Dag Lindgren¹ and Bo Karlsson²

1Department of Forest Genetics and Plant Physiology, Swedish University of Agricultural Sciences, S - 901 83 UMEÅ, Sweden.

This document is mainly based on a report to the National Board of Forestry, Sweden (Lindgren 1992). The report aimed at surveying available Swedish genetic materials of Norway spruce (Picea abies) and evaluating their potential for use in seed production. The emphasis in the report was on reducing seed production costs.

Later developments have made cost reductions even more urgent. The interest in seed orchards has dropped for a number of reasons, which are discussed below. The price for timber and demand for plants decreased during the 90's because of the world recession. Current forest growth projections suggest that annual growth is much higher than annual cut, and this raises doubts about the need of improved seeds. The focus has changed from production towards environmental issues. To abstain from actions to increase production is now regarded as good environmental practice and gives immediate savings, which is an attractive combination. When foresters encounter problems that they cannot explain, they tend to blame the provenance. The improved production of genetically improved stocks is, although economically highly significant, difficult to see with the naked eye in field trials. The predicted gains of the initial selections have become lower and the problems better known and more evident as more knowledge about seed orchards has accumulated. A system with governmental subsidies for establishment of seed orchards has been discontinued. Plant and seed production are now regarded as commercial activities and not as a governmental responsibility. Seed and plant supply for long rotation forestry are not the best targets for strict application of the principles of market economy. This will most likely lead to decisions that are not optimal for Sweden.

Swedish seed orchards have recently been discussed and reviewed by Wilhelmsson et al. (1993). Norway spruce seed orchards have a smaller impact on Swedish forest production than projected when they were established. The adult Norway spruce seed orchards (Table 1) were established mainly in the 60's. The Norway spruce seed orchards are not expected to have reached full production during the period mentioned in Table 1. A second batch of plus tree selections from forests for new seed orchards was made in the 80's. Norway spruce seed orchards have always had lower priority than pines, partly because of problems like irregular flowering and insects, and partly because the big companies have demanded pines.

Table 1. Adult seed orchards (cf. Eriksson & Palmér 1991).

Area, ha Harvest 1979-88, kg % of plants "today"

Scots pine 488 19 754 60

Norway spruce 206 2 958 20

We experienced some minor problems in making a uniform national compilation of the young seed orchards. These problems can be predicted to grow if the "market" becomes less regulated. Compilations of this kind are needed for many reasons (learn by experience, predict possible future seed shortage, OECD-demands etc.) and do not require much work. The National Board of Forestry, Sweden, has recently accepted responsibility to keep a National Seed Orchard Register.

Seed orchards are planned and decided several decades ahead of the harvest and use of the seeds. Neither the seed demand three decades ahead, the seed production, nor the technical development can be forecasted with any precision.

One part of a strategy under uncertainty should be diversification. Ideas may turn out good or bad, who knows? If no variation is allowed, even retrospectively, it may never be known that there was a better way.

There is a considerable chance the seed demand will be less than projected. This is what has happened during recent years in Sweden. If variable seed supply options are created it is more likely that better options than worse options will be used. An elite seed orchard can be expanded in the future at a cost by intensive management or vegetative propagation of the seeds. Thus, it seems to be a good strategy even in a situation with strict economic constraints to invest some of the resources in rather intensive concepts. On the other hand, there are inexpensive ways to create options to get cheap improved seeds, and these could be used for meeting a large fraction of the projected seed demand.

From genetic and biological points of view, the situation has improved, and instead of relying on old phenotypes there are now argue amounts of tested materials (see next section) as well as a working long-term breeding program. Much of the best parts of the breeding population are juvenile, and this will remain so permanently in the future as long as the breeding program continues.

We believe that the future niche for conventional grafted seed orchards of Norway spruce will be a small one. The seeds will be expensive and still genetically outdated and the seed production is not reliable. There is a need for alternatives.

The existing breeding materials are described in greater detail by Werner & Danell (1994, this volume) and Karlsson (1994, this volume).

Table 2. Progeny test by number of grafted parents tested in different years for Norway spruce in southern Sweden.

Y e a r

Mating design -1974 1975-79 1980-84 1985-90 1991-

Known father 50 75 75

Wind pollination 75 2264 >75

 The current situation is that there is a rather small base material represented in old trials with known fathers and a great many young wind-pollinated progeny-tests (Table 2).

In 10 of the adult orchards large full sib families were created by artificial crossings in grafted seed orchards, usually according to rather complex mating designs, and sometimes resulting in more than 1000 descendants from each clone.

For 7 adult orchards wind-pollinated seeds collected in the orchard were used, results were presented by Hannerz (1993). For the second batch of plus tree selections, wind-pollinated seeds were collected in the forest as fast as possible after the identification of the plus tree.

