i4G.RICULTURE ROOM
CCNTROLLED EXPOSURE TESTS CIS PIRCI I
PLYWOOD INDICATE DURAQILITY O F
WATER RCS I_~
T 1 AMOS'
UNITED STATES DEPARTMENT OF AGRICULTUR E
FOREST SERVIC E
LFOREST PRODUCTS LABORATOR Y
IL--
Madison,
Wisconsi n
In Cooperation with the University of Wisconsi n
November 1938
CONTROLLED EXPOSURE TESTS ON BIRCH PLYWOOD INDICATE DURABILITY OF WATE R
RESISTANT GLUE JOINTS "
By Don Brouse, Enginee r
ti
Introductio n
Previous tests have demonstrated that age, in itself, is not a
cause of failure in well made glue joints (1)2 . and that the commonly use d
woodworking glues are capable of producing joints that are permanentl y
durable so long as the conditions of service do not exceed certain limits .
The limits of the service conditions under which joints made with the mor e
common glues can be expected to remain strong and serviceable vary wit h
the different adhesives and have been defined by tests previously re ported (2) . It has been demonstrated further (3) that when a glue joint ,
originally well made, fails in service the cause may be chemical hydrol ysis of the glue itself, destruction of the glue by micro-organisms, o r
softening and weakening of the glue from absorption of water . These fac tors are usually combined with and exaggerated by the mechanical stresse s
developed on the joints as the wood changes dimensions because of change s
in moisture content . In joints that are not well made, the mechanica l
stresses are sometimes sufficient to cause failure without weakening o f
the glue material itself . The significance of hydrolysis and its rela tion to formulation has been developed for casein glues (4+) and the general principles are believed to apply to other water resistant protein
glues . It has been shown that adding toxic chemicals to the glue whe n
mixed, or treating the wood with them after gluing, greatly reduces th e
severity of mold attack (5) .
In these previous tests, nothing was found, however, that re duced to a significant degree the failures due to softening and weakening
of the glue from absorption of moisture, combined with the mechanical action of the swelling and shrinking stresses in the wood on the glue joint .
For highly water resistant and durable glue joints, a glue must be capabl e
of producing strong joints in the dry condition and be very resistant t o
softening, to hydrolysis, and to mold attack . The common woodworkin g
glues are capable of producing strong joints in the dry condition bu t
soften to a greater or less degree when soaked in water . Some of the mor e
unusual adhesives previously available, such as the cellulose esters ,
soften very little, if at all, in water but lack the properties necessary
to produce strong joints in the dry condition .
l Presented at a meeting of the Wood Industries Division, at High Point ,
N. C ., Sept . 22-23, 1938, of The American Society of Mechanica l
Engineers . (Published under the title "Exposure Tests on Plywood "
in Mechanical Engineering, Vol . 60, Nov . 1938, No . 11, pp . 852-856 .
? Figures in parentheses refer to the literature cited .
R1185
r
The development of the artificial resin glues provided adhesive s
capable of producing high dry strengths together with very high resistanc e
to softening in water and, consequently, a high degree of resistance t o
mechanical stresses . Results of some of the foreign tests (6) as wel l
as early experience at the Forest Products Laboratory with a phenoli c
resin in alcohol solution (7) were very promising as far as propertie s
of joint were concerned . When the resin glue manufacturers had develope d
their products to a point where they could be offered at a price that th e
woodworker was willing to pay, the field was opened for a very narke d
advance in the production of glued wood products possessing a high degree of resistance to all forms of exposure involving moisture and a
nearer approach to the desired goal of a glue joint capable of with standing any treatment that the wood itself could stand .
The evaluation of glued joints requires a long period of exposure to different conditions because methods have not yet been perfected by which ultimate durability may be predicted with certaint y
from short laboratory tests . Shortly after the synthetic glues wer e
offered on the domestic market, therefore, the Forest Products Laborator y
started long-time tests with a view to answering the questions of the actual quality and durability of joints that might be obtained by the us e
of these newer types of woodworking glues .
