investigation practice Data are collected from 35 sites where the fabric of clay soils is related to laboratory and field behaviour. The sites are chosen to range over the geological series from Recent to Ordovician time and to include applications to dams, road embankments, cuts, retaining walls, foundations and culverts. The cases illustrate the inadequacy of conventional site investigation procedure for major civil engineering projects with respect to the relevant properties of natural foundation strata.It is shown to be essential first to examine, describe and record the fabrics of consecutive soil samples and if sampling is relevant to use the knowledge of the overall geology, fabric details, and water levels in relation to the engineering problem to decide the location, quality, and size of specimens for element and model tests. The quality and size of specimens influence the drilling technique.Brief mention is made of consecutive and continuous sampling, of stereo photomacrographic technique, of equipment for sampling and testing 250 mm dia. specimens and of a centrifuge large enough to accept models representative of natural fabric.The proposals, which are reviewed in the light of earlier work of Hvorslev, Harding, Terzaghi, and Peck, are adaptable to the class of investigation appropriate to the problem, the geology and the available finance.Financial arrangements which avoid overall competitive tendering for site investigations are considered to be essential if improvements in the quality of site reports are to be achieved.On a rassemble les donnees relatives a 35 sites pour lesquels le comportement des sols au laboratoire et in situ est fonction de la nature de l'argile. Les sites ont Cte choisis de man&e a recouvrir les differentes dpoques geologiques allant de l'epoque recente a I'Ordovicien et de maniere a comprendre les applications aux barrages, aux talus de routes, aux murs de souttnement, aux fondations et aux canaux.Ces exemples demontrent l'inadaptation des methodes classiques d'dtudes des sites pour les projets importants de genie civil en ce qui concerne les proprittes des couches de fondations naturelles. On montre qu'il est essentiel de commencer par examiner, decrire et noter la texture des echantillons isoles et si ceux-ci sont representatifs il faut alors utiliser les don&es fournies par la geologic g&&ale, les details de la texture? les niveaux d'eau, en fonction des problemes de gtme civil poses pour decider du choix de l'emplacement, de la qualite et de la taille des carottes pour les essais sur modeles ou en laboratoire. La qualite et la taille des carottes influencent la methode de forage a utiliser. On cite rapidement le carottage continu ou discontinu, les techniques de stereo photomacrographie, le materiel de prise d'echantillons et d'essais pour carottes de 250 mm de diamttre et une centrifugeuse suffisamment grande pour recevoir les modeles representatifs d'une structure naturelle.ies suggestions faites, examinees en fonction des travaux anterieurs de Hvorslev, Harding, Terzag...
