West Virginia Seed Sources of Balsam Fir

Research Bulletin 1191-99

Methods

In August 1981, seed was collected (Figure 3) from 15 trees in stands of balsam fir at four locations in West Virginia - Blister Run in Randolph County, Blister Swamp in Pocohontas County, Stone Coal Run in Tucker County, and Canaan Valley State Park in Tucker County. In the spring of 1984, an additional collection was made from 20 trees in a stand near Cortland Road north of the State Park in Canaan Valley (Figure 4; Table 1). Since one of the major objectives was to evaluate differences between and within areas, seed was collected and kept separate for individual trees (half-sib families) having a range of phenotypic characteristics at each location. Typically, the stands were located in or adjacent to bogs or swamps having small, running streams flowing through them; waters were not stagnant (Figure 5).

Figure 4. Locations of balsam fir stands in West Virginia. Blister Run Blister Swamp Stone Coal Run Canaan Valley-State Park Canaan Valley-Cortland Road

Figure 5. Natural stands of balsam fir at Blister Run (top) and Canaan Valley-Cortland Road (bottom).

In the spring of 1983, seed of 10 half-sib families from each of the 1981 collections was seeded in flats filled with sterile sand in a greenhouse at the OSU/OARDC Wooster, Ohio, campus. In addition, seed of two stand collections of Fraser fir (Abies balsamea var. fraseri) from Roan Mountain, North Carolina seed sources, which had been purchased from commercial dealers, was seeded. These collections were included in the study to provide a reference for comparison of trees of the West Virginia sources of balsam fir with a species/variety that has been widely planted and accepted as a Christmas tree. Following germination, seedlings were transplanted into 2-inch x 2-inch x 10-inch open-ended paper containers that had been filled with potting mix. Seed of eight half-sib families of the Canaan Valley-Cortland Road collection were handled in the same way except they were seeded in the greenhouse in the fall of 1984.

The planting site for the study is located at the USDA North Appalachian Experimental Watershed Laboratory near Coshocton, Ohio. The area had been an old field that had not been used for field crops or pasture for several years but was mowed periodically to control briars and other woody vegetation. The site slopes gently to the southwest and the soil is classified as a moderately well-drained Coshocton silt loam, which formed in residuum of interbedded acid shales, siltstone, and coal underclays.

Prior to tree planting, three-foot-wide strips along planting rows were sprayed with glyphosate (2.5 quarts a.i.a.). After trees were planted,areas within rows received annual applications of herbicides (simazine and atrazine, 2 pounds a.i.a. each); areas between rows were also mowed two or three times per year. Although the initial herbicide spraying killed most of the existing vegetation, there was substantial reinvasion/sprouting of woody species, particularly blackberry (Rubus sp.). Following planting, late summer applications of glyphosate and Garlon 3A were used to control that vegetation.

In the spring of 1985, trees of the 10 half-sib families from the first four West Virginia locations, plus the two Fraser fir stand collections, were planted. One-tree plots were used in each block (replication), with 10 feet between rows of trees and six feet between trees within rows. (This planting arrangement was used to accommodate conversion of the planting to a seed orchard at a later date.)

During planting, space was left at random locations in each block for adding trees of the eight half-sib families from the Canaan Valley-Cortland Road collection which were added to the planting in the spring of 1986. In all, there were 50 trees per block (48 half-sib families from the five West Virginia collections plus seedlings of the two Fraser fir stand collections), 24 blocks, and 1,200 total trees.

Periodic measurements and evaluations were made on individual trees, and statistical analyses were made using analyses of variance to test:

• Variation between half sib-families from different locations.
• Variation between all individual half-sib families.
• Variation within individual half-sib families.

In addition, correlation analyses were used to test the relationships between results of evaluations made in different years. When subjective ratings were used to rate tree shape and density, foliage colors and spring bud break, all evaluations were done by the author to minimize personal bias and perceptual differences.