Species which bear ripe seed in adequate quantities at all times present little problem to the experienced seed collector, but such species are few. Some species bear seed throughout the year but only a little at any one time e.g. P. merkusii in Indonesia (Keiding 1973), and this makes seed collection slow and expensive. In the majority of species the seeding season is concentrated within a few weeks and the collector's objective is to collect as much of the crop as possible within the short period while the seeds are mature but the fruits have not yet fallen or dehisced. Large indehiscent or fleshy fruits may be collected from the ground but even in these cases collection must be done quickly to avoid losses from animals, fungi or premature germination. Prior planning of collecting activities is therefore essential in order to ensure that operations are conducted as quickly and efficiently as possible in the limited time available. Collection in accessible and easily observable plantations or seed orchards reduces the need for careful preparation. On the other hand, collection in inaccessible, multispecific natural forests or sampling a number of different seed sources within a widely spread species calls for very careful planning if trained collecting teams are to operate with the right equipment in the right place at the right time. International seed expeditions encounter special problems, as they often operate in several different countries, each with its own regulations, and they seek to supply the varying needs of many using countries.
Selection of species for planting often presents no problem. In a simple afforestation project, which uses a proven well adapted species and provenance and obtains the seed from a local seed source, the choice is automatic. But not infrequently afforestation objectives change, e.g. emphasis may shift from sawlog to pulpwood or fuelwood production, or unexpected disease problems may arise. In East Africa Pinus radiata featured prominently in planting programmes until it was severely attacked by the needle blight Dothistroma pini in the 1960's. Thereafter its large-scale planting had to be abandoned and the planting programmes of resistant species, such as P. patula and Cupressus lusitanica, were expanded in compensation.
For large-scale collections, data on seed demands by species need to be assembled some months in advance. Most species need a year or more in the nursery. Estimates of seed demand must therefore be made about two years before planting in the field. Only rarely will the seed collectors and the seed users be the same persons. More often, whether collection is done by forest services or by private collectors, they will be supplying the needs of several different users. A centralised organization is needed to solicit demand estimates from the several planting agencies and to consolidate these by species and provenance. Consolidated regional or global estimates of seed demand are much more difficult to compile than those for single countries, but recently attempts have been made to do this for tropical conifers (Nikles 1979) and for western North American conifers (Barner 1978, OECD 1979).
The word “provenance” has been used in somewhat different ways by different authors. In its simplest use, it is “The place in which any stand of trees is growing” (OECD 1974). When applied to seeds, the meaning is frequently extended to include “The area where the mother trees of the seeds were growing”. In cases where seeds are collected from an exotic plantation or “derived provenance” (Jones and Burley 1973), there has been some inconsistency in usage; some authors would define provenance as the place in which the immediate parents were growing as exotics, others would confine its use to the place where the original progenitors were growing in natural forest. Provided that seed origin data give the data on the full pedigree, including both the location of the original natural progenitors, the location of the immediate parents and the location of any intermediate generations, e.g. Cmpt. K2 Elburgon, Kenya (immediate parents) ex Cmpt. 16 Nelspruit, South Africa ex Los Reyes, Hidalgo, Mexico (original natural forest progenitors), the practical forester will not worry as to which is, strictly, a provenance and which is not.
Over the last half century evidence has continually accumulated that, within a botanical species, significant genetic variation in forest trees is frequently associated with geographic differences between the places where they are growing. This is particularly the case where geographic displacement is associated with climatic or soil changes. Thus the word “provenance” is increasingly applied to areas characterized by the genetic nature of the populations growing there rather than their location alone, e.g. “The geographic source or location to which plants are native and within which their genetic characteristics have been developed through natural selection” (Zumer-Linder 1979). For the purpose of seed collection, the ideal provenance, as described by Barner (1975a), would:
Be composed of a community of potentially interbreeding trees of similar genetic constitution (and of significantly different genetic constitution from other provenances).
Be sufficiently large for collection of reproductive material in quantities significant for forest practice.
Be defined by means of boundaries which can be identified in the field.
Although it is not yet possible, in most cases, to delineate the boundaries of natural provenances, there is ample evidence in many tropical as well as temperate species that significant genetic differences exist between them. In the case of derived provenances growing in plantations, boundaries can be defined much more easily and, after a generation or two of conscious selection by man, these “land races” often differ significantly from the original natural provenance.
Increasingly foresters recognise the vital importance of provenance and specify the precise provenance which they need to plant on a given site, not just the species. Even within a single country distinct provenances or races of a species are recognised; some may be morphologically distinct, others which look alike may differ in their adaptability to specific sites. A good example is the large number of provenances of Tectona grandis recognised in India. Seed collecting teams must therefore expect an increasing number of orders with a breakdown by provenance as well as species. This trend is to be encouraged but it does complicate collecting operations, since it is clearly more time-consuming to collect say 20 kg of seed from each of 10 different locations 100 km apart than to collect 200 kg in one area. Another difficulty is in deciding on the limits of a provenance. Often a provenance is named from the nearest village and there is no evidence to show whether there is a significant shift in gene frequencies in the populations 1 km, 10 km or 100 km away from the original collecting point. Some attempts have been made to define the limits of provenances or seed zones of a few north-temperate conifers (Barner 1978) and similar studies have been made more recently on the delineation of provenance regions of Pinus caribaea and P. oocarpa in Honduras (Robbins and Hughes 1983). Provenance regions of Eucalyptus camaldulensis have been defined in terms of the main drainage systems in Australia (Turnbull 1973) but very little work has been done on this aspect in the tropical hardwoods. The need to collect more than one provenance of a species calls for increased care in planning field operations. Some widespread species flower and fruit a few weeks earlier at low than at high elevations and at lower than at higher latitudes. Knowledge of a species' phenological variation in relation to geography will assist the collector to choose the most appropriate sequence of collecting sites so as to extend the total duration of useful collections (Kemp 1975 b).
