Leslie H. Groom

AbstractThe structural perfonnance of wood trusses, which

are now commonplace in light-fraine construction, isdictated in part by the mechanical properties of thetruss-plate joints. However. little Infonnation existsquantifying the effect of enVironmental conditions ontruss-plate joint properties. The main objective of thispaper was to quantify the effect of moisture cycling onthe mechanical properties of truss-plate joints. Asecondary objective was to evaluate the possibility ofretarding the degenerative effects of moisture cyclingby means of an adhesive applied to the teeth of thetruss plate immediately before assembly. The resultsindicate that mild moisture cycling decreases truss-plate joint strength and stiffness by approximatelytwice that of specimens subjected to a constant mois-ture content. More severe moisture cycling acceleratedthis decrease by a factor of about three. Addition of anadhesive to the teeth decreased plate backout andincreased the mechanical properties of the truss-platejoints. The increase in initial stiffness caused by theadhesive was minimal after moisture cycling, and theimprovement in ultimate load remained substantialeven after eight severe moisture cycles.

ing it possible to assign design values based on theirmechanical behavior.

Recommended design values for truss membersare dictated primarily by the matertals compoSing thetruss. the type of truss. and the expected loadingduring the life of the truss (26). The design plan laidout by architects and engineers for a particular struc-ture generally dictates the type of trusses that will formthe framework of the roof and floor systems. Therequired mechanical properties of the trusses aredetermined by defining the overall lifespan of thestructure and by the loading as prescribed by localbuilding codes. The truss manufacturer engineers thetrusses based on required truss mechanical proper-ties that meet or exceed current Truss Plate InstituterrPI)designvalues.

Truss performance is in most cases a reflection ofthe properties of the joints (31). Failure of trusssystems occurs primarily at the joints because ofstress concentrations. whereas large deflections intruss systems are generally the result of small jointdisplacements (22). Thus. evaluation of truss perform-ance must begin with an understanding of the me-chanical behavior of truss-plate joints.

In recent years. models of truss-plate joint behaviorbased on finite element analysis (3.8) and beam-on-inelastic-foundation theory (9.11) have been devel-oped that define the static structural behavior of woodmembers and truss plates based on mechanical prop-erties. However. both models concentrate on the fun-damental mechanisms of stress transfer and Ignoresecondary variables such as load duration and cyclicfluctuations of wood moiSture content (MC).

tight-frame construction has been moving awayfrom the complete on-site assembly practices of the1960s and 1970s and toward assembly of prefabri-cated sections. Hundreds of millions of trusses havebeen constructed In the past two decades In suchstructures as residential. commercial. and farm build-Ings: Parallel floor trusses and roof trusses withpitched top chords are now commonplace. with rooftrusses representing the majority of the low-rise resi-dential roof market. The joints of these all-woodtrusses are connected almost exclusively using l1ght-gauge steel plates with die-punched teeth. commonlyreferred to as truss plates. Truss plates are proprie-tary. with different sizes. types. and assemblies. How-ever. they exhibit stmilar structural properties. mak-

The author is a Research Forest Products Technologist.

USDA Forest Service. Southern Forest Expt. Sta.. 2500

Shreveport Hwy.. Pinev1lle. LA 71360. The use of trade orann names In this publication is for reader infonnation anddoes not imply endorsement by the USDA of any productor service. This paper was receiVed for publication In March1993.e Forest Products Society 1994.

Forest Prod. J. 44(1):21-29.

21FOREsr PRODUcrs JOURNAL Vol. 44. No. I

Trusses are often found In the most envIronmen-tally volat11e areas: attics and basements. Joist andrafter MC In these enVironments has been shown tofluctuate seasonally by anywhere from 10 percent(5.6) to as much as 18 percent (2.4.24). Even trussessubject to sheltered outdoor conditions experiencedMC values that ranged from 10.6 to 15.5 percent overa I 0 -year period (3 1 ). 1b18 type of m 0 Is ture cycling cansignificantly decrease truss performance. Unfortu-nately. the effect of moisture cycl1ng has not beenquantified. As a result. design values have beenadopted that are conseIVatlve due to the l1m1ted dataavailable. 1b18 makes for Inefficient use of lumberwhose mechanical properties are continually decreas-Ing.

for by dJvtding the average load of 1) five moIsture-con-trolled specimens (MC - 15%:t 4%) by 3.0. or 2) fivemoisture-response specimens (MC - 15% assembly.then 7%. then greater than 10%) by 2.5. Adequateunderstanding of environmental vaI1ables on truss-plate joint behavior is necessary to assign designvalues for truss manufacture.

