holding the germination tray at 10 °C for 7 days. This ductivity, which can be measured with a conductivity meter
test mimics the cool damp conditions of soil in early spring. (figure 25). Bonner and Agmata-Paliwal (1992) reported on
At the end of the period, the germination trays are trans- the use of conductivity for tree seeds and found that results
ferred to the appropriate temperature for germination. The have poor repeatability for precise estimates but possibly
higher the percentage of germination, the more vigorous the would work for general estimates of classes as poor, low,
seedlot is said to be. High-germinating lots have the needed intermediate, or high viability.
strength or vigor to pass the period of stress and still have Several statistics have been put forward to use speed of
energy for high germination when conditions are favorable. germination as an indicator of vigor. The faster a seedlot
This is analogous to the case of the tree seedlot that drops in completes germination or reaches its peak, the more vigor-
germination following prechilling. Those that drop in germi- ous it is said to be. The simplest indicator is days to 90% of
nation after prechilling are weak. total. For example, if the final germination is 88%, the indi-
The accelerated aging test is conducted with the stress of cator would be how many days it takes to reach 79% germi-
high temperature and moisture. The given weight of seeds is nation. A lot that reaches 79% in 12 days would be more
placed in a small box with a screen tray that suspends the vigorous than one that takes 16 days. To use this statistic,
seeds over a reservoir of water (figure 24). These boxes are counts must be made quite frequently, even daily, or the
then placed in an aging chamber at 40 to 43 °C for 72 to data must be interpolated to determine the number of days
288 hours, depending on the species. Whichever temperature to the specified germination.
is chosen in this range, the variation must be virtually nil to Czabator’s factor (1962), developed for use with south-
ensure repeatability of the results. The water-jacketed incu- ern pines, combines the maximum daily average germina-
bator has been determined by organized tests among labora- tion, called the peak value, and the average daily germina-
tories to be the best device to give this necessary strong con- tion at the end of the test to form one statistic called the
trol over the test conditions. At the end of the period, the germination value. Germination is counted frequently, at
seeds are planted and tested for germination under the stan- least every third day, and the cumulative germination on
dard conditions. each day is divided by the number of days that the test has
The electrical conductivity test has been widely tested in been run in order to compute the mean daily germination
agriculture but has not been adopted as routine practice for that day. For example, if on day 22 the cumulative ger-
except in a few specialized areas. In this procedure, seeds mination is 88, the mean daily germination is 4. This mean
are soaked individually or in bulk. Deteriorated or dead
seeds leak electrolytes more readily than high-vigor seeds.
Figure 24—Chapter 5, Seed Testing: the accelerated
aging test is conducted by placing seeds in a plastic box
Figure 23—Chapter 5, Seed Testing: as seed viability with a water reservoir and holding them at 40 ºC for
decreases, the proportion of live low-vigor seeds increases. 72 hours.

