F

Deflection fatigue of Ti-6Al-7Nb, Co-Cr, and gold alloy cast clasps

Ahmad Mahmoud, BDS,a Noriyuki Wakabayashi, DDS, PhD,b

Hidekazu Takahashi, DDS, PhD,c and Takashi Ohyama, DDS, PhDd

Division of Oral Health Sciences, Graduate School, Tokyo Medical andDental University, Tokyo, Japan

Statement of problem. There is little information about the deflection fatigue of clasps in relation to stressdistribution.

Purpose. The aim of this study was to investigate the fatigue resistance and permanent deformation of castclasps made of titanium and other dental alloys and to relate the fatigue resistance with the calculated stressvalues.

Material and methods. Twenty-five Ti-6Al-7Nb, 25 Co-Cr, and 15 Type IV gold alloy clasps were subjectedto cyclic deflection of preset values of 0.25 mm, 0.50 mm, or 0.75 mm, for 106 cycles (n = 14). Finite elementmodels were created to calculate principal stresses within the specimens. Fatigue life, retentive force, andpermanent deformation were recorded, and the fracture locations were determined microscopically. The resultswere characterized in relation to the stress within the clasps. One-way analysis of variance and Tamhane’s post-hoc tests were used to compare the results of the 9 material-deflection groups (a=.05).

Results. Ti-6Al-7Nb clasps exhibited significantly less permanent deformation than the other clasps underrelatively greater deflections, indicating better adaptation to the tooth surface. However, the fatigue life of theTi-6Al-7Nb clasps under 0.75-mm deflection, with the stress above the alloy’s 0.2% yield strength, wassignificantly shorter than those under smaller deflections. The gold clasps showed significantly longer fatigue lifethan the other clasps under the 0.50-mm deflection. High-stress areas within the fatigue clasp specimenscoincided with the fracture locations. The probabilities of fatigue fracture and permanent deformation wereclosely related to the material strengths and the preset deflections.

Conclusion. To minimize fatigue failures, the cast clasp should be designed with consideration of the stressesdistributions within the clasps. (J Prosthet Dent 2005;93:183-8.)

CLINICAL IMPLICATIONS

The Ti-6Al-7Nb and gold clasps demonstrated fatigue resistance that allows placement inundercut greater than 0.25 mm, which is suitable for Co-Cr alloy clasps. These clasps may beindicated when esthetics or periodontal health is a primary concern.

Ti-6Al-7Nb alloy has been developed as an alterna-tive to Co-Cr and other existing removable partial den-ture alloys because it offers excellent biocompatibility,

This study was supported by Grants-in-Aid for Scientific Research,12771170 (N.W.) and 14571840 (N.W.), from The Ministry ofEducation, Science, and Culture of Japan. Presented at the 110thScientific Meeting of the Japan Prosthodontic Society, Nagano,Japan, October 2003.

Presented at the 82nd General Session and Exhibition of theInternational Association for Dental Research; an Arthur R.Frechette Research Award Finalist, Hawaii, March, 2004.

aResident, PhD student, Removable Prosthodontics, Department ofMasticatory Function Rehabilitation.

bResearch associate, Removable Prosthodontics, Department ofMasticatory Function Rehabilitation.

cAssociate Professor, Advanced Biomaterials, Department of Re-storative Sciences, Division of Oral Health Sciences.

dProfessor and Chair, Removable Prosthodontics, Department ofMasticatory Function Rehabilitation.

EBRUARY 2005

good resistance to abrasion, and other advantages forrelatively low cost.1-4 Like other titanium alloys, theTi-6Al-7Nb alloy has an elastic modulus that is approx-imately half of Co-Cr alloy, which increases its flexibil-ity.5 The increased flexibility allows the retentive clasparms to be placed into larger undercuts on abutmentsthan Co-Cr alloy clasps, and this may be importantwhen esthetics or periodontal health is a primary con-cern.