Thus. the primary objective of this study was toquantify the effect of fluctuations in wood MC ontruss-plate joint behavior. Spedftcally. this studycompared the effect of Severity and number of mois-ture cycles with truss-plate backout. joint stiffness.and joint strength. A secondary objective was to inves-tigate ma1nta1n1ng interfada1integrtty of truss-plateteeth and wood by applying an epoxy adhesive at theinterface and noting the changes in truss-plate jointmechanical behavior.

Materials and procedure

Ezperimental deailnA randomized block design. summarized in Table

1. was used to m1n1mize the high variability custom-arily seen in lumber mechanical properties. Two sepa-rate blocks were constructed to evaluate the effect ofwood MC ranges and number of moisture cycles onvarious physical and mechanical properties of truss-plate joints. The two blocks were standard truss-platejoints and truss-plate JOints with an adhesive appliedat the tooth-wood interface.

Each block evaluated two factors: the severtty ofMC levels to which truss-plate joint specimens weresubjected and the number of MC cycles. The severityof MC level was evaluated by subjecting 10 replicatesfor each factor to the following MC conditions: 1) heldconstant at 12 percent MC: 2) cycled between 9 and15percentMC: and 3) cyc1edbetween 5 and 19 percentMC. The rate at which subsequent moisture cyclesdegraded truss-plate joints was evaluated by subject-ing 10 replicates to O. 1.2.4. or 8 MC cycles. ThIs madefor a total of ISO truss-plate joint specimens per block(3 vaJY1ng MC ranges x 5 possible numbers of MCcycles x 10 replicates! cell- ISO specimens). Althoughthe material used to construct the replicates in theblocks was from different boards. the boards werechosen to have comparable specific gravity (SG) andstiffness values. thus allowing for a relative compara-bility between blocks.

A workshop on structural wood research (13) iden-tified long-term response to loading and variable en-Vironment as a research topiC of high priority. Al-though moisture cycling has been shown tosign1ftcantly affect mechanical properties of nailedjoints (20).l1m1ted data exist that relate enVironmentalconditions and truss-plate jOints. In one study. truss-plate joints made of laminated veneer lumber and pineand subjected to an outdoor enVironment for 1 yearshowed a substantial decrease in stiffness andstrength (19). The decrease in mechanical propertieswas due in part to "plate backout, W a phenomenon in

which the truss plate backs out of the wood membersbecause of shrtnking and swe1l1ng stresses. Platebackout stgnitlcantly weakens the interfaCial contactbetween the truss plate teeth and wood. Plate backouthas also been obseIVed in other studies (10.21.27).

Several studies have evaluated the effect of chang-ing MC under a constant load for timber connectors(7,14.16-18.29). The results of these studies share onecommon theme: the combined effect of changtng en-Vironment and loading s1gniflcantly decreases trussstiffness. In most cases, strength was also stgnift-cantly reduced. However. the research neither quan-tified the effect of moisture cycling on stiffness orstrength nor defined the level at which moisturecycling proved sign1ficant.

The lack of quantifying data regarding enViron-mental and loading effects presents a d1fflculty intruss design. Design specifications for duration ofloading based on the Madison Curve are available fromTPI (26). The effect of moisture cycling is compensated

JANUARY 199422

4~"'~

l~~t";o-~ ~~-i( l [-r

'f l.

JiI; ~-

Figure 1. - Typical truss-plate joint used in this study, showingrestraining clamps and LVDT.

Joint construction and testinal'1gure 1 shows tl1e standard truss-plate joint used

in this study. In accordance witl1 TPI-8S (26). all teethwere removed within a lumber end distance of O.Sinch. The teetl1 were removed witl1 a milling bit tl1atcut tl1e teetl1 at tl1e plate surface. Thirty-Six pairs ofteetl1 were present in the upper member and 24 pairsof teetl1in tl1e lower member. This asymmetry ensuredfailure in tl1e lower member. which was equipped witl1linear vartable displacement transducers (LVD11.

