Canonical Biplot Statistical Analysis to Detect the Magnitude of the Effects of Phosphates Crystallization Aging on the Color in Siliceous Conglomerates A. C. In˜ igo, 1 * J. Garcı´a-Talego´ n, 2 S. Vicente-Tavera 3 1 Department of Enviromental Degradation Processes and Recovery, IRNASA-CSIC of Salamanca, c/ Cordel de Merinas, 40-52, 37008-Salamanca, Spain 2 Department of Geology, University of Salamanca, Salamanca, Spain 3 Department of Estadistic, University of Salamanca, Salamanca, Spain Received 20 December 2011; revised 18 May 2012; accepted 7 June 2012 Abstract: The Canonical Biplot method is used to deter- mine the magnitude of the effects on the D chromatic coor- dinates (DL*, Da*, Db*) and the parameter DE*, where (DL*, Da*, Db*) are the difference in the values of the sample after each aging cycle and the value of the untreated sample and DE* ¼ [(DL*) 2 þ (Da*) 2 þ (Db*) 2 ] 1/2 . We performed a study of the changes in color produced by two types of artificial aging procedures on four varieties of siliceous conglomerates from Zamora (Spain) that have traditionally been used in construction and later renovations in historical buildings in the zone. To accomplish this, 25 cycles of the following types of acceler- ated artificial aging were carried out: (a) freezing/thawing and cooling/heating (T1) and (b) combined freezing/thaw- ing and cooling/heating þ salt (phosphates) crystallization (T2).The results of the statistical study applied (Canonical Biplot) allowed us to differentiate the magnitude effect on the color of the surface brought about by T1 as compared with T2. These effects (p \ 0.05) were observed in all but one (ochre conglomerate) of the varieties, but mainly in the variable governing red hue (Da*) and yellow hue (Db*). Ó 2012 Wiley Periodicals, Inc. Col Res Appl, 39, 82 – 87, 2014; Published online 16 October 2012 in Wiley Online Library (wileyonlineli- brary.com). DOI 10.1002/col.21779 Key words: Canonical Biplot; color difference metrics; preservation; restoration; aesthetics INTRODUCTION The Canonical Biplot 1 is similar to Multivariate variance analysis (MANOVA). Gabriel 2 defined MANOVA-Biplot as a graphical method to study the variables responsible for the discrimination between groups, involving the spe- cific Biplot characteristics. 3–5 The Canonical Biplot extend the MANOVA-Biplot by adding confidence circles based on univariate Student t-tests to perform post hoc analysis of each variable. 6 Likewise, it gives us an approximation of the effect size between the groups to compare projected onto the factorial planes. On the prin- cipal factorial planes, we have drawn arrows that measure the changes (effects) observed in the increments studied in the T2 treatment as compared with T1. Because it is impossible to measure the process of dete- rioration in zones with different pathologies present on historical buildings, and since it is crucial to know how this evolves with time, accelerated aging techniques in climatic chambers under controlled conditions are used. Such evolution will depend on the intrinsic characteristics of the materials as well as on the external conditions affecting the stone, either inside (microclimates created inside buildings) or outside (climate, pollution, anthropic factors, etc), 7 which may to a large extent affect the struc- tural integrity of the buildings if prompt and correct measures are not taken as regards the causes of the deteri- oration occurring in the stone. *Correspondence to: A. C. In˜igo (e-mail: adolfo.inigo@irnasa.csic.es). V VC 2012 Wiley Periodicals, Inc. Contract grant sponsor: Ministry of Education and Science; Contract grant number: CGL2007-62168BET, Contract grant sponsor: FEDER, Contract grant sponsor: Ministry of Science and Innovation; Contract grant number: MAT2010-20660. 82 Volume 39, Number 1, February 2014