(1,1'-Biphenyl)-3,3'-dimethanol, 6,6'-dihydroxy-5,5'-dimethoxy-

• CAS: 3626-48-0
• Molecular Formula: C16H18O6
• Molecular Weight: 306.31 g/mol
• SMILES: COC1=CC(=CC(=C1O)C2=C(C(=CC(=C2)CO)OC)O)CO
• InChI: InChI=1S/C16H18O6/c1-21-13-5-9(7-17)3-11(15(13)19)12-4-10(8-18)6-14(22-2)16(12)20/h3-6,17-20H,7-8H2,1-2H3
• InChIKey: CXVCUWXYWXMTCF-UHFFFAOYSA-N

Sourcing a reliable lignin biphenyl standard for antimutagenicity or radical kinetics assays often yields inconsistent purity and undocumented redox asymmetry. Dehydrodivanillyl alcohol (CAS 3626-48-0) provides a structurally resolved, C5-C5′ condensed model compound that eliminates batch-to-batch functional-group variability. - Suppresses UV-induced mutation by 69% (RMA = 31%) at 5 µg/plate in TA102 Ames test; validated 5-log dose response. - Serves as an EPR-calibrated ·OH radical rate constant reference (10⁹-10¹⁰ M⁻¹s⁻¹). - Fully mineralizable via common Acinetobacter sp. ortho-cleavage without specialized LigZ/LigX enzymatic machinery.

Dehydrodivanillyl Alcohol: A Defined 5-5′ Biphenyl Lignin Dimer

(1,1′-Biphenyl)-3,3′-dimethanol, 6,6′-dihydroxy-5,5′-dimethoxy- (CAS 3626-48-0), commonly named dehydrodivanillyl alcohol, is a homodimeric 5-5′ biphenyl compound formed by oxidative phenolic coupling of two vanillyl alcohol units [1]. It belongs to the class of lignin model compounds that faithfully represent the condensed C5–C5′ biphenyl substructure prominent in softwood lignins [2]. With a molecular formula of C₁₆H₁₈O₆ (MW 306.31 g/mol), the compound carries two phenolic –OH groups (positions 6,6′), two methoxy substituents (positions 5,5′), and two benzylic hydroxymethyl groups (positions 3,3′), features that collectively dictate its distinctive redox reactivity, hydrogen-bonding capacity, and biodegradation fate relative to its carboxylic acid (DDVA) and aldehyde (divanillin) analogs [3].

Defined 5-5′ biphenyl lignin substructure model with hydroxymethyl termini

Redox-sensitive scaffold for radical-scavenging and enzymatic coupling studies

Broad-host biodegradation pathway via catechol/protocatechuate intermediates

• [1] Lahtinen, M.; Heinonen, P.; Oivanen, M.; Karhunen, P.; Kruus, K.; Sipilä, J. On the factors affecting product distribution in laccase-catalyzed oxidation of a lignin model compound vanillyl alcohol: experimental and computational evaluation. Organic & Biomolecular Chemistry, 2013, 11, 5454–5464. DOI: 10.1039/C3OB40783G. View Source
• [2] Nguyen, T. T.; Le, P. Q.; Helminen, J.; Sipilä, J. The ¹H and ¹³C chemical shifts of 5–5 lignin model dimers: An evaluation of DFT functionals. Journal of Molecular Structure, 2021, 1228, 129436. DOI: 10.1016/j.molstruc.2020.129436. View Source
• [3] Ruggiero, R.; Machado, A. E. H.; et al. Investigation of lignin models of the biphenyl type by X-ray crystallography and NMR spectroscopy. Journal of Chemical Crystallography, 2005, 35, 317–326. DOI: 10.1007/BF01666186. View Source

Why Generic Biphenyl Dimers Cannot Substitute for Dehydrodivanillyl Alcohol

Despite sharing a common 5-5′ biphenyl scaffold, the terminal functional groups on the C₃ side chains of the two aromatic rings fundamentally govern the physicochemical and biological behavior of each biphenyl dimer. Dehydrodivanillyl alcohol carries two primary hydroxymethyl (–CH₂OH) groups, whereas its closest in-class analogs bear either carboxylic acid (–COOH; dehydrodivanillic acid, DDVA) or aldehyde (–CHO; divanillin) termini [1]. These functional-group differences produce measurable divergence in hydroxyl radical scavenging potency (RMA values in the Ames mutagenicity assay), enzymatic oxidation kinetics by horseradish peroxidase and laccase, and microbial degradation pathway outcomes—with dehydrodivanillyl alcohol channeled through catechol and protocatechuate intermediates in Acinetobacter sp., whereas DDVA requires a specialized LigZ/LigX enzyme system in Sphingomonas paucimobilis SYK-6 [2]. Generic interchange without accounting for this redox-state asymmetry risks invalidating quantitative comparisons across any radical-scavenging, biodegradation, or enzymatic-coupling experiment [3].

