JARQ 28, 268-275 (1994) A Laboratory Method for Predicting the Durability of Tropical Hardwoods Koichi YAMAMOTO* and Lay T HONG** '"Wood Chemistry Division, Forestry and Forest Products Research Institute (Tsukub a, Ibaraki, 305 Japan) '"'"Forest Research Institute Malaysia (Kepong, 52109 Kuala Lumpur, Malaysia) Abstract The possibility of predicting the resistance of tropical hardwoods to fungal decay was studied A multiple regression equation for predicting the natural durability of tropical timbers was proposed The weight loss of 24 commercially popular Malaysian hardwoods was obtained by a modified ASTM D2017 soil block method using white-rot fungi, Corio/us versicolor, Ganoderma lucidum, Pycnoporus coccineus, and a brown-rot.fungus, Tyromyces palustris The percentage of weight loss induced by T palustris was selected as a criterion variable (Y) Various properties of the woods which could affect the durability of timbers, such as extractives content, lignin content, pH, density, water absorption capacity, and morphological characteristics were analyzed The factors selected as independent variables based on the significance of coefficients for the weight loss induced by T palustris were the density (Xl), water absorption capacity (X2), methanol extractives content (X3), and pH (X4) The weight loss calculated by the equation was the lowest in Cheng al (Neobalanocarpus heimii) The regression equation obtained was Y = 18.546 - 37 028 (Xl) + 0.016 (X2) - 1.056 (X3) + 8.056 (X4) The durability of the timbers estimated by the equation matched the natural durability obtained from field stake tests with minor differences Discipline: Forestry and forest products Addition al key words: decay fungus, extracti ves, natural durability Introduction Natural durability is one of the important factors to ana lyze the nature of wood The durability of timbers is determined based on data obtained through field trials over a certain period of time Very often these trials involve long-term exposure of the timbers to biodegrading organisms in the field Such field tests give good estimates of the natural durability because the timbers are exposed to both types of biodegrading fa una and flora, i.e insects and fungi, but it takes a long time before meaningful data can be obtained Although laboratory met hods have been used for studies on the resistance of timbers to both fungal and insect attacks, they also require a long period of time and technical skill for dealing with fungi and insects Therefore simple and rapid laboratory methods to predict the natura l durabi lity of tropica l timber species sho uld be developed > The decay resistance of wood has been shown to be associated with the amount of extractives •15•20 •24>, the type of lignin · 10 •16>, lignin content 19>, density 13 •17 •20>, 269 Y, and wood anatomy 19 •22 •23 > In order to st udy the possibility of predicting the resistance of tropical hardwoods to funga l decay without using fu ngi, various chemical and physical properties of the 24 timber species related to the durab ility, i.e the amount of ex tractives, lignin co ntent, pH, dens ity, water absorption capacity, and morphological characteristics were determined A multiple regression equatio n for predicting the natural durability was proposed by using several factors as independent variables Materials and methods The 24 tropical timber species used in this study and shown in Table I were selected from the list of "Commercial timbers of Peninsular Tobie Trade name Balau Chengal Giam Keranji Merbau Rcsak Kapur Kempas Kerning Mata ulal Punah Rcngas Bintangor Durian Jclutong Meranti bakau Mcranti, dark red Meranti, white Meranti, yellow Mera wan Mersawa Pcrupok Ram in Rubbenvood Malaysia" 11• A modified version of the ASTM 02017 testing procedure for decay 21 was app lied to evaluate the resistance of the timbers LO the decay induced by white-rot fungi, Corio/us versicolor (strain FPPRI 1030), Ganoderma l ucidum (FFPR I GI I), Pycnoporus coccineus (FFPR I Ps lh), and a brown-rot fungus, Tyromyces pa/ustris (FFPRT 0507) Six wood blocks/timber/fungus, 2.0 by 2.0 by 0.5 cm, were subjected LO the fungal decay rest The percentage of weight loss of each timber species obtained by the test du ring J weeks of incubation with the fungi was compared to the fo llowi ng characteristics to derive a mu lti ple regression equation for the estimation of natural durabi li ty Various characteristics of the timber species, namely cold water ex tract ives content (JIS P8004 - 1976) 12), hot water Average weight loss percentage or 24 timber species subjected to decay by fuugi for 12 weeks Scientific name Shorea sp Neobala11ocarp11s heimij Hopea sp D iali11111 sp lntsia palembanica Vatica sp D ryobalanops aro111atica K oompassia mataccensis DipterocaqJuS sp Kokoo11 N 271 Y,111111111010 ct "'·: Prediction of D11mbili1y of Tropical Ht1rdh'oocls extrac tives content (JJS P8005 -1 976) 12 >, ethanol-benzene extractives contelll (HS P8010-1976) 12 >, methanol extractives content, Klason lignin content (T APPi standards T 13wd - 74) 18>, pH value (wood to water ratio of I: 10l , density, water absorption capacity (JIS Z2J04-1957) 12>, proportion of wood elements, and ratio of cell-wall area of wood fiber were determined for each of the 24 timber species The measurements of the wood element proportion and the fiber cell-wall area were carried out on cross-sections in 15 J(m th ick The natural durability of tropical timber species has been examined by field stake tests in different cnvironments 1•3 •4 ·1L,a • >, The est imated natural durability was compared to the durability classification obtained in the field tests Uesults and discussion T he average percentage of weight loss of the 24 timber species subjected to decay induced by T palustris, C versico/01~ P coccineus, and G /ucidum is given in Table I The value of weight loss was greater in most of the timber species degraded by T paluslris compared wit h the other fungi Various characterist.ics measured such as cold water extractivcs content (%), hot water extractives contem (%), ethanolbenzene extractives content (% ), methanol extractives content (%), Klason lignfa content (%), pH value, density (g/cm 3), water absorption capacity for one day (mg/cm 2), proportion of fibers (%), proportion of parenchyma cells (%), proportion of vessels (%), and ratio of cell-wall area of wood fiber (OJo) are indicated in Table The correlation coefficients between the percentage of weight loss and the various parameters were generally higher in the samples subjected to decay by T pa/ustris than in those by the other fungi (Table 3) Therefore, the percentage of weight loss induced by T palustris was selected as a criterion variable (Y) Table shows that there was a significant correlation between the values of the parameters such as density, water absorption, contents of ethanol-benzene extractives and methanol cxtractivcs, pH, and fiber cell-wall area and the weight loss for all fungi Among the kinds of extractivcs content, methanol extractives content was an important factor for every fungus used Decay resistance in tropical hardwood timbers appeared to be mainly Table Correlation coefficients between weight loss percentage induced by decay fungi and various parameters Weight loss (0/o) Parameters Tyromyces palustris -0.626·· Density Water absorption 0.562° Cold water extraccives -0.623*" -0.623 Hot wacer excractives Ethanol-benzene extrnctives -0.589*" -0.685 Methanol extraccives pH 0.562** Lignin content (0/o) - 0.157 0.304 Percentage or fibers -0 181 Percentage or parenchyma -0.35 Percentage or vessels -0.461* Ratio of fiber cell wall * Significant at 5% level, •• Signil'icant at l "lo level Corio/us versicolor Pyc11oporns coccineus Ganoderma /ucidwn -0.683** 0.541** -0.538 -0.594 -0.557** - 0.615 0.535** -0.136 0.389* -0.295 -0.391* -0.610** -0.551*" 0.579** -0.341 -0.383* -0.399* -0.410* 