An epoxy resin witl1 a cure time of IS minutes wasapplied witl1 a sponge to tl1e tips of the truss plate teetl1immediately before pressing into tl1e wood memberspreconditioned to an MC of 12 percent. An average of0.0031 g of epoxy was applied to each tootl1. witl1 tl1eresin spread along tl1e tootl1length during pressing.Truss plates were pressed into tl1e wood members soas to make intimate contact but not so deeply tl1at tl1eplate surfaces became embedded in tl1e wood.

After assembly. all 300 truss-plate joints wereplaced in a hygrotl1ermally controlled standards roomat 79. F and 66 percent relative humidity (RH) for about4 weeks to allow for relaxation of stresses induced bypressing. After conditioning. 50 standard truss-platejoints and 50 adhesive-interface joints each wereplaced in one of three hygrotl1ermally controlled unitsset to cycle between tl1e following MC ranges: 9 to ISpercent. S to 19 percent. and a constant 12 percent.The conditions for tl1e 9 to IS percent moisture cyclingranged from approximately l00.F. 40 percent RH to8S.F. 8S percent RH. The MC conditions for tl1e S to19 percent moisture cycling ranged from approxi-mately 110.F. IS percentRH to 8S.F. 9S percentRH.respectively. Ten standard truss-plate joints and 10adhesive-interface truss-plate joints were removedfrom tl1e hygrothermany controlled units at tl1e com-pletion of 0 (no cycling). 1.2.4. and 8 cycles and placedin tl1e standards room for approximately 4 weeksbefore testing. The constant 12 percent specimenswere held at a constant MC in tl1e standards room fortl1e same duration as tl1e O. 1. 2. 4. and 8 cyclespecimens. The constant 12 percent MC specimensserved as a control. allowing distinction between tl1eeffects of time and moisture cycling. Mid-deptl1 MCwas monitored during cycling witl1 a handheld surfacemoisture meter. It should be noted that all moisturewas atmospheriC in nature; at no time were tl1e jOintsexposed to direct surface water.

Loading was applied in tension by a 30.000-pound-capacity. screw-driven crosshead testing maChine(Fig. 1). A constant displacement rate of 0.01 Sin./min.was applied to produce failure in S to 10 minutes. Thespecimen was attached to tl1e testing maChine witl1universal joints to eUminate potential moments pro-duced by misalignment. The upper member was fur-tl1er restrained by tl1e attachment of two C-clamps.which made end-matching of tl1e wood blocks unne-cesary. Joint slip was monitored by LVDTand load wasmonitored witl1 a 10.000-pound-capacity load cell.Four simultaneous readings were taken every secondfrom LVDT. located in pairs at tl1e two opposite sides

Lumber and truss plate.The standard tnlss-plate joints were constnlcted

With No.2 grade KD 15 southern pine lumber repre-senting a narrow range ofSG and modulus of elasticity(MOE). MOE of each board was estimated by measur-ing centerline deflection over a simple span. SG wasbased on volume at time of test. MC was determinedon an approximately 1.5- by 1.5- by 3.5-inch block ofwood removed adjacent to the tnlss plate after testing.Dimensions of the boards were nominal 2 by 4 inchesby 18 feet in length. The boards were sampled from alocal lumber retailer in central Louisiana.

To determine the effect of an adhesive interface. asecond block of replicates was constnlcted that wassim1lar in size and properties to the standard tnlss-plate joint block of replicates. Each block consisted ofan equal amount of pith-associated and non-pith-as-sociated lumber. The SG of the standard block was0.488 and the MOE was 1.76 x 106 psi. The adhesiveinterface block had an SG of 0.485 and an MOEof 1.86x 106 pSi.

The die-punched tnlss plates used in this studywere Gangnail GN20 type plates made of 20-gaugegrade C sheet metal. 3 by 4 inches. with an average of8.0 teeth per square Inch. and an average tooth lengthof 0.360 inch.

23FOREST PRODUCTS JOURNAL Vol. 44. No. I

EFFECT OF MOISTURE DIFFERENTIAL ON PLATE BACKOUT- ~ - of..,-.. per - , - 10 EFFECT OF MOISTURE DIFFERENTIAL ON INmAL STIFFNESS-"';-. - of .., ~ - ,- - 10

iIe

. I , , . s . , . .NUMBER OF MOlmJRECVas

Figure 5. - Stiffness between 700 and 1,100 pounds for stand-ard truss-plate joints subjected to 0, 1, 2, 4, or 8 cycles at aconstant 12 percent MC, cycled between 9 and 15 percent MC.and cycled between 5 and 19 percent MC.

