Previous Page Table of Contents Next Page

Chapter 9 SEED TESTING (Contd.)

Combining purity and Germination Tests

Purity tests of most commercial seed lots of Eucalyptus are not made because it is difficult or impossible to separate the seed from the chaff of some species (Boland et al. 1980). Species in which seed and chaff are extremely similar in size, weight and colour include E. cloeziana, E. regnans and E. delegatensis. other small-seeded genera in which separation of pure seeds is difficult are Alnus, Betula, Populus and Salix. Even when separation is possible in these species, it is very time-consuming. In the seed laboratory in Canberra it takes only 6–7 minutes to set up a test of eucalypt seed with 4 weighed replicates, but to carry out a purity separation and set up a 4 × 100 seed test takes from 20 to 50 minutes, depending on the degree of difficulty in separating out the chaff (Turnbull 1983). The small size of seed also precludes a cutting test to determine the number of full but ungerminated seeds at the end of a test. For these reasons tests are best made on replicates by weight and results recorded as numbers of germinated seeds per unit weight of the impure mixture of seed and inert material. A squash test may also be used to give a rough estimate of viability.

Where seed is sown broadcast, the practising forester is primarily interested in the number of plants he can expect to get from a given weight of the seed lot he receives. Provided he is told that 1 kg of “impure” seed should produce 36,000 germinated seedlings, he is not worried whether this is caused by a combination of 90 % purity × 80 % pure seed germination or 80 % purity × 90 % pure seed germination. For local tests, therefore, if seed laboratory staff is insuffient, there is no objection to omitting the purity test even on species for which it would be practicable. In the case of direct sowing into individual containers, however, there are obvious advantages in having separate figures for purity and germination, since sowing is by numbers of (one to several) seeds per container rather than weight of seed per m2 of bed or tray.

Indirect Tests of Viability

Estimating the germination potential of a seed lot by actually germinating a sample of it is often the method most relevant to practical forestry. But the tests take several weeks to complete and for some species pretreatment may take some additional weeks or months. For this reason much research has been conducted to find other methods by which seed viability can be estimated accurately but much more rapidly than by germination testing. The following account of these methods closely follows that of Turnbull (1975 d).

The objects of quick viability tests are:

Only two methods, the topographical tetrazolium test and the embryo excision test were previously accepted by the International Seed Testing Association as official methods for some species of seeds. ISTA has recently accepted the X-ray method as a valid alternative to the cutting test for the detection of empty and insect-damaged seeds. The following tests can be applied depending on the circumstances:

Cutting test

The simplest viability testing method is direct eye inspection of seeds which have been cut open with a knife or scalpel. If the endosperm is of normal colour with a well developed embryo, the seed has a good chance of germinating. This test is not very reliable. Seeds with milky, unfirm, mouldy, decayed, shrivelled or rancid-smelling embryos and abortive seeds that have no embryo can be judged as non-viable without much difficulty (Bonner 1974). But it is not possible to distinguish moribund, recently dead or recently injured seeds which still appear the same as sound ones. The cutting test, as already mentioned, is used at the end of a germination test to determine the apparent viability of ungerminated seeds; it is also a useful tool in estimating the size and maturity of the seed crop before collection (Chapter 3) and the efficiency of methods used in processing.

In the philippines good correlation has been found between cutting and germination tests in fairly large-seeded species such as Leucaena, Intsia bijuga and Lagerstroemia speciosa (Seeber and Agpaoa 1976), but germination % was consistently 10–20 % less than the percentage of sound seeds on cutting test.

Topographical tetrazolium test

The tetrazolium method is only one of a number of biochemical tests which have been developed for seed testing. The various tests have been briefly reviewed by Moore (1969). The tetrazolium test was introduced in 1942 by G. Lakon in Germany.

In this method living cells are stained red by the reduction of a colourless tetrazolium salt to form a red formazan. The method emphasizes the need for a knowledge of the soundness of individual embryo parts for predicting the development of embryos into countable seedlings (Moore 1973).

The testing procedure is described in detail in the ISTA Rules (ISTA 1976) which approve the test for some species of hardwoods and conifers which germinate slowly by regular germination methods. Normal practice is to soak the seeds in water for about 20 hours, then cut or puncture the seed coat to facilitate entry of the 1% aqueous solution of tetrazolium (TZ) and immerse the seeds in the dark for 48 hours (Bonner 1974). The process can be greatly speeded up by cutting through the seed at a distance of one third from the micropyle and placing it in a Vitascope vacuum machine for only half an hour. This method gives satisfactory results in Denmark but interpretation of results requires more experience in the operator than the method of seed immersion followed by excision of the stained embryo (Knudsen 1982). The test is done on 4 replicates of 100 seeds each (ISTA 1976).