Test Procedur e
At the time the first of these exposure tests on artificia l
resin glues were started by the Laboratory the technic of handling th e
artificial resin glues was not widely known and was limited largely t o
the manufacturers or the individuals who were developing or promotin g
the respective adhesives . For the purpose of these tests it was decided that the gluing should be done by the manufacturers rather tha n
by the Laboratory . Yellow birch veneer (1/16" in thickness) selecte d
for smoothness, firmness, straightness of grain, and freedom from . defects was sent to the manufacturers with the request that they glu e
approximately 20 panels, each 3-ply, 3/16" x 12" x 12", under gluin g
conditions that they believed most favorable for their particular pro duct . Included in the tests were four artificial resins (P-l ) P-2 ,
P-3 and p--) +) reported to be of the phenol-aldehyde type and on e
artificial resin (V-1) reported to be a vinyl ester .
4W
Four types of exposure tests, as described later, were used .
The number of specimens glued with resin P-2 was insufficient to pro vide for all four exposure tests . The specimens available were subjected to the soaking-drying exposure because this test is the mos t
important of the four in disclosing the unusual properties of th e
resin glues . The omission of specimens glued with resin P-2 fro m
three of the four tests necessitates the use of two sets of "contro l
averages" shown in the sixth column of Table 1 .
R11S5
-2-
u
1
.
9s
T
7
II
'sr
Table 1 . - A summary of average test values a
Aned upon subjecting specimens of birch plywood to different conditions c
exposure .
-7--
:Average of :
Resin
Resin
: Resin ,
Resin
Glue
Resin
: , P-4
:P-1, 2, 3 •
V-1
P-1
P-2
:
P-3
Time o f
:
and 4
exposure
------------------ ------------ ------------ ----------- ------------- --------- - ----------- Test Valued'.
404-27
435-100
590-74
479-86
:(514-79i
473-77
None Dry controls :
:(494-84,5
4
375-65
:(426-6
275-5
460-73
329-92
: 443-67
48 hrs . Wet:(402-74411
Blood
Casei n
447-48
418- 5
379-51
265- 2
368-28
344-26
196- 0
151-0
126-2
Soaked continuously in water
2-1/2 months
5
8
12
18
24
30
36
42
48
N
"
456-62
398-90
421-100
327-69
396-73
399-100
278-67
304-72
:
:
:
291-100
233-98
231-94
:
412-74
387-60
399-22
332-90
284-34
277-70
268-58
228-100
229-73
:
:
:
350-44
375-75
352-14
301-58
331-45
243-46
278-31
242-68
242-86
157-48
401-65
402-83
: 355-48
365-51
: 354-78
: 268-49
286-58
267-75
: 234-95
206-7 2
:
:
:
:
246-6
239-8
214-2
251-3 8
223-52
183-24
341-31
189-20
179-42
:
117-64
:
241-100
211-9 3
:
340-100
:
251-100
154-64
49-0
8-0
0-0
0(25 )
305-82
305-74
307-66
238-62
232-90
Exposed continuously to 97% relative humidity
2-1/2 months
5
12
18
24
30
36
42
48
356-45
450-98
328-99
335-80
:
1
1
:
463-100
477-98
446-100 5
423-1005
448-100 :
:
372-82
434-89
382-98 ,
326-100-
339-73
319-100
340-100 6
N
414-96
464-100
212-100353-100
360-65
354-60
342-44
332-63
294-625
290-33
242-99
275-100
194-902
180-100
377-69
423-86
; 378-81
1
:
381-80
387-87
359-69
311-94
350-96
263-9 6
: 286-100
1
:
1
277-5
263-12
223-0
301-33
262-40
258-22
234-2
245-42237-1 8
223-37
:
81- 6
397-90
295-100
256-100
:
0-0
0-0
0(12 )
215-100
158-100 ,
162-OO
131-100
91-100
Exposed to a repeating cycle : 2 days soaking followed by 12
days drying in 30% relative humidity .
