SEED COLLECTION TECHNIQUES

 by D.J. Boland

Seed collection involves a wide range of activities from the initial planning of collections through to seed extraction from the fruit. The term seed collection is not particularly appropriate as it is usually the fruit that are first collected from which seed is later extracted. The aim here is to cover briefly a range of activities connected primarily with research collections in natural stands such as planning the trip, timing of collections, sampling of trees, seed collection methods and extraction. Most available published information relates to seed collection of temperate species but increasingly attention is being given to seed collection in tropical species for which a lot less is known.

This account on seed collection techniques ranges widely from Australian to world techniques to provide a broad account of problems and solutions to seed collection in numerous species. It is hoped that the lessons and knowledge gained can be adopted or modified to suit conditions experienced by seed collectors in East Africa.

PLANNING AND TIMING OF COLLECTIONS

Considerable effort should be put in to the planning and timing of seed collections. A successful mission is usually one which has been well planned. Some typical specific questions to consider during the planning stage include:

SAMPLING

In any collection programme the collector must consider the range of available strategies that could be adopted to sample a species over its natural distribution. In this regard the two most important concepts to the seed collector are the concepts of how to sample populations and how to sample trees within populations. The first concept involves provenance which can be defined as the source of the material collected and is usually designated by the place name of the collection site. In this account we shall consider seed collections for species introduction trials and for species provenance trials.

1.     Species introduction trial

Edwards, (1963) suggested that for an unknown species several provenances should be tried, including at least one from:

(a)     Optimum site. A sample taken from the part of the natural distribution where a species exhibits its optimal development can be valuable but because of genotype-environment, interaction, the phenotype in the natural stand may bear little relationship to the performance of the tree under different conditions. In western America the largest and straightest Pinus contorta trees are found at high altitudes in the Sierra Nevada ranges, the coastal low altitude trees are small and frequently of poor form. However, plantations of this species in Europe, New Zealand etc., have proved that it is the interprovenances which give the most productive forests.

(b)     Matching site. The sampling of 'matching' sites involves a study of the environmental factors throughout the species natural range and attempting to locate an area,or areas, where there is a similarity with the site where the trial will be established.

(c)     Marginal site. Populations near the boundaries of the natural distribution are often growing under environmental conditions which differ from those where the main populations occur. Other, often severe, selection pressures may operate on the margins which can produce populations physiologically and morphologically very different from the other populations. Marginal populations which tolerate a greater frequency of droughts are likely to be more adaptable to drier areas than the other populations. The idea of marginal sites assumes that the limits of distribution of a species are the result of limiting climatic or edaphic factors and that populations close to the boundary are those which have had maximum selection intensity applied to them through the limiting conditions. The sampling of marginal sites is very important when determining the extent of genetic diversity in a species.

The sampling of 'optimum sites', 'matching sites' and 'marginal sites' is generally carried out in an attempt to save resources such as time, labour and land which are required for complete provenance trials. It is applicable if there are a large number of possible species for a site and when some indication of the species potential is required. The choice of seed sources for initial screening trials of introductions requires careful study of the natural distribution of the species and the environmental factors operating within the range.

2.     Provenance trials

(a)     Strategy for sampling provenances

Where there is some indication that a species may be of value on a particular site, but the extent of the variation in it largely unknown, it is usual to conduct range-wide provenance trials using many seed sources widely scattered throughout the existing range. The number of seed sources sampled will depend on the extent of the natural distribution, the genetic diversity of the species, and the resources available to sample and carry out the research.

When a species is being sampled for the first time and where little or nothing is known of the variation pattern sampling frequency should not be high. Samples are taken at rather widely separated intervals through relatively large areas. This type of sampling has been referred to as coarse-grid sampling and the purpose is to determine the major features of distribution patterns.

