university of copenhagen

Københavns Universitet

Guide to handling of tropical and subtropical forest seed

Schmidt, Lars Holger

Publication date:
2000

Document Version
Publisher's PDF, also known as Version of record

Citation for published version (APA):

Schmidt, L. H. (2000). Guide to handling of tropical and subtropical forest seed. Danida Forest Seed Centre.

Download date: 10. Dec. 2015

 SEED TESTING 11

SEED TESTING
Contents 11.1 Introduction 1

11. 2 Terminology 2

11.3 Timing Seed Testing 3

11.4 Sampling 4
11.4.1 Drawing samples 5
11.4.2 Reduction of sample size for testing 8

11.5 Simple Seed Testing 10

11.6 Standard Seed Testing Parameters 11

11.6.1 Seed weight 11
11.6.2 Purity 13
11.6.3 Moisture content 15
11.6.4 Viability and germination 16

11.7 Other Seed Testing 23

11.7.1 Vigour test 23
11.7.2 Seed health testing 27

REFERENCES 28

Appendix A11.1 Procedure for tetrazolium testing 30

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 Other Chapters of the book Guide to Handling of Tropical and Sub-Tropical Forest
Seed by Lars Schmidt soon available on www.dfsc.dk

Chapter 1: Introduction
Chapter 2: Seed Biology, Development and Ecology
Chapter 3: Planning and Preparation of Seed Collections
Chapter 4: Seed Collection
Chapter 5: Fruit and Seed Handling between Collection and Processing
Chapter 6: Seed Processing
Chapter 7: Phytosanitary Problems and Seed Treatment
Chapter 8: Seed Storage
Chapter 9: Dormancy and Pretreatment
Chapter 10: Germination and Seedling Establishment
Chapter 12: Genetic Implications of Seed Handling
Chapter 13: Microsymbiont Management
Chapter 14: Seed Documentation
Chapter 15: Trade and Transfer of Forest Seed
Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’
by Lars Schmidt, Danida Forest Seed Centre. 2000.

2 SEED TESTING
 SEED TESTING 11

SEED TESTING
11.1 Introduction Seed testing is used for control of quality parameters during seed
handling, and test results are submitted to customers as documenta-
tion on seed quality (chapter 14 and 15). Standard parameters such
as seed weight, purity, and germination or viability enter as factors in
the calculation of seed demand (section 3.4.2, table 3.5). Since seed
is sold on a weight basis, these seed quality parameters are essentially
also economic parameters. Moisture content in seeds is particularly
important during storage (chapter 8).

Quality tests may be carried out at intervals from harvest until the
seeds leave storage to be dispatched or sown in the nursery (Yue-Luan
1993). Most tests are simple non-standardized tests serving as practi-
cal guidelines during daily seed handling and nursery practice. Such
'simple tests' range from simple cutting tests which are little more
than observations, to moisture tests using calibrated field moisture
meters. In these cases a quick result is the more important and some
inaccuracy will be tolerated.

In order to make data of different seed lots and species comparable,

e.g. for marketing and research, tests must be standardized and rep-
licable. The widely adopted methods of standardized seed testing
follow the prescriptions of the International Seed Testing Association
(ISTA)1, which were first formulated in 1931 and now are revised
and updated every three years. The ISTA rules (ISTA 1996) contain,
in addition to standard seed testing procedures, specific guidelines
for a number of species. Specific guidelines on testing tropical and
subtropical forest species occur in a recently published ISTA hand-
book (ISTA 1998). In addition ISTA has published a number of more
elaborate handbooks on individual seed testing procedures (see list
of references).

Standard seed testing requires a fairly advanced seed laboratory, capa-

ble of conducting all routine tests according to the international rules.
1
Several American countries In particular germination tests require a high investment in equipment
follow the rules of Associa-
tion of Official Seed Analysts
for germination chambers with control of temperature, light and
(AOSA) which, however, differ moisture. Such investments are far beyond the possibility of most
only in minor aspects to the field stations to which this guidebook is mainly addressed. Therefore,
ISTA rules. advanced methods of seed testing will only be superficially described

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 1

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 in this book. It should, however, be stressed that although it may not
be practically possible to perform tests according to the international
standards, the ISTA prescriptions contain much information on e.g.
pre-treatment and germination conditions which are also useful under
less strict test procedures. Although advanced testing procedures are
omitted here, a certain basic knowledge is necessary to be able to
interpret results from the seed laboratory. Standard parameters such as
seed weight, purity and viability are therefore thoroughly explained.
A fairly detailed description of sampling is given since the principle
of sampling is very important for both simple and advanced testing.
In the case of standard testing, seed suppliers are normally the ones
to draw samples from seed lots and forwarding them for testing by
authorized seed testing laboratories.

For more detailed and elaborate information on how to carry out

seed testing, reference is made to above mentioned ISTA publica-
tions, publications from ASEAN Tree Seed Centre Project (ASEAN
1991, Yue-Luan 1993, Bhodthipuks et al. 1996), and Danida Forest
Seed Centre (Poulsen 1993 and 1994)

11. 2 Terminology Seed testing is an analysis of some physical parameters and the physi-
ological quality of a seed lot, based on a small representative sample.
The ‘quality’ (here strictly referring to physiological quality, in contrast
to the genetic quality) is the measure of potential performance of a
seed lot under optimal conditions. Seed testing includes a number
of parameters such as seed weight, purity, viability, germination and
moisture content, each with its own test procedure as will be defined
and outlined below. In the strict sense, testing always implies a standard
procedure, which may be subjected to statistical analysis. In common
terms, ‘testing’ applies to anything with the character of a more thor-
ough examination. In order to emphasize the distinction between the
strict and common understanding of the term ‘test’, this book distin-
guishes between ‘standard testing’ and ‘simple testing’, where the former
Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

refers to tests carried out according to ISTA or AOSA rules.

Seed testing is carried out on a sample that is a small representative

part of the seed lot. The terminology of the various units of sampling
is outlined in section 11.4. According to ISTA (1986) a seed lot is
defined as a ‘stated portion of the consignment assumed to be reason-
by Lars Schmidt, Danida Forest Seed Centre. 2000.

ably uniform’. What is ‘reasonably uniform’ is ultimately the judge-

ment of the seed handler. Normally seeds of the same provenance
and the same seed source, collected approximately the same date are
bulked before processing and hence typically make up a seed lot. It is
impractical to keep too many seed lots separate, and if a reasonable
level of uniformity exists, seed lots from different collections may
be bulked into one larger lot. Provenances should, however, always
be kept separate. It may also be reasonable to split up a large collec-
tion into smaller seed lots. This is typically the case for very large
quantities of seeds (ISTA rules set an upper limit for the size of seed
lots, typically 1000 kg, or 5000 kg for very large seeded species), or
if different parts of the seed lot are exposed to different conditions
likely to influence uniformity e.g. during processing or storage.

2 SEED TESTING
 11
A standard test has a certain design, which describes how a test is
carried out. It always has a number of replicates which are similar
tests carried out on the same number of seeds from the sample.

SEED TESTING
Five replicates mean that five identical tests are carried out with an
equal number of seeds. Replications allow calculation of statistical
parameters such as mean and variance, and minimize the chances
of an erroneous result.

11.3 Timing Seed Preliminary tests to assess maturity and seed quality at harvest are de-
Testing scribed in chapter 3. It may be relevant to carry out a test on moisture
content and possible fungal and insect infestation immediately after
harvest to guide preliminary handling before processing e.g. the need
for further drying, measures to prevent or reduce fungal damage. If fruits
or seeds are temporarily stored for a prolonged period before process-
ing, an interim test may indicate possible deterioration, and whether
continuous storage under the given conditions is safe. This test contains
the same elements as testing after harvest, i.e. moisture content and
infection rate.

If processing is carried out as one continuous procedure, testing

during processing is normally not applicable. However, purity tests
may be carried out at intervals during seed cleaning to suggest how
far cleaning should continue, and moisture content analyses may
determine the necessity for further drying (Yue-Luan 1993).

A more thorough test may be carried out between final processing and
storage. This test usually contains all the standard elements such as
purity, seed weight, moisture content and viability. As these data are
the ones normally to be submitted with the seed lot, the tests should
as far as possible be carried out by an authorized seed laboratory.

If the seed lot is stored for a prolonged period, or shorter period

under conditions where viability is likely to be impaired, viability
test(s) during storage may be applicable. Assessment of viability after
a certain time of storage may also help to predict the potential storage
life of the seed lot (cf. chapter 8). Where seed lots are dispatched to
customers, the result of the most recent viability test is obviously the
one to accompany the seed lot. Where seeds are stored under ambient
conditions, moisture-content analysis during storage is recommended
to detect possible increase in moisture content. A seed health test
may be carried out before storage to detect possible seed-borne pest
and pathogens. Vigour tests may be relevant for seed lots which show
high germination capacity under test conditions, but where some
deterioration is suspected e.g. after prolonged storage. A guideline
on timing of seed testing is summarized in table 11.1.

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 3

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 Table 11.1.
Seed testing during the
period of seed handling.