In the late 70's it was decided to make a new batch of seed orchards. However the large investment in progeny testing of the older selections could hardly be utilized at all, as too few clones were tested and the tests were too young. More clones could have been tested earlier if wind-pollinated progenies had been harvested in connection with the selection. In retrospective, it may be concluded that much too large effort was invested in testing too few clones. The current idea at that time was to aim for the stars (viz. significance stars). Today, the program is aiming more for high gain combined with a sufficient genetic base, and then it does not pay to produce large families at the cost of testing many parents. It is not unlikely that further reductions of progeny size for parental ranking would be rational. The large initial experiments have a large value for scientific reasons and for learning more, so they may not be regarded as a waste of money. Nonetheless, as a breeding operation they were not efficient.

 During the 80's many spruce genotypes have been tested as clones multiplied by cuttings (Table 3).

 Table 3. Number of Norway spruce clones for which field trials have been established in southern Sweden by year.

-1980 1981-85 1986-90

1630 3323 3342

 The intention was that these tested clones should be used for clonal forestry. However, they may have a bigger potential as parents of the production and breeding populations. Clone tests are more powerful than progeny tests for evaluation of breeding values with small tests and limited non-additive variance (Lindgren 1992, Appendix 9). We actually suggest that there is no need for the concept of "progeny testing" for species where vegetative propagation is easy, progeny testing can and should be replaced by clonal testing.

 Cone harvest from Norway spruce is not dependent on level ground and wide spacing (as it is with Scots pine). The concept of clonal forestry has resulted in there being a large number of tested cutting clones in field trials. Seed production areas similar to a forest can be established using tested clones multiplied by cuttings. This could be achieved at a much lower investment cost than the present grafted seed orchards. The sacrifice in genetic gain need not be high, if any. It is recommended that a high percentage of the projected future need of Norway spruce seeds will be produced in this way instead of by intensively managed grafted seed orchards.

 However, as part of the strategy it is also suggested that some of the seeds should be produced on intensively managed grafts of the very best tested genotypes in the breeding population. The best grafted alternatives are likely to be genetically better than the best cutting alternatives (cf. Lindgren 1992, Appendix 9), there may be a few outstanding progeny-tested trees which are too mature to be propagated by cuttings but which are desirable to use for a long time and grafts may be more suitable for intensive management. If only a few seeds are produced (perhaps after controlled crossing), they can be expanded by cuttings if the foresters are willing to pay much for plants at that time.

 The Norway spruce seed orchards are basically established according to a concept developed more than half a century ago for Scots pine by Syrach Larsen in Denmark, which turned out to be rather efficient and viable for Scots pine (cf. Table 1). The biology connected with cone harvesting differs widely between Norway spruce and Scots pine. Norway spruce is characterised by infrequent bumper crops, most years it is not worthwhile collecting the few cones, while Scots pine produces seeds and is harvested in almost all years. For Scots pine, the cones are spread out over the surface of a rather wide crown, while for Norway spruce the cones are much more concentrated to a narrow volume at the top of the tree. The Scots pine cones are firmly attached to the shoots and there is no other practical way to get them down than to manually remove them, which requires level ground to support a ladder or a machine-operated platform. Wide spacing is needed to allow the movement of such equipment. The Norway spruce cones are heavier and more loosely attached. They can be shaken off with a tree shaker. They can be raked down using a helicopter with a cone-raker. The top can be cut off and lowered down to the ground when there is abundant cone set. The tree will usually survive and recover from this drastic treatment, perhaps before the next bumper crop. The operation can be done by a helicopter-operated scissors, or by a tree climber, or by a regular tree harvester leaving a six metres tall "stump" (once in a decade). Alternatively, some cone-bearing trees can be felled in a conventional way, and the effects on wood production may not be particularly large.

 The requirement for level and smooth ground has forced foresters out on farmland, and that has created problems and costs. The land is expensive and subject to competition with non-forest interests and under different laws and administration than forest land. The seed orchard is seen as an alien element, bringing weeds to the farm fields, awkward landscaping, etc. The flat land is often in need of expensive draining operations. The flat land is frost-exposed. The competing vegetation is vigorous. On non-agricultural level land costs for smoothing (removal of stones) may be high. For Norway spruce all these problems can be avoided by not requiring level and smooth land and abstaining from the option of manual cone-picking.

 Grafts are very expensive. As they are expensive and sensitive to damage, expensive protective actions like fencing are often found justified. For Norway spruce, cutting propagation is a much cheaper alternative. A problem has been that maturation problems make it impossible to use an adult plus tree as ortet. However, in the present situation in Sweden this is actually not any particularly severe disadvantage. There are many clone trials (Table 3). The predicted genetic gain, when selected good tested cutting clones are used, is not much below grafted alternatives (Lindgren 1992, Appendix 9). In principle, grafts cannot be genetically inferior to cuttings, as grafting is always possible when cutting propagation is possible, but it is also possible on mature material. Thus, it is realistic to accept a loss in gain. If the seed production forests are intensively thinned, higher gain may be obtained with cuttings than with grafts if the alternatives are compared at the same price. On the other hand, the genes of mature superior genotypes can always be rejuvenated for cutting propagation via sexual crosses. At present, most tested clones in Sweden have been preserved as juveniles in hedges, but clones may be propagated from older material for seed production stands than for ordinary forests, so juvenility is less constraining.

