THE ZONALITY OF OCCURRENCE OF KNOTS AND RELATIONS BETWEEN THEIR LOCATION AND SIZE IN LARGE-DIMENSIONED SPRUCE STEMS IN SOUTH-EASTERN NORWAY

Anna Barszcz¹, Peder Gjerdrum^2
1 Department of Forest and Wood Utilization, Agricultural University of Cracow, Poland
² Norwegian Forest and Landscape Institute, Ås, Norway

ABSTRACT

The present research concerned selected characteristics of knottiness of large-dimensioned spruce timber (of trees with the minimum breast-height diameter of 40 cm over bark) from three stands, over one hundred years old, located in south-eastern Norway. Fourteen sample trees were felled and measured, and knots which occurred in them were classified into three categories of healthiness and three categories of tightness with the surrounding wood. Sound and tight knots were dominant in the timber under analysis. All categories varied significantly in relation to average diameters, relative diameters (related to stem thickness at the points where the knots occurred) and the relative height of location along merchantable boles. The researches analysed the significance of relations of diameters or relative diameters of knots with the relative height of their location along merchantable boles. The relations, confirmed by coefficients of correlation close to or higher than 0.50 were described by equations of regression, most often by the expotential model. The frequency of occurrence of knots of different categories in sections of 0.1 of merchantable bole length allowed for distinguishing characteristic quality zones on the spruce stems: from the tree bottom to 0.3 of merchantable bole length inclusive, from 0.4 to the upper end of a merchantable bole – in respect of knot healthiness; as well as from the tree bottom to 0.3, from 0.4 to 0.6 and from 0.7 to the upper end of a merchantable bole – in respect of tightness of knots with surrounding wood. The results were compared with analogous data obtained earlier for spruce timber in mature stands in the Polish Beskidy Mts.

Key words: large-dimensioned spruce timber, Norway, categories of knots, diameters of knots, location on stems, quality zones.

INTRODUCTION

A tendency noted in most European countries, including Poland, is towards an increase in the share of stands of older age classes and, as a result, towards an increase in timber resources on stem. The issues of quality, e.g. in the aspect of the occurrence of timber defects and possibilities of processing such timber were undertaken within COST E40 Action. The present research, concerning large-dimensioned spruce timber quality, was financed from the means of COST E40 Action as part of a short-term scientific mission in Norway applying the methods accepted by COST.

Knots are among the most important timber defects. Depending on their sizes, location and structure due to their healthiness and degree of tightness with surrounding wood, they affect the result of timber quality classification to various degrees, which, consequently, influences further possibility of commercial use and achievable price. For many years research has been conducted on the occurrence of knots in timber as well as on direct and indirect factors affecting timber knottiness. Confirmation, by means of research, of the regularities connected with the locations and sizes of various categories of knots and description of the relations of these characteristics should provide the basis for creating tables or computer programmes for the quality classification of round timber which would be useful in timber industry.

Norway spruce, which is the subject of the present research, is one of very variable species. It creates numerous forms influenced by external conditions; there are also many geographical varieties of this species [10]. Research on the features of spruce knottiness, important for practical purposes, shows a number of regularities and constant relations whose significance was confirmed by statistical analyses [1,2]. Significant differences in the height of location of knots of different categories were also pointed out between large-dimensioned spruce trees – with the breast-height diameter over 40 cm and thinner ones [3].

The aim of the present research is analysis of selected characteristics of knottiness of large-dimensioned spruce timber from south-eastern Norway as well as comparison, in the form of discussion, of the results obtained with the results of analogous research carried out earlier in mature spruce stands of the Beskidy Mts in Poland [1].

MATERIAL AND METHODS

Following consultation in the Norwegian Forest and Landscape Institute in Ås, three mature (over one hundred years old) spruce stands were selected while felling was being performed there as scheduled in 2007. The stands represented the environmental and forest conditions that are typical of south-eastern Norway, with varied altitude and site characteristics:

• stand 1: Østfold District, Hobøl Commune, about 60 km to the south-east from Oslo, private forest of natural origin, altitude of 80 m, flat area, site G17 (medium fertile), volume  410 m³·ha^-1, age of the trees felled: about 150 years;

• stand 2: Buskerud District, Lier Commune, about 25 km westward from Oslo, private forest of natural origin, altitude of 480 m, a slope with a 15% gradient, site G15 (medium fertile/poor), volume 350 m³·ha^-1, age of the trees felled: about 105 years;

• stand 3: location, ownership and origin as in stand 2, altitude of 420 m, a valley bed with a north-western slope, site G20 (fertile), volume 350 m³·ha^-1, age of the trees felled: about 160 years.