In contrast with provenances, the boundaries of individual stands are commonly well defined. In many cases the stands are being managed for seed production e.g. by thinning. Often they are in plantations. Seed orchards are a special case, designed for seed production before they are planted and managed continuously for that purpose. The problem with seed stands and seed orchards is thus not one of identification, but that their area may be insufficient to supply all the demands for seed placed on them. If this is likely to occur, it is advisable to require seed users to name second or third choices, in both stands and provenances, in case their full needs cannot be met from their first choice.
Seed users need to define the quantity of seed needed of each species, provenance or stand. For this it is necessary to know the area of plantation to be established annually and the initial spacing to be used, together with an estimate of losses and culls in the nursery, of replacements needed after planting to achieve full stocking, and of the number of germinated seedlings to be expected from each kg of seed sown. An example of the type of calculation is shown in Table 3.1.
Information on planting area and initial spacing is usually available from Plantation Management Plans, while some guidance to germination rates is available in published documents (e.g. FAO 1975a). Whenever possible, local experience on variation between provenances and planting sites should be used to refine estimates based on average conditions. For example, seeds of two provenances of Picea abies weigh, respectively, 6 gm and 12 gm per 1000 (Barner 1981); seed of Eucalyptus cloeziana collected in the moist coastal forests of Queensland averages 100,000 – 400,000/kg, whereas seed from dry inland wood lands averages only 35,000 – 65,000/kg (Turnbull 1983). In Italy it was found that in nursery trials on several eucalypt species the number of plants produced as a percentage of viable seed ranged from 18 % for E. robusta to 46 % for E. camaldulensis (Giordano and Gemignani 1961). Similarly, differences in climate, soil and incidence of pests and diseases can have a big effect on the rate of losses in different nurseries and plantations, whether or not there are any differences in the efficiency of management. So it may be necessary to apply an appropriate “locality correction factor” or “nursery recovery factor” to arrive at an accurate estimate of seed requirements for a particular plantation project. This aspect is discussed further in Chapter 9. Before sending in his final order for seed to a central seed unit or commercial seed merchant, the plantation project manager should deduct those quantities of seed already in stock or likely to be available by collection from older plantations within the project area.
3.1 Provenance Regions for Pinus caribaea and P. oocarpa in Honduras. (source Robbins and Hughes 1983)
3.2 Cone crops of Douglas-fir, Vancouver Forest District, 1935–1974. Over the 40-year period, only eight cone crops have been considered “collectable”, i.e. heavy enough to justify mobilizing large scale collections. The period between collectable crops has ranged from 2 to 8 years. (Canadian/British Columbia Forestry Services)
3.3 Example of a cone-cutter for seed crop estimation on a longitudinal section. (USDA Forest Service)
3.4 Seed content is estimated by counting good seeds on one surface of each of several sliced cones. (USDA Forest Service)
An alternative approach, favoured in some countries, is for the project manager to specify the number of plantable plants he needs to raise and to leave it to the seed officer to decide on the weight of seed to be collected and issued, in the light of germination tests of current seed lots and of known nursery recovery factors. As an example see Appendix 1A, seed form 12.
If good annual crops of seed can be guaranteed, it is satisfactory to order enough seed each year to produce nursery stock for the area scheduled for planting in about two years' time. Storage space can then be kept to a minimum. For species which show periodicity in seeding, however, there are big advantages in acquiring several years' supply in a single good seed year, when the seed will be both cheaper and better. This course of action can only be of practical value if local seed storage facilities are adequate to maintain seed viability over the interval between good seed years. Knowledge is needed for each species or provenance on both the likely interval between good seed years and the rate of loss of viability of the seed under existing storage conditions (Turnbull 1975 a). If heavy seed years are separated by several successive years of seed crop failure and if storage at ambient temperature results in rapid loss of viability, the alternatives may be to build a refrigerated storage room or to change the species.
TABLE 3.1 ESTIMATING SEED DEMAND
| 1. | Species | (a) Pinus kesiya | (b) Tectona grandis | ||
| 2. | Plants per ha | ||||
| (a) | Number planted | 1670 (3 × 2 m) | 1111 (3 × 3 m) | ||
| (b) | Additional for replacing field casualties | ||||
| - percent | 15 | 35 | |||
| - number of plants | 250 | 389 | |||
| (c) | Total requirement - plantable plants | 1920 | 1500 | ||
| (d) | Additional for nursery losses and culls | ||||
| - percent | 20(1 | 62.5(2 | |||
| - number of plants | 480 | 2500 | |||
| (e) | Total requirement - germinated seedlings | 2400 | 4000 | ||
| 3. | Estimated number of germinated seedlings per kg of seed received 3) | 32000 | 500 | ||
| 4. | Basis for estimate in 3. | 1980 crop ex Zambia, dewinged and cleaned | 1979 crop ex Trinidad, with involucres removed | ||
| 5. | Number of kg of seed needed per ha of plantation | 0.07 | 8.0 | ||
| (13.3 ha/kg) | (0.12 ha/kg) | ||||
| 6. | Annual planting area (ha) | 12000 | 5000 | ||
| 7. | Annual requirement of seed (kg) | 900 | 40000 | ||
Seed-bearing of many forest trees is rather irregular from year to year. One year with a heavy crop (a “seed year” or “mast year”) may be followed by one or several years with a poor seed crop or none at all (Morandini 1962). This habit of periodicity in seeding is an important factor to consider when planning seed collecting operations. The collection of seeds in a good year confers a number of advantages. There can be a high intensity of selection of seed bearers, the cost of collection is lower, due to the concentration of the crop, and the seeds will usually be of higher germinative capacity and will retain their viability longer than those collected in a poor seed year (Turnbull 1975 a, Seal et al. 1965). Damage from insects affects a smaller proportion of seeds in a good than in a bad seed year. A heavy seed crop usually reflects a previous heavy production of pollen, to which all or most trees in the stand have contributed. Collection in a good seed year therefore conserves a higher proportion of the genetic diversity among male parents than collection in a bad year which follows pollination from only a small number of trees.