108 • Woody Plant Seed Manual

daily germination increases with each day of germination temperature must be very tightly controlled. Just a degree or
two difference in temperature can change the speed of ger- 5
until the period of maximum germination has ended and
then decreases. The largest value of the mean daily germina- mination, effecting the value of those statistics that rely on
tion is called the peak value. Figure 26 shows a graph of a germination speed. The accelerated aging test was difficult
typical germination. Initially only a few seeds germinate, to standardize until chambers were developed that had virtu-
followed by a period of rapid progress, and finally a slow- ally no variation in temperature. Difficulty in standardizing
down period and an end of germination altogether. laboratory tests and the lack of clear and consistent interpre-
Germination value is computed by multiplying the peak tation to the field has prevented the operational use of vigor
value by the mean daily germination. Lots that have higher testing.
For a more complete list of literature references and a
germination values are generally considered more vigorous.
detailed explanation of the vigor testing procedures, refer to
Another characteristic of more vigorous lots is that they
the Association of Official Seed Analysts’ Seed Vigor
store for longer periods of time without loss of germination.
Testing Handbook (AOSA 1983) and the International Seed
Therefore, if 2 seedlots have equal germinations, the one
Testing Association’s Handbook of Vigour Test Methods
with the lower vigor might be considered for first use,
(ISTA 1995). Bonner (1998) has also made a thorough
because the germination of this lot will likely decrease
review of vigor testing specifically for tree seeds.
faster than the lot with higher vigor. This approach would
give the greatest potential number of seedlings. Lower vigor Chemical Staining for Viability
seedlots will loose viability even under ideal conditions in The tetrazolium staining procedure mentioned in the
the freezer. vigor section is useful in estimating the viability of dormant
Tetrazolium staining has been tried also as a vigor test seeds, especially very dormant ones. This test involves soak-
(Moore 1976). Because of the highly subjective nature of ing the seeds first in water so that they imbibe fully and
this test and the great amount of experience it requires to soften for cutting. A moistened seed will take up the stain
administer, it has never been widely used as a vigor test and more rapidly. A variety of methods are used to open seeds. It
never successfully with forest plants. As stated in the next is extremely important that no damage occur to the embry-
section, tetrazolium staining can be used to successfully onic axis when a seed is cut. The embryonic axis is the radi-
estimate viability for very dormant species or for other hard- cal and the plumule. The meristematic regions are here, and
to-germinate species. their condition needs to remain unaltered until they are care-
The problem with vigor analysis is that it has proven to fully examined. These are the areas where the embryo must
be difficult to standardize, apparently because the test condi-
tions are so exacting. In tests involving germination, the
Figure 26—Chapter 5, Seed Testing: germination curves
of 3 pairs of seedlots, with high, moderate, and low germi-
nation [Link] upper curve in each pair represents the
Figure 25—Chapter 5, Seed Testing: the conductivity more vigorous lot in the pair because the germination is
meter is used to measure seed viability or vigor by estimat- completed sooner.
ing the amount of cations lost from deteriorating seeds.

Chapter 5: Seed Testing • 109

 grow in order to produce a normal seedling. Usually forceps Figure 27—Chapter 5, Seed Testing: seeds are cut open
5 carefully to prepare them for tetrazolium staining.
and sharp single-edged razor blades are used to cut open the
seeds (figure 27). For seeds with harder or stony seedcoats,
a variety of vises, hammers, and clippers are used to cut
through or remove the seedcoat (figure 28).
The solution that is used to make a tetrazolium (TZ) test
is colorless. It is made by dissolving 2,3,5-triphenol tetra-
zolium chloride in a phosphate buffer at pH 7.4, which is the
optimum pH for the TZ reaction. The buffer is necessary to
compensate for any pH imbalance in the TZ salt, the water,
or possibly the seeds. The colorless solution is taken up by
the prepared seeds and then reacts with respiratory enzymes
(that is, dehydrogenases) to form an insoluble light pink
(magenta) precipitant. Tissues that are alive and respiring
will stain, and those that are not alive will not. For a detailed
discussion of this procedure, refer to the AOSA Handbook
on Tetrazolium Testing (AOSA 2000). The TZ test can be
completed in 4 to 48 hours, depending on the amount of
preparation time required and the rate of staining. Figure 28—Chapter 5, Seed Testing: vises, hammers, or
Tetrazolium staining has proven useful with many clippers are used to cut through or remove hard or stony
species that have deep dormancy, including tuliptree seedcoats for conducting tetrazolium or excised embryo
(Liriodendron tulipifera L.), baldcypress (Taxodium dis- tests.
tichum (L.) Rich.), Rocky Mountain juniper, and sumac
(Rhus L.) species. However, for a few species with very
deep dormancy, there will be no staining (Vivrette 1995)
unless the seeds are prechilled.