Harvest Processing Before storage During storage

Simple tests Maturity, health Health (cutting test)
(cutting test)
Moisture content
Moisture content (moisture meter)
(moisture meter)
Purity (screening)
Standard tests Seed weight Viability/germin-ability
Purity Moisture content
Moisture content (oven
method)
Viability/germination (TTZ,
X-ray, germination etc.)
Special tests Health test Health test
(Vigour) Vigour

11.4 Sampling Whether a sample is to be submitted for standard testing or is used

for more simple tests, it must comply with the basic rule of being
representative of the whole seed lot in any aspect to be tested. A
sample should thus have the same average seed size, moisture content,
viability etc. as the whole seed lot. The higher the degree to which a
sample is representative of the seed lot from which it was taken, the
better the test results can be valid for the whole lot. The importance of
proper sampling becomes more evident when expressed negatively: if
a sample is not representative, then the quality of the seed lot cannot
be concluded from the test results and the whole exercise is wasted.
A practical way of testing representativeness is to carry out test on
two individual samples: if sampling is ideal, the results of two indi-
vidual samples should give the same result with regard to all tested
aspects within the magnitude of statistical error. Thorough theoretical
Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

background and practical guidelines on sampling are found in ‘ISTA

handbook on seed sampling’ (ISTA 1986).

Sampling is subject to a number of practical difficulties because a

seed lot is never fully homogeneous. For example, seeds stored in
bags or containers tend to stratify themselves according to gravity and
by Lars Schmidt, Danida Forest Seed Centre. 2000.

any other physical features during handling (ISTA 1986). Further, the
environment of the immediate vicinity of the seed influences seed
characters. The external environment will have a higher influence
on seeds located at an outer or upper position of the seed lot than
those located in the interior. Hence, a sample taken from the top of
a container may contain seeds which are on average smaller, lighter,
drier or have different viability than the average seed. Also impuri-
ties tend to be stratified by the impact of mechanical handling. If a
seed lot contains a lot of inert matter, purity percentage taken from
the top and the bottom of containers or bags may be very different
(Peterson 1987). The characters of individual seeds vary in many as-
pects that may influence their being representative, e.g. size, maturity
and infestation. Therefore a sample must be large enough to cover

4 SEED TESTING
 11
the full variation within the seed lot. Yet the individual test rarely
comprises more than 5 replications of 25-100 seeds.

SEED TESTING
All techniques applied in sampling aim at obtaining samples which
are representative of the lot from which they were taken. Two main
pre-conditions help to assure homogeneity:

1. Variations of seed characters within the seed lot should be

as small as possible. Therefore, apparently different seed lots
should be kept separate and tested individually. Variations
typically occur between different provenances, growth sites
and degrees of maturity (cf. the concept of ‘seed lot’).
2. Homogeneity within the seed lot should be assured by
thorough mixing. For large seed lots sub-samples should
be taken from several locations within the seed lot. Where
a large seed lot is stored in several individual containers,
sub-samples are taken from each of these containers and
according to their relative size, if their size varies. However,
for seed lots stored in more than six containers, special
rules apply for sampling frequency; samples may be taken
from a smaller number of containers than the total, but
the containers from which the samples are taken should be
selected in an unbiased way.

11.4.1 Drawing In principle, there are two ways of drawing test samples: (1) by sub-
samples sequent divisions after mixing, and (2) by triers (fig. 11.2) taking out
samples from different parts of the seed lot and then mixing them
into a larger sample.

Mixing and division

Small seed lots of e.g few kilos can usually achieve a high degree of
uniformity by hand mixing. For larger lots, both hand and mechanical
mixing can be quite unreliable. Especially when mechanical mixing is
employed, both seed and impurities tend to distribute themselves in
the lot according to physical features such as size and weight. The risk
of stratification depends both on mixing procedure, seed and debris.
Stratification is, obviously, a higher risk for seeds with a high variation
in morphological characters. A seed lot of pines where only part of the
seeds have lost their wings will for instance typically be distributed with
the winged seeds on top and the de-winged seeds at the bottom.

Proper mixing is most easily assured manually: the seed lot is

poured onto the floor or other smooth surface. Seeds are mixed by
manually shovelling or raking from side to side. When the lot has
been manually mixed, it is divided into equal parts. Each part is put
into a container and the lots are then simultaneously poured into
a larger container. The procedure may be repeated once or twice
(Willan 1985). It is necessary to be at least two persons to carry out
the mixing, and the quantity is limited to what can be easily lifted
by the persons.

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 5

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 Figure 11.1.
Manual mixing of a
seed lot prior to sampling.

Larger seed lots are initially mixed as above and then shovelled into
a pile. The pile is then divided into four parts. One part is taken out
and mixed as above, or, if still too large to handle: mixed, piled and
divided until it achieves a reasonable size. The procedure of division
is employed because seed and debris tend to stratify while being piled
up. Relatively small lots or manually divided lots may be divided into
the required size by one to several successive divisions in a mechani-
cal divider. Several types are available. They are normally used for
subdividing submitted samples during seed testing (see below), but
can also be used for drawing samples from seed lots. The principle
of mechanical dividers appears from fig. 11.5. Some dividers separate
seed lots into halves. The size of the sample may then be further
reduced by several successive divisions. Other types of dividers are
capable of separating the seed into several equal sections in one
procedure. One part is then used as a sample for testing.

If seeds are stored in several containers and bags, yet belonging to

the same seed lot, each part of the seed lot should contribute with a
proportionally equal amount of seeds to the sample, i.e. a container
with 20 kg of seeds should contribute twice as much to the sample
as one containing 10 kg.

Triers

Thorough mixing of large seed lots prior to sampling is laborious and

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

time consuming. A less thorough mixing (although a certain level

of homogeneity should be achieved) may be allowed if sub-samples
are drawn from different positions in the seed lot. In practice these
sub-samples (in the ISTA terminology called ‘primary samples’) may
be taken directly from containers or bags by the help of ‘triers’. The
design of commonly used trier types is illustrated in fig. 11.2. A trier
by Lars Schmidt, Danida Forest Seed Centre. 2000.

consists of two tubes, one fitting outside the other like a sleeve. The
two tubes have equal size rectangular slots along their side to allow
seeds to flow in. The inner tube may be a continuous tube or trans-
versely separated into several compartments or pockets by partings
under each hole. The latter type is preferred since it gives a more
even sampling, especially where seeds are not freely flowing (e.g.
rough, angular or winged seeds). In the open tube type, seeds that
enter into the top slots may tend to fall through the tube into the
lower parts if these have not already been filled with rapidly inflow-
ing seeds (Edwards and Wang 1995). This error may be reduced by
inserting the trier into the container in a slightly slanting position
(ISTA 1986). Opening and closing the holes or pockets of the trier is
carried out by turning or sliding the inner tube with a handle. Where

6 SEED TESTING
 11
the inner tube has no partings between the pockets (i.e. a continuous
tube), the trier can be emptied through an upper opening. The end
of the trier is tapered or pointed to allow its easy insertion into the

SEED TESTING
seed lot. Several sizes of triers are available, the diameter and length
vary according to seed size.
Figure 11.2.
Various types of triers used for
sampling in large seed lots.

Sampling by the help of a trier follows five steps.

1. Close the holes of the trier by sliding or turning the outer

sleeve.
2. Insert the trier right down to the bottom of the bag, container
or pile of seeds while keeping the holes closed.
3. Open the holes by sliding or turning the outer sleeve; move
the trier slightly to assure that the seeds flow into the holes
freely and fill the trier.
4. Close the holes by turning into closed position and pull out
the trier.
5. Empty the trier into a pail. Trier types with no partings between
the holes can be emptied through the top holes. If the trier has
separate pockets, it must be emptied through the holes.

If the seeds flow freely into the holes or pockets, a representative sam-
ple will be taken from each level of the container. If the containers
or bags contain the same amount of seed, each sample will take out
the same percentage of the total. If the content of containers varies,
the sample size should be adjusted accordingly. With containers of
say 25 and 50 kg, one sample may be taken from the 25 kg container
every time two samples are taken from each 50 kg container. Each
sub-sample taken with the trier is put into a container for further
handling as described below.

Triers are quick and convenient tools for sampling large seed lots.
They can be used for all free-flowing seeds of small or intermediate
size. However, large, rough, winged, angular and other seed types
which are too big or do not easily enter into the trier must be sam-
pled by other methods, e.g. by removing portions at different levels
of the seed lot by hand, or by the above mentioned mixing and
dividing method.

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 7

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 11.4.2 Reduction of Several terms apply to the different types of samples during the
sample size sampling process (fig. 11.3): a primary sample is a small quantity
for testing of seed taken out from a single position in the seed lot, e.g. by the
help of a trier. When a number of primary samples taken from dif-
ferent parts of the lot are bulked, they make up a composite sample.
Usually this sample is several times larger than the sample actually
needed for testing. For official seed testing, ISTA (1995) has issued
prescriptions for the quantity of seeds needed by the seed laboratory
to carry out standard tests. The quantity, which varies with species,
is called the submitted sample. This sample is further reduced in the
laboratory to a working sample according to the quantity required
for the individual test (ISTA 1986, ISTA 1998).

