U. S. Department of Commerce Research Paper RP 2087 National Bureau of Standards Volume 44, April 1950
Part of the Journal of Research of the National Bureau of Standards
Thermal and Moisture Expansion Studies of Some Domestic Granites By Arthur Hockman and Daniel W. Kessler
As a part of a study of the physical proper ties of b uilding stone, Lh ermal expansion determinations we re made on 48 sa mples of domestic granites by t he d iffcrcntial inter fcrometcr method (w ith an inlerfcrogJ;aph r ccording attachmcnt) over t he tcmperaturc range - 20° to + 60° C. Thermal coe fficients computed between - 20° and + GO° C I'anged from 4.810 8.3 X lO- 6 per d eg C w ith an avcrage of 6.2 X 10- 6 pCI' deg C. Coe ffi cic nts ob La incd on cool ing (GO ° to 0° C) averaged G.7 X 10- 6 p el' d eg C. Expansion cu r vcs dra\\"Il [01' all samplcs indicatc sligh t irrcguJ arilies in lhc 0- to 20-dcg rangc in the healing CUI'I'cs of at least G5 pOl'ce nt of thc samples. These irrcgulari t ies are p r obably callscd by m oi;; lurc changes in lhc sample during Lhc tesL i\[oislurc cxpansion dcLc rminat ions \\'crc madc on t hc 48 sa mples bv mcans of l hc T uckerman oplical strain gagc ovcr a 2"I-hollr pcriod at co nstanL tcmper aturc. Thc cxpan sions obtained r ange d from 0.000-1 to 0.009 percenL and avcraged 0.0039 per ce llt. Advcrsc cff cls r cs ul t ing from continued thcrmal and moi -ture expansion and contrac
tion of !!:ran ite arc bel ic \'cd (0 be of SO IllC significance ill monulllcntal ~ trll cl l1l' es lhat arc
cxpcclcd to la~ L for long perio::!s o f limc.
I. Introduction l'es uHing from the II ual diurnal Lernpe l'ature cllaoges call e internal stresses wh ich, after Granite is known (0 bc one of th e most durablp numerous repetitions, may have a weakening effeeL of structural matcri als. For thi s rca on it has on (h e stone. II also seem likely that the granite been an ideal cllOi c;e for monuments, public build may be a rrectecl by Lhe unequal expansion of the ings, shrines, and structures that arc cxpected to eliA'e rent mineral co nstituents, and t he fact that last for cenLu ries. Therc arc grani tc structlll'cs the pri neipa l constituen ts, namely feldspar and in this country that have existed for periods up quartz, expand unequally along differenL crystal to 200 yr without sho\\-ing serious signs of dis lographic axe. integration, yet there is evidcnce to show t hat Expansion cfl' ects clu e to absorption of water by some granite structlll'{,s have begun to deteriorate granite have not been generally understood or in less than 50 yr. appreciated. The experimental data acq uired The durability of granite may be adversely in this study definitely prove th aL granite llllder affected by one or more of many physical and goes a change in length on becoming wet. As in chemical agencies. The effect of natural temper the case of thermal expansion, the ]'epeated ex ature changes in contributing to the deterioration pansion and contraction res ul ting from wetting of the stone has been recognized by geologists and drying may have a weakening eerect on the and engineers.' It is well known that the rapid stori e. M oist ure grad ien ts resulting from partial heating of granite, as in the case of a fire , will \\-et ti ng of the tone cause internal stresses, which usually cause serious spaUing. 1\ot so apparent in th e co urse of time may adversely affect its is the fact that gradients in a graniLe s Lructure durability.
1 T he supposition that rocks in ge neral arc seriously affected by natural The pLll'pose of this study was to accumulate changes ill temperature is strongly chall enged by some geologists. For data on the thermal expansion characteristics of "xample sec E. 13Jackwelder, Tbe insolation hypothesis or rock weathering Am. J. Sci. 2 6, 970 ( 1933) . grallite over the temperature range produced by
Expansion of Domestic Granites 395 the usual weather exposures and also on the TABLE 1. Summary of thermal ex pansion data on granite expansion due to the absorption of water. Most obt ained by previous investigators of the published data on thermal expansion 'of Mean co granite were computed from measuremen ts taken effi cient of No. of Temperature linear ther- Date of at larger temperature intervals than those to specimens ran ge a m~ l expan- Inves tigator publi c tested SIOn per ation which masonry is normally subjected , and these degree CXIO ' values can be misleading when applied to the seasonal range. In this study coefficients of °C linear thermal expansion were determined over L ______- 14 to +39 8. 25 Bartlett. ______. 1832 L ______+10 to 98 8.22 Adie ______1836 the range of - 20° to + 60° C (_4° to 140° F) , 4 ______8.85 Heade ______1886 ll ______o to 100 b 7.0 \Vatcrtown ArsenaL ______1895 Since there are very few data published on the L ______o to 100 10. 2 Bald win-'V iseman & 1909 moisture expansion of granite, this property was Gr iffi th . determined on the entire series of samples tested L ______+ 19 to 172 8.6 Wheeler. ______1910 20 ______0 to 100 8. 1 Griffith ______1936 for thermal expansion. This report cites a few 5______3 to 60 6. i Willis & DeReus ______1939 examples of granite structures that show some 4______-20 to 60 6. 0 Johnson & ParsoDs ______1944 deterioration result ing from temperature and n The ran ge for which t he coeffi cients were calculated. moisture eff ects. b Specimens tested in wet condition.
II. Previous Investigations stones, marbles, and one sample of granite (P eter A review of the literature on thermal expansion head, Scotland). The temperature range was m easurements of gra~it e reveals that in 1832, + 20° to 300° C. In 1910, Wheeler [6] studied Bartlett [1] 2 seeking the cause of fissures in coping th e eff ects of six repeated heatings from + 1 go to joints in a granite structure, m easured the ther 1,000° C of samples of Rhode Island granite, mal expansion of a sample of Massachusetts Canadian diabase, and of Italian marble from granite, and samples of m arble and limestone, + 19 ° to 500° C. His specimens were 94 in, long, and the tempera The more recent investigations of 'thermal ture range was - 14° to +39° C . (See table 1 expansion include those made in 1936 by Griffith for results of previous investigations). Also [7]. Among the 106 samples of American rocks about this time, Adie [2] m easured the expansion tested by him, about 20 represent granites gen of two granites (Aberdeen gray and P eterhead erally used for building and monumental purposes. red from Scotland), also samples of marble, sand His range was room temperature to 260° C. stone, and greenstone, 11 is temperature range In 1939, Willis and D eR eus [8] included the was + 10° to 98° C. Adie recognized that a m easurem en ts of five samples of granite in their sample of greenstone expanded permanently study of concrete aggregate, The temperature after several successive heatings. R eade [3] in range was + 3° to 60° C. In 1944, Johnson and 1886, included four samples of grani te in his ther Parsons [9], a pplying a test procedure similar to mal expansi on studies of several types of rocks that used in the present investigation, included but did not state the temperature range, In the m easurem ents of four granite samples in their 1895 at the Watertown Arsenal [4] thermal ex study of thermal expansion of concrete aggregates. pansion tests were made on eleven samples of Comparatively few investigators have studied granite acquired from eight states. Measured the moisture expansion of building ston e. It bars, 24 in. by 6 in. by 4 in. were placed in cold seems that in 1886 Schumann [10] (original refer water (32° F ), then in hot water (212° F ), and ence unobtainable) was the first to make a syste back in the cold bath. These expansion results m atic study of moisture expansion of natural eviden tly represen ted the combined eff ects of stone. Hirschwald [11] repeated th ese experi temperature and moisture. m en ts. Using a micrometer microscope, th e latter In 1909 Baldwin-Wiseman and Griffith [5] determined length changes of several stone speci determined permanent length changes, due to m ens after 2 weeks immersion in water. The heating, for several samples of sandstones, lime- average expansion obtained for one sample of granite and four samples of basalt was 0.029 2 Figures in brackets indicate thc literature references at the end of this paper. per'cen t.