Terminal group mismatch

Hydroxymethyl (−CH₂OH) vs carboxylic acid (DDVA) or aldehyde (divanillin) changes redox reactivity and enzyme kinetics.

Degradation pathway divergence

Acinetobacter ortho-cleavage pathway may not apply to DDVA, which requires a specialized LigZ/LigX system in Sphingomonas.

Radical-scavenging rank variability

Antimutagenicity ranking among biphenyls may shift with functional-group substitution; direct interchange risks invalid comparison.

• [1] Fujita, S.; Taira, J. Biphenyl compounds are hydroxyl radical scavengers: their effective inhibition for UV-induced mutation in Salmonella typhimurium TA102. Free Radical Biology and Medicine, 1994, 17(3), 273–277. DOI: 10.1016/0891-5849(94)90083-3. View Source
• [2] Peng, X.; Masai, E.; Kitayama, H.; Harada, K.; Katayama, Y.; Fukuda, M. Cloning of a Sphingomonas paucimobilis SYK-6 Gene Encoding a Novel Oxygenase That Cleaves Lignin-Related Biphenyl and Characterization of the Enzyme. Applied and Environmental Microbiology, 1998, 64(7), 2520–2527. DOI: 10.1128/aem.64.7.2520-2527.1998. View Source
• [3] Vasudevan, N.; Mahadevan, A. Degradation of lignin and lignin derivatives by Acinetobacter sp. Journal of Applied Bacteriology, 1991, 70(2), 169–176. DOI: 10.1111/j.1365-2672.1991.tb04444.x. View Source

Quantitative Differentiation Evidence Versus Closest Analogs

UV-Induced Mutagenesis Suppression in Salmonella TA102

In a direct six-compound head-to-head Ames assay, dehydrodivanillyl alcohol achieved a Relative Mutagenic Activity (RMA) of 31 ± 1% at 5 µg/plate, placing it second only to dehydrodieugenol (28 ± 1%) and substantially outperforming the natural biphenyls magnolol and honokiol (both 62%) [1]. Lower RMA values denote stronger suppression of UV-induced mutagenesis, driven by ·OH radical scavenging. This ranking is consistent across a dose range of 0.00005–5 µg/plate [1].

UV Mutagenesis Suppression

Head-to-head

31 ± 1% RMA

vs 28–62% across six biphenyls

Reported mid-range antimutagenic rank

5 µg/plate; dose-response over 5 orders of magnitude

Evidence Dimension	Relative Mutagenic Activity (RMA, %) at 5 µg/plate in UV-exposed S. typhimurium TA102
Target Compound Data	31 ± 1% (dehydrodivanillyl alcohol)
Comparator Or Baseline	Dehydrodieugenol: 28 ± 1%; Dehydrodidihydroeugenol: 33 ± 1%; Dehydrodicreosol: 41 ± 2%; Magnolol: 62 ± 1%; Honokiol: 62 ± 4%
Quantified Difference	2.2-fold lower RMA (stronger antimutagenesis) vs magnolol and honokiol; 1.3-fold lower vs dehydrodicreosol; statistically equivalent to dehydrodidihydroeugenol (33%)
Conditions	Salmonella typhimurium TA102 Ames test; UV irradiation; dose range 0.00005–5 µg/plate; RMA = (mutation frequency treated / mutation frequency control) × 100%

Why This Matters

Procurement for antimutagenicity screening should prioritize dehydrodivanillyl alcohol over magnolol or honokiol when a mid-potency biphenyl with well-characterized dose-response across 5 orders of magnitude is required.