Figure 11.3.
Procedure of sampling. Primary samples are drawn from the seed lot and mixed into a composite sample. The composite
sample is reduced to a submitted sample to be forwarded to the seed laboratory for testing. In the seed laboratory work-
ing samples are drawn for the individual test. The same working sample may be used for more than one test if the test is
not destructive e.g. purity test → germination or moisture content.
Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

Reduction of sample size, whether composite sample to submitted sam-

ple, or submitted sample to working sample, must be conducted under
observation of the same strict rules of maintaining representativeness
and being unbiased as in the previous steps of sampling. Because the
quantity of the composite sample is relatively small, it requires little effort
to maintain the samples in a homogeneous state. The methods of sample
by Lars Schmidt, Danida Forest Seed Centre. 2000.

reduction are similar to those employed for sampling from small seed
lots. In its simplest form, seeds are spread on a plane table in an even
layer. The sample can then be divided in halves once or several times
by a ruler or other straight tool (Peterson 1987) (fig. 11.4). Mechanical
dividers are convenient accessories for unbiased reduction of sampling
size (fig. 11.5). They can be used for most seed types, although very large,
rough or winged seeds may cause some problems. Such seeds are most
easily separated by the above mentioned ‘ruler parting’.

ISTA (1996 and 1998) provides rules for minimum working samples
for seed analysis. To carry out basic tests of purity, seed weight, mois-
ture content and viability/germination analysis roughly 2500-5000
seeds are needed, depending on seed size. However, for very small

8 SEED TESTING
 11
seeded species, a sample size of less than 1-5 g is impractical, although
it may contain much more seed in number than actually required.
For large seeded species, reduction of sample size to a minimum of

SEED TESTING
500 seeds is acceptable. ISTA (1995) proposes submitted samples
to the laboratory being twice the size of the total required working
samples. Examples of the weight of some submitted samples are
listed below.

Figure 11.4.
A seed sample can be
reduced to a working sample
by repeated halving (Robbins).

Figure 11.5.
Some mechanical dividers used for accurate and unbiased reduction of sample
size. The conical divider (left) divides the sample into two or more equal size parts
according to number of spouts at the bottom, the riffle type divider (right) divides
the sample into two parts of equal size.

Species Submitted sample Species Submitted sample

Acacia nilotica 1100 g Dryobalanops oblongifolia 24000 g
Acacia senegal 550 g Gliricidia sepium 835 g
Acacia tortilis 420 g Khaya nyasica 2500 g
Afzelia quarzensis 25 kg Khaya senegalensis 1600 g
Cedrela odorata 165 g Swietenia macrophylla 2400 g
Ceiba pentandra 500 g Tamarindus indica 3600 g
Dalbergia melanoxylon 840 g Ziziphus mauritiana 3500 g
Table 11.2.
Examples of weight of submitted samples for seed testing, given that each submitted sample
contains 2x minimum weight of working sample (2500 seeds) (from ISTA 1998).

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 9

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 11.5 Simple Seed Where seeds are to be sown locally and immediately after collection
Testing or processing, expensive testing rarely makes sense (Yue-Luan 1993).
Yet, information on seed quality is still very useful for nursery op-
eration. Further, regular simple tests often serve as a valid guideline
during seed handling. Simple tests do not follow the strict conditions
prescribed for standard testing. Therefore comparison of test results
becomes less valid. Yet simple tests often suffice when information is
only required for an individual seed lot, e.g. to determine the need for
further cleaning or drying, or to state the physiological quality. Simple
testing encompasses the same parameters as those used in standard
tests, which will be described more detailed in section 11.6.

Seed weight. There is no short cut as compared to the standard

method; number of seeds and their weight must be known to cal-
culate number of seeds per kg. An electric digit scale is preferred. If
such equipment is not available, weighing may be carried out on a
balance. There are two methods:

1. Put a known number of seeds on one side of the balance and

weights on the other until it balances.
2. Put a known weight on one side and add seeds on the other
until it balances.

Count the seeds and calculate seed weight as described in section

11.6.1.

Purity. We are interested in knowing the fraction of pure seed, not

the composition of other matter. Therefore we weigh a sample of
seeds with impurities, then separate the two fractions and weigh one
of them (which in practice should be the seeds as it is usually the
larger fraction and as such gives a more exact weight).

1. Pour the sample into one bowl of the balance against a known
weight e.g. 250 g. Remove all the entire pure seeds manually with
Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

tweezers. Throw the impurities out and weigh the seed fraction.

2. As above but sorting is done by spreading the sample on

a table cloth or the like. Impurities may also be removed
by blowing, sifting or letting the seeds roll down a slanting
cloth frame as described in chapter 6. However, care should
by Lars Schmidt, Danida Forest Seed Centre. 2000.

be taken during sorting not to lose ‘pure seeds’ because the

sample is small.

Moisture content. Quick measurement of moisture content may be

carried out with the aid of a moisture meter as described in appendix
A5.3. It is important that the moisture meter is initially calibrated by
several measurements on samples where moisture content has been
measured by the standard method (see below). Microwave drying is
a quick method of moisture measurement for large seeds for which
moisture meters are less applicable. The seeds are cut into small pieces
before drying in the micro-oven for 5-6 minutes (ISTA 1991, Bonner et
al. 1994). The sample is weighed before and after drying and moisture
content calculated as after the oven drying method (see below).

10 SEED TESTING
 11
Germination and viability. For discussion of the two terms, reference
is made to section 11.6.4. The simplest viability test is the cutting
test. Seeds are cut longitudinally through the embryo. For some

SEED TESTING
seeds soaking in water facilitates cutting. Seeds with firm, white or
light green, healthy looking embryo are deemed viable (Bonner et
al. 1994). Viability tested by the tetrazolium test (TZ) is described in
appendix A11.1. It can be carried out at ambient temperature and
does not require sophisticated laboratory equipment. Germination
trials in nursery plots give valuable information on germination un-
der field conditions and are often the only method for recalcitrant
seed. Because such trials are essentially strongly influenced by the
environment, comparison of different seed lots are normally subject
to large errors. Seedbeds or pots are prepared as during normal nurs-
ery operation. Where dormancy is known or suspected, appropriate
pretreatment should be carried out prior to sowing. It is advisable to
sow the seeds in a design that permits easy counting of germinants.
Figure 11.6.
Some equipment for simple
seed testing. A balance, B mag-
nifying glass, C secateurs, D
screen, E moisture meter,
F tweezers.

11.6 Standard As stated above, this book will not cover the practical methods of
Seed Testing standard seed testing, but rather explain the background of informa-
Parameters tion provided by seed testing. Detailed information on methods of
analysis, prescriptions, definitions etc. should be sought in specific
seed testing handbooks, in particular the newest version of ISTA
international rules for seed testing (ISTA 1996 or later). Reference is
also made to ISTA 1998 and several papers from the ASEAN Forest
Tree Seed Centre Project (Yue-Luan 1993, Bhodthipuks et al. 1996,
and ASEAN 1991). Generally, standard seed testing is carried out
whenever seeds are to be stored for a prolonged period, or whenever
seeds are to be traded by authorized seed suppliers.

11.6.1 Seed weight Seed orders are given by weight, seedlings planted by numbers. Seed
weight is therefore, together with purity and germination percent-
age, important when calculating seed demand for a given planting
programme (table 3.5). Further, seed size may be correlated with

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 11

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 vigour and hence be an indirect measure of potential performance
(see further below).

There are two ways of indicating seed weight: either in number of

seeds per kg (or for small seeds occasionally per 100 grams), or in
weight in grams for 1,000 seeds. Because it is not always clear whether
the former figure refers to a pure sample, seed testing always indicate
the 1000 (pure) seed weight (tpsw). The figure can easily be trans-
formed to number of pure seeds per kg.

Examples:

a. 1,000-seed weight of Eucalyptus camaldulensis is 1.5g. Number

of seeds per kg is: 1000seeds/1.5g x 1000 g = 666,000 seeds.

b. Pinus caribaea contains 3500 pure seeds per kg. 1,000-seed

weight is: 1000g/3.5 = 285 g.

Seed weight is usually calculated on replications of samples of 100

seeds. For very large seeds, calculation is conveniently based on a
smaller number. In official seed analysis, variance analyses are carried
out based on several replicates of 100 seeds. The figure expresses the
variation in seed weight within the sample. Seed weight analysis uses
the same criteria as purity test for what may be included as ‘seed’ in
the calculation (see below).

Seed weight varies both with seed size and density, and there are
reasons to be observant on factors that may influence these, especially
when comparing figures. For example, the term ‘seed’ is in some cases
subject to some confusion. Clear indications of what is the tested unit
are sometimes necessary e.g. with or without wings, arils or pericarp.
In Swietenia macrophylla seed weight of winged seeds is about 2100
seeds per kg, while for de-winged seeds it is about 2300 seeds per kg.
Moisture content influences density and thereby seed weight. Seed
Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

weight of de-winged S. macrophylla roughly increases from above 2400

per kg at 5% m.c. to 2300 seeds per kg at 9% m.c.