Periodicity is well documented for many temperate conifers. For example on average Pinus sylvestris bears an abundant crop every 2 – 3 years and Pseudotsuga menziesii every 4 – 6 years in the UK. Since the period between the good crops is not regular, a general rule of collecting three years' sowing requirements whenever a species bears a heavy cone crop is there recommended by Seal et al. (1965).
Periodicity in tropical species is less well documented. Irregular mast years in Triplochiton have a great influence on the regeneration, or lack of regeneration, of that species (Howland and Bowen 1977), although periodicity in pests and diseases (the weevil Apion and the fungal smut Mycosyrinx) may play as big a part in seed production as periodicity in flowering (Jones 1975). Poor seed years have been recorded in Pinus caribaea and P. oocarpa (Kemp 1973), and P. merkusii (Keiding 1973). In other species periodicity is not marked. Tectona grandis has generally good flowering each year, although exceptionally good seed years are observed in some localities on a three or four year cycle (Murthy 1973). Gmelina arborea starts to seed early, from age 3 in the Philippines to age 7 in Nigeria and usually produces regular abundant crops (Greaves 1981) but poor seed years have been recorded, at least in some provenances (Lauridsen 1977). Pinus kesiya bears abundant crops every year within its indigenous range and as an exotic if planted in the appropriate climate (Armitage and Burley 1980). Cassia siamea, Acacia mearnsii, Cupressus lusitanica and ornamentals such as Delonix regia and Jacaranda mimosaefolia are other species which can be expected to flower and fruit profusely each year. Periodicity may vary considerably between species within the same genus. Among eucalypts, E. grandis, E. saligna and E. camaldulensis usually bear heavy seed crops every two to three years, while E. gomphocephala and E. maculata only seed heavily at longer intervals (Turnbull 1975 e). Dipterocarps in Malaysia have heavy seed years at unpredictable intervals of one to six years (Ng 1981). Periodicity and flowering patterns of eucalypts can change when they are grown as exotics. Eucalyptus maculata and E. citriodora bear much larger crops more regularly when grown in plantations.
Even in good years, flowering may vary substantially from one locality to another. Sometimes individual trees of a stand are on different cycles, some flowering abundantly one year, others in the next (Krugman et al. 1974).
For those species which are known to exhibit periodicity in flowering and fruiting, it is highly desirable to visit the stands to be collected well in advance of the fruiting season, in order to assess in which of them the next seed crop promises to be heavy enough to justify the cost of collection. Too little is known about which external factors have a decisive affect on flowering to allow prediction of future seed crops on the basis of climate. Estimation of the crop can best be done by counting flowers or young fruits on a sample of the trees in the collection stands. Assessment of the abundance of flowering can give a preliminary estimate of the potential seed crop, but may be misleading if there are subsequent severe losses e.g. from insects, wind or poor pollination. In Eucalyptus regnans trapping studies indicated that only about 15 % of flower buds and 30 % of flowers can be expected to develop into mature fruits (Turnbull 1975 e). In species such as the pines, which take two years from pollination to ripen their fruits, a count of one year old cones can give a useful indication of the next year's crop (Stein et al. 1974), which can be confirmed by further inspection a month or two before collection is due to begin. Recurrent personal inspection of the future crop is the ideal and presents no problem if collection is carried out in seed orchards or easily accessible plantations or natural forest. It may be difficult or impossible for collecting teams working in inaccessible areas or for international expeditions working in several countries. In such cases the team leader may have to rely on reports from an experienced correspondent or on estimates made during the previous year's collections. If reliable local information is not available, a special reconnaissance may be justified in advance of the expensive collecting expedition.
Where the main seed collecting areas are located in less accessible places, it is useful to maintain permanent phenological plots in easily accessible locations within the same forest type, to act as biological indicators. Such plots should be monitored regularly so as to obtain a record of the timing and intensity of flowering and fruiting. The performance of the plots will indicate to collectors the best times to go into the less accessible areas of the district, to check for flowering. The boundaries of a phenological district that can be usefully served by a phenological plot must be defined by experience. In Malaysia, a sample of 86 Dipterocarp trees in an artificially established arboretum at the Forest Research Institute, Kepong, is monitored monthly or fortnightly and the percentage of trees flowering in a month or year is used as an index of Dipterocarp flowering for that month or year (Ng 1981). This index gives a fairly reliable indication of Dipterocarp phenology within the State of Selangor (about 8000 sq. km) in which Kepong is located. Since the phenological plot in Kepong is within 10 minutes' walking distance from the researchers' offices and laboratories, a lot of time and money is saved which would otherwise have to be spent on transportation and organization of field trips.
Where it is necessary to count the fruits or cones, binoculars or telescopes are an essential aid. They should be of high optical quality. In binoculars a wide field of view should be combined with only moderate magnification; a minimum of 50 mm aperture and magnification x7 or x8 are suitable. The normal method is to make counts on a representative sample of seed trees dispersed throughout the whole seed source. It is necessary to take the sample from within the stand, because the perimeter trees always fruit more heavily than the trees within the seed source (Seal et al. 1965). Counting may be done from the ground (Seal et al. 1965) or by climbing neighbouring trees (Machaniček 1973). Fruits are counted from one side of the crown only and the number counted is converted to an estimate of the total crop for that sample tree by a correction factor which varies according to the species and the abundance of the crop. In Czechoslovakia (counting from climbed trees) a factor 1.6 is used for Abies alba, which is constant irrespective of the size of crop, because the clustering of the cones near the top of the tree makes counting in this species reliable. For Picea abies the factor varies according to the average number of cones counted per tree; for 1 – 40 cones it is 1.4, for 41 – 70 cones it is 1.8 and for more than 70 cones it is 2.5 (Machaniček 1973). In the UK (counting from the ground) a factor of 4 is used for Pinus, Larix and Pseudotsuga, while for species with very numerous small cones the practice is to scan say one tenth of the crown from one side and use a factor of 20 (Seal et al. 1965).
