Introduction

The selection of best performing trees (plus trees) for target traits is essential in tree breeding programs (Zobel and Talbert 1984, Ruotsalainen 2014). Contrary to easily measurable phenotypic selection traits, some traits which has economic importance (i.e. Wood strength and density) require considerable efforts for determination. Breeding of wood quality characters requires methods for fast screening of large numbers of trees (Ruotsalainen 2014). There is a need for assessing a large number of trees and their families for economically significant traits, such as strength and density. However, traditional methods for assessing for wood strength and density are relatively expensive, they are required laboratory works and restrict the numbers of samples that can be processed. In addition, traditional strength measurement methods involve cutting of the sample trees. Since there is a strong linear relationship between wood density and strength of tree species (Steffenrem et al. 2007), wood density/specific gravity is used as a relatively easily measured indicator parameter of wood quality (Panshin and de Zeeuw 1980, Haygreen and Bowyer 1996, Zhang 1997). Depending on different works investigating density-strength relations for pine species r² is high, i.e. Kiaei (2011), found r² value of 0,73 between compression strength (Cs) and density, 0,71 between MOR (Modulus of rupture) and density for eldar pine (Pinus eldarica Medw.); Zhang (1997) reported higher values between Cs and specific gravity for horsetail pine (Pinus massoniana Lamb.) (r²=0,86) and for yunnan pine (Pinus yunnanensis Franch.) (r²=0,85); Guller (2007), reported strong linear relations (r²=0,90 for Cs; r²=0,68 for MOR) for red pine (Pinus brutia Ten.). Moreover, positive moderate to high genetic correlation was reported between density and wood strength for pine species (Kumar 2004, Missanjo and Matsumura 2016).

The relationship of the mechanical properties with wood density/specific gravity at species level, to some extent, is different from species to species. In general, maximum crushing strength in compression parallel to the grain in a species from the one softwood category (softwoods with abrupt transition from earlywood to latewood) appears more closely related to specific gravity as compared to the species from the other softwood category (softwoods with gradual transition within a ring) (Zhang 1997).

Red pine (Pinus brutia Ten.), having a large commercial use, was given the highest priority and is defined as target species for intensive tree breeding studies (Yildirim et al. 2011, Kandemir 2013).

Tree improvement programs for red pine (Pinus brutia Ten.) have mainly focused on breeding of trees with superior growth characteristics (Isik 1998). However, wood quality is important for end user (Barnett and Jeronimidis 2003) indeed, particular attention should be given to wood technological characteristics to be considered in the tree selection process (Nocetti 2008, Yildirim et al. 2011). In Turkey, there are relatively few studies on heritability of wood properties compared to growth characteristics. The preliminary results of a study conducted by Yildirim et al. (2011) suggested that selection for some wood characteristics (ie. specific gravity determined on small clear samples with 0,42 ± 0,07 individual (hi²) (and 0,55 ± 0,03 family mean heritability- h_hs²) and growth traits in Turkish red pine could be practiced at early ages for short rotation (about 30 years) industrial plantations (Yildirim et al. 2011). For the same species medium to high heritability values (hi² between 0,45 and 0,74; h_hs² between 0,68 and 0,80) were reported for wood density (determined by x-ray densitometry) and ring density components (except for latewood density) suggesting that these characters are under moderate to strong genetic control, and thus, by selecting populations having high wood density would lead to genetic improvement can be achieved in red pine (Pinus brutia Ten.)(Guller et al. 2011).There is a need to include wood quality traits in tree selection program to ensure future wood supplies that have the appropriate mechanical properties for primarily structural applications and other end uses.

There has been an increasing interest in developing and using cost-effective non-destructive/semi-destructive technologies to evaluate the strength of standing trees all around the World (Bucur 2003, Wu et al. 2007, Brashaw et al. 2009).