In the species list of multipurpose trees and shrubs, Carlowitz (1991)

indicates variation in seed weight of several species, in some as large
as 5-10 times. Although some variation may be ascribed to different
by Lars Schmidt, Danida Forest Seed Centre. 2000.

units used (winged/de-winged, seed, pyrene etc.) and possibly mois-

ture content, seed weight is known to be one of the very variable seed
characters within species. It is influenced by genetic, developmental
and environmental factors (chapter 2).

Relatively high seed weight is often desirable since it is often cor-

related with rapid germination and good seedling establishment
(Griffin 1972, Sorensen and Campbell 1993). It should, however, be
noted that direct comparison between provenances is rarely valid,
since variation caused by genetic differences may overshadow pos-
sible vigour differences.

12 SEED TESTING
 11
11.6.2 Purity In common terms purity is an expression of how ‘clean’ the seed lot
is. A purity analysis at a certain stage of processing may serve as a
guideline for the necessity of further cleaning. However, as explained

SEED TESTING
in chapter 6, it is in practice never possible to achieve a completely
clean or pure seed lot by mechanical processing, because the physical
characters of other seed or inert matter may be so similar to those of
the seeds in question that separation is impossible. Yet information
on the actual composition of the seed lot is important for the seed
handler, both in relation to the factors of seed price and seed demand
as mentioned above, and because any type of impurity may hamper
practical sowing. Certain types of impurity may also harbour infec-
tive fungi, which in turn could hamper seed quality, especially when
associated with high moisture content (chapter 6 and 8).

Purity of a seed lot indicates in percentage how large a fraction is

made up of pure seeds of the species in question, and how much is
made up of inert matter and other seeds. Impurities may be any non-
seed material (leaf, flower, fruit fractions, soil etc.), small fractions
of seeds of the actual species, as well as seeds of other species. ISTA
(1996) specifies the pure seed fraction to contain:

1. Intact seeds of the actual species as well as dead, shrivelled,

diseased, immature and pre-germinated seeds.
2. Achenes and similar fruits e.g. samaras, with or without
perianth and regardless of whether they contain a true seed,
unless it is apparent that no true seed is contained.
3. Fractions of broken seeds, achenes etc. which are more than
half of the original size. However, seeds of e.g. legumes and
pines that have the entire seed-coat removed are regarded as
inert matter.

During purity analysis each ‘pure seed’ fraction (1-3) is separated from
the working sample. Purity is expressed as the weight percentage of
pure seed fraction over the total weight of the working sample:

Purity = Weight of pure seed (g) x 100

%
Total weight of working sample (g)

Since ‘pure seed’ may include both dead and empty seed, plus dam-
aged seed, purity does not tell anything about viability. Forest seeds
are often collected by hand either by harvesting directly from the
tree or, for large seeds, by picking from the ground. In both methods
the risk of contamination with foreign seed is small. Large seeds are
generally easy to clean and purity analyses are therefore often omit-
ted (Yue-Luan 1993). However, for smaller seeds contamination with
other seeds may occur e.g. during processing. Therefore impurities
are normally separated into ‘inert matter’ and ‘other seed’. Each
component is reported as a percentage of total weight (ISTA 1998).

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 13

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 Table 11.3.
Weight Percentage
Example of fractions
indicated in a purity Working sample 60 g 100 %
test for Pinus merkusii. Pure seed 54 g 90 %
Other seed 1g 1.7 %
Inert matter 5g 8.3 %

Figure 11.7.
Examples of ‘pure seed’ defini-
tions according to ISTA (1991).
Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’
by Lars Schmidt, Danida Forest Seed Centre. 2000.

14 SEED TESTING
 11
11.6.3 Moisture Moisture content (m.c.) is crucial in connection with storage and
content longevity. Since moisture content of seeds tends to vary with at-
mospheric humidity (chapter 5 and 6), it is important that exposure

SEED TESTING
to varying humidity is minimized before testing. Therefore, seeds
should be packed in waterproof material as quickly as possible after
sampling. In order to avoid the possibility of water condensing on
the seed when removed from cold store, seed should be allowed to
reach ambient temperature before the container is opened.

Under laboratory testing, seed moisture is measured by the oven-

drying method, which is the direct method, prescribed by ISTA
and described below. This method can also be used for calibrating
moisture meters for indirect measurement of moisture content (ap-
pendix A5.3). The indirect methods provide very quick results, which
can be used as a guide during seed handling, e.g. to determine the
necessity for further drying.

Moisture content of a sample is the loss of weight when it is dried in

accordance with the prescribed rules. It is expressed as a percentage
of the weight of the original sample (ISTA 1996). This is the fresh
weight basis. Moisture content measurement contains the following
components:

1. Container (heat resistant) including cover is weighed (M1).

2. Seeds are ground or cut into smaller fractions before drying
to assure that moisture can escape from the interior.
3. The seeds are placed in the container and weighed together
with the container (M2).
4. Seeds are placed in an oven at 103 +/- 3°C for 17 +/- 1
hour.
5. After drying, the seeds are placed in a desiccation chamber
while cooling (to avoid reabsorbtion of moisture from the
atmosphere).
6. After cooling, the seeds plus container are weighed again
(M3).

The moisture content (fresh weight basis) is calculated: m.c. =

(M2-M3) x 100 %
(M2-M1)

Seeds typically contain far less moisture in the seed-coat and possibly
pericarp than in the embryo and endosperm. Hence, processing may
influence moisture content both directly in terms of drying rate, and
indirectly in connection with extraction and possible de-winging. For
example, moisture content of a seed sample of Swietenia macrophylla
would normally be smaller if the seeds are not dewinged before treat-
ment, since in that case the entire dry wing contributes to the seed
weight, yet little to the total moisture content.

The method anticipates that the total loss of weight is caused by

evaporation of water. In practice other volatile compounds such as

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 15

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 oil and resin are also lost during drying which, in seeds rich in these
compounds, contribute to an overestimation of the moisture content.
Despite this potential source of error, the oven-drying method is still
used as a standard for these seeds, but the seed handler should be
aware of the likely overestimation of the real moisture content when
testing oil or resin rich seeds (Poulsen 1994).

The above methods all refer to calculation of moisture content on

fresh weight basis. Moisture content expressed as loss of moisture in
percentage of dry weight (dry weight basis) is little used but it still
occasionally appears in the literature. A conversion scale is shown in
fig. 11.8. It should be noted that percentages above 100 may occur
when calculated on dry weight basis while such figures are obviously
impossible when calculated on fresh weight basis.
Figure 11.8.
Approximate conversion scale
of moisture content calculated
on dry weight converted to
moisture content on fresh
weight basis (Willan 1985).

11.6.4 Viability and A high germination percentage is obviously desirable for the nursery-
germination man, anything other than pure germinable seed is waste. Therefore a
germination or viability test should indicate the potential germinabil-
ity which, with proper handling, should reflect expected germination
in the nursery. Germination potential is most directly determined
in a germination test: under the appropriate conditions everything
that can germinate should germinate. Germination tests are widely
used in both standard seed testing and more informal simple nurs-
ery tests. However, the tests have several limitations, some of which
may either over-estimate or under-estimate the actual germination
Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

potential of a seed lot. Three situations where germination tests are

less applicable are the following:

• Where seeds have a very short viability. Duration of a

germination test is typically 3-5 weeks. For short-lived
recalcitrant seed significant loss of viability may take place
by Lars Schmidt, Danida Forest Seed Centre. 2000.

during the test period. Hence, the germination percentage

obtained by the test may not be valid for the seed lot from
which it was taken because the viability of the seed lot has
declined during the test period.
• Where germination is delayed or suppressed by deep
dormancy. If pretreatment has been insufficient to overcome
dormancy, germination may be low even if seeds are viable.
• Where fast test results are required. Especially for slow
germinating species (some species take several months
to germinate) the duration of a germination test may be
inconvenient. Where a seed lot is to be dispatched soon after
collection, there is often not enough time for a germination
test.

16 SEED TESTING
 11
Where germination tests for some reasons are inconvenient or un-
reliable, or where shortage of standard germination facilities limits
the use of germination tests, germination potential may be tested by

SEED TESTING
indirect methods viz. viability tests. These tests do not prove that
seeds are germinable, only that they are (most likely) alive. It is es-
sential to distinguish between the two terms ‘viability’ (percentage)
and ‘germination’ (percentage); the two terms refer to different types
of test result, and they are not synonymous. Although prescribed test
procedures aim at creating concurrence between the two types of
tests, some divergence may occur. Seeds deemed viable may not be
germinable because of an advanced stage of deterioration (reduced
vigour) or dead tissue in vital parts of the embryo. Another example
is tetrazolium staining which indicate live tissue (see below). Young
(immature) seed may stain normally by this procedure although they
have not achieved germinability. On the other hand, viability tests
are not always inferior to germination tests. In the three situations
listed above a viability test is preferred. Viability tests are also used as
a supplement to germination tests in order to examine the character or
quality of seeds that have not germinated during the standard test.