396 Journal of Research --,
More r ecent studies of moisture expansion of in the rock, the more abundant usually bein g a building stone include those made by Matsumoto potash feldspar (orthoclase or microcline or both). (12), Stradling (1 3), Royan (14), and K essler [1 5]. The other is soda-lime feldspar (plagioclase) , and However, the reports of these investig ators did may include one or more of the following spec ies: not include any data on granite. albite, oligoclase, and andesine. In most granites, An examination of all the publish ed data on quartz ranks second in abundance to the feldspars, thermal and moisture expansion of granites de a ferro-magnesian mineral (biotite, hornblende, or scribed in the preceding paragraphs reveals; (1) pyroxene) ranking third. The term "commercial there is a scarcity of data on th e thermal expansion granite" may include such closely r elated rocks characteristics of domestic granites over the sea as gneiss, syenite, monzonite, etc. "Black gran sonal range of temperature; (2) inves tigators took ites", also a commercial d esigna tion, include rock readings at wide temperature intervals and few species such as diabase, norite, and gabbro, which made observations on cooling; (3) there arc prac are somewhat distantly rela ted to the true gran tically no data available on dimensional changes ites. This study includes several of th ese "re in domestic granites due to absorption of moisture. lated" granites . Of the 48 granites tested, potash feldspar was th e most abundan t mineral in 35 Description of Samples III. samples, soda-lime feldspar in 10 , and quartz in 3. The 48 samples used in this study were collected :Mica was the most abundant of th e accessory for a previous investigation at this Bureau of other minerals in 39 samples, hornblende in 4, and pyrox physical properties of th e domes tic granites [16] . ene in 2. The samples represent the deposits of th e principal The descriptive portion of table 2 gives the producing districts of 17 states. Most of these sample numbers, source, classification, texture, granites have b een used for architectural, monu and major mineral constituents of all samples m ental, 0 1' structural purposes. tested. The " classification" of a granite is derived Geologically, granite is designated as a crystal from the most abundant accessory mineral, sllch line rock of igneous origin, and is composed of as "biotite granite," or "hornblende granite", feldspar, quartz, and one or more accessory min etc. T extures r efers to th e average size of the erals such as mica, hornblende, pyroxene, ctc. feldspar grains and follows th e scale given at th e Generally, there are two kinds of feldspar present end of th e table.
T A BLE 2. Description, source, classifi cation, textw ·e, composition, thermal expansion coeiJicients, and moisture ex pansion percentag es of all samples tested
( I) (2) (3) (4) (5) (6) (7)
'l~ h or m a 1 ex pans io n- average Moisture coc rTicient of li near therm al cxpansion- expansion per degree C X IO 6 length ______~--_- ~~ '; I~;;,~~e Sam M ajor M ineral Constituents H eati ng Cooling sion in ple Source Class ifi cation Texture b (in descend in g order o[ water [or No.- abundance) ------24 hI' at constant tempera - 20° to 60° 0° to 60° 60° to 0° ture (21. 5° ± .5° C) --1------·1------Percent 1 Vinalhaven , Me ______Biotite granite ______Fine to medium ___ Ort hoclase , q uartz, oli goclase 5. 8 6. I 6. 4 0.0018 and biotite. 2 ____ _do ______Olivine llori te______F ine ______Soda-li me [eldspars (labrador- 5. 9 6. 1 6. 1 . 0008 ite to bytownite), hypers thcnc,oIi vine and magneti te. 5 _____ do ______Bio ti te b o rnbl e nd e _____ do ______Orthoclase and microcline, 6.5 6. 7 7. 0 . 0036 granite. smoky quartz , oligoclase, biotite, hornblende. • 8 M t. D esert, M e ______Biotite granite ______do ______Orthoclase , smoky quartz, 4. 8 4.9 5. 2 . 0028 oli goclase and b(ot ite. 9 Jonesboro, M e ______do ______M edium to coarse _____ .do ______6. 8 7,2 7. 4 . 0020 See footnotes at end of table.