The number of sample trees used for cone counting varies according to the size of the stand. In the UK five trees are sampled in small seed sources less than 0.5 ha and the number increases progressively to 20 trees for seed sources of more than 4 ha (Seal et al. 1965). Czechoslovakia uses a series of plots, in each of which about 5 dominant trees are climbed to scan a further 10 – 15 neighbouring trees (Machaniček 1973). In Tasmania extensive eucalypt seed sources are sampled at the rate of 1 tree per ha (Turnbull 1975 e).
The results of cone or fruit counting are applied to the stand and expressed as a numerical value on a scale running from total crop failure to exceptional seed years (Morandini 1962, Turnbull 1975 a). With experience it may be possible to define the criteria for an economically collectable crop in quantitative terms e.g. for Pinus sylvestris in the UK a minimum of 25 seed trees per ha each bearing at least 300 – 400 cones is specified (Seal et al. 1965). For conifers in Arizona and New Mexico Schubert and Pitcher (1973) defined “Few” cones per tree as 1 – 20, “Many” as 21 – 160 and “Loaded” as over 160 cones per tree. Quantitative classifications of this type will clearly vary to a great extent according to species, provenance and site conditions.
More often use is made of qualitative scoring assessments which rely on the experience of the assessor. In Washington and Oregon (1982) five ratings are used, as shown below:
RATING EXPLANATION
| Other than true firs: | ||
| 5 | Heavy | - Good crop of cones on all exposed crowns of most trees. |
| 4 | Medium | - Good to medium crop on ¾ of exposed crown on most trees. |
| 3 | Light | - Good to fair crop on ½ of exposed crown of ½ of trees. |
| 2 | Very light | - Some cones on some trees. |
| 1 | Failure | - No cones to a few scattered on a few trees. |
| True firs: | (upper ⅓ of crown) | |
| 5 | Heavy | - Good crop of cones on most upper branches of most trees. |
| 4 | Medium | - Good to medium crop on most trees. |
| 3 | Light | - Few cones on many trees. |
| 2 | Very light | - Few cones on scattered trees. |
| 1 | Failure | - Nonexistent |
A rating of 4 or 5 is good prospect for all pickers.
A rating of 3 has possibilities for more experienced pickers.
A rating of 1 or 2 is poor prospect for all pickers.
Each year the State Forest Service publishes estimated average cone crop ratings by species and geographic area, for the benefit of individual pickers. They are based on surveys of a number of different stands and the average for each area is expressed to one place of decimals. For example in 1972, a poor year, the best rating was 2.5 for Tsuga heterophylla in the Western Cascade of Oregon, while 1.0 or total failure was recorded for some species in more than one area.
In Tanzania a four class scale is used and seed crop estimates are made twice, one at flowering and the other about a month before seed collection (Pleva 1973). The classes are:
| 0 | - | no seed crop. Trees without flowers and fruits. |
| 1 | - | weak seed crop. Flowering and medium size seed crop on free growing trees and trees on free borders of stands. |
| 2 | - | medium size seed crop. Flowering and very good crop on free growing trees and on free borders of stands, trees within the stands bearing crop at the top of crowns. |
| 3 | - | very good seed crop. Flowering and very good crop on most trees. |
In Sweden forecasts of the cone and seed germinability of Pinus sylvestris and Picea abies have been made annually for about 80 years. Estimates are given separately for the different combinations of latitude (in 1° steps) and altitude (in 100 m steps) (Simak and Remröd 1976).
The methods described above provide an estimate of the cone or fruit crop. It is necessary to relate this to seed production by examining the contents of a sample of the fruit crop. Fruits may develop normally to maturity whether one or one hundred of the contained ovules have been successfully fertilized and undergone normal development; in parthenocarpous species fruits can mature without containing any sound seeds at all. The number of fruits is therefore not always a good guide to the number of seeds.
The method generally recommended is to cut cones or fruits lengthwise and count the number of seeds which can be seen on one cut surface (Morandini 1962, Seal et al. 1965, Stein et al. 1974). Special cone-cutting knives have been designed for this purpose. One or two cone samples from each of 20 to 100 trees in an area are suggested for southern pines in the USA (Wakeley 1954), 5 or 10 cones from each of 10 trees in the UK (Seal et al. 1965). Only normal seeds should be counted. Underdeveloped seeds which often occur at the top and base of cones are not included (Stein et al. 1974). The number of full seeds counted which denote a good crop varies according to species, for example 6 or more indicates a good crop in Pseudotsuga, 14 or more a good crop in Picea sitchensis (Douglass 1969, Stein et al. 1974). The relationship of the total number of full seeds per cone to number of full seeds exposed on a cut surface is known for some species, for example a factor of x4 or x5 is appropriate for Pseudotsuga in the western USA (Greathouse 1966).
The average number of seeds per fruit for many tropical species is not known and needs to be established under local conditions; they vary from one seed per fruit in e.g. most Dipterocarps to several hundred per fruit in Anthocephalus. With multiseeded fruits, it is likely that the number of seeds developed will vary in accordance with climate, soil fertility and the age of the parent trees. The first crops borne by young trees nearly always contain fewer sound seeds per fruit than those borne by the same trees when fully mature.
Examination of a sample of seeds in the fruit serves the additional purposes of indicating the state of development or maturity of the seeds (see the following section) and the incidence of damage by pests or diseases.
A final decision on whether the seed crop is sufficiently heavy to justify collection in a particular year should be dependent on both the rating of the fruit or cone crop and the results of the cutting test for determination of full seed content. An example of a recording form which combines these two assessments is shown in Appendix 1C11.