Several types of viability tests are available. The most common ones
are cutting test, tetrazolium, X-ray, excised embryo, and hydrogen
peroxide test, which are described below. Of these methods only
tetrazolium, hydrogen peroxide and excised embryo tests actually
prove a life manifestation, in the first case as the activity of a meta-
bolic enzyme complex, in the latter as a directly observable embryo
development. It should be emphasized that all types of viability test
are subject to some subjectivity in the interpretation of results. Vi-
ability tests are generally less applicable to very small seeds such as
eucalypts, and for exised embryos, the method is practically impos-
sible (Boland et al. 1980).

Cutting test

Cutting tests are never used as the sole viability test in standard testing,
but rather to examine the conditions of non-germinated seeds in a
germination test. The method is, however, widely used in simple seed
testing, both during collection and processing. It is described in section
11.5 in connection with simple tests. In a cutting test, viable and dead
seeds are distinguished visually, which in practice means that seeds that
are empty, insect-damaged, under-developed or showing other distinct
signs of damage are deemed non-viable, and the remaining portion
viable (although actual life manifestations are not proven).

Tetrazolium

The tetrazolium (TZ) test is the most widely adopted biochemical

method to examine seed viability. The method is also called topo-
graphical tetrazolium test (TTZ) to emphasize that specific areas of
the seed are examined rather than just general evidence of viability
(Enescu 1991). A thorough description of the theory and practice of
tetrazolium testing is given by Moore (1985), Yu and Wang (1996),
and Enescu (1991). The practical method is summarized in appendix

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 17

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 A11.1. TZ test is especially useful as an alternative to germination
test for species that require long periods of pretreatment to overcome
dormancy (e.g. several temperate species), but the test is also widely
used as a quick test for species with less complex dormancy. Although
the method is in principle applicable to all seed types, interpretation
of results becomes extremely difficult for very small seeds.

The principle of TZ test is as follows: dehydrogenases are a group of

metabolic enzymes in living cells. During the reduction processes in the
metabolicly active cells dehydrogenases release hydrogen. The hydrogen
is able to reduce an applied pale yellow solution of 2,3,5-triphenyl tetrazo-
lium chloride or bromide (TZ) to a stable, bright red triphenylformazan.
Hence, the formation of red formazan is an indication of dehydrogenase
activity, which is in turn an indication of viability. Because staining of
tissue is local, it is possible to distinguish living (red-coloured) and dead
(colourless) parts of the seed. Where dead (necrotic) tissue occurs only
superficially in cotyledons, while the radicle stains normally, the seeds
may still be viable. On the other hand, even small patches of necrotic
tissue in the vital part of the embryo normally means that the seed would
not be able to germinate. The exact evaluation of these partly stained
seeds requires a fair amount of experience. Moore (1985), Yu and Wang
(1996) and Enescu (1991) contain examples of TZ staining pictures of
a number of seeds, which may serve as guidelines for interpretation of
the results of TZ staining.

Seed embryos are likely to stain whether they are dormant or not, and
damaged but not necrotic tissue may stain normally. Therefore the
result of the TZ test is likely to include the three classes in the germi-
nation test: normal seedlings, abnormal seedlings, and live but not
germinated seeds (including hard seeds) (ISTA 1998). Yu and Wang
(1996) found good concordance between TZ test and germination
in comparative studies of different viability tests on several tropical
tree species. A precondition for application of the TZ test is that the
seeds are mature i.e. physiologically germinable. Immature seeds may
Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

stain normally because they contain live cells, but would give poor
results in a germination test. Another source of error is that seeds
infected by fungi may stain because of the metabolic activity of the
fungi and not the plant cells. However, such fungal cells generally
stain dark brownish-red, not bright red as live sound plant cells do
(Mittal 1997, pers. comm.).
by Lars Schmidt, Danida Forest Seed Centre. 2000.

X-radiography

X-radiography is a quick test to differentiate empty, under-developed,

insect or physically damaged seeds from morphologically intact and
healthy seeds by the aid of X-rays (ISTA 1996). A thorough descrip-
tion of principle and practice in X-radiography in tropical tree seeds
is found in Simak (1991) and Saelim et al. (1996). X-rays are electro-
magnetic waves with wavelengths of 0.05-100Å2 (visible light approx.
2
Wavelengths of light are usu- 4000 - 8000 Å). The seeds are placed between the X-ray source and a
ally indicated in Å (angstrom) photosensitive film or paper. When the seeds are exposed to X-rays
or nm (nanometre). 1 Å = of low energy (longer wavelength, approx. 1 nanometre), an image
0.1nm, or 1 nm = 10Å. (radiograph) is created on the film/paper. Photographic processing

18 SEED TESTING
 11
converts the radiograph into a visible picture. Since X-rays are non-
destructive, seeds examined by the X-radiographic method may also
be used in direct germination tests.

SEED TESTING
Because normal X-radiography is unable to distinguish aged or physi-
ologically damaged seeds from sound seeds, the best correlation of the
X-ray and germination test is found where empty seeds, insect damage
or other physical damage contribute largely to possibly reduced seed
quality. The method is especially useful where such damage or lack of
development are not apparent on the exterior of the seed, e.g.:

1. Empty seeds in pines, eucalypts and others, where the seed

develops into full size even if it contains no embryo.
2. Insect infested seeds where no entry hole is visible, e.g.
legume seeds infested by bruchids or conifers or eucalypts
infested with chalcids (e.g. Megastigmus spp.).
3. Seeds enclosed by a hard fruit structure, e.g. drupes or
samaras, where the pericarp or endocarp bears no sign of
presence or condition of the enclosed seed(s). X-radiography
may reveal both number of seeds in such fruits and their
condition.
4. Seeds where internal mechanical damage to the embryo may
have occurred e.g. during processing.
5. Seeds with shrunken or underdeveloped embryos e.g.
immature seeds.

X-radiography is especially useful for estimating viability of recalcitrant

seeds because they are short-lived and their germination potential has
to be determined quickly (Saelim et al. 1996). Chaichanasuwat et al.
(1990) found good conformity between X-radiography and germina-
tion test for Peltophorum pterocarpum. In comparative studies of differ-
ent viability tests, Bhodthipuks et al. (1996) found that X-radiography
generally over-estimated viability as compared to germination tests.
Laedem et al. (1995) found good correlation between X-radiography
and germination test in Dalbergia cochinchinensis and Pinus kesiya, both
of which had a high seed quality, while there was low correlation
for Pinus merkusii in which the germination percentage was low. It
may from these observations be generalized that the method is less
applicable to seed lots of low physiological quality.

Figure 11.9.
X-radiographs used for seed
quality analysis.
A) Pinus kesiya , the picture
reveals some seeds with rudi-
mentary embryos and some
empty seeds.
B) Albizia procera , the picture
reveals seeds infected with
bruchid beetle.
From Saelim et al. 1996.

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 19

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 Application of specific contrast chemicals e.g. BaCl2, AgNO3, NaI,
or KBr to the seed before X-ray enhances the possibility of evaluating
viability of tissue. Because these chemicals stain differently in live and
dead tissue, the X-ray contrast (XC) method gives a different image
of live and dead seed (or seed tissue) similar to the TZ test (Saelim et
al. 1996, Simak 1990). However, since X-radiographs are black and
white, interpretation of the results requires even more experience
than the TZ test.

Excised embryo test

This method is used where seeds germinate very slowly or where the
seeds are deeply dormant and require long pretreatment. It may also
be used where the nature of dormancy and hence pretreatment is
not known. The principle of the test is that the embryo is manually
excised from the seed-coat and possible endosperm under aseptic
conditions, placed on filter or blotting paper and incubated in ger-
mination cabinets at 20-25°C. The result of the excised embryo test is
germination percentage under incubation. It should be noticed that
in order to be statistically valid, also seeds with damaged, deformed,
discoloured or lacking embryos must be included in the final calcula-
tion (ISTA 1996). Because the embryo is surgically excised from the
seed during the operation, it requires a certain minimum size of seed
and embryo before it is practically possible. The method is thus not
applicable to very small seeds.

The excised embryo test is a transition form to a true germination

test, since the embryos are evaluated on radicle development that
is essentially an early germination event. However, the germination
process is concluded before the seeds develop into seedlings that
could be evaluated for normal growth, as is done during normal
germination tests.

Hydrogen peroxide test

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

This test is another viability test, which forms a transition to a ger-

mination test. It is described in detail in Bhodthipuks et al. (1996).
During the test, initial germination is evaluated after application of
the chemical hydrogen peroxide (H2O2). The purpose of applying
H2O2 is to increase the supply of oxygen to speed up initiation of
by Lars Schmidt, Danida Forest Seed Centre. 2000.

germination. The application is sometimes used as a pretreatment

(chapter 9). The H2O2 test method is illustrated in fig. 11.10. Seeds
to be tested are initially soaked in a 1% solution of the chemical for
8-12 hours. They then have a small piece of their seed-coat removed
at the radicle end and are incubated for a period of about 7 days.
The solution is changed after about 3 days. Incubation is conducted
under dark conditions as the chemical is very light sensitive. Seeds
are considered viable when radicles emerge from the cut end.