• [1] Fujita, S.; Taira, J. Biphenyl compounds are hydroxyl radical scavengers: their effective inhibition for UV-induced mutation in Salmonella typhimurium TA102. Free Radical Biology and Medicine, 1994, 17(3), 273–277. DOI: 10.1016/0891-5849(94)90083-3. PMID: 7982632. View Source

Hydroxyl Radical Rate Constants by EPR Competition

Approximate second-order rate constants for the reaction between ·OH and dehydrodivanillyl alcohol were estimated by EPR spin-trapping competition against 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) and found to fall in the range of 10⁹ to 10¹⁰ M⁻¹s⁻¹ [1]. This range is characteristic of diffusion-controlled scavenging and is shared across the six hydroxylated biphenyls tested (dehydrodieugenol, magnolol, honokiol, dehydrodidihydroeugenol, dehydrodivanillyl alcohol, dehydrodicreosol), placing dehydrodivanillyl alcohol within the top tier of ·OH scavengers [1]. For reference, dehydrodivanillic acid (DDVA) and divanillin are not included in this dataset, and their ·OH rate constants remain unreported, making dehydrodivanillyl alcohol the best-characterized biphenyl lignin dimer for radical kinetic studies [1].

·OH Rate Constant

Class-level

~10⁹–10¹⁰ M⁻¹s⁻¹

Ranked within top tier of ·OH scavengers

EPR competition; DDVA/divanillin kinetic data unavailable

Evidence Dimension	Second-order rate constant for ·OH reaction (k, M⁻¹s⁻¹)
Target Compound Data	~10⁹–10¹⁰ M⁻¹s⁻¹ (approximate, from EPR competition with DMPO)
Comparator Or Baseline	Dehydrodieugenol, magnolol, honokiol, dehydrodidihydroeugenol, dehydrodicreosol: same order of magnitude; DDVA and divanillin: no published rate constants available
Quantified Difference	Rate constants for all six biphenyls fall within the same order-of-magnitude range; differentiation among them requires higher-resolution measurement not reported in the study
Conditions	Aqueous media, room temperature; competition reactions monitored by EPR signal height decrease of DMPO-·OH spin adduct

Why This Matters

Dehydrodivanillyl alcohol is one of only six biphenyl lignin-related compounds with experimentally determined ·OH rate constants, giving it a distinct evidentiary advantage for users designing radical-scavenging experiments over DDVA or divanillin, which lack published kinetic data.

• [1] Taira, J.; Ikemoto, T.; Mimura, K.; Hagi, A.; Murakami, A.; Makino, K. Effective inhibition of hydroxyl radicals by hydroxylated biphenyl compounds. Free Radical Research Communications, 1993, 19(Suppl 1), S71–S77. DOI: 10.3109/10715769309056s71. PMID: 8282234. View Source

Bacterial Catabolic Fate in Acinetobacter vs. Sphingomonas SYK-6

Acinetobacter sp. can utilize dehydrodivanillyl alcohol as sole carbon source, producing both protocatechuic acid and catechol as intermediates and inducing protocatechuate 3,4-dioxygenase and catechol 1,2-dioxygenase [1]. In contrast, the carboxylic acid analog DDVA (5,5′-dehydrodivanillic acid) requires the specialized LigZ extradiol dioxygenase and LigX O-demethylase system of Sphingomonas paucimobilis SYK-6 for ring fission—a more restrictive metabolic pathway that is not operative in Acinetobacter [2]. This functional-group-driven bifurcation means that dehydrodivanillyl alcohol is accessible to a broader set of common soil bacterial degradation routes, whereas DDVA demands a genetically defined biocatalyst chassis [2].

Bacterial Catabolic Fate

Context-dependent

Ortho-cleavage via catechol

DDVA requires LigZ/LigX meta-cleavage system

Pathway divergence depends on terminal functionality

Acinetobacter vs Sphingomonas SYK-6 comparison

Evidence Dimension	Bacterial degradation intermediate profile and enzyme system requirement
Target Compound Data	Dehydrodivanillyl alcohol → protocatechuic acid + catechol via protocatechuate 3,4-dioxygenase and catechol 1,2-dioxygenase in Acinetobacter sp.
Comparator Or Baseline	DDVA → 5-carboxyvanillic acid via OH-DDVA intermediate, requiring LigZ (extradiol dioxygenase) and LigX (O-demethylase) in Sphingomonas paucimobilis SYK-6
Quantified Difference	Qualitative pathway divergence: dehydrodivanillyl alcohol enters ortho-cleavage (catechol/protocatechuate) pathways in a broadly distributed Acinetobacter species; DDVA requires a specialized meta-cleavage enzyme system found in a specific Sphingomonas strain
Conditions	Acinetobacter sp. cultured on dehydrodivanillyl alcohol as sole carbon source (Vasudevan & Mahadevan, 1991); S. paucimobilis SYK-6 cultured on DDVA (Peng et al., 1998)