Coarse grid sampling has been used often in sampling many of the tropical pines or the eucalypts for the first time. The aim is to collect from populations which will represent the major variation in the species. Practical difficulties such as lack of access to part of the species range may rule out sampling on a regular grid or strip pattern. Often there is no regularity of distance or direction between successive sampling sites. The number of populations to be sampled is determined in advance in accordance with the time available. The samples are usually taken from near established roads or tracks in an irregular pattern determined by stand differences, and marked changes in ecological or topographical situations. Aerial photographs can sometimes be used to save time in picking out possible collection sites in unknown country or in difficult terrain.

A variation of the coarse grid system which takes advantage of a known distribution of the species, makes some concessions to ease of access and is less subjective is described by Green, (1971). Sampling localities for a study designed to provide basic genetic variation information in Eucalyptus obliqua were selected systematically by superimposing a 70-mile square grid on a map of the known range of the species. With some adjustments, twenty-two generalised locations where the species was known to occur were obtained. These locations were visited and precise localities chosen according to the abundance of seed crops, the absence of serious disturbance, and convenient access.

Given that only a certain amount of time and money is available to make seed collections of say 400 trees for any given species across its natural distribution, then some hard decisions have to be made on whether to sample few provenances with relatively large number of individuals per provenance or many provenances from diverse sites but with few individuals per provenance. Most collections involve this sort of rational and the decision to make is difficult. The central question is whether greater variation is among rather than within a provenance. The answer is not always clear-cut.

Once we have selected each provenance the next step is to seriously consider our selection of seed trees within the provenance.

(b)     Strategy for sampling trees within provenance

The actual trees to collect from will depend on the object of the collection. IUFRO has provided some recommendations on collections from individual trees (FAO, 1969).

(1)     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;

(2)     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;

(3)     Seed trees to be at least seed fall distance apart from each other. A distance of 100m 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 for Eucalyptus (Boland et al., 1980),

(4)     Individual seed trees to be marked,

(5)     Collect equal numbers of cones, fruits or seeds per tree, and,

(6)     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.

In some collections the botanical identity of some individuals can be difficult to resolve in the field and for this reason it is a good policy to keep individual seedlots and voucher botanical specimens. Once specimens are identified, seed from individuals within a provenance can be bulked. This concept is important for eucalypts and acacias in Australia, for dipterocarps in S.E. Asia and for pines in Central America where differences between species within genera can be difficult to discern in the field.

SEED COLLECTION METHODS AND EQUIPMENT

There is a great variety of methods and equipment available for collection of fruits and the choice depends on a number of factors which, following Robbins et al., (1981), may be summarized as follows:

(a)     Relative size and numbers of the fruiting units. In the case of 1 - 3 large seeds inside a dehiscent or indehiscent fruit (e.g. Aesculus, Tectona), collection can be done most easily by awaiting natural fall of seed or fruit and collecting from the ground;

(b)     Characteristics of the fruit: size, number, position and distribution of fruits and resistance of peduncles to shaking, pulling, breaking or cutting, interval between ripening and opening;

(c)     Characteristics of the tree: diameter, shape and length of bole, bark thickness; shape of crown; size, angle, density and resistance to breakage of branches; density of foliage and depth of crown;

(d)     Characteristics of the stand:. distribution and stocking of trees (e.g. isolated trees, open or dense stand, density of understorey and ground vegetation); and,

(e)     Characteristics of the site: slope, accessibility.

The various methods of collection may be classified into the following: (1) Collection of fallen fruits or seeds from the forest floor; (2) Collection from the crowns of felled trees; (3) Collection from standing trees with access from the ground; (4) Collection from standing trees with access by climbing.

1.     Collection of Fallen Fruits or Seeds from the Forest Floor

(a)     Natural seedfall

Collection from the forest floor of fruits which have fallen after natural ripening and abscission is common practice with a number of large-fruited genera. It is cheap and does not require highly skilled labour. Temperate genera commonly collected from the ground are Quercus, Fagus, Castanea and tropical genera include Tectona, Gmelina, Triplochiton and several genera among the dipterocarps. Eucalyptus globulus fruits are collected from the ground in Spain but never by this means in Australia.