. 12 J' S' '"NUMBER OF MOISnJRE CYCLES

Rgure 4. - Average plate backout values for standard truss-plate joints subjected to 0, 1, 2, 4, or 8 cycles at a constant 12percent MC, cycled between 9 and 15 percent MC, and cycledbetween 5 and 19 percent MC.

altered by the presence of gaps between the plywoodand stud. The mechanism of nail bending is analogousto that of tooth bending. Groom (9) mod1fted Chou'smodel to accommodate boundary conditions morerepresentative of a truss-plate tooth in a wood matJ1x.with results that support this finding.

VaI1abWty of the load -slip traces was most greatlyaffected by addition of the adhesive interface. withjoints contatntng an adhesive interface being slightlyless vaI1able. VaI1abll1ty increased with the severity ofthe moisture cycling and the number of moisturecycles for both standard and adhesive-interface jOints.VartabWty of the truss-plate joints will be covered ingreater deta1lin a USDA Forest SelVice General Tech-nical Report to be published at a later date.

Plate backoutDimensional changes in lumber caused by mois-

ture hysteresis have been shown in previous studiesto cause truss plates to back out of wood(10.19.21.27). Plate backout as described in the pre-vious studies was quantified in this study. Truss-platejoints subjected to a constant 12 percent MC backedout about 0.004 inch. which is the equivalent of about1 percent of the total tooth length (Fig. 4). over a10-month time frame.

In contrast. the first 9 to 15 percent MC cycleresulted in a plate backout of slightly less than 2percent of tooth length. Subsequent mild moisturecycling resulted in some additional plate backout.which appeared to stabllize after the fourth moisturecycle at somewhere between 3 to 4 percent of toothlength. The more severe 5 to 19 percent MC cyclingresulted in drastic plate backout. Truss plates backedout approx:tmately 4 percent of tooth length after thefirst moisture cycle and continued to back out at analmost linear rate. Plate backout at the eighth and finalmoisture cycle was 0.045 inch. or about 17 percent oftooth length.

the teeth can be seen. Thus. even under ideal condi-tions. the txuss-plate joints exhibited slightly dimin-ished load-slip traces after post-assembly.

This reduction in mechanical behavior is magnifiedunder conditions of moisture hysteresis (Figs. 3b and3c). The first moisture cycle seems to have the greatesteffect on truss-plate joint mechanical behavior. butsubsequent moisture cycling continues to have adegradatory effect. This effect appeared to be levelingoffwith the mUd moisture cycling but still had a rathersignificant impact on the severely moisture-cycledspedffiens.

The moisture hysteresis also appears to affect thefundamental manner in which stresses are trans-ferred between the wood and truss plate. The time-de-pendent behavior of the data in Figure 3a shows thatthe load-slip traces change only in magnitude but notshape. as the traces are curvilinear from initial loadingto failure. However. both the magnitude and shape ofthe load-slip traces in Figures 3b and 3c are alteredby subsequent moisture cycles. The load-slip tracesfor moisture-cycled specimens are linear up to about50 percent of the ultimate load. at which time thetraces revert back to curvilinearity. The degree oflinearity appears to be related to the severity of themoisture cycling. with the highest degree of linearityoccurring in the most severely moisture-cycled speci-mens. Load-slip traces are most noticeably differenti-ated at low load levels. indicating a substantial de-crease in initial joint stiffness for moisture-cycledspecimens.

An alteration in the shape of the load-slip traces ismost likely attributable to either a change in theinterfadal contact between the wood and tooth or tobacking out of the truss plate with subsequent mois-ture cycles. This finding seems consistent with theresults of Chou (1). who found that load-slip traces ofnails connecting plywood and studs are significantly

2SFOREST PRODUCTS JOURNAL Vol. 44. No. J

Joint stiffness

FIgure 5 shows the effect of moisture hysteresis onthe initial stiffness of truss-plate joints. measuredbetween 700 and 1,100 pounds. Simple rheologicaleffects can be seen in the specimens subjected to aconstant 12 percent MC. Exposure to constant hy-grothermal conditions resulted in a slow. steady de-cline of initial stiffness. Although the initial stiffnessof standard spedffiens continued to decline even aftereight moisture cycles. the rate of decline was small.Thus. the constant MC spedffiens seemed to stabilizeat approximately 75 percent of the original stiffness.