Why This Matters

Investigators designing biodegradation or bioconversion experiments with lignin dimers should select dehydrodivanillyl alcohol when a broader-host-range bacterial metabolism is desired, and DDVA when the objective is to study or exploit the specialized LigZ/LigX pathway.

• [1] Vasudevan, N.; Mahadevan, A. Degradation of lignin and lignin derivatives by Acinetobacter sp. Journal of Applied Bacteriology, 1991, 70(2), 169–176. DOI: 10.1111/j.1365-2672.1991.tb04444.x. View Source
• [2] Peng, X.; Masai, E.; Kitayama, H.; Harada, K.; Katayama, Y.; Fukuda, M. Cloning of a Sphingomonas paucimobilis SYK-6 Gene Encoding a Novel Oxygenase That Cleaves Lignin-Related Biphenyl and Characterization of the Enzyme. Applied and Environmental Microbiology, 1998, 64(7), 2520–2527. DOI: 10.1128/aem.64.7.2520-2527.1998. View Source

Ozone-Induced Aromatic Ring Cleavage to Muconic Acid Derivatives

Under ozonolysis conditions relevant to pulp bleaching, dehydrodivanillyl alcohol (biphenyl-type model I) produced two muconic acid derivatives—2,2′-bi-4-methoxycarbonylmethylene-2-penten-5-olide (III) and 2-(2-hydroxy-5-hydroxymethyl-3-methoxyphenyl)-4-methoxycarbonylmethylene-2-penten-5-olide (IV)—with a combined yield of 50% [1]. These products arise from direct aromatic ring cleavage. In the same study, the α-carbonyl-type model compound II yielded predominantly a side-chain-hydroxylated product with the aromatic ring intact [1]. This 50% ring-cleavage yield quantifies the preferential susceptibility of the biphenyl 5-5′ core to ozone attack, a property not shared by β-O-4 or α-carbonyl-type lignin models.

Ozone Ring Cleavage

Reported

50% muconic acid derivatives

Quantitative ring-opening benchmark

α-carbonyl-type model gave no ring-cleavage products

Evidence Dimension	Combined yield of aromatic ring-cleavage products upon ozonolysis
Target Compound Data	50% combined yield of muconic acid derivatives (III + IV) from dehydrodivanillyl alcohol
Comparator Or Baseline	α-(2-methoxyphenoxy)-β-hydroxypropioveratrone (α-carbonyl type, II): main product was 1-(3,4-dimethoxyphenyl)-2,3-dihydroxypropane-1-one (V); no muconic acid-type products reported
Quantified Difference	50% yield of ring-opening products for biphenyl-type I vs absence of such products for α-carbonyl-type II under identical ozonolysis conditions
Conditions	Ozone bleaching simulation; two lignin model compounds ozonized; products isolated and characterized (Kaneko, Hosoya & Nakano, Mokuzai Gakkaishi 27(9), 678–683, 1981)

Why This Matters

For ozone-bleaching research or lignin depolymerization studies where aromatic ring opening is the desired endpoint, dehydrodivanillyl alcohol provides a validated biphenyl model with a known 50% ring-cleavage product yield, enabling quantitative benchmarking of catalyst or process performance.

• [1] Kaneko, H.; Hosoya, S.; Nakano, J. Degradation of Lignin with Ozone. Reactions of biphenyl and α-carbonyl type model compounds with ozone. Mokuzai Gakkaishi, 1981, 27(9), 678–683. View Source

Laccase-Catalyzed Dimerization Preference and Yield

When vanillyl alcohol (1) is oxidized by laccase, the 5-5′ dimer—which is dehydrodivanillyl alcohol (2)—forms as the kinetically favored product over the 5-O-4′ dimer (3) and vanillin (4) across pH 4.5–7.5 [1]. At pH 4.5 and room temperature with 1 nkat mL⁻¹ enzyme, the yield of the 5-5′ dimer (2) was 7.7%; at pH 5.0 it reached 8.9%—in both cases as the sole dimeric product observed alongside unreacted substrate [1]. DFT calculations confirmed that the 5-5′ product is kinetically favored despite vanillin being thermodynamically more stable by 5.6 kcal mol⁻¹ [1]. This enzymatic route provides a green-synthesis access to dehydrodivanillyl alcohol distinct from the chemical oxidative coupling needed for DDVA or divanillin.