Although many of non-destructive methods which are used for determination of wood strength are indirect methods (Brashaw et al. 2009), Fractometer can perform direct strength measurements on 5 mm diameter cores (Goetz et al. 2002, Chiu et al. 2006, Matsumoto et al. 2010). The advantages of the device are that it produces relatively fast results obtained relatively fast, portable, and easy to use (even in the field), causes less damage on trees (Lin et al. 2007, Matsumoto et al. 2010). Required increment core sample for the device also give a chance to see and compare visually growing performance of the investigated trees. Due to its numerous advantages, the Fractometer has been employed for assessing the wood strength of standing trees for various purposes, in general by arborist to monitor trees (Mattheck et al. 1995, Dolwin 1996, Matheny et al. 1999, Lin et al. 2004, Chiu et al. 2006, Matsumoto et al. 2010). There are few works which compares standard strength measurements performing on clear samples and Fractometer measurements. Matsumoto et al. (2008) compared Fractometer measurements and static test results for bending and compression strength of two Japanese softwoods (sugi (Cryptomeria japonica (L.f.) D.Don) and akamatsu, japanese red pine (Pinus densiflora Siebold & Zucc.)) at green condition using small clear specimens according to Japan Industrial Standard. They reported no significant relationship between bending strength of core samples and static bending strength of small clear specimens. In contrast, significant positive relationship was reported for compression strength suggesting, the standard small clear sample compression strength of the woods can be estimated non-destructively by using core samples and Fractometer. Few years later, Matsumoto et al. (2010) reported a relatively high relationship between Fractometer-measured compressive strength and basic density for five Indonesian plantation species (kauri (Agathis sp Salisb.), merkus pine (Pinus merkusii Jungh. & de Vriese), silver wattle (Acacia mangium Wild.), american mahogany (Swietenia sp. Jacq.), and teak (Tectona grandis L.)). Kraler and Beikircher (2013) stated that the Fractometer device can be used for a preliminary assessment of wood strength rather than determining/predicting the strength value which is performed by standard test method.

Cown (2005) mentioned about GF Plus system in operation (Cown 2005) whereby individual traits are ranked as growth, straightness, branch frequency, disease resistance, wood density, and spiral grain. Among the different wood traits wood density which is determined easier than wood strength and which has been regarded as an important indicator of wood quality has been given the highest priority for tree improvement for many tree species worldwide (Zobel and van Buijtenen 1989, Cown 2006) and generally researchers have been focused on estimating of wood strength by using density as an independent variable. Innovations on portable non-destructive/semi-destructive devices and practical procedures for direct measurement of wood strength in the field, might shift that approach vice-versa.

The study's theoretical approach postulates that if it is possible to estimate wood strength based on the linear relationships between density (as an independent variable) and wood strength (as a dependent variable), the reverse should also hold true. The primary aim of this study was to assess the practical usability of the fractometer as an initial evaluation tool in the field for selecting red pine (Pinus brutia Ten.) trees based on certain wood quality traits. Specifically, the focus was on density, with the intention of showing possibility for integrating these traits and the device into pine tree breeding programs.

Materials and methods

Sample materials and applied method

Red pine (Pinus brutia Ten.) increment cores (5 and 10 mm in diameter, defect-free) from 1,30 m height of experimental trees were used as sample material. The primary data for this study were obtained as part of a project (Project number: TUBITAK 110-O-560). Increment core samples were collected from 5 experimental stands for the project, as shown in Figure 1 and Table 1. Since all experimental stands were protected areas used for seed source and genetic trials, there was no option for tree harvesting in the project. Indeed, considering previously published results regarding the relationships between Fractometer measurements and standard tests, which suggest that it can only be used for a rough estimation of standard strength test results for standing trees. Small clear sample tests were not conducted as part of the project, and this study did not include any comparisons with standard tests.

In the initial stages of the project, the plan was to acquire a high-speed motorized coring device to collect a larger number of cores. However, as the project commenced, the manufacturer of the intended device ceased production of the unique equipment. Consequently, an alternative approach was adopted for high-speed incremental coring. Classical 5 mm and 10 mm increment borers were purchased, and their steel cores were fitted with two fundamental steel adapters (designed by Bilgin Icel specifically for this application). These adapters were then affixed to two distinct rotary devices, both of which were acquired from the local market.