In each stand, treated as the sample plot, among the trees selected for felling as part of scheduled harvesting, 14 sample trees were randomly selected (respectively 5, 4 and 5 trees in stand 1, 2 and 3). The selection criterion was the breast-height diameter of minimum 40 cm over bark (Table 1).

After felling and delimbing, each tree was measured for its merchantable bole length to the diameter not smaller than 7 cm over bark [9], timber diameters at each 1 m of the merchantable bole length and the height of the crown base. Next, over ½; of the stem girth, diameters of all knots sized ≥ 1 cm were measured and their distances from the lower end of the stem were noted. The rule of measurement of minimum diameters was applied [8]. The knots were classified according to their healthiness (sound, unsound and rotten knots) and according to the degree of their tightness with the surrounding wood (tight, partially intergrown and not tight knots) [7].

The next stage consisted in calculating the volume of merchantable boles and the relative height of the crown bases. By means of interpolation, timber diameter was determined at the point of occurrence of each knot registered. Other values calculated for each knot included: the relative height of its location along the merchantable bole (as the relation of the distance of a knot from the lower end of the stem to the merchantable bole length) and the relative diameter (as the relation of the knot diameter to the timber diameter at the point where the knot occurred).

Further analyses included description of the statistical characteristics of the features of each knot category and, by means of the Kolmogorov-Smirnov test (α=0.05), determination of the accordance of the features with the normal distribution.

The following stage was comparison of knot groups in categories of healthiness and degrees of tightness with the surrounding wood. The items compared included: diameters, relative diameters and the relative height of the location of knots. Since the variants of the above characteristics were not homogenous the non-parametric Kruskal-Wallis test (α=0.05) was applied; analyses were extended by means of the multiple comparison test.

The length of merchantable boles was divided into 10 equal sections and, for each section, the average diameters and relative diameters of knots as well as the frequency of the occurrence of knots were calculated (the number of items along the full girth of the timber and in %). The final stage consisted in analyzing the dependence of the diameters and relative diameters of knots on the relative height of their location along a merchantable boles. The coefficient of correlation R and the coefficient of determination R² (α=0.05) were used to choose the best matching dependency model.

The frequency of occurrence of the knot categories under analysis in sections of 0.1 m of the merchantable bole length was the basis for distinguishing characteristic quality zones on the stems of the spruce trees under analysis. The data were analysed in relation to the results of analogous research conducted in the mature stands of the Beskidy Mts [1].

RESEARCH RESULTS

The largest average length and volume of merchantable boles as well as the largest absolute and relative height of the crown base characterized the trees on sample plot 1. The smallest length and volume of merchantable boles was noted on plot 3; the absolute and relative height of the crown base on this plot was small, similar to the values found on plot 2 (Table 1).

Considering the combination of the categories of healthiness and the categories of the degree of tightness with surrounding wood, 7 detailed categories were distinguished in the material under research. Among them the dominant ones were sound, tight knots (over 64%) whereas unsound, partially intergrown knots and rotten, partially intergrown knots were noted in smallest numbers (each category: less than 1%). The remaining categories constituted several % of the total number of knots (Fig. 1).

In the aspect of healthiness exclusively, sound knots were clearly dominant (Fig. 2). Their diameters, relative diameters and relative height of their location reached the highest values while the lowest values of the features in question were noted for rotten knots (Table2). Considering only the degree of tightness with the surrounding wood, the dominant knots were the tight ones (Fig. 3). In this division, the largest diameters were noted for partially intergrown knots whereas the largest relative diameters and relative height of location was found in the case of tight knots. Not tight knots had the lowest values of all features. The relative diameters in all knot categories showed a larger variability than their diameters; the largest variability was often found for the relative height of location of knots on stems.

In most cases, analysis of the accordance of the distribution of the features of knots with the normal distribution yielded a negative result. The only exception were: relative diameters of unsound knots as well as relative diameters and relative height of location of  partially intergrown knots. Moreover, a lack of homogeneity of variance was found by means of the Leavene test (F from 6.988 to 77.558, p=0.000). The results of these analyses were decisive for the application of non-parametric statistical tests in further stages of the present research.