Expansion of Domestic Granites 391 TABLE 2. Description, sow'ce, classification, texture, composition, thermal expansion coe.tJicienis, and moist1lre ex pansion percentages of all samples tested- Contin ued
(I ) (2) (:3) (4) (0) (6) (i ) l ' h e r m a ] expansion- average NloisLtlTl" cOf'fficient of linear thrrmal ex pansion- ex pansion per degree C X 10° length ______changr due toimmC'r- Sam· ::-1 ajol' Mineral Constituents H eating Cooling f: ion in pIe Source Classification Texture b (in descending order of No. a. water lor abundance) 24 hI' at constant tC lnpera ture (21.5" ± .5° C) ------1----_.------Percent ]0 Long Con', ?\lc ______Biotite--muscoyite granite._ Fine to ll1 cdiuIlL __ l\1icrocline and orthoclase, 6.6 6.5 7. 1 0.0068 quartz, OligoclasE' , biotite and muscovite. II XOl'th Jay, Me ______Biotite-lTIusoovitegranite_. Fine ______do ______6. 2 0.3 6.5 . 0028 ]2 Stonington, ::\1c ___ . ___ Biotite granite ______Coarse ______Orthoclase and micrcclinc, 0.9 6.0 6.4 .0020 smoky Quartz, oligoclase, and biotite. ]3 Frankfort, Nle ______do ______. __ j\lfedium ______do ______6.6 7. 1 7.2 . 0048 16 Highpine, M e ______do ______Coarse ______lVficrocline and orthoclase, D. 8 D. 9 6.3 .0080 smoky quartz, oligoclase, and biotite. 1i Concord, N. H ______~lu sco"ite -biotitegranite -- Fineto medium __ _ Microcline and orthoclase, 7. 7 i. i 8. 4 .00<8 quartz, oligoclase-a lbi te, m usco\'ite, and biotite. 18 R edstone, C'I. H ______Biotite-granite ______Coarse ______Orthocla;c, smoky Quartz (oli - 5. 4 0.0 6. J .0022 goclase-albi te), and biotite. 21 Th1ilfo rd,)J. H ______Quartz monzonite ______Fine ______JVlicroclineand orthoclase, oli- 5.7 6.0 63 .0030 goclase, smoky quartz and biotite. 24 Fitzwilliams,:-.J. B _____ Biotit e -muscov i te _____ do. Orthoclase and microclin e, 5. 4 5 . .) 6.2 .0084 granite. quartz, Oligoclase, biotite and musco\'ite. 25 Barre, Vt ______Biotite granite ______do. ______Orthoclase, smoky quartz, 6.3 6. I 6.8 .00.;0 oli goclase-albite, biotite and m uscovite. 27 Derby, Vi. ______Quartz mon zonite ______Fine to mediulll ___ Smoky quartz, oligoclase, mi- i . l 7.2 i . .0026 crocline and orthoclase, bio tite and muscodte. 28 \\-oodbury, Vi. ______Biotitc granite ______Coarse ______Th1icrocline and orthoclase, 0. 1 6.2 6.6 .0012 quartz, oligoclase , biotite and muscodtc. 30 Windsor, VL ______Hornblende biotite Fine to medium -- Orthoclase, smoky quartz , 0.3 0.6 0.6 .0038 granite. plagioclase (oligoclase-albite) horn bien de and biotite. 33 Quincy, ]\{ass ______Ricbeckite-aegirite granite_ Coarse ______Orthoclase, s]noky quartz, 0.3 0.3 5.9 .0008 riebeckitc, aegirite and albite. 37 W. Chel msford , Mass __ Muscovite-biotit e Medium ______l\1icrocline and orthoclase, 5.5 D.6 6.2 .0032 granite-gneiSS. oligoclase, quartz, rutile, muscovite and biotitc. 39 Chester, Mass ______Biotite-mu s cov i te Medium to fin e ___ M icroclinc, quartz, plag ioclase, 6. 0 6.6 7. 2 .0066 granitc-gneiss. biotite and muscovite. 40 Milford , Mass ______Biotite granite ______Coarse ______1\1icrocline and orthoclase, 0. 9 6.4 6.6 .0020 quartz, plagioclase (albite oligoclase) , and biotite. 41 Rockport, lIf ass ______Hornblendegranite ______do ______Orthoclase and microcline, 5. 4 5.7 5.9 .0038 smoky quartz, h ornblen de, and plagioclase (albite oli goclase) . 43 Ansonia, Conn ______IVfu scovite-biotite Fine ______Microcline and orthoclase, i . 5 7.5 8.0 .0058 granite-gneiss. quartz oligoclase, muscoyitc and biotite. 45 East Lyme, Conn ______Quartz monzonite ______do ___ ~ ______Quartz, oli goclase, microcline i .O ,. 7.5 .00.;0 and orthoclase, biotite and m uscovite. See footnotes at end of table.
398 Journal of Research T A BLE 2. D escription, source, classification, te.rture, composition, ther mal expansion coe. f!i cients, and moisture ex pansion percentages oJ all samples tested-Continued
( I) (2) (3) (4 ) (5) (6) (i)
'I' h (' r m a l expa nsion- avera gp 1\![oisture coc fli ciel1t of ii n('ar thermal C'x nansio l1 - ex pansion PC I' d eg ree CX lO° ' Ie n gth cha.nge d uC' to im mer- Sam l\i( ajor lWi llC'l'aJ C OIl Sl iLuenr s H eating Cooling sion in o le 8 0I1 1'c(' Cla ssificat io n Texture b (in desce nding order of w ater for XO. n abun dance) 24 h r at constant tem pera ture (2] .5° ± .5° C) ------_._------Percen t 46 \Yaterford, Conn ______Quartz mon zonite ______F ine ______Oligocln se, microclinc and 01"· 5.3 5. 4 6. 0 0. 0068 thoclase, quartz, biotite and muscovite. 47 Dra nford , Conn . ______B io tite gra nit e gne iss. __ l\1ed iu m to coarse. ~1icro clin e and or thoclase, 5. 3 5. 3 6. 2 .0080 Quartz oli goclase, biotite and muscovite . 53 !\U . Airy, N. C ______Biotit e gra nitc ______l\l ecliu TIl ______Oligoclase, orlhoc}a:3c and 6.0 6.2 6. 4 . 0020 microcline, quartz and bio tite. 56 Salisbury, X. C ______do ______do ______Orl hoclase with a sma ll 7. 7.3 i. 7 .0020 amount of microcline, pla gioclase, quartz a nd biotite. 59 R ioll , S. C ______do ______do ______Orthoclase and m ic roclillc, G.5 6.6 Ii. 9 .0066 pl ag i oc la s e (o li gocla~(' ) , quartz and biotite . 62 :-
'Sample numbers correspond to the se ri al numbers listed in lable 1 of RP]320 J ournal of R esearch N B S ~5 , 161 (1940) R P 1320. b rr ex ture refers to the a\"erage s ize of the feldspar grains, accord ing to thc following des ign ation : fi ne, less than 0.5 em; medium, 0.5 to ] em; coarse, slightly larger t han 1 Clll.