Justice (1972) states that, while the tetrazolium procedure is good in principle, its practical use in routine testing is limited by many problems, including: difficulty in staining of some seeds; necessity of cutting of dissecting seeds to permit observation of stained parts; poor agreement with results of germination tests in some cases, especially for seed of low germination capacity; lack of uniform interpretation of staining and difficulty in interpreting the significance of different degrees of staining; and an increase in man-hours required to test 400 seeds compared to regular germination tests.

The necessity for an experienced analyst for the successful use of this “common sense test” is admitted by Moore (1973). There is little doubt that the test can be useful for testing the viability of certain species, providing trained staff are available to prepare the seeds and evaluate the results.

Excised embryo test

By this method, the seeds are soaked for 1 – 4 days and the embryos are then excised from the seeds and placed on moist filter paper or blotter discs in petri dishes. The tests are placed in the light at a constant temperature of 20°C. The condition of the embryos is examined daily. Depending upon the species and lot differences, the tests can be terminated after only a few days, up to a maximum of 14 days, or as soon as distinct differentiation into viable and non-viable embryos can be made.

The excised embryo test is similar to germination tests in that it measures the quality of the seed by their actual germination. In addition it allows some measure of the embryo dormancy to be made, by counting those seeds which, although not growing normally, have grown slightly, remained firm and have kept their colour for the test period. The test is not valid for previously germinated seeds and must not be applied to samples which contain any dry germinated seeds. The success of the test requires considerable skill and experience in the operator and the ISTA rules restrict it to only a few species.

In a comprehensive study, Schubert (1965) compared the excised embryo method with the tetrazolium method for determining the viability of dormant tree seeds. He concluded that the tetrazolium method should receive preference over the embryo excision method but that improvements in the tetrazolium test should be made by providing for the use of bactericides and stronger reducing solutions to resolve doubts in weakly stained tissues.

Radiographic methods

Radiography was first used to determine seed quality over 70 years ago (Lundstrom, 1903, cited by Kamra 1964). The studies of Simak and Gustafsson (1953) highlighted the X-ray technique as a diagnostic method of tree seed analysis. The X-ray contrast method which uses various contrast or radiopaque agents was developed and applied successfully to species of Pinus and Picea (Simak 1957; Kamra 1963 a, 1963 b).

The X-ray method permits the detection of empty seeds, mechanical damage and abnormally developed internal seed structures, measurement of the thickness of the seedcoat and assessment of the seed viability when combined with a contrast agent.

The X-ray contrast method is based on the principle of semipermeability. When seeds are treated with a contrast agent, for example aqueous BaCl2 or vaporous CHC13, their living tissues are able to prevent its entry due to their semi-permeability, but the dead tissues become impregnated. The impregnated tissues absorb X-radiation more intensively than the unimpregnated ones and thus appear lighter on the film than the unimpregnated ones. The contrast permits living and dead tissue to be located in the seed and an estimation of its viability (Kamra 1964). There are now possibilities of using non-toxic water, instead of toxic BaCl2 or CHC13, as a contrast agent for testing seed viability (Simak 1982).


9.15 X-ray radiograph of teak fruits showing the variation in the number of locules (two to six). (S.K. Kamra)




9.16 X-ray radiographs showing embryo and endosperm classes in coniferous seeds. (M. Simak)

ONeither embryo nor endosperm (= empty seed)
IEndosperm, embryo cavity developed but no embryo observed.
IIPEndosperm, one or more small embryos the length of which does not exceed their breadth (“point embryos”).
IIEndosperm, and one or several embryos, none of which is longer than half of the embryo cavity.
IIIEndosperm, and one or more embryos, the longest of which measures between half and three quarters of the embryo cavity.
IVEndosperm, with one fully developed embryo, completely or almost completely occupying the embryo cavity. Diminutive embryos rarely occur.
AThe endosperm almost fills the seed coat to capacity and easily absorbs the x-rays.
BThe endosperm fills the seed coat incompletely and is often shrunken or otherwise deformed. The x-ray absorption is inferior to that of class A.
AbSeed with abnormally developed endosperm or embryo.
JSeeds damaged by insects, containing larvae (JI) or their excrement (Je).

The development of soft X-ray equipment has greatly simplified the operation (Belcher 1973). Complicated photographic equipment is not necessary and pictures can be made with polaroid film which provides clear and detailed radiographs within 30 seconds (Edwards 1973).

X-ray radiography has been successfully applied in determining the number of seeds in fruits of teak (Tectona grandis) and for studying their degrees of development (Kamra 1973). The technique has been tried on the fruits or seeds of sixty tropical forestry species and the results show that it can be reliably applied in processing such seeds (Kamra 1974, 1976, 1980).

A technique of stereoradiography as a supplement to the X-ray contrast method for use in seed quality testing has been developed by Kamra, Meyer and Wegelius (1973). The chief advantage of stereoradiography is that it is possible for the observer to have a three-dimensional view of the object from a pair of radiographs. In this way, the exact topographical location of the contrast agent in the seed can be reliably determined. This increases the information which can be obtained from radiographs and adds to the analytical accuracy.