One of the devices operated using a portable electric generator, while the other utilized a rechargeable drill (Figure 2a and Figure 2b). The adapted devices performed effectively when used on younger, smaller diameter trees. However, they encountered frequent obstructions in larger trees, particularly during the process of obtaining 10 mm cores (for x-ray densitometry) in seed stands.

Consequently, the proportion of increment cores collected from younger trees was greater than that from older ones in the total of one hundred sampling trees. 1,30 meters of trees were marked, and then 5 mm and 10 mm increment cores were taken as close to each other as possible.

Data sets

The overall dataset (obtained 100 trees, 2 different cores in diameter, and x-ray and Fractometer measurements per tree) was randomly divided into two equal parts, designated as modelling (M) and testing (T) datasets. All experimental stands were represented by a minimum of three trees in both datasets. However, due to the unequal representation of experimental stands in the datasets, the impact of this unequal distribution on the correlation between fracture strength and density was additionally assessed by segregating data from younger stand and older stands (Figure 4, below; A: data from AK genetic trial , and B: data from BM, SK, BP, and G stands).

Determination of wood density

X-ray densitometry dataset, which was validated at INRA laboratories through the collection of cores from the same trees were utilized in the project, and was employed for model estimations. A mass attenuation coefficient of 3,03 cm²/g for red pine (Pinus brutia Ten.) (Guller 2010) was applied to the formula provided below (Equation 1). Fracture strength (FS) was measured using the IML Fractometer (Figure 3) in laboratory conditions. Measuring bending and compression strengths with the Fractometer is straightforward.

The recommended approach involves initiating measurements at the bending station using intact increment cores. Subsequently, the next step involves utilizing broken segments (cut to the measurement interval of 6-10 mm) for the compression unit. However, conducting measurements on the compression unit with such a small sample size can pose challenges in the field, contingent upon topography and forest structure. Conversely, the bending unit employs entire increment cores.

After each core is fractured, data is measured and automatically recorded for each broken part. Therefore, utilizing the bending unit alongside a simple model allows for swift preliminary assessments of tree density and strength. This approach proves advantageous and more feasible for tree breeders during the process of tree selection in various scenarios. This constituted the primary rationale for selecting fracture strength as an independent variable for modeling in this study.

Furthermore, Tang et al. (2016) reported a positive and significant correlation between bending and compressive strength, as measured with the fractometer, for certain subtropical urban trees.

Acclimatization for 8 % equilibrium moisture was applied in all samples to eliminate the moisture content effect between x-ray densitometry and Fractometer samples. Moisture- control samples (which were the same size and the same species, but different than study samples), were used for determining target moisture level. Internal condition of the chamber was determined by preliminary experiment to prevent cracks and, required time to reach the target moisture (8 % here) a couple weeks in acclimatization chamber, and the ending time determined by checking moisture of the control samples. To prevent shape changing, each sample was taped into a core holder.

The resultant scatter plot which exhibited definable trend was linear and, supported of the usage of the simple linear functional form. Linear regression model was constructed for estimating of wood density by using relationship between fracture strength and wood density. Fifty samples were used for constructing the statistical model (linear regression). The equal sample size was used to test the model.

Paired sample t-test (Davis 2013, McDonald 2014) used to compare model and x-ray densitometry results. Kolmogorov-Smirnov and Shapiro-Wilk tests were used to check normality of distribution (Additional Table1 for statistics). ANOVA test was applied to reveal the statistical difference among groups for the four determined parameters of which were Fracture strength (FS), density (D), annual ring width (ARW) and fracture angle (FA). The level of confidence level was chosen as 95 % for all statistical analysis. Equation 1 (Bucur 2003).

ρ: Density of sample

Results and discussion

Descriptive statistics for the two investigated groups (1: Data set A for genetic trial and 2: Data set B for seed stands) are shown in Table 2. There was not a significant difference in terms of simple linear regression model between two parameters for separated data sets (A, B) and combined ones (A+B) (Figure 4). Therefore, the combined data were used for all evaluations.