20 SEED TESTING
 11
Germination

During germination tests, seed quality is measured directly as the abil-

SEED TESTING
ity of the seed to germinate under optimal germination conditions of
temperature, moisture and light. It is anticipated that germination is
not impeded or delayed by possible dormancy. Therefore seeds should
be pretreated before a germination test. Germination under the ISTA
standard test is subject to strict prescriptions to pretreatment methods
and germination conditions (ISTA 1996). Germination is normally
carried out in germination cabinets under controlled environment.
The conditions prescribed by ISTA include the following variables:

• temperature (level and regime, e.g. constant day and night or

fluctuating)
• light (+/- light or period of day/night cycles)
• substrate (sand (S), top of sand (TS), top of paper (TP),
between paper (BT) and pleated paper (PP) (parentheses
correspond to abbreviations used in ISTA prescriptions)).

Figure 11.10.
Procedure for hydrogen perox-
ide viability test. From Laedem
1984.

The ISTA rules also indicate days of first and last count in order to
standardize the duration of the test period. Although these rules are
standardized for laboratory tests, they may also serve as guidelines
for more informal nursery tests. Germination is carried out on the
‘pure seed’ fraction (see section 11.6.2), which on one hand excludes
all ‘other seeds’, on the other hand includes large, damaged seeds.

Germination is defined as ‘the emergence and development of the

seedling to a stage where the aspects of its essential structures indicate

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 21

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 whether or not it is able to develop further into a plant under favour-
able conditions in the soil’ (ISTA 1996). This means for tree seeds
a root system, shoot axis, cotyledons, and terminal bud. The exact
criteria of evaluation vary slightly between species, e.g. in eucalypts a
seed is considered to have germinated when the radicle has developed
normally and the cotyledons have emerged from the seedcoat and
have unfolded (Boland et al. 1980).

Germinated seeds are counted regularly during the prescribed ger-

mination period from the indicated ‘first count’ to ‘final count’.
Counting once per week is usually sufficient, but species with rapid
germination may be counted and removed every two days. Removal
of germinants is done in order to facilitate subsequent countings
and to avoid possible fungal spread. Both ‘normal’ and ‘abnormal’
germinants are counted, registered and removed during the period.
At the end of the period all ungerminated seeds are examined. The
final test result is grouped into the following classes:

1. Normal germinants. The cumulative number of seeds

which have developed into seedlings of normal and healthy
appearance with all essential structures of a seedling. This
also includes seedlings where possible damage is caused by
secondary infection.
2. Abnormal germinants. The cumulative number of seeds
which have germinated during the test period but in which
the seedlings show abnormal or unhealthy appearance e.g.
lacking essential structures such as cotyledons, or being
discoloured or infected by seed-borne pathogens (primary
infection).
3. Ungerminated seeds. Seeds which have not germinated
by the end of the test period. These are grouped into the
following sub-classes:

a. Hard seeds, which are seeds that remain hard because

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

they have not imbibed (normally because of insufficient

pretreatment).
b. Fresh seeds, which are seeds that have not germinated
although they appear firm and healthy.
c. Dead seeds, which are seeds that are soft, or showing other
signs of decomposition.
by Lars Schmidt, Danida Forest Seed Centre. 2000.

d. Other seeds, e.g. empty seeds.

Category a. and b. may be germinable but dormant. Their correct

status may be further determined by viability test. If the number of
viable but not germinated seeds is high, a new germination test fol-
lowing new pretreatment may be appropriate.

The final evaluation of the germination test is reported as germi-

nation percentage or germination capacity, which counts ‘normal
germinants’.

22 SEED TESTING
11.7 Other Seed
11
Two additional seed quality tests shall be mentioned viz. vigour and
Testing seed health tests. Their rationale is discussed below. Neither of the
methods are carried out as routine tests by seed laboratories. Unlike

SEED TESTING
the above tests there are no strict adopted standard procedures for
conducting the tests and evaluating the results. However, in some
cases results from germination or viability tests can be used for evalu-
ating both seed vigour and health.

11.7.1 Vigour test The main limitation of the germination test is its inability to detect
quality differences among seed lots with high germination percent-
ages. Vigour test is a more sensitive test, which aims at detecting such
differences. Several definitions of seed vigour exist, the two most
common ones are formulated by ISTA and Association of Official
Seed Analysts (AOSA):

1. Those seed properties, which determine the potential for

rapid, uniform emergence, and development of normal
seedlings under a wide range of field conditions (AOSA
1983).
2. The sum of the properties which determine the potential
level of activity and performance of the seed or seedlot
during germination and seedling emergence. Seeds which
perform well are termed ‘high vigour seeds’ (ISTA 1995 cit.
Perry 1981).

Practical application of a vigour test is mainly related to field perform-

ance and storability. Since germination tests are carried out under
optimal germination conditions, the test results express the germination
potential, i.e. likely germination under ideal conditions of tempera-
ture, light and humidity. This figure may be quite different from the
performance under stressed field conditions. As discussed in chapter
8.5, seeds undergoing natural ageing during storage are likely to lose
vigour at a faster rate than they lose viability. For the individual seed,
decline of vigour may be manifested in reduced germination capacity
under suboptimal conditions. Hence, a standard germination test of a
low-vigour seed lot may show relatively high germination, while a test
conducted under stressed conditions (which probably in most cases
would better reflect real field conditions) may show comparatively
poor germination (ISTA 1995). Further, a low vigour seed lot has in
itself comparatively short storability, whereas a high vigour seed lot
will produce a high number of normal seedlings under a wide range of
environmental conditions and it will store better (i.e. lose viability at
a lower rate). As can be seen from fig. 11.11, there is high conformity
between viability and vigour for high vigour seed lots.

Since seed vigour is a concept describing several characteristics associ-

ated with seed performance, and not a single measurable character,
there is no adopted standard procedure to measure vigour, but rather
a number of options. Vigour tests are described thoroughly in ISTA
(1995) and AOSA (1983). Some of the methods are summarized
below.

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 23

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 Figure 11.11.
Relation between viability
and vigour over a period of
time. The relation may be ex-
pressed as germination under
a certain set of conditions
after a certain ageing period.
The viability curve represents
germination under optimal
conditions while the vigour
curve expresses germination
under stressed conditions. Ex-
ample: after 8 months’ storage,
germination under optimal
conditions is about 80% while
under stressed condition it is
only about 30%.

Velocity of germination

Germination percentage only states the percentage of the seeds that have
germinated during the test period (germination capacity), not whether
germination occurred during the first or last part of the test period. Under
field conditions rapid germination is obviously an advantage for seedling
establishment. Speed of germination is an expression of seed vigour. It is
anticipated that high-vigour seeds germinate faster than low-vigour seeds
under any conditions. Speed of germination can be calculated from the
daily germination records. An example of records of a daily count of
a four-week germination test is presented in fig. 11.12. The velocity of
germination, termed ‘germination energy’ can be expressed in various
ways from the germination results (cit. Willan 1985).

1. As the percentage of tested seeds that germinate within a

given period, shorter than the total test period, e.g. 7, 14 or
Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

21 days, depending on species. In the example in fig. 11.12,

48% have germinated after 14 days, 60% after 21 days.
2. As the percentage of tested seeds that germinate up to the
time of peak germination, which is the highest number of
germinants appearing in a given 24 hour period. The peak of
the records in fig. 11.12 occurs at day 10, up to which 26%
by Lars Schmidt, Danida Forest Seed Centre. 2000.

have germinated.
3. As the number of days required to reach 50% of the final
germination percentage. It is 11 days in fig. 11.12.
4. As the average germination speed over the full test period,
based on daily counts. This is calculated by the following
formula:
Cumulative daily germination percentage
Σdaily germination speed (DGS)
= number of test days
total test period (days)
total test period (days)

24 SEED TESTING
 11
Some attempts have been made to combine germination capacity
and germination energy into a single figure, expressing both total
germination and germination speed. The term ‘germination value’

SEED TESTING
was introduced by Czabator (1962) and later evaluated and modified
by Djavanshir and Pourbeik (1976). However, the term is not much
used, since it is usually preferred to report germination capacity and
germination energy in separate figures.

Figure 11.12.
Daily recorded germination
in a germination test of Pinus
caribaea. Columns indicate
the average percentage of
germinants registered each test
day during the full test period.
Curve indicates the cumulative
percentage of germinants over
the period (Data from Paul
1972, quoted in Willan 1985).

Conductivity test

The conductivity test is based on the assumption that a disintegration

of cell membranes in the seed takes place during seed deterioration.
During the early stages of imbibition the cell membranes reorganize
and repair possible damages. Delayed repair or failure to overcome
such membrane damages causes leakage of electrolytes from the
imbibing seeds. A conductivity meter can measure the leakage of
electrolytes into the water in which seeds are imbibed. Since high-
vigour seeds are able to reorganize their membranes more rapidly
and repair any damage to a greater extent than low-vigour seeds,
electrolyte leakage can indicate level of vigour. Hence, low conduc-
tivity indicates low electrolyte leakage and thus high vigour, high
conductivity accordingly indicates low vigour (ISTA 1995).