The X-ray method is a useful one and is likely to play an increasing role in seed testing. Earlier models of X-ray machines were expensive, but recent models, particularly from Japan are much cheaper and now cost less than a cabinet germinator. Improvements in photographic films and paper have speeded up the process and simplified the interpretation, so that technicians can be trained easily to produce consistent results. ISTA has accepted the method as a valid alternative to the cutting test for the detection of empty and insectdamaged seeds. It also shows considerable promise for distinguishing between viable and non-viable seeds among “full seeds” (Simak 1980, Simak and Sahlén 1981). For certain temperate conifers it has been possible to obtain good correlation between the Development Class (DC) of seeds, based on the development of both embryo and endosperm, and their germinability. Figure 9.16 illustrates the DCs which have been defined for conifers, and Table 9.4 illustrates the germinability corresponding to each class, for Pinus sylvestris and Picea abies (Simak 1980).

T a b l e 9.4

Germinability of fresh collected, undamaged seeds (in %) belonging to different DCs, selected by radiography

(Material: Widespread seed samples from the whole of Sweden. Germination test: Jacobsen's apparatus, temperature 23°C constant, light 1000 lux 8 hours/day).

Pinus sylvestris05088990  54368
Picea abies03682970157192

For key to Development Classes, see Fig. 9.16.

Hydrogen peroxide

Hydrogen peroxide (H2O2) has a stimulating effect on seed germination and has been used in a rapid test for germination of several conifers in the western USA (Bonner 1974). Seeds are soaked overnight in 1% H2O2. The seedcoat is then cut open to expose the radicle tip and the seeds put back into 1% H2O2 in the dark at alternating temperatures (20° and 30°C). Counting and refreshment of H2O2 is done after 3 or 4 days and final assessment after 7 or 8 days. Radicle growth of 5 mm or more is scored “evident”, 0 – 5 mm “slight” and no growth means a nonviable or empty seed (Danielson 1972 cited by Bonner 1974). The test is quicker but less reliable than a normal germination test (usually producing a more rapid and higher final germination), slower but simpler to perform than excised embryo and easier to interpret than TZ.


9.17 X-ray radiograph of Pinus caribaea seed. Most seeds have very poorly developed gametophyte and embryo and are dead. A few germinable are indicated by the black outlines. (M. Simak)

9.189.18 Quercus seeds cut in halves for oven drying in moisture determination. (USDA Forest Service)
9.19 Dole electric seed moisture meter used in the USA. (USDA Forest Service)
9.20 Electric moisture meters used in Denmark. A: Jacoby's infrared. B: Super-matic. C: Mettler. (DANIDA Forest Seed Centre)

Testing Moisture Content

The importance of moisture content in affecting the longevity of seeds in storage has been stressed in Chapter 7. In order to control the operations of drying (or moistening) of seeds in preparation for storage and to check the stability of moisture content during storage, it is clearly essential to have reliable methods of measuring the amount of moisture in a given sample.

Methods of determining moisture content of seeds have been classified by Justice (1972) into (a) basic methods in which the moisture is driven out of the seeds by heat and measured by the loss of weight of the original material, or the weight or volume of the condensed moisture, and (b) practical methods designed for rapid routine work and standardized against one or more of the basic methods. Probably all the moisture cannot be driven out of seeds without driving out small amounts of other volatile constituents or causing chemical changes in the material which would result in weight changes. In applying any method, therefore, it is necessary to adhere closely to the prescribed procedure in order that the results of all tests made by that method will be comparable.

Until recently ISTA prescribed three possible procedures:

(1) Drying in an oven for 17 hours at 103°C (2) Drying in an oven for one to four hours at 130°C and (3) Toluene distillation. Method (2) is applicable only to certain agricultural seeds and method (3), previously used for Abies, Cedrus, Fagus, Picea, Pinus and Tsuga, has now been eliminated because it had ceased to be used in practice (ISTA 1981 c). This leaves only method (1) - the “Low constant temperature oven method” as applicable to forest trees.