Excised Embryo Testing

The excised embryo test is done on the embryo after it is
removed from the seed (Flemion 1948; Heit 1955). In this
germination test, the embryo has been freed from the
restriction of the seedcoat and nutritive tissue (figure 29).
Therefore, a germination that would take many months and
be incomplete can be complete in 10 to 14 days. The ashes
(Fraxinus L.), maples (Acer L.), and cherries and plums
(Prunus L.) are some of the genera that are tested by
embryo excision. Because the embryos are very vulnerable
to infection once excised, the test must be done under strict- nificant injury or broken apart. The main advantage of this
ly clean (axenic) conditions. The work surface and all tools, method over the TZ test is that the evaluation is less subjec-
hands, and germination dishes should be washed carefully, tive; the growing embryo is actually observed in most cases.
perhaps with absolute ethanol. If “clean” embryos mold eas- Therefore, a direct reading on the growth potential of the
ily, then cleaning procedures must be reviewed for effective- seedlot can be obtained.
ness and the work area examined for sources of microbial A complete excision is not always required. Russian-
contamination. Generally, however, sterilization procedures olive (Elaeagnus angustifolia L.) (Belcher and Karrfalt
such as autoclaving are not required. The procedures for 1979) and western white pine (Hoff and Steinhoff 1986)
excising the embryos are similar to those used in preparing respond well to only partial removal of the seedcoat.
seeds for tetrazolium. Greater care is needed, however, However, white pine seeds require prechilling before their
because the embryo must be removed intact without any sig- seedcoats are cut open.

110 • Woody Plant Seed Manual

X-Radiography is more expensive and slower to develop (over 45 minutes)
5
X-radiography is very useful in forest seed analysis. It and a light table is required to see the images. Usually the
provides a very rapid and accurate analysis of the internal film or paper is placed in a paper or vinyl cassette so that it
structure of seeds, identifying empty, insect-damaged, or may be handled in the light. This cassette can result in some
poorly developed seeds (figure 30). It is an immense help in loss of clarity of the image, especially with small seeds;
judging maturity, determining how many bad seeds should using the film in the dark where the seeds can be laid direct-
be remove, and detecting any mechanical injury. It is more ly on the emulsion gives a noticeably superior image.
accurate than cutting tests in many cases, requires much less Seed work is usually done with x-rays in the range of 10
time, and is nondestructive (AOSA 1979; Simak and others to 30 kvp (that is, kilovolt potential), which is is the amount
1989). X-radiography was first applied to tree seeds by of penetrating power the x-rays have. The exact kilovolt
Simak in Sweden. The use of contrast agents has improved potential depends on the equipment and the seeds in ques-
the ability of the x-ray test to discriminate between viable tion. Trial and error is necessary to find the best combina-
and nonviable seeds with some species (Kamra 1963; Simak tion. Too high a kilovolt potential and the seeds will not be
1957; Vozzo 1978). A contrast agent enters damaged areas visible or will appear too dark. Too low a kilovolt potential,
of the seed differentially from nondamaged areas, making and the image will lack detail and be too light. Some small
the damaged areas more radiopaque. They will then appear seeds need to be x-rayed at a low kilovolt potential to give
as bright areas on the radiograph. Aqueous solutions of the correct penetration but need a long exposure to produce
heavy salts such as iodine or barium chloride and vaporous a radiograph with enough density to provide good contrast.
agents such as chloroform have been used as contrast X-ray inspection cabinets are manufactured that operate in
agents. this very low kilovolt potential range for examining small
Radiographs can be made on Polaroid® film, x-ray items such as seeds. They are designed for total protection
paper, or x-ray film. Polaroid film is useful if no darkroom of the operators, with complete lead shielding and safety
is available, because the film is developed in the light, just interlocks on the door (figure 31).
like a Polaroid photograph. The disadvantages of Polaroid X-radiography has proven useful for studying the seeds
are high cost, short shelf life, and lack of detail. X-ray paper from many wild species, which often can be empty or poor-
is fast to use but does require a simple darkroom. It is less ly formed. Some laboratories test every lot of seeds they
expensive than Polaroid, has a shelf life of several years in receive with x-rays and get a good initial evaluation.
cold storage (3 °C), and much better resolution. The best X-radiography can be of great value in evaluating germina-
resolution is obtained with x-ray film. X-ray film, however, tion test results, because it is much faster than cutting open