 The time aspect is important. It seems possible that cuttings will flower later than grafts. There are no reliable data on that. General impressions are mainly based on comparison with grafts from mature ortets in seed orchard environments and cuttings from juvenile ortets in clone trials. Such a comparison is very unfair to the cuttings. Experimentation is needed, e. g. the same clones could be multiplied in both ways, starting with the same mother plants (some variation in maturation stage), and compared in the same environments (both forest and archive). Until data are available, a conservative opinion may be justified, thus we may assume that cuttings in the forest will produce seeds at age 25-30 years to the same extent as grafts at age 15 years. A long delay may be the most severe draw-back with juvenile cuttings. However, grafted seed orchards have a long lifetime (actually much longer than desirable from a breeder's point of view), while a seed production forest may be harvested only once rather early, so throughout the orchards entire life-time, the time loss may not be large, if any.

 Traditional grafted Norway spruce seed orchards fit extremely poorly into a market-economy. There is no complementary of alternative economic return from the investment than the seeds, so there is very little flexibility from a managerial point of view. The capital costs of expensive land have to be charged on a few kg of seeds. The risk of a complete failure (e. g. biological failure, loss of market, or some ideas enforced by some international authority) with no return on investment is evident. The time span between the investment and the return is extremely long (longer than for Scots pine). This makes the investment very hard to justify by calculations involving the type of interest rates charged by a bank. As the flowering is irregular, expensive seeds have to be stored for many years, with high interest costs to be charged on the seeds. The improved seeds compete on the market with unimproved seeds of suitable provenances, which can be bought very cheaply from abroad (unimproved Scots pine seeds are not that cheap). Many Norway spruce plants are bought by private persons, and immediate private consumption may be more attractive than the glory of a higher yield, which someone else will harvest in the far future. For Scots pine, there are more companies and trained foresters involved, who have been taught to use the land well, who think about the future in a more institutional way, and who do not spend their own money. Thus Scots pine seed orchards are less problematic.

 Seed production forests fit much better into a market-economy than conventionally grafted seed orchards. The land and the investment are much cheaper. There is another valuable product produced, viz. wood. The seed price needs to be charged only marginal costs, not the full cost. Forest production will be reduced by the cone collecting operations, but only some trees will be affected, the most valuable part (the bottom 6 m of the trunk) will be less affected than other parts, and this cost appears first when - and only if - cones are collected.

 There are some extra costs compared to a normal stand. The stands must be registered and documented for half a century, they may become "forgotten". The best available clonally tested material must be used.

 There are a number of options, which give added benefits, but also extra costs if chosen. Dry "pine sites" may be used to stimulate flowering and to improve pollen isolation, but that has a cost in lost forest production (perhaps partly compensated by cheap land). Fertilizers may be applied to force fast development to cone-producing size. Intensive thinning may be applied. Clones may be identified (mapped) to make a more efficient thinning possible. Clonal rows may be considered, there is a cost in increased selfing level, but also a gain in the option of a more genetic thinning and selective cone harvest without the cost and complication of a map. Flowering stimulation could be applied, and if some trees do not survive, it is not any particular disaster.

 The combined clonal and progeny tests of the breeding population could be planted so that they can be converted into seed production stands after final evaluation. This would give good improved seeds as a side effect of long-term breeding at a minimal extra cost, and with genetic thinning the material would be the best available at the moment. The distance in time to the breeding population would be a few years instead of decades.

 Seed production areas could be established as a combined clonal and progeny test with denser spacing and more clonal copies on the site than required for long-term breeding. This would create a seed production unit which may be as expensive as a conventional seed orchard, but would produce genetically better seeds. This idea was developed in greater detail by Lindgren and Werner (1989).

Acknowledgements. We are grateful to a number of persons for sharing information and ideas with us.

 Eriksson, U. & Palmér, C.-H. 1991. Skogsfröplantager i Sverige. Institutet för Skogsförbättring, Årsbok 1990.

 Hannerz, M. 1993. SkogForsk, Avelsvärden Nr 9.

 Karlsson, B. 1994. Twenty years of clonal forestry with Norway spruce breeding in Sweden. Paper planned for this volume.

 Lindgren, D. 1992. Produktion av förädlat granfrö. (Production of improved Norway spruce seeds for Sweden, Summary and part of contents in English). Department of Forest Genetics and Plant Physiology, Swedish University of Agricultural Sciences, Arbetsrapport 40:1-98.

 Lindgren, D. & Werner, Ö. 1989. Gain generating efficiency of different Norway spruce seed orchards designs. Institutet för Skogsförbättring, Rapport 11:189-206.

 Werner, M. & Danell, Ö. 1994. On the problem of Picea abies breeding in Sweden. Paper planned for this volume.

 Wilhelmsson L, Eriksson U & Danell Ö 1993. Produktion av förädlat frö (Production of genetically improved seed, In Swedish with English summaries). The Forestry Research Institute of Sweden. Redogörelse nr 3. 52 pp.

This file is an improved version of a file from November 21,1993 "granfrö2.pa3".

Dag Lindgren