The test should be made on two samples of about 5 g each, drawn from the working sample including impurities, not on pure seeds. Large seeds should be ground, broken or cut into small fragments to facilitate drying and a good rule of thumb is that any seeds that average over 10 mm in diameter or length should be broken (Bonner 1981). The samples should be weighed and placed in metal containers, well-spaced to facilitate air circulation, within an oven which is maintained at a temperature of 103° ± 2°C for 17 ± 1 hours. At the end of that period the seed should be placed in a desiccator to cool for 30 – 45 minutes and then reweighed. The relative humidity in the laboratory where the final weighing is done should be less than 70%, to avoid rapid re-absorption of moisture. The difference in MC of the two samples should not exceed a stated tolerance %. If it does, a further pair of samples should be tested; otherwise the mean of the two samples is the final result. The tolerance earlier prescribed by ISTA for all species was 0.2% but, as pointed out by Gordon (1979) and Bonner (1981), a single tolerance figure is not applicable to all species. At the 1983 ISTA Congress in Ottawa, the tolerances approved for moisture tests in tree seeds were agreed as follows:

Sample ConditionTolerance %
Small seeds, moisture <12%, e.g. Picea, Alnus0.3
Large seeds, moisture <12%, e.g. Carya0.4
Small seeds, moisture >12%0.5
Large seeds, moisture 12 to 25%0.8
Large Seeds, moisture > 25%, e.g. Quercus2.5

For tropical tree seed laboratories wishing to conform to ISTA rules, this relaxation of tolerances will be a considerable help.

The calculation of moisture content should be made on a wet weight or fresh weight basis (see pp. 122–124) i.e.

Although wet weight basis is prescribed by ISTA and is becoming increasingly the standard form for expressing moisture content, it is not yet universal. To avoid any doubt, the method of calculating moisture content should be stated explicitly on any certificate or statement of results.

As explained by Gordon and Rowe (1982), provided that the initial fresh weight of a seed lot is measured and the initial moisture content (wet weight basis) calculated by oven-drying a sample, any new MC reached as a result of drying (or wetting) can be calculated directly from the new weight of the seed lot; there is no need for further oven-drying of samples at the new MC. The desired weight of the seed lot to be achieved through drying (or wetting) can be calculated by multiplying its initial weight by the initial dry matter percentage and dividing by the desired dry matter percentage.

e.g.(1)If initial wet weight of a seed lot = 50 kg and the MC (wet weight basis), determined by oven-drying a sample, is 25%, the oven-dry weight = 75% of wet weight = 37.5 kg.
(2)If a period of drying reduces the wet weight to 46.5 kg, the new MC
(3)If it is desired to reduce the MC (wet weight basis) to 10%, then desired oven-dry weight will be 90% of new wet weight and the seed lot must be further dried until its wet weight

Electric moisture meters give rapid estimates of seed moisture but they are not considered accurate enough for official seed testing. Their rapid operation does make them very useful in certain situations, for example they should be accurate enough to check tree seed moisture as a guide to drying seeds for storage (Bonner 1974, 1981). Meter readings are converted to seed moisture content by means of charts supplied by the manufacturer or developed from calibration curves in the laboratory for the species in question. Most meters will not measure moisture above 15 – 20 % and require a minimum of 90 – 100 g of seeds for a test (Bonner 1981). A locally made, cheap and portable electric moisture meter has been used successfully in Thailand for several years to measure the MC of rice grains (Kosol 1984) and could be used for tree seed of similar size. It measures electric capacitance and uses a 9-volt battery as the source of power.

Electric moisture meters are well suited to small seeds, but cannot be used for large seeds such as Juglans or Quercus; winged seeds such as Fraxinus are also difficult to measure (Bonner 1978). Large or winged seeds can be rapidly dried in a microwave oven. If the oven is preheated, drying can be completed in 5 minutes and weighing in 6 minutes immediately afterwards if in an electronic balance, or after 30 – 45 minutes' cooling in a desiccator if weighing is done in an ordinary balance (Bonner and Turner 1980). Results can be expected to be within 7 % at a probability of 0.05 in the case of large seeds of high MC such as Quercus, and within 2 % in Fraxinus and Carya, as compared with the more accurate results of the slower conventional methods.

A simple and cheap method of drying seeds quickly is to use an infra-red lamp (Gordon and Rowe 1982). A weighed sample is subjected to heat from an infra-red lamp with such an intensity that it loses all its moisture, without being burnt, in about 20 minutes. When the loss of weight ceases, the new weight is measured and the percentage loss calculated.

An up-to-date account of the measurement of tree seed moisture content has been published recently (Bonner 1981).

Other Tests

Other qualitative tests or observations may be made as the need arises, but do not call for detailed prescriptions. In many cases they can be combined with the purity test. They include:


There are several methods of determining whether the seeds are of the species stated. They are:

  1. Positive identification of the parent trees and their certification preferably on the basis of herbarium samples.

  2. Identification of the seeds by use of an analytical key or by comparison with a reference collection.

  3. Identification of the seedling. This may be the only way of determining if the seedlot is contaminated by hybrids or a mixture of two or more species with similar seed characteristics. A key and reference collection of seedlings will assist in identification (Turnbull 1975 d).

Authenticating seeds as to provenance is not possible for most species, but some progress has been made in this field for Pseudotsuga and Abies (Bonner 1974) and the use of isoenzyme techniques may open new possibilities (Burley 1976).

Damage, health

During the purity test, the operator should be alert to the incidence of mechanical damage and pathogenic infestation, which may indicate the need to improve the methods of transport or processing in use.

