A TREE IMPROVEMENT PROGRAMME FOR SOUTHERN HARDWOODS ¹

by
R.C. Purnell and R.C. Kellison ²

RECENT PROGRESS

The rate of planting of hardwoods in the southern U.S. is increasing significantly; it is currently 4,000 ha annually. The site specificity of hardwoods necessitates the use of 4 – 5 species and it is imperative to use the best genetic stock available.

The N.C. State University-Industry cooperative hardwood research programme began in 1963, since when it has undergone several changes. Initially superior trees were selected using an index that combined assessment of a number of traits into a single score for each tree. The traits used in the index were: volume, pest resistance, crown conformation, leader dominance, bole straightness, bole pruning, freedom from epicormic branching and branch angle. This system was used in preference to the comparison of candidate select trees with neighbouring “check” trees, because most commercially important southern hardwoods occur in mixed and uneven-aged stands where comparison selection systems are inappropriate.

A total of 709 trees representing 27 species were selected, using the index system. Many of the selected trees were preserved in clone banks, while superior sweetgum (Liquidambar styraciflua L.), sycamore (Platanus occidentalis L.), yellow-poplar (Liriodendron tulipifera L.), and green ash (Fraxinus pennsylvanica Marsh.) were used to establish clonal seed orchards (Table 1). The progeny from these selected trees have been included in open-pollinated genetic tests.

Table 1. HARDWOOD SEED ORCHARDS ESTABLISHED BY MEMBERS OF THE N.C. STATE HARDWOOD RESEARCH PROGRAMME

The selection of phenotypically superior southern hardwood trees that are also genotypically superior is difficult, even when using the index system, because natural hardwood stands are composed of many species and have originated from both seedlings and sprouts. Trees from sprout origin are often inadvertedly selected in preference to trees from seedling origin, because sprouts have superior bole and crown characteristics which are features of vegetative propagules (Sweet and Wells, 1974). Phenotypic selection is also difficult in hardwood stands because many stands have suffered from dysgenic selection that has resulted from repeated, incomplete harvests.

The Hardwood Research Cooperative is circumventing these problems and limitations by selecting trees that have above-average phenotypic value for that site but not necessarily meeting the phenotypically superior standard required by the index selection. The genotypic value of the above-average trees is then determined in open-pollinated genetic (mother-tree) tests. Information from these tests will be used to identify the best trees for grafting into clonal seed orchards. Using this extensive selection system, an additional 513 trees from 4 commercially important hardwood species are being tested in open-pollinated genetic tests (Table 2).

Table 2. NUMBER OF ABOVE-AVERAGE PHENOTYPES BEING TESTED IN OPEN-POLLINATED GENETIC TESTS

THE FUTURE

The need exists to integrate the trees selected and under test in the intensive and extensive selection programmes into a cohesive improvement programme. Despite the large number of species that are represented by selected trees, only sycamore, sweetgum, green ash, water oak and willow oak have enough commercial importance across the southern United States to warrant a regional programme. Since the total area planted each year of any single species is limited, the best option for an improvement programme is to rely upon an open-pollinated breeding scheme. This offers the opportunity for some genetic gain at a cost that is generally lower than a control-pollinated programme, especially for species where floral biology is poorly understood. However, increasing relatedness and accompanying inbreeding depression may require a shift to more control-pollinated breeding in the future.

For maximum efficiency and genetic gain, a tree improvement programme has two distinct phases - a breeding phase and a production phase. The breeding phase allows continuing genetic improvement into advanced generations and the production phase allows mass production of improved material in seeds or vegetative propagules. The individuals used in the production phase are a subset of the best individuals from the breeding population.

Breeding Phase

To maintain genetic diversity and integrity of geographic seed sources, the southeastern United States has been divided into breeding regions for each commercially important species. The breeding regions are zones that have genetically similar trees and similar physiographic and environmental conditions. For example, the range of sweetgum east of the Mississippi River, the working territory of the Hardwood Research Cooperative, has been divided into seven breeding regions (Figure 1) The selected trees will be utilized only within the region from which they originate, until tests reveal their suitability outside the breeding region.

Figure 1. Proposed breeding regions for sweetgum.