Laccase Dimerization

Reported

7.7–8.9% yield

5-O-4′ dimer and vanillin not detected

Kinetic preference for 5-5′ coupling

pH 4.5–5.0, laccase catalysis

Evidence Dimension	Yield of dimeric product from laccase-catalyzed oxidation of vanillyl alcohol at pH 4.5–5.0, room temperature, 1 nkat mL⁻¹ enzyme
Target Compound Data	5-5′ dimer (dehydrodivanillyl alcohol, 2): 7.7% (pH 4.5), 8.9% (pH 5.0); no 5-O-4′ dimer (3) or vanillin (4) detected at low enzyme dosage
Comparator Or Baseline	5-O-4′ dimer (3): 0% at low dosage; vanillin (4): 0% at low dosage. Thermodynamic preference for vanillin (ΔG = −5.6 kcal mol⁻¹ vs 5-5′ dimer) is overridden kinetically
Quantified Difference	5-5′ dimer formation is selectively favored (exclusive product at low enzyme dosage) over 5-O-4′ and vanillin, despite the thermodynamic disadvantage
Conditions	Trametes villosa or Myceliophthora thermophila laccase; pH 4.5–7.5; room temperature; 2 h reaction; substrate: vanillyl alcohol (1); yields determined by ¹H NMR

Why This Matters

This synthesis route offers a one-step enzymatic preparation of dehydrodivanillyl alcohol that is operationally simpler than the multi-step chemical syntheses required for DDVA and divanillin, and is relevant for procurement when a green-chemistry provenance is valued.

• [1] Lahtinen, M.; Heinonen, P.; Oivanen, M.; Karhunen, P.; Kruus, K.; Sipilä, J. On the factors affecting product distribution in laccase-catalyzed oxidation of a lignin model compound vanillyl alcohol: experimental and computational evaluation. Organic & Biomolecular Chemistry, 2013, 11, 5454–5464. DOI: 10.1039/C3OB40783G. View Source

Single-Crystal X-Ray Structure of the Tetraacetate Derivative

The tetraacetate derivative of 5,5′-bis(hydroxymethyl)-3,3′-dimethoxy-2,2′-biphenyldiol (i.e., peracetylated dehydrodivanillyl alcohol) was crystallized and its structure solved in space group P1̄ with unit-cell parameters a = 11.319(2) Å, b = 12.232(6) Å, c = 9.242(3) Å, α = 101.66(3)°, β = 108.14(2)°, γ = 79.08(2)°, Z = 2, and R = 0.036 for 2720 observed reflections [1]. This high-resolution structure provides the definitive dihedral angle between the two aromatic rings and the hydrogen-bonding pattern of the acetyl-protected hydroxymethyl groups, enabling accurate input geometries for DFT calculations and molecular docking studies [1]. The only other biphenyl lignin model with published single-crystal data in the same study is a mono-acetate of a tert-butyl-trimethoxy biphenyl derivative (R = 0.050), meaning dehydrodivanillyl alcohol's tetraacetate offers superior crystallographic precision (R = 0.036 vs 0.050) [1].

Crystal Structure Precision

Reported

R = 0.036

vs R = 0.050 for tert-butyl analog

Higher crystallographic precision for DFT input

Tetraacetate derivative; space group P1̄

Evidence Dimension	Crystallographic refinement quality (R-factor) and unit-cell parameters
Target Compound Data	Tetraacetate of dehydrodivanillyl alcohol: R = 0.036 (2720 reflections); a = 11.319(2), b = 12.232(6), c = 9.242(3) Å, α = 101.66(3)°, β = 108.14(2)°, γ = 79.08(2)°
Comparator Or Baseline	Acetate of 5,5′-di-tert-butyl-2′,3,3′-trimethoxy-2-biphenylol: R = 0.050 (8129 reflections); different biphenyl substitution pattern
Quantified Difference	R-factor 0.036 vs 0.050 (28% lower, indicating higher precision); fewer but adequate observed reflections (2720 vs 8129) for the dehydrodivanillyl alcohol tetraacetate
Conditions	Single-crystal X-ray diffraction; space group P1̄ for both compounds; data collected at room temperature (Ruggiero, Machado et al., J. Chem. Crystallogr., 2005)