Dependent (density) and independent variables (fracture strength) mean values of model data set were 0,547 gcm^-3 ± 0,035 and 48,3 ± 9,89 MPa respectively. There was a considerably high r² value (r²=0,74; adjusted r²=0,732) between fracture strength and density (Figure 4, Table 3). Pearson correlation coefficient between the two parameters (for model training data set) was 0,859 (Table 3). The visual parallelism of measured and estimated densities of the model testing data set (Figure 5), and high correlation coefficient (0,96) between them, clearly indicates the power of model estimates.

The mean density values of measurements and estimates were found as 0,546 gcm^-3 and 0,543 gcm^-3 respectively with 0,003 gcm^-3 standard deviation, which meant the simple model is very successful for predicting mean density of population. On the other hand, because of the fact that the samples of model data set and testing data set were even randomly selected from the total amount of samples, both data set included half-siblings (from AK trial area). Therefore, this might be the reason of high parallelism between statistical estimates and x-ray measurements or at least contribute it.

Depending on ANOVA statistics for the two groups (Trees from seed stands and trees from genetic trial), with p values of 0,031<0,05, there were only a statistically significant difference for ARW (Table 4 and Table 5). However, values clearly indicate that the trees which have the highest fracture strength also have the highest density in ranking (Table 6 and Table 7).

Matsumoto et al. (2010) reported high coefficient of correlation between fractometer-measured compressive strength and basic density for five Indonesian plantation species, and based on their results for strength values of clear samples, they mentioned that fractometer could be applied for the selection of trees with high strength properties. Results from this study agreed with that idea.

Depending on literature knowledge, genetic correlations are useful to make inferences and prediction about indirect responses of one trait from the selection of other one (Neyhart et al. 2019). If two traits are genetically correlated, a response to selection would also be expected in a second trait (Neyhart et al. 2019).

Therefore, the result of study and other previous findings on high positive linear density-strength relations for pine species (Zhang 1997, Guller 2007, Kiaei 2011), furthermore positive moderate to high genetic correlation previously reported between density and wood strength for pine species (Kumar 2004, Missanjo and Matsumura 2016), all reveal that even a selection based solely on wood fracture strength, which is easily measurable with fractometer in the field, would result in improvement for density trait of red pine (Pinus brutia Ten.).

“The breaking resistance of wood is best represented by the fracture moment and the energy that is expended to break the wood sample. This energy is a function of the fracture moment and the fracture angle. The fracture angle measures the elasticity of the wood sample” (Ganesan and Abdul-Hamid 2010: referring to the book titled as Body language of trees, a handbook for failure analysis by C. Mattheck).

All these mean that if it is possible to determine breaking resistance, fracture angle and density of wood in the field, then it would be possible to make estimation for other wood quality traits at site depending on the relations among all traits. While looking at the ranking tables (Table 6 and Table 7), although there is not an exact matching in ranking among all parameters higher fracture strength means higher density, but does not means higher FA.

Elasticity of wood is a very complex topic and is affected not only by density but also different wood anatomical properties (i.e. microfibril angle) (Haygreen and Bowyer 1996). There is an indication of weak correlation between wood density and MOE for red pine (Pinus brutia Ten.) (Guller 2007).

Although the following indication of Steffenrem et al. (2007) for Norway spruce as “The strong correlations found between MOE and MOR, both on individual tree level and family-mean level, suggest that they can be treated as the same trait in practical tree breeding” found in the literature, we did not see any clear relationship between FS and FA (Table 6 and Table 7).

Some defects occurred by natural or artificial reasons, for example drying of wood may cause micro cracks, may have negative effect on elasticity (Haygreen and Bowyer 1996). The drying process of increment cores in the study might have such negative effects, particularly on larger increment cores. Therefore, the relationship should be investigated on fresh samples (increment cores which keep their moisture content). There is a previously published work of the author on the same site (Guller et al. 2011).

