When I thus determined the primitive crystal and integrand molecule of bardiglione, the Tableau Compartif des Résultats de la Crystallographie et de l'Analyse chimique of M. l'Abbé Haüy had not appeared. This work informs me, that its learned author had also determined the form of the primitive crystal, as well as of its integrand molecule: and at the same time I perceive, with infinite regret, that; as he was not more fortunate than myself with respect to the height of these two molecules, the other dimensions he assigned them differ from those to which I had been led by my study of this substance. According to this distinguished mineralogist, the base of the rectangular tetrahedral prism is not a square, but a rectangle, the sides of which are to each other as 16 to 13.4 (fig. 5) ; the prism is divisible in a direction parallel to its diagonals by planes, the intersection of which forms angles of 100° 6′ and 79° 54′ (fig. 6). Thus it would be divisible into rectangular rhomboidal prisms, the bases of which would be rhombs of 100° 6′ and 79° 54′. But these prisms, it appears to me, cannot in any way be considered as the integrand molecules of the tetrahedral prism with rectangular bases, considered as the primitive one; for this prism being also divisible in a direction parallel to each of its sides, as shown at fig. 7, this second division would separate each of the prisms with rhomboidal bases of 100° 6 and 79° 54′ into four rectangular trihedral prisms, of which two opposite ones would have for their base an isosceles triangle of 100° 6′, and the other two an isosceles triangle also of 79° 54′, fig. 5. Thus these molecules would be of two different forms, and consequently could no more be considered as the integrand molecules of the substance, than those with rhomboidal bases of 100° 6′ and 79° 54′.
On the other hand, this substance presents a very particular and interesting fact, likely to mislead, with respect to the primitive