Expansion of Domestic Granites 399 IV. Test Method for Thermal Expansion on its upper surface. Another interferometer Measurements plate, a, is supported approximately 0.5 · mm above plate b by a perforated fused quartz tube 1. Apparatu s 42.65 mm high , the latter also resting on plate e (a) Differentia l Interferometer and surrounding th e specimen, d. Plate a is When the usual interferometer method is applied polished on both faces. Plates band e are polished to measure the expansion of fine grained mate on the upper faces and fine ground on the lower, rials, a relatively small specimen, 2 to 6 mm in with the exception of a small area on plate e, h eigh t is generally used [17]. However, in order which is polished on both faces in order to produce to make the interferometer applicable to the the temperature fringes. The expansion of the thermal expansion measurement of medium and specimen on heating causes plate b to move coarse grained grani tes it was found necessary to toward plate a. This results in a movement of use a test specimen considerably larger than the the interference fringes in the fi eld. (See section conventional small one. IV- 6 for the details of calculations.) For this purpose, a differential interferometer, (b) Thermal Chamber similar to that applied by Saunders [18], was used to make the expansion measurements. This The thermal chamber (fig. 2) was the same as method is a modification of Fizeau's differential that used by Johnson and Parsons [9]. A brass interferometer [19] in which th e three adjustable cylinder, 9 in. long and 3 in. in diameter was screws used by Fizeau to support the top plate inserted through the top of an ordinary electric were replaced by a perforated fused quartz tube refrigerator into a tank of kerosine, which sur of a specific height. rounded the freezing coils. The heating element In these tests the specimen is 38.42 mm high consisted of a hard rubber tube wound with 30 (avg. size), T -sh aped in cross section, and with a ft of No. 24 Nichrome wire. A removable copper maximum width and depth of 12.5 mm. The cup, supported from the refrigerator top by steel sides are perforated to reduce its heat capacity rods, held all the components of the interferom to a minimum. Figure 1 is a schematic diagram eter. Just beneath the bottom thermometer of the interferometer arrangemen t used in all plate were inserted the three variable junctions tests. The specimen rests on a glass thermometer of a copper-constantan thermocouple. Three plate, e, and supports an interferometer plate, b, slide wire resistors, an ammeter, and a toggle switch were placed in series with th e heating r------~~------Q element and the 110-v power source for control of the heating current. b ------~~~- (c) Recording Mechanism A recording apparatus (fig . 2) previously constructed and used by Johnson and Parsons [9], +------c was a simplified form of the one originally de signed by Saunders [20]. This apparatus replaced d ------' the eyepiece of the ordinary Pulfrich type inter ferometer viewing instrument and thus eliminated the tedious job of counting fringes visually. A strip of 35-mm high speed photographic film was drawn from an ordinary camera cartridge over e rollers by a sprocket geared to a synchronous motor, at the rate of 1}~ in. per hr. A lens fo cused the interference pattern through a yellow filter upon a screen with a slit }~ mm wide, which al
FIGURE 1. Diagram of differential interferometer. lowed only a very narrow portion of the image to reach the slowly moving film. A side tube with a a, Top interferometer plate; b, lower inteferometer plate; c, fu sed Quartz tube supporting tO I) plate; d, test specimen; e, t hermometer plate. slotted mirror placed at a 45° angle, permitted
400 Journal of Research 21
.3
27
I 4 10 Plan VIew Z9 SecfionA-A or SId e - lube 31
34
FIG UHE 2. Diagram oj viewing instrument, Tecording camem, and thermal chamber.
1, M agazine: 2, film cartridgo; 3, AIm rollers; 4, feed sprocket ; 5, take-UI) s pool ; G, pulley to take-up spool; 7, drive shaft ~ea recl to synchronous molor; 8, rewind knob; 9, slit (aperture); 10, light-ti ght cover; 11 , camera eyepiece and lenses; 12, side-t.ube eyepiece; 13, yell ow filter; 14 , objective lens; 1.5, focus adjustment; 16, mirror support; 17, mirror; 18, side- tube eyepiece lenses; 19, s pectrum tube; 20, 90 0 prism; 21, objective; 22, heating co il terminals; 23, double glass; 24 , thermocouple leads; 25, de::;s icant tubes; 26, steel supports; 27, kerosine; 28, brass cyli odor; 20, copper cup; 30, specimen; J J, rr rri ge rator cooling co il ; 32, glass ring; 33, hard rubber tube; 34 , nichrome wire; 35, interferometer plates. visual observations for checking purposes during equilibrium temperatures, the latter being meas the operation of the recorder (see reference [20] ured by a three-junction thermocouple located for details of the recording instrument). just beneath the plate. The fringe numbers obtained in this manner were plotted against the 2. Temperature Determinations temperatures, and a fringe-temperature curve Temperatures were measured with an optically drawn. From this curve it was possible to ground thermometer plate (fig. 1, e) often referred determine the temperature corresponding to any to as an interference or refraction thermometer specifically numbered fringe. During each test [21]. This plate gave a satisfactory set of straight the temperature fringes were recorded simul interference bands that shifted at a definite rate taneously with the expansion fringes. whenever it was heated or cooled. In order to 3. Preparation of the Specimens calibrate this thermometer plate it was placed, together with the other components of the inter Figure 3, a, b, e and d illustrates the stages in ferometer, into the thermal chamber, and a ther the fabrication of a typical test specimen. A mometer fringe count determined for a number of rectangular prism, approximately 12.5 mm by
Expansion of Domestic Granites 401 12.5 mm in cross section and approximately 3S.5 the in terferometer was placed, very carefully, into mm in length was first cut from the granite sample the thermal chamber. The chamber was closed, by means of a 6-in. circular diamond saw. Sec and after a preliminary fast heating to 6S o C to tions were then cut away with the same saw Jeay stabilize the air films between th e points of contact ing a T-shaped piece having sides approximately of the specimen and of the quartz tube with the 2.5 mm thick (fig. 3, b and c). Next, with the aid interferometer plates [22], the cooling system was of small finc silicon carbide grinding wheels, the started and allowed to operate overnight. Next specimen illustrated in figure 3, d was co mpleted. morning after th e final aclj ustmen ts were made on The ends of the finished spec imen were ground so the viewing instrument and recording apparatus, that there were three polished points at thr the film was inserted, and the initial temperature bottom to rontact the glass thermometer plate, determined by the thermoeouple. The reeorder and three points at the top to support the lower was then started, cooling system turned off , the interferometer plate. Similarly, thc quartz tube heating eurrent turned on and increased at 15- used to support the upper interferometer plate min in tervals to produee a uniform rise in temper (fig. 1) had three polished points on its top and ature of approximately 0.4 0 C per minute. \iVhen bottom surface. the final temperatu re was reached (approximately With a fine emery stone and a micrometer gage 65 0 C ), the cooling system was again turned on, the heigh ts of the "legs" of the specimen were and the heating cu rren t reduced at regular inter eq ualized to within 0.005 mm. The legs of the vals until the temperatu're fell just below 00 C. quartz supporting tube were adjusted to a height The recorder was then stopped, the chamber re of 42.65 mm. By setting up all the various parts heated to room temperature to prevent eondensa Of the interferometer and examtning through a tion inside, and the spec imen removed. Finally viewing instrument, the heights of the legs of th e the film was rewound in to the cartridge and specimens were adjusted further to give the removed for development. desired interference pattern. The finished speci mens, which averaged 3.2 g in \,'eight, were dried 5. Description of the Interferogram at SO D C for 4 days and placed in a desiccator The interference pattern Khen seen through the until ready for test. viewing instrument appears as shown in figure 4 . Test Procedure 4, A. When the eyepiece of the instrument is replaced with the recording apparatus, only a After the desired fringe pattern was obtained, narrow portion of the pattern is focused on the th e copper cup containing all the components of moving film (fig. 4, B ). Figure 4, C and Dare schematic drawings of portions of a typical inter ferogram resulting from the fringe pattern shown in figure 4, A and B .3 The fringes become a series of bands that are parallel to the direction of the movemen t of the film only so long as thermal equilibrium is maintained, such as duri.ng the first few minutes of a test. The upper portion of figure 4, C illustrates this condition. When the temperature begins to rise and the specimen ex
FRONT SIDE pands, there is a corresponding movement of temperature and expansion hinges. To prepare the interferogram for the counting of fringes, D u T penciled lines arc drawn parallel to a sli t lin e,t a b c d through each intersection of an even numbered FIG URE 3, Diagrams showing sllccessive steps i n the pre pa thermometer fringe with the thermometer reference ration of a test speci men. line. Where each of these penciled lines crosses 3, R ectangular prism cut from largrr sample; b, sections cut away to fo rm a T·shaped specimen; c, view of b after grinding it to form " legs"; d , \-ie ws of 3 Expansion data were read from a 3X enlargement of t he intcrferogram. fini shed specimen showing points on legs and sections ground away to reduce 4 A slit line is produced on a film by turning off the source of monochromatic its mass. light for about 30 sec during a test, resulting in a narrow underexposed strip.