The population which is abbreviated as AK-D here, reported in the paper as the best performing population for ARW and shared the top performance with the mid-altitude population M1 at Kepez site (abbreviated as AK-M here) in terms of wood density. The ranking shows that some populations (i.e. AK-D: one of low altitude populations) can show superior growth, wood density and strength and it is possible to select those kind of populations in the field by using Fractometer (Table 6).

Evaluation of individual tree density in the field is also an important topic for practice. The results of paired sample t-test with p=0,034<0,05 (Table 8), indicated that although mean values of the two density groups (model and x-ray densitometry) were very close, the model did not provide exact matching in individual base at 95 % confidence level.

The highest difference between model estimated density (0,555 gcm^-3) and x-ray densitometry measurement (0,530 gcm^-3) for the same sample in testing data set was found as 0,025 gcm^-3 while the lowest difference was found as 0,001 gcm^-3 (Figure 6). Therefore, estimation of individual tree x-ray density by using the simple linear model is not exactly match.

Although statistical test (paired sample t-test) identified differences at individual base, considering the range of numerical results there is not practically important difference in terms of having pre-estimates for tree selection in practice. Results are showing the clear correlation between density and fracture strength even relatively small number of data and additionally project experiences indicated that there is a high possibility for development of a new model, of which individual estimates would be more powerful, by considering possible intervening factors and foreseeable causes of few high individual deviations in the data set.

For example, increment core samples were acclimatized to have equal moisture content for both x-ray densitometry and fractometer samples. During the moisture losing period larger cores were affected from different forces more than shorter ones. Occurrence of some invisible cracks might have caused higher decrease on fracture strength of such size of samples.

For further research, evaluation of relation between density-fracture strength on fresh samples and building new models by using the data would give better opportunities for individual estimates. On the other hand, even core sampling theoretically applied at the same trees and at the same height (the closest place to each other), in fact two different cores sampled depending on their requirements for the two compared way. In some cases, to obtain a defect free sample sampling repeated.

Another possible reason for few deviated FS values, which are not accepted as outliers in the study because of the size of model data set, could be the modified increment coring devices. Probett et al. (2013) indicated that better surface smoothness was obtained on radiate pine (Pinus radiata D. Don) increment cores by working with high-speed increment coring device than classical increment borers, and they made possible cause and effect relationship between lower FS values of samples which were taking with classical borers and their rough surface.

That opinion means the coring device influences fractometer measurements. In this work classical increment borer steel cores were adapted to two different devices which had different torque. Although complete and satisfying visual surface quality obtained for samples, different torque of the two devices might have resulted in different internal stress causing lower fracture strength for some samples.

During increment coring steps of the project, we only focused on taking unbroken and smoot surface cores and recording planned information about sampled trees. Unfortunately, we did not plan recording of which core extracted by which device. Although we know that generally 10 mm cores were extracted by high torque and 5 mm increment cores were extracted by lower one, we do not have exact recorded information on field sheets.

Therefore, we can only mention about two different coring device and possibility of using high torque device for taking some of samples in 5 mm diameter. This is the most probable reason of few highly deviated FS data that might affect power of the model and distorting parallelism of model estimates and x-ray densitometry results at some points (Figure 5).

It is crucial to notice the presence of similar and very close numeric results, along with the highest density difference which is only 0,025 g/cm³, supporting the potential of Fractometer and modeling as a practical approach for density estimation. Although future mix models using Fractometer data may have a high probability of achieving more close individual estimations, unfortunately, these models are unlikely to accurately estimate early and late wood band densities, especially within the narrow annual rings of pine species.

Conclusions

The mean density values of x-ray measurements and model estimates found to be quite close in terms of numerical values. Furthermore, there is a high correlation between fracture strength and wood density. All these results indicated that with a selection of higher fracture strength, an improvement in density can be achieved for Pinus brutia. However, to apply this method successfully, it is necessary to emphasize the importance of the surface smoothness of the increment cores.

Defects, drying defects even fine surface cracks, may affect breaking resistance value. Therefore, it is recommended to consider all these situations in future studies. Particularly investigating all relations by working on fresh increment cores recommended.