Some species in the tropics carry some ripe seed at all times of the year. Even in these there is often a period of maximum seed production, when collection will be cheapest and seed quality highest. In other species, and especially in the temperate zone with its marked distinction between summer and winter, ripe seed is borne for a limited period, often during the autumn. For many species there is good information on average dates of the seeding season, but these averages may not be sufficiently accurate for planning collection in a particular year. The period between seed maturation and seed dispersal is often short, whereas the effects of climate in a given year may displace the dates of seeding by several weeks from the average. In the temperate zone an early spring and dry summer can cause very early seed ripening, while strong, dry winds cause rapid dispersal of the ripe seeds. Cool, wet weather, on the other hand may delay ripening and dispersal by weeks or months (Stein et al. 1974). In the dry tropics there are similar annual variations in the dates of the onset of the dry season and of the rains. It is therefore necessary in each year to check the correct timing of collection by examination of the crop itself.
The reconnaissance of the size of the seed crop, made 1 – 2 months before seed collection as described on p. 28 will also give some indications as to how the seeds are maturing. Conclusions from it should be twofold e.g.: -“Stands A, B and C: Very light crops, not worth collecting this year. Stands X, Y, Z: Good crops, seeds probably mature in 4 weeks' time”. A final check on the ripeness of seed must, however, be done at the time of collection.
The humid tropics present special problems because seasonality effects are usually subtle or absent and the period of maximum seed production is uncertain. After detection of flowering in a stand from which it is desired to collect seeds, it is important to carry out periodic reconnaissance to check on the progress of fruit maturation. An efficient schedule of reconnaissance requires prior knowledge of the length of time between anthesis (flower opening/pollination) and fruit maturity. Among Malaysian trees, the period from anthesis to fruit maturity varies from 3 weeks for Pterocymbium javanicum to 11 months for Diospyros maingayi (Ng and Loh 1974). In the exotic Brazil nut, Bertholletia excelsa, the period is 15–16 months (Lambourne 1930). In Malaysia it is recommended that if the maturation period is X weeks, the development of the crop should be checked at 1/2 X and 3/4 X weeks after flowering. A fixed and arbitrary schedule of, say, once a month, will result in the collector being too late for a fast-maturing fruit like Pterocymbium javanicum and too wasteful of effort in the case of a slow-maturing fruit like the Brazil nut.
Apart from the exceptional case of deliberate collection of immature seeds (discussed below), seed collectors need to be able to time collection for the period when seeds (but not necessarily fruits) are fully ripe but before they are dispersed by fruit dehiscence or fruit consumption by animals. To achieve this aim collectors must be able to distinguish ripe from unripe seeds. Several different methods have been used for the recognition of seed maturity. None of them work perfectly on all species and a good deal of experience or research is needed to determine the best method, or combination of methods, for a hitherto unfamiliar species. They may be divided into those which are of direct application in the field and those which need laboratory equipment. The latter may be valuable in providing a check on the field methods, but are unlikely to be of practical use to the collector unless the collection site happens to be close to the laboratory, which may be the case for some seed orchards.
(a) Dry weight. The most generally accepted measure of maturity is the time when the seed has reached its maximum dry weight, a point called physiological maturity. This means that nutrients are no longer flowing into the seed from the mother tree (Harrington 1972). Maximum fresh weight does not indicate physiological maturity because the maturing seed begins losing water while nutrients are still being accumulated and biochemical processes are continuing.
Recurrent dry-weight determination of a series of seed samples can be made and the results extrapolated to the remaining crop, but this method is slow and therefore seldom used.
(b) Chemical analysis. Biochemical changes take place as seeds mature but relatively little is known for most species. Chemical indices of seed maturity have been determined in a few species e.g. content of crude fat and protein-nitrogen, which increase five and four times respectively from immaturity to physiological maturity, are the best chemical indices for Fraxinus pennsylvanica. But they have no advantages over an examination of the embryo and the colour change of the fruit, and the extra trouble to perfom the analyses does not appear to be justified (Bonner 1973 b). Rediske (1969) found that Pseudotsuga seeds were physiologically mature when the content of reducing sugars dropped to 14 mg/g.
(c) X-ray radiography. The examination of the development of the embryo and endosperm of sample seeds by means of X-ray radiographs is a quick and relatively straightforward method of assessing seed maturity, provided that suitable facilities and skilled technical staff are available (Turnbull 1975 a). The technique has been used successfully for Tectona (Kamra 1973) and a number of other tropical species (Kamra 1974), as well as for temperate species such as Pinus strobus (Wang 1973). It has the disadvantage of requiring relatively expensive equipment and relies heavily on the judgment of the seed analyst for reliable results (Turnbull 1975 a).
(d) Moisture content of fruits. Water loss of maturing cones and fruits occurs in many species and is closely related to the maturity of the seed. Seeds of Picea glauca are considered ripe when moisture content falls below 48 % (Cram and Worden 1957), of Larix decidua at 25–30 % (Messer 1963, 1966) and of Pinus sylvestris when it falls to 43–45 % (fresh weight basis) (Schmidt-Vogt 1962, Remröd and Alfjorden 1973). However, determination of moisture content by drying in an oven suffers the same disadvantage of slowness as determination of dry weight.
(e) Specific gravity of fruits. As moisture content of fruits and cones decreases with maturation, so does specific gravity or density, the ratio of unit weight to unit volume, decrease. Unlike moisture content, it is not too difficult to determine approximate specific gravity in the field by flotation in liquids of known specific gravity. Specific gravity indices of maturity have been established for cones of a number of coniferous species, and the cone to be tested is placed in a liquid in which it will float if mature and sink if immature (Stein et al. 1974). Various mixtures of kerosene (SG = 0.80), light SAE 20 motor oil (SG = 0.88) and linseed oil (SG = 0.93) have been used to prepare flotation liquids having a designated specific gravity. Tests must be made immediately after cones are picked from a tree. Specific gravity indices have proved reliable for some temperate conifers e.g. an S.G. of 0.74 for Picea glauca (Cram and Worden 1957), but not for several southern hardwoods in the USA (Bonner 1972).