Clasps undergo permanent deformation and fatiguefracture under repeated flexures caused by denture in-sertion and removal and mastication.6-9 The fatiguelife of cast clasps made of commercially-pure titaniumwas reported to be shorter than that of Co-Cr andgold alloy clasps.5 However, the fatigue fracture andpermanent deformation of cast clasps made of titaniumalloy has not been sufficiently assessed in relation tostress distribution, and little is known about how these

THE JOURNAL OF PROSTHETIC DENTISTRY 183

THE JOURNAL OF PROSTHETIC DENTISTRY MAHMOUD ET AL

clasps would function in long-term clinical use.Permanent deformation and fatigue fracture are causedby the stress created in the clasp.10,11 The stress distribu-tion may depend on the elastic modulus of the alloy, di-mensions and curvature of the clasp,12,13 and theamount and direction of defection in relation to theabutment undercut.14

The purpose of this in vitro study was to investigatethe fatigue of the Ti-6Al-7Nb clasps compared toCo-Cr and gold alloy clasps. The fatigue life, retentiveforce, and permanent deformation of clasps under cyclicdeflections of different magnitudes were evaluated, anddeflection fatigue was characterized in relation to thestress distributions within the clasps.

MATERIAL AND METHODS

A metal cylinder with a 10-mm outer diameter wasused for the adaptation of a pre-formed tapered wax pat-tern (Rapid-Flex; Degussa, Dusseldorf, Germany). Theclasp arm pattern originated from a plate and curvedaround the cylindrical surface 120 degrees in a singleplane (Fig. 1, A). Average width and thickness of theclasp arm at 30 degrees were 0.92 mm and 0.97 mm, re-spectively, while those at 120 degrees were 0.73mm and0.79 mm, respectively. The plate served as an attach-ment for fixation to a fatigue-testing machine (250N;Shimadzu Corp, Kyoto, Japan). The original patternwas duplicated using a split mold to make acrylic resinpatterns (Pattern resin, GC, Tokyo, Japan). A sphericalbead 0.5 mm in diameter was glued onto the innersurface of each clasp tip to provide a point of force appli-cation. The patterns were invested and cast using Ti-6Al-7Nb (T-alloy Tough; GC) with a magnesia-basedinvestment (Selevest CB; Kobelco, Hyogo, Japan),Co-Cr (Biosil L; Degussa) and a Type IV gold alloy

Fig. 1. A, Schematic illustration of fatigue test specimen.Specimen consists of clasp arm, loading sphere, and plate forfixation to testing machine. B, Finite element model of clasparm. Dimensions of eachmodel were based on test specimen.

184

(Degulor M; Degussa), both with a phosphate-bondedinvestment (Biosint-supra; Degussa). The compositionsand material properties of these alloys are shown inTable I. The Ti-6Al-7Nb was cast using argon arc melt-ing technology with a centrifugal vacuum pressure cast-ing machine (Valcan-T; Shofu, Kyoto, Japan). The Co-Cralloy was cast using high-frequency induction meltingtechnology with a centrifugal casting machine (SA-2000;Sankin, Tokyo, Japan). The gold alloy was cast usingelectric resistance melting technology with a centrifugalcasting machine (TS3; Degussa). The casting proce-dures were determined following the manufacturers’ in-structions for the metals and investments. Recoveredcastings were lightly cleanedwith airborne-particle abra-sion using 80-mm aluminum oxide particles. Each goldalloy clasp was heat-treated according to the manufac-turer’s instructions.