402 Journal of Research ------
e When the maximum temperature Wtl S reacil ed ------, and cooling began , th e direction of th c fringe / ' / j f '. movement for both expansion a nd temperature d i j iUf" ' \. ..' was r eversed (fi gure 4, D, 7/). As til e frin ges re \ d /
'..'" I crossed tbeir respective refe rence marks t li ey were ...... _----- assigned tll eir prev ious numbers . A B 6. Calculation of Results
Since th e total number of expansion Crin ges passing during a test was affectcd by th e expansion 0 0 0 0 of the quartz tube and th e interferometer plate, b, -0' · k 0 0 0 as well as that of the grani te sp ecimen (fig. 1), it 0 0 0 0 \I-as n ecessary to develop a formula that would 0 0 0 0 include all these factors. The fo llowing formula 0 0 0 . _D.. _. n was used for th is purpose: h 0 _Q- _ I 0 0 0 0 0 0 0 0 0 0 0 0 0 0 where L = lellgth of specimell aL in itial temperature C D T1 in millimeLers; FIG UR I, 4. IntelJeromeler and inlelJel'ogram. t.L= lin ear expansion of spec imen in mm/m lu from T1 to a ny lligh er temperature, A, Interferometer as 5('(' 11 through the '"ic wi ng ins trlln1t~n t ; H, C'XPOS lI1'(' slit magnified to show frin ges; (' , sect ion of i nt(,l' r(' r o~ram at start of test; ]), T2 ; middle section of illtcl'fcrogl' 'rem perature Temperature Expansion fringe No. fringe No. 7. Reliability of Results ------° C In order to ch ec k th e pe rformance of tbe 0 - 22.2 0 2 - 10.6 6. 9 apparatus, samples of three d ifferen t materials, 4 + 0.6 11. 5 n amely, porcelain, glass, a nd a metal alloy were 6 11. 6 15.5 testeel for th ermal expansion . These values were 8 22. 1 21 . 0 compared with those obtained on identical ma- 10 32. 4 27.0 12 42. 2 33. 8 5 Data taken from the table of air corrections listed in reference [17] . 14 5 1. 8 41. 0 , EX l}ansion coe ffi cient of the fused quartz was determined by the usual 16 6l. 0 49. 3 interferometer method on a small sample cut from the original piece of quartz tubillg. The ex pansion coe ffi cient of the interferometer plate b (Vycor hrand glass) was taken from reference [18]. Expansion of Domestic Granites 403 E u " :::l. ..J- E ~I C\J -20 o 20 40 60 -20 0 20 40 60 TEMPERATURE °c FIGURE 5. Thermal ex pansion curves of 48 samples of granite. 0 , H eating points; e, cooling points; fi gures refer to sample numbers in table 2. terials by the Thermal Expansion Section of this V. Thermal Expansion Results and Bureau. The results were as follows: Discussion Coeffi cien t of linear thermal ex pan· sian per deg C from - 20 0 to + 600 1. General C Sa mple Table 2, column 6 gives the m ean coefficients of Hidnert 7 Authors linear thermal expansion for three different ------1------temperature ranges of all samples tested, and Be (Porcelain) ______4.5 X lO- 6 4.4 X 10- 6 figure 5 shows their characteristic h eating and BG (Glass) ______8.7 X 10- 6 8.6 X IO- 6 cooling curves. The coefficients obtained for the BF (Alloy) ______2.5 X lO- 6 2.5X 10- 6 48 samples ranged from 4.8 to 8.3 X 10- 6 per deg C 0 0 7 Chief, Thcrmal Expansion Section , National Bureau of Standards. between -20 and + 60 C with an average of 6.2 X 10- 6• The coefficients of 84 percent of the The results agree within the limits of experi samples in this same temperature range were m en tal error . between 5 and 7 X 10- 6 per deg C. For the 404 Journal of Research temperature range 0° to 60 ° C, the average of th e moisture is absorbed. Absorption of water during coefficients was 6.3 X 10- 6 and on cooling (60° to the early heating part of the cycle migh t account 0° C), it was 6.7 X 10- 6 per deg C. for the increased slope of the expansion curve in that range, while loss of moisture at higher tem 2. Irregularity of Expansion peratures is consistent with the reduced rate of A general view of all the expansion curves as expansion occurring at about 10° C. illustrated in figure 5 reveals more or less smooth The abrupt change in slope observed in some of regular curves on both h eating and cooling. How the curves at 0° C or slightly above might indi ever, on closer scrutiny it can be readily seen that cate that some of the absorbed moisture was in a at least 65 percent of the samples expand at a frozen state below that temperature. If the certain rate until the region of 0 to 10 deg is sample absorbs water while being cooled to - 20° reached, then change to a slower rate for about 10 C, there is ample opportunity for the water to deg and finally resume a new rate of expansion become frozen since the sample is maintained at that continues regularly until th e maximum this low temperature for at least 9 hrs. The temperature is reached. On th e other hand, th e relative high thermal coefficient of ice migh t con cooling curves of all the samples show li ttle tribute to the over-all expansion at the lower evidence of irregularities. Samples 17 , 24 , 46 , 47, t emperatures, while a gradual melting with 65, and 97 show some of th e more pronounced attendant loss in volume would reduce the ob instances of irregular h eating curves. These served slope on further heating. irregularities were reproducible on r etesting of Further evidence to show that irregula ri ties are the samples. The irregular expansion appears to be due to moisture changes in the sample during the test. a If water is prevented from entering the sample during a test the irregularities disappear. Three samples showing irregular expansions were re tested after having been dried at 80° C for 4 days and coated with a syn thetic resin water-proofing. The resul ting curves (fig. 6) showed virtually no irregularities after the treatment. It might also be noted here that the testing of nonabsorptive materials, such as glass and metal, produced E smooth coinciding curves on both heating and .....