The main disadvantages of collection from natural fruit shedding are the risks of collecting immature, empty or unsound seeds, of seed deterioration or premature.germination if collection is delayed, and uncertainty in identifying the mother trees from which seed is collected. Seeds in the first fruit to fall naturally in the season are often of poor quality (Morandini, 1962; Aldhous, 1972). In Thailand shedding of teak fruit starts in March, but observations have shown that the most viable fruits are shed in the latter part of the season, so collection is usually postponed until April (Hedegart, 1975). Clearing the forest floor of vegetation and debris and laying sheets can help. If carefully timed, this operation will also eliminate much of the risk of collecting empty or non-viable seed. Sound fruit should be gathered as soon as possible after they have fallen to avoid damage or losses from insects, rodents or fungi and premature germination. This is of particular importance in the moist tropical forest. Observations have indicated that many of the seeds of the more important dipterocarps lose their viability within a few days of shedding and studies on Shorea platyclados in Malaysia demonstrated that seedlots collected from the ground included considerably more defective seeds than lots collected from the standing tree (Tang, 1971). Collection from the ground must, therefore, be perfectly timed with seedfall.

In the Jari region of Amazonia in Brazil, Woessner and McNabb, (1979) found that collection of green or yellow fruit of Gmelina arborea from the forest floor gave the best results in operations for collecting about 10 000kg of seed a year. They could be stored temporarily in sacks during transit from the field to the fruit processing depot, without serious loss of viability. Older brown or black fruit ferment and heat in the sacks and rapidly lose viability. Collecting teams are instructed to collect only the fresh green and yellow fruit. 50kg of fruit can be collected per man in an 8 hour working day and yield about 3kg of dried stones. Similar results have been obtained in Malaysia where green and yellow fruit collected from the ground gave over 90% germination, but brown fruits, on the other hand, produced only 53% (Mohammad and Ibrahim, 1980).

Seeds of some hard-coated species may remain viable on the forest floor for years, especially in temperate conditions. In Hungary, seed of Robinia pseudoacacia is collected from the forest floor under 30-year-old stands in the Pusztavacs forest district (Keresztesi, 1979). In South Africa, seed of Acacia mearnsii can remain viable for several years on the ground under seed orchards and collected at any preferred time.

(b)     Manual shaking

Trunks of small trees and low branches may be shaken directly by hand. Higher branches may be shaken by means of a long pole and hook or by a rope. This method has produced good results in Cordia alliodora and Cedrela species. Several arid zone Australian acacias are also treated in this fashion.

(c)     Use of a rope

Involves an initial operation to pass the rope over the branch to be shaken. The same method is used to hoist a saw or pulley into the crown. A thin line is attached to a weight which is projected over the branch by hand or by catapult. For higher branches the line may be attached to an arrow, which is shot from a bow, or to an iron rod shot from a calibre .22 rifle. A light nylon line such as a fishing line of 50lbs (23kg) breaking strain is suitable and the weight or projectile used should be heavy enough to drop to the ground pulling the line with it over the branch. Once the end of the line has reached the ground, the weight or arrow can be detached and a nylon cord of 3-4 mm attached instead; the line is drawn back over the branch, pulling the cord with it. The two ends of the loop can then be pulled together to shake the branch. The cord should be positioned towards the end of the branch where it will have the maximum shaking effect and not close to the bole where the branch is thickest.

(d)     Collection of seed after dispersal

This usually involves nets or sheets to catch falling seeds and has been used to collect seed of Acacia aneura in central Australia (Doran et al., 1983).