Fatigue test

Twenty five Ti-6Al-7Nb, 25 Co-Cr, and 15 gold al-loy cast specimens were divided into different experi-mental groups according to preset clasp tipdeflections, 0.25 mm, 0.50 mm, and 0.75 mm, whichwere designed to compare the clasps’ functions underdifferent undercuts to the tooth surfaces. The 3 groupsof the Ti-6Al-7Nb clasps based on deflections were de-noted by codes T25 (n = 5), T50 (n = 10), and T75(n = 10). The 3 groups of Co-Cr clasps denoted byC25 (n = 5), C50 (n = 10), and C75 (n = 10), andG25 (n = 5), G50 (n = 5), and G75 (n = 5) were usedfor the gold alloy clasps. Each specimen was fixed tothe testingmachinewith screws and subjected to a sinus-oidal cyclic deflection generated by the radial directionforce at the tip of each clasp arm at a frequency of 5 Hz.The load/deflection curve was monitored, and the testwas terminated when the maximum force was reducedto less than 15% of the initial load, or when 106 cycles15

were completed. The permanent deformation wascalculated as the difference in piston tip-loading spheredistance between the position at the beginning of eachcycle and the position at the first cycle. The actual deflec-tion was thereafter obtained by subtracting the perma-nent deformation from the preset deflection. One-wayanalysis of variance (ANOVA) and Tamhane’s T216,17

post-hoc tests were conducted to determine the differ-ences in the permanent deformations, applied forces,and number of loading cycles between the 9material-de-flection groups were significant or not. After the fatiguetest, an optical microscopic (VH-5000; Keyence Corp,Osaka, Japan) examination was performed to locate thefracture site in each clasp specimen.

Finite Element Analysis

By the use of the preprocessor of the finite elementmethod (FEM) computer software (ANSYS 7.1 FEM;ANSYS Inc, Canonsburg, Pa), a 3-dimensional finite

VOLUME 93 NUMBER 2

THE JOURNAL OF PROSTHETIC DENTISTRYMAHMOUD ET AL

Table I. Composition and mechanical properties of alloys evaluated

Ti-6Al-7Nb Co-Cr Type IV gold

Composition (wt %) 86.5 Ti, 7 Nb, 6 Al,

0.5 others

62.5 Co, 30.5 Cr, 5 Mo, 1 Si,

0.4 Mn, 0.3 C, 0.3 N

70 Au, 4.4 Pt, 13.5 Ag, 8.8 Cu,

2 Pd, 0.1 Ir, 1.2 others

Modulus of elasticity (GPa) 123 220 90

0.2% Proof stress (MPa) 890 710 620

Tensile strength (MPa) 950 900 740

Percentage elongation (%) 5 6 17

All data were based on information provided by manufacturer for each respective material except for modulus of elasticity of Type IV gold alloy.18 Note: recorded

yield strength value is affected by offset value and can be substantially different from proportional limit.24

element model for each fatigue test clasp was createdbased on the width and thickness of each specimen,which were measured at 7 representative locations usinga measuring microscope (MM-60; Nikon, Tokyo,Japan) (Fig.1, B). Each model was meshed by 8-nodehexahedral elements. In the simulation, each clasp tipwas displaced by the average of actual deflections inthe radial direction while the base area at the clasp’sshoulder was fixed. Elastic moduli of 90 GPa,18 220GPa, and 123 GPa (Table I) were input into the pro-gram to simulate the gold, Co-Cr, and Ti-6Al-7Nb al-loys, respectively. A Poisson ratio of 0.33 was used forall of the alloys.19 The friction coefficient between therod and the clasp surface was assumed to be 0.2.14 Theprincipal stress distribution within each clasp modelwas calculated using the post processor of the FEMcomputer software.

RESULTS

Fatigue resistance

Figure 2 displays the fatigue life of each specimen. Foreach test group, a representative specimen was chosen,and its permanent deformation and force required fordeflection are respectively shown in Figures 3 and 4 asfunctions of loading cycles. In the early stage, instantincreases in permanent deformations were recorded forall specimens. Thereafter, all of the clasps with 0.25-mmdeflections, except for 1Co-Cr clasp, sustained con-stant forces and permanent deformations and survived106 cycles without fracture. One Co-Cr clasp witha 0.25 mm deflection and all of the other clasps withlarger deflections also sustained almost constant forcesand permanent deformations after the early stages.However, those clasps that fractured before the 106

cycles demonstrated sudden force decreases accompa-nied by dramatic deformation increases shortly beforefracture. The average numbers of loading cycles for fail-ure, the average loads for deflection, and the averagepermanent deformations are listed in Table II.