Accelerated ageing

Accelerated ageing (AA) is a stress test with two main applications

in practical seed handling: (1) to predict the potential storage life of
seed, and (2) to assess vigour of a seed lot. The first application is
discussed in connection with storage in appendix A8.1. AA is based
on the assumption that if seeds deteriorate at a certain predictable
rate under a given set of storage conditions (mainly as a function
of temperature and humidity), then deterioration will occur much
faster under sub-optimal conditions of increased temperature and/or
humidity. The basic assumption is that the same process of deterio-
ration which takes place during a natural (slow) ageing period will
occur during a short period when seeds are exposed to unfavourable
conditions (Delouche and Baskin 1973). In other words: natural de-
terioration is simulated and compressed into a short convenient test
period. Under such conditions high-vigour seed lots will show only
slight decline in germination while low-vigour seed lots will decline
markedly after exposure to AA (Elam and Blanche 1989).

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 25

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 AA has proven a useful method for comparison of parameters
related to seed deterioration. However, it is quite evident that at
least some factors have a larger influence in AA than during natural
ageing mainly because of higher humidity. Micro-flora (fungi) and
repair mechanism of cell organelles are two factors apparently more
prevailing under AA conditions than under natural ageing (Priestley
1986).

Stress test

A number of methods have been used to evaluate seed and seedling

performance under stressed conditions, which are all attempted to
simulate single stress factors occurring under field condition. These
tests are germination tests carried out under sub-optimal conditions
and hence differing from normal germination tests. The type of suit-
able stress factor depends on species and, except for the exhaustion
test, the factor most likely to be encountered in the field. The methods
of conducting stress tests are described thoroughly in ISTA (1995) and
AOSA (1983). The methods are mentioned only briefly here.

During the Hiltner test the ability to overcome physical stress is evalu-
ated by germinating the seeds under a 3-4 cm thick layer of crushed
brick stone or gravel. Cold test evaluates the ability to germinate and
grow under low temperatures. This test is frequently used for temper-
ate species but is also suitable for tropical and subtropical highland
species. High temperature and water-stress are other variable factors
likely to reflect difference in vigour.

The exhaustion test is based on the principle that seeds germinated in

darkness do not carry out photosynthesis but rely entirely on nutrients
derived from the seed. The germinants become etiolated, and after
a specified test period the dry weight of the seedlings is measured.
Seedlings derived from high vigour seeds have the highest dry weight
(Poulsen 1993). Obviously this method is not applicable to seeds that
Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

require light for germination (cf. sections 9.5.5 and 10.5.3).

Seedling growth

The standard germination test only distinguishes between normal

and abnormal germinants. Variations in seedling size and vigour
by Lars Schmidt, Danida Forest Seed Centre. 2000.

are likely to occur within the category ‘normal germinants’. Since

initial growth is highly influenced by the seed, evaluation of seedling
vigour, expressed e.g. as dry weight or evaluated in size classes, is in
turn an expression of seed vigour. Comparison of different seed lots
must obviously be carried out under strict observation of standard
germination conditions and duration of test period. The latter implies
that seedlings must not be removed during the test as is customary
during normal germination evaluation.

11.7.2 Seed health Seed health is indirectly revealed during viability, germination or
testing vigour tests since infected seeds are often unable to germinate or
appear non-viable when examined by e.g. X-ray, TZ test or other

26 SEED TESTING
 11
viability test methods, or they germinate slowly and produce poor
seedlings. However, in some instances a more thorough examina-
tion of the presence, and type of seed-borne pests and pathogens

SEED TESTING
is relevant. Especially in international transfer of seed, where there
is a risk of introducing seed borne organisms together with seeds, a
special health test may be required. Methods of seed health testing
are described in Richardson (1990), and the ISTA rules (ISTA 1996)
provide general guidelines on health testing.

The level of seed health testing varies from simple assessment of

infection rate by visual examination of the seed sample under a
stereo microscope, to thorough examination and species identifica-
tion after incubation.

Assessment of insect infestation rate may be carried out as part of

X-radiography or a cutting test as described under viability. Where
identification of insects is required, it is often necessary to acquire
adult specimens. Since the insects present inside the seeds are often
in the larva or pupal stage, incubation under conditions that promote
their development may be necessary.

Fungal spores present on the surface of the seed may be detected by

microscopical examination of an aqueous suspension after washing
the seeds in a small quantity of water (Desai et al. 1997). However,
most fungal examination requires pre-incubation under warm moist
conditions. Incubation is normally conducted on blotter paper, sand
or agar plates. Sand and blotter paper are used where germination is
desired. After a few days’ incubation fungal growth may be visible
on seed-coats or as symptoms appearing on the seedlings. It should
be stressed that certain pre-treatment methods e.g. sulphuric acid or
hot water used for breaking physical dormancy should not be used in
seed health testing as they may kill possible fungi and hence interfere
with the result. Where pretreatment of such seeds is necessary, they
should be mechanically scarified. During the agar method seeds are
placed on the surface of a sterile nutrient agar gel during incubation.
The fungi will grow and form a colony on the agar plate, which may
be identified by its colour and type of growth (Desai et al. 1997).

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 27

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 REFERENCES AOSA 1983. Seed vigor testing handbook. Contribution no. 32 to the handbook on
seed testing. Association of Official Seed Analysts. USA.
ASEAN 1991. Standard germination test. Training Course Proceedings No. 2. ASEAN-
Canada Forest Tree Seed Centre Project. Muak Lek, Thailand
Bhodthipuks, J., Saelim, S. and Pukittayacamee, P. 1996. Hydrogen peroxide (H2O2)
testing for viability of tropical forest tree seed. In: Rapid Viability Testing of Tropi-
cal Tree Seed (Bhodthipuks et al. eds.). Training Course Proceedings No. 4: 11-15.
ASEAN Forest Tree Seed Centre Project. Muak Lek, Thailand.
Bhodthipuks, J., Pukittayacamee, P., Saelim, S., Wang, B.S.P. and Yu, S.L. (eds) 1996.
Rapid Viability Testing of Tropical Tree Seed. Training Course Proceedings No. 4.
ASEAN Forest Tree Seed Centre Project. Muak Lek, Thailand.
Boland, J.D., Brooker, M.I.H. and Turnbull, J.W. 1980. Eucalyptus seed. CSIRO.
Australia.
Bonner, F.T., Vozzo, J.A., Elam, W.W. and Land, S.B. Jr. 1994. Tree seed technology
training course, instructors manual. US. Dept. Agric. Southern Forest Experiment
Station.
Carlowitz, P. G. von. 1991. Multipurpose trees and shrubs. International Council for
Research in Agroforestry (ICRAF), Kenya.
Chaichanasuwat, O., Wang, B.S.P. and Wasuwanich, P. 1990. Evaluating seed quality
of Peltophorum pterocarpum by X-radiography and germination. In: Tropical Tree Seed
Research (Turnbull, J. ed.). Proceedings of an International Workshop held at the
Forestry Training Centre, Gympie, Qld., Australia, 1989. pp. 68-72.
Chaiyasit, L., Piewluang, C., Wasuwanich, P. and Pukittayacamee, P. 1990. X-radio-
graphic determination on seed germinability and early seedling survival in Pterocarpus
macrocarpus Kurz. The Embryon, 3:1, 1-5.
Czabator, F.J. 1962. Germination value: An index combining speed and completeness
of pine seed germination. Forest Science 8: 4, 386-396.
Delouche, J.C. and Baskin, C.C. 1973. Accelerated aging techniques for predicting the
relative storability of seed lots. Seed Sci. and Technol. 1: 427-452.
Desai, B.B., Kotecha, P.M. and Salunkhe, D.K. 1997. Seeds handbook, biology, pro-
duction, processing and storage. Marcel Dekker, Inc.
Djavanshir, K. and Pourbeik, H. 1976. Germination value - a new formula. Silvae
Genetica 25: 2, 79-83.
Duangpatra, J. 1991. Seed vigour testing. In: Standard Germination Test. Training
Course Proceedings No. 2: 49-52. ASEAN-Canada Forest Tree Seed Centre Project.
Muak Lek, Thailand.
Edwards, D.G.W. and Wang, B.S.P. 1995. A training guide for laboratory analysis of
forest tree seeds. Pacific and Yukon Region Inf. Rep. BC-X-356. Canadian Forest
Service.
Elam, W.W. and Blanche, C.A. 1989. Accelerated aging: a potential vigour test for
multipurpose tree seeds. In: Tropical Tree Seed Research (Turnbull, J.W. ed.). Proceed-
Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

ings of an International Workshop held at the Forestry Training Centre, Gympie,

Qld, Australia, 1989. pp. 63-67.
Enescu, V. 1991. The tetrazolium test of viability. In: Tree and shrub seed handbook,
chapter 9. International Seed Testing Association ISTA. Zurich.
Griffin, A.R. 1972. The effects of seed size, germination time and sowing density on
seedling development in Radiata Pine. Australian Forest Research, 5(4), 25-28.
ISTA 1986. ISTA handbook on seed sampling. International Seed Testing Association.
by Lars Schmidt, Danida Forest Seed Centre. 2000.