2-1/2 months
5
8
12
18
24
30
36
42
48
N
N
r
477-97
429-74
524-89
495-60
493-76
438-92
476-66
348-99
392-46
345-28
:
:
346-99
397-98
427-72
382-99
402-80
362-46
394-36
279-80
370-78
291-65
:
:
:
:
:
448-24
487-32
412-100
483-47
427-82
446-49
468-54
351-61
512-83
312-68
490-20
368-91
411-46
433-32
400-46
437-78
273-10
205-14
198-3
187-28
421-80
41-63
474-69
428-67
431-71
421-66
403-42
286-65
362-37
298-53
211-1
90- 2
1PDWMM
58-0
28-0
0(19)
:
to a repeating cycle : 2 weeks in 97% relative humidity
followed by 2 weeks In 30% relative humidity .
508-a
553-70
395-47
270-5
: 530-86
370-14
466250-4
204-2
451-79
367-8
428-59
408-66
483-68
162-11
! 465-98
419-73
226-8
:
372-28
404-49
466-71
346-21
52-0
354-22
: 461-53
64-10
489-59
520-95
306-15
: 415-66
77-2
344-,31
: 445-47
120-8
453-41
489-80
268: 387-6j
92-6
328-20
237-0
204-0
93-0
43-0
0-0
42-0
0(25 )
383-56
356-99
356-98
332-36
359-8 2
330-7 3
267-2
251-0
50-2
397-0
407-29
298-0
323-40
233-0
270-0
129- 0
135- 0
105- 0
Exposed
2-1/2 months
N
5
n
8
12
18
24
30
36
42
48
577-77
497-93
465-89
;g: 24
Z401-56
3
541-77
0
420-88
537-70
393-100 5
:
0(12 )
314-30
279-5 9
292-3 4
305-57k
IFiret figure in each pair of values is joint strength in pounds per square inch ; the second figure is wood failur e
in percent . Each value is an average of 5 specimens . Figure in parenthesis represents time in months when las t
2 specimen failed .
-Averages of P-1, 3 and 4
4verages of P-1, 2, 3, and 4 .
-Tested wet after soaking in water at room temperature for 48 hours .
Slight evidence of wood rot .
-Marked evidence of wood rot .
Z M 34079 F
For comparative purposes, twenty similar test panels each wer e
glued at he Forest Products Laboratory with casein and blood'-albumi n
glue .
The casein glue was the casein-lime-sodium silicate combinatio n
described in several publications (8) of the Forest Products Laboratory
and the blood-albumin glue was the paraformaldehyde-ammonium hydroxide blood albumin combination also described in Laboratory publications (8) .
For the gluing done at the Laboratory, the veneer was con ditioned to approximate equilibrium with 65 percent relative humidity ,
giving a moisture content of approximately 12 percent . The gluing wit h
casein glue was done in the conventional way with conditions adjuste d
to fall within limits favorable to the production of good joints (s) .
The panels glued with blood albumin were spread and pressed at room
temperatures, allowed to remain under pressure over night, and the n
hot-pressed the following morning for 10 minutes at a temperature o f
approximately 260° F . and under a pressure of 200 pounds per square inch .
The procedure after gluing was identical whether the panel s
were glued at the Laboratory or received from the manufacturers . Al l
panels were conditioned to approximate equilibrium in a room at 65
percel, relative humidity and then cut into standard plywood tes t
specimens (9) . Each panel yielded 30 test specimens, giving a total o f
600 specimens for each glue . Five specimens from each panel were teste d
dry and 5 were tested wet after soaking for 4s hours in water at'roo m
temperatures . If the test values from any panel were low and erratic ,
that panel was eliminated from further tests . After eliminating defec tive panels, the dry test values of the specimens from the remaining
panels were averaged for each glue and these averages were used as a
basis of comparison throughout the tests . The same procedure wa s
carried out to obtain the wet test values, although these were not use d
as a base for comparison . For each glue, the specimens were then mixe d
together to insure random sampling and divided into four groups of 7 5
specimens each, one group for each of the following tests :
1 . Soaked continuously in water at room temperatures .
s0° F .