One-wayANOVA showed that the differences in loadrequired for deflection, the permanent deformation, andthe number of cycles to failure among the 9material-de-flection test groupswere significant (P,.05). The results

FEBRUARY 2005

of the post-hoc comparison between the 9 material-deflection experimental groups are shown in Figure 2and Table II. It should be noted that one C50 claspwith the exceptionally greater number of cycles to failurewas considered an outlier andwas excluded from the sta-tistical analysis (Fig. 2). Postoperative microscopic ob-servations indicated that the fracture sites of all of thefailed clasps were located between 48 and 91 degreesfrom the clasp arm shoulder, with a mean value of 716 11 degrees (Fig. 5). There were no significant differ-ences in the fracture sites among the 9 test groups.

Finite element analysis

In the finite element models, the sites where themax-imum principal stresses occurred were located between39 and 85 degrees from the clasp shoulder, with anaverage of 706 11 degrees. The average principal stressdistribution for each test group is illustrated in Figure 6.The stress values detected between 35 and 90 wererelatively constant and above 85% of the maximumstress. There were no significant differences in the sitesof the maximum stress among the clasp materials or

Fig. 2. Fatigue life distributions of specimens as function ofnumber of cycle. Each vertical line segment represents claspspecimen. Bar with asterisk was considered outlier and notincluded in statistical analysis. Left superscript letters of eachtest group represents homogeneous subsets resulting frompost-hoc comparison. Each subset represents groups withinsignificant differences. (P>.05).

185

THE JOURNAL OF PROSTHETIC DENTISTRY MAHMOUD ET AL

Table II. Means (standard deviations) of fatigue test and FEM results

Material

Preset deflection

(mm) Ti-6Al-7Nb Gold alloy Co-Cr

Test group code / Number of specimens* 0.25 T 25 / 5 G 25 / 5 C 25 / 5

0.50 T 50 / 9 G 50 / 5 C 50 / 8

0.75 T 75 / 9 G 75 / 4 C 75 / 9

Cycles to failure 0.25 No failure No failure 4 did not fail

0.50 22,469(9,876) 120,500(18,014) 28,929(10,358)

0.75 2,875(1,000) 32,375 (8,845) 15,806 (5,312)

Permanent deformation (mm)** 0.25 9.6 (4.0) a 13.4 (2.4) a, b 14.0 (3.5) a, b

0.50 27.4 (3.9) c 40.0 (8.9) b, c, d 114.4 (12.7) e

0.75 64.7 (11.4) d 132.0 (9.6) e 298.1(21.2) f

Load to deflection (N) 0.25 6.0 (0.8) 6.3 (0.8) 12.0 (0.9)

0.50 13.9 (1.9) 12.3 (1.4) 21.0 (1.6)

0.75 21.9 (1.8) 18.5 (1.9) 25.9 (2.9)

Max. principal stress (MPa)*** 0.25 362 (17) 296 (14) 640 (23)

0.50 706 (47) 559 (27) 1095 (52)***

0.75 1070 (26)*** 773 (23)*** 1254 (56)***

*Exclusion of some specimens due to errors in testing is reason for deviation of number of specimens from originally assigned 5 or 10 specimens.

**Subscript letters (a to f) indicate homogeneous subsets resulting from post-hoc test. Insignificant difference in mean permanent deformation was evidenced

among groups of identical subset (P>.05).

***Recorded maximum stress value exceeding respective material’s 0.2% yield strength.

among the preset deflections. The average maximumprincipal stresses calculated are shown in Table II.

