Fredrik SCHLYTER-1,2, Qing-He ZHANG-3,4, Milos KNIZEK-5,
Paal KROKENE-6, Goran BIRGERSSON-7.
1- Animal Ecol, Dept. Ecology, Lund Univ, S-223 62 LUND, Sweden,
2- Chemical Ecology, Dept Plant Prot Sci Swedish Univ Agric Sci P.O. Box 44,
S-230 53 ALNARP, Sweden,
3- N Forest Pest Bio control Ctr, Gen Stat Forest Pest Ctrl, Min Forestry, No.58
N Huanghe Street, SHENYANG 110034, P. R. China,
4- Dept Environ Sci & Forestry, FCT, Univ Nova Lisaboa, P-2825 MONTE DE CAPARICA,
Portugal,
5- Forestry and Game Mngmt Res Inst, CZ-156 04 PRAHA 5 ZBRASLAV, Czech Republic
6- Norwegian Forest Res Inst, N-1432 AS-NLH, Norway
7- Chemical Ecol, Dept Botany Göteborg Univ, S-413 l9 GÖteborgRG, Sweden.
A west palearctic population (Norway/Sweden) was known to contain a blend of about 10: 90 of E-myrcenol: ipsdienol (Id: EM) and this blend was shown as highly competitive to natural pheromone emitted by unpaired males in spruce logs (Schlyter et al. 1993). To find the optimum ratio for best attraction of the two pheromone components, a range of baits with 0, 1, 10, 50, 90, 99 and 100 % of ipsdienol was tested in four sites. In NE China, a ratio near 90% Id of Id & EM was most attractive to a population in Xinggan Range in 1992. However, another population in Inner Mongolia (N China) showed a reversed preference in 1993 and 1994, with best results for the 10 and 50% Id baits. Tests in 1994 in Bohemia (Czechia) and Norway confirmed the preference of western populations for a blend between the 50 and 90% Id ratio. A cluster analysis showed the Iriner Mongolia responses 1993 and 1994 to be the most divergent from west Palearctic and NE China rersponses. Norway results from 1993 and 1994 cluster closest and near both Czech and NE China results.
A non-linear regression (NLR), based on a Gaussian response profile, allowed a specific pin-pointing of both the peak of response (location of optimum) between experimental values and a measure of how wide the peak was (width of response window). Interpreting the Gaussian curve fitted as a standard normal distribution, the peak equals the mean (µ or x ) and the window the standard deviation (o or S.D.). The average fitted values were x ± 1S.D. (r 2): Norway 0.66 ± 0.15 (0.88), Czechia O.54 ± 0.17 (0.69), NE China O.75 ± 0.02 (0.97), Inner Mongolia O.3410.05 (0.86). The NLR procedure is applicable to a wide range of data sets and provides an objective measure for both response peak and window. For Ips duplicatus the peak location varies, with a possible relict population as an outlier, while the response window seems to vary less in width.