2 . .. Exposed continuously to 97 percent relative humidity a t
3 . Exposed to a repeating cycle consisting of 2 days soaking
in water at room temperatures followed by drying for 12 days in 3 0
percent relative humidity at 80° F .
A
straight soybean glue was included in the tests but the result s
are not reported because the glue was of a type developed for and
used extensively on softwoods but not recommended by the manufac turer for use on dense hardwoods . As might be expected, the re sults on birch were lower than those shown for casein glue an d
serve mainly to illustrate the fallacy of using on dense hardwood s
a glue of insufficient mechanical strength .
Rlls5
-3-
4• Exposed to a repeating cycle consisting of 2 weeks in 97
percent relative humidity at 80° F . followed by 2 weeks in 30 percen t
relative humidity at 80° F .
From each group, five specimens were tested at intervals o f
2-1/2, 5, 8, 12, 18 1 24, 30, 36, 42, and 48 months . For the cyclic exposures, nos . 3 and 4, the testing was done at the end of the "dry half t"
of the cycle . Specimens from exposure no . 1 were tested wet as soon a s
possible after removal from soakfng and specimens from exposure no . 2
were tested promptly upon removal from 97 percent relative humidity .
The test values for each 5 specimens were averaged and the averages ar e
shown in table 1 .
The tests are still in progress while this report is bein g
prepared, with sufficient specimens remaining to continue the study
through the 60th month .
In preparing the char t a 1 successive average values wer e
plotted as percentages of the
average value from the dry tests .
This procedure permitted easier .Nparison of the rates of failure sinc e
all lines then start from the same origin of 100 . Percentage of woo d
failure was based on a visual inspection of the broken glue joint an d
an estimation of how much of the failure occurred in the wood as compared to the failure in the glue line itself . The amount of wood failure was expressed in percentage of the total joint area and in th e
charts, therefore, the wood failures were plotted as recorded .
offal
In these tests, the specimens were unprotected, the dimensions were small, and the specimens spaced on rods to permit circulation of water or air during the exposure cycles . The wood, therefore ,
probably attained approximate equilibrium with the exposure condition s
at each period of the exposure cycle and the stresses developed on th e
glue joints approached the maximum that could be expected under th e
conditions prevailing .
Result s
One of the first impressions gained in this study was the
importance of the amount of wood failure developed when testing th e
joints . With the more commonly used woodworking glues the percentag e
of wood failure was low after any appreciable exposure to moisture .
With many of the artificial resin glues, however, the amount of woo d
failure developed, usually exceeded 50 percent and often approache d
100 percent . For this reason it seems important to include wood fail use values in tables and charts if an accurate picture of the qualit y
of joint is to be presented . If the percentage of wood failure developed on test approaches 100, obviously the strength of the woo d
in shear rather than the strength of the bond determines the tes t
value obtained .
Rl185
J4-
Testno .1 . Continuous soaking in water . The wet test value s
(table 1) indicate primarily the degree to which the joints are weakene d
by early softening of the glue, particularly if low strength values ar e
combined with low wood failures . When long continued, however, soakin g
in water serves as an approximate measure of the rate or the degree t o
which the joints weaken by the hydrolysis of the glue itself (3) .
Conforming to the results of previous experiments (3), th e
casein glue hydrolyzed at such a rate that all the joints had faile d
completely at the end of 25 months and at the end of lg months th e
average test value was nearly down to zero (table 1 and figure 1) .
This was a casein glue comparatively low in alkalinity . A casein glue
high in alkalinity would be expected to hydrolyze more rapidly .
As might be expected from a consideration of their chemica l
composition, the artificial resin glues used in these tests did not sho w
a tendency to weaken when soaked continuously in water . If that tendenc y
is present at all, it is masked by the more rapid weakening of the woo d
itself . When the average figures are plotted (figure 1) they show a
gradual decrease in strength but the average percentage of wood failur e
at the end of four years is some 72 percent, an amount approximatel y
equal to the average wood failure developed in the original dry tests .