Comparison of the knot groups distinguished on the basis of 3 basic categories of healthiness and 3 categories of degrees of tightness with surrounding wood allowed for the conclusion that all groups significantly differed from one another both in the aspect of size parameters in absolute and relative units as well as in the aspect of the relative height of location. As a result of analysis by means of the Kruskal-Wallis test, the H_crit values obtained ranged from 786.0 to 3370.4 and the p < 0.001. The multiple comparison test (Table 3), which followed, allowed for the use of the above-mentioned knot categories in further analysis.

Divided into groups of healthiness, knots showed obvious differences in location on the merchantable boles of the sample trees. The largest number of sound knots (Fig. 4) occurred within crowns of the trees, which were on average based on 0.35 of the length of a merchantable bole; 60% of such knots occurred in the section from 0.6 to 0.9 of merchantable bole length. Their diameters over the length up to 0.6 of a merchantable bole were about 20 mm (the largest ones were in sections 0.3 and 0.5) and decreased above it. Relative diameters systematically increased towards tree tops. In the last, highest part of merchantable bole, there was a significant increase of relative diameters, which can result from larger taper of stems close to the tree top.The relation between the diameters and relative diameters of knots in this category and their location on stems were described by means of the expotential model: y= exp (a+bx), which proved to be the most adequate. The parameters of both equations are presented in Table 4.

Most unsound knots were found in sections 0.2 – 0.3 of a merchantable bole, i.e. below the crown, and then within the crown – in section 0.5 (Fig. 5). Their diameters in the upper half of a merchantable bole showed relatively low values and increased toward the top. Relative diameters increased irregularly towards the top. The low coefficients of correlation between the location and the diameters (R= 0.322) and relative diameters of knots (R= 0.372) indicate the average strength of the relation; that is why the dependencies were not described by equations of regression.

Rotten knots, most frequently occurring in the butt-end section of stems (Fig. 6), were characterized by a small differentiation of diameters and an increase in relative diameters towards the tree tops. Due to a very low coefficient of correlation between the relative height of location and the diameters of these knots (R=0.044), the expotential equation was applied only to describe the influence of location along stems on relative diameters (Table 4).

The following stage consisted in analysis of knots as categorized according to the degree of their tightness with surrounding wood. Similarly to sound knots, the tight ones were also mostly (nearly 70%) located in the section from 0.6 to 0.9 of merchantable bole length (Fig.7). Their diameters showed a decreasing tendency from section 0.2 towards the top of the stem. The relative diameters increased towards the top. The coefficient calculated indicates a high-level correlation between the diameters of the knots in the examined category and their location along the merchantable bole. The dependency was described by means of the expotential equation (Table 4). The relation between relative diameters and the height of location was weaker (R=0.401).

Except the lowest and the top sections, partially intergrown knots were characterized by regular distribution along the merchantable bole (Fig. 8). Their diameters did not show distinct tendencies towards change along the merchantable boles. The coefficients of correlation between the location and the size parameters (both the absolute and the relative ones) were low and amounted to, respectively: 0.358 and 0.392; the relations were not described by equations of regression.

Numerous in the bottom zone of the stem (except section 1), not tight knots clearly decreased in number towards the tree top (Fig. 9). Their diameters had small variability, relative diameters gradually increased towards the top. The relation of the diameters of this category of knots with the height of their location turned out to be very weak (R=0.064). In order to describe the relations of relative diameters and the location of knots, the multiplicative model (y= ax^b) turned out to be the most suitable (Table 4).

Considering all knots together, without the division into categories, it was noted that their number increased up to section 0.8 inclusive (Fig. 10). In the part near the top, the frequency of occurrence of knots decreased. The diameters increased up to section 0.6 inclusive and then decreased. The relative diameters systematically grew towards the top. The relation between the diameters and the relative height of location of knots was generally weak (R=0.194). The relation of the relative diameters to the height of location of knots along merchantable boles was found to be stronger (Table 4).