Figure 29—Chapter 5, Seed Testing: embryos of peach Figure 30—Chapter 5, Seed Testing: x-radiography can
(Prunus persica L.) have been removed from their seedcoats be used to quickly determine how many seeds are empty,
for an excised embryo test of viability. damaged, or poorly developed.

Chapter 5: Seed Testing • 111

 seeds that failed to germinate. Empty seeds will never ger- seed. With slash and longleaf pines, cutting can reveal
5
minate, and damaged or poorly developed seeds will seldom embryos that have initiated chlorophyll (turned green), a
germinate. The excised embryo or tetrazolium test for diffi- result of a seedlot having been held too long at high mois-
cult-to-cut seeds can be speeded up by x-radiography. The ture. This is a sure sign of a weakened seedlot.
seeds are first placed on the x-ray film or paper in a manner Hydrogen peroxide (H2O2) has been used as a quick test
that will allow the comparison of the exact image to the for western conifers (Ching and Parker 1958). In this test,
exact seed. This is done by placing the seeds on an addition- the seeds are floated in solution of 1.0% hydrogen peroxide
al piece of paper before placing the paper on the x-ray film overnight. The radicle ends are then clipped and the seeds
or paper. If orientation of the seed is important, as in incubated in the dark at 20 to 30 °C for 10 to 12 days.
double-seeded fruits such as dogwood, the seeds can be Counts of germinates are made at 3 to 4 days and at 10 to
placed on adhesive tape and that then laid on the paper. The 12 days. The hydrogen peroxide solution is changed at the
seeds should be oriented so that both seeds in the fruit can first count.
be viewed and the tape prevents them from turning. After
the radiograph is made, the seeds are gently slipped off the Sowing Rates
x-ray paper so that the seeds are kept in order for cutting. A sowing rate is the amount of seeds sown in a unit area
Only those seeds that are morphologically sound in the of nursery bed to produce the desired number of seedlings.
radiograph need to be cut. The following formulas show how seed test data are used to
determine this rate.
Other Quick Tests
As stated in the introduction, cutting tests are very limit- Weight of seeds to sow in a nursery bed
ed in their application. However, they can provide useful (width x length) is equal to
information on full seed percentages and the condition of
(bed width x bed length x seedlings desired per area) ÷
the internal structures. For example, color of the tissue can-
(germination x seeds per weight x purity x survival factor)
not be determined in a radiograph, which is only black and
white. Seeds that are cut and found to be dark are not likely Number of seeds needed to sow per area of nursery
to germinate. New and unfamiliar images in a radiograph bed is equal to
require cutting the seed to determine what is actually in the (seedlings desired per area) ÷ (germination x survival
factor)

Figure 31—Chapter 5, Seed Testing: a cabinet x-ray

In both of these formulas, the survival factor is the ratio
system is a safe and simple way to make radiographs of
seeds. of the number of seedlings expected to the number of viable
seeds planted. It is derived from experience in the given
nursery and should be constantly updated with new informa-
tion collected from history plots. History plots are perma-
nent sample plots in a nurserybed used for carefully moni-
toring the number of seeds sown and the number and quality
of seedlings produced (Landis and Karrfalt 1987). For
example, if 100 seeds are sown on a square foot, germina-
tion is 80% in the laboratory, and 60 seedlings actually grow
on the square foot, then the survival factor is 60 ÷ 80 or
.75 (75%).
Computing sowing rates for containers is somewhat dif-
ferent, because we must predict the probability of an empty
cell in the container. The probability that a container cell is
empty is equal to 1 minus the probability that at least 1
seedling is in the cell. Sowing 1 seed per cell, this probabili-
ty is 1 minus germination. With a 90% germination, the
probability of an empty cell following single-seed sowing is
0.1. In sowing 2 seeds per cell, the probability of no