Mild moisture cycling between 9 and 15 percentMC resulted in a much more abrupt decline in initialstiffness. The first moisture cycle resulted in a de-crease in stiffness of 35 percent and the second cyclestiffness had decreased by 47 percent. Although therewas a rapid decrease in initial stiffness in the mildlymoisture-cycled spedffiens. the rate of decrease instiffness slowed after several moisture cycles. with alower l1m1t appealing to be about 40 percent of theoriginal stiffness. A s1m11ar but more degradative effectwas seen with the more severe 5 to 19 percent MCcycles, in which stiffness decreased 50 percent afterthe first moisture cycle and stab1l1zed at approxi-mately 30 to 35 percent of the ortginal stiffness. Theseresults are consistent with those of Wilkinson (28),who showed that creep of truss-plate joints was sen-sitive to moisture cycling.

Comparison of plate backout (Fig. 4) and initialjoint stiffness (FIg. 5) suggests that initial truss-platejoint stiffness is not fully explainable by plate backout.Large decreases in initial stiffness as a result ofmoisture cycling appear to be caused by plate backoutand. either indiVidually or in concert. by the localizedrelaxation of the wood surrounding the teeth and areduction in the intimacy of contact between the woodand teeth. Wilkinson (30) noted that the elastic-bear-

ing constant decreased 38 percent after one moisturecycle between 10 and 20 percent MC. suggesting thatsome localized relaxation may be occurring.

Ultimate loadThe most important property of truss-plate joint

mechanical behav1or from a design standpoint isstrength. which also seems to be sign1flcantiY affectedby moisture cycling (Fig. 6). The rheological effects onultimate load of specimens that were subjected to aconstant MC showed a net loss in strength of about11 percent over a 9-month time frame. It appears thatail upper limit of strength reduction for a constant MCwould be approximately 20 percent. for a long-termstrength of about 136 lb./tooth pair. However. thisvalue 1s only an approximation. because the jointswere continuing to degenerate.

The mild moisture-cycled specimens demon-strated a s1m11ar but magnifted degenerative effect.Overall strength was reduced by almost 19 percent bythe end of the eigI:lth cycle. The upper limit of sttengthreduction for the mildly moisture-cycled specimensappears to be about 25 percent. or about 127lb./tooth. Specimens cycled between 5 and 19 percentMC demonstrated an accelerated degradative re-sponse. An overall strength reduction of 31 percentafter the eighth cycle was observed. which is compa-rable to the 27 percent reduction seen by McAlister(19) for I-year outdoor exposure of truss-plate joints.The rate of the degradative response appeared to slowslightly with each subsequent cycle. but not enoughthat long-term response could be projected with anydegree of rel1abll1ty. Perhaps a s1m1lar expeIimentaldesign conducted over a longer time frame would helpdefine this upper limit.

It is worth noting that there is a relationshipbetween the amount of plate backout (Fig. 4) andultimate load (Fig. 6). suggesting that strength oftruss-plate joints may be a function of plate backout.

SUMMARY OF MOISTURE EFFECT ON PLATE BACKOUTMC-I~.'-I~._S-I"(.--,-)EFFECT OF MOISTURE DIFFERENTIAL ON ULTIMATE lOAD- Jo'-. ... of Npi- PO' - ..- a 10

7~.. 1-8

I ~t

~ .l

~

!

§~ ,~g

~

..

- _!-~- ~~

a-t '1'" j.' t t

NUMBBOFMOImIaECYQ.2S

Figure 7. - Comparison of plate backout br standard truss-platejoints and trussoplate joints with adhesive subjected to O. 1. 2. 4,or 8 cycles at a constant 12 percent MC. cycled between 9 and15 percent MC. and cycled between 5 and 19 percent MC.

. I J J."""NUMBER OF MO~E CYa.ES

Figure 6. - Ultimate load for standard truss-plate joints sub-jectedtoO, 1, 2, 4, orB cycles at a constant 12 percent MC, cycledbetween 9 and 15 percent MG, and cycled between 5 and 19percent MC.

JANUARY 199426

--

il.. .