As a general conclusion, statistical modelling and Fractometer data can be used for pine tree breeding programs as a practical preliminary assessment way of wood density in the field for the purpose of population selection.

Acknowledgements

This study mostly consists of data and experiences obtained from a completed project of author financially supported by The Scientific and Technological Research Council of Turkiye (Project number: TUBİTAK 110-O-560:Determining wood density and annual ring properties of standing Turkish red pine using non-destructive methods) and author personal works. I am very appreciated for financial support and all kind of contribution related the project. Special thanks to Philippe Rosenberg and Manuela Romagnoli. Some additional measurements and statistical analysis which were not included in the project final report, done later for this paper. The study is previously presented and published only as an abstract form in the proceedings of the meeting ISFOR 2017 in Turkiye.

References:

Barnett, J.R.; Jeronimidis, G. 2003. Wood Quality and its Biological Basis. Blackwell Publishing Ltd: Oxford, UK. ISBN 1-84127-319-8. https://books.google.com.tr/books?id=NhwSMryGzY0C&pg=PR3&source=kp_read_button&hl=en&redir_esc=y#v=onepage&q&f=false

Brashaw, B.K.; Bucur, V.; Divos, F.; Gonçalves, R.; Lu, J.; Meder, R.; Pellerin, R.F.; Potter, S.; Ross, R.J.; Wang, X.; Yin, Y. 2009. Nondestructive testing and evaluation of wood: A Worldwide research update. Forest Products Journal 59(3): 7-14. https://www.fpl.fs.usda.gov/documnts/pdf2009/fpl_2009_brashaw001.pdf

Bucur, V. 2003. Nondestructive characterization and imaging of wood. Springer-Verlag: Berlin Heidelberg. ISBN 978-3-540-43840-3. https://doi.org/10.1007/978-3-662-08986-6

Chiu, C.M.; Wang, S.Y.; Lin, C.J.; Yang, T.H.; Jane, M.C. 2006. Application of the Fractometer for crushing strength: Juvenile-mature wood demarcation in Taiwania (Taiwania ryptomerioids). Journal of Wood Science 52(1): 9-14. https://jwoodscience.springeropen.com/counter/pdf/10.1007/s10086-005-0723-x.pdf

Cown, D. 2005. Understanding and managing wood quality for improving product value in New Zealand. New Zealand Journal of Forestry Science 35(2/3): 205-220. https://www.scionresearch.com/__data/assets/pdf_file/0020/59114/08_COWN.pdf

Cown, D. 2006. Wood quality in standing timber - evolution of assessment methods in plantations. Wood Structure and Properties ́06. Edited by S, Kurjatko.; J, Kúdela.; R, Lagaňa. Arbora Publishers: Zvolen, Slovakia. 11-17. ISBN 80-968869-4-3.

Davis, C. 2013. Chapter:6. Basic Statistical Testing. ISBN: 9780643107120. In: SPSS for Applied Sciences. CSIRO publishing: Australia. https://www.publish.csiro.au/book/6900

Dolwin, J.A. 1996. Evaluation of internal defects in trees and the legal implications. Arboricultural Journal 20(2):173-178. https://doi.org/10.1080/03071375.1996.9747112

Ganesan, S.K.; Abdul Hamid, M. 2010. Survey of wood strength properties of urban trees in Singapore using the fractometer II. Journal of Tropical Forest Science 22(1): 97-105. https://www.jstor.org/stable/23616695

Goetz, K.; Bethge, K.; Mattheck, C. 2002. The Fractometer II - a mobile wood-testing device. Institut für Materialforschung. Report nummer FZKA-6704. https://publikationen.bibliothek.kit.edu/270051621/3814177

Guller, B. 2007. The effects of thinning treatments on density, MOE, MOR and maximum crushing strength of Pinus brutia Ten. Wood. Annals of Forest Science 64: 467-475. http://dx.doi.org/10.1051/forest:2007024

