Compacidad en espacios reflexivos

Permutational multivariate analysis of variance (PERMANOVA) produced a significant separation (P=0.0151) of the fungal communities present at the three levels of annual rainfall (<800 mm, dry; 800–1200 mm, intermediate; >1200 mm, wet). Use of CAP provided further evidence of the importance of rainfall, strongly separating the three groups (P=0.0008). In Figure 2, Axis 1is seen to separate the sites of intermediate rainfall from the other sites, whereas Axis 2 separates the wet sites from the others. PERMANOVA and CAP analyses of fungal communities provided only marginal, rather thanstrong, evidence for a significant association between communities and altitude (P=0.0583 and 0.0949, respectively). Similarly, using DistLM, no evidence was found that fungal communities have a significant correlation to temperature (P=0.262 using a composite score from PCA), whereas using a composite PCA score evidence was found of an association with rainfall (P=0.0158), thereby supporting the results from PERMANOVA and CAP that rainfall is the most significant environmental variable influencing fungal community assemblages.

Significant differences in fungal assemblages as a function of needle age were also obtained using permutational multivariate analysis of variance. From PERMANOVA, pairwise differences were indicated between 1-yr-old needles and fallen needles (P=0.0159) and between 2-yr-old needles and fallen needles (P=0.0184), with marginal

evidence of an assemblage difference between 3-yr-old needles and fallen needles (P=0.0725). CAP, which is an alternative way of testing whether the multivariate means are different for the four different age classes, was unequivocal in indicating a significant difference between age classes (P=0.0181 [trace test] or 0.006 [test based on first squared canonical correlation]), with 9 out of the 11 fallen needle samples being correctly classified. No 1-yr-old or 2-yr-old needle sample was correctly classified, and only 3 of the eight 3-yr-old needle samples were correctly classified.

Figure 2.CAP analysis of fungal communities in 34 pooled DNA samples from 3 rainfall categories (dry, intermediate, wet) of P. radiata plantations in Tasmania.

-0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4

CAP

2

CAP 1

This suggests that fallen needles have a different fungal assemblage to that of the three age classes, with their OTUs being more distinct than those in the needles retained by the trees. Nevertheless, there were some species, e.g. Lophodermium pinastri, which can be considered ubiquitous, as it was present in 14 of the cloning samples, 9 of which were in fallen needles. The sites at which L. pinastri was not present in the fallen needles sample were Tower Hill, Styx, and Branches Creek, the three driest sites.

VI.4. DISCUSSION

Needle fungal communities varied significantly between the 12 different plantation sites of 5 year old P. radiata that were investigated in Tasmania. Other studies also indicate similar significant variation in the needle fungal communities of another Pinus species (P. sylvestris) across different sites and locations in the Massif Central in France, colonized by the same or similar needle cast fungi active in Tasmania, viz. Lophodermium pinastri and Cyclaneusma minus (Gourbiere et al. 2001, Gourbiere and Debouzie 2003).

The present study found that temperature at a site, irrespective of whether it is an average, minimum or maximum temperature, does not strongly influence the composition of needle fungal communities. In contrast, statistical analysis using PERMANOVA, CAP, and DistLM of climate data after applying PCA to their standardised values, showed that rainfall was the key factor correlating with fungal assemblages at the 12 sites. In forest health surveys subsequent to this study (Wardlaw, unpublished data), only the two wettest of the 12 sites (Inglis River and Oonah) were reported to have a serious level of SNC. These two sites in northwest Tasmania are in the same general area where a previous study found other plantations that had SNC (Podger and Wardlaw 1990b). The average annual rainfall of these two wettest sites during the nine years preceding the survey, 1439 and

1353 mm, respectively (see Table 1), falls squarely within the predicted range of 1200- 2000 mm that was determined by Podger & Wardlaw (Podger and Wardlaw 1990a) from their analysis of the climatic characteristics of the disease distribution.

In another study (Prihatini et al. 2014 ), a Pinus radiata marker aided selection trial planted in 1999 in Oonah, Tasmania (a locality also used in the present study but in a different, and younger, plantation), three replicates of pine needle samples from four different needle cast disease classes, three full-sib families with known breeding values for SNC resistance, and three different needle ages, for a total of 108 pooled samples, were collected in spring 2007 and had their needle fungal communities analysed in a similar manner to that in the present study. Lophodermium pinastri, Dothistroma septosporum and Cyclaneusma minus occurred frequently (12, 16 and 19 needle samples, respectively), but notably 17 species of hither to unknown members of the Teratosphaeriaceae were identified, of which Teratosphaeriaceae sp. 23 and Teratosphaeriaceae sp. 03 occurred in 33 and 12, respectively, of the 36 needle samples. Teratosphaeriaceae sp. 23, a presumed endophyte, was detected in healthy trees as well as those affected by SNC, but Teratosphaeriaceae sp. 03, in contrast, was detected most frequently in needles of the highest disease class and never in healthy needles. These Teratosphaeriacae species were detected only by DNA sequences and were not isolated in the present study, and, to our knowledge, have not previously been isolated.

Podger and Wardlaw (1990a) postulated that SNC was due to the pathogenic activity of one or a suite of pathogens that switched from an endophytic phase following an as yet unidentified but ephemeral stress. This endophyte may not necessarily be a primary pathogen but secondary to an environmental stress when conditions are cold and wet, the most likely candidate being L. pinastri due to its ubiquitous nature. Of the four

putative causal agents of SNC (S.geniculata, C. minus, D.septosporum and L. pinastri), only L. pinastri was strongly associated with needles of trees with SNC (Prihatini et al. 2014b) . It is suggested that this Lophodermium species is present throughout these plantations and if the right stress is there, it can act as a pathogen, perhaps in collaboration with Teratosphaeriaceae sp. 03.The latter species is not ubiquitous but seems to be present at the sites that will subsequently develop SNC and it also has an association with trees that have higher levels of SNC symptoms (Prihatini et al. 2014b).