The vinyl resin is not included in the averages plotted on the chart but ,
like the other artificial resins in this respect, it did not appear t o
weaken any more rapidly than the wood when soaked continuously in water .
The artificial resin glues, however, are not unique in thei r
ability to resist exposure of this type . The paraformaldehyde--bloo d
albumin glue applied by the hot pressing method does not hydrolyze a t
all rapidly . So far as resistance to continuous soaking is concerned ,
specimens glued with hot-pressed, paraformaldehyde blood glue performe d
as well as the specimens glued with artificial resins .
Test no . 2 . Continuous exposure to97percent relativ e
humidity . The conditions of this exposure are favorable to the devel opment of fungi . Under these conditions, molds will attack an unpro tected protein glue rapidly and wood destroying fungi will cause rottin g
of non-durable and unprotected wood specimens .
The casein glue used contained no chemicals of sufficien t
toxicity to retard mold growth and the joints made with casein glu e
failed rapidly, dropping to 19 percent of their original strengt h
after 2-1/2 months (figure 2) . After 5 months, the casein joints wer e
so weak that they broke in test before a measurable load could b e
applied. All specimens glued with casein glue had failed completel y
by the end of 12 months .
The artificial resin glues of the phenolic type appear to b e
resistant to attack by micro-organisms . The strength test values decreased slowly but the percentage of wood failure increased, indicat ing that the wood was failing more rapidly than the glue and no t
R1185
-5-
establishing clearly whether the glue itself had been weakened . At the
end of 1$ months, visual evidence of rot could be detected in the specimens glued with phenolic resins and by 36 to 48 months the wood ha d
rotted to a very marked degree . Apparently then, the phenolic glue s
themselves were not attacked by micro-organisms but the presence of a
phenolic glue line did not afford protection sufficient to prevent ro t
in 3-ply, 3/1b inch birch plywood .
The resistance of the vinyl resin joints was not clearly established by these tests . At the end of 4s months the test values ha d
decreased to some 55 percent of their original value . The specimen s
were clearly rotted but the percentage of wood failure did not increas e
to the extent that might have been expected from the amount of ro t
present .
Against mold action, the resistance of the hot pressed paraformaldehyde blood joints appeared to be more than equal to the resistanc e
of yellow birch to wood destroying fungi . The fact that 100 percen t
wood failure was developed in all tests of blood glue joints in thi s
exposure at the 5th month and thereafter indicated that the decreas e
in test values was due to a weakening of the wood rather than failur e
in the glue itself . The specimens were not examined microscopically fo r
evidence of wood destroying organisms nor were cultures made but visua l
evidence of rot could be detected by the lSth month and rotting was ver y
marked by the end of three years . In tests of this type, additiona l
information on the resistance of the glue line itself might have bee n
obtained if tests had also been carried out with a more durable specie s
of wood, such as the heartwood of southern cypress, redwood, or wester n
red cedar .
Two conclusions from this test should be emphasized :
1. The resistance of the phenolic resins to attack by mircoorganisms appeared entirely satisfactory but not unique for hot-presse d
blood glues (that contained paraformaldehyde) also proved resistan t
to this type of exposure .
2. The presence of a glue line resistant to fungi did no t
prevent rotting of the plywood . The production of plywood resistan t
to this type of exposure requires a glue resistant to fungi and a
species of wood resistant to wood destroyers, or a treatment of the
wood with effective preservatives .
a
Test no . 3 . Exposure to a repeating cycle that consisted o f
soakin : in water for 2 d s followed b dr-in_ for 12 d- s in 0 -ercen t
relative humidity .
This exposure, which approached most nearly t o
exterior exposure conditions, was one that brought out most clearly th e
superiority of the hot-pressed phenolic resin glues over the other glue s
used in these experiments . The casein glue joints had lost something ove r
40 percent of their strength by the first test period of 2-1/2 month s
(Figure 3) and had failed almost completely at the end of 1$ months ,
although the last remaining specimens did not fall apart until the 30t h
111185
-6-
O
month . Compared with other similar tests on casein glues, these joint s
were more than usually durable . Similar exposures in other tests hav e
caused casein joints to fail completely as early as three months (2), (3) .