(f) Examination of seed contents. Examination of seed contents exposed by cutting open fruits or cones lengthwise can be a reliable and simple method of assessing seed ripeness, provided the operator is experienced. Most embryos and endosperm pass through an immature “milk” stage, followed by a “dough” stage when the tissue becomes more firm. Mature seeds have a firm white endosperm (where present) and a fully developed firm embryo (Turnbull 1975a).
(g) Colour of fruits or cones. Colour changes in fruit or cone provide a simple and, in some species, reliable criterion for judging seed maturity, but the operator must be experienced in the characteristics of the species concerned. In common with the specific gravity method, it involves no destruction of the seeds in the sample examined. Colour changes are usually from the green of the immature fruit or cone to various shades of yellow, brown or grey, and this may be accompanied by hardening of cone scales or of the pericarp of dehiscent or woody fruits. Since the seed normally matures before the fruit, it is advisable in some species to time collection at an earlier rather than a later stage of the colour change. Colour change was found to be the most reliable indicator of maturity for general practice in several southern hardwoods in the USA (Bonner 1972). It has also given good results in a number of temperate conifers. In Malaysia Tamari (1976) found that the best results were obtained by timing collection of Dipterocarp fruits when the wings turned brown but before the fruit itself changed colour.
In Thailand cone colour is used as a guide to optimum time of collection of pines, but differs according to species. In Pinus kesiya collection starts when cones have hardened and the colour is changing from green to brown in proportions of 50 : 50. In Pinus merkusii optimum time of collection is reached when the majority of cones are brownish and some have started to open (Granhof 1975). Trials with the Zambales (Philippines) provenance of P. merkusii have shown not only that extraction is a much more lengthy and expensive operation with green than with brown cones but also that the seed extracted has lower germination rate (Gordon et al. 1972). Experience with P. caribaea in Honduras is similar (Robbins 1983a).
Abscission and shedding of fruits is usually a sign of fruit maturity and it might be assumed that it also indicates a high content of sound, mature seeds. This is not always the case. The first seeds or fruits which fall naturally are often of poor quality (Morandini 1962), in which case it is advisable to reject them and to postpone collection until the peak and the latter half of the season. In Thailand fruits of Tectona grandis start to be shed in March but observations have shown that the most viable fruits are the last to be shed, so collection starting only in April is recommended (Hedegart 1975). The first fruits of Dipterocarp species that fall upon ripening are usually defective and collecting should be delayed until the greater portion of the fruit has fallen (Seeber and Agpaoa 1976).
It is general practice to collect seeds when they are mature, because they have a higher germinative energy and a greater longevity in storage than immature seeds. An alternative method is to collect fruits prior to ripening and to store them in relatively cool, well ventilated conditions which permit afterripening of the seeds within the fruit. It has shown promise on a research scale in a number of species.
There are several reasons for the interest in developing techniques for artificial ripening (Turnbull 1975 a). They are: -
to extend the collection season. The short period between seed maturity and dispersal may place an excessive demand on the availability of seasonal labour and in some areas unfavourable weather conditions in the collection period may aggravate this situation. Lengthening the period available for collection permits better organization of the collections and allows skilled personnel to pick more of the crop. It can be particularly valuable in research, where a large number of seed lots have to be collected from widely scattered localities. Griffin (1974) used the technique in his study of Douglas fir provenances.
to avoid damage to the seed crop by insects and other pests. Insects, birds, rodents and other pests frequently damage or destroy seeds and fruits when they have reached maturity. Early collection may be one method of avoiding these losses. Damage and deterioration of seeds is usually worse during the period while they lie on the forest floor than during any other stage in their history; any reduction in this period will improve their subsequent viability and longevity in storage.
to salvage immature seed collected inadvertently. Untrained collectors of seeds often begin picking fruits and cones too early in the year before they are fully mature. Artificial ripening provides a method for handling this material.
The development of techniques for after-ripening of immature seeds will require more research, before they can be applied to a wide range of species. However, where a problem of rapid dispersal or of seed pests exists and provided the earliest time for safe collection of immature fruits can be established, such techniques can be very beneficial. Some successful examples are mentioned on pp. 90–91.
If it can be assumed that the seed collector has received clear directions from the seed user as to which species and provenances and, in some cases, from which stands he is to collect, it is still his responsibility to select the individual trees for collection. Criteria will vary considerably according to whether the collections are large-scale for afforestation projects or small-scale for research purposes.
Identification of species presents no problems in monospecific plantations, but is essential and may be difficult in mixed natural forest, especially where very similar species of the same genus occur mixed together as happens with pines in Mexico and Central America, eucalypts in Australia and dipterocarps in S.E. Asia. Unless identification is certain, it is often advisable to collect herbarium specimens as well as seed.
In large-scale collections emphasis is on collecting as much seed as possible, as quickly and cheaply as possible, rather than on very careful selection of the parent trees. It is, nevertheless, essential to avoid collecting seeds from very poor phenotypes or seeds which prove to be empty or non-viable. Useful guidelines are listed by Stein et al. (1974), and form the basis of the following: -
Collect seed only from healthy vigorous trees of reasonably good form that are making average or better growth.
Where possible, collect from mature or nearly mature trees. Overmature trees should be avoided, since seeds from them may be of low viability.
Avoid isolated trees of naturally cross-pollinating species, since these are likely to be self-pollinated. Seeds are likely to be few, of low viability, and any seedlings produced are frequently weak or malformed.
Avoid collecting in stands containing numerous poorly formed, excessively limby, off-colour, abnormal or diseased trees.