Calculation of Results

The following examples indicate the type of calculations required in the various stages of seed testing.


Weight of full working sample 62.52 g
Weight of pure seed 56.89 g
=91 %

Seed weight

The weight of 1000 seeds may be calculated as follows:


(a) Seed weight determination carried out on 8 × 100 seeds from the pure seed component of the purity test.

Replicate No.12345678TotalMean
Weight (g)3.813.693.753.793.823.723.713.7930.083.76

Since this is considerably less than the maximum of 4.0 prescribed by ISTA, the sample is judged to be homogeneous and no further sampling is needed.

Weight of 1000 seeds = 3.76 × 10 = 37.6 g


(b) Seed weight determination carried out on 1000 seeds from the pure seed component of the purity test, without replication.

Weight of 1000 seeds = 37.6 g

The number of seeds per unit weight may be derived as follows:


Test made on 4 × 100 seed replicates from the pure seed component of the purity test.

Replicate No.  1  2  3  4TotalMean
No. germinated on completion of test7985768832882
Sound seeds on cutting test  4  3  6  3  16  4

The range in number of germinated seeds from the largest to smallest of the replicates is 88 - 76 = 12. Reference to Table 9.3 on p. 220 shows that the maximum tolerated range for a mean germination of 82% is 15. Since the actual range is less, the sample is accepted as homogeneous.

Germination %=82%
Viability %=82 + 4 = 86%

Viable seeds per unit weight. A combination of the Viability % and weight of pure seed will give a figure for the number of viable seeds expected per unit weight of pure seed, while use of the germination % will indicate the number of germinable seeds. Incorporation of a factor for purity % will express the numbers in terms of the number of viable or germinable seeds per unit weight of “impure” seed.

 Pure Seed
 per gper kg
No. of viable seeds26.6 × 86 ÷ 100 = 22.922,900
No. of germinable seeds26.6 × 82 ÷ 100 = 21.821,800
 Impure Seed
 per gper kg
No. of viable seeds22.9 × 91 ÷ 100 = 20.820,800
No. of germinable seeds21.8 × 91 ÷ 100 = 19.819,800

In the case of very small-seeded species for which the purity test is impracticable, the number of germinated seeds per unit weight of impure seed is determined directly by test. Figures for numbers of pure seeds per unit weight are not obtainable for this type of seed. Although usually expressed as “viable seeds per g”, it should be noted that a cutting test is also impracticable for these small seeds, so strictly speaking the figures refer to germinable seeds. As an example:

Weight of replicate of impure seed (E. grandis) 0.10 g

Replicate No.1234TotalMean
No. germinated on completion of test6573637127268

No. of germinable (“viable”) seeds per g = 680
No. of germinable (“viable”) seeds per kg = 680,000

Germination energy. Calculation of germination energy and energy period depends on the criterion used to define this. Table 9.5 gives an actual example extracted from paul (1972). As mentioned above, the energy period may be arbitrarily defined in advance but is normally much less than the full period of the test. A single assessment is sufficient in this case. If, in the present example, the energy period had been defined as 12 days, then

Energy period = 12 days

If, on the other hand, the energy period is taken as that up to the day of peak germination, then daily assessment is necessary as shown in the table and

Energy period = 10 days

Table 9.5

Germination Test Sheet (extracted from Paul 1972)

Species: Pinus caribaea var. hondurensisTest No. 26/72
Seed Lot No. 85/71 
Date sown: 2/11/72Place: Mantin Nursery
Date completed: 30/11/72Germination Percent: 64%

Days after SowingSub-Samples (4 × 100 seeds)Daily TotalCumulativeTotalCumulative Total as % of total seeds.Mean daily germination %.Daily Total as % of germinable seeds.Cumulative total as % of germinable seeds
  1  -  -  -  -  ------
  2  -  -  -  -  ------
  3  -  -  -  -  ------
  4  -  -  -  -  ------
  5  -  -  -  -  ------
  6  -  -  -  -  ------
  7   -  -  -   -  ------
  8  6  8  7  829  29  7.250.9112  11
  9  5  9  9  831  6015.001.6712  23
10101113104410426.002.6017  40
11  9  810  93614035.003.1814  54
12  6  5  7  52316340.753.40  9  63
13  3  5  6  41818145.253.48  7  71
14  5  3  2  31319448.503.46  5  75
15  4  2  1  31020451.003.40  4  79
16  2  4  2  41221654.003.38  5  84
17  2  1  1  1  522155.253.25  2  86
18  1  2  2  3  822957.253.18  3  89
19  2  -  -  1  323258.003.05  1  90
20  2  1  2  -  523759.252.96  2  92
21  -  1  1  -  223959.752.85  1  93
22  2  1  2  -  524461.002.77  2  95
23  2  -  1  1  424862.002.70  2  97
24  -  1  1  1  325162.752.61  1  98
25  1  2-  -  325463.502.54  1  99
26  -  --  -  -25463.502.44-  99
27  -  --  -  -25463.502.35-  99
28  1  -  1  -  225664.002.29  1100
Totals63646861256   100 
Cut test  5  2  4  5  16     