The cooperating organizations within each breeding region will continue to select and evaluate trees from natural stands and unimproved plantations, using the extensive selection methods until each breeding region has selected at least 500 above-average trees per species. The open-pollinated genetic tests of the 500 above-average trees established with the extensive selection method will be used to select the superior trees for the breeding population, based on family and individual tree information. The two best trees from each of the best 50% of the families will be the parents of the second-generation breeding population. Three commercial check lots common to each region will be included in all genetic tests of that region.

Each genetic test will consist of 50 families and will be planted at two locations. Each location will have ten replications with six trees per family per replication. The six trees from each family in a replication will be divided into three groups of two, and the three groups will be scattered throughout the replication so that a different set of families will surround each group (Figure 2). The groups will allow for a 50% thinning in the tests without losing any families and will also improve cross-pollination among families. Genetic tests will be computer-designed from N.C. State University. The cycle will then be repeated for subsequent generations, with individuals selected for the best breeding population being chosen from the previous breeding population on the basis of family and individual tree information (Figure 3). Clonal seed orchards will also be rogued on the results of the genetic tests.

Figure 2
Experimental layout (ith replication). Design of the 6-tree partial noncontiguous plots in the second generation genetic tests.

N. C. STATE HARDWOOD RESEARCH COOPERATIVE TREE IMPROVEMENT BREEDING SCHEDULE

Figure 3. N. C. State Hardwood Research Cooperative Tree Improvement Breeding Schedule

Provisions have been made for continued selection of superior trees from unimproved plantations to increase the gene pool of the breeding population for future generations.

Production Phase

The production phase of the improvement programme involves the mass production of improved stock and will utilize the very best subset of trees from the breeding population. The improved material can be mass-produced either by seeds from seed orchards or by vegetative propagules. Vegetative propagation has a tremendous potential for the southern hardwoods; however, more research is needed before it is accepted for mass-producing the improved stock.

Clonal seed orchards are favoured above seedling seed orchards within the Hardwood Cooperative because the design of the first-generation genetic tests is not conducive to conversion to a seed orchard. The second-generation genetic tests are more conducive for converting to seedling seed orchards because the individuals from a family are located in noncontiguous plots instead of row plots. However, the gains from these second-generation seedling seed orchards still would not exceed those of clonal orchards because clonal orchards could utilize the best individuals from every genetic test in the breeding region (500 families), while the seedling seed orchard would be restricted to one test (50 families). For the second-generation seed orchard it would be best to utilize the very best individuals from each genetic test in the breeding region.

More improvement could possibly be obtained in advanced generation seed orchards by inserting selected trees from breeding populations from other breeding regions. The feasibility of seed source movement from one region to another will be investigated when breeding populations are established. Progeny from five superior trees from each genetic test within a region will be combined for test establishment in each of the other breeding regions. The design of this trial will be like that of the mother-tree study, comprising six replications of ten-tree row plots.

The design of the clonal seed orchards will be similar to existing hardwood seed orchards, with the size of the orchard depending upon the seed productivity of each species. Dioecious species like green ash pose special problems in production seed orchard design. Generally each planting spot in a seed orchard is both male and female; but for green ash, males and females will have to be positioned for maximum pollination efficiency.

For the first generation an open-pollinated system for intermating the selected trees has been adopted, but in later generations control-pollination may be necessary to avoid inbreeding depression and reduced gains from the seed orchards. A control-pollinated breeding scheme is particularly applicable to an operational planting programme utilizing vegetative propagules.

REFERENCES

Sweet, G.B. and L. Wells. 1974 Comparison of the growth of vegetative propagules and seedlings of Pinus radiata. New Zealand Jour. For. Sci. 4(2): 399–409.

¹ A shortened version of a paper presented at the 17th Southern Forest Tree Improvement Conference, University of Georgia, Athens, June 6–9, 1983.

² Graduate Research Assistant and Director, respectively, Hardwood Research Cooperative, School of Forest Resources, North Carolina State University, Raleigh, N.C., USA.

Note from the Editor - Although limited to species of SE United States, the principles and methodology presented in the present paper could be applicable to tropical hardwood species, in which many of the problems and constraints expressed in the paper also apply.