Guller, B. 2010. Kızılcam’da (Pinus Brutia Ten.) Odun yogunluğunun x-ray yogunluk olcer ile belirlenmesi. Suleyman Demirel Universitesi Orman Fakultesi Dergisi A(2): 97-109. https://dergipark.org.tr/tr/download/article-file/195756

Guller, B.; Isik, K.; Cetinay, S. 2011. Genetic variation in Pinus brutia Ten.: Wood density traits. BioResources6(4): 4012-4027. https://bioresources.cnr.ncsu.edu/BioRes_06/BioRes_06_4_4012_Guller_IC_WoodDensityTraits_2015.pdf

Haygreen, J.G.; Bowyer, J.L. 1996. Forest Products and Wood Science. 3 Edition. Iowa State University Press, US. ISBN 0-81382-256-4

Isik, F. 1998. Genetic variation, heritabilities and genetic gain from Pinus brutia open-pollinated progeny trials. Southwest Anatolia Forest Research Institute. Tech. Bull.:7. ISBN 975-8273-08-6. https://www.ogm.gov.tr/batiakdeniz/yayinlarimiz-sitesi/Yayinlarimiz/Teknik%20B%C3%BClten/Teknik%20B%C3%BClten%20No%2007%20K%C4%B1z%C4%B1l%C3%A7amda%20(Pinus%20brutia%20Ten.pdf

Kandemir, G. 2013. TURKEY: First National Report on Forest Genetic Resources Country. Country Report for the FAO First State of the World’s Forest Genetic Resources for Food and Agriculture. FAO publication. https://www.fao.org/3/i3825e/i3825e71.pdf

Kiaei, M. 2011. Anatomical, physical, and mechanical properties of eldar pine (Pinus eldarica Medw.) grown in the Kelardasht region. Turkish Journal of Agriculture and Forestry 35(1): 31-42. https://doi.org/10.3906/tar-1001-552

Kumar, S. 2004. Genetic parameter estimates for wood stiffness, strength, internal checking, and resin bleeding for radiata pine. Canadian Journal of Forest Research 34(12): 2601-2610. https://doi.org/10.1139/x04-128

Kraler, A.; Beikircher, W. 2013. Non-Destructive and Semi-Destructive Test Methods for Strength Determination of Aged Wood. Advanced Materials Research 778: 385-392. https://doi.org/10.4028/www.scientific.net/amr.778.385

Lin, C.J.; Wang, S.Y.; Chiu, C.M. 2007. Crushing strength sampling with minimal damage to Taiwania (Taiwania cryptomerioides) using a fractometer. Wood and Fiber Science 39(1): 39-47. https://wfs.swst.org/index.php/wfs/article/download/1136/1136/0

Lin, C.J.; Wang, S.Y.; Chiu, C.M. 2004. Assessment of crushing strength in Taiwania using the fractormeter. Taiwan Forest Prod Indus 23(1): 23-31. https://cir.nii.ac.jp/crid/1574231875251576960

Matheny, N.P.; Clark, J.; Attewell, D.; Hillery, K.; Graham, A.W.; Posner, G. 1999. Assessment of fracture moment and fracture angle in 25 tree species in the United States using the fractometer. Journal of Arboriculture 25(1): 18-23. https://doi.org/10.48044/jauf.1999.003

Matsumoto, K.; Ishiguri, F.; Lizuka, K.; Yokota, S.; Yoshizawa, N. 2008. Evaluation of bending and compression strength of wood using Fractometer. Wood Industry 63(8): 358-363. https://agris.fao.org/search/en/providers/122558/records/64724a9053aa8c89630557db

Matsumoto, K.; Ishiguri, F.; Wahyudi, I.; Takashima, Y.; Shimizu, K.; Lizuka, K.; Yokota, S.; Yoshizawa, N. 2010. Application of Fractometer for wood property evaluation in five Indonesian plantation species. Bulletin of the Utsunomiya University Forests 46: 1-6. https://agriknowledge.affrc.go.jp/RN/2010792439.pdf