DISCUSSION

Based on the loads required to deflection, both Ti-6Al-7Nb and Type IV gold alloy would be clinically ad-

Fig. 3. Change of permanent deformation for representativespecimen from each group. Solid lines: Ti-6Al-7Nb clasps;dotted lines: gold alloy clasps; dashed lines: Co-Cr clasps.

186

vantageous due to low rigidity when compared to theCo-Cr clasps, which is expected to have aminimum pos-sibility of traumatic overloading to the abutment toothduring insertion and removal.12 Fatigue fracture is com-monly believed to occur at cyclic stress levelsmuch lowerthan that needed to cause failure on a single applicationof load.19 However, it was previously claimed that the

Fig. 4. Load required for deflection as function of deflectioncycles for representative specimen from each group. Solidlines: Ti-6Al-7Nb clasps; dotted lines: gold alloy clasps;dashed lines: Co-Cr clasps.

VOLUME 93 NUMBER 2

THE JOURNAL OF PROSTHETIC DENTISTRYMAHMOUD ET AL

Co-Cr alloy could withstand a stress slightly above itsproportional limit without failure over infinitely manycycles.20 Results of the present study were not in conflictwith this claim. Four out of 5 Co-Cr clasps tested under0.25-mm preset deflections, which had maximum ten-sile stresses (average of 640 MPa) slightly below the al-loy’s reported 0.2% yield strength (Fig. 6), survivedover 106 cycles. However, all of the other Co-Cr claspsunder larger preset deflections with stresses above theyield strength failed long before reaching 106 cycles.All of the Ti-6Al-7Nb and gold alloy clasps with 0.5-mm deflections and maximum stresses below the mate-rials’ yield strengths failed before reaching 106 cycles.Furthermore, the fatigue lives of Ti-6Al-7Nb clasps un-der 0.75 mm deflection, with the maximum calculatedstress above its yield strength, remarkably decreased.These results indicate that the fatigue limits of these al-loys are well below their yield strengths.

At relatively high deflection amounts, the materialdifferences greatly affected the magnitude of the perma-nent deformations of the clasps. The Ti-6Al-7Nb claspshad considerably smaller permanent deformations whencompared to the Co-Cr clasps under all preset deflec-tions greater than 0.25 mm. They also showed muchsmaller permanent deformations than the gold alloyclasps under the 0.75 mm deflection. This may be be-cause of thematerial’s high ratio of yield strength to elas-ticmodulus,21which indicates a relatively largerworkingrange (7.24 for the Ti-6Al-7Nb, 3.23 for Co-Cr,and 6.89 for gold alloy). These results indicate betteradaptation of Ti-6Al-7Nb clasps to tooth surfaceswhen designed to engage undercuts greater than0.25 mm. A relatively higher load was needed to gen-erate a larger permanent deformation in the clasp (21.9

Fig. 5. Fractured Co-Cr specimen tested under 0.50 mmpreset deflection with enlarged view.

FEBRUARY 2005

N for T75) than the other materials tested because ofthe greater yield strength of the Ti-6Al-7Nb alloy.This provides the Ti-6Al-7Nb clasps better resistanceto permanent deformation, which is advantageous insituations of accidental overloading.9,22

Greater permanent deformation of Co-Cr clasps ledto reduction of the actual deflections, and this contrib-uted to the extension of the fatigue lives because smalleractual deflections would reduce the maximum tensilestress within the clasps.12 Permanent deformation wasobserved in the clasps that had maximum stresses belowthe yield strengths of the alloys (Table II). It is suspectedthat these deformation values were partially due to thesurface flattening of the load application sphere at theclasp tip, which was confirmed in postoperative micro-scopic observations.