0.615** -0.433 0.263 -0.069 -0.400* -0.484** -0.465** 0.517 -0.315 -0.315 -0.375* -0.375* 0.626•* -0.45 2* 0.236 -0.023 -0.383* -0.407* 272 JARQ 28(4) 1994 T able Density Water ab Cold w exl Hot w ext Et-ben ext Metha ext pH Lignin Fibers Parenchyma Vesse ls Fiber wall Density Water ab Cold w ext Hot w ext Et-ben ext Mctha ext pH Lignin Fibers Parcnchyma Vessels Fiber wal l Corch,tion ma lrix among lhc charnc1eris1ics of tim bers Density Water ab Cold w ext Hot w ext Et-ben ex1 Metha ext -0.372 0.343 0.382 0.384 0.375 -0.194 0.1 09 -0.199 0.264 0.130 0.802 -0.176 -0.244 -0.383 -0.303 0.360 -0 198 -0.004 0.204 -0.204 -0.192 0.979 0.864 0.949 -0.344 -0.274 - 0.555 0.393 0.532 0.267 0.850 0.978 -0.415 -0.201 - 0.480 0.328 0.468 0.274 0.905 -0.419 -0.365 -0.509 0.29 0.556 0.346 -0.489 -0.21 l -0.492 0.334 0.485 0.310 pH Lignin Fibers Parenchyma Vessels Fiber wall -0.241 0.213 -0 104 -0.261 -0.255 0.390 -0.257 -0.38 -0.081 -0.802 -0.867 -0.470 0.4[7 0.495 0.37 due to the presence of polyphenolic compounds which are soluble in methanol The wood element proporiion was not ou tstandingly important Lignio content was not closely associated with decay resistance The type of lignin is a factor more important in the slower degradation of woods than is the lignin content, or t he anatomical structure itse1r 10> The factors selected as independent variables based on the significance of the coefficients for weight loss and multicollinearity were the density (X l ), water absorption capacity (X2), methanol extractives contents (X3), and pH (X4) Although the ratio of the cell-wall area of fiber showed a highly significant correlation coefficient, it was noi used because it was not a independent variable Table shows the correlation matrix among the factors studied The fiber cell-wall area ratio showed a high correlation coefficient with the density Multiple regression analysis enabled co derive the following regression equation for predicting natural durability: Y = 18.546 - 37 028 (XI)+ 0.016 (X2) - 1.056 (X3) + 8.056 (X4) where Y; estimated percentage of weight loss induced by Tyromyces palustris, XI; density, X2; water absorption, X3; methanol extractives contents, X4; pH The partial correlation coefficients of XI, X2, X3 and X4 were - 0.522**, 0.374*, - 0.509** and 0.348, respectively A multiple correlation coefficient of 0.86 and coefficient of determination of 74 were obtained for the regressors, X I , X2, X3, X4, indicating that these factors, XI, X2, Ya111a111010 er al.: Prediction of Durability of Tropical Hardwoods X3, X4 accounted ror 740Jo of the variation in weight loss as a measure of resistance The measured value of resistance of 24 timber species indica1ed by the weight loss val11es ob1ained in the laboratory decay test and I he estimated values obtained by the equation are shown in Table These sets of values showed small differences excep1 for Merawan, Keruing and Kapur Merawan showed a su rprisingly large difference between the measured and calcula1ed values Hence the measured dnrabilily of Merawan was classified as very high, and the estimated one as moderate Merawan was classified into a durable group, bu1 in which the durability was reduced after extraction, according to rhe cluster analysis using the weight loss of the unextracted and ext ractecl wood blocks 21 > Table Measured nnd ci1lculuted weight loss values of 24 timber species used Trade name Balau Chengal Giam Keranji Merbau Resak Kapur Kempas Kcruing Mala ulat Punah Rengas Bintangor Durian Jelutong Meranti bakau Mcranti, dark red Meranti, white Meranti, yellow Mcrawan Mcrsawa Perupok Ramin Rubbcrwoocl Measured Calculated value value 3.2 0.0 1.3 2.9 0.2 7.6 4.4 19.5 45.1 39.0 46.0 2.8 27 47.7 54.3 37.2 37.1 46.9 44.7 3.7 44.0 48.7 57.4 58.6 9.0 -10.4 0.4 11.6 12.8 1.8 24.2 23 25.4 31.0 51.5 1.8 40.0 37.0 47.4 32.2 32.6 37.3 35.5 34.8 41.6 52.0 56.9 50.6 Error 5.8 J0.4 0.9 8.7 12.6 5.8 19.8 3.6 19.7 8.0 5.5 1.0 12.2 10.7 6.9 5.0 4.5 9.6 9.2 31.1 2.4 