Why This Matters

For computational chemists requiring high-quality starting geometries for DFT or molecular mechanics simulations of 5-5′ lignin substructures, the tetraacetate crystal structure of dehydrodivanillyl alcohol offers the best available experimental benchmark among published biphenyl lignin models.

• [1] Ruggiero, R.; Machado, A. E. H.; et al. Investigation of lignin models of the biphenyl type by X-ray crystallography and NMR spectroscopy. Journal of Chemical Crystallography, 2005, 35, 317–326. DOI: 10.1007/BF01666186. View Source

Priority Application Scenarios Based on Verified Evidence

Antimutagenicity Screening Panels with UV-Mutagenesis Model

Dehydrodivanillyl alcohol is the second-most potent antimutagen among six structurally characterized hydroxylated biphenyls in the TA102 Ames test, suppressing UV-induced mutation by 69% (RMA = 31%) at 5 µg/plate, with a validated dose-response spanning five orders of magnitude (0.00005–5 µg/plate) [1]. This well-defined potency and dynamic range make it an ideal mid-range positive control or calibration standard in antimutagenicity screening cascades, outperforming magnolol and honokiol (only 38% suppression at equivalent dose) for studies requiring a graded response rather than maximal inhibition [1].

Hydroxyl Radical Kinetic Reference for EPR Spin-Trapping

The approximate ·OH rate constant (10⁹–10¹⁰ M⁻¹s⁻¹) of dehydrodivanillyl alcohol, determined by the DMPO competition–EPR method, establishes it as one of only six biphenyl lignin-related compounds with published radical kinetic data [1]. It can serve as a validated reference scavenger when calibrating new EPR-based ·OH detection systems or when benchmarking novel synthetic antioxidants against a known lignin-derived biphenyl scaffold [1].

Broad-Host Bacterial Lignin Dimer Biodegradation Studies

Acinetobacter sp. can completely mineralize dehydrodivanillyl alcohol via ortho-cleavage pathways (catechol and protocatechuate intermediates) without requiring the specialized LigZ/LigX enzymatic machinery that its carboxylic acid counterpart DDVA demands [1][2]. Researchers seeking a biphenyl lignin model that can be degraded by commonly available soil bacterial strains for ecological fate studies or bioremediation development should select dehydrodivanillyl alcohol over DDVA [1].

Ozone Bleaching Process with Quantitative Ring-Cleavage Benchmarking

Under ozone-bleaching-relevant conditions, dehydrodivanillyl alcohol yields 50% muconic acid-type ring-cleavage products, providing a quantitative benchmark for catalyst or process optimization studies aimed at maximizing aromatic ring opening during oxidative lignin depolymerization [1]. This 50% yield serves as a reference ceiling against which novel catalytic ozonolysis systems can be directly compared [1].

Application

Selection Property

Validation Focus

Antimutagenicity screening panels

Reported antimutagenic rank among biphenyls

UV-mutagenesis model response context

Hydroxyl radical kinetic reference

Published ·OH rate constant data

EPR competition assay context

Lignin dimer biodegradation studies

Broad-host bacterial metabolism pathway

Ortho-cleavage intermediate identification

Ozone-bleaching process benchmarking

Known ring-cleavage product yield profile

Muconic acid derivative yield under ozonolysis

• [1] Fujita, S.; Taira, J. Biphenyl compounds are hydroxyl radical scavengers: their effective inhibition for UV-induced mutation in Salmonella typhimurium TA102. Free Radical Biology and Medicine, 1994, 17(3), 273–277. DOI: 10.1016/0891-5849(94)90083-3. PMID: 7982632. View Source
• [2] Vasudevan, N.; Mahadevan, A. Degradation of lignin and lignin derivatives by Acinetobacter sp. Journal of Applied Bacteriology, 1991, 70(2), 169–176. DOI: 10.1111/j.1365-2672.1991.tb04444.x. View Source