112 • Woody Plant Seed Manual

seedling in a cell drops to .01, but now there are 81 cells seedlot when the buyer takes possession of it. Ideally the
5
(.90 × .90) out of 100 that will have 2 seedlings per cell. tests should be run by a disinterested third-party laboratory
Double seedlings per cell requires thinning to 1 seedling per that is well qualified to do the tests. The results of informal
cell for proper growth. Thus, in container nurseries it is nec- analysis, such as the cut test, should never be accepted as
essary to choose between empty spaces and thinning. In this the true measure of the worth of a seedlot.
example, to go from 10 empty cells to 1 empty cell per 100 Which tests are important to request, and how should
cells, 81 seeds were wasted. For 10,000 seedlings, 8,100 they be used? Moisture content is important for 2 reasons.
seeds would be wasted, which would be 112 to 224 g (3.9 to First, the seeds need to be at a proper storage moisture con-
7.8 oz) of seeds when the seeds number 72,300/kg (32,900 tent to ensure viability. Orthodox seeds need to have a mois-
seeds/lb). Thinning also requires more labor and may be ture content below 10% and recalcitrants usually above
dysgenic by favoring early germinating genotypes. Sowing 25%. Second, it must be remembered that extra water can be
extra containers is another strategy followed to compensate added to the seeds and distort the true value of the lot. One
for empty cells. The empty cells are still present but enough kilogram of a seedlot with 10,000 seeds/kg (22,000 seeds/lb)
seedlings are produced without the problems of thinning. at 7% moisture content would contain 70 g (2.5 oz) of water
The purity and seed per weight are still important to the and 930 g (32.6 oz) of dry seeds. A similar lot of 10,000
seeds/kg at 9% moisture would have 90 g (3.2 oz) of water
container grower, because they will be used to compute the
and 910 g (31.9 oz) of dry seeds. In an accurate comparison
amount of seed to prepare. The following formula can be
between the lots (that is, dried to equal moisture content of
used to calculate how many seeds are in a unit weight of
7%), the mass of the seeds in the second lot is 20 g (2%)
seeds:
less than that of the first. Although both seedlots might
appear to have seeds of the same size, the lot with higher
weight of seeds × purity × no. of seeds per unit
moisture would actually have slightly smaller sized seeds.
weight
Because the water is free, adding extra moisture can be a
good way for the seller to increase profits.
Example: 1 kg of seeds at 98% purity, 33,000 seeds/kg: The number of pure live seeds per weight is a calcula-
1 kg x 0.98 x 33,000 pure seeds/kg = tion that is often helpful in assessing the value of a seedlot.
32,300 pure seeds In this procedure, germination, purity, and seed weight are
all considered. Consider a seedlot with germination of 90%,
or 1 lb of seeds at 98% purity, 15,000 pure purity of 98%, seed weight of 18,600 seeds/kg (8,500
seeds/lb: seeds/lb). The number of pure live seeds per kilogram is
1 lb x 0.98 x 15,000 seeds/lb = then .90 × .98 × 18,600 = 16,400 seeds/kg (7,500 seeds/lb).
14,700 pure seeds This is the same value as a lot that has 95% germination,
96% purity, and 18,000 seeds/kg (8,200 seeds/lb) (.95 × .96
To sow 10,000 cells with 1 seed each, 10,000 seeds are × 18,000 = 16,400) but a higher value than a lot with 97%
needed, which is 10,000 seeds divided by 32,300 seeds/kg = germination, 84% purity, and 17,400 seeds/kg (7,900
0.31 kg (10,000 divided by 14,700 = .69 lb). Double sowing seeds/lb) (.97 × .84 × 17,400 = 14,200). Number of pure live
10,000 cells would take 2 x 0.31 = 0.62 kg (2 x .69 = 1.4 seeds per weight tells the grower the potential number of
lb). Combining these steps yields the following formula: seedlings and removes at least some ambiguity in comparing
the value of different seedlots. If the maximum number of
(number of seeds to sow per cell x number of cells to sow) potential plants is the most important factor, the first 2 lots
÷ (purity x seeds per unit weight) are superior to the third, although the third has a higher via-
bility. Alternatively, as might be the case for a container
In the double-sowing example, this would be 2 × 10,000 ÷ nursery, the high germination could be the most important
factor, and the third lot would be chosen over the first two,
.98 x 33,000 seeds/kg (.98 × 14,700 seeds/lb) for a total of
even though potentially fewer trees can be produced from it.
0.62 kg (1.4 lb) of seeds required.
Minimum standards are usually set for all quality values.
Buying and Selling Seeds Those minimums depend on the type of nursery, the general-
Current test data are essential. To be current, the data ly available quality for a species, the desirability of the seed
should not be more than 9 months old. The more recent the source, and other factors. In general, the higher its quality,
test, the more likely it is to reflect the true condition of the the more a seedlot is worth in the nursery. High germination