~

Figure 8. - Comparison of stiffness between 700 and 1,100pounds for standard truss-Plate joints and truss-pjate joints withadhesive subjected to 0, 1, 2, 4, or 8 cycles at the following MCs:(a) constant 12 percent; (b) between 9 and 15 percent; and(c) between 5 and 19 percent..discernible difference between the constant 12 per-cent MC spedmens and the 9 to 15 percent MCspecimens. prtmal1ly because of the small degree ofbackout.

Addition of an adhesive had the greatest effect onthe stiffness of the truss-plate joints (Fig. 8). Althoughthe stiffness of the truss-plate jOints was InitiallyIncreased by a factor of almost two. the gain In stiffnessdissipated over time (Fig. 8a). The gain In stiffness wasgreatly affected by moisture cycltng. With most of theIncreased stiffness being negated after the first mois-ture cycle. However. the adhesive did seem to add apennanent element that Increased stiffness by someftxed amount. The degree to which the truss-platejoints were stiffened by the adhesive seemed to beInversely related to the severity of the moisture cycles.With the most severely moisture-cycled specimensexhibiting the least benefit.

Moisture cycl1ng of truss-plate joints also has as1gn1ficant effect on ultimate load (Fig. 9). Truss-platejOints With an adhesive tooth/wood Interface degradedWith successive moisture cycles at approximately thesame rate as joints With no adhesive. However. appli-cation of the adhesive Interface did Increase the load-canytng capadty of the Joints. As was the case WithInitial stiffness. the level of increase appears to betnfluenced by the severity of the moisture cycles. Withthe smallest Increase under the most severe cycles.

It should be noted that the adhesive used In thisstudy did provide adequate adhesion. However. visualexamination of failed Joints With an adhesive Interfaceshowed that approximately 5 percent of the surface ofthe teeth was covered with wood. indicating that adifferent adhesive system coupled With surface prepa-

AU spedffiens In this study failed because of toothwlUtdrawal. demonstrating that the withdrawal forcesexceeded the resistance to these forces. Not only dotruss-plate joints with backed-out plates have lesssurface area with which to resist withdrawal forces.but the gap between Ute plate and the wood alsoproduces wlUtdrawal forces because of the eccentI1ctransferal of stresses. SupportJve evidence is gIven byKanamort et aI. (12). who showed that Ute withdrawalresistance of common nails In Japanese larch de-creased by about half after five moisture cycles be-tween approXimately 8 and 15 percent MC.

meet of adhwvebetween tooth and wood

Sliker (25) has shown that an adhesive bond be-tween the shank of a natl and wood Increases thelateral load resistance by approximately 50 percentS1m11ar Ondings have also been reported for glued-bolted joints (23). Lambuth (15) also reported im-proved mechanical performance of truss-plate jointswith the addition of an adhesive to the teeth immedi-ately preceding assembly. Thus. a secondary objectiveof this study was to examine the effects of appiytng anadhesive Interface between the teeth of the truss plateand the wood.

Plate backout for the truss-plate joints with andwithout an adhesive Interface Is summal1Zed In Figure7. The adhesive used In this study retarded platebackout In the specimens cycled between 5 and 19~t MC. reducing backout from 17 percent of totaltooth length to 12.5 percent Thus. It appears thattruss-plate joints subjected to large vanations In MCmay benefit from an adhesive Interface between toothand wood. Addition of the adhesive resulted In no

FOREsr PRODUcrs JOURNAL 27Vol. 44. No.1

6. Duff. J .E. 1978. Moisture content of a joist floor over a crawlspace. Res. Pap. SE-I89. USDA Forest Serv.. SoutheasternForest Expt. Sta. 12 pp.

7. Feldborg. T. 1989. TImber joints in tension and nails in With-drawal under long-tenn loading and altemaung humidity. For-est Prod. J. 39(11/12):8-12.

8. Foschi. RO. 1977. Analysis of wood diaphragms and trusses.Part 11: Truss-plate connections. Canadian J. of CMl Engineer-Ing 4:353-62.

9. Groom. L.H. 1989. Experimental verlftcaUon and nonlinearmodeling of truss-plate Joints by Runge-Kutta numel1cal tech-nique. Ph.D. diss. Oregon State Untv.. Corvallis. Oreg.

10. . 1991. Effect of adhesive appIJed to the tooth-woodInterface on metal-plate oonnecuons loaded In tension. ForestProd. J. 41(4):33-5.