At the end of four years, joints made with paraformaldehyd e
blood glue still retained something over 20 percent of their original dry
strength but the test values showed in general a steady and consisten t
decrease . The average percentage of wood failure never exceeded 44 percen t
and, after the 18th month, no wood failure could be detected by visua l
inspection of the broken specimens . Decreasing test values and no woo d
failure indicates a weakening of the glue line . The trend at the 48t h
month, therefore, indicated that the blood glue joints were approachin g
ultimate failure . Similar tests, carried out previously on blood albumi n
joints, resulted in total failures at from 25 to 30 months (2 ) 3) .
Onthe other hand, joints made with phenolic resins have retained an average of 60 percent of their original dry strength throug h
four years of alternate soaking and drying . More important, the averag e
percentage of wood failure in the 4 year specimens was some 53 percent .
The average test values appeared to be decreasing but the fact that a
high percentage of wood failure continued to be developed indicate d
that the severe exposure may have been weakening the wood . At the 48t h
month there was no indication that the joints would fail more rapidl y
than the wood itself . These results lend encouragement to the hope tha t
glues may now be available that can withstand exposures as severe as ca n
be resisted by the wood .
In this very severe exposure, the vinyl ester glue lin e
(table 1) appeared to lack the necessary strength when wet and the joint s
weakened at a rate approximating that for casein glue joints .
Test no . 4 . Ex .osure to a reoeatin c cle of 2 weeks e osur e
to 97 percent relative humidity followed by 2 weeks exposure to 3 0
percent relative humidity . From the nature and rate of failuresit appears that the primary cause of failure in this test cycle was attac k
by micro-organisms . The casein joints failed more rapidly than they di d
in the soaking-drying test (table 1) yet it was improbable that th e
mechanical stresses involved were more severe . The exposure serves t o
illustrate the probable performance of joints with the different glue s
exposed in service to dampness and warmth for a period followed by a
period of dryness .
As in the tests involving continuous exposure to high humidity ,
the resistance of paraformaldehyde blood glue was satisfactory . At th e
end of 4 years the average test value was some 68 percent of the original dry test value and the percentage of wood failure was over 50 per cent (figure 4) .
Slight evidence of wood rot could be detected by
visual inspection, indicating that the decrease in test values was due ,
at least in part, to a loss of mechanical strength of the wood .
R1185
-7-
JT4ENOTN WOOD
VALUES FAILURE
RESIN.---•
•
- ■
BLOOD 0---50 - ■
CASE/N tom--• 5e
-
RESIN
----- _-„
_-_-
-s
/BLOOD
•
/2
/5
/B
2/
24
27
30
TIME OF EXPOSURE (MON TNS)
36
33
39
42
48
45
FIG . 1 RATE OF DETERIORATION OF GLUE JOINTS WHEN SOAKED CONTINUOUSLY IN WATE R
■
~,~
.