Although the calculated average maximum principaltensile stresses in clasps of the C50, C75, T75, andG75 groups exceeded the yield strengths of the respec-tive alloys (Table II), no specimen in these groups failedduring the first cycle, and all survived at least 1500cycles. The permanent deformations did not have nota-ble increases after the first few cycles, and the deforma-tions remained stable until just before fracture. Theresidual stresses and strain hardening induced by thepermanent structure deformations after the deflectioncan explain these findings.19,23 The residual stressesgenerated in the clasp after the first few cycles shouldhave comprised compressive stress on the inner surface

Fig. 6. Average principal stress distribution for each testgroup. Each volume was divided into 9 different colorsaccording to stress levels. Red zone indicates greatest stressregion (tension), blue zone indicates smallest. Scales belowshow maximum and minimum scale values, as well asboundary values between levels. Bold numbers indicatereported 0.2% yield strengths (Table I).

187

THE JOURNAL OF PROSTHETIC DENTISTRY MAHMOUD ET AL

and the tensile stress on the outer surface of theclasp.19,23 Besides strain hardening, these stresses couldfurther increase the clasp’s apparent yield strength andresistance to further permanent deformation.19,23 Itshould also be noted that the stresses were calculatedin themodels by inputting the actual deflectionswithoutconsidering these residual stresses. This may result inhigher calculated tensile stresses than those actuallypresent in the specimens. Post-insertion adjustment orre-adaptation of a clasp to the tooth surface may causepermanent deformation in the opposite direction tothat seen in this study and may generate some tensile re-sidual stress on the inner surface. This may subsequentlyincrease the susceptibility to further permanent defor-mation with the reduction of actual deflection, whichcould apparently extend the clasp’s fatigue life.5

The fracture locations in all of the failed specimenswere found to be within the range where the stress wasmore than 85% of the maximum stress. This impliesthat the stress distribution within the clasp was closelyrelated to the fracture location. However, since an areaof more than 85% of the maximum stress is relativelywide and stress is almost constant in this area, the siteof the maximum stress could not be used as a good pre-dictor of the fracture site.

It should be noted that the experimental condition inthis study may differ from that encountered clinically.For example, fixed direction and magnitude were usedfor the experimental loading in this study, althoughthe real direction may be different and the loading con-ditions can actually fluctuate.14 Also, the fatigue test inthis study was conducted in dry air atmosphere whichis different from the oral invironment.19 Further studiesare needed to develop a sound base for a computerizedmethodology of design optimization that should mini-mize the failure possibilities of cast clasps.

CONCLUSIONS

The gold alloy clasps exhibited significantly longer fa-tigue lives, while the Ti-6Al-7Nb clasps showed signifi-cantly greater resistance to permanent deformationunder cyclic deflections. The results of the fatigue ana-lyses suggest that the Ti-6Al-7Nb and gold clasps aresuitable for use with undercuts greater than 0.25 mm.However, a reduction in the fatigue resistance of theTi-6Al-7Nb claspwas clear when the tensile stress withinthe clasp exceeded its yield strength.

REFERENCES

1. Wang TJ, Kobayashi E, Doi H, Yoneyama T. Castability of Ti-6Al-7Nb alloy

for dental casting. J Med Dent Sci 1999;46:13-9.

2. Watanabe K, Miyakawa O, Takada Y, Okuno O, Okabe T. Casting behav-

ior of titanium alloy in a centrifugal casting machine. Biomaterials 2003;

24:1737-43.

188

3. Iijima D, Yoneyama T, Doi H, Hamanaka H, Kurosaki N. Wear properties

of Ti and Ti-6Al-7Nb castings for dental prostheses. Biomaterials 2003;24:

1519-24.

4. Kobayashi E, Wang TJ, Doi H, Yoneyama T, Hamanaka H. Mechanical

properties and corrosion resistance of Ti-6Al-7Nb alloy dental castings.

J Mater Sci Mater Med 1998;9:567-74.

5. Vallittu PK, Kokkonen M. Deflection fatigue of cobalt-chromium, tita-

nium, and gold alloy cast denture clasp. J Prosthet Dent 1995;74:412-9.

6. Keltjens HM, Mulder J, Kayser AF, Creugers NH. Fit of direct retainers in

removable partial dentures after 8 years of use. J Oral Rehabil 1997;24:

138-42.