Chapter 5: Seed Testing • 113

 is of great value in a bareroot nursery but indispensable to to be measured: germination, seed weight, purity, and mois-
5
the container grower, who seeks to avoid wasting seeds by ture content. Moisture needs to be monitored during storage
double-sowing or having empty growing space in single to be sure it is being properly maintained. If there are no
sowing. However, for some species, 60% germination might changes in moisture content, then seed weight and purity
well be typical and expecting 98% germination is not rea- will not change and viability will change very slowly, if at
sonable. all. Some annual monitoring of seed moisture is necessary
to ensure that storage conditions are being adequately main-
Test Limitations and Variation tained. The viability can be retested at 3- to 5-year intervals
with a current test always done no more than 6 to 9 months
Standardized laboratory tests are designed to give maxi-
before sowing. If a longer time passes before sowing, some
mum values with minimum variation, allowing the results to
deterioration could occur, resulting in changes in germina-
be repeated. Without repeatability, there would be no stan-
tion. Determining viability in some seeds takes a long time,
dard by which to compare different seedlots. Assessments
and thus it is important to schedule adequate lead time into
would become even more difficult if seed tests were con-
the production schedule.
ducted in random fashion to mimic field conditions.
Although the correlation between laboratory and field ger-
Commercial Trade of Tree, Shrub, and
minations is frequently low, experience has shown that seed-
Native Plant Seeds
lots with higher germination scores will, over time, give
more germination in the field. Further improvement in pre- Official rules for testing seeds developed by the
dicting field performance may result from improvements in International Seed Testing Association or the Association of
the techniques of vigor testing. Official Seed Analysts are very important to the orderly buy-
Because seeds are biological, they usually are quite vari- ing and selling of seeds. These rules give standardized pro-
able in size and performance. Natural things simply show cedures that can be repeated with acceptable variation no
more variation than is usually seen in manufactured items. matter what laboratory conducts the test. In addition, the
Test results, therefore, can vary more than perhaps is expect- seed testing associations conduct comparative tests among
ed and still be accurate. For example, a seed weight of themselves to verify that the procedures are being applied
16,000 seeds/kg (7,300 seeds/lb) might be reported as uniformly and within tolerable limits of error. Such a system
15,600 (7,100) in a second test, and both results are in fact is important: it assures sellers that they are offering good
accurate. Tolerance tables can guide decisions on whether seedlots for sale, and it reassures buyers by giving them reli-
test results are comparable. Some of these tables are suitable
able information on which to base their purchases. There are
for all types of seeds, whereas others vary depending on the
also consumer protection seed laws in some countries or
type of test and the kind of seeds tested.
states requiring that seeds offered for sale meet certain mini-
Scheduling Seed Tests mum standards or be accurately labeled as to their quality.
Without a repeatable system of testing procedures, such
When should a seed test be conducted? Tests are needed laws would be impossible to enforce. Both the consumer
to formally determine the quality of the seedlot upon com-
and the reputable seed dealer would suffer. Uniformity of
pletion of the conditioning process. All variables then need
the rules combined with uniformity in applying the rules
equals order in the marketplace.