_
0_--O
B
CA,IE/N
0-
s
6
}
9
/2
/5
/B
/LUR E
■
a
- .49
BLOOD
S
3'
WOOD
VALUES
RESIN
84000
RESI N
CASEIN
0
LEGEN D
J 1RENOTN
11ENg"
s
s
2/
24
27
30
TIME OF EXPOSURE (moNTNs)
'~
36
33
S9
42
48
45
FIG . 2 RATE OF DETERIORATION OF GLUE POINTS WHEN EXPOSED CONTINUOUSLY TO 97 PER CENT RELATIVE HUMIDIT Y
I
~~~_
'\
\~
nn
~
irnnvai
LEGEND
STRENGTH W00 0
vALUEJ FAIL UR
N
fRESI
BLOOD
H
n'
CASE/M ~• -0
-
s
CASE/N
3
/2
15
/8
~
2/
24
27
30
TIME OF EXPOSURE (MONr .S)
33
36
39
42
45
43
FIG . 3 RATE OF DETERIORATION OF GLUE JOINTS WHEN EXPOSED TO A REPEATING CYCLE OF 2 DAYS SOAKING FOLLOWE D
BY 12 DAYS ' DRYING
A
LEGEND
.rrxmaTN WOOD
VALUES FAILURES
RESIN
0
■
8[000 o- -- 0
CASE/N
- -0
0
CASEIN
~'\
s
9
•\
S
20
1
Q
W
0
0
no . 4
y
3
5
9
/2
15
LB
2/
24
27
30
TIME OF EXPOSURE(MONTHS)
33
36
39
42
45
/00 =
Bo
r
48
O
60 W
40
to 4
0
O
-'
0
RATE OF DETERIORATION OF GLUE JOINTS WHEN EXPOSED TO A REPEATING CYCLE OF 2 WEEKS IN 97 PER CENT RELATIVE HU MIDITY FOLLOWED BY 2 WEEKS IN 30 PER CENT RELATIVE HUMIDIT Y
Z M 34768 F
I
As might have been expected from their behavior in other tests ,
phenolic resin joints were not seriously affected by this exposure . A
relatively high percentage of wood failure was developed at each tes t
throughout the four years, Whatever the decrease in average test value s
it appeared to be due to a decrease in the strength of the wood rathe r
than failure in the glue . Visible signs of wood rot could be detected i n
some of the specimens, indicating again that a glue line resistant t o
fungi does not offer sufficient protection to the wood against fungu s
attack .
The joints made with the vinyl ester did not appear to b e
affected by mold and they withstood this exposure much better than the y
did the soaking and drying cycles .
Summar y
Joints made with different artificial resin glues of the phenolic type have satisfactorily withstood four years of exposure to extremely severe test conditions . These glues did not appear to soften o r
hydrolyze on continuous soaking in water and the joints were not affecte d
by molds although the presence of the mold-resistant glue line did no t
protect the wood itself from the action of wood--destroying fungi . Afte r
4 years of soaking and drying the specimens still developed a high percentage of wood failure in test .
The one vinyl ester included appeared to be sufficiently resistant to hydrolysis and to mold attack but lacked the strength required to withstand stresses caused by repeated wetting and drying .
Previous experiments with the older woodworking glues wer e
confirmed in that : (1) The blood-glue formula containing paraformaldehyde was sufficiently resistant to molds and hydrolysis but it lacke d
the strength to withstand indefinitely the mechanical stresses set u p
by repeated wetting and drying ; (2) casein glues were readily subjec t
to failure from hydrolysis, mold action, and mechanical stresses whe n
the glue was softened by absorption of water .
I
81185
-g-
Literature Cite d
4
It
(1) "Age and Strength of Glue Joints" . By Don Brouse ,
Woodworking Industries, June 1931, pp . 32-33 .
(2) "Serviceability of Glue Joints" . By Don Brouse ,
Mechanical Engineering, Vol . 60, pp . 306-8
(Apr . 1938) .
(3)
"Behavior of Casein and Blood Glue Joints Under Differen t
Conditions of Service" by Don Brouse . Furnitur e
Manufacturer, Sept . 1934, pp . 9--11 .
(4)
"Casein and Its Industrial Applications" by Edwin Sutermeister .
Chemical Catalog Company, New York City, 1927, pp . 169-217 .
(5) "Increasing the Durability of Plywood" by Don Brouse .
Mechanical Engineering, Vol . 53, pp . 664-66 (Sept . 1931) 0
(6)
"Holzvergitung Burch Kunstharzverleimung" by P . Brenner and
0 . Kraemer . Publication No . 12 . Fachausschuss fUr
Holzfragen, Berlin, 1935 . 40 pp.
(7)
"Spread of Condensite" by W . L . Jones, F . P . L . unpublished .
report, May 24, 1922 .
(S )
"The Gluing of Wood" by T . R . Truax, U . S . Dept . Agric .
Bulletin 1500 .
(9 )
"Gluing Wood in Aircraft Manufacture" by T . R . Truax, U . S .
Dept . Agric . Tech . Bulletin 205 .
d
a
R1185
-9-