Mattheck, C.G.; Breloer, H.; Bethge, K.A.; Albrecht, W.A.; Zipse, A.W. 1995. Use of the Fractormeter to determine the strength of wood with incipient decay. Journal of Arboriculture 21(3):105-112. https://joa.isa-arbor.com/request.asp?JournalID=1&ArticleID=2670&Type=2

McDonald, J.H. 2014. Handbook of Biological Statistics. 3rd ed. Sparky House Publishing: Baltimore, Maryland. http://www.biostathandbook.com/pairedttest.html

Missanjo, E.; Matsumura, J. 2016. Genetic improvement of wood properties in Pinus kesiya Royle ex Gordon for sawn timber production in Malawi. Forests7(11): e253. https://doi.org/10.3390/f7110253

Neyhart, J.L.; Lorenz, A.J.; Smith, K.P. 2019. Multi-trait improvement by predicting genetic correlations in breeding crosses G3-Genes. Genomes Genetics 9(10): 3153-3165. https://doi.org/10.1534/g3.119.400406

Nocetti, M. 2008. Miglioramento genetico delle piante per la produzione legnosa conparticolare riferimento ai caratteri del legno. Forest@ - Journal of Silviculture and Forest Ecology 5: 112-120. https://foresta.sisef.org/contents/?id=efor0506-0050112

Panshin, A.; de Zeeuw, C. 1980. Textbook of wood technology. McGraw-Hill Inc.: New York, US. 122p.

Probett, P.S.E.; Craig, C.S.; Probett, B.J. 2013. In situ structural timber strength measurement advances using qualitative resistography and quantitive resisto- fractometry. 18th International Nondestructive Testing and Evaluation of Wood Sym. Madison-WI, US. https://inspectapedia.com/structure/In_Situ_Structural_Timber_Strength_Measurement_Advances_Probett_2013.pdf

Ruotsalainen, S. 2014. Increased forest production through forest tree breeding. Scandinavian Journal of Forest Research 29(4): 333-344. https://doi.org/10.1080/02827581.2014.926100

Steffenrem, A.; Saranpää, P.; Lundqvist, S.O.; Skrøppa, T. 2007. Variation in wood properties among five full-sib families of Norway spruce (Picea abies). Annals of Forest Science 64: 799-806. https://doi.org/10.1051/forest:2007062

Tang, A.M.C.; Chu, P.P.L.; Leung, M.W.K.; Chu, L.M.; Liao, W.H. 2016. Evaluating wood strength properties of subtropical urban trees using fractometer II. Journal of Tropical Forest Science 28(3): 249-259. https://www.jstor.org/stable/43856529

Wu, H.X.; Eldridge, K.G.; Matheson, A.C.; Powell, M.B.; McRae, T.A.; Butcher, T.B.; Johnson, I.G. 2007. Achievements in forest tree improvement in Australia and New Zealand 8. successful introduction and breeding of radiata pine in Australia. Australian Forestry 70 (4): 215-225. https://doi.org/10.1080/00049158.2007.10675023

Yildirim, K.; Ozturk, H.; Siklar, S.; Balkiz, O.D.; Kaya, Z. 2011. Strong genetic control of high wood specific gravity in young progenies of Pinus brutia: Potential of early selection for industrial plantations. Silvae Genetica 60: 249-258. https://doi.org/10.1515/sg-2011-0033

Zobel, B.J.; Talbert, J. 1984. Applied forest tree improvement. New York: John Wiley & Sons. 505p. https://blackburnpress.stores.yahoo.net/apfotrim.html

Zobel, B.J.; Buijtenen, J.P. 1989. Wood Variation Its Causes and Control. Springer Series in Wood Science. Springer: Berlin, Heidelberg. https://link.springer.com/book/10.1007/978-3-642-74069-5

Zhang, S.Y. 1997. Wood specific gravity-mechanical property relationship at species level. Wood Science and Technology 31:181-191. https://link.springer.com/content/pdf/10.1007/BF00705884.pdf

Author notes

^♠Corresponding author: bilginicel@comu.edu.tr