Introduction   Chemical modification of wood has been successfully employed to improve dimensional stability and decay resistance of both solid wood and wood based panels (Papadopoulos and Traboulay 2002; Papadopoulos and Hill 2003; Papadopoulos and Gkaraveli 2003). In all cases considered in the literature, chemical modification employed in small clear wood specimens. In many cases, samples with the same number of growth rings were preferentially chosen and carefully selected so that growth rings were parallel to the tangential face, to ensure rapid reagent penetration into the cell lumen (Papadopoulos and Hill 2003). However, deviations from normal structure are not common, since trees are living organisms and are subject to various influences throughout their life span. When wood is looked upon as a raw material, most abnormalities adversely affect its service value; these are commonly called defects (Panshin and DeZeeuw 1980). From the wood utilisation point of view, defects are also certain normal characteristics of all trees, namely knots and pith. Despite an exhaustive search of the literature, there was no studies published dealing with the modification of wood with defects. Consequently, the purpose of this paper was to look at the effect that some defects may have on chemical modification of wood. The defects studied here in were: false rings of pine wood (Pinus sylvestris), discontinuous rings of pine (Pinus nigra), compression wood of pine (Pinus sylvestris) and tension wood of beech (Fagus sylvatica).

Experimental   Chemical modification reactions were performed as described previously (Papadopoulos and Hill 2003).

Results and discussion   The weight percent gain (WPG) values of wood with defects and normal wood, modified with acetic anhydride at 100°C for a 7 h period, are summarised in Table 1. Some interested points have come out from the Table. It can be seen that the presence of false and discontinuous rings did not affect the chemical modification reactions, bearing in mind that wood represents an exceedingly complex structure upon which to perform chemical reactions. Compression wood on the other hand, had a clearly adverse effect on chemical modification, since the WPG value was reduced by 61.7% compared to the normal wood. Two possible explanations can be offered for this behaviour. Firstly, it is known that chemical modification takes place upon the accessible hydroxyl groups of cellulose and hemicelluloses. Since the cellulose content of compression wood is less than that of normal wood (Panshin and DeZeeuw 1980; Siau 1984), less hydroxyl groups became available for the reaction to take place and this may affect the modification procedure. Secondly, it is known that the longitudinal permeability is reduced due to the reduced lumen size in the compression wood tracheids (Stamm 1964). It is also interesting to notice the high standard deviation values, which are almost four times higher than those of normal wood. In the case of tension wood, no definite observation can be made. The WPG values were lower than those of normal wood, but this can easily be attributed to the heterogeneous structure of the wood and to the fact that the differences between tension and normal wood are much less marked that those between compression wood and normal wood (Panshin and DeZeeuw 1980). Again, it is also interesting to notice the high standard deviation values, which are almost three times higher than those of normal wood.

Wood history	WPG	(%)
Pine wood with false rings	19.8	(0.9)
Normal pine wood	21.3	(0.4)
Pine wood with discontinuous rings	19.2	(1.9)
Normal pine wood	18.7	(0.6)
Compression pine wood	14.1	(2.7)
Normal pine wood	22.8	(0.7)
Tension beech wood	12.9	(2.5)
Normal beech wood	15.2	(0.9)