(e)     Animal caches

Animals sometimes gather together cones or fruit as a food supply and these caches may be raided for the seed, but this source of seeds is confined to limited areas. Squirrel caches are an important source of coniferous seed in western North America. Squirrels usually locate their caches year after year in the same places. Typically, they are found in damp areas near springs, small creeks or marshes, on northern exposures, and in decayed wood or duff or around old fallen trees. A single cache may contain from a few cones to many bushels. Fresh cones on the ground are a sign of squirrel activity; piles of cone scales and cores may indicate a nearby cache (Stein et al., 1974, Dobbs et al., 1976).

Ants sometimes gather seeds together and in North Africa they have been observed to accumulate large piles of Acacia seeds. Any seeds collected from rodent or insect caches should be tested for soundness by cutting test or other means.

2.     Collection from the Crowns of Felled Trees

One method of collecting large amounts of seed is to synchronise it with normal commercial fellings carried out during the seed ripening season. There are obvious advantages in this but some disadvantages include:

The collection of fruit from wind-thrown trees is generally undesirable, as little selection can be applied and there may be a bias towards trees with characteristics which pre-dispose them to wind damage.

3.     Collection from Standing Trees with Access from the Ground

(a)     By Hand

In the case of shrubs or low-branched trees, fruit can be picked directly from the branches by the collector while standing on the ground (Morandini, 1962). Examples are Crataegus, Sorbus and Ilex spp. in temperate zones (Aldhous, 1972), the smaller acacias and mallee eucalypts in Australia and many of the small drought-resistant species of the arid and semi-arid zones. Smaller fruit are generally harvested directly into a basket, bag, bucket or other container held or worn by the picker (Stein et al., 1974).

(b)     Cutting, breaking and sawing

For branches out of arm's reach, a variety of long-handled tools is available to enable the collector to reach the fruit from the ground. A pole and hook may be used to pull branches down within reach. Long-handled rakes, saws, hooks or pruning shears are used to pull off or sever individual fruit or fruit-bearing branchlets. Light, rigid bamboo, aluminium or plastic poles 4-6m in length are common. In order to reach beyond the 6-8m range of single poles, multistage telescopic poles with a shear on the end have been developed. Robbins et al., (1981) have noted that, in some species, fruit or cones on the lowest branches may yield little seed, because of lack of pollination in that position, and that it is therefore preferable to collect fruits from at least halfway up the crown. Ability to use long-handled tools efficiently from the ground is much affected by the density and form of the crown in individual trees.

A rope can be thrown or pulled over a branch to break off the seed-bearing branch rather than to shake it. A thicker rope is needed than for shaking. The method is not recommended for general use. It damages the tree, allows access to pest and diseases and, in the case of pines and other species which take two years to mature their seeds, destroys the next year's seed crop while collecting the current year's.

Several types of flexible saw have been used successfully to sever branches from the ground. One model, described by Anon., 1979, consists of 3ft long flexible cutting cable fitted with precision-set carbon steel teeth and two 35ft polypropylene control lines. A sand-filled safety weight is used to project one of the control lines over the branch. An earlier model, the "commando saw". now no longer in production, was effective in severing eucalypt branches in Australia (Boden, 1972). Two operators could bring down branches up to 20cm in diameter quickly and easily.

The method is not applicable to trees with acutely angled branches such as Eucalyptus tereticornis. The other limitation depends on the efficiency of projecting the line over the desired branch.

(c)     Use of rifle

Another method of severing seed-bearing branches is to shoot them down with a large calibre rifle. The method was successfully used to shoot out the tops of white spruce trees in seed production areas in north eastern U.S.A. (Slayton, 1969). The topping of the trees was found to be not only less expensive than climbing but also the cones could be collected at the best stage of development because of the short time in which the operation could be completed. More recently, shooting off branches or tops from a helicopter has yielded promising results in Canada.

In Australia the collection of small samples of eucalypt and Araucaria seed from tall trees has been accomplished efficiently using a .222 or .243 or .308 calibre rifle with x4 telescopic sights (Boland et al.,1980). Branches up to 15cm diameter could be brought down. "Pointed soft-point" ammunition is more effective than "hollow-point" when used with a .308 rifle.