11. . 1992. Nonlinear modeling of truss-plate jOints. J. ofStructural Engineering 118(9):2514-31.

12.. Kanamorl. K.. A. Chino. and Y. Kawarada. 1978. Studies onWithdrawal resistance of nail: Effect of changIng moisture con-tent in wood and ume after nail driving. Hokkaido Forest Prod.Res. Institute 67:103-28.

13. Keenan. F.J.. S.K. Suddarth. and S.A. Nelson. 1983. Woodtrusses and other manufactured structural components. In:Proc. of the Workshop on Structural Wood Research. Am. Soc.of CMl Enginee11ng. New York. pp. 175-84.

14. Kunesh. R.H. and J.W. Johnson. 1968. Strength of multiplebolt joints: Influence of spacing and other variables. Rept. T - 24.Forest Res. Lab.. Oregon State Untv.. Corvallis. Oreg. 19 pp.

15. Lambuth. A.L. 1987. Adhesive/nail plate truss assembly.United States Patent No. 4.659.604. April 21. 1987. 9 pp.

16. Leicester. R.H. and G.F. Reardon. 1979. Performance charac-teristics of some commercial trusses. In: Proc. of the 19th ForestProd.. Res. Conference. Mdboume. Australia. Paper 12:418-20.

17. . . and K.B. Schuster. 1979. Toothed-plateconnector joints subjected to long duration loads. In: Proc. ofthe 19th Forest Prod. Res. Conference. Melboume. Australia.Paper G-l :3.

18. Mayo. A.P. 1980. Long-term performance tests on trussedrafters. Building Research EstabIJshment. Princes RlsboroughLab.. Aylesbury. England. Report CPl/80. 10 pp.

19. McAlIster. RH. 1990. Tensile loading charactel1stics of tnJssplate joints after weathering and accelerated aging. Forest Prod.J.4O(2):9-15.

20. Noren. B. 1968. Nailed joints-Their strength and I1gidltyundershort-term and long-tenIlloading. Report 158B. The NationalSwedIsh Institute for Building Research. Stockhohn. Sweden.BOpp.

21. Qualle. AT. and F.J. Keenan. 1979. Tnlss plate testing InCanada: Test procedures and factors affecting strength proper-ties.In: Proc. Metal Plate Wood Tnlss Conference. St. lhuts. Mo.Forest Prod. Res. Soc.. Madison. WIs. pp. 105-112.

22. Reardon. G.F. 1971. A stnJctural analysis of frames Withsem1-11gid Joints. Dtv. of Forest Products. CS1RO. Australia.Technological Pap. No. 59.

23. Riberholt. H. 1980. Steel bolts glued Into glulams. lUFROMeeting. Oxford. England. Pap. 11:242-51.

24. Sherwood. G.E. 1985. CondensaUon potentJalln high thermalperfonnance waDs - hot. humtd sununer climate. Res. Pap.FPL-455. USDA Forest Serv.. Forest Prod. Lab.. Madison. WIs.

25. Sl1ker. A. 1970. Creep In nailed wood-metal tensionjolnts. WoodSc1. 23(1):23-30.

26. Truss Plate Institute. 1985. DesIgn SpeCiftcation for Metal PlateConnected Wood Tnlsses. 'IPI-85. Truss Plate InsUtute. Madi-son. WIs.

27. Wight. M. 1977. Moisture content effects on strength of tnJssplate connector Joints. B.Sc.F. thesis. Faculty of Forestry. Univ.of Toronto. Ont.. Canada.

28. W1lk1nson. T.L. 1966. Moisture cycling of tnJssed rafter Joints.Res. Pap. FPL-67. USDA Forest Serv.. Forest Prod. Lab.. Madl-SOD. WIs.

29.

30.

. 1971. TheoreUca11ateral resistance of nailed jOints.J. of the Structural DIvision. ASCE 97(STS):1407-22.

. 1974. Elastic bearing constant of wood: Effect ofmoisture content conditions. Res. Pap. FPL-235. USDA ForestSeIV.. Forest Prod. Lab.. Madison. Wis.

. 1984. lcngume perfonnance of trussed rafters withdI1J'erent connection systems. Res. Pap. FPL-444. USDA ForestSeIV.. Forest Prod. Lab., Madison. Wis.

31

29FOREST PRODUCTS JOURNAL Vol 44. No.