Often it will be necessary to compromise between seed production and phenotypic appearence. No seed should be collected from excessively coarse-branched, vigorous “wolf” trees, even though they often bear a large crop, while trees of exceptionally good form sometimes bear so little seed that they do not warrant the trouble of collection. The bulk of seed will come from trees which are “average or better than average” in both form and seed production.
Although few studies have been made of the reproductive biology of tropical trees, the occurrence of some species at a very low stocking (less than one every km2) suggests that these must be naturally self-pollinated. Seed collection from such trees is free from disadvantages associated with collections from isolated trees of natural cross-pollinators.
In small-scale collections for research, selection of trees will depend on the precise objective of the planned research. Provenance research is now receiving a good deal of attention in many countries. Advice of IUFRO on provenance seed collections includes the following recommendations on collections from individual trees (FAO 1969): -
Collect from not worse than dominant and co-dominant trees of average quality, within “normal” rather than “plus” stands. Collections from superior phenotypes, if made, should be kept separately.
Collect from a minimum of 10 trees, preferably from 25 to 50 in the stand. If the stand is very variable, increase the number of trees. Record the number of trees and the approximate percentage which they form of the stand.
Seed trees to be at least seed fall distance apart from each other. A distance of 100 m has been adopted for Pseudotsuga. This is to reduce the risk of collecting from half-sib parents. In Australia a minimum distance of twice tree height is used as a practical rule of thumb (Boland et al. 1980).
Individual seed trees to be marked.
Collect equal numbers of cones, fruits or seeds per tree.
In normal first stage provenance collections, seed from individual trees may be mixed together. If special studies on individual genotypes are to be done, seed from each tree should be kept separate.
Foresters are interested in the variation within populations and provenance as well as the variation between them. In the case of exotics, one approach for the introducing countries is to study between provenance differences under local conditions first, and to investigate variation between individuals in the best provenance, by means of progeny trials, only at a later stage after the provenances locally best adapted have been identified. If progeny trials are the object of seed collection, it is essential to keep seed from individual trees separate at all stages of collection, transport, processing, nursery and field planting.
The preservation of the identity of individual trees through the collection and extraction phase often requires considerably more effort than bulking the collections. If the effort is made, then certain advantages accrue, as listed by Turnbull (1975b):
It permits biosystematic study of genetic variation both within and between populations. McElwee (1969) states that seed collections must be kept separate by individual trees throughout the test from collection to outplanting, as provenance tests are weakened by combining seed within a stand, making it impossible to distinguish between seed source and individual variation. However, in large provenance tests with many individual trees, it may be beyond the resources of the investigator to retain the identity of the parents.
It is possible to manipulate to equalize the amount of viable seed from each tree in the provenance mix if the seed must be bulked prior to sowing.
It is not always possible to detect trees bearing hybrid seeds in the field. This is particularly the case with eucalypts. If seedlots are kept separate then, following the raising of small samples from each tree, any showing evidence of hybridization can be eliminated before the main trial is established.
In seed collection in clonal seed orchards, the unit of identity to be kept separate is often the clone rather than the individual ramet. In Zimbabwe the practice of keeping individual clonal seedlots separate has been followed for many years and it is considered that it more than justifies the additional costs and efforts involved over bulked orchard collections. The advantages can be summarized as follows:
Maintaining individual clonal identities at all stages from collections through to storage enables one to act with the minimum of delay on most information as it becomes available. The more obvious areas where action can be taken quickly are when rogueing becomes necessary or when susceptibility to pests or diseases becomes apparent, and the undesirable seedlots can be isolated or discarded.
Seedlots can be made up and supplied to suit specific sites, making use of the most recent information on genotype/environment interaction from progeny test analyses.
In conjunction with (b) above, and using the test results of each clone, composite seedlots can be made up in such a way as to give equal clonal representation in the final planting stock, avoiding the dominating effects that some clones frequently have in a bulked seedlot.
In general, it increases the options available to the user.
Collections are also made for the purpose of attempting to conserve the gene pool ex situ either as seed in long-term storage or in planted conservation stands. Because exact knowledge of gene frequencies in the indigenous populations is largely lacking, collecting for gene conservation must be based mainly on commonsense. Similar methods as for provenance collections are likely to be appropriate, with the exceptions that: -
A somewhat larger number of trees per genepool should be sampled. Estimates are in the range or 50 to 100 (e.g. Nikles 1974, Marshall and Brown 1974).
The sample should be a strictly random one and include poorer than average as well as better than average trees, in order to capture as much as possible of the total genetic variation. The only restriction on this principle is the impossibility of sampling trees which are bearing no seed.
An additional measure to ensure the fullest possible genetic diversity within the seed collected is to collect in a better than average seed year. The better the seed year, the better the representation of male parents contributing pollen as well as of female parents supplying seed.
The quantity of seed to be collected from each provenance will usually be larger, because the recommended area for a conservation stand, 10 ha, is much greater than the total area of one provenance in a single provenance trial.
One part of planning is the timely assembly of clear information on the nature and magnitude of the seed collection tasks - number of species and provenances, seed quantities, location of stands, best dates to collect etc., as described above. The other part is to select and assemble the resources needed to do the job. Details of the various resources which may be useful are discussed in subsequent chapters. During the planning stage, the leader of collecting operations needs to check the preparations for field work under the following headings:
Organization of collecting teams. Known or estimated output of collecting teams needs to be related to the quantity of seed, number of stands and length of season, in order to determine the required number and size of teams. For example provenance collections of Pinus kesiya totalling 3000 kg of cones from 16 stands can be completed in 30 days by one team in Thailand (Granhof 1975). If the full season is estimated at 45 days and there is a demand for a total at 9000 kg of cones from the same 16 stands, two teams will be needed to do the job. If planning can be done sufficiently in advance, it may give an opportunity to train additional climbers if they are needed. It is desirable to have at least one tree-climber on the permanent staff, who can be responsible for looking after climbing equipment and for training new temporary climbers. In the field the climbers should be organized in small teams with a foreman in charge of each. In Honduras a team of 6 pairs (with one climber and one ground assistant or anchorman in each pair) has been found to be a suitable size (Robbins et al. 1981).