3.3 2.5 8.0 273 The estimated values from the regression equation were used for determining the natura l durability classificat ion Timber species arc usually grouped into classes in terms of their durability 9> The natural durabi lil'y is grouped into classes in Malaysia 1• 11 •14>, and classes in USA >, England 3l, and Japan 13 > The service life in the classification is also quite different depend ing on the exposure tests carried out in different environments The service life in the highly durable class exceeds JO years in Malaysia 1>, years in Japan 13>, and 25 years in USA > and England 3> In this experimenL, a timber with less lhan 5% of estimated weight loss was classified as highly durable (code I); 5- l 50Jo weight loss as durable (code 2); I - 35o/o weight loss as moderately durable (code 3); 35 -45% weight loss as nondurable (code 4); and over 450Jo weight loss as perishable (code 5) The estimated durability classification thus obtained matched the classificat ion obtained from field stake tests with minor differences (Table 6) Therefore the regression eq uation could be used to estimate the durnbi lity of a timber species whose resistance to decay is not known yet and where such data are required before field exposure tests arc completed The natural durability of tropical timber species determined by field stake tests carried out in different countries is markedly different in some timbers such as Balau 1•3 •4 • 11 • 1.1,1 (Table 6) It is generally recognized that there is a considerable variation in durability between differem samples of the same species, and some samples of certain timber species fall into a class above or below that in which they are normally classified 9> Balau which is classified as highly durable in Malaysia 1> and Japan 13> is classified as nondurable in USA 4> This discrepancy is due Lo lhe fact that in lhese stake tests differem species in a group arc util ized and each species displays its own natural durability It is important 10 assess the natura l durability of each timber species using the JARQ 28(4) 1994 274 Table Comparison· or na lural du rabilily classific.:llion of 24 timber species Natural durability classificat ion Trade name A I) Oalau Chengal Giam Karanji Merbau Resa k Kapur Kem pas Kerning Mata ular Punah Rengas Bintangor Durian Jelurong Mcranti bakau Meranii, dark red Meranti, white Merant i, yellow Merawan Mersawa Perupok Ramin Ru bberwood I I Ma laysia B II) I I 4 CH) USA 41 3 Japan 13> (4) I I England·1' 3 1-2 3 I J -3 I J 3 3 3 3 4 4 2 3 3 4 This report 3-4 3-4 3 2-3 4 4 5 4 3 4 3 4 4 5 laboratory method proposed Conclusion The regression equation obtained in this study could be used lo estimate the durabili ty of a timber species whose resistance to decay is not known yet and where such data are required before ricld exposure tests are completed The regression eq uation could be further improved by using additional data from a large number or timber species References I) Anonymous (1975): Properties and uses of commercial timbers of Peninsu lar Malaysia In Malaysian Forest Service Trade Leanct No 40 I 4 3 3 3 4 4 5 5 5 Malaysian Timber Industry Board 2) Anonymous ( 1977): Standard methods of accelerated laboratory test of natural decay resistance of woods /11 Annual book of ASTM standards Vol.04.09 Phi ladelphia, American Society for Testing and Materials 3) Anonymous (1979): Timbers of the world Vol I Timber Research and Development Association The Construction Press, England 4) Chudnoff, M (1984): Tropical t imbers in the world Agricu lture handbook No 607 United States Department of Agriculture Forest Service, USA 5) Da Costa, E W 8., Rudman, P & Gay, F J (1961): Relationship of growth rate and related factors to durability in Tec1011a gmndis Emp For Rev , 40, 308 - 319 6) Da Costa, E W 8., Rudman, P & Deverall, F J (1962): Inter-tree variation i 11 decay resistance of karri (£11calyp1us diversicolor F Muell.) J Inst Wood Sci., 2, 48 - 55 Yt1111