A disadvantage of the rifle method is that very strict safety precautions must be observed and that there are limitations to where a rifle may be used, for example, not near roads or built up areas. Also crowns of some species such as Araucaria and Picea may be considerably damaged by this technique.

In shooting down branches it is usually necessary to steady the rifle on a tripod or to rest the stock against a tree or the side of a vehicle. A clear line of sight is required and this can be a limiting factor in dense forests. It is usually best to shoot at right angles to the branch and to sever the bark on the underside with the first shot to avoid branch hang-ups. The bark on the upper side is then cut and finally shots are placed at intervals across the branch. It is important to select branches which will fall unobstructed to the ground. Horizontal branches are more readily detached than ascending branches. The shots should be positioned to take advantage of branch leverage. The method is best suited for collecting research quantities of seed from a heavy seed crop clustered on branches or tops too inaccessible to be conveniently reached by other means.

4.     Collection from Standing Trees with Access by Climbing

There is a limit to the height to which long-handled tools can be used for collecting seeds or fruit from the ground. Near that limit the operation consumes much time and energy but produces little seed. For tall trees which cannot be felled therefore, climbing is often the only practical method of collecting. Some men are excellent natural climbers, while good training and good equipment can render collection by climbing an efficient and safe, albeit energetic, operation. For convenience the operation may be described under the following sections: (a) Climbing into the crown by way of the bole, (b) Climbing into the crown directly, (c) Climbing and picking of fruits within the crown.

(a)     Climbing into the crown by way of the bole

     (1)     Climbing with minimum equipment. Climbing without mechanical aids is practised in a number of countries. In the Philippines some seed collectors climb barefooted or with the help of a rope which ties both feet together and presses them against the trunk of the tree (Seeber and Agpaoa, 1976). Other modifications are for the climber to cut successive notches in the bole with a hand-axe to support his feet, or to hammer in a series of iron spikes about 20cm long which are later withdrawn for re-use as he descends. Both of these methods are physically exhausting, whether or not a safety belt is used, and do some damage to the tree. Climbing tall branchless boles with hands and feet involves a considerable safety hazard and the risks may tempt climbers to prefer collecting from the most easily climbable trees which are often silviculturally the least desirable. It is preferable to introduce one or other of the special climbing aids now available.

     (2)     Climbing irons or spurs, which are attached to the climber's boots, offer a light and inexpensive means of safer and more efficient climbing, if combined with safety belt, strap and line, safety helmet of glass fiber and heavy leather gloves. The lightness of the spurs (less than 1kg a set) makes them particularly suitable for use in inaccessible stands in roadless country, where all equipment must be carried on foot. They have been found to be the most efficient method for climbing trees of Pinus kesiya and P.merkusii in Thailand, (Granhof, 1975) and for P.caribaea and P.oocarpa in Honduras and are in common use in many countries, especially for conifers (Robbins et al., 1981).

          The main disadvantage of spurs is the damage they do to the bark, particularly of thin-barked species. If climbing is only occasional, this should not be excessive, but frequent climbing of the same tree, e.g. for pollination and seed collection in seed orchards, is liable to cause an unacceptable degree of damage; other climbing methods should then be preferred.

     (3)     Ladders. For heights from about 8 to 40 metres, vertical scaling ladders in several sections provide safe and convenient means of climbing the bole to the live crown. They can be made of a variety of materials including wood, aluminium, magnesium alloy, etc., but each section must be light enough to be easily pulled up by the climber. The legs of the bottom section can be placed on adjustable platforms for greater stability. The length of each section varies between 1.8 and 3m and its weight should not exceed 3-5kg.

     (4)     The Swiss tree bicycle or "Baumvelo" is a device for climbing tall straight trees which are branch-free to the live crown. It is lighter to transport than sectional ladders but heavier than climbing irons. It does no damage to the tree. It is suitable for use on stems with diameters ranging from about 30-80cm.