None of the tested P. kesiya sources reach the height and diameter growth of the control 5/87, which over the 7 year trial period has produced about 50 m3/ha more wood than the best P. kesiya source.
The Madagascar (esp. Morarano) and Vietnamese sources are the best in regard to growth. There is not much difference in growth vigour (no statistical differences) among the tested Vietnamese sources, but the trial results indicate Xuan Tho, Hothien and Thac Prenn as the best.
The Zambia CSO source has a volume production at about the same level as the Vietnamese and Madagascan sources, and is generally a good choice of seed source, also considering the good stemform (see further discussion below).
The Philippine sources are also above average in growth, but have less growth vigour than the Madagascan and Vietnamese sources. The height growth is comparable to the Madagascan and Vietnamese sources, but diameter growth is slower making volume production lower than would be expected based on height alone.
The John Meikle CSO source from Zimbabwe is not among the top-ranking sources. The seedlot originates from only one clone in the orchard, and this might explain the somewhat disappointing performance. A more representative seedlot, sampling all clones in the orchard, would probably perform better.
The Thai and Chinese sources are intermediate to below average. There is a rather large variation in the performance, best Thai source being the Doi Suthep and best Chinese source Simao, but generally they do not compare favourable to the above-mentioned sources.
The P. yunnanensis sources perform poorly. This is also true for the Aungban and Zokhua sources of P. kesiya from Myanmar.
Production potential, based on the best performing P. kesiya sources and based on growth in the first 7 years, is 20-22 m3/ha/year.
In general the adaptive traits show no clear picture and no clear geographical patterns as seen for the growth characters.
Survival percentage is high for all provenances. The statistical analysis reveals significant differences, but the analysis has to be interpreted with caution, as model assumptions are not fulfilled. In any case, differences among top-ranking provenances are probably small.
The results on cone-setting should not be given too much importance. The trial is still young and cone-setting only starting. Furthermore there may be no obvious (direct) correlation between adaptation and cone setting. The statistical analysis reveals significant differences, but there are no clear patterns of variation.
Finally, the analysis of foxtails reveals significant differences, but again, with no clear patterns of variation.
The general poor quality of P. kesiya with coarse branches and poor stemform is often referred to as a key constraint for a wider use of the species in plantation forestry. Improvement of quality would therefore be important in any future tree breeding.
The results (esp. the analysis of stemform and branch diameter) clearly illustrate the quality issue. The controls in general perform better than the P. kesiya sources.
The results, however, also illustrate that quality improvements can be achieved by selection and tree breeding. The seedlot from the clonal seed orchard in Zambia compares well with the controls, and have significantly better stemform than all other P. kesiya sources. It is also among the best in branch diameter.
Among the un-improved sources, the Madagascar land races (esp. Morarano) perform well in stemform and better than the Vietnamese sources. They are again better than the Philippine ones. The Philippine sources, on the other hand, have smaller branches than the Madagascar and Vietnamese sources.
The Philippine sources have low wood density, which is surprising both considering the relatively slow diameter development of these sources and results from other trials where the Philippine sources have been among the best. Thai and Chinese sources are generally ranked high in wood density, and better than the Vietnamese and Madagascar sources. However, considering growth and other quality features (esp. stemform) of these sources, they cannot be recommended.