Characterisation of 19 almond cultivars on the basis of their free amino acids composition L. Helena Seron, a E. GarrigoÂs Poveda, b M. S. Prats Moya, b M. L. Martõ Â n CarratalaÂ, b V. Berenguer-Navarro b & N. GraneÂ-Teruel b * a Departamento de Quimica, Centro Ciencias Exatas e Tecnol., Univ. Federal de Sao Carlos., Caixa Postal 676, Sao Carlos, SP Brazil b Departamento de QuõÂmica AnalõÂtica, Universidad Alicante. P.O. Box 99, 03080 Alicante, Spain (Received 7 January 1997; accepted 4 April 1997) Several multivariate techniques (ANOVA, principal components analysis, cluster, linear discriminant function) were applied to the chromatographic data on free amino acids for a set of 19 almond cultivars. It is concluded that the cultivars studied have signi®cantly dierent free amino acid patterns. The information provided by multivariate techniques using amino acids is complementary to that obtained from fatty acid composition. Spanish cultivars can be successfully clas- si®ed as a single class dierent from the rest. Cultivars Texas, Genco and Wawona have such peculiar patterns that each of them can be considered as a consistent class. Cultivars Titan and Non Pareil associate either with the Spanish cultivars or with the American in function of amino acids processed. # 1998 Elsevier Science Ltd. All rights reserved INTRODUCTION Multivariate techniques have been proved to be an interesting set of powerful tools to analyse and recog- nise hidden patterns in complex matrices of experimen- tal data, and in this way to characterise similar objects. These techniques are of paramount importance, espe- cially in food chemistry, since they can be used to dis- tinguish or classify samples according to their geographical origin and varieties. Actually, multivariate techniques have been successfully applied to discrimi- nate alcoholic distillates (Ortiz et al., 1993), wines (Shi- moda et al., 1993; MartõÂn et al., 1996) coee (TomaÂs and Molins, 1990), tea (Varela et al., 1996), orange juice (Robards and Antolovich, 1995) and apple juice (Pilando and Wrolstad, 1992). These techniques are also recommended for selecting reference materials in food analysis (Pennickx et al., 1996). Amino acids are a rich fraction of many natural foods, and sometimes this has been exploited for dier- entiation purposes. In a previous paper (Prats and Berenguer, 1994), the ability of free amino acids to establish signi®cant dierences within a limited number of almond cultivars from the same region has been reported. The aim of the present paper is to verify the extent of this statement using a set of 19 selected and well-de®ned almond cultivars from very dierent origin, but cultivated in the same experimental ®eld and under the same conditions, in order to exclude accidental sources of dierentiation. MATERIALS AND METHODS Samples The set studied consisted of the following 19 cultivars: eight Spanish (Malaguen Äa MA, Peraleja PE, Atocha AT, Del Cid DC, Desmayo Largueta DL, Ramillete RA, Marcona MR, and a hybrid CE obtained at CEBAS (Centro de EdafologõÂa y BiologõÂa Aplicada del Segura, Murcia), four American (Texas TE, Non Pareil NP, Titan TI, Wawona WA), three Italian (Genco GE, Tuono TU, Cristomorto CR), one Australian (Chellaston CH), one Tunisian (Achaak AC), one from the Caucasian region (Primorskyi PR), and one French (Ferragnes FE). The trees corresponding to the abovementioned Spanish cultivars (generally very diversi®ed) were selec- ted and accepted as the most authentic representatives of these cultivars some years ago by a technical com- mittee (Felipe et al., 1984). Such selected cultivars have been replicated in other experimental research ®elds in Spain. All the samples for the present study are a mix- ture of almonds taken from dierent trees grown under the same conditions in CEBAS (Centro de EdafologõÂa y BiologõÂa Aplicada del Segura) experimental ®elds. All Food Chemistry, Vol. 61, No. 4, pp. 455±459, 1998 # 1998 Elsevier Science Ltd All rights reserved. Printed in Great Britain PII:S0308-8146(97)00083-6 0308-8146/98 $19.00+0.00 455 * To whom correspondence should be addressed.