7. Saito M, Notani K, Miura Y, Kawasaki T. Complications and failures in re-

movable partial dentures: a clinical evaluation. J Oral Rehabil 2002;29:

627-33.

8. Hofmann E, Behr M, Handel G. Frequency and costs of technical failures

of clasp- and double crown-retained removable partial dentures. Clin

Oral Investig 2002;6:104-8.

9. Carr AB, McGivney GP, Brown DT. McCracken’s removable partial pros-

thodontics. 11th ed. St. Louis: Elsevier; 2005. p. 79-115.

10. Vallittu PK. Fatigue resistance and stress of wrought-steel wire clasps. J

Prosthodont 1996;5:186-92.

11. Gapido CG, Kobayashi H, Miyakawa O, Kohno S. Fatigue resistance of

cast occlusal rests using Co-Cr and Ag-Pd-Cu-Au alloys. J Prosthet Dent

2003;90:261-9.

12. Yuasa Y, Sato Y, Ohkawa S, Nagasawa T, Tsuru H. Finite element analysis

of the relationship between clasp dimensions and flexibility. J Dent Res

1990;69:1664-8.

13. Sato Y, Yuasa Y, Akagawa Y, Ohkawa S. An investigation of preferable ta-

per and thickness ratios for cast circumferential clasp arms using finite el-

ement analysis. Int J Prosthodont 1995;8:392-7.

14. Sato Y, Abe Y, Yuasa Y, Akagawa Y. Effect of friction coefficient on Akers

clasp retention. J Prosthet Dent 1997;78:22-7.

15. Wiskott HW, Nicholls JI, Belser UC. Stress fatigue: basic principles and

prosthodontic implications. Int J Prosthodont 1995;8:105-16.

16. Tamhane AC. Multiple comparisons in model I one way ANOVAwith un-

equal variances. Comm Statist Theory Methods 1977;6:15-32.

17. Tamhane AC. A comparison of procedures for multiple comparisons of

means with unequal variances. J Am Statist Assoc 1979;74:471-80.

18. Anusavice KJ, Cascone P. Dental casting and soldering alloys. In:

Anusavice KJ, editor. Philips’ science of dental materials. 11th ed. St.

Louis: Elsevier; 2003. p. 604.

19. Dieter GE. Mechanical metallurgy. 3rd ed. New York: McGraw-Hill; 1986.

p. 48, 375-431.

20. Bates JF. Studies related to the fracture of partial dentures; flexural fatigue

of a cobalt-chromium alloy. Br Dent J 1965;118:532-7.

21. Brantley WA. Wrought alloys. In: Anusavice KJ, editor. Philips’ science of

dental materials. 11th ed. St. Louis: Elsevier; 2003. p. 623.

22. Warr AJ. Numerical system of clasp design. J Prosthet Dent 1961;11:

1105-11.

23. Beer FP, Johnston ER, Dewolf JT. Mechanics of materials. 3rd ed. New

York: McGraw-Hill; 2002. p. 209-306.

24. Morris HF, Asgar K, Rowe AP, Nasjleti CE. The influence of heat treat-

ments on several types of base-metal removable partial denture alloys. J

Prosthet Dent 1979;41:388-95.

Reprint request to:

DR NORIYUKI WAKABAYASHI

REMOVABLE PROSTHODONTICS, DIVISION OF ORAL HEALTH SCIENCES

GRADUATE SCHOOL, TOKYO MEDICAL AND DENTAL UNIVERSITY

1-5-45 YUSHIMA, BUNKYO

TOKYO 113-8549

JAPAN

FAX: 181-3-5803-584

E-MAIL: wakabayashi.rpro@tmd.ac.jp

0022-3913/$30.00

Copyright � 2005 by The Editorial Council of The Journal of Prosthetic

Dentistry.

doi:10.1016/j.prosdent.2004.11.011

VOLUME 93 NUMBER 2