The number of unsound and rotten knots in sections of 0.1 of merchantable bole length was similar: the largest number of these knots was in sections 0.2 and 0.3; higher up it decreased (Fig.11). In sections 0.9 – 1.0 no rotten knots were found. The number of sound knots grew dynamically from the bottom of stems up to section 0.8 inclusive (on average there were about 35 knots along the girth of this section); higher up their number decreased. With regard to the average height of the crown base, which was at 0.35 of merchantable bole length in the analysed trees, it may be observed that in the lower stem parts without living branches the differentiation of the healthiness of knots was considerable, with a large share of unsound and rotten knots. This analysis allowed for distinguishing characteristic zones along spruce merchantable boles. Lower stem parts up to 0.3 of merchantable bole length inclusive, i.e. zone I,  had a similar share of knots of all healthiness categories. Higher up, i.e. from 0.4 to the thinner end of a bole, called zone II, sound knots were dominant and the share of the other categories decreased.

Analysis of the frequency of occurrence of knots as categorized according to the degrees of their tightness (Fig. 12) allowed for an observation that still within the crown, even up to section 0.5, numerous not tight knots were found. From the bottom of the stem up to 0.4 of merchantable bole length, the number of these knots in sections was higher than the number of tight knots. In sections 0.8 to 1.0, the number of partially intergrown and not tight knots was similarly small; in the last section near the top not tight knots were absent. In sections 0.7 to 0.9, tight knots were very numerous: their frequency was from 20 to 33 items/section. The distribution of knots in categories of tightness with surrounding wood along merchantable boles allows for distinguishing the following zones. Zone I included the butt-end, up to 0.3 of merchantable bole length, where the dominant knots were the not tight ones. Zone II, located higher, can be divided into sub-zones: IIa – up to 0.6, with large differentiation of knot categories and the middle of which was the intersection point of the curves of tight and not tight knots (0.45 of merchantable bole length) and IIb – above 0.6, with a large number of tight knots and the disappearance of the other categories.

Considering merchantable bole length (on average 24.5 m), the zones characteristic of trees under analysis are: zone I, up to 0.3 of bole length inclusive is 7.3 m long and its upper diameter is 32.3 cm and zone II, up to the upper end of a merchantable bole, is 17.2 m long and its upper diameter is 7 cm over bark. The two sub-zones of zone II, distinguished for the frequency of occurrence of knots in different categories of tightness, are of the following length: sub-zone IIa, 0.4 to 0.6 inclusive, is 7.3 m long (its upper end is 14.6 m from the tree base and has the diameter of 24.5 cm); and sub-zone IIb, up to the upper end of a merchantable bole, is 9.9 m long.

Assessment of timber technical quality requires the number of knots per unit of stem length. In the material under analysis, over the whole length of a single spruce tree (in relation to the full girth of timber), there were on average 253.2 knots. An average of 10.4 knots were found on each meter of length of a merchantable bole.

Considering the frequency of occurrence of all categories of knots together, it was concluded that merchantable bole sections from 0.5 to 0.9 (Fig. 13) contained the largest number of knots; the most – about 37 ones (when the adopted threshold of measurement of their diameters was ≥ 1 cm.) were found in section 0.8

DISCUSSION

Comparison of the results received in the present research with the data obtained in earlier studies in mature spruce stands of the Beskidy Mts in Poland [1] allows for a number of observations.

The average tree height as well as the absolute and relative height of the crown base were smaller in spruce in Norway. However, the sample trees were thicker, and therefore had higher taper, than the material from Poland. The average merchantable bole length was 27.7 for Polish spruce while in Norway it was by about 12% smaller. Both groups had very similar average volume of the merchantable bole of a single tree, i.e. about 1.52 m³ in Norway and about 1.59 m³ in the Beskidy Mts. A larger taper and a smaller height of the crown base in Norway is connected with e.g. the geographical position of the stands under research. Numerous earlier analyses had led to conclusions that, as environmental conditions move away from optimum, e.g. when geographical width increases, the taper and crown length increase in Norway spruce [6]. For the same reason, large dimension trees are more frequently found in Polish than in Norwegian stands.

As they were shorter, the merchantable boles of spruce in Norway had a larger number of knots. Along the merchantable boles of Polish spruce [1], 204.2 knots were noted over full girth and 7.8 items on 1 running meter, which is respectively by 20 and 25% fewer than in Norway. The Norwegian material had a much higher share of sound knots and a lower share of unsound ones. In both groups of material, rotten knots showed an almost identical share in relation to the sum of all knots (slightly over 12%). The material in Norway had more tight knots, fewer not tight knots while the share of partially intergrown knots in the spruce of both countries was very similar (over 6%).