<.1 " cooling. ..J Since the uncoated samples were dry at th e start Expansion of Domestic Granites 405 probably due to moisture in some form is indi feld spars. Furthermore, each mineral expands cated by the fact that six samples showing no at differen t rates along different r l" ystallographic irregularities had an average absorption value of axes. Since a typical granite of th is investiga 0.08 percent 8 and a moisture expansion value of tion contains four major mineral constituents 0.0026 percent. On the other hand, the six plus several minor ones, all occurring in llandom samples showing the most pronounced irregula [" orientation, it would seem impossible to form a ities had an average absorption value of 0.25 per precise correlation bet""een mineral constituents cent and a moisture expansion value of 0.0072 and expansivity. Granites of similar m ineral percent. The average absorption and moisture composition but from different quarries gave expansion values for all samples were 0.20 percen t different results. and 0.0039 percent, respectively. Potash feldspar was the most abundant mineral constituent in 35 of the 48 samples tested. The 3 . Effect of Mineral Composition average of the thermal coeffi cients obta.ined for these 35 samples was 6.2 X IO - 6 per deg C between The thermal expansivity of granite is a function - 20 0 and + 60 0 C. For the 10 samples contain of the expansivities of its mineral constituents . ing soda-lime feld spar as the most abundant There are some data available on the thermal minerfl! co nstituent, the average of the coeffici rnts expansion coefficients of the chief minerals in was 6.3 X 10- 6 per deg C in this samr range. granite. The values in table 3 indicate that the Quartz wa s the most abundant mineral consti t expansion coeffi cients of quartz diff er from those uent in 3 of the 48 samples tested. The average of the feld spars, and that the coefficien ts of potfLsh of the coeffi cien ts for these samples was 7.0 X 10- 6/ feldspar are different from those of the soda-lime deg C between - 20 0 and + 60 0 C. Since the average of the coefficients of quartz measured in TAB LE 3. Nlean coe.ffi cients oj lineaT theTmal expansion JOT some oTie nted minC1·als contained in (fTanite all directions is higher than t.hose of the feldspars, it seems logical that sa mples with an abundance ~1ea.n coefficient of quartz would have higher than average coeffi of linear ther mal expansion cients. per degree Mineral Orientation C X IO ' Source 4. Thermal Expansion Data for Other Building - 20° to + 60° to Stones + 60° 0° Table 4 shows the range in linear ther-ma] expan Quartz ______Perpendicular to c-ax is._ 13.5 13.5 Johnson & Par- sidn coeffi cients determined for some other types sons [9]. Do ______Parallel to c-axi s ______7.4 7. 7 Do. of stone used for strll ctural purposes. ~\'f joro el i no ___ Parallel to b-a xi s ______I.! 1. 3 Do. Do ______P erpendicular to b-axi s __ 3. 5 3. 2 Do. Do ______P arallel to a-axi s ______17. I 17.3 Do. VI. Test Method for Moisture Expansion Oligoclase ______do ______3. I 3.7 D o. 1. Specimens . 20 ° to 100° For convenience in preparation, cylind rical Ort hoclase . ______d o ______15. 0 Kozu and Sa ik i specimens, 2 in. in diameter and 2 ;~ in . long, were [23]. cored from the original granite samples. Two Do ______Parallel to b-a xis ______0. 0 Do. plane surfaces opposite each other were ground on Plagioclase ___ P arall el to a-axis ______11. 2 Kozu and Ueda [24] . each specimen , one to provide a base and th e Do ______P erpendicula r to 010 ____ 3.8 Do. other for attachment of the gage. H ornblende __ Perpendicular to 100 ____ 0. 2 D o. Do ______P arallel to b ______7.5 Do. 00 ______P arallel to c ______i. 2 D a. 2 . Apparatus Oli yine ----- P arallel to a ______. 5. 0 Kozo, lieda & T sumori [25] . A brass tank large enough to accommodate Do .. _____ P arallel to b ______10. 0 D o. Do ______P arallel to c ______8.8 Do. two specimens was enclosed within a well insLl lated box with a cover having three holes. Two 8 A bsorption values were determined by measuring the perccntage gain in of the holes were so spaced as to provide openings weigh t of cylindrical specimens, 2 in. in dia meter a nd 272 in. Jong, after 48· for reading the gages, and the other was fitted hr immersion in water at room temperature. 1'I oisture expansion tests arc described in section V fl . with a 2-hole rubber stopper for a thermometer 406 Journal of Research T AB LE -1. -111 ean coe.fficients of linear ther mal expansion for sufficiently, the specimens were tal,en into a con some building stones other than granite troll ed temperature room. The gages wcre then mounted and the knife edges held in firm con tact M:can coefficient of li nea r with the metal strips by small rubber bands placed t hermal expan sion per dcgrrc around the specim ens and gages. Thc specimens M aterial Sonrer Orien tat ion CXIO' wi th the gages were then placed in the brass tank and left overnigh t to reach an equilibrium tcm perature (about 210 C) . A flask of distiJl ed water Limesto ne .. R ock wood, Ala __ Parallel to bedding O. i -0.1 was also left standing near th e specimens. Abou t plane. 16 hI's later the gages were read and water pourcd Do_.. ___ _. ___ do ______Perpendicular to bed· 1. 7 2.9 ding plane. into the tank until the level reached the tops of Do_.. ____ B edford , Ind ______Parall el to bedding 2. S 3.4 the specimens. Gage readings were taken at plane. Do ___ .. _ _. ... do _...... Perpendicular to beel· 4. 2 4.2 approximately I-hI' inter vals for 8 hI's and a final d ing plane. reading after 24 hI's. ~ I arb l e .. _.. South Do\-cr , N. Y _ "A" direction (ran- S. i 9.2 dom ). During any test, the tempcrature of the water Do ...... do ...... Pcrpen dicLil ar to "A n 5. 2 5. 9 did not vary more than ± 0.5 deg C. The maxi direction. Sa ndsto ne __ M cDermott, Ohio .. Parall el to bedding 9. 2 10.2 mum diff erence in tcmperaturc of the water be plane. twcen the time of th e first and last reading was Do ...... do ...... Perpendicular to bed· 9.5 10.5 d ing plane. 0.2 deg C. E rrors due to thermal expansion of the gage or spec im en werc thcreby virtuall y el im inated. and funnel. rrwo calibra ted Tuckerman op tical To verify the fact tha t thc length changes strain gages [26], of 2-in. gage length, reading to obse rved in the stone spec imens were caused only 2 X 10- 6 in. were used for the expansion measure by the absorption of wa te r, two metal cylindrical ments (fi g. 7). The knife edges of the gages were spccim ens were tes ted in the manner described. mounted on thin brass strips attached to the speci The greatcst change obse rved in four tests wa,s men with water-proof cement. 2 X lO- G in . Sin cc the gages arc sensitive only to this amount, it can bc assumed that no significant 3 . Procedure dim cnsional change took place during these tests. The specim ens were dri ed at 1050 C for 24 hr and cooled in a desiccator. The brass strips VII. Moisture Expansion Results were then attached to one plane face of each sp eCIm en an d, when the cemen t had hardened Ta ble 2, column 7, gives the linear moisture expansion resul ts for the 48 samples of granite. PRISM ADJUSTER The valu es ra,nge from a min imum of 0.0004 pcrcen t to a maximum of 0.0090 percen t, wi th an average of 0.0039 percen t. Although the immer LOZ ENGE sion period fo r each test was 24 hI'S, fi ve specim ens exhibited maximum expansion in 1 hI' , and 28 specimens between 2 and 8 hI'S. Figure 8 gives moisture expansion curves for nine samples, together with th e corresponding absorption curves. These how results typical for all samples tested. Both the expansion and ab GRANI T E SPECIMEN sorption curves tend to rise sharply during the fi rst few hours and then quickly level off , but there seems to be littlc correlation between the amounts METAL SUPPORTS of expansion and absorption . The moisture ex SIDE END pansion values were also compared with density, F I GURE 7. Diagram oj T uckennan optical strain gage porosity, and mineral composition but no correla mounted on granite specimen. tion was found with these properties. Expansion of Domestic Granites 407 'fA RLE 5. j\I[oisture expansion values of some building stones other than granite Exoan Type of stone Source sion ------_·_------1-·---- Percent Limestone ______Rockwood, Ala ______0.0032 Do ______Bedford. Ind______. 