Zurich, Switzerland.
ISTA 1991. Tree and shrub seed handbook. The International Seed Testing Associa-
tion (ISTA). Zurich, Switzerland.
ISTA 1995. Handbook of vigour test methods. The International Seed Testing As-
sociation (ISTA). Zurich, Switzerland.
ISTA 1996. International Rules for Seed Testing, Rules 1996. International Seed Test-
ing Association (ISTA). Seed Science and Technology 24 (supplement). Zurich,
Switzerland.
ISTA 1998. ISTA Tropical and sub-tropical tree and shrub seed handbook (Poulsen
K.M., Parratt, M.J. and Gosling, P.G., eds.). International Seed Testing Association
(ISTA). Zurich, Switzerland.
Laedem, C.L. 1984. Quick tests for tree seed viability. B.C. Ministry of Forests and
Lands Research Branch. Victoria, B.C.
Laedem, C.L., Bhodthipuks, J. and Clark, J.M. 1995. Effects of stratification and
temperature on the germination of Dalbergia cochinchinensis, Pinus kesiya and Pinus
merkusii. Journal of Tropical Forest Science 7(3): 355-370.

28 SEED TESTING
 11
Mittal, R.K. 1997 (pers. comm.). Senior Scientist (Pathology), Indian Council of
Agric. Res., Almora, U.P.
Moore, R.P. 1985. Handbook on tetrazolium testing. International Seed Testing As-
sociation (ISTA). Zurich, Switzerland.

SEED TESTING
Nydam, J. 1997. (pers. comm.) Danish Seed Testing Laboratory.
Paul, D.K. 1972. A handbook of nursery practice for Pinus caribaea var. hondurensis and
other conifers in West Malaysia. Wkg. Paper No. 19, FO: SF/MAL 12, UNDP/FAO
Kuala Lumpur.
Perry, D.A. 1981. Handbook of vigour test methods. International Seed Testing As-
sociation (ISTA). Zurich, Switzerland.
Peterson, J. 1987. Seed testing procedures for native plants. In: Germination of Austral-
ian native plant seed (Langkamp, P. (ed.). pp 31-45. Inkata Press, Melbourne.
Poulsen, K.M. 1993. Seed quality. Concept, measurement and methods to increase
quality. Lecture Note C-14. Danida Forest Seed Centre. Denmark.
Poulsen, K.M. 1994. Seed testing. Lecture Note No. C-8. Danida Forest Seed Centre.
Denmark.
Poulsen, K.M. 1996. Ways and recommendations on how to test seed quality of tropical
tree seed species. In: Innovations of Tropical Tree Seed Technology. Proceedings of
the IUFRO Symposium of the Project Group P.2.04.00, ‘Seed Problems’ Arusha,
Tanzania. pp. 233-241.
Priestley, D.A. 1986. Seed ageing, implications for seed storage and persistence in the
soil. Comstock Publ. Ass., Ithaga and London.
Richardson, M.S. 1990. ISTA handbook on seed health testing. The International
Seed Testing Association (ISTA). Zurich, Switzerland.
Saelim, S., Pukittayacamee, P., Bhodthpuks, J. and Wang, B.S.P. 1996. X-radiography
testing for viability of tropical forest seed. In: Rapid Viability Testing of Tropical
Tree Seed (Bhodthipuks, J. et al. eds.). Training Course Proceedings No. 4: 17-31.
ASEAN Forest Tree Seed Centre Project. Muak Lek. Thailand.
Simak, M. 1990. Testing of subtropical and tropical forest tree seeds by X-radiography.
In: Tropical Tree Seed Research (Turnbull, J.W. ed.). Proceedings of an International
Workshop held at the Forestry Training Centre, Gympie, Qld., Australia, 1989.
pp. 72-77.
Simak.M. 1991. Testing of forest tree and shrub seeds by X-radiography. In: Tree and
shrub seed handbook. Chapter 14. The International Seed Testing Association
(ISTA). Zurich, Switzerland.
Sorensen, F.C. and Campbell, R.K. 1993. Seed weight - seedling size correlation in
coastal Douglas-fir: genetic and environmental components. Can. Jour. For. Res.
23(2), 275-285.
Willan, R.L. 1985. A guide to forest seed handling. FAO Forestry Paper 20/2.
Danida/FAO.
Yu, S.L. and Wang, B.S.P. 1996. Tetrazolium testing for viability of tree seed. In:
Rapid Viability Testing of Tropical Tree Seed (Bhodthipuks, J. et al. eds.). Training
Course Proceedings No. 4: 33-58. ASEAN Forest Tree Seed Centre Project. Muak
Lek, Thailand.
Yue-Luan, Hor. 1993. Seed testing for selected tropical trees in ASEAN region. Re-
view Paper No. 2. ASEAN - Canada Forest Tree Seed Centre Project. Muak Lek,
Thailand.

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 29

by Lars Schmidt, Danida Forest Seed Centre. 2000.
 Appendix A11.1 This appendix gives practical guidelines for conducting viability tests
Procedure for with tetrazolium. The procedure as described here has been compiled
tetrazolium testing from Yu and Wang (1996), Moore (1985) and ISTA (1996). For detailed
information reference is made to these sources.

Since TZ measures the activity of metabolic enzymes (dehydro-

genases) in living cells, it is necessary that seeds are imbibed and
incubated at a temperature allowing active metabolism during the
test. For most species it is necessary to pre-moisten the seeds, either
slowly (between or on top of moist paper) or fast (soaking in water)
until fully imbibed. Seeds with a hard seed-coat must be scarified
(punctured) to allow imbibition. Seeds enclosed within a hard peri-
carp must be scarified or extracted prior to imbibition.

Tetrazolium chloride or bromide is available as a salt which is dis-

solved in water before use. The salt as well as the solution are light
sensitive and must be stored in a dark bottle and/or in darkness.
To avoid light exposure the bottles may be wrapped in light-proof
material such as alu-foil. Solutions stored cold may retain their
strength for several months. The chemical is normally not re-used
but discharged after completion of the test. Tap water (clean) may be
used for the solution provided the pH is within the range 6.5 - 7.5. If
the pH is higher or lower, a buffer solution prepared with KH2PO4
and Na2HPO4 should be used. A buffer solution is prescribed by
ISTA (1996):

Solution 1: Dissolve 9.078 g of KH2PO4 in 1 litre of water

Solution 2: Dissolve 9.472 g of Na2HPO4 in 1 litre of water,
or 11.876 g of Na2HPO4 x 2H2O in 1 litre of water

Mix two parts of solution 1 with three parts of solution 2. If the

solution is not clear, add a drop of alcohol to clear it. Check the
pH (6.5-7.5).
Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’

For most species a 1% TZ solution is used. This concentration is

achieved by dissolving 1 g TZ salt in 1 litre of water or buffer. The
practical steps of the TZ test are as follows:

1. Prepare hard seeds for imbibition by scarification, puncturing

or extraction.
by Lars Schmidt, Danida Forest Seed Centre. 2000.

2. Pre-moisten seeds by soaking, or between or on moist paper

at approx. 20°C for 3-48 hours (depending on species, cf.
ISTA prescriptions); drain off water.
3. Immerse seeds in the TZ solution. The seeds should be
completely covered.
4. Incubate seeds in the TZ solution in darkness at 30-35°C for
1-24 hours (depending on species, cf. ISTA prescriptions).
5. Wash seeds in distilled water and place them on moist filter
paper until evaluation.
6. Evaluate staining.

The duration of incubation in TZ should as far as possible follow

the guidelines as listed by e.g. ISTA (1996). For species where no

30 SEED TESTING
 11
information is available, duration must be determined by experi-
ence. Too short incubation implies insufficient staining. Too long
incubation makes evaluation difficult since the tissue tends to be

SEED TESTING
very dark coloured and dead tissue dissolves. Evaluation should be
done shortly after termination of incubation; the seeds are meanwhile
stored in the dark. A few hours’ exposure to light, e.g. in connection
with evaluation is, however, not critical (Nydam 1997, pers. comm.).
Seeds fully stained are classified as viable; those not stained as non-
viable. Seeds only partly stained are thoroughly examined. Only
those with fully stained embryos, or with only minor portions of the
cotyledons unstained, are classified as viable. Detailed guidelines on
the evaluation are given in the above references.

Tetrazolium chemicals have been under suspicion for having carci-

nogenic effects. Though there is at present no confirmation to that,
it is advised to handle the chemical with care i.e. using mask and
ventilation during handling of the powder and rubber gloves during
handling of dissolved chemical.

Extract from ‘Guide to Handling of Tropical and Subtropical Forest Seed’ 31

by Lars Schmidt, Danida Forest Seed Centre. 2000.