The relation between knot locations and sizes was more visible in the spruce in the Beskidy Mts. Significant relations between diameters or relative diameters of knots and their relative height of location – with the coefficient of correlation R close to or higher than 0.50 – occurred more seldom in the material from Norway (6 equations of regression) than in that from Poland (9 equations). The coefficients of correlation for the relations in question were higher for spruce in the Beskidy.

Knots (altogether) with the largest diameters were noted in spruce in Norway in the bottom part of the crown, i.e. in the first three sections from the crown base towards the top. An identical result was also obtained for spruce in the Beskidy [1]. A similar regularity was found for spruce in north-eastern France by Colin and Houllier [4], and then by Vestøl and Høibø [11] in south-eastern Norway.

Analysis of spruce in the Beskidy allowed for describing, by means of equations of regression, the relations between the relative height of the crown base and the relative height of location on stems of all distinguished knot categories (except rotten knots). The coefficients of correlation (in the range from 0.478 to 0.699) assumed positive values each time. The same character of relations, although not described by equations of regression, may be observed in the Norwegian material, where – with lower crown bases – the relative height of location of all knot categories was smaller than in the Polish Beskidy.

With regard to the frequency of occurrence of different knot healthiness categories in sections of 0.1 merchantable bole length, it may be observed that both in Poland and in Norway the maximum number of knots per 1 section – considering in turn rotten, unsound and sound knots – increased towards tree tops (the difference between unsound and rotten knots was only slight in Norwegian material). A regularity was also observed that in spruce in Norway these maximum numbers in each knot category are found by 0.1 of merchantable bole length lower than in Polish spruce. The border between zones I and II, distinguished on the basis of the frequency of occurrence of knot healthiness categories in the Polish material at 0.6 of merchantable bole length, was, however,  located much lower, i.e. in section 0.3, in spruce stands in Norway. The length of zone I of a merchantable bole that was analogous to spruce in Norway (up to 0.3) was found in thinned spruce stands (aged 40 to 60 years) in the Beskidy Mts [2].

Similarity in the course of the curves of knot frequencies in Polish and Norwegian spruce is visible in the zonality of knots as categorized according to the degrees of their tightness with surrounding wood. As in the case of healthiness categories, also in this categorization the maximum numbers for Norwegian spruce occurred by 0.1 of merchantable bole length lower (for not tight knots the shift was by 0.2). As a result, the border of zone I in the material from mature spruce stands in Norway was established to be by 0.1 of bole length lower than in Poland. The point of intersection of the curves of frequencies of tight and not tight knots, which in Poland was found at 0.6 of merchantable bole length (i.e. 1 section higher than the crown base), was noted in Norway to be located at 0.45 of bole length, i.e. also 1 section higher than the crown base. The upper border of sub-zone IIa, found in Norway at the height of 0.6, is located at 0.7 of merchantable bole length in Poland [1].

The issue discussed in the present study is of practical significance. Giefing [5], after Taffe and Lippemeier, notes that knots influence the quality of spruce sawn timber in 75-90%. Much attention is currently being paid to the research on the features of knottiness and possibilities of obtaining timber with small, sound and tight knots, surrounded by broad zones devoid of knots. Spruce timber with such advantages may be obtained, among others, by means of the optimum number of seedlings per area unit, by regulating stand density in various stages of stand development and, above all, by tree pruning. The present research, as well as previous studies done in the Beskidy Mts [1], shows a relation between the relative length of tree crowns as well as the distribution and overlapping of the zones of occurrence of different categories of knots. Therefore, in order to produce high quality timber, it is necessary to focus on the formation of appropriate proportions between crown length and tree height as part of stand management. Tree pruning is a significant measure in this respect. When performed in a rational way [5]  pruning is an exceptionally effective way of augmenting the timber value, whose increase may even reach several hundred per cent and make it possible to obtain special assortments, in which the features of knottiness have a key role.

SUMMING UP AND CONCLUSIONS

1. Analysis of features of knottiness in large-dimensioned spruce from mature stands located in south-eastern Norway showed that, among three categories of knot healthiness and three categories distinguished on the basis of degrees of tightness of knots with surrounding wood, sound and tight knots dominated in the examined material. These knots were characterized by the largest relative diameters and height of location along merchantable boles. Sound and partially intergrown knots had the largest diameters. A relatively high share was noted for rotten and not tight knots. All knot categories differed significantly in their size parameters (absolute and relative ones) and in their location on tree stems.