0028 56 . ",...... ------Do ______Cedar Park, Tex. ______. 0015 ~ c 56 , Marble ______Gantt's Qu arry, Ala ______. 0010 Do ______Whitestone, Ga______. 0025 . 53 Do ______Swanton, Vt. ______. 00lO ~ ------~~------~ Sandstone ______Amherst,Ohio.______. 013 Do______Glenmont, Ohio ______. 010 Do ______McDermott,Ohio______. 044 z o iii ....- ______!LZ ______- ~ z z o ~ ~ usually much lower than those required to produce x Il. w cr o rupture, frequent repetition and reversal of direc if> CD tions may ultimately cause fractures. These q: stresses in a compact material like granite are probably greater than in more porous materials. Two cases can be cited in which rather large spalls occurred at corners of granite masonry o where the superimposed loads were not sufficient l~ to account for the fractures. It seems possible that a complication of stresses due to temper 2 atm'e effects might occur in such parts of struc tures. o 2 4 6 8 24 One case of cracking has b een examined where TIME ,HOURS the facade of a building is made of a granite with unusually large quartz crystals. The polished FIGU R E 8. Moisture ex pansion and correspondfng absorp tion curves of nine granite samples. columns show considerable cracking on the parts most exposed to the sun, while the shaded parts r 0 , Moistnre ex pansion points; . , a bsorption points; fi gures refer to sample numbers in table 2. appear to be ent.irely sound. Several cubic inches of granite have crumbled away from the corner Table 5 shows the moisture expansion values of a column base that' is freely exposed to the obtained for several samples of other types of afternoon sun. This case provides evidence that building stone, tested in the manner described. unequal expansion of the minerals might be causing the deterioration. VIII. Thermal and Moisture Expansion in The moisture expansion values obtained in this Relation to Durability of Granite study of granites are not large in comparison with values obtained on certain other types of building Merrill [27] cites several cases of rock weather materials. In general the total expansion of ing, including granite, which he attributed to granites due to moisture (using average values) is heating and cooling effects. Temperature changes equivalent to the thermal expansion produced by a can produce internal stresses in two ways: (1) a temperature change of about 6 deg C. It was not thermal gradient due to heating or cooling of one possible to determine by the methods used in this surface causes unequal expansion of layers at study if a wet specimen expands at the same rate, different depths; (2) the various mineral con due to a thermal change, as a dry specimen. It stituents of granite have different expansion rates can be assumed that the two types of expansion among themselves and also in different crystal occur simultaneously, and the total expansion of a lographic direct.ions. Stresses caused by both wet granite can be computed as the sum of the conditions come into action simultaneously and moisture and thermal changes. Using the aver augroent each other. Although such stresses are age values for the thermal and moisture expan- 408 Journal of Research sions it is found that a 100-ft course of granite beeome apparent on certain monument'al struc may 'expand 0.05 in. on becoming wet and 0.08 tures after a long period of time. in. due to a 10 deg C increase in temperature, giving a total of 0.13 in. If the course is con IX. Summary strained, there will be an elastic deformation of l. Thermal and moisture expansion determin a 0.13 in. , and one may estimate the stress produced tions were made on 48 samples of domestic granites where the modulus of elasticity (E value) is rei)~'esentative of those used in important struc known. Assuming an E value of 7,000 ,000 tures. Complete descriptions are given of the Ib/in. 2, the resulting compressive stresses would apparatus and test procedures used in making the be 760 Ib /in.2, which is high in comparison with the measurements. The major mineral constituents of superimposed weight stresses in most structures, the samples were determined by petrographic but rather low in comparison with strength of methods. granite. 2. The linear thermal expansion coefficients for A more serious eff ect might result on coping all samples ranged from 4.8 to 8:3 X 10- 6 and aver courses where the masonry has more freedom of aged 6.2 X 10- 6 per deg C for the range - 20 ° to movement. The combined effect of thermal and +600 C (- 4° to 140° F) . This average value is moisture expansion may cause each block to move smaller than the average (7.7 X 10- 6) ob tained by toward the ends of the structure; but when con previous investigators, which may be accounted traction occurs, there is insufficient tensile strength for in part by the lower temperature range explored in the mortar joints to pull the blocks back to their in the present study. The coefficients of 84 per original position, hence cracks at the joints arc cent of the samples were between 5 and 7 X 10- 6 produced. These cracks become partly filled wi th pel' deg C. dirt and sand from the mor tar and the next ex 3. Some irregularities in the heaL ing curves were pansion results in moving the blocks still further. evident in about 65 percent of the samples tested. Open joints allow water to en ter, and freezing may These irregularities, which OCCUlTed somewhere cause rupture of the mortar in joints at lower between 0° and + 20 ° C, were probably due to levels. moisture changes in the sample during the test. Schaffel' [28] points out that a stress gradient is 4. Although a definite correlation between produced by moisture from rains because the stone thermal expansion and mineral composition could is soaked to only a slight depth. He believes that not be established from the values