Two new chemical analyses, combined with optical data, confirm the conclusion of J. Wyart (1933) that the unit-cell formula of scolecite is Ca₈Al₁₆Si₂₄O₈₀.24H₂O. The limit of natural substitution of Ca by Na₂ appears to be about 1 atom Na per unit cell, and by K₂ about 0·1 atom. The crystal elements as deduced from X-ray and goniometric measurements agree excellently. The cell size is a 18·48, b 18·95, c 6·54 Å., β 89° 21′, and the space-group S⁴. A detailed optical study of one of the analysed specimens was made. Base-exchange experiments were made, introducing Na·, K·, Li·, NH₄·, Ag·, and Tl· in place of Ca··, and the effect on the optical properties studied; in several cases the products were natrolites, confirming the view that natrolite and scolecite are isostructural. Vapour-pressure study showed that on partial dehydration a transition to metascolecite occurs at a degree of hydration dependent on the temperature of the experiment; except at low temperatures, the transition occurs considerably before the composition CaAl₂Si₃O₁₀.2H₂O, commonly attributed to metascolecite, is reached. In both scolecite and metascolecite the water consists of two distinct groups, in scolecite a more volatile one of sixteen and a less volatile one of eight mols. per unit cell; in metascolecite the group of eight is the more volatile. The vapour-pressure constants of both groups have been ascertained for metascolecite, and of the first for scolecite. There is no evidence that anhydrous metascolecite constitutes a third crystal phase. The vapour-pressure data tend to support the crystal-structure suggested for scolecite by W. H. Taylor, C. A. Meek, and W. W. Jackson, rather than that proposed by L. Pauling. The material described by H. Michel as natural metascolecite was incorrectly identified.