Inspection of the pattern of germination suggests that rejection of all seeds germinating after peak germination would result in the rejection of an excessive proportion (60 %) of the potentially germinable seeds, while acceptance of all germinable seeds would unduly prolong the period of test and probably result in the inclusion of some seedlings of very poor vigour. A sensible rule of thumb applicable to the type of germination pattern in this example would be to define the energy period as lasting until daily germination falls to less than 25 % of peak. With this definition

Energy period = 16 days

Percentage of total germinable seeds which germinate within the energy period = 84 %.

Another commonsense measure of germination energy, used in Zimbabwe (Seward 1980), is the percentage of germination when mean daily germination (cumulative germination divided by time elapsed since sowing date) reaches its peak. In the present example (Table 9.5) the peak of mean daily germination percent is 3.48 %, the energy period 13 days and the germination % is

Percentage of total germinable seeds which germinate within the energy period measured in this way

Germination Value. Calculation of germination value, by the methods of Czabator (1962) and Djavanshir and Pourbeik (1976), is shown in Table 9.6, using the data from Table 9.5.

Table 9.6

Calculation of Germination Value

(Methods of (1) Czabator and (2) Djavanshir and Pourbeik)

Days since sowingDaily Germination PercentCumulative Germination PercentDaily Germination Speed (or Mean Germination) (Col. 3 ÷ Col. 1) DGSNumber of counts DGS/N (Col.5 ÷ Col.6)
  8  7.25  7.250.91  0.9110.91
  9  7.7515.001.67  2.5821.29
1011.0026.002.60  5.1831.73
11  9.0035.003.18  8.3642.09
12  5.7540.753.4011.7652.35
13  4.5045.253.4815.2462.54
14  3.2548.503.4618.7072.67
15  2.5051.003.4022.1082.76
16  3.0054.003.3825.4892.83
17  1.2555.253.2528.73102.87
18  2.0057.253.1831.91112.90
19  0.7558.003.0534.96122.91
20  1.2559.252.9637.92132.92
21  0.5059.752.8540.77142.91
22  1.2561.002.7743.54152.90
23  1.0062.002.7046.24162.89
24  0.7562.752.6148.85172.87
25  0.7563.502.5451.39182.86
26  0.0063.502.4453.83192.83
27  0.0063.502.3556.18202.81
28  0.5064.002.2958.47212.78

(1) Czabator method

Germination value = Final DSG × Peak Value DGS = 2.29 × 3.48 = 7.97.

(2) Djavanshir and Pourbeik method

Germination value = (Final DGS (N) × (Final Cumulative Germination %/10) = 2.78 × 6.4 = 17.79.

Indirect tests of viability

Similar methods should be used as for germination tests i.e. 4 replicates should be used, the results tested for homogeneity and the mean number of apparently sound full seeds (cutting test) or stained embryos (tetrazolium test) expressed as a percentage of the total pure seeds tested.

Moisture content (example of small seeds <12% MC)

 Original weightOven-dry weightDifference = 
moisture content
% MC
 (wet wt. basis)
Sample 15.655.140.51
Sample 24.924.470.45

Difference between samples (0.12%) is less than the 0.3% prescribed by ISTA, so no further sampling is needed.

Interpretation of Results

It is important that results of tests should lead to decisions and action. For example an abnormally low purity % may indicate a further cleaning operation for an organization which collects, processes and stores its own seed; or it may lead to a contested account if the seed was bought and the purity is less than that specified by the seller. Determination of moisture content is necessary before deciding whether a seed lot is suitable for immediate storage or should be subjected to further drying.

Most important of all are the various measures of germination potential and of the numbers of plants which can be produced from a unit weight of seed. This affects both nursery management and the achievement of afforestation targets, because it enables the somewhat crude estimates of average requirements shown in Table 3.1 (on p. 26) to be refined in the light of more precise data on the actual seed lots being used in a given year.

The results of the seed tests approved by ISTA make no forecast of survival and mortality after the stage of germination. Germination energy is the only measure which attempts to do this but it is entirely dependent on the subjective choice of an appropriate criterion by which to define the Energy Period. Extrapolation of seed testing results to nursery management is, strictly, outside the scope of this manual, but it is worth brief discussion here because it is of vital importance to treat seed management as part of a continuous chain of operations including plant production, plantation establishment and plantation management, and not as an end in itself.