obtained be frequent repetition of such stresses migh t produce cause ·of the heterogeneous composition of grani te, injurious effects. Observation on granite build the res ul ts indicated that granites with excep ings shows that the surface of granite often scales Lionally hi gh percentages of quartz had higher off, but the scaling is usually confined to small than average coefficients. areas of the lower co urses. This can be explained 5. T he moisture expansion values for a 24-h1' by the assumption that salts are carried up in to soaking period at constant temperature, ranged the granite by ground water; and crystals arc from 0.0004 to 0.009 percent with an average of formed in the pores. The expansive action of the 0.0039 percent for all samples. The expansion salt crystals in forming causes tensile stresses increased rapidly during the first few hours only; perpendicular to the exposed face in sufficient thereafter it increased very slowly or remained amounts to cause the flaking. It seems likely constant. Little correlation was fonnd between that a compressive stress parallel to the face result moisture expansion and mineral composition, ing from a moisture gradient would aggravate the absorption, specific gravity, or porosity. flaking action. 6. It is probable that the repeated thermal ex Based on sueh studies as have been made on pansion and contraction of granite resulting from structures, it is the authors' belief that most natural weather conditions, is one of the many granites are not affected as seriously by thermal factors contributing to the eventual disintegra and moisture expansion as by other physical and ti on of the stone. rrhese adverse effects of tem chemical agencies. D eterioration due to expan perature are due mainly to stresses produced by sion probably does no t manifest itself appreciably the unequal expansion of layers at different depths during the normal life of a building, bu t it may a nd by the unequal expansion of the mineral con- Expansion of Domestic Granites 409 877106- 50--5 ------~ ~~ - - - stituents. The deterioration resulting from ther [12] T. :vJ:atsumoto, A study of the effect of moisture con mal expansion progresses very slowly and probably tent upon the expansion and contraction of plain does not manifest itself appreciably during the and reinforced concrete, Univ. of II!. Eng. Exp. Sta., Bu!. No. 126 (Urbana, 1921). normal life of a building, but it may become ap [13] R E. Stradling, Effects of moisture changes on build parent on certain monumental structures after a ing materials, Dept. Ind. Sci. Research, Bid. long period of time. R esearch BuL 3 (London, 1928). There is some evidence that moisture expansion [1 4] N. Royen, Naturstenens tekniska egenskaper och dess an vandning till hysbyggnader och dkydds may contribute to the weathering of granite but rumsanlaggningar, Trans. Chalmers Univ. of T ech to a lesser degree than thermal expansion. How nology, Gothenburg, Sweden No.8 (1942). ever, when moisture and thermal expansions occur [15] D . Vol. Kessler, A. Hockman, and R. E. Anderson, simultaneously they may cause masonry units to Physical properties of terrazzo aggregates, NBS be moved slightly out of position and injure the Building Materials and Structures Report BMS98 (1943). mortar joints. [16) D . W. Kessler, H. Insley, and W. H. Sligh, Physical mineralogical and durability studies on the building The authors thank P eter Hidnert and James B. and monumental granites of t he United States, J . Saunders of this Bureau for their very helpful R esearch N BS 25, 161 (1940) RP1320. [17] G. E. Merritt, The interference method of measuring suggestions regarding the thermal expanSIOn thermal expansion, J. R esearch NBS 10, 59 (1933) method used in this inves tigation. RP515. [18) J . B. Saunders, Expansivity of a Vycor brand glass, J . X. References Research NBS 28, 51 (1942) RP1445. [19) H. Fizeau, Uber die Ausdehnung starrer Korper durch [1) VlT. H. C. Bartlett, Experiments on the expansion and die Warme, Ann. Physik 128, 564 (1866). contraction of building stone, Am. J. Sci. & Arts 22, [20] J . B. Saunders, An apparatus for photographing inter 136 (1832). ference phenomena, J . R esearch N BS 35,157 (1945) [2) A. J . Adie, On t he expansion of different kinds of RP1668. stone from an increase of temperature with a de [21) A. Q. Tool, D. B. Lloyd, and G. D . Merritt, Dimen scription of the pyrometer used in making t he ex sional changes caused in glass by heating cycles, periments, Trans. Roy. Soc. Edinburgh 13, 366 J. Am. Ceramic Soc. 13, 641 (1930). (1836). [22) J. B. Saunders, Interferometer measurements on the [3) T . M. R eade, Origin of mountain ranges (Taylor and expansion of iron, J . R esearch NBS 33, 75 (1944) Francis, Ltd. London, 1886). RP1597. [4) Report on tests of metals, etc. VlTatertown Arsenal, [23] Kozu and Saiki, Sci. Rep . Toh. Uoiv. 3, 2, 205 (1925) U. S. War Dept., 322 to 418 (1895). a s quoted in the Handbook of Physical Constants, (5) W. R. Baldwin-Wiseman and O. W. Griffi th, Physical Geo!. Soc. Am. Special Papers, No. 36, 33 (1942). properties of building materials, Proc. Inst. Civil [24] Kozu and Ueda, Proc. Imp. Acad. Japan 9, 262 (1933); Eng. 3, 179, 290 (1909). 10, 25 (1933) as quoted ii1 the Handbook of Physi· [6) N. E. Wheeler, On the thermal expansion of rock at cal Constants, Geo!. Soc. Am. Special Papers, No. high temperatures, Trans. Roy. Soc. Can. 3, 4, 26, 33 (1942). 19 (1910). [25] Kozu, Ueda, and Tsumori, Sci. R ep . Toh. Univ. 3, [7) J . H. Griffith, Thermal expansion of typical American 10, 222, (1934), as quoted in the Handbook of rocks, Iowa Eng. Expt. Sta. Bulletin 128 (1936). Physical Constants, Geo!. Soc. Am. Special Papers, [8) T . F. Willis and M. E. DeReus, Thermal volume No. 36, 33 (1942). change and elasticity of aggregates and their effect [26] L. B. Tuckerman, Optical strain gages and exten on concrete, Am. Soc. Testing Materials 39, 919 someters, Proc. Am. Soc. Testin g Materials 23, pt . (1939) . 2, 602 (1923). [9] W. H . Johnson and W. H . Parsons, Thermal expan [27] G. P. Merrill, Rocks, rock-weathering and soils, p. sion of concrete aggregate materials, J . Research 158to 162 (Macmillan Co., New York, N. Y.1906.) NBS 32, 101 (1944) RP1578. [28] R. J . Schaffer, The weathering of natural building [10) Schumann, Protocoll de Verhandlungen del' IV Gener stones, Dept. Ind. Sci. Research, Bldg. Research al-Versammlung des Vereins Deutscher Cement Spec. R eport No. 18, London (1932). Fabrikanten (1881). [11) J . Hirschwald, Handbuch del' bautechnischen Ges teingsprufung (Gebriider Borntraeger, Berlin , 1912). WASHINGTON, October 27,1949. 410 Journal of Research