2. Relations between the size of diameters and the relative height of location turned out to be statistically significant in the case of sound and tight knots. More of such relations were confirmed for relative diameters and the relative height of location; it concerned: sound, rotten and not tight knots as well as all knots considered together. These dependencies were described by equations of regression, in most cases by the expotential model.

3. Relative diameters of all categories of knots increased towards the tree top, which was noted in sections of 0.1 of merchantable bole length. In the stem part from the stem bottom to 0.8 of merchantable bole length inclusive, the number of knots overall per one section increased. This was due to an increased growth of the number of sound and tight knots. The number of rotten, unsound and not tight knots reached the maximum in section 0.2 and then systematically decreased toward the tree top. Partially intergrown knots did not show a distinct tendency towards change in their number along merchantable boles.

4. The frequency of occurrence of knots in their categories of healthiness allowed for the division of merchantable boles into the following zones: I – up to section 0.3 of a merchantable bole inclusive, with a distinct share of knots of all categories; and II – above zone I, to the upper end of a merchantable bole, with the dominance of sound knots. The distribution, on the stems of the spruce trees under analysis, of knots in categories of degrees of their tightness with surrounding wood allowed for the distinction of the following zones: I – up to section 0.3 of a merchantable bole inclusive, with a dominant share of not tight knots; IIa – from 0.4 to 0.6 inclusive with the presence of all kinds of knots and IIb – above section 0.6 with the dominance of tight knots.

5. The analysed material revealed the relations of the features of knottiness with the relative height of the crown base (located at 0.35 of merchantable bole length). Moreover, numerous analogies were found between the features of knottiness of spruce in south-eastern Norway and spruce in the Beskidy Mts in Poland.

REFERENCES

1. Barszcz A. 1995. Ocena jakości surowca drzewnego i zmienności wad drewna w rębnych drzewostanach świerkowych na terenie Beskidów [Assessment of timber quality and timber defect variability in mature spruce stands in the Beskidy Mts]. Zesz. Nauk. AR Kraków. Rozpr. 199 [in Polish].

2. Barszcz A. 1999. Knottiness as a factor differentiating characteristic quality zones on spruce stems in thinned stands. Sci. Pap. Agric. Univ. Pozn., Forestry 2, 13-21.

3. Barszcz A., Michalec K. 2007. Spruce timber quality in trees with different thickness from the Beskidy Mts. Acta Sci. Pol., Silv. Colendar. Rat. Ind. Lignar. 6(3), 5-15.

4. Colin F., Houllier F. 1991. Branchiness of Norway spruce in north-eastern France: modelling vertical trends in maximum nodal branch size. Ann. Sci. For., 48, 679-693.

5. Giefing D. F. 1999. Podkrzesywanie drzew w lesie [Pruning of trees in the forest]. Wyd. AR w Poznaniu, Poznań.

6. Modrzyński J. 1998. Ekologia. W: Biologia świerka pospolitego [Ecology. In: Biology of Norway spruce], Boratyński A.i Bugała W. Bogucki (eds), Wyd. Nauk. Poznań, 303-426 [in Polish].

7. PN-79/D-01011. 1980. Drewno okrągłe. Wady [Round timber. Defects]. Wyd. Normalizacyjne, Warszawa [in Polish].

8. PN-EN 1310. 2000. Metody pomiaru cech [Methods of measurement of features]. PKN, Warszawa [in Polish].

9. PN-93/D-02002. 1993. Surowiec drzewny. Podział, terminologia i symbole [Round wood. Classification, terminology and symbols]. Wyd. Normalizacyjne Alfa, Warszawa [in Polish].

10. Tomanek J. 1980. Botanika leśna [Forest botany], PWRiL, Warszawa [in Polish].

11. Vestøl  G. I.,  Høibø O. A. 1998. Prediction of  knot diameter in Picea abies (L.) Karst.). In: Single-tree models of knot properties in Norway spruce (Picea abies (L.)Karst.). Doctor scientiarum theses 34. Agric. Univ. Norway.

Accepted for print: 14.07.2008

Responses to this article, comments are invited and should be submitted within three months of the publication of the article. If accepted for publication, they will be published in the chapter headed 'Discussions' and hyperlinked to the article.