114 • Woody Plant Seed Manual

 References 5

Anderson HW, Wilson BC. 1966. Improved stratification procedures for Hart JR, Golumbic C. 1962. A comparison of basic methods for moisture
western white pine seed. Pub 8. Olympia: Washington State Department determination in seeds. Proceedings of the International Seed Testing
of Natural Resources. Association 27: 907–919.
AOSA [Association of Official Seed Analysts]. 1979. X-ray handbook. Hart JR, Golumbic C. 1966. The use of electronic moisture meters for
Lincoln, NE: AOSA. determining the moisture content of seeds. Proceedings of the
AOSA. 1983. Seed vigor testing handbook. In: Contrib. 32. Handbook on International Seed Testing Association 31: 201–212.
seed testing. Lincoln, NE: AOSA. 88 p. Heit CE. 1955. The excised embryo method for testing germination quality
AOSA. 1996. Rules for testing seeds. Journal of Seed Technology 16(3): of dormant seed. Proceedings of the International Seed Testing
1–113. Association 27: 907–919.
AOSA. 1995. Uniform classification of weed and crop seeds. In: Larson AL, Hoff RJ, Steinhoff RJ. 1986. Cutting stratified seed of western white pine
Wiersema JH, Handwerker T, eds. Contrib. 25. Handbook on seed test- (Pinus monticola Dougl. ex D. Don) to determine viability or to increase
ing. Lincoln, NE: AOSA. [Link] Planters’ Notes 37(1): 25–26.
AOSA. [Link] testing handbook. Contrib. 29. Handbook on seed ISTA [International Seed Testing Association]. 1996. International Rules for
testing. Lincoln, NE: AOSA. 302 p. Seed Testing, 1996. Seed Science and Technology 21(Suppl.): 1–288.
Belcher EW. 1975. Influence of substrate moisture level on the germination ISTA. 1995. Handbook of vigour test methods. 117 p.
of seed of selected Pinus species. Seed Science and Technology 3(3/4): Jones L. 1960. Rapid moisture determination of tree seed with an
597–604. electronic [Link] Planters‘ Notes 43: 7.
Belcher EW. 1978. Aspects of seed quality. In: Proceedings, Western Forest Justice OL, Bass LN. 1978. Principles and practices of seed storage. Agric.
Nursery Council and Intermountain Nurseryman’s Association Handbk. 506. Washington, DC: USDA Agricultural Research Service: 26.
Combined Nurseryman’s Conference and Seed Processing Workshop. Kamra SK. 1963. Studies on a suitable contrast agent for the x-ray radiog-
1978 October; Eureka, CA. D.54–D.59. raphy of Norway spruce seed. Proceedings of the International Seed
Belcher EW. [Link] effect of seed condition and length of stratification on Testing Association 28: 197–201.
the germination of loblolly pine [Link] Planters’ Notes 46(4): Karrfalt RP. 1987. Measuring tree seed moisture content now and in the
138–142. future. In: Landis TD, eds. [Link]. Rep. RM-151. Meeting the Challenge
Belcher EW, Karrfalt RP. 1979. Improved methods for testing viability of of the Nineties. Proceedings, Intermountain Forest Nursery Association;
Russian olive seed. Journal of Seed Technology 4(1): 57–64. 1987 August 10–14; Oklahoma City, OK. Fort Collins, CO: USDA Forest
Bonner FT. 1972. Measurement of moisture content in seeds of some Service, Rocky Mountain Forest and Range Experiment Station.
North American hardwoods. Proceedings of the International Seed Landis TD, Karrfalt RP. 1987. Improving seed-use efficiency and seedling
