- P is the significance level of the factor (n.s.: not significant; *<0·05; **<0·01; ***<0·001). % is the percentage of variability explained by the factor. Factors that have not been considered in the model are represented with a dash (–).

- a roentgen: resistant; S: prone.
- b Level of inoculated seedlings.
- c Mean wilting payment ± SE.
- d Letters title homogeneous communities from the Fisher’s LSD test (P = 0·05).

## Hydraulic conductivity and susceptability to cavitation

Vulnerability to cavitation (P_{fifty} and P_{80}), Kx_{maximum} and absolute conductivity (Kx) did not differ significantly among the types of crosses (Fig. 1; Table 3). Loss of conductivity began at ?0·3 MPa and progressed at a similar rate in all crosses, i.e. there were no differences in the slope of VCs (P = 0·87; Table 3).

- a WP 20 d.a.i., wilting percentage 20 days after inoculation; VC slope, ‘a’ parameter of the exponential sigmoid: PLC = 100/(1 + exp[a(??b)]); P
_{50}, applied pressure at which the sample loses 50% hydraulic conductance; P_{80}, applied pressure at which the sample loses 80% hydraulic conductance; Kx_{max}, maximum xylem specific conductivity; VL_{max}, maximum vessel length; bVL, vessel length distribution parameter; WD, wood density; VD, vessel diameter; VTA, vessel transectional area; THC, relative theoretical hydraulic conductance; VF, vessel frequency; (t/b) 2 , resistance to implosion; PGV, percentage of grouped vessels; VPG, vessels per group; VGA, vessel groups per area; CLVF, contribution of large vessels (VD >70 ?m) to flow; CMVF, contribution of medium vessels (40 < VD < 70 ?m) to flow; CSVF, contribution of small vessels (VD <40 ?m) to flow. - b Roentgen: resistant, S: vulnerable. Suggest worth ± SE. Emails name homogeneous communities contained in this a variable (P = 0·05, Fisher’s LSD means).
- cP-really worth from the anova.
- *log-switched in order to meet anova standards; **inverse-switched to meet up anova criteria.

Despite P_{80} and Kx_{max} not differing between crossing types, these variables were positively correlated with WP 20 d.a.i. for the 24 selected trees (P < 0·05; Table S1). Nevertheless, the coefficient of correlation was low in both cases (R 2 < 0·20).

## Anatomical has

Maximum vessel length (VL_{max}) ranged from 69 to 118 mm. S ? S trees had 30–40% significantly longer conduits and a higher percentage of longer vessels (Fig. 2a; Table 3). There was a negative correlation between Kx_{max} (log-transformed) and bVL (R 2 = 34·5, P = 0·0026; Table S1): plants with shorter vessels had lower conductivity.

S ? S progeny showed the widest vessels (VD; Table 3), and were unique in having vessel diameters greater than 90 ?m (Fig. 2b). The progeny of the S ? S cross also had larger VTA, and a THC twice as high as the other two groups (Table 3). CLVF, CMVF and CSVF did not differ among crossing types (P > 0·05, Table 3). As expected, Kx_{max} (log-transformed) was positively correlated with THC (R 2 = 32·6, P = 0·0035; Table S1) and VD (R 2 = 28·8, P = 0·0068; Table S1). In addition, R ? R individuals showed a significantly higher VF (c. 20%) and a greater (t/b) 2 (P < 0·05; Table 3). Meanwhile, S ? S saplings had significantly higher PGV (Table 3). There were no differences in WD, VPG or VGA between the groups (P > 0·05; Table 3).

Both ?_{pd} and ?_{md} progressively decreased after DED inoculation (Fig. 3a). Seventeen d.a.i., ?_{pd} had dropped more than 0·25 MPa and 47 d.a.i. c. 1 MPa, independent of the type of crossing (Fig. 3a). ?_{md} dropped from ?1 MPa to almost ?3 MPa in S ? S progeny at the end of the experiment singleparentmeet aanmelden. From the thirteenth d.a.i., ?_{md} of R ? R cross progeny was significantly different from S ? S cross progeny.