Organization of transport. Collecting teams need to cut to a minimum the time spent in moving between one site and the next. Transport must be available where and when it is needed. If necessary, extra vehicles may be temporarily hired. In roadless country, advance arrangements may be needed to employ extra unskilled workers to assist in carrying equipment, tents, etc.
Organization of equipment. Choice of equipment will vary greatly according to local conditions. The steeper and less accessible the terrain, the simpler and lighter should be the equipment. Whereas highly mechanised equipment such as tree shakers or mechanised platforms may be appropriate in large seed orchards on flat land, only light portable equipment is practical where natural stands are 4 – 6 hours' walk from the nearest road (Granhof 1975). Apart from collecting tools, safety clothing, first aid equipment and plenty of bags and sacks should be provided.
Organization of records. Meticulous recording and labelling is essential to good collecting. Appropriate labels and forms need to be designed well in advance and printed in adequate numbers (Sompherm 1975a). For examples see Appendix 1.
Organization of permits. These are not normally required for forest services collecting in government forest reserves, but may be needed when collecting on private land, in National Parks and special reserves, or in another country. Even if formal authority is not needed, it is often advisable to inform local communities of proposed operations in advance.
Organization of seed extraction. Arrangements for rapid movement of fruits from collection site to extractory may be needed, involving the advance organization of transport. The seed extractory staff must be advised when to expect the fruits. If some preliminary sun-drying of fruits in the forest is planned, polythene sheeting or tarpaulins will be needed.
An example of the logistics of collecting Pinus caribaea seed is given in Appendix 5.
Some species are currently of greater importance as plantation trees in exotic situations than they are in the source countries. Examples are Gmelina arborea, Pinus radiata, many sub-tropical and tropical pines and many eucalypts. When the natural range of such species extends through several different countries and the seed is in demand by many introducing countries, international action may be the most effective way of organising collection and distribution of seed. This is particularly true in the case of collections for provenance trials and for the establishment of ex situ conservation stands, seed stands and pilot plantations, for which only small or moderate quantities of seed are needed but sources must be precisely identified and fully documented. Examples of international collections of this type are those organised by IUFRO for western North American conifers (Fletcher and Barner 1978), by CFI Oxford for Central American pines and hardwoods and by the DANIDA Forest Seed Centre for Tectona and Gmelina.
Kemp (1975 b, 1976) has described those problems in seed collection which are accentuated in the international context. These include especially the difficulty of obtaining accurate information in advance, on which to base the detailed plan of operations, and problems concerned with crossing international borders, customs regulations, language differences and so on. There are many uncertainties regarding access, travel and transport of material which may only be resolved when the collection is actually in progress. This is particularly true in many tropical countries, in which precise information on the natural distribution, variability, flowering and fruiting times, etc. of the species concerned may be incomplete or entirely lacking. While this makes precise planning of operations more difficult, it also calls for long and careful consideration of the many possible situations that may be actually encountered and the preparation of contingency plans to meet such situations, as well as to allow for unknown emergencies that may arise. Some particular aspects which must be considered in international expeditions are briefly stated below.
(1) Objectives. Because an international expedition is likely to be undertaken on behalf of many countries, it may be attempting to meet several different objectives at the same time, such as collection of seed for provenance trials, bulk seed for large plantations of selected sources, collection from individual phenotypes and so on. Expeditions involving a lengthy period of travel in distant areas are very expensive and must therefore be used as fully as possible to meet different needs, but nevertheless these needs may to some extent conflict. Since the time available for collection is always limited, a choice may have to be made between spending longer at an individual site, for detailed sampling or bulk collection, and covering more sites more quickly. For this reason, the objectives must be clearly defined in advance and an order of priority established in case it is necessary to choose between them.
(2) Local Regulations. Most countries have regulations governing collection, export, introduction and perhaps movement of seed. Official permits may be required for any of these procedures and failure to obtain the necessary documents may seriously delay current operations as well as jeopardizing future collecting by international teams. Similarly the expedition staff may require personal documents, such as entry visas, work permits and international health certificates. Some equipment may also be subject to importation restrictions, particularly firearms and, perhaps “walkie-talkie” radios, if these are to be used.
(3) Involvement of local staff. International expeditions can benefit enormously from the active involvement of local personnel. They can assist in interpretation and their knowledge of local geography and customs can be extremely valuable. They may also be in a position to reconnoitre seed crops in advance of further collections in subsequent years. In return, arrangements may be made for travel expenses of local personnel to be borne by international funds and for the host country to receive a part of the seed collected. Such arrangements should be agreed on both sides in advance.
(4) Equipment. Decisions on the most appropriate equipment to use, and the essential items to be supplied to the expedition in advance, are more difficult without previous knowledge of the area and local conditions. Transport of bulky equipment by air is expensive, but equally a delay in the start of operations once the expedition is in the field, due to lack of needed equipment, can be very costly of both funds and the limited time available for collection. A list of equipment, proposed by Kemp (1976), is reproduced as Appendix 6.
(5) Timing of operations. It is preferable to collect only in a year of abundant seed production, since this offers much greater freedom of action in the choice of stands and trees, as well as the possibility of obtaining more seed for a given cost. However, international expeditions must be prepared far in advance of the time when crop assessment is possible and the complex arrangements involved make it very difficult to alter plans at a late stage. Even in a year of abundant seed production, best results can only be achieved if the actual season for collection is accurately known, and this also is more difficult for an international expedition, based elsewhere, to determine in advance.
(6) Field Records. Because of the difficulty and high cost of obtaining data once the expedition has left the area, particularly if it might involve returning there to make fresh observations, accurate and full records of the sites and collection details are most important. See Appendix 1 B for examples.