Testing Association 37(3): 975–983. quality through the use of history [Link] Planters‘ Notes 38(3):
Bonner FT. 1974. Determining seed moisture in Quercus. Seed Science and 9–15.
Technology 2(3): 399–405. Lanquist KB. 1965. Calibration charts for Radson No. 200 moisture meter.
Bonner FT. 1981. Measurement and management of tree seed moisture. Tree Planters‘ Notes 73: 11–12.
Pap. SO-177. New Orleans: USDA Forest Service, Southern Forest Moore RP. 1976. Tetrazolium seed testing developments in North America.
Experiment Sation. 10 p. Journal of Seed Technology 1(1): 17–30.
Bonner FT. 1984. Tolerance limits in measurement of tree seed moisture. Moreno R. 1985. Maximum germinants/units of seed as a tool for calculat-
Seed Science and Technology 12(3): 789–794. ing bareroot and container sowing calculations. In: Proceedings, Western
Bonner FT. 1992. Moisture content chapter In: Gosling PG. Report of the Forest Nursery Council/Intermountain Nurseryman’s Association; 1984
Forest Tree and Shrub Seed Committee 1989–1992. Seed Science and August 14–16; Coeur d’Alene, ID: 105–108.
Technology 20(Suppl. 1): 83–84. Piotto B. 1994. Effects of temperature on germination of stratified seeds of
Bonner FT. 1998. Testing tree seeds for vigor: a review. Seed Technology three ash species. Seed Science and Technology 22(3): 519–529.
20(1): 5–17. Simak M. 1957. The x-ray contrast method for seed testing: Scots pine—
Bonner FT, Agmata-Paliwal A. 1992. Rapid tests of seed quality in Picea Pinus sylvestris. Meddelanden fran Statens Skogsforskningsinstitut 47(4):
species by the leachate conductivity method. In: DeHayes DH, Hawley 1–22.
GJ, eds. Genetics in forest biology; Proceedings, 1st Northern Forest Simak M, Bergsten U, Henriksson G. 1989. Evaluation of ungerminated
Genetics Association Conference; 1991 July 23–25; Burlington,VT. Berea, seeds at the end of germination test by radiography. Seed Science and
KY: Northern Forest Genetics Association: 69–75. Technology 17(2): 361–369.
Buszewicz G. 1962. A comparison of methods of moisture determination Stein WI. 1967. Laboratory seed tests—are they doing the job? In:
for tree seeds. Proceedings of the International Seed Testing Association Proceedings, 1967 Annual Meeting of Western Reforestation
27: 952–961. Coordinating Committee. Portland, OR.
Ching TM, Parker MC. 1958. Hydrogen peroxide for rapid viability tests of Vivrette N. 1995. Personal communication. Carpinteria, CA: Ramson Seed
some coniferous tree seed. Forest Science 4: 128–134. Laboratory.
Czabator FJ. 1962. Germination value: an index combining speed and com- Vozzo J. 1979. Radiopaque agents for seed research with Juglans nigra. In:
pleteness of pine seed germination. Forest Science 8: 386–396. Frank Bonner, ed. Proceedings, Symposium on Flowering and Seed
Flemion F. 1948. Reliability of the excised embryo method as a rapid test Development in Trees; 1978 May; Mississippi State University, MS.
for determining the germinative capacity of dormant seeds. Starkville, MS: USDA Forest Service, Southern Forest Experiment
Contributions of the Boyce Thompson Institute 15: 229–241. Station and others: 172–280.

Chapter 5: Seed Testing • 115

5

116 • Woody Plant Seed Manual