Note on analysis of quartz grain dimensions in foliated greywackes

Quartz grain dimensions, measured parallel to an internal reference system (reference aspect ratio, RAR), were analyzed using the ,,reduced major axis, (RMA) linear regression analysis. In contrast to least-squares analysis, this statistical technique does not distinguish between dependent and independent variables. Application of the RMA analysis in conjunction with RAR values can be most useful for comparisons between foliated greywackes in which different grain scale processes, such as pressure solution and rigid-body rotation, were active. The RAR/RMA analysis reflects grain alignment and is therefore also useful for cleavage classification. In addition to the above, the application of RAR/RMA analysis for the determination of strain is discussed. Robin strains are compared with the arithmetic and harmonic means from RAR's and it is concluded that the arithmetic mean RAR produces a reasonable estimate of longitudinal strain in these rocks. *) Author's address: Dr,B. A. VAN DER PLUIJM, Department of Geological Sciences, University of Michigan, 1006 C. C. Little Building, Ann Arbor, MI 48109, USA. R~sum~ Les dimensions des grains de quartz, mesur~s parall~lement ~t un syst~me de r~f&ence interne (,>reference aspect rations: RAR) ont ~t~ trait~es par la m&hode de la rSgression lin~aire (,>reduced major axis<<: RMA). Contrairement 5 la m&hode des moindres carr~s, cette technique statistique ne fait pas de distinction entre les variables d@endantes et ind@endantes. Uapplication de ce type de m&hode s'av~re tr~s utile 5 la comparaison de grauwaekes schisteuses dans lesquelles les dimensions des grains peuvent &re la consequence de processus diff&ents, tels que la dissolution (pressure solution) et la rotation. L'analyse RAR/RMA traduit l'alignement des grains et peut, de ce fait, 8tre utilis~e aussi 8 l'appr~ciation du type de schistositC D'autre part, l'auteur discute l'application de l'analyse RAR/RMA ~ la d&ermination de la d~formation finie. Si on compare les d~formations ~>Robin<< aux moyennes arithm& tique et harmonique des RAR, on peut conclure que ies moyennes RAR arithm&iques fournissent une estimation raisonnable de la d~formation longimdinale.