The testing programme
During the initial work, the need for progeny-testing was soon recognised and a series of four small field tests of Sitka spruce families was established in 1953-4. Although these included less than 20 families, they were to become important in later years when the relationship between early growth and mid-rotation volume was established (Gill, 1987).
The next tests planted comprised the series of four Scots pine tests already discussed, planted in 1957. Whilst these did provide early indications of breeding values and heritability, annual planting programmes did not begin until 1966-7 for all the major species. The capacity for large scale plant production and progeny test planting grew from this time and an intensive programme ran steadily over the next 25 years.
When the primary concentration on Sitka spruce together with major testing programmes for Scots pine, lodgepole pine, hybrid larch and occasionally Corsican pine are taken into account, annual planting programmes of up to 30 hectares of conifer progeny tests involving up to 500 different families took place.
The data in the table below demonstrate the scale of the field progeny testing work during the period when it reached its peak in the mid 1980s.
|Sitka spruce||Scots pine||Lodegepole pine||Larch||Total|
|Field-test planting||No. of tests planted||14||5||12||0.3||33|
|No. of families per year||186||87||97||15||492|
|Field assessments||No. of measured assessments||30||10||24||7||77|
|No. of subjective assessments||18||10||25||4||61|
|Total No. of assessments||48||20||49||11||138|
|Total trees assessed||103 789||28 577||82 881||11 872||250 823|
Tree breeders strive to obtain estimates of breeding values as quickly as possible from as small a number of trees per family as possible in order to reduce the overall cost of a breeding programme. Forest stage progeny-testing is by far the most expensive stage in terms of manpower and resources. Knowledge regarding the optimum progeny test design, the number of sites within Britain required to test a parent tree adequately and the optimum selection ages, was scarce in the early years.
Design of progeny-tests
Randomised block designs were generally used for progeny testing, but plot size varied initially between 8-tree line plots and 16-tree (4x4) square plots with between two and six replications. In 1968 an experiment was established with Sitka spruce to compare a range of linear and square plot sizes. This indicated that plot sizes of up to 10 trees were the most efficient, and accordingly an experimental design within sites of eight tree row plots with five replications was used routinely from 1977 (Johnstone and Samuel, 1978). This requirement for 40 trees per family per site on three sites (120 trees in total) became the standard design for all subsequent progeny tests from which calculation of breeding values of original plus-trees (parent trees) was intended.
Suitable progeny test sites were selected by foresters based at one of the research outstations. Site confirmation would always be required by the relevant project leader before a site could be prepared for planting. In addition to the outstations operated by the branch, the services of other outstations operated by Silviculture (North and South) Branches were also used.
Bare-land sites which had not previously been planted with trees were fairly readily obtainable until the early 1980s when areas of restocking had to be increasingly considered. It was clear that a more sensitive experimental design was required to compensate for the increased heterogeneity of a restocking site. Whereas it seemed quite possible to find bare-land sites with superficial homogeneity over the dimensions of a typical replicate (normally around 1600 m2), the random location of stumps and accumulations of brash, together with variable type of cultivation, meant that homogeneous units were vastly reduced in size on restocking sites.
The Generalised Lattice Design (also referred to as the 'Alpha Design') (Patterson and Williams, 1976), which sub-divided each replicate into incomplete blocks typically of just five or six plots (total dimensions 160 to 200 m2), was introduced in 1983. Plot size remained at 8-tree row plots and there were still five complete replications.
Number of test-sites
Initial trials were planted on a wide range of sites (eight for Sitka spruce, fove to six for pines) felt to be representative of major planting zones for the different species within Great Britain. Combined analysis across all planting sites of height data collected six years after planting in some of the first Scots pine and Sitka spruce progeny tests indicated broad similarities between groups of sites and the number of testing sites was subsequently reduced to the three representing the most contrasting site types. For example, with Sitka spruce these were:
- Those in which the growing season is short and a cold climate additionally influences growth (sites in North-East Scotland)
- Those in which the growing season is long, low temperature rarely limits growth and frost damage is minimal (characterised by South Wales and South-West England)
- Intermediate sites in which these extremes are less prominent (the borders between Scotland and England have been commonly used).
With appropriate modification for particular species, representative sites based on these criteria have been used since 1975.
Assessment of field experiments
During the early phase of progeny-testing, a fairly intensive schedule of field assessments was used. This followed those already established in the provenance testing programme. Height was often measured at planting and at the end of the 1st, 3rd, 6th and 10th growing seasons. After this it became more difficult and expensive to measure total height of each tree, and diameter at breast height (DBH) was commonly adopted on a five yearly basis. Annual measurement of height in the Sitka spruce population study (see below) indicated that there were many changes in rank in progeny performance up to the 6th year after planting but that after this age, rankings became more consistent. Gill (1987), in a study of data from one of the original Sitka spruce progeny tests planted in 1953, found that height at 6th year was strongly correlated with volume at 27 years. Much time and effort was therefore saved by making a 6th year height assessment the first and definitive measure of growth rate in progeny tests. Backwards selection of plus trees for the breeding population was therefore based on selection at this age with some minor adjustments based on DBH at 10 or 15 years.
The data from the population study also indicated that there were very poor correlations between nursery height and later field performance. This led to the abandonment of routine nursery height measurement in the late 1970s.
Following breeding value estimation through progeny testing, re-selection of the best parents, based on the performance of their progeny, was carried out to create a breeding population. The very best of the breeding population would be selected to constitute the production populations which would inter-mate to give the improved propagules for the next generation of forests. The work in these areas will now be considered separately for each species.
A summary of progress in the main species appears in the table below. This table also refers to the seed origin or provenance testing programme which began as early as 1926 but only became part of the remit of the branch in 1987.
|Sitka sruce||Scots pine||Corsican pine||Larch||Douglas fir||Lodgepolepine|
|Total no. of tests||392||143||38||72||48||156|
|Planting years (19-)||67-98||57-87||64-89||59-97||59-95||66-89|
|No. of years in which planting took place||32||20||7||26||15||19|
|No. of families in test||2594||1174||935||330||575||578|
|% selections tested||90||86||92||39||70||12|
Provenance/seed origin testing
Previous: The early years