The factors which the nurseryman needs to consider in converting seed testing results to nursery production include the following:

  1. Field germination in an operational nursery will often differ from test germination. It may be considerably lower than that in the ideal conditions of a laboratory test and somewhat lower than that in a research nursery. It should differ very little from a test carried out in the same nursery in advance of the main sowings. Observations on the differences between laboratory and nursery germination in trials in the Sudan have been reported by Wunder (1966). Differences varied with species and in some cases speed of germination was affected more than the final numbers germinated. Selected examples are:

    SpeciesLaboratory germination % in daysNursery germination % in daysPretreatment
    Acacia albida  76607060 
    Cedrela odorata100  38518 
    Prosopis chilensis100157938  2 minutes in conc. H2SO4
    Acacia raddiana100  330  460 minutes in conc. H2SO4

    Variation between nurseries may be associated with a number of different climatic, soil or cultural factors. As an example Roney and Brown (1978) found that germination of Pinus ponderosa was 38 % better if the seeds were covered with a depth of 1.5 cm of grit than if covered with a depth of 0.4 cm. Frequency of watering also has a significant effect on germination (Costales and Veracion 1978).

  2. Some losses may occur in the seed beds.

  3. Some losses may occur during pricking out or transplanting.

  4. Some losses may occur in the transplant beds.

  5. Some of the least vigorous surviving plants may be culled at planting time.

All of these factors together combine to make the number of plantable plants raised per unit weight of seed much less than the number of germinated seeds indicated by testing. The nurseryman must make an equivalent increase in the quantity of seed sown. The term “plant percent” or “tree percent” is frequently used to combine these factors. It is defined by Ford-Robertson (1971) as “The %, by number, of seeds in a given sample that develop into seedlings at the end of a given period, generally the end of the first growing season”. Bonner (1974) specifies “plantable seedlings” and this is commonly understood, since it adds the important concepts of usefulness and vigour to mere survival. For practical purposes it is preferable to broaden the usage to cover the whole period in the nursery i.e. “the %, by number, of seeds that develop into plantable plants (transplants or seedlings) at the end of a given period, generally the end of the nursery period at the time of field planting”. This is much more appropriate for tropical conditions where many species are pricked out within a week of germination and spend a further 6 – 12 months in transplant beds.

In practice, the record of plant yields obtained in past years from any given nursery provides the most useful basis for estimating future production from that locality (Aldhous 1972). Records based on production on other soil types or in other climatic conditions are very little help. Where experience is lacking, as in the early years of new nurseries, only rough “guesstimates” are possible. In Zambia a reduction factor of 20 % was allowed for the difference between laboratory and field germination in both pines and eucalypts (Allan and Endean 1966), i.e. if laboratory tests indicated that 1 kg of seed produced 100,000 germinated seedlings, it was assumed that only 80,000 germinated seedlings would be obtained in the nursery; a further reduction factor of 10 % was made for pricking-out losses in pines and 15 % in eucalypts. Some additional, but unspecified, allowance would be needed for later culling. Paul (1972) proposed a 20 % allowance for seedbed losses plus 15 % for pricking-out and transplant losses and culls. The nursery recovery factor, which converts test germination percent to plantable plant percent, varies considerably with species, seed lot and year. In Zimbabwe a nursery recovery factor of 70 % has been provisionally adopted for pines and of 25 % for eucalypts, equivalent to 30 % and 75 % nursery losses respectively (Seward 1980). Seeber and Agpaoa (1976) have shown that, in Pinus kesiya, the lower the germination % the greater the gap between the proportion of seeds which germinate and those which grow into plantable plants.

 Seed lot 1Seed lot 2
Number of pure seeds per liter30,00030,000
Purity %       95      95
Germination %       90      60
Plant %       60      30
Nursery recovery factor %                          
Number of plantable plants/ liter of seed30,000 × 0.95 × 0.60 = 17,10030,000 × 0.95 × 0.30 = 8,550

For Pinus patula Wormald (1975) reports that a survey of a number of countries raising the species in large quantities indicated a plant percent of about 33 % compared with an average of 80 % for germination tests, a recovery factor of about 45 %. He remarks that “From these figures it would appear that one usable plant is obtained for about every 3 seed that are sown. This compares well with the British Forestry Commission who have, over the last 15 years, improved their seedling recovery from one in ten to one in four (Gordon and Tee 1973)”. It confirms the statement in Goor and Barney (1976) that “Actually, in nursery and direct seeding practice, this amount (quantity of seed needed based on germination tests) would be doubled or tripled to compensate for seedling losses in the first few years”. The nursery recovery factor must be based on personal experience but, because of vagaries in weather, seed lots etc., it can never be very precise. Because of this there is a limit to the precision to be demanded from routine (as opposed to research) testing in the laboratory.

Using the example shown on p. 232–233, plant percent may be calculated as follows:

 Nursery Recovery FactorGermination
Germinable Seeds per kgPlantable Plants per kg
Nursery 1
Many years' experience in raising the species





Nursery 2
New, on difficult site, without experience in raising the species





On the assumption that both nurseries were required to raise 500,000 plants, nursery 1 would need

while nursery 2 would need

If seeds are to be sown into seed beds for later transplanting (rather than direct into pots) in addition to calculating the total amount of seed to be sown, the nurseryman will need to calculate the weight of seed to be sown per seed bed. For example if

Size of one seed bed = 10 × 1 m= 10 m2
Desired final stocking of seedlings= 2400/m2
Number of seeds/kg of pure seed= 26,600
Purity percent= 91% or 0.91
Germination percent= 82% or 0.82
Expected seedbed recovery rate 
(Seedlings which survive until transplanting as a proportion of germinated seeds reported by test)=0.65

then the required sowing rate

In the case of nursery 2 above, the number of seedbeds needed

The concept of “effective kilogram” of seed, which is in use now in several countries, has been found useful in planning sowing programmes and in calculating seed prices (Aldhous 1972). The “effective kilogram” is defined as the weight of seed of any particular seed lot which can be expected to produce the same number of viable seeds (as used in UK) or of plantable plants (as used in Zimbabwe) as would be produced by one kilogram of standard seed; this number is determined for each species from the average of previous experience. In Zimbabwe separate standard seedling recoveries have been established for (a) Orchard seed (b) Select and ordinary seed e.g. for Pinus elliottii the standard seedling recovery (= plantable plants per kg of seeds) is 15,500 for orchard seed and 14,500 for select or ordinary seed (Seward 1980).

The Kilogram Effective Factor (KEF) is the ratio of standard seedling recovery to actual seedling recovery of a given seed lot. It can be calculated from the equation:

Using the previous example (26,600 pure seeds/kg, purity factor 91 % or 0.91, germination factor 82 % or 0.82 and nursery recovery factor 65 % or 0.65) and assuming a standard seedling recovery of 15,000/kg,

The KEF can then be used, in conjunction with standard seedling recovery, to calculate the weight of seed required to raise any given number of plants. The equation is:

For example, to raise 1.5 million plants from the above seed lot

The 116 kg actual seed weight is equivalent to 100 effective kilograms or 100 kg of standard seed. Only when KEF = 1.0 are actual and effective seed weights identical.

In the case of eucalypts and other small-seeded species in which purity analysis is not done, the KEF equation is modified as follows:

For E. grandis, the KEF of a particular seed lot might be

and the quantity of seed needed to raise 1.5 million seedlings would be:


The main testing of a seed lot is done after processing and before storage or immediate despatch to nurseries. If a seed lot remains in storage for any length of time, it is essential to retest the germination or viability before use in case there has been any deterioration with time. Many seed centres retest annually, setting aside a representative sample of the whole seed lot in a small container in advance, so as to avoid recurrent opening of the bulk seed containers. Purity % does not need retesting and moisture content should be retested only if there is reason to believe that leaks have developed in sealed containers.

Special measures have been suggested for recurrent testing of agricultural seed in storage for long-term genetic conservation (Ellis et al. 1980). The authors suggest that, in view of the high value of the stored germ plasm and the need to avoid unnecessary wastage of it in testing, a sequential sampling system may be more economic of material than the standard ISTA method of 4 × 100 replicates. The objective is to detect loss of viability in the early stages of ageing, i.e. as soon as it falls to below 80–90% of the initial rate and to initiate rejuvenation by growing out the seed to produce a new generation. The proposed method would be equally suitable for conservation of forest genetic resources although, because of the much longer generation cycle in trees, conservation as growing plants will be more important than in agricultural crops, as compared with conservation as seed.

Special Considerations for Recalcitrant Seeds of Tropical Rainforests

Many of the tests prescribed in this chapter depend for their usefulness on the ability to keep the bulk of the seed alive during the period while the sample is being tested. Seeds of most species in tropical rainforests are recalcitrant and lose viability at such a rapid rate that any certificate of quality is already obsolete at the time it is issued. In normal practice all recalcitrant seeds are sown in a nursery as soon as possible after collection (Ng 1983).

For purposes of documentation, a germination record is usually maintained, based on a random sample of the seeds being sown. This sample (conveniently 50–100 seeds) is sown separately from the rest, in an enclosure protected by wire mesh, to keep out birds, rats and other pests. The medium (usually soil) should be the same as that used in germinating the bulk of the seeds.

Germination should be monitored daily until no more seedlings emerge and the investigator is satisfied that all ungerminated seeds are decayed. For most rainforest seeds, the end point is reached within a few weeks, nevertheless there may be a significant number of surprises. In view of the paucity of knowledge of rainforest seeds, it is desirable to record the progress of germination from beginning to end, regardless of time. Any arbitrary cut-off point e.g. 21 or 28 days, is too restrictive in this present exploratory phase of rainforest